HANDBOOK OF EMOTIONS

HANDBOOK OFEMOTIONS

Third Edition

EDITED BY MICHAEL LEWIS,JEANNETTE M. HAVILAND-JONES,

AND LISA FELDMAN BARRETT

THE GUILFORD PRESSNew York London

© 2008 The Guilford PressA Division of Guilford Publications, Inc.72 Spring Street, New York, NY 10012www.guilford.com

All rights reserved

No part of this book may be reproduced, translated, stored in a retrieval system,or transmitted, in any form or by any means, electronic, mechanical, photocopying,microfilming, recording, or otherwise, without written permission from the Publisher.

Printed in the United States of America

This book is printed on acid-free paper.

Last digit is print number: 9 8 7 6 5 4 3 2 1

Library of Congress Cataloging-in-Publication Data

Handbook of emotions / edited by Michael Lewis, Jeannette M. Haviland-Jones,Lisa Feldman Barrett. — 3rd ed.

p. cm.Includes bibliographical references and index.ISBN 978-1-59385-650-2 (cloth : alk. paper)1. Emotions. 2. Emotions—Sociological aspects. I. Lewis, Michael, 1937 Jan. 10–

II. Haviland-Jones, Jeannette M. III. Barrett, Lisa Feldman.BF561.H35 2008152.4—dc22

2007042332

About the Editors

Michael Lewis, PhD, is the University Distinguished Professor of Pediatrics and Psychiatryand Director of the Institute for the Study of Child Development at Robert Wood JohnsonMedical School, University of Medicine and Dentistry of New Jersey. Dr. Lewis has writtenor edited more than 30 books on developmental psychology, and he was rated as number onein scholarly impact in the field of developmental psychology in a survey published in Devel-opmental Review in 1995. He is coeditor of The Handbook of Developmental Psychopathol-ogy. His book Altering Fate: Why the Past Does Not Predict the Future (1998) was a finalistfor the Maccoby Book Award. He recently coedited a new text, Introduction to Infant Devel-opment, Second Edition.

Jeannette M. Haviland-Jones, PhD, is Professor of Psychology and Director of the HumanEmotions Lab at Rutgers, The State University of New Jersey. She has written extensivelyabout emotional development for over 25 years and coauthored The Hidden Genius of Emo-tion with Carol Magai. Her long-time research interests have included lifespan changes indisplays of emotion, gender differences in emotions, and the organizing effects of emotion oncognition and on personality. Recently she has initiated research on the chemosensory aspectsof emotion communication, including studies on the emotional environment.

Lisa Feldman Barrett, PhD, is Professor of Psychology and Director of the InterdisciplinaryAffective Science Laboratory at Boston College, with appointments at Harvard MedicalSchool and Massachusetts General Hospital. Her major research focus is on the nature ofemotion from social-psychological, psychophysiological, cognitive science, and neuroscienceperspectives. Dr. Barrett has published over 80 papers and book chapters. She has edited,with Peter Salovey, The Wisdom in Feeling: Psychological Processes in Emotional Intelli-gence, and, with Paula Niedenthal and Piotr Winkielman, Emotion and Consciousness. She isthe recipient of the Career Trajectory Award from the Society for Experimental Social Psy-chology; the James McKeen Cattell Award from the James McKeen Cattell Fund and the As-sociation for Psychological Science; and a fellowship from the American Philosophical Soci-ety. She is currently Co-Editor in Chief of the journal Emotion Review.

v

Contributors

Jo-Anne Bachorowski, PhD, Department of Psychology, Vanderbilt University, Nashville,Tennessee

Lisa Feldman Barrett, PhD, Department of Psychology, Boston College, Chestnut Hill,Massachusetts

John E. Bates, PhD, Department of Psychological and Brain Sciences, Indiana University,Bloomington, Indiana

Gary G. Berntson, PhD, Department of Psychology, Ohio State University, Columbus, Ohio

Eliza Bliss-Moreau, BA, Department of Psychology, Boston College, Chestnut Hill,Massachusetts

George A. Bonanno, PhD, Department of Counseling and Clinical Psychology, TeachersCollege, Columbia University, New York, New York

Leslie R. Brody, PhD, Department of Psychology, Boston University, Boston, Massachusetts

John T. Cacioppo, PhD, Department of Psychology and Center for Cognitive and SocialNeuroscience, University of Chicago, Chicago, Illinois

Linda A. Camras, PhD, Department of Psychology, DePaul University, Chicago, Illinois

Gerald L. Clore, PhD, Department of Psychology, University of Virginia, Charlottesville,Virginia

Michael A. Cohn, BA, Department of Psychology, University of Michigan, Ann Arbor,Michigan

Karin G. Coifman, MA, Department of Counseling and Clinical Psychology, Teachers College,Columbia University, New York, New York

Nathan S. Consedine, PhD, Department of Psychology and Intercultural Institute on HumanDevelopment and Aging, Long Island University, Brooklyn, New York

Leda Cosmides, PhD, Department of Psychology and Center for Evolutionary Psychology,University of California at Santa Barbara, Santa Barbara, California

A. D. (Bud) Craig, PhD, Atkinson Research Laboratory, Barrow Neurological Institute,Phoenix, Arizona

vi

Matthew Davidson, BA, Department of Psychology, Columbia University, New York,New York

Brian T. Detweiler-Bedell, PhD, Department of Psychology, Lewis and Clark College, Portland,Oregon

Jerusha B. Detweiler-Bedell, PhD, Department of Psychology, Lewis and Clark College,Portland, Oregon

Michael A. Diefenbach, PhD, Department of Urology and Department of OncologicalSciences, Mount Sinai School of Medicine, New York, New York

Ed Diener, PhD, Department of Psychology, University of Illinois at Urbana–Champaign,Champaign, Illinois

Kenneth A. Dodge, PhD, Department of Public Policy Studies, Duke University, Durham,North Carolina

Seth Duncan, MA, Department of Psychology, Boston College, Chestnut Hill, Massachusetts

Serah S. Fatani, MA, Department of Psychology, DePaul University, Chicago, Illinois

Agneta H. Fischer, PhD, Department of Social Psychology, University of Amsterdam,Amsterdam, The Netherlands

Jennifer E. Fite, MA, Department of Psychological and Brain Sciences, Indiana University,Bloomington, Indiana

Mark Frank, PhD, Department of Communication, University at Buffalo, State Universityof New York, Buffalo, New York

Barbara L. Fredrickson, PhD, Department of Psychology, University of North Carolina atChapel Hill, Chapel Hill, North Carolina

Nico H. Frijda, PhD, Department of Psychology, University of Amsterdam, Amsterdam,The Netherlands

Jackson A. Goodnight, BS, Department of Psychological and Brain Sciences, IndianaUniversity, Bloomington, Indiana

Laura Goorin, MA, Department of Counseling and Clinical Psychology, Teachers College,Columbia University, New York, New York

Leslie S. Greenberg, PhD, Department of Psychology, York University, Toronto, Ontario,Canada

James J. Gross, PhD, Department of Psychology, Stanford University, Stanford, California

Jonathan Haidt, PhD, Department of Psychology, University of Virginia, Charlottesville,Virginia

Judith A. Hall, PhD, Department of Psychology, Northeastern University, Boston,Massachusetts

Paul L. Harris, DPhil, Graduate School of Education, Harvard University, Cambridge,Massachusetts

Jeannette M. Haviland-Jones, PhD, Human Emotions Lab, Department of Psychology,Rutgers, The State University of New Jersey, New Brunswick, New Jersey

Marc W. Hernandez, PhD, Department of Psychology, University of Chicago, Chicago,Chicago Illinois

Martin L. Hoffman, PhD, Department of Psychology, New York University, New York,New York

Contributors vii

Randall Horton, PhD, Department of Comparative Human Development, University ofChicago, Chicago, Illinois

Alice M. Isen, PhD, Department of Psychology and Johnson Graduate School of Management,Cornell University, Ithaca, New York

Tiffany A. Ito, PhD, Department of Psychology, University of Colorado, Boulder, Colorado

P. N. Johnson-Laird, PhD, Department of Psychology, Princeton University, Princeton,New Jersey

Craig Joseph, PhD, Department of Comparative Human Development, University of Chicago,Chicago, Illinois; Department of Psychology, Northwestern University, Evanston, Illinois

Josh Joseph, BA, BS, Department of Psychology, Columbia University, New York, New York

Dacher Keltner, PhD, Department of Psychology, University of California at Berkeley,Berkeley, California

Margaret E. Kemeny, PhD, Department of Psychiatry, University of California at San Francisco,San Francisco, California

Elizabeth A. Kensinger, PhD, Department of Psychology, Boston College, Chestnut Hill,Massachusetts

Hedy Kober, BS, Department of Psychology, Columbia University, New York, New York

Ann M. Kring, PhD, Department of Psychology, University of California at Berkeley,Berkeley, California

Jeff T. Larsen, PhD, Department of Psychology, Texas Tech University, Lubbock, Texas

Joseph E. LeDoux, PhD, Center for Neural Science, New York University, New York,New York

Elizabeth A. Lemerise, PhD, Department of Psychology, Western Kentucky University,Bowling Green, Kentucky

Howard Leventhal, PhD, Institute for Health, Health Care Policy, and Aging Research,Department of Psychology, Rutgers-The State University of New Jersey, New Brunswick,New Jersey

Michael Lewis, PhD, Institute for the Study of Child Development, Department of Pediatrics,Robert Wood Johnson Medical School, University of Medicine and Dentistry of New Jersey,New Brunswick, New Jersey

Kristen A. Lindquist, BA, Department of Psychology, Boston College, Chestnut Hill,Massachusetts

George Loewenstein, PhD, Department of Social and Decision Sciences, Carnegie MellonUniversity, Pittsburgh, Pennsylvania

Richard E. Lucas, PhD, Department of Psychology, Michigan State University, East Lansing,Michigan

Diane M. Mackie, PhD, Department of Psychology, University of California at Santa Barbara,Santa Barbara, California

Carol Magai, PhD, Department of Psychology and Intercultural Institute on HumanDevelopment and Aging, Long Island University, Brooklyn, New York

Antony S. R. Manstead, DPhil, School of Psychology, Cardiff University, Cardiff, Wales,United Kingdom

David Matsumoto, PhD, Department of Psychology, San Francisco State University,San Francisco, California

viii Contributors

John D. Mayer, PhD, Department of Psychology, University of New Hampshire, Durham,New Hampshire

Clark R. McCauley, PhD, Department of Psychology, Bryn Mawr College, Bryn Mawr,Pennsylvania

Suzanne M. Miller, PhD, Division of Population Science, Psychosocial and BehavioralMedicine Program, Behavioral Research Core Facility, and Behavioral Center of Excellence inBreast Cancer, Fox Chase Cancer Center, Philadelphia, Pennsylvania

Marvin Minsky, PhD, Department of Electrical Engineering and Computer Science, and MITMedia Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts

Jennifer Mize, BA, Department of Psychology, Boston College, Chestnut Hill, Massachusetts

Paula M. Niedenthal, PhD, Laboratory for Social and Cognitive Psychology, National Centerfor Scientific Research and University of Clermont-Ferrand, Clermont-Ferrand, France

Keith Oatley, PhD, Cognitive Science Program, University College, University of Toronto,Toronto, Ontario, Canada

Arne Öhman, PhD, Psychology Section, Department of Clinical Neuroscience, KarolinskaInstitute, Stockholm, Sweden

Andrew Ortony, PhD, Department of Psychology, Northwestern University, Evanston, Illinois

Maureen O’Sullivan, PhD, Department of Psychology, University of San Francisco,San Francisco, California

Michael J. Owren, PhD, Department of Psychology and Center for Behavioral Neuroscience,Georgia State University, Atlanta, Georgia

Jaak Panksepp, PhD, Department of Veterinary and Comparative Anatomy, Pharmacology,and Physiology, College of Veterinary Medicine, Washington State University, Pullman,Washington

Ellen Peters, PhD, Decision Research, Eugene, Oregon

Elizabeth A. Phelps, PhD, Department of Psychology, New York University, New York,New York

Kirsten M. Poehlmann, PhD, Department of Neurosciences, University of California at SanDiego, San Diego, California

Matthew Porter, PhD, Department of Oncological Sciences, Mount Sinai School of Medicine,New York, New York

Scott Rick, PhD, Operations and Information Management Department, The Wharton School,University of Pennsylvania, Philadelphia, Pennsylvania

Paul Rozin, PhD, Department of Psychology, University of Pennsylvania, Philadelphia,Pennsylvania

James A. Russell, PhD, Department of Psychology, Boston College, Chestnut Hill,Massachusetts

Carolyn Saarni, PhD, Department of Counseling, Sonoma State University, Rohnert Park,California

Peter Salovey, PhD, Department of Psychology, Yale University, New Haven, Connecticut

Daniel L. Schacter, PhD, Department of Psychology, Harvard University, Cambridge,Massachusetts

Avgusta Shestyuk, PhD, Helen Wills Neuroscience Institute, University of Californiaat Berkeley, Berkeley, California

Contributors ix

Michelle N. Shiota, PhD, Department of Psychology, Arizona State University, Tempe, Arizona

Richard A. Shweder, PhD, Department of Comparative Human Development, University ofChicago, Chicago, Illinois

Eliot R. Smith, PhD, Department of Psychological and Brain Sciences, Indiana University,Bloomington, Indiana

Robert C. Solomon, PhD (deceased), Department of Philosophy, University of Texas at Austin,Austin, Texas

Peter N. Stearns, PhD, Department of History and Office of the Provost, George MasonUniversity, Fairfax, Virginia

Michael Stefanek, PhD, Behavioral Research Center, American Cancer Society, Atlanta,Georgia

Nancy L. Stein, PhD, Department of Psychology and Department of Comparative HumanDevelopment, University of Chicago, Chicago, Illinois

Jan E. Stets, PhD, Department of Sociology, University of California at Riverside,Riverside, California

John Tooby, PhD, Department of Anthropology and Center for Evolutionary Psychology,University of California at Santa Barbara, Santa Barbara, California

Tom Trabasso, PhD (deceased), Department of Psychology, University of Chicago,Chicago, Illinois

Jonathan H. Turner, PhD, Department of Sociology, University of California at Riverside,Riverside, California

Tor D. Wager, PhD, Department of Psychology, Columbia University, New York, New York

Arlene S. Walker-Andrews, PhD, Department of Psychology and Office of the Provost,University of Montana, Missoula, Montana

Sherri C. Widen, PhD, Department of Psychology, Boston College, Chestnut Hill,Massachusetts

Patricia J. Wilson, PhD, Department of Psychology, La Salle University, Philadelphia,Pennsylvania

x Contributors

Preface

When we published the first edition of the Handbook of Emotions in 1993, research in emo-tions was beginning to blossom. Since then, its study has been recognized as an essential as-pect of any study of humankind. At the beginning of the new century, the value of studyingemotions—as they interface with cognition, with personality, and with social and health is-sues—has become obvious to all. The second edition of the Handbook was an attempt tomeet this new challenge by allowing us to consider the work already accomplished and to be-gin exploring the new work that lies ahead.

The first edition was rated by Choice as an “Outstanding Academic Book” in 1995. Thethird edition continues this tradition, as it seeks to be both a compendium of what we havelearned and an introduction to new areas that have caught our imagination. For those whoare just beginning to take an interest in the field of emotion research, this volume should pro-pel their interest further, with new areas to investigate and old questions to be viewed in newways. For those of us who continue to work in the field and who once labored as isolatedscientists, this third edition demonstrates the strength and vitality of the field as new areasopen up and new investigators make notable contributions. Such a development reflects thelarge community of scholars who are now invested in research and conceptual questionsabout emotions. These research activities are only enhanced when such a gathering of schol-ars reaches the mass it has now attained.

This edition contains new chapters by authors who made different contributions to thefirst two editions, updates to chapters in the first and second editions that reflect the researchof the last 15 years, and chapters by new contributors that reflect the expanding understand-ing of how emotions interface with other aspects of behavior. Thus this edition includes workby pioneers of the field who are continuing to expand their original ideas, as well as efforts bynew investigators whose contributions to the study of emotion are well documented. In thisthird edition, we have maintained our interest in “Interdisciplinary Foundations,” adding re-vised chapters on several topics and a new chapter on economics to this first part of the book.Part II, “Biological and Neurophysiological Approaches,” now includes new chapters onneuroimaging and olfaction. Whereas in the second edition “Social and Personality Issues”was considered in a single part of the book, we have now divided this section into two sepa-rate parts, reflecting how emotions have increasingly become a part of the study of the socialworld as well as personality. Part III, “Developmental Changes,” now includes a new chapteron life course in emotional development. Both the “Cognitive Factors” and “Health and

xi

Emotions” parts have been expanded, including new chapters on artificial intelligence andthe role of emotions in the neuroendocrine and immune systems. The final part, “Select Emo-tions,” also now reflects the interest in positive emotions. Our efforts to look at emotionsfrom a broad as well as a more focused perspective allow the reader access not only to the roleof emotion in diverse fields, but also to more focused views of work in specific emotions.

Although this edition of the Handbook follows the same basic organization of the firstedition, there has been such an increase in the possible topics that could have been includedthat we greatly regret having to omit some areas. We have chosen some areas because we be-lieve that they will emerge as significant; others because they have already been enormouslyproductive; and still others simply to provide balance to the various perspectives. Bringingtogether the broadening content of the Handbook and working with this large, diverse groupof scholars were great challenges and privileges. It is to our contributors that the volumeowes its strength.

We continue to hope that scholars, clinicians, and students from a variety of disciplineswill find our efforts helpful. Finally, we would like to express our appreciation to StaceyNapoli, without whom this edition would not have been possible.

MICHAEL LEWISJEANNETTE M. HAVILAND-JONESLISA FELDMAN BARRETT

xii Preface

Contents

PART I. INTERDISCIPLINARY FOUNDATIONS

1. The Philosophy of Emotions 3ROBERT C. SOLOMON

2. History of Emotions: Issues of Change and Impact 17PETER N. STEARNS

3. The Sociology of Emotions 32JAN E. STETS and JONATHAN H. TURNER

4. The Affective Brain and Core Consciousness:How Does Neural Activity Generate Emotional Feelings?

47

JAAK PANKSEPP

5. The Psychologists’ Point of View 68NICO H. FRIJDA

6. The Clinical Application of Emotion in Psychotherapy 88LESLIE S. GREENBERG

7. Emotions, Music, and Literature 102P. N. JOHNSON-LAIRD and KEITH OATLEY

8. The Evolutionary Psychology of the Emotions and Their Relationshipto Internal Regulatory Variables

114

JOHN TOOBY and LEDA COSMIDES

9. The Role of Emotion in Economic Behavior 138SCOTT RICK and GEORGE LOEWENSTEIN

xiii

PART II. BIOLOGICAL ANDNEUROPHYSIOLOGICAL APPROACHES

10. Emotional Networks in the Brain 159JOSEPH E. LEDOUX and ELIZABETH A. PHELPS

11. The Psychophysiology of Emotion 180JEFF T. LARSEN, GARY G. BERNTSON, KIRSTEN M. POEHLMANN,TIFFANY A. ITO, and JOHN T. CACIOPPO

12. Vocal Expressions of Emotion 196JO-ANNE BACHOROWSKI and MICHAEL J. OWREN

13. Facial Expressions of Emotion 211DAVID MATSUMOTO, DACHER KELTNER, MICHELLE N. SHIOTA,MAUREEN O’SULLIVAN, and MARK FRANK

14. A “Nose” for Emotion: Emotional Information and Challengesin Odors and Semiochemicals

235

JEANNETTE M. HAVILAND-JONES and PATRICIA J. WILSON

15. The Neuroimaging of Emotion 249TOR D. WAGER, LISA FELDMAN BARRETT, ELIZA BLISS-MOREAU,KRISTEN A. LINDQUIST, SETH DUNCAN, HEDY KOBER, JOSH JOSEPH,MATTHEW DAVIDSON, and JENNIFER MIZE

16. Interoception and Emotion: A Neuroanatomical Perspective 272A. D. (BUD) CRAIG

PART III. DEVELOPMENTAL CHANGES

17. The Development of Facial Expressions: Current Perspectiveson Infant Emotions

291

LINDA A. CAMRAS and SERAH S. FATANI

18. The Emergence of Human Emotions 304MICHAEL LEWIS

19. Children’s Understanding of Emotion 320PAUL L. HARRIS

20. The Interface of Emotional Development with Social Context 332CAROLYN SAARNI

21. Young Children’s Understanding of Others’ Emotions 348SHERRI C. WIDEN and JAMES A. RUSSELL

22. Intermodal Emotional Processes in Infancy 364ARLENE S. WALKER-ANDREWS

23. Long-Lived Emotions: A Life Course Perspectiveon Emotional Development

376

CAROL MAGAI

xiv Contents

PART IV. SOCIAL PERSPECTIVES

24. Gender and Emotion in Context 395LESLIE R. BRODY and JUDITH A. HALL

25. The Cultural Psychology of the Emotions: Ancient and Renewed 409RICHARD A. SHWEDER, JONATHAN HAIDT, RANDALL HORTON,and CRAIG JOSEPH

26. Intergroup Emotions 428ELIOT R. SMITH and DIANE M. MACKIE

27. Empathy and Prosocial Behavior 440MARTIN L. HOFFMAN

28. Social Functions of Emotion 456AGNETA H. FISCHER and ANTONY S. R. MANSTEAD

PART V. PERSONALITY ISSUES

29. Subjective Well-Being 471RICHARD E. LUCAS and ED DIENER

30. Temperament and Emotion 485JOHN E. BATES, JACKSON A. GOODNIGHT, and JENNIFER E. FITE

31. Emotion Regulation 497JAMES J. GROSS

32. Emotional Complexity 513KRISTEN A. LINDQUIST and LISA FELDMAN BARRETT

PART VI. COGNITIVE FACTORS

33. Emotional Intelligence 533PETER SALOVEY, BRIAN T. DETWEILER-BEDELL,JERUSHA B. DETWEILER-BEDELL, and JOHN D. MAYER

34. Some Ways in Which Positive Affect Influences Decision Makingand Problem Solving

548

ALICE M. ISEN

35. Advances in Modeling Emotion and Thought:The Importance of Developmental, Online, and Multilevel Analyses

574

NANCY L. STEIN, MARC W. HERNANDEZ, and TOM TRABASSO

36. Emotion Concepts 587PAULA M. NIEDENTHAL

37. Memory and Emotion 601ELIZABETH A. KENSINGER and DANIEL L. SCHACTER

Contents xv

38. A Framework for Representing Emotional States 618MARVIN MINSKY

39. Appraisal Theories: How Cognition Shapes Affect into Emotion 628GERALD L. CLORE and ANDREW ORTONY

PART VII. HEALTH AND EMOTIONS

40. Emotions and Health Behavior: A Self-Regulation Perspective 645MICHAEL A. DIEFENBACH, SUZANNE M. MILLER, MATTHEW PORTER,ELLEN PETERS, MICHAEL STEFANEK, and HOWARD LEVENTHAL

41. Emotions, the Neuroendocrine and Immune Systems, and Health 661MARGARET E. KEMENY and AVGUSTA SHESTYUK

42. Health-Promoting and Health-Damaging Effects of Emotions:The View from Developmental Functionalism

676

NATHAN S. CONSEDINE

43. Emotion Disturbances as Transdiagnostic Processesin Psychopathology

691

ANN M. KRING

PART VIII. SELECT EMOTIONS

44. Fear and Anxiety: Overlaps and Dissociations 709ARNE ÖHMAN

45. The Development of Anger and Hostile Interactions 730ELIZABETH A. LEMERISE and KENNETH A. DODGE

46. Self-Conscious Emotions: Embarrassment, Pride, Shame, and Guilt 742MICHAEL LEWIS

47. Disgust 757PAUL ROZIN, JONATHAN HAIDT, and CLARK R. McCAULEY

48. Positive Emotions 777BARBARA L. FREDRICKSON and MICHAEL A. COHN

49. Sadness and Grief 797GEORGE A. BONANNO, LAURA GOORIN, and KARIN G. COIFMAN

Author Index 811

Subject Index 832

xvi Contents

PA R T I

INTERDISCIPLINARYFOUNDATIONS

C H A P T E R 1

The Philosophy of Emotions

ROBERT C. SOLOMON

“What is an emotion?” That question wasasked in precisely that form by William James,as the title of an essay he wrote for Mind wellover 100 years ago (James, 1884). But philoso-phers have been concerned about the nature ofemotion since Socrates and the “pre-Socratics”who preceded him, and although the disciplinehas grown up (largely because of Socratesand his student Plato) as the pursuit of rea-son, the emotions have always lurked in thebackground—often as a threat to reason and adanger to philosophy and philosophers. One ofthe most enduring metaphors of reason andemotion has been the metaphor of master andslave, with the wisdom of reason firmly in con-trol and the dangerous impulses of emotionsafely suppressed, channeled, or (ideally) inharmony with reason. But the question “Whatis an emotion?” has proved to be as difficult toresolve as the emotions have been to master.Just when it seems that an adequate definitionis in place, some new theory rears its unwel-come head and challenges our understanding.

The master–slave metaphor displays twofeatures that still determine much of the philo-

sophical view of emotion today. First there isthe inferior role of emotion—the idea thatemotion is as such more primitive, less intelli-gent, more bestial, less dependable, and moredangerous than reason, and thus needs to becontrolled by reason (an argument that Aris-totle and other enlightened Athenians used tojustify the political institution of slavery, aswell). Second, and more profoundly, there isthe reason–emotion distinction itself—as if wewere dealing with two different natural kinds,two conflicting and antagonistic aspects of thesoul. Even those philosophers who sought tointegrate them and reduce one to the other(typically reducing emotion to an inferior ge-nus of reason, a “confused perception” or “dis-torted judgment”) maintained the distinctionand continued to insist on the superiority ofreason. It was thus a mark of his considerableiconoclasm that the Scottish skeptic DavidHume (1739/1888), in the 18th century, fa-mously declared that “reason is, and ought tobe, the slave of the passions.” But even Hume,despite an ingenious analysis of the structure ofemotions, ultimately fell back on the old mod-

3

els and metaphors. His work remains an exem-plary celebration of reason, even while chal-lenging its limits.

Philosophy is a historical discipline. It is con-strained and defined as much by its past as byany particular field of phenomena. Philosophi-cal theories and debates today cannot be un-derstood or appreciated without some under-standing of philosophy’s rich and convolutedpast. Even when a philosopher pretends to un-derstand the phenomenon of emotion “in it-self,” or to analyze the language of emotionwithout reference to history or to any earlierattempts at analysis, both the wisdom and thefolly of generations of accumulated reflectionand argument are nevertheless inescapable. Al-though one might impatiently demand fromthe outset that one “define the terms” beforethe current discussion commences, the truth isthat a definition emerges only at the end of along discussion, and even then it is always ten-tative and appropriate only within a limitedcontext and certain models of culture and per-sonal character.

In what follows, I have tried to sketch asomewhat selective history of philosophical at-tempts to understand emotion, followed by abrief summary of questions still central tophilosophical debate. Given the nature of phi-losophy and its emphasis on reason, however,we would expect that the focus of most philo-sophical analysis has been and remains themore cognitive aspects of emotion, with thephysiological and to a certain extent the socialand behavioral dimensions of emotion dimin-ished or even denied. The dialectic in philoso-phy, accordingly, tends to go back and forth inits rediscovery of these often neglected dimen-sions. Sometimes emotions are dismissed asmere feelings and physiology, utterly unintelli-gent, even subhuman. In reaction, emotions arethen ascribed the virtues of true wisdom; theyare defended as the proper masters of reasonand even the very foundation of our being-in-the-world. Most philosophers, however, try tofind some more moderate, multidimensionalposition.

One might object that philosophical theoriesof emotion tend to be “armchair” speculation,devoid of the empirical support supplied by so-cial scientists. However, this objection ignoresthe fact that philosophers, contrary to theirown self-styled reputations as men and womenof pure reason, have emotions themselves, andin most (but not all) cases a sufficiently rich

repertoire of emotions to fund and support adozen theories of emotion. As Descartes (1649/1989) said in his introduction to the subject,“everyone has experience of the passions with-in himself, and there is no necessity to borrowone’s observations from elsewhere in order todiscover their nature.” Ultimately, there is nojustification for the century-old feud betweenphilosophy and psychology. Their history is infact the same, and the phenomenon of emotionlies equally open to both of them.

THE HISTORY OF THE PHILOSOPHYOF EMOTION

Although the history of philosophy has oftenbeen described as the history of the develop-ment of reason—for example, by the great19th-century German philosopher G. W. F.Hegel—philosophers have never entirely ne-glected emotion, even if they have almost al-ways denied it center stage. It would be a mis-take, however, to put too much emphasis onthe term “emotion,” for its range and meaninghave altered significantly over the years, due inpart to changes in theories about emotion. Sotoo, the word “passion” has a long and variedhistory, and we should beware of the mislead-ing assumption that there is a single, orderly,natural class of phenomena that is simply des-ignated by different labels in different lan-guages at different times. The language of“passion” and “emotion” has a history intowhich various feelings, desires, sentiments,moods, attitudes, and more explosive re-sponses enter and from which they exit, de-pending not on arbitrary philosophical stipula-tion but on an extensive network of social,moral, cultural, and psychological factors.Thus we will often find that the focus is notemotion as such, but rather some particularclass of emotions or particular emotion and itsrole in the manners or morals of the time.

The emotions as such, accordingly, do notform one of the three aspects of Plato’s (ca.428–347 B.C.) tripartite soul as defined in TheRepublic (1974). There are reason, spirit, andappetite; not only does what we call emotionseem divided between spirit and appetite, but,considering Plato’s discussion of eros as thelove of the Good in his dialogue The Sympo-sium (1989), there are emotions involved inreason as well. Aristotle (384–322 B.C.), bycontrast, did seem to have a view of emotion as

4 I. INTERDISCIPLINARY FOUNDATIONS

such, but although he had a mania for taxono-mies, he spent relatively little time listing or an-alyzing the emotions—as he did, for example,the virtues and the various kinds of birds. In hisRhetoric (1941), however, he defined emotion“as that which leads one’s condition to becomeso transformed that his judgment is affected,and which is accompanied by pleasure andpain. Examples of emotion include anger, fear,pity, and the like, as well as the opposites ofthese.”1 (He did not tell us what these “oppo-sites” might be.) Aristotle discussed certainemotions at length, notably anger, which he de-scribed in remarkably modern terms. In theRhetoric he defined anger as “a distressed de-sire for conspicuous vengeance in return for aconspicuous and unjustifiable contempt ofone’s person or friends.” He added that “angeris always directed toward someone in particu-lar, for example, Cleon, and not toward all ofhumanity,” and mentioned (if only in passing)the physical distress that virtually always ac-companies such emotion.

The key to Aristotle’s analysis, however, isthe notion of a “slight.” This is the cause of an-ger, and it may be an instance of “scorn, spite,or insolence.” Aristotle made allowances foronly imagined slights (in other words, unwar-ranted anger is nevertheless anger), and he gavea central place to the desire for revenge, thusintroducing a behavioral component at theheart of the emotion. We might note that Aris-totle, who was so precocious in so many disci-plines, seems to have anticipated most of themain contemporary theories. His analysis ofanger includes a distinctive cognitive compo-nent, a specified social context, a behavioraltendency, and a recognition of physical arousal.He even noted that physical or psychologicaldiscomfort—sickness, poverty, love, war,breached expectations, or ingratitude—yields apredisposition for anger. It is worth noting thatAristotle had little to say of “feeling,” presum-ably not because the Greeks were anesthetic,but rather because what we (inconsistently) call“affect” and inner sensation generally held lit-tle interest for them and played no significantrole in their language or their psychology.

Perhaps the most important single point tomake about Aristotle’s view of emotion is thefact that his analyses make sense only in thecontext of a broader ethical concern. Angerwas of interest to him because it is a natural re-action to offense and a moral force, which canbe cultivated and provoked by reason and rhet-

oric. (Thus its inclusion in a book on thattopic.) Anger (and several other emotions, no-tably pride) is also prominent in Aristotle’sclassical list of virtues in his Nicomachean Eth-ics (1941), where he discussed in some detailthose circumstances in which it is appropriateto get angry, those in which it is not, and whatamount or intensity of anger is justified. Hesuggested that forgiveness may be a virtue, butonly sometimes. He also insisted that only foolsdon’t get angry, and that although overly angrypeople may be “unbearable,” the absence ofanger (aimed at the right offenses) is a vicerather than a virtue. In this as in all else, Aris-totle defended moderation, the “mean betweenthe extremes.” So too, he discussed fear atlength in the Ethics with regard to courage,which is not fearlessness or “overcoming” fearso much as it is having just the right amount offear—not being either foolhardy or cowardly.The emotions, in other words, are central andessential to the good life, and the analysis oftheir nature is part and parcel of an ethicalanalysis.

So too, in Roman times, we find the conjunc-tion of ethics and emotion in the philosophy ofthe Stoics (see Sorabji, 2003, and Nussbaum,1994). But whereas Aristotle took emotion tobe essential to the good life, the Stoics analyzedemotions as conceptual errors, conducive tomisery. In modern terms, the Stoics Seneca andChrysippus developed a full-blooded cognitivetheory of the emotions two millennia ago (seeespecially Seneca, 1963). Emotions, in a word,are judgments—judgments about the worldand one’s place in it. But the world of Romansociety was not a happy or a particularly ratio-nal place. (Seneca served under the EmperorNero, and ultimately committed suicide at hisbehest.) And as the Stoics saw the world theylived in as out of control and beyond any rea-sonable expectations, they saw the emotions,which impose such expectations on the world,as misguided judgments about life and ourplace in the world. The emotions consequentlymake us miserable and frustrated. Accordingly,the Stoics made a careful study of the compo-nent judgments that compose the emotions—the presumptuousness of moral judgmentin anger, the vulnerability of love, the self-absorption of security in fear. The alternativewas seen as “psychic indifference,” or apatheia(apathy). The Stoics believed in a “higher” rea-son, one transcending the vanities of the socialworld. But they felt that the best life in that

1. The Philosophy of Emotions 5

world could be achieved only by gettingstraight about the ultimate pointlessness ofemotional attachments and involvement.

Throughout the Middle Ages, the study ofemotion was again typically attached to ethics,and it was central to Christian psychology andthe theories of human nature in terms of whichthe medievals understood themselves (seeHyman & Walsh, 1973). There were elaborate,quasi-medical studies of the effects of the vari-ous “humours” (gall, spleen, choler, and blooditself) on emotional temperament, but therewere (as there were among the Stoics) espe-cially rich studies of the cognitive and “cona-tive” aspects of the emotions. Emotions wereessentially linked with desires, particularly self-interested, self-absorbed desires. And so theChristian preoccupation with sin led to elabo-rate analyses of those emotions, passions, anddesires designated as sins (notably greed, glut-tony, lust, anger, envy, and pride; sloth, per-haps, is a special case). The tight linkage be-tween the study of emotion and ethics isparticularly evident in the curious observationthat the highest virtues, such as love, hope, andfaith, were not classified as emotions as such,but were rather elevated to a higher status andoften (e.g., by Thomas Aquinas) equated withreason. The old master–slave metaphor re-mained alive and well, and as some emotionswere seen as sins, the highest virtues couldhardly be counted among the mere emotions.

Reviewing the ancient and medieval litera-ture on emotion, René Descartes (1596–1650)was provoked to write that what they taughtwas “so slight, and for the most part so farfrom credible, that I am unable to entertain anyhope of approximating the truth excepting byshunning the paths they followed” (1649/1989). Descartes is typically recognized as the“father” of modern philosophy, and, in a morescholarly vein, as the bridge between the scho-lastic world of the Middle Ages and our own.But Descartes was fundamentally a scientistand a mathematician, awed by “the naturallight of reason” and fascinated by the uniqueautonomy of the human mind. Accordingly, hedisdained the bodily and the bestial, insistingthat the mind is a separate “substance” fromthe body (and that beasts therefore do not haveminds). The separation of mind and bodyproved to be a famously difficult problem forDescartes and his successors, however, and no-where was that problem more evident than inhis attempt to deal with the emotions.

Thoughts about mathematics may be clearly“in” the mind, as stomach contractions are inthe body, but an emotion seems to require theinteraction of mind and body in an undeniableway. Accordingly, Descartes defended a theoryin his treatise On the Passions of the Soul(1649/1989), in which the mind and body“meet” in a small gland at the base of the brain(now known as the pineal gland), and the latteraffects the former by means of the agitation of“animal spirits” (minute particles of blood),which bring about the emotions and their phys-ical effects in various parts of the body. But theemotions also involve not only sensationscaused by this physical agitation, but percep-tions, desires, and beliefs as well. Thus overand above the physical agitation and familiarsensations, the emotion of hatred ultimatelyarises from the perception of an object’s poten-tial harmfulness and involves a desire to avoidit. Accordingly, it is not as if an emotion ismerely a perception of the body; it may also be,as Descartes put it, a perception of the soul(e.g., a perception of desire), and some percep-tions (as in dreams) may in fact be of thingsthat do not exist at all.

An emotion is one type of “passion,” forDescartes defined the passions in general as“the perceptions, feelings or emotions of thesoul which we relate specifically to it, andwhich are caused, maintained, and fortified bysome movement of the [animal] spirits.” Thepassions in general are distinguished from“clear cognition,” and render judgment “con-fused and obscure.” Emotions are particularlydisturbing passions. And yet emotions can beinfluenced by reason. For example, writing ofcourage, Descartes stated:

To excite courage in oneself and remove fear, it isnot sufficient to have the will to do so, but wemust also apply ourselves to consider the reasons,the objects or examples which persuade us thatthe peril is not great; that there is always more se-curity in defense than in flight, that we shouldhave the glory and joy of having vanquished,while we should expect nothing but regret andshame for having fled, and so on.

And so the physiological account gives wayto a cognitive account, and the emotions movefrom being merely bodily to becoming an es-sential ingredient in wisdom: “The utility of thepassions consists alone in their fortifying andperpetuating in the soul thoughts which it isgood that it should preserve, and which with-


out that might easily be effaced from it.” Howthen can there be “bad” emotions? “The harmis that they fortify these thoughts more thannecessary, or they conserve others on which it isnot good to dwell.” Somewhat bewildered bythe physiology (though he was at the cuttingedge of the science of his times), Descartes ulti-mately tended to a value-oriented analysis ofemotion. His six “primitive” passions—wonder, love, hatred, desire, joy, and sadness—are not meaningless agitations of the animalspirits, but ingredients in the good life.

Baruch (Benedict) Spinoza (1632–1677)might well be considered to be a latter-dayStoic, like Chrysippus and Seneca in ancientRome. Just as the Stoics saw the emotions asmisguided judgments about life and our placein the world, Spinoza too saw the emotions as aform of “thoughts” that, for the most part,misunderstand the world and consequentlymake us miserable and frustrated. But unlikethe Stoics, Spinoza did not aspire to that “psy-chic indifference” known as apatheia; rather, inhis Ethics (1677/1982), he urged the attain-ment of a certain sort of “bliss,” which can beachieved only once we get straight our thinkingabout the world. In particular, we have to giveup the idea that we are or can be in control ofour own lives, and adopt instead the all-embracing idea of ourselves and our minds aspart of God. Most of the emotions, which arepassive reactions to our unwarranted expecta-tions of the world, will leave us hurt, frus-trated, and enervated.

The active emotions, by contrast, emanatefrom our own true natures and heighten oursense of activity and awareness. Spinoza, likethe Stoics, developed an early version of thecognitive theory of emotion. But Spinoza alsodefended a grand and complex metaphysics, inwhich all substance is one, and mind and bodyare but dual “aspects” of one and the same be-ing. Accordingly, he did not face Descartes’sformidable “mind–body” problem; althoughhe himself would not have predicted this, heanticipated some of the subtle emotion–brainresearch that is being carried out today.

David Hume (1711–1776) was one of themost outspoken defenders of the Enlighten-ment, that very vocal and often rebellious intel-lectual movement that challenged old ortho-doxies, elevated science and put religion on thedefensive, attacked superstition and irrational-ity in all quarters, practiced and encouragedvigorous debate and discussion, and put a pre-

mium on the virtues of reason. But Hume, incarrying out the directives of reason to chal-lenge, debate, and question, came to questionthe role and capacities of reason itself, and inparticular the power of reason to motivate eventhe most basic minimum of moral behavior. “Itis not against reason,” he declared in one of hismost outrageous proclamations, “to prefer thedestruction of half the world to the scratchingof my finger” (1739/1888). What motivates usto right (and wrong) behavior, Hume insisted,were our passions, and rather than being rele-gated to the margins of ethics and philosophy,the passions deserve central respect and consid-eration.

Accordingly, he gave the passions the largemiddle portion of his great first book, A Trea-tise of Human Nature (1739/1888). Unfortu-nately, however, most philosophers then andsince have preferred to read the first and thirdparts, on knowledge and ethics, and to ignorethe central position of the passions.

Hume’s theory is especially important notonly because he challenged the inferior place ofpassion in philosophy and questioned the roleof reason. He also advanced a theory of thepassions that, although limited and encum-bered by his general theory of mind, displayeddazzling insight and a precocious attempt tograpple with problems that would only be for-mulated generations later. Hume, like many ofhis contemporaries and predecessors, definedan emotion as a certain kind of sensation, orwhat he called an “impression,” which (as inDescartes) is physically stimulated by themovement of the “animal spirits” in the blood.Such impressions are either pleasant or un-pleasant, but the differentiation of the manyemotions is not to be found in the nature ofthese impressions as such. Rather, the impres-sions that constitute our emotions are alwaysto be located within a causal network of otherimpressions and, importantly, ideas. Ideascause our emotional impressions, and ideas arecaused in turn by them. The pleasant impres-sion of pride, for example, is caused by the ideathat one has achieved or accomplished some-thing significant, and the impression in turncauses another idea, which Hume described asan idea of the self, simpliciter.

The emotion, in other words, cannot beidentified with the impression or sensationalone, but can only be identified by the wholecomplex of impressions and ideas. What Humeacknowledged with his emphasis on the essen-


tial place of ideas in emotion is what we nowcall the cognitive dimension of emotion, in ad-dition to the physiological (“animal spirits”)and merely sensational (“impression”) aspectsof emotion. Moreover, his inclusion of the sec-ond idea of the self in his analysis of prideindicates his grappling with the notion ofintentionality (the “aboutness” of emotions)—an effort further reinforced by his somewhatobscure insistence that the connection betweenan emotion (the impression) and this conse-quent idea is “original” or “natural,” or some-thing more than the merely causal associationsthat form the usual bonds between ideas andimpressions.

The emotions, for Hume, form an essentialpart of ethics. There are good emotions and bademotions. Pride, he declared, is a good emotion.Humility, its opposite (an unpleasant feelingbrought about by the idea that we are inade-quate or deeply flawed beings), is a bad emotion,a “monkish” emotion. Here we can see again theextent to which, as so often, a theory of emotionserves to grind some larger philosophical ax—inthis case, Hume’s Enlightenment attack on reli-gion. In this regard too, we might mention an-other aspect of Hume’s moral philosophy, fol-lowed in kind by his illustrious Edinburgh friendand colleague Adam Smith (1723–1790, also theauthor of The Wealth of Nations [1776/1976],the bible of modern capitalism).

Hume and Smith both defended the impor-tance of what they called “the moral senti-ments” (see Smith, 1759/1976), the foremostof which is sympathy, our ability to “feel with”other people and appreciate (if not suffer with)their misfortunes. Sympathy, they argued, is auniversal feature of human nature (counteringand mitigating the self-interest that Smith inparticular famously championed in The Wealthof Nations), and it is the bedrock foundation ofsociety and morality. Emotion, in other words,is not an embarrassment or part of the refuse ofthe human psyche, but rather the very essenceof human social existence and morality. It isnot to be unfavorably contrasted and opposedto reason, but, on the contrary, is to be cele-brated and defended along with it.

Immanuel Kant (1724–1804) was also achampion of the Enlightenment, but althoughhe too questioned the capacities and lim-its of reason, he was uncompromising in itsdefense—against Hume’s skepticism, againstany attempt to replace reason by irrationalfaith, and against any attempt to ground ethicsin fleeting human feeling instead of the univer-

sal and necessary dictates of reason. Thus Kantreinforced the crucial distinction between rea-son and what he called “the inclinations”(emotions, moods, and desires) and dismissedthe latter (including the moral sentiments) asinessential to morals at best and intrusive anddisruptive at worst. And yet, although Kant feltno need to develop a theory of emotion to ac-company his elaborate and brilliant “critiques”of reason, his position on the “inclinations” ismore ambiguous than is usually supposed, andhis respect for “feeling” more significant. Itwas Kant, a quarter-century before Hegel (whois credited with it), who insisted that “nothinggreat is ever done without passion,” and it wasKant, in his Critique of Judgment (1793/1953,concerned in part with art and aesthetics), whocelebrated the importance of shared (“inter-subjective”) feeling in the appreciation ofbeauty and the awe with which we try to com-prehend the wonder of God’s creation. Indeed,even Kant’s central notions of respect and hu-man dignity—the very heart of his rationalistethics—are sometimes suggested to be mattersof feeling as well as reason, thus calling intoquestion the harshness of his ruthlessly dividedself. When his successor Hegel took over thereins of German philosophy in the early 19thcentury, the overstated distinction between rea-son and passion was again called into question,and Hegel’s own odyssey of reason (in anepochal book called The Phenomenology ofSpirit [1807/1977]) has rightly been called a“logic of passion” as well.

Friederich Nietzsche (1844–1900) was a phi-losopher for whom passion was the watchwordand reason a source of suspicion. He was theculmination of a long line of “Romantics,” be-ginning with the Sturm und Drang poets of theprevious century and continuing through thephilosophy of Nietzsche’s own favorite in-fluence, the neo-Kantian pessimist ArthurSchopenhauer. Nietzsche anticipated the globalskepticism and conceptual chaos of the 20thcentury; like Freud, who admired him, he de-scribed (and celebrated) the darker, more in-stinctual, and less rational motives of the hu-man mind. Accordingly, in his On theGenealogy of Morals (1887/1967), he praisedthe passions and, in an ironic twist, describedthe passions as themselves having more reasonthan Reason. But this was not to say that allpassions are wise; some, he declares, “drag usdown with their stupidity,” and others, notablythe “slave morality” emotion of resentment,are devious and clever but to a disastrous


end—the “leveling” of the virtuous passionsand the defense of mediocrity. Nietzsche neverdeveloped a “theory” of emotions, but his dis-tinctions were remarkable in their insight andsubtlety. His celebration of passion scared thewits out of a great many philosophers in Eu-rope, however, who saw more than enoughpassion and irrationality in World War I andthen the rise of National Socialism in Germany.Accordingly, the ancient celebration of reasonwould once more rule philosophy, and emotionwas again relegated to the sidelines.

In the 20th century, one can trace the fate ofemotion in Western philosophy through twovery different tracks. In North America and inEngland, the emotions were given short shrift,in large part because of the newly exaggeratedemphasis on logic and science. The great Brit-ish philosopher Bertrand Russell gave elabo-rate praise to love and passion in the openingpages of his autobiography (1967), but in hisphilosophy he said virtually nothing aboutthem. Of course, the nature of emotion was amajor concern of William James and the youngJohn Dewey in the early years of the century,but with James’s emphasis on the physiologicalnature of emotion (he argued [1884] that anemotion is a sensation or set of sensationscaused by a physiological disturbance, which inturn is prompted by some “perception” orother), coupled with the subsequent and quiteunfortunate split between philosophy and psy-chology as academic disciplines, questionsabout emotion were relegated to the realm ofpsychology (where they were also treated withless than the full respect due them). Indeed, thefirst major attention to emotion in Anglo-American philosophy came in midcentury,when an ethical theory named “emotivism”came to dominate both the English and theNorth American scene. But emotivism, whichwas part and parcel of an across-the-boardphilosophical purgative known as “logical pos-itivism,” was essentially a dismissal of ethical(and many other) questions in philosophy as“meaningless” (i.e., unscientific and withoutverifiable solutions). Emotion came back ontothe stage of philosophy, but only as the butt ofthe argument: Ethical statements were viewedas meaningless because they were seen as noth-ing but expressions of emotion.

During the same period in Europe, however,the emotions enjoyed more attention. FranzBrentano (1874/1971) succeeded the British“moral sentiment” theorists in attempting tofound an ethics on a foundation of emotions.

(Sigmund Freud was one of his students.) Fol-lowing the “phenomenology” of EdmundHusserl (1838/1960), Max Scheler (1916/1970), Martin Heidegger (1927/1962), andmore recently, Paul Ricoeur (1950/1966) devel-oped ambitious philosophies in which emo-tions were given a central place in human exis-tence and accorded with considerable respect.Heidegger, in particular, defended what he gen-erally called “moods” as our way of “beingtuned” to the world. In the shadow of WorldWar II, Jean-Paul Sartre offered the slim butimportant The Emotions: Sketch of a Theory(1939/1948), followed by his magnificent tomeBeing and Nothingness (1943/1956), which in-cludes embedded within its many pages a num-ber of detailed “phenomenological” analysesof emotion. Sartre’s conception of emotionsas “magical transformations of the world”—willful strategems for coping with a difficultworld—added a new “existential” dimensionto the investigation of emotion. But, predict-ably, philosophy in both France and Germanyturned again to other interests, although thestudy of emotion continued despite the peren-nial shift in fashions.

In Anglo-American philosophy, however, thefortunes of emotion were also to change. In anarticle simply entitled “Emotion” (indicatinghow rarely the topic had even been broached),Errol Bedford (1956/1964) addressed theAristotelean Society in London on the nature ofemotion and the errors of thinking of emotionsas “feelings.” The essay might have sat on theshelves gathering dust except for the fact thatthe then dean of Oxford philosophers, J. L.Austin (1956–1957/1964), took it upon him-self to remark on one of Bedford’s claims. (Aus-tin’s own essay was not about emotions at all.)Austin’s attention kept the article alive and oc-casionally anthologized until the 1960s, whenthe subject seemed to come to life again.

Today, one finds a rich variety of argumentsabout emotions on both sides of the AtlanticOcean and the English Channel. Given the na-ture of philosophy and its current concern withepistemological matters, it is again not surpris-ing that the focus is on the conceptual struc-tures of emotion, rather than the sensory, so-cial, or physiological aspects of emotion. Butthere has been a reaction even within philoso-phy to the “hypercognizing” of emotion; con-sequently, there has been a serious effort to joinforces with psychologists, neurologists, anthro-pologists, and moral philosophers to obtain amore holistic theory of emotion.


SOME PHILOSOPHICAL QUESTIONSABOUT EMOTION

What is an emotion? Because philosophy is adiscipline concerned with the essential natureand the “definition” of things, the basic ques-tion facing theories of emotion in philosophy isstill the question posed by James and answered,in a fashion, by Aristotle. It is, on the face of it,a quest for a definition, a conceptual analysis.But it is also a much larger quest for an orienta-tion: How should we think about emotion—asintrusive, as essential to our rationality, as con-stitutive of meaning, as dangerous, as dispens-able, as an excuse for irresponsibility, or as amode of responsibility? Which of the evidentaspects of emotion—that is, the various sen-sory, physiological, behavioral, cognitive, andsocial phenomena that typically correspondwith an emotion—should we take to be essen-tial? Many philosophers hold onto the old“Cartesian” view that an emotion cannot lackits “subjective” or “introspective” aspect, al-though what this means (and how accessible orarticulate an emotion must be on inspection) isitself a subject of considerable dispute, for in-stance, in Freud (1915/1935), Sartre (1943/1956), Lyons (1980), and de Sousa (1987).But many philosophers have become skepticalabout such subjective essentialism and, liketheir associates in the sciences, have pushedthe analysis of emotion toward more public,observable criteria. Accordingly, philosophershave formulated their own versions of behav-iorism, physiologism, and social constructiontheory, for example, although they have not al-ways been mutually aware of their counter-parts in the social sciences, especially.

The seemingly self-evident Cartesian de-mand that first-person experience is inelim-inable is evident, nevertheless, even among themost radical philosophical behaviorists. For in-stance, Gilbert Ryle (1951) chastised philoso-phers for their “myth of the ghost in the ma-chine” and suggested that many emotions aremere “agitations” (much as Descartes had in-sisted) and dispositions to behave in certainways. But Ryle did not give up the idea thatsome of the symptoms of emotion consist of“tingles and itches” or some such “feeling.”Can one have an emotion without feeling?What is a “feeling”? According to WilliamJames (1884), it is a specifiable sort of sensa-tion, the sensation of one or more “visceral dis-turbances”—changes in the body due, for in-

stance, to the stimulation of the autonomic ner-vous system. The great virtue of the Jamesiantheory is that it ties down the nature of emo-tional “feelings” to quite particular and there-fore verifiable bodily responses. Unfortunately,the Jamesian theory has often been shown to bewrong, at least in its details (e.g., Cannon,1929/2003; Dewey, 1894/2003; Schachter &Singer, 1962). How specifically are emotionalfeelings tied to physiological processes? To besure, whatever goes on in the mind must nowbe supposed to have some correlate and causein the brain, but can we not and should we notdescribe the “phenomenology” of those feel-ings quite apart from their brain correlationsand causes? Some theorists have tried to savefeeling theory by employing the vague, general(and technical) notion of “affect” and its cog-nates (“affective tone”) (Stocker, 1996). But dosuch terms do anything more than cover up theproblem with another word, whose meaningcan only be explained by “the kind of feelingyou get when you have emotion X”? It is a heu-ristic mistake to suppose that such feelings areindescribable or “ineffable,” whether out ofexcessive romanticism (as if understanding un-dermines passion) or dismissive scientism (whytalk about feelings if we can’t experimentallytest them?)

Most feelings have at least an “as if” famil-iarity (“It feels as if I’d known her for years” or“It felt as if he had shot me through the heart, itwas so sudden and so traumatizing”). Manyfeelings have a distinctive structure, which (notsurprisingly) emerges in the thoughts (and thenin the verbal expressions) of the emotions.Thus we should identify the experience of hav-ing an emotion (as opposed to just a simple“feeling”) as embodying thoughts, judgments,and other cognitive elements. In general, oneshould ask how much cognition and learningare presupposed in the feelings that we identifyas emotions. It may well be that at least some“basic” emotions are largely to be explained byreference to one or another neurophysiological“affect programs,” but even the most basicemotions involve or come to involve“intentionality”—an engagement with theworld. And this involves perception and someknowledge, as well as the abilities to act in theworld. Thus an emotional experience is not justa Jamesian sensation, but a complex awarenessof one’s engagements in the world and one’stendencies to act in it. (There may be, however,a kind of “borrowed intentionality” peculiar to


emotional sensations; Goldie, 2000.) Whetheror not there are Jamesian sensations that ac-company such engagements and action tenden-cies (and no doubt usually there are), the emo-tion is first of all a mode of engagement. But itshould be said that we are not always fullyaware of our engagements in the world, nor arewe usually fully cognizant of our feelings.

Thus there is room for “unconscious”emotions—a grab-bag concept that embraceseverything from the Freudian “Unconscious,”to the fact that we misidentify and fail to recog-nize our own emotions, to the more bracingclaims of some neuroscientists that an emotionis essentially not conscious and that awareness(if it happens at all) comes late in the neurologi-cal game. As for the feelings (the sensations)themselves, Freud was right in wonderingwhether it makes any sense to claim that theycan be unconscious, given that their whole ex-istence seems to be to be experienced. But thisembodies just the confusion that one would ex-pect with such a grab-bag concept of the un-conscious: On the one hand, an unfelt feelingor sensation makes no sense, but the observa-tion that we are not always aware that we havea certain feeling or sensation is quite evident.And because the experience of an emotion isso complex (engaging one’s beliefs about theworld, oneself, and other people, as well as anynumber of preferred scenarios and outcomes),it is obvious how and why we may not recog-nize an emotion when we have it. But whatdoes seem essential to all emotions, includingthose that are most “basic,” is some sense ofwhat is going on in the world, some “cogni-tion,” whether or not one is or even can be (re-flectively) aware of it. Thus no neurologicalsyndrome or “affect program” can be an emo-tion if it does not engage the world in someway, perhaps by virtue of a more or less auto-matic “appraisal” or some possibly subliminalstimulus. The affect programs typical of, say,fear and anger do not actually constitute fearand anger if there are no appropriate apprais-als, beliefs, or judgments accompanying them.A person may well feel flushed, uncomfortable,and “as if ” he or she wants to flee or start afight with someone—but if there is no fearfulobject (more precisely, if the person has nosense of a fearful object), or if there is nothingobjectionable, frustrating, or offensive (to theperson), then those feelings do not count asfear and anger (or even as “feeling afraid” or“feeling angry”).

Recent advances in neurology have disclosedilluminating structural and functional patternsin the central nervous system that are corre-lated with, and that under experimental condi-tions bring about, certain emotional reactions.Do these patterns dictate the structure of an ad-equate theory of emotion, or are those findingsbut one more set of (contingent) consider-ations for inclusion in an all-embracing theory?Whatever the case, it is now clear that philoso-phers cannot ignore or neglect the rich neuro-physiological literature on emotions. Indeed,there is now a interdisciplinary subfield inphilosophy called “neurophilosophy,” whichmakes the new neurology central to any ade-quate analysis of emotion and “the mind”(Churchland, 1986). Philosophers may con-tinue to argue that Aristotle knew all aboutemotions even though he did not know any-thing about the brain, but they do so at theirperil—and in the face of the obvious fact thatamong the factors that have altered the historyof philosophy and its concepts most radicallyhave been new advances in previously un-known or undeveloped sciences.

Virtually all emotions get expressed (how-ever minimally) in behavior. Should behavioraltendencies or sequences of actions or certainbasic gestures be taken as essential? A greatdeal of detailed work in psychology has shownthe enormous subtlety and the seemingly“hard-wired” nature of basic patterns of facialexpression. And yet philosophers remain skep-tical about the implied shift in conception fromthe emotion to a symptom of emotion. Theemotion would seem to be the experience, theperception, the awareness—what is expressed,not the expression itself. On the other hand,many philosophers of a somewhat behavioristbent (following Wittgenstein’s later Philosophi-cal Investigations [1953] and Gilbert Ryle’sThe Concept of Mind [1951]) have suggestedthat an emotion is nothing but its behavioralexpression, though certainly not a single ges-ture but an open-ended sequence of actions. Anemotion is not a “ghostly inner event,” accord-ing to Ryle, but a “multitrack disposition” tobehave in any number of recognizable ways. Sotoo, philosophers have tried to understandemotion not as an inner feeling but as a value-laden description of a social situation. ThusErrol Bedford (1956/1964) suggested in his pi-oneering article that the difference betweenshame and embarrassment, for example, is notsome shade of difference between internal


qualia, but the difference between two con-trasting descriptions of the situation.

What remains at the core of all such theories,however, is an awareness that all emotions pre-suppose or have as their preconditions certainsorts of cognitions—an awareness of danger infear, recognition of an offense in anger, appreci-ation of someone or something lovable in love.Even the most hard-headed neurological orbehavioral theory must take account of the factthat no matter what the neurology or thebehavior, if a person is totally unaware (andnot just “consciously unaware”) of a certainstate of affairs or facts, he or she cannot havecertain emotions. If neurologically inducedrage does not include some object of anger, thatreaction (whatever else it may be) cannot beanger. So too, Freud’s “free-floating anxiety”would count as an emotion only insofar as itdoes indeed (as Freud [1915/1935] argued)have an object, albeit “unconscious.” Philoso-phers (following Aristotle and the scholasticsof the Middle Ages) have come to call this the“formal object” of emotion, and one mightwell think of this as the minimum essential setof “beliefs” defining an emotion type and amore or less specific kind of emotional experi-ence. The formal object of fear, to take an obvi-ous case, is a fearful object, together with the atleast minimal beliefs constituting the awarenessof the presence or threat of such an object.

Other emotions are more complicated and,accordingly, are more often topics of philo-sophical debate and disagreement. Angerwould seem to require a formal object involv-ing an offense, but some authors would allowfrustration alone to count as anger (Gordon,1987). Still others would argue that anger is toa large extent “socially constructed,” and itsmanifestation in any given culture will there-fore be more or less specific (though probablynot unique) to that culture (Averill, 1985). Jeal-ousy is more difficult still, for its object seemsto involve not only a threatened loss but a per-petrator as well (perhaps the threatened objectas a perpetrator too), and possibly the largersocial situation in which jealousy involves notonly loss but humiliation as well (Neu, 1980).But although the exact natures of the formalobjects and requisite beliefs of various emo-tions are matters of lively debate (and there iseven more doubt and debate over the very ideaof a generalized formal object for emotions assuch), the presumption is that every emotionmust have a cognitive basis and an object

(intentionality). There is some corollary debateconcerning the status of moods and mood-likeemotions (e.g., joy), which do not have a deter-minate object, but it can be argued that moodsdo have an object—namely, the world as awhole. “A depressed man lives in a depressedworld,” wrote Wittgenstein (1953).

There is also considerable debate over thenature of cognition itself. Beliefs seem to be es-tablished states and therefore lack the sponta-neity that characterizes many emotions. Beliefsalso seem to be too fully articulate for the unre-flective reaction that characterizes most emo-tions. For that reason, some theorists prefer theconcept of “judgment” or “evaluative judg-ment” (e.g., the ancient Stoics; Solomon, 1976,2003; Nussbaum, 2003), while others preferthe term “thought” (e.g., Spinoza, 1677/1982;Neu, 1977, 1999). Psychologists seem to prefer“appraisal,” although it should be said that thesocial sciences have been much more keen tounderstand “appraisal” in a multilevel fashionthan philosophers, at least until recently (Prinz,2004). Others have preferred the less cog-nitively committal notion of a way of seeing(“seeing as”)—sometimes as a rejection of thecognitive view, but more appropriately, per-haps, as a refinement of it (Calhoun, 1984,Roberts, 2003). The nature of an emotionalcognition, and whether it must be fully con-scious or capable of articulation, remain mat-ters for considerable debate. Indeed, if certainholistic suggestions can be worked out, it maybe that the very distinctions that philosophershave so long presupposed among cognition,behavior, physiology, and feeling are them-selves inadequate and ought to be integratedinto a single picture (Damasio, 1994).

One point of general agreement among phi-losophers is that emotions have intentionality.“Intentionality” is a technical notion, but itscommon-sense meaning can be captured by theidea that emotions are always “about” some-thing or other. One is always angry aboutsomething; one is always in love with someoneor something (even if one is also “in love withlove”); one is always afraid of something (evenif one doesn’t know what it is). Thus we canunderstand the “formal object” of an emotionas its essential intentionality—the kind of ob-ject (event, person, state of affairs) to which itmust be directed if it is to be that emotion. Butintentionality has also been the object of philo-sophical consternation for over a century now,because despite its appeal as a way of under-


standing the nature of perception and othermental “acts” (which gets us away from theimage of images or representations “in” themind), intentionality has its own peculiar com-plications (Kenny, 1963; Searle, 1983). Somephilosophers have argued for a more dynamicand action-laden concept than “aboutness,”perhaps “motility” or “engagement” (asabove) (Merleau-Ponty, 1962/1994; Solomon,2003). Many philosophers accept the idea ofintentionality, but try to integrate it into a neo-Jamesian analysis of emotions as physiologyplus sensation, by way of an intricate causaltheory of perception (Prinz, 2004). It could beargued that in so doing, they are denyingintentionality with the right hand while endors-ing it with the left.

Most troubling for philosophers is the obvi-ous fact that an emotion may be “about” somenonexistent, merely imagined object. The ob-ject of fear may be nowhere around. The imag-ined threat in jealousy may not exist. The per-son one still loves may be dead. (Indeed, theproblem seems to remain whether the loverknows of the death or not. In either case, theemotion is directed at a person who is in no po-sition to receive it.) Moreover, the object of anemotion would seem to be one and the sameobject, whether it exists or not. (It is one andthe same devil that is the object of a child’s fear,whether the devil exists or not.) Thus the onto-logical status of the intentional object of emo-tion causes considerable commotion, particu-larly in the area of aesthetics and “make-believe” (Walton, 1990). In recent decades,many Anglo-American language-oriented or“analytic” philosophers have reduced theseemingly mysterious notion of intentionalityto the supposedly more manageable notion of“intensionality,” a precisely defined feature ofcertain sorts of sentences (Dennett, 1978,1991). But whether intensionality does in factcapture the necessary features of intentionalityis itself a topic of considerable debate; at least itseems to confuse the language in which we de-scribe emotions with the nature of the emo-tions themselves (Searle, 1983).

Philosophers have also become concernedwith the “why?” of emotions—their functionand their explanation. Most of the work herehas been done on the explanation of particu-lar instances of emotion, although a few in-vestigators have recently tackled the muchlarger question of the evolution and functionof emotions as such (de Sousa, 1987;

Gibbard, 1990). Particular instances of emo-tion seem to be subject to two different sortsof explanations. On the one hand, becausethey are intentional and essentially involvebeliefs (as well as desires, needs, attitudes,and values), emotions seem to require an ex-planation that invokes a person’s belief andattitudes toward the world. A person is angrybecause he believes that so-and-so wrongedhim, or someone is saddened because she hasfound out that she has just lost a loved one,and so on. But this cannot be a complete ac-count of emotional explanation. We also ex-plain emotions by citing the fact that a per-son has been sleepless all week, or is ill, orhas been given some medication. In otherwords, explanation of emotion may cite anunderlying cause that may or may not makemention of the object of emotion. The causemay be physiological—for example, an un-derlying state of irritability, an ingested drug,or a direct surgical stimulation of the brain.The cause may be some state of affairs or in-cident that “triggered” the person’s emotion,but this may not be the object of the person’semotion, nor need he or she have any mem-ory or awareness of it. (Again, “subliminal”messages presumably work this way.)

How is causal explanation to be reconciledwith an explanation in terms of beliefs andattitudes? Many philosophers have tended toemphasize the importance of one form ofexplanation over the other, or to reduce all ex-planations to either causal explanations orbelief-and-desire, “reason”-type explanations.The latter sort of explanation provides a fulleraccount of the intentionality of an emotion bydescribing not only its formal object (“He’s an-gry because he’s been offended”) but the spe-cific details of the situation, as well as the per-son’s beliefs and various attitudes. The formersort of explanation invokes an underlyingcause that may or may not make mention ofthe object of emotion. Very often, however, thecitation of a cause of emotion (its initiatingstimulus or “trigger”) and the account of theobject of the emotion will be nominally thesame (“He got mad because she stepped on histoe”). The problem that has been addressed bymany philosophers and in recent years has be-come the subject of intense debate betweenthose who prefer some version of the in-tentionality model of emotion and those whodemonstrate a strong preference for a morebiological model of emotion (Rorty, 1980;


Nissenbaum, 1985; Griffiths, 1997; Prinz,2004).

The cognitive basis of emotions also raisesanother question, one that was often a matterof deep concern for earlier philosophers: thequestion of the rationality of emotions. Manythinkers have written as if the emotions werenot only irrational but also nonrational—noteven candidates for cognition. Accounts ofemotions as mere feelings or physiological pro-cesses would make them nonrational (one can-not have a “stupid” headache, except by wayof a roundabout complaint about its inconve-nience). Aristotle, on the other hand, simply as-sumed that an emotion can be appropriate orinappropriate, foolish or prudent, not just onthe basis of whether or not it is acceptable inthe circumstance in question (though that so-cial dimension is certainly essential), but on thebasis of the perceptions, beliefs, desires, andsituation of the individual. The fact that emo-tions consist at least in part of cognitionsmeans that they can be evaluated in terms ofthe same epistemic, social, and even ethical cri-teria that we use to evaluate beliefs and inten-tions: Are they appropriate to the context? Dothey consider the facts of the matter? Are theirperceptions fair and their evaluations reason-able? Indeed, the argument is now prevalentand persuasive that emotions cannot be under-stood without grasping their reasons, and thesereasons in turn give us a basis for evaluation(de Sousa, 1987; Greenspan, 1988). The cur-rent debate, however, concerns how these rea-sons are to be understood, and whether the ra-tionality of emotions can indeed be fairlycompared to the evaluation of more fully delib-erative, articulate activities.

The rationality of emotions also moves tocenter stage the question of emotions and eth-ics that we have been following through thehistory of philosophy. How does emotion enterinto ethical understanding, and how do ourethics affect our emotions? One thing is clear:The commingling of emotions and ethics is notgrounds for dismissing either ethics or emo-tion, as the midcentury “emotivists” suggested.But it is worth noting that a new conception ofthe emotional foundations of ethics has takenroot in the Anglo-American tradition and, inan appropriate irony, has taken the name“emotivism” (Gibbard, 1990). Of course, oneof the questions that remains, left over from thecharge that emotions are “subjective,” is thatemotions vary too much from culture to cul-

ture to provide a firm basis for ethics; in otherwords, they are “relative.” But though philoso-phers alone cannot answer the empirical ques-tion of the universality or relativity of emo-tions, they can and should clear away thedogmatic assumptions and mistaken concep-tions that have often occupied philosophy inthe past. There is nothing in the nature ofemotion (including the human brain, whichchanges significantly with experience and var-ies considerably from person to person) that as-sures universality, but neither is it so obviousthat emotions differ so much from place toplace either. (This is indicated not only by stud-ies of facial expression, but by the logic of the“human condition” and its more general fea-tures.) Indeed, the whole question of “humannature” is once again up for grabs.

One of the most critical questions about hu-man nature is the extent to which we cantranscend our own biology. In one sense, ofcourse, this is absurd, but in another it is per-fectly plausible. Human beings did not evolveto fly, but through technology we now have“frequent flyer” programs. Human beings didevolve with anger and jealousy as putativelyuniversal emotions, but the question remainswhether we can overcome our anger and jeal-ousy, or perhaps even eliminate them from ouremotional repertoires. This raises the questionof emotions and choice, and challenges the sup-position that we are passive regarding ouremotions. Sartre (1939/1948, 1943/1956) sug-gested that the emotions are both choices andstrategies, but many philosophers who do notshare Sartre’s extreme voluntarism would agreethat emotions are indeed ways of coping,whether inherited through natural selection orcultivated in the less articulate practices of a so-ciety. But are we at the mercy of our emotions?Do we simply “have” them, or do we perhapsto some extent cultivate them and “do” themourselves? Obviously, a good deal of ethics andour attitudes toward ourselves depend on this.The study of emotion in philosophy is, accord-ingly, not a detached and marginal discipline,but the very core of our inquiry into ourselvesand our own natures. It was Socrates, the greatchampion of reason, who took as his mottosthe slogan at Delphi (“Know thyself ”) and therather extreme injunction that “The unexam-ined life is not worth living.” But part of thatknowledge, surely, is our understanding andappreciation of our emotions, which are, afterall, much of what makes life worth living.


NOTE

1. This and other quotations from Aristotle in thischapter have been translated by Jon Solomon.

REFERENCES AND FURTHER READING

Aristotle. (1941). The basic works of Aristotle (R.McKeon, Ed.). New York: Random House. See alsoW. Fortenbaugh (1975). Aristotle on emotion. Lon-don: Duckworth.

Austin, J. L. (1964). Pretending. In D. Gustafson (Ed.),Essays in philosophical psychology. Garden City,NY: Doubleday/Anchor. (Original work published1956–1957)

Averill, J. R. (1985). The social construction of emotion,with special reference to love. In K. Gergen & K. Da-vis (Eds.), The social construction of the person(pp. 89–109). New York: Springer-Verlag.

Barbalet, J. (1999). William James’ theory of emotions.Journal for the Theory of Social Behavior, 29, 251–256.

Bedford, E. (1964). Emotion. In D. Gustafson (Ed.), Es-says in philosophical psychology. Garden City, NY:Doubleday/Anchor. (Original work published 1956)

Ben-Zeev, A. (2000). The subtlety of emotions. Cam-bridge, MA: MIT Press.

Brentano, F. (1971). Psychology from the empiricalstandpoint. London: Routledge. (Original work pub-lished 1874)

Calhoun, C. (1984). Cognitive emotions? In C. Calhoun& R. C. Solomon (Eds.), What is an emotion? NewYork: Oxford University Press.

Cannon, W. B. (2003). Bodily changes in pain, hunger,fear and rage [Excerpt]. In R. Solomon (Ed.), What isan emotion? (2nd ed.). New York: Oxford UniversityPress. (Original work published 1929)

Charland, L. (2002). The natural kind status of emo-tion. British Journal for the Philosophy of Science,53, 511–537.

Churchland, P. S. (1986). Neurophilosophy. Cambridge,MA: MIT Press.

Damasio, A. (1994). Descartes’ error. New York:Putnam.

Dennett, D. (1978). Brainstorms. Cambridge, MA: MITPress.

Dennett, D. (1991). Consciousness explained. Boston:Little, Brown.

Descartes, R. (1989). On the passions of the soul (S.Voss, Trans.). Indianapolis, IN: Hackett. (Originalwork published 1649)

de Sousa, R. (1987). The rationality of emotion. Cam-bridge, MA: MIT Press.

Dewey, J. (2003). The theory of emotion [Excerpt]. InR. Solomon (Ed.), What is an emotion? (2nd ed.).New York: Oxford University Press. (Original workpublished 1894)

Ekman, P. (1973). Darwin and facial expression. NewYork: Academic Press.

Ekman, P. (2003). Emotions revealed. New York: TimesBooks.

Ekman, P., & Davidson, R. J. (Eds.). (1994). The natureof emotion. New York: Oxford University Press.

Ellsworth, P. (1994). William James and emotion: Is acentury of fame worth a century of misunderstand-ing? Psychological Review, 101, 222–229.

Elster, J. (2000). Alchemies of the mind. Cambridge,UK: Cambridge University Press.

Freud, S. (1935). The unconscious (C. M. Baines,Trans.). In Essays in metapsychology. London:Liveright. (Original work published 1915)

Frijda, N. (1986). The emotions. Cambridge, UK: Cam-bridge University Press.

Gibbard, A. (1990). Wise choices, apt feelings: A theoryof normative judgment. Cambridge, MA: HarvardUniversity Press.

Goldie, P. (2000). The emotions: A philosophical explo-ration. Oxford: Clarendon Press.

Gordon, R. M. (1987). The structure of emotion. Cam-bridge, UK: Cambridge University Press.

Greenspan, P. (1988). Emotions and reasons. NewYork: Routledge.

Griffiths, P. (1997). What emotions really are. Chicago:University of Chicago Press.

Hamlyn, D. W. (1978). The phenomenon of love andhate. Philosophy, 53, 5–20. Includes a discussion ofBrentano’s theory.

Hegel, G. W. F. (1977). The phenomenology of spirit (A.N. Miller, Trans.). Oxford: Oxford University Press.(Original work published 1807)

Heidegger, M. (1962). Being and time. New York:Harper & Row. (Original work published 1927) Seealso the following explications of Heidegger: C.Guignon (1984). Moods in Heidegger’s Being andtime; H. Dreyfus, (1991). Being-in-the-world: Acommentary on Heidegger’s Being and time. Cam-bridge, MA: MIT Press.

Hobbes, T. (1994). Leviathan. Indianapolis, IN:Hackett. (Original work published 1651)

Hume, D. (1888). A treatise of human nature (L. A.Selby-Bigge, Ed.). Oxford: Oxford University Press.(Original work published 1739) See also A. Baier(1991). A progress of sentiments: Reflections onHume’s treatise. Cambridge, MA: Harvard Univer-sity Press. And see also D. Davidson (1976). Hume’scognitive theory of pride. Journal of Philosophy, 73,733–757.

Husserl, E. (1960). Cartesian meditations (D. Cairns,Trans.). The Hague: Nijhoff. (Original work pub-lished 1938)

Hyman, A., & Walsh, J. (1973). Philosophy in the Mid-dle Ages. Indianapolis, IN: Hackett.

James, W. (1884). What is an emotion? Mind, 9, 188–205.

James, W. (1950). Principles of psychology. New York:Dover. (Original work published 1890)

Kant, I. (1953). Critique of judgment (J. H. Bernard,Trans.). New York: Hafner. (Original work published1793)


Kenny, A. (1963). Action, emotion and will. London:Routledge & Kegan Paul.

Lane, R. (1999). The cognitive neuroscience of emotion.New York: Oxford University Press.

Lazarus, R. S. (1994). Emotion and adaptation. NewYork: Oxford University Press.

LeDoux J. (1996). The emotional brain. New York: Si-mon & Schuster.

Lyons, D. (1980). Emotion. Cambridge, UK: CambridgeUniversity Press.

Merleau-Ponty, M. (1994). The phenomenology of per-ception. (C. Smith, Trans.). New York: Routledge.(Original work published 1962)

Neu, J. (1977). Emotion, thought and therapy. Berkeley:University of California Press.

Neu, J. (1980). Jealous thoughts. In A. Rorty (Ed.), Ex-plaining emotions. Berkeley: University of CaliforniaPress.

Neu, J. (1999). A tear is an intellectual thing. New York:Oxford University Press.

Nietzsche, F. (1967). On the genealogy of morals (W.Kaufmann, Trans.). New York: Random House.(Original work published 1887)

Nissenbaum, H. (1985). Emotions and focus. Stanford,CA: CSLI.

Nussbaum, M. (1994). The therapy of desire. Princeton,NJ: Princeton University Press. A sympathetic anddetailed discussion of the Stoics on the place of emo-tion in life.

Nussbaum, M. (2003). Upheavals of thought. Cam-bridge, UK: Cambridge University Press.

Ortony, A., Clore, G., & Collins, A. (1988). The cogni-tive structure of emotions. New York: CambridgeUniversity Press.

Plato. (1974). The republic (E. M. A. Grube, Trans.). In-dianapolis, IN: Hackett.

Plato. (1989). The symposium (A. Nehamas & P. Wood-ruff, Trans.). Indianapolis, IN: Hackett.

Prinz, J. (2004). Gut reactions. New York: Oxford Uni-versity Press.

Ricoeur, P. (1966). The voluntary and the involuntary(E. Kohak, Trans.). Evanston, IL: Northwestern Uni-versity Press. (Original work published 1950)

Roberts, R. (2003). Emotions. New York: CambridgeUniversity Press.

Robinson, J. (1995). Startle. Journal of Philosophy,92(2), 53–74.

Robinson, J. (2005). Deeper than reason. New York:Oxford University Press.

Rorty, A. (1980). Explaining emotions. In A. Rorty(Ed.), Explaining emotions. Berkeley: University ofCalifornia Press.

Russell, B. (1967). The autobiography of Bertrand Rus-sell (Vol. 1). Boston: Little, Brown.

Ryle, G. (1951). The concept of mind. New York:Barnes & Noble.

Sartre, J.-P. (1948). The emotions: Sketch of a theory (B.Frechtman, Trans.). New York: Philosophical Li-brary. (Original work published 1939) See the com-

mentary in R. C. Solomon (1988). From Hegel to ex-istentialism. Oxford: Oxford University Press.

Sartre, J.-P. (1956). Being and nothingness (H. Barnes,Trans.). New York: Washington Square Press. (Origi-nal work published 1943) See also J. Fell (1965).Sartre’s theory of the passions. New York: ColumbiaUniversity Press.

Schachter, S., & Singer, J. (1962). Cognitive, social andphysiological determinants of emotional state. Psy-chological Review, 69(5), 379–399. For a goodphilosophical rejoinder, see R. M. Gordon (above),Ch. 5.

Scheler, M. (1970). The nature of sympathy. New York:Archon. (Original work published 1916)

Searle, J. (1983). Intentionality: An essay in the philoso-phy of mind. Cambridge, UK: Cambridge UniversityPress.

Seneca. (1963). De ira. Oxford: Loeb Classical Library,Oxford University Press.

Smith, A. (1976). Theory of the moral sentiments. Ox-ford: Oxford University Press. (Original work pub-lished 1759)

Smith, A. (1976). An inquiry into the nature andcauses of the wealth of nations. Indianapolis, IN:Liberty Classics. (Original work published 1776)For a good study of the relation between Smith’sethics and his economic theory, see P. Werhane(1991). Ethics and economics: The legacy of AdamSmith for contemporary capitalism. Oxford: Ox-ford University Press.

Solomon, R. (1976). The passions. Notre Dame, IN:University of Notre Dame Press. See also R. Solomon(1988). About love. New York: Simon & Schuster;R. C. Solomon & K. Higgins (Eds.). (1991). The phi-losophy of (erotic) love. Lawrence: University ofKansas Press.

Solomon, R. (Ed.). (2003). What is an emotion? (2nded.). New York: Oxford University Press.

Solomon, R. (2004). Not passion’s slave. New York:Oxford University Press.

Sorabji, R. (2003). Emotion and peace of mind. NewYork: Oxford University Press.

Spinoza, B. (1982). Ethics. (S. Shirley, Trans.). Indianap-olis, IN: Hackett. (Original work published 1677)See also A. Rorty (1991). Spinoza on the pathos oflove. In R. C. Solomon & K. Higgins (Eds.), The phi-losophy of (erotic) love. Lawrence: University ofKansas Press.

Stocker, M. (1996). Valuing emotions. Cambridge, UK:Cambridge University Press.

Thalberg, I. (1977). Perception, emotion and action.New Haven, CT: Yale University Press.

Walton, K. (1990). Mimesis as make-believe. Cam-bridge, MA: Harvard University Press.

Williams, B. (1973). Morality and the emotions. InProblems of the self. Cambridge, UK: CambridgeUniversity Press.

Wittgenstein, L. (1953). Philosophical investigations.London: Routledge & Kegan Paul.


C H A P T E R 2

History of EmotionsIssues of Change and Impact

PETER N. STEARNS

Charting changes in emotion raises an intrigu-ing set of analytical challenges, and at the sametime yields vital understandings of emotion it-self. Here is the core of the history of emotionsas a field, as increasingly widely practiced byhistorians and kindred sociologists and anthro-pologists. A specific example will illustrate theclaim, highlighting the main features of the an-alytical strategy, which can then be explored ina more generalized fashion.

Studies of emotional emphases in WesternEurope in the wake of the Protestant Reforma-tion have highlighted the pervasiveness of anatmosphere of melancholy—held to be an ap-propriate religious demeanor, given the snaresof this world and the ravages of sin. Paintingsfrequently seized on melancholy, and diaryentries characteristically portrayed dolefulnessand grief. This tone began to change, however,with the 18th century, when what can now beseen as a dominant modern Western insistenceon cheerfulness emerged. Christian values weregradually recast, with a smiling demeanor heldto be an appropriate tribute to a beneficentGod. Commercial signals gained new strength,

with material goods seen as producing hap-piness, and a cheerful disposition as mostsuitable for successful business transactions. Asone historian has noted, even improved den-tistry entered in, with people more willing tosmile openly as the quality of their teeth im-proved during the century of the Enlighten-ment (Jones, 1996).

Once launched, the new insistence on cheer-fulness gained ground steadily, particularly (asmany European observers noted) in the UnitedStates. By the late 20th century it affected laborrelations as well as commercial interactions,with workers expected to display a sunny dis-position as proof of their employability. Mediaextended the emphasis on cheer, and opera-tions like Disney World—where trams from theparking lots still pipe in messages about howhappy everyone is—showed the ubiquity of thenorm. Cheerfulness also affected cross-culturalcontacts: Franklin Roosevelt, at Yalta, was dis-concerted when he could not make the dourStalin smile, but later, when McDonald’s beganto penetrate the former Soviet Union, Ameri-can standards of grinning service pervaded the

17

training standards, imposed, against local cus-tom, on Russian sales personnel. PervasiveAmerican cheerfulness had its obvious down-side, however, imposing constraints that un-doubtedly helped explain comparatively highrates of depression in the population as a whole(Kotchemidova, 2005).

Here, in a thumbnail sketch, is the historicalcontribution to emotions research at its best:capable of identifying a significant new trendand tracing its evolution; working to determinethe relationship between cultural standards(where change can be readily charted) and ac-tual personal emotional experience; providingat least strong indications of causation, in thiscase associated with cultural shifts (the En-lightenment) and the emergence of consumer-ism; and demonstrating impact, both in publicspheres such as politics or labor relations(where emotional standards have obvious con-sequences) and in more private life (wheresome of the limitations of insistent cheerfulnessmay be particularly clear). In this instance—and there are several different historical scenar-ios, depending on the emotion in question—origins of novel emotional standards go wellback in time, but historical analysis shows theintensification and widening ramifications ofthese standards as they continue to shape con-temporary emotional life. Arguably—and thisis the key point in relation to emotions researchmore generally—the contemporary Americanemphasis on cheerfulness simply cannot beadequately understood without the historicalcontribution.

Historical work on emotions as an explicitsubfield is now two decades old as and contin-ues to expand rapidly, dealing with processes ofchange in emotional standards and emotionalexperience, or, somewhat more complexly,with emotional continuities amid changingcontexts. Historians may also be interested in athird focus: seeking to grasp the characteristicemotional styles of a particular period, in andof themselves, as a means of both enriching theportrayal of that past time and launching theprocess of comparing one previous period toanother. Ultimately, however, the analyticalgoals center on change, either in emotionsthemselves or in the environments in whichthey operate. Here, correspondingly, is the cen-tral justification for adding history to the list ofdisciplines seriously engaged in emotions re-search. For if emotions change in significantways—and historians and others have conclu-

sively demonstrated that they do, as in the caseof modern cheerfulness—then the process mustbe grappled with as part of evaluating emo-tional expressions even in the present time.Adding change to the variables involved inemotions research means adding complexity,but it is empirically inescapable and providesan essential perspective for assessing the resultsof other social science research on emotion,such as the findings emanating from sociologyand anthropology.

Examining change involves establishingbaselines, so that new trends can be carefullyevaluated against real, rather than assumed orimagined, past standards—a task that histori-ans at their best handle quite well. It involvesassessing the causes of change and also its re-sults, in personal emotional lives but also inlarger institutions (such as law, education, themedia, or the political process).

THE DEVELOPMENTOF EMOTIONS HISTORY

Explicit historical research is still a relativenewcomer. Theorists from other social sciencedisciplines provided frameworks for historicalassessment that were long ignored by most his-torians themselves. Thus Norbert Elias’s (1938/1982) classic work on how new levels of “civi-lization” began to constrain more spontaneousemotional expressions, beginning with theWestern European aristocracy by the 17th cen-tury, focused attention on a key turning pointnow being widely explored. More generally,the constructivist theory of emotion, generatedby several social psychologists as well as sociol-ogists, argues that emotions should be inter-preted primarily in terms of the social functionsthey serve; constructivists also correctly notethat as social functions frequently change,emotions will shift substantially as well(Averill, 1980, 1982). Some emotions may dis-appear as part of this process, and others maynewly emerge. Here is a richly suggestivehistorical framework—only rarely, however,fleshed out by detailed historical research, anduntil recently largely ignored by professionalhistorians. The constructivist approach, includ-ing the common attention to cultural contextas a functional area or as an intermediary be-tween function and emotion, independently af-fecting ideas about emotional experiences andthe vocabularies used to phrase them (Gordon,


1989), in fact tallies closely with recent histori-cal work.

Historians themselves moved into researchon emotion hesitantly. The great French socialhistorian Lucien Febvre called over 70 yearsago for a “historical psychology” that would“give up psychological anachronism” and “es-pecially detailed inventory of the mental equip-ment of the time” (Febvre, 1933/1973). His ap-peal was not quickly heeded. A number ofcultural historians dealt with past styles andrituals that had strong emotional components.Johann Huizinga’s (1927) masterful portrayalof the late Middle Ages contained a wealth ofdata relevant to emotions history, and evenmore limited studies of popular protest or reli-gious life offered important emotional insightinto the past periods involved (Stearns withStearns, 1985). Explicit focus on emotion,however, was lacking. Most historians contin-ued to emphasize the conscious actions and ra-tional decisions of their subjects. Even the ad-vent of social history, bent on detailing theactivities and interests of groups of ordinarypeople, did not quickly break this mold. In-deed, it could confirm it, as social historians in-sisted on rescuing ordinary people from accu-sations of mob impulsiveness and so stressedtheir transcendent rationality (e.g., in protestsituations). Ordinary people may have mentalworlds different from those of elites, accordingto the pioneer social historians, but they are noless careful choosing methods appropriate totheir goals. Emotion, in this formulation, wasnot a significant variable.

The advent of psychohistory in the 1960sbrought attention to the role of emotions in thepast, but on a very limited scale. Mostpsychohistorians, from the great Erik Erikson(1958) to more recent practitioners, have con-centrated on biography and have utilized alargely Freudian theoretical framework. Theyhave linked emotional characteristics to histor-ical developments—thus Erikson translates Lu-ther’s tense relationship with his harsh fatherinto Lutheranism’s preoccupation with an an-gry and omnipotent God—but they have notdealt with emotional change, and they havetended to enmesh emotional factors in a rigidand unchanging psychodynamic. Furthermore,while psychohistorians generated interestingwork, their approach never won wide accep-tance within the historical discipline, and (be-cause of pervasive Freudianism) had a limitedreception in other fields as well. (See, however,

Gay, 1984–1986, for a more ambitious psycho-historical treatment of the emotional lives andstandards of the Victorian bourgeoisie.)

The direct antecedents of historical researchon emotion, however, awaited a final ingredi-ent, provided by the 1970s through the matu-ration of social history as the leading branch ofhistorical inquiry. By this point, historiansthroughout the United States and Western Eu-rope increasingly focused not only on the activ-ities and value systems of ordinary people, butalso on institutions and behaviors in additionto formal politics, as the central stuff of thepast. New topics meant new materials, andalso promoted the analysis of change as thedominant mode of historical presentation, dis-placing the mere narration of political and mili-tary events.

From social history, in turn, the issuesemerged that led a number of historians to con-sider emotional patterns as central to their taskand that produced increasing confluence withother disciplines dealing with the social con-texts of emotional life. Social historians inevi-tably developed a strong interest in familyhistory. Initially they focused on “objective”features of family organization—size, house-hold composition, marriage age, and the like—where indeed important changes could betraced. Quickly, however, concern about theemotional quality of family relationships be-gan to shape research agendas. Discussion ofaffective parental ties with children followed,for example, from analysis of the impact ofchanges in family size. A general linkageemerged between reductions in birth rate andgreater affectionate intensity between parentsand individual children, although which camefirst was (and is) not always easy to discern.Other aspects of household composition re-lated to emotional factors. When the propertypower of older family members began to de-cline, in a more commercial economy in whichindependent jobs for younger adults becamemore abundant, affective links between youngadults and older parents might well improve.Finally, efforts to explain changes in marriagepatterns—in rates of marriage, ages at mar-riage, and age ratios between partners—generated attention to the emotional implica-tions of courtship behaviors and subsequentspousal relationships. (For a solid survey of thefamily history field, see Mintz & Kellogg,1989; see also Gillis, 1995.) The social historyof emotion—the effort to trace emotional

2. History of Emotions 19

norms in groups of relatively ordinary peopleand their impact on key institutions of dailylife—was born above all from the progressiveextension of family history.

By the late 1970s, various studies were di-rectly confronting the emotional aspects offamily history. Historians working on France,Britain, Germany, and colonial North Americauncovered a pronounced increase in familial af-fection in the late 17th and 18th centuries, con-trasting with the more restrained emotionaltone seemingly characteristic of families in ear-lier centuries. John Demos (1970), a historiandealing with colonial New England, noted aneffort in 17th-century Plymouth to keep fami-lies free from the angry bickering still toleratedamong neighbors; this effort to control angerwas accompanied by the encouragement ofconjugal love. European families may have tol-erated outbursts of anger as part of appropriatefamily hierarchy for a slightly longer time—it ispossible that in the unsettled conditions of thecolonies, preservation of family harmonyproved particularly important in NorthAmerica—but a similar evolution set inthroughout much of Western Europe by the18th century (Stone, 1977; Flandrin, 1979;Trumbach, 1978; Shorter, 1975). Child-rearingmethods that had focused on breaking chil-dren’s wills, reflecting parental anger at animal-like offspring and generating intense if neces-sarily repressed anger in turn, yielded to greaterreliance on affectionate persuasion, althoughthe change was gradual and uneven. Mothersbegan to be defined as central ingredients of thenetwork of familial affection. Romantic lovebegan to influence courtship and marital ex-pectations; the absence of love even served, bythe 18th century, as a valid reason for the disso-lution of engagements. On the eve of its declineas an economic unit, the family began takingon important new emotional functions andbegan to generate new expectations (Leites,1986). Although the rise of various kinds oflove headed the innovation list (Stone, 1977),other emotions entered in. Here was the con-text in which 18th-century family manuals be-gan to urge repression of anger within the fam-ily, particularly enjoining men to treat theirwives, children, and servants with appropriatedecorum.

While the expansion of family history beganto introduce emotional change as an explicithistorical topic—indeed, a central issue in deal-ing with the rise of new kinds of family rela-

tionships in the 17th and 18th centuries—another kind of social history promoted atten-tion to other emotional issues. Here French his-torians led the way, in contrast to Anglo-Saxondominance in the pioneering family historystudies. A field of “mentalities” researchemerged, focusing on deeply held popular be-liefs about self, environment, and society,which were expressed more frequently in ritualbehavior than in formal declarations of princi-ple. Historians of mentalities probed what or-dinary people really meant by their religiousobservances, often discovering that beneath aChristian veneer a variety of magical ingredi-ents still held sway. Emotional beliefs, or emo-tional components of other beliefs, increas-ingly engaged this field of inquiry. RobertMuchembled (1985) and Jean Delumeau(1978, 1989) emphasized the high level of fearcharacteristic of French peasants from the Mid-dle Ages to the 18th century, expressed in vari-ous religious and magical practices and in festi-val rituals. Anxiety about death, about cropfailure, and about violence generated intensecommunity practices that might relieve fears ofthe outside world. Delumeau, in particular,painted a picture of popular religion domi-nated by the need to control constantly over-spilling fear. Delumeau also argued, however,that as with family emotion, the 18th centurysaw a pronounced change in popular emo-tional life: Growing confidence about measuresthat could control the natural and social envi-ronment reduced the need for fear-managingrituals, leading to a shift in religious emphasisand a redefinition of fear that (in a processDelumeau did not himself trace) would lead ul-timately to the 20th-century formulation offear as an interior emotion focused on inwarddemons. Mentalities historians also dealt withrelationships between elite and popular beliefsystems. Here too they emphasized a signifi-cant change opening up in the 17th and 18thcenturies. Elite Europeans began to lookaskance at a leisure tradition in which they hadonce willingly shared. A key focus of their dis-may was emotional spontaneity—those occa-sions when emotion generated physical behav-iors, such as crowd frenzy, ribald dances, orexuberant sports, that now seemed both vulgarand disorderly. Correspondingly, the elitelaunched a variety of disciplinary and legalmeasures designed to curb spontaneity, andwon some success in denting the traditional fes-tivals of European peasants and artisans


(Burke, 1978; Mitzman, 1987). A historiandealing with colonial Virginia has subsequentlytraced a somewhat similar process of elite–mass divergence over emotional spontaneity inleisure, taking shape in the later 18th century inparts of North America (Isaac, 1982).

In various ways, in sum, analysis of emo-tional change and its impact had become ines-capable by the late 1970s. Without launching aspecific subfield concentrating on the history ofemotions, social historians were vigorously en-gaged in dealing with several facets of emo-tional change, with familial emotions, fear, andspontaneity heading the list. A number of topi-cal inquiries, initially directed toward other is-sues (e.g., family life or leisure), pointed con-clusively both to the existence of substantialemotional change in the past and to the impor-tance of this change in grasping key passages insocial history. Attention centered particularlyon the early modern period, with analysesof the several different ways Europeans andNorth Americans were changing their emo-tional rules or seeing these rules changed dur-ing the 17th and 18th centuries. The link be-tween these findings and Elias’s earlier modelof increasing civilized restraint was not imme-diately drawn, but it soon added a theoreticalingredient to the emerging picture.

A MATURING FIELD

The history of emotions emerged as an explicitand increasingly polished research area duringthe 1980s. Earlier findings continued to vivifythe field, but a number of features were addedor redrawn. In the first place, growing numbersof social historians began dealing with the his-tory of emotions in and of itself, rather thanas an adjunct to family or mentalities study.Changes in a particular emotion, and the rela-tionship between these changes and other as-pects of a historical period, began to constituterespectable (if still clearly innovative) historicaltopics. Social historians also expanded the listof emotions that could be subjected to histori-cal scrutiny. Along with love and fear, anger,envy, jealousy, shame, guilt, grief, disgust, andsadness began to receive significant historicalattention, and interest in augmenting the rangeof emotions considered as part of research onhistorical change continues strong. Significantwork more recently has addressed not onlycheerfulness but also nostalgia.

The contexts in which emotional changecan be explored have also been elaborated.Predominant attention, particularly in Anglo-American research, continues to go to familysettings and related emotional socialization,but studies of emotional change have now dealtwith workplace relations, religion, leisure andits emotional symbolism, and legal standardsand uses of the law to reflect new emotionalnorms. A growing body of work also dealswith the intriguing relationships between emo-tions and modern consumerism.

In addition, the maturation of emotions his-tory has involved growing recognition of aneed to modify some of the impulses towardreifying stark contrasts that characterizedmuch of the initial work. Premodern families,for example, are no longer seen as emotionallycold. Affection for children is not a modern in-vention, nor—despite the fascinating argumentof a feminist French historian (Badinter,1980)—is mother love. Recognition of somebiological constants in emotional expressions,and simply more extensive data probes,has modified the earlier picture or sharppremodern–modern emotional dichotomies.Better use of theory has come into play. Histo-rians using Jerome Kagan’s (1979) findings onchild rearing can understand that evidence ofsevere physical discipline in the past, oncetaken as a sign of emotional distancing, is infact compatible with real affection. Change hascontinued to organize historical research onemotion, but change is now seen as more subtlethan it was previously regarded as being. Forexample, one study intriguingly complicatesthe emotional history of the European MiddleAges, contesting undue emphasis on unre-strained anger, and in the process chal-lenging key assumptions of the Elias school(Rosenwein, 2006).

The process of reassessing initial overstate-ments generated some interesting byways.Some revisionists began to argue that certainemotional relationships do not change; LindaPollock (1983), most notably, tried to demon-strate that European parents manifested consis-tent love for their children from the 16th cen-tury onward, though in fact her evidenceclustered around 1700. In another importantvariant, Philip Greven (1977; see also the laterextension of the argument, Greven, 1991) pos-ited three basic emotional socialization styles incolonial North America, which have since per-sisted; change was involved in establishing the


initial variety, but thenceforward continuityhas prevailed. Angry parents in the 1990s havebeen trapped in the same culture that generatedtheir predecessors in 1750. These approacheshave not captured dominant historical atten-tion, but they have compelled more sophisti-cated treatment of change.

Love offers a clear illustration of the currentapproach. Historians now realize that their ini-tial impulse to contrast economically arrangedmarriages with modern romance was overlysimple. Economics remains a factor in modernlove, and love entered into premodern court-ship. The nature and experience of love weredifferent, however. Love in 17th-century West-ern Europe was less intense and less individu-ally focused than would become the norm inthe 18th and 19th centuries. The system of ar-ranged marriage led to groups of young menand women stimulating each other emotion-ally, for an individual could not be singled outprior to final arrangement. The 18th-centurydecline of arranged marriages cut into thegroup-oriented experience of premarital excite-ment; this shift soon led to an unprecedentedassociation of love with privacy and with one-on-one intensity. (For an intriguing study ofchanges and in emotional expectations in 18th-century courtships, see Eustace, 2001.) Finally,expressions of love pulled away from a tradi-tional range of vigorous bodily manifestations.Suitors in Wales stopped urinating on theirfiancées’ robes as a sign of affection; kissing be-came gentler, biting far less common. The rela-tionship of love and the body, in other words,changed substantially (Gillis, 1985; Leites,1986; Stearns & Stearns, 1988). This meansthat a new definition of love—a modern kindof romantic love—did indeed emerge in the late17th and 18th centuries. The significance of thechange is, if anything, enhanced by its fullerdefinition, even if the complexity increases aswell. Similar modifications of initial generaliza-tions about grief (e.g., over infant death) andparent–child affection have generated moresubtle, but also richer, definitions of whatemotional change entails (Lofland, 1985;Rosenblatt, 1983). A crucial part of this in-creased sophistication has resulted from histo-rians’ growing recognition of distinctionsbetween emotional standards—the “feelingrules” or emotionology that describes sociallyprescribed emotional values, and often the cri-teria individuals themselves use to evaluatetheir emotional experience—and emotional ex-

perience itself. Both topics are important, butthey are not the same. The rise of official ap-proval of love in courtship and marriage is gen-uinely significant—it began to influence legalreactions to marital distress, for example(Griswold, 1986)—but it is not the same thingas a rise of the experience of love. The actualexperience may have changed less, or at leastdifferently, than the new standards imply. His-torians of emotion still try to deal with both as-pects of their subject, but in distinguishingbetween culture and experience they greatlyimprove their precision.

Finally, maturation of emotions history hasinvolved increasing interaction between histo-rians and other scholars working on the socialcontext of emotion. The revival of attention toemotions research in sociology brought new in-terest in the issue of emotional change fromthis camp. North American sociologists andsocial psychologists dealt with a number ofchanges in emotional standards in the 20thcentury, using many of the same materials his-torians themselves relied upon (Cancian, 1987;Shields & Koster, 1989; Cancian & Gordon,1988). More recent additions to the list of soci-ologists’ contributions to emotional history in-clude the work on cheerfulness, and also an im-portant study of gender, shyness, and change(McDaniel, 2003). European sociologists, par-ticularly in the Netherlands, have taken asomewhat longer view. Relying heavily on theElias (1938/1982) framework, they have dealtwith new forms of emotional control in earliercenturies and particularly have tried to place20th-century patterns of emotional manage-ment, including a new informality, in the con-text of earlier shifts (de Swaan, 1981; Wouters,2004). Systematic interaction between emo-tions historians and psychologists remains lim-ited, but the rise of a minority school, deeplycommitted to exploring the cultural construc-tion of mental states, provides one vigorousconnection. Interdisciplinary concern for therelationship between language and emotion,including linguistic change, is another connec-tion that involves psychology, though also cul-tural studies and other social sciences (Harré &Stearns, 1995; Gergen, 1998). Finally, althoughmost anthropologists dealing with emotionscontinued to focus on durable cultural tradi-tions, several major studies, such as RobertLevy’s (1973) work on Tahiti, also treated al-terations in emotional expression under the im-pact of such changes as missionary contact.


Historians, for their part, have become moreaware of relevant work in other fields.

ISSUES OF RANGE AND OF THEORY

Growing interest in the history of emotions hasgenerated research in a variety of historical ar-eas, though from a disciplinary point of viewimportant imbalance remains both chronologi-cally and geographically. A handful of classicalhistorians have dealt with emotion in antiquity.The medieval field is richer. A substantial liter-ature exists on the rise and subsequent impactof chivalric love. A 1997 presidential addressto the American Historical Association, by aleading medievalist, specifically invoked theimportance of changes in emotional formsbetween medieval and early modern Europe(Bynum, 1997). The study of medieval angeropens this period to new complexities, from thestandpoint of emotions history (Rosenwein,2006). But further work on earlier periods re-mains highly desirable, if only to provide fullerperspective on modern trajectories.

Geographic gaps remain even more trou-bling, and (in contrast to some purely contem-porary work, such as Salovey’s [1991] volumeon jealousy) there is almost no explicitly com-parative historical analysis where emotions areconcerned. Mark Elvin and others have exam-ined emotions history in China (Elvin, 1989,1991). A history of Maoism notes the Chair-man’s appeal for emotional reconfiguration onthe part of peasants, toward releasing the angernecessary to fuel revolution (Solomon, 1971).Advances in the history of childhood, particu-larly in China and Latin America, also have im-plications for emotions history, and there ispromise of further gains in the future. But thechallenge of extending explicit work on emo-tional change to non-Western areas and tocomparative treatment remains compelling.

Theoretical work may also have laggedsomewhat, aside from continuing utilization ofthe Elias framework or ideas about informal-ization. There are theoretical challenges inabundance, but with a few exceptions histori-cal advances from the 1990s onward have em-phasized empirical case studies. Only WilliamReddy’s (2001) ambitious work stands out,seeking to apply psychological and social sci-ence theory to his own framework, applied inturn to emotional changes and their politicaland legal consequences during the French revo-

lutionary period. Various American historianshave also speculated about the special role ofpsychology in shaping national emotional life(Pfister & Schnog, 1997)—an inquiry that hasboth theoretical and comparative implications.Finally, the increasing clustering of studies ofconsumerism and emotions sets the basis formore extensive theoretical statements. Heretoo, however, opportunities have not been fullyrealized, and the gaps also affect the relation-ship between historical findings and work inother fields such as psychology.

KEY CHANGESOF WESTERN SOCIETY

Significant historical findings do already exist,providing both basis and model for additionalwork and for interdisciplinary collaborations.The richest literature, steadily expanding, ap-plies to modern Western history (Western Eu-rope and the United States), beginning about1500 and extending to the present. Three peri-ods command primary attention and generatethe most extensive findings, for significantemotional change emerges in sporadic bursts,not consistently in every decade.

Historical research continues to embellishthe picture of a fundamental transformation inemotional standards during the early moderncenturies, and particularly the 17th and 18thcenturies. Imaginative research on the Germanpeasantry has even discerned some symptomsof emotional change, toward fuller identifica-tion of an emotional self, in the century after1500 (Sabean, 1984). In addition to refine-ments in the understanding of changes in pa-rental and marital love, and to the ongoingwork on fear and spontaneity, historians havemade a number of other changes to the earlymodern transformation model. John Demos(1988), looking at New England in the colonialand early national periods, traces a shift frompervasive use of shame in dealing with childrenand miscreant adults to guilt, from the 18th tothe early 19th century. As community cohesiondeclined, parents had to find new ways to inter-nalize behavioral guidelines; they were able touse newly intense love as the basis for instillinga greatly heightened level of guilt. A compara-ble shift, toward guilt rather than public sham-ing, describes innovations in the principles ofsocial discipline and criminal justice in thesame period. The work on the balance between


cheerfulness and acceptable sadness adds an-other important dimension. On yet anotherfront, Alain Corbin (1986) describes a vasttransformation in the emotion of disgust: AsFrenchmen from the 18th century onward be-gan to manifest intense disgust at a new rangeof objects, they began using the emotion to mo-tivate a variety of new sanitary and cosmeticbehaviors and to justify new social distinctionsbetween the washed and the unwashed. Hereemotional change is directly linked to alteredexperience of the senses. This kind of emotionsresearch, linked to the mentalities approach,has expanded into historical inquiries aboutchanges in gestures and humor (Bremmer &Roodenburg, 1997).

Research on 19th-century emotions historyhas become increasingly active, though it lacksthe focus of the early modern framework. Tosome extent, the very notion of a great trans-formation in the 17th and 18th centuries over-shadows findings on the 19th century, as manydevelopments served to amplify and dissemi-nate to new social groups the basic trajecto-ries established earlier. Yet amplification cancarry important new messages. North Ameri-can studies on the apotheosis of mother love gowell beyond 18th-century findings concerningnew expectations for parental affection (Lewis,1989), and the standards applied to children, interms of anticipated emotional reward, esca-lated as well (Zelizer, 1985). Several importantstudies on the 19th-century version of romanticlove similarly point to novel and distinctive fea-tures (Stearns, 1989; Lystra, 1989). Love be-came for many in the American middle classesvirtually a religious ideal, involving self-ab-negation and worshipful devotion to the other;18th-century standards had not sought somuch. Jealousy was redefined in this process,as a largely female emotion and a contradictionof proper selflessness in love; older ideas ofjealousy in defense of honor fell by the way-side. Grief gained new attention and vast newsymbolic expression in Victorian funeral prac-tices (Houlbrooke, 1989). Anger received moreexplicit condemnation, particularly in the fam-ily setting. Gender distinctions urged total sup-pression of anger in women, but an ability tochannel anger toward competition and righ-teous indignation in men (Stearns & Stearns,1986).

Victorian emotional patterns thus providedno overall new direction, but they did adjust

prior trends to the new sanctity of the family inan industrial world; to new social class divi-sions; and to the new need to define emotionaldistinctions between boys and girls, men andwomen. Whereas love was seen as uniting menand women in common emotional goals, nega-tive emotions became highly gender-linked; inaddition to jealousy and anger, the conquest offear was redefined to serve purposes of genderidentity (Stearns, 2006). Although the primaryfocus has been on the new standards urged in19th-century Western society and within themiddle class, spelled out in a surge of newkinds of prescriptive literature, various evi-dence suggests considerable behavioral impact.Men and women did have, with some fre-quency, the kind of love experiences now rec-ommended, while adjusting their premaritalsexuality accordingly; they did work towardappropriate training of children concerning an-ger and fear. Still not entirely defined, the 19thcentury stands as a rich source of materials onemotional history and as the scene of a numberof significant modifications in norms and expe-rience.

The 20th century, finally, has receivedsharper definition from the historical perspec-tive. Several emotions historians, to be sure,trace a variety of oscillations in 20th-centurystandards without an overarching theme. Someanalysis concentrates on the need to refute fac-ile modernization ideas that urge (against virtu-ally all available findings) that the 20th centuryshould be seen simply in terms of increasingopenness of emotional expression, as older re-pressions have gradually fallen away. Signifi-cant work also stresses continuities from the19th century, particularly in the gender-linkedquality of certain emotional standards; women,for example, continue to be held to a particu-larly self-sacrificing image of love, even as malestandards may have changed (Cancian, 1987).A few historians have argued for continuitypure and simple, from the 19th or even the18th century (Kasson, 1990; Flandrin, 1979).

Two related approaches focus a number ofthe most important current findings about the20th century. The first, emanating particularlyfrom several Dutch sociologists, grapples withthe problem of growing emotional informalityand apparent liberalization, in a context thatcontinues to insist on a great deal of self-control. Spontaneity has revived, but withinstrict (if unacknowledged) limits. The general


argument is that most Westerners have learnedso well the lessons of restraint of violence andof unwanted sexuality that they can be alloweda good bit of informal emotional idiosyncrasyas part of personal style. Rules of emotional ex-pression have become more complex, and judg-ments are made about appropriate emotionalpersonalities on the basis of various individualinteractions rather than rigid and hierarchi-cal codes (de Swann, 1981; Wouters, 1992;Gerhards, 1989). Increasing democratization isalso part of this shift, as emotional standardsused in the 19th century to separate respectableand unrespectable classes are now more widelyenjoined (Wouters, 2004). The second ap-proach similarly urges that the 20th centuryconstitutes a period of considerable and rea-sonably coherent change in emotional stan-dards. It focuses on implicit attacks on 19th-century emotional formulas, beginning in the1920s and extending over a transition periodof several decades (Stearns, 1994). Hostility tonegative emotions has increased. Gender link-ages, though by no means absent, have beenmuted in favor of more uniform standards ofemotional control. The importance of manag-ing emotions through talking out rather thanactive expression has become a dominanttheme. Reference to embarrassment in frontof others has come to supplement guilt andshame as enforcement for emotional normality(Stearns & Stearns, 1986; Stearns, 1989).Amid a host of specific changes, including newemphasis on avoiding rather than masteringfear as part of building character, the dominanttheme is a new aversion to undue emotional in-tensity (Shields & Koster, 1989; Stearns &Haggerty, 1991). The decline of the acceptabil-ity of open grief is a key index to the new emo-tional regime. Even good emotions have dan-gers; earlier icons of intense emotion, such asthe idealized Victorian mother or fervid roman-tic love, have come in for substantial criticism.In this context, Arlie Hochschild (1997) hasnoted how work sometimes provides an easieremotional climate than home life. Also vital isthe relationship between increasingly overtsexuality and emotional formulations; sexualprowess may compete with emotional links.Growing approval of envy has adjusted 19th-century standards to the needs of a consumereconomy (White, 1993; Shumway, 1998;Seidman, 1991; Matt, 1998; Hochschild,1997).

PROBLEMS AND RESPONSES

None of the three chronological focal pointsof historical research on emotion is entirelyworked out. Gaps, disagreements, and issues ofsynthesis persist. One of the obvious currentchallenges involves pulling together diversefindings on emotional transformation, even forthe early modern period, toward a fuller under-standing of relationships among different fac-ets of change. Historians also grapple with in-dications of change in the most recent decadeor two, where their perspective needs to becombined with data emanating from the socialand behavioral sciences: How fast, and howfundamentally, are emotional formulationschanging now in a society such as the UnitedStates? Do trends mainly launched in the sec-ond quarter of the 20th century continue toshape a predominant American Western emo-tional framework?

Positioning historical work in interdisciplin-ary context remains an issue, as historiansthemselves oscillate between interesting theirdisciplinary colleagues in the relevance of emo-tions research to more standard topics, such asthe consequences of the French Revolution,and working across boundaries. Active collab-oration with sociology is gaining ground, de-spite the continued impulse of some Americansociologists to treat the pre-20th-century pastin terms of undifferentiated traditionalism.And there is also friendly overlap with anthro-pology and with some work on the philosophyof emotion.

Contacts with psychology, in contrast, areless well developed, despite their mutual im-portance in terms of theory. The maverick“discursive” or constructionist school useshistorical data readily, and contributes inturn; moreover, work by some individuals,such as Stephanie Shields, directly enrichesemotions history. In principle, additional casestudies could be developed that examine cul-tural context and change along with moredurable emotional expressions and functions(comparative work could be revealing here),to determine more precisely where the bound-ary lines are and to improve the statement ofwhat changes in emotions and emotionalstandards involve. But the dominant culturesand funding sources of history and its disci-plinary allies (such as anthropology andmuch sociology) on the one hand, and psy-


chology on the other, so far inhibit the desir-able interactions. What should be a knowl-edge frontier awaits additional settlers.

Furthermore, some obvious problems are en-demic to the history field itself. Finding dataappropriate for dealing with the emotionalstandards, and even more the emotional expe-rience, of dead people is no easy task. The dis-tinction between professed values and actualemotions helps. Historical research has becomeprogressively more inventive in finding materi-als on emotional standards and in interpretingthem through nuances of language and choicesof metaphor, as well as through explicit mes-sage. Changes in word meanings (“temper,”“lover”) and outright neologisms (“sissy,”“tantrum”) provide direct testimony. Absenceof comment may sometimes prove revealing, asin the avoidance of elaborate jealousy discus-sions in Victorian culture. Utilization of diaryevidence, available in Western society from the17th century and also in 19th-century Japan(Walthall, 1990), provides insight into internal-ization standards and self-evaluations, thoughthere are problems of representativeness. (Foran instance of work using diaries and letters tocalculate the role of love in 18th-century Amer-ican courtship, see Eustace, 2001.) A growingnumber of historians utilize rituals and variousethnographic evidence to get at emotional ex-pressions in the past (Gillis, 1988). Historyadds greatly to the cases available for assessingthe social contours of emotion, and while itsservice as laboratory has some undeniableempirical complexities, major strides in datasources continue.

Emotions history also enters researchers intoversions of debates important in other socialscience fields. The results complicate the field,but also allow historical work to contribute tolarger issues. Research on emotional changeobviously provides yet another confrontationbetween definitions of basic emotions—biologically predetermined, though perhapsvariable in target and expression—and empha-sis on the cultural preconditions of emotionalexperience. Like other emotions researchers,historians, as they have gained in theoreticalsophistication, participate in these discussionsfrom various vantage points; the fact of signifi-cant change in aspects of emotional perceptionadds a vital dimension to the larger debates,challenging excessive focus on inherent basicresponses.

In addition, emotions history generates sometheoretical issues of its own, associated withthe focus on tracing change. These issues meritfurther exploration, not only in history but inother areas of emotions research as historicalfindings are increasingly taken into account.Problems of timing constitute one example.When emotional standards begin to change in asociety, how long does it take for key groups tointernalize the changes, at least to some signifi-cant extent? Are there generalizable factorsthat speed or delay the response? For example,advice givers in the United States began early inthe 19th century to urge that parents not usefear as a disciplinary tool with children. Butmanualists were still arguing against the “bo-geyman” style a century later, implying thatmany parents still held out; and studies of ruralareas in the 1930s reveal explicit and onlymildly embarrassed use of the ploy. Yet changedid come: By the 1950s, most prescriptive liter-ature no longer judged the warnings necessary.The issue is not whether change occurred, butat what pace, and what factors determined thetiming. Another instance involves implementa-tion of standards in the public sphere: Availablefindings in the 19th and 20th centuries, againin the United States, suggest a three- to five-decade lag between significant middle-classacceptance of standards (about marital love,jealousy, or grief) and translation of thesestandards into relevant laws about divorce,jealousy-provoked crime, or grief-related dam-age suits (Stearns, 1994). Again, can we devisemore general models to describe the probablespeed of change, at least in modern societies, orwill we be confined (as is currently inescapable)to case-by-case judgments? A similar problemarises in dealing with the interaction of domi-nant prescriptions, issued by leading religious(or, in modern cases, scientific) popularizers,and the effective emotional standards of sub-cultures (ethnic or socioeconomic). Historianshave done much better with middle- and upper-class emotions history than with immigrant orlower-class.

A final set of historically generated theoreti-cal issues involves the relationship between rec-reations and emotions. Historical work makesit increasingly clear that cultural expressions—in theater, or reading matter, or ritual, orsports—sometimes serve to train individuals indominant emotional norms. Middle-class par-ents in the United States about 1900 believed


that boxing was a good way to teach boys toretain and express anger (Stearns & Stearns,1986), while confining its intensity to appro-priate targets. In other instances, however, cul-ture can be used in reverse fashion, as an outletfor emotions that are proscribed in daily life.Chinese love poems issued from a soci-ety highly intolerant of love in actual youthrelationships (Goode, 1959). Spectator sportsallow 20th- and 21st-century men to ventemotions that they know are normally inap-propriate, despite greater acceptability in thepast. Historians of leisure talk about “compen-sations” for the daily restrictions of contempo-rary life, but their approach needs further inte-gration with emotions history. The historicalperspective is not the only means of entry intothese issues of cultural–emotional relation-ships, but it provides a growing list of signifi-cant and diverse examples.

THE STRENGTHS OFTHE HISTORICAL APPROACH

Emotions history, despite its various limita-tions, has already generated a number of im-portant findings about changes in standardsand their relationship to aspects of emotionalexperience. The direction of changes in threemajor periods in modem Western history, al-though by no means fully captured, is becom-ing increasingly clear. The results in turn addevidence and issues to a number of basic dis-cussions in emotions research, highlighting thephenomenon of change as an essential analyti-cal factor.

In this context, historical analysis permitsdeeper exploration of the causation operatingin the social context of emotions—an area sug-gested by constructivist theory, but not system-atically probed. Historical research deals withthe factors that induce new emotional formula-tions, permitting a kind of causation analysisthat differs from and is more extensive thanthat possible in cross-cultural comparisons ofrelevant variables. In the major cases exploredthus far, historians have picked up on the roleof shifts in larger beliefs in inducing new emo-tional standards. For example, the ProtestantReformation encouraged reevaluation of emo-tions within the family, while new elite cultureprompted reassessment of popular spontaneity.The role of changing expertise in the 20th and

21st centuries provides an opportunity to as-sess cultural causation of another sort.Economic and organization systems providethe second major strand of causation. In-creased commercialization prompted new at-tention to family emotionality in the earlymodern period, as relationships among otheradults became more competitive (Nelson,1969). The separation of home and workprompted emotional reevaluations in the 19thcentury. Most of the leading judgments on ba-sic shifts in the 20th century point to the im-pact of new organizational experiences andstyles, attendant on the rise of a service econ-omy, corporate management hierarchies, andmass consumerism. A causation topic of grow-ing importance, where the historical approachto change over time is essential, involves therole of modern media in both shaping and re-flecting public and even private emotion—helping to organize, for example, grief or fearreactions. Definitive statements of causationremain elusive, particularly in terms of assign-ing priority and precedence to one set of vari-ables over another, but the analytical task hasbeen engaged. It involves evaluations that,though rooted in history, inevitably apply toother social research on emotion as well.

Historical research also encourages renewedattention to the impact of emotion and emo-tional standards, again in a context of change.Emotions research in other disciplines some-times assumes that further understanding ofemotions themselves constitutes a sufficientend result. The growing group of researchersinterested in emotions history certainly seeks toadd to this understanding, and accepts it as amajor goal. Historians, however, are typicallyinterested in relating one facet of the humanexperience to others, so it is natural that theyseek to discuss the effects of emotional changeon other aspects of society, whatever the timeperiod involved. The effort to distinguish be-tween emotional standards and outright expe-rience adds to this inclination, for new stan-dards often have measurable impact—on thelaw, for example—even when basic emotionalexperience may remain more obdurate.

Historians of emotion have consistentlycommented on the interaction between emo-tional change and other facets of family life,such as marriage choice or birth rate decisions.History is proving to be a crucial means of im-proving the articulation between gender and


emotion. Because gender is in large measure acultural construct and varies greatly over time,historical research is central to the determina-tion of the origins and results of particular gen-der formulas for emotional expression. One ofthe key findings of researchers dealing with the19th and 20th centuries thus involves recogniz-ing the central importance of gender distinc-tions in Victorian emotional prescriptions, andthen their reconfiguration beginning in the1920s. Emotional standards also intertwinewith power relationships, even aside from gen-der. Emotions are used regularly to enforce,and sometimes to conceal, such relationships.Research on emotion in the early modern pe-riod deals extensively with the interrelationshipbetween changes in this area and revision offamily hierarchies. Research on differences inrates and directions of emotional change, evenin the 21st century, provides new insight intothe often hidden hierarchies of contemporarysocial and economic life. It also explains thechanging emotional bases for collective protestwithin the configuration of power—includingthe decline of emotionally charged protest, thelate 1960s excepted, from the 1950s to thepresent in Western society (Moore, 1978).

Dealing with emotions through change alsofacilitates evaluation of shifting emotionalproblems—another kind of consequencebroadly construed. Seeing the disadvantages ofintensifying pressures for cheerfulness, for ex-ample, is easier when cheerfulness itself isgrasped as an evolving emotional goal. Emo-tions history can clearly be linked with analysisof changing disease patterns and categories.

Historical research on emotion generates im-portant new data, evaluative tools, and theo-retical perspectives for emotions research moregenerally. It provides prior examples of emo-tional reassessments and an explicit historicalvantage point for evaluating current directionsof change. Emotions history, increasingly en-sconced in the broader field of social historydespite its newcomer status, thus becomes partof an interdisciplinary inquiry into the constit-uents of emotional experience and the role ofemotion in social life. The history of emotionsadds challenge and complexity to the study ofemotion by introducing the factor of change asa central ingredient. Emotions history also pro-vides many of the tools necessary to deal withthe issue of change and to use its analysis tomove toward fuller understanding of the waysin which emotions develop and function.

GROWING MOMENTUM

Work in emotions history from the 1990s on-ward suggests how the field is gaining ground,beginning to tackle some of the major items onthe research agenda. Use of findings by histori-ans not explicitly concerned with emotion alsosuggests important cross-fertilization—for ex-ample, in work on changing patterns of spousalabuse (Del Mar, 1996). Although the break-through in interdisciplinary collaboration withpsychologists has not yet occurred, particularlyin real collaborative research, the discussionswith social psychologists as well as avowedconstructionists are encouraging. Newer re-search also provides sophisticated treatment ofchanges in scientific paradigms of emotion andtheir impact on wider emotional expectations(Dror, 1998).

Equally important are the extensions withinhistory itself, including the coverage of addi-tional emotions such as envy, the increased at-tention to medieval history, and the fuller ex-plorations of the complexities of 20th-centurychange. Some big tasks still await attention,such as comparison and wider exploration ofemotions history in Asian, African, and LatinAmerican contexts. But emotions historians inthe United States are beginning to grapple withone vital task: the evaluation of the relations ofsubgroups to dominant European Americanmiddle-class standards. Emergence of a distinc-tive African American religious–emotional sub-culture took on new significance about 1900,for example, because its trajectory contrastedso markedly with changing patterns in middle-class emotionology (Phillips, 1998). Differ-ences in timing between Catholic emotionalculture and its Protestant counterpart provideimportant insights into uses of fear (Kelly &Kelly, 1998; Griffith, 1998).

Another growth area involves extending theinquiry in the consequences of emotional cul-ture and emotional change. Here, as in explor-ing change itself, historians offer considerableexperience in dealing with causation and withinterlocking factors. Corbin’s (1986) work hasshown the impact of emotional reactions inhelping to shape social class relations, butalso sanitary policies, in 19th-century France.Research on changing patterns of friendshipsuggests another vital impact area (Rotundo,1989; Rosenwein, 2006). Fuller understandingof 20th-century change in emotional expecta-tions is being applied to parent–child relations,


as well as to wider phenomena such as consum-erism (Stearns, 1998). American work has ex-tended the idea of changes in emotional cultureas links between large structural shifts—therise of a corporate, service-based economy, forexample—and alterations in law, politicalbehavior, and even military policy (Stearns,1997, 1999; Wouters, 1992). Connections be-tween changes in emotional style (Wouters,1996) and shifts in the crime rate are being ex-plored. Clearly, this diverse use of emotionalpatterns as cause, though still tentative, is amajor new frontier in the field.

A number of historians and kindred sociolo-gists are attempting to further the contribu-tions of emotions history by defining topicsbased explicitly on contemporary patterns, andmoving backward toward historical explana-tion and assessment, in a movement currentlylabeled “behavioral history” (Stearns, 2005).Thus novel preferences for house size and fur-niture purchases in the United States are evalu-ated in terms of changes in nostalgia (explainedin turn through emotional reactions to shiftingresidential patterns). A challenging opportu-nity for behavioral history emerges concerningpublic, and in some cases private, experiencesof fear in the contemporary United States.Several sociologists have posited counterpro-ductive levels of fear (Glassner, 1999), butthey encounter difficulties in explaining theemotions—their primary focus is on irresponsi-ble media—and in locating them in terms ofchange. Emotions history, applied to currentfear behaviors, can do a better job in measuringpublic fears over time (e.g., comparing reac-tions to Pearl Harbor to those roused by theterrorist attacks of 9/11), and also in assessingcausation; the media role thus makes no senseif it is not juxtaposed with broader shifts in thesocialization of fear. Emotions history can alsojoin in evaluating the consequences of currentAmerican fear and—tentatively—in exploringremediation (Stearns, 2006). Here is a signifi-cant new frontier for interdisciplinary inquiry.

Emotions history has carved a noticeable(though hardly dominant) niche in the largerhistory discipline, and a measurably increasingrole in interdisciplinary work. It has consider-ably expanded available knowledge about therange of emotional cultures and behaviors. Ithas fed important theories about processes ofchange and, increasingly, the consequences ofchange. Its interactions with other relevant dis-ciplines have produced mutually significant re-

sults. Real challenges remain, however—in so-cieties, periods, and topics still unexplored, aswell as in the pressing need for comparativework and for fuller cross-disciplinary collabo-ration. The vitality of the subfield suggests thepossibility of addressing these broader researchneeds.

REFERENCES

Averill, J. R. (1980). A constructivist view of emotion.In R. Plutchik & H. Kellerman (Eds.), Emotion: The-ory, research, and experience: Vol. 1. Theories ofemotion (pp. 305–339). New York: Academic Press.

Averill, J. R. (1982). Anger and aggression: An essay onemotion. New York: Springer-Verlag.

Badinter, E. (1980). L’amour en plus: Histoire del’amour maternel. Paris: Flammarion.

Bremmer, J., & Roodenburg, H. (Eds.). (1997). A cul-tural history of humor. Cambridge, MA: Polity Press.

Burke, P. (1978). Popular culture in early modern Eu-rope. New York: New York University Press.

Bynum, C. (1997). Wonder. American Historical Re-view, 102, 1–26.

Cancian, F. M. (1987). Love in America: Gender andself development. Cambridge, UK: Cambridge Uni-versity Press.

Cancian, F. M., & Gordon, S. (1988). Changing emo-tion norms in marriage: Love and anger in U.S.women’s magazines since 1900. Gender and Society,2(3), 303–342.

Corbin, A. (1986). The foul and the fragrant: Odor andthe French imagination. Cambridge, MA: HarvardUniversity Press.

Del Mar, D. (1996). What trouble I have seen: A historyof violence against wives. Cambridge, MA: HarvardUniversity Press.

Delumeau, J. (1978). La peur en Occident, XIVe–XVIIesiecles: Uné cite assiégée. Paris: Fayard.

Delumeau, J. (1989). Rassurer et proteger: Le senti-ment de sécurité dans l ’Occident d’autrefois. Paris:Fayard.

Demos, J. (1970). A little commonwealth: Family life inPlymouth Colony. New York: Oxford UniversityPress.

Demos, J. (1988). Shame and guilt in early New Eng-land. In C. Z. Stearns & P. N. Stearns (Eds.), Emo-tion and social change: Toward a new psychohistory(pp. 69–86). New York: Holmes & Meier.

de Swaan, A. (1981). The politics of agoraphobia: Onchanges in emotional and relational management.Theory and Society, 10(3), 359–385.

Dror, O. E. (1998). Creating the emotional body: Con-fusion, possibilities, and knowledge. In P. N. Stearns& J. Lewis (Eds.), Emotional history of the UnitedStates (pp. 173–196). New York: New York Univer-sity Press.

Elias, N. (1982). The history of manners (E. Jephcott,


Trans.). New York: Pantheon Books. (Original workpublished 1938)

Elvin, M. (1989). Tales of the Shen and Xien: Body-personal and heart-mind in China during the last 150years. Zone, 4, 266–349.

Elvin, M. (1991). The inner world of 1830. Daedalus,120(2), 33–61.

Erikson, E. (1958). Young man Luther. New York:Norton.

Eustace, N. (2001). The cornerstone of a copious work:Love and power in eighteenth-century courtship.Journal of Social History, 34(3), 517–546.

Febvre, L. (1973). A new kind of history. New York.Harper & Row. (Original work published 1933)

Flandrin, J. L. (1979). Families in former times (R.Southern, Trans.). Cambridge, UK: Cambridge Uni-versity Press.

Gay, P. (1984–1986). The bourgeois experience: Victo-ria to Freud (2 vols.). New York: Oxford UniversityPress.

Gergen, K. J. (1998). History and psychology: Threeweddings and a future. In P. N. Stearns & J. Lewis(Eds.), Emotional history of the United States(pp. 15–32). New York: New York University Press.

Gerhards, J. (1989). The changing culture of emotionsin modern society. Social Science Information, 28,737–754.

Gillis, J. R. (1985). For better for worse: British mar-riages, 1600 to the present. New York: Oxford Uni-versity Press.

Gillis, J. R. (1988). From ritual to romance: Toward analternate history of love. In C. Z. Stearns & P. N.Stearns (Eds.), Emotion and social change: Toward anew psychohistory (pp. 87–122). New York: Holmes& Meier.

Glassner, B. (1999). The culture of fear: Why Americansare afraid of the wrong things. New York: BasicBooks.

Goode, W. J. (1959). The theoretical importance of love.American Sociological Review, 24(1), 38–47.

Gordon, S. L. (1989). The socialization of children’semotion: Emotional culture, competence, and expo-sure. In C. Saarni & P. Harris (Eds.), Children’s un-derstanding of emotion (pp. 319–349). Cambridge,UK: Cambridge University Press.

Greven, P. J., Jr. (1977). The Protestant temperament:Patterns of child-rearing, religious experience and theself in early America. New York: Knopf.

Greven, P. J., Jr. (1991). Spare the child: The religiousroots of punishment and the psychological impact ofphysical abuse. New York: Knopf.

Griffith, R. M. (1998). “Joy unspeakable and full ofglory”: The vocabulary of pious emotion in the nar-ratives of American Pentecostal women, 1910—1945. In P. N. Stearns & J. Lewis (Eds.), Emotionalhistory of the United States (pp. 218–240). NewYork: New York University Press.

Griswold, R. L. (1986). The evolution of the doctrine ofmental cruelty in Victorian American divorce, 1790–1900. Journal of Social History, 20, 127–148.

Harré, R., & Stearns, P. N. (Eds.). (1995). Discursivepsychology in practice. London: Sage.

Hochschild, A. R. (1997, April 20). There’s no place likework. The New York Times Magazine, pp. 50—55,81–84.

Houlbrooke, R. (Ed.). (1989). Death, ritual and be-reavement. New York: Routledge/Chapman & Hall.

Huizinga, J. (1927). The waning of the Middle Ages.London: Arnold.

Isaac, R. (1982). The transformation of Virginia, 1740–1790. Chapel Hill: University of North CarolinaPress.

Jones, C. (1996). The great chain of buying: Medical ad-vertisement, the bourgeois public sphere, and the ori-gins of the French Revolution. The American Histori-cal Review, 10(1), 13–40.

Kagan, J. (1979). The growth of the child: Reflectionson human development. New York: Norton.

Kasson, J. F. (1990). Rudeness and civility: Manners innineteenth-century urban America. New York: Hill& Wang.

Kelly, T., & Kelly, J. (1998). American Catholics anddiscourse of fear. In P. N. Stearns & J. Lewis (Eds.),Emotional history of the United States (pp. 259–282). New York: New York University Press.

Kotchemidova, C. (2005). From good cheer to “drive-by smiling. Journal of Social History, 39, 5–38.

Leites, E. (1986). The Puritan conscience and modernsexuality. New Haven, CT: Yale University Press.

Levy, R. I. (1973). Tahitians: Mind and experience in theSociety Islands. Chicago: University of ChicagoPress.

Lewis, J. (1989). Mother’s love: The construction of anemotion in nineteenth-century America. In A. E.Barnes & P. N. Stearns (Eds.), Social history and is-sues in human consciousness (pp. 209–229). NewYork: New York University Press.

Lofland, L. (1985). The social shaping of emotion:The case of grief. Symbolic Interaction, 8(2), 171–190.

Lystra, K. (1989). Searching the heart: Women, men,and romantic love in nineteenth-century America.New York: Oxford University Press.

Matt, S. J. (1998). Frocks, finery, and feelings: Ruraland urban women’s envy, 1890–1930. In P. N.Stearns & J. Lewis (Eds.), Emotional history of theUnited States (pp. 377–395). New York: New YorkUniversity Press.

McDaniel, P. A. (2003). Shrinking violets and CasparMilquetoasts: Shyness, power, and intimacy in theUnited States, 1950–1995. New York: New YorkUniversity Press.

Mintz, S., & Kellogg, S. (1989). Domestic revolutions:A social history of American family life. New York:Free Press.

Mitzman, A. (1987). The civilizing offensive: Mental-ities, high culture and individual psyches. Journal ofSocial History, 20, 663–688.

Moore, B. F. (1978). Injustice: The social basis of obedi-ence and revolt. White Plains, NY: Sharpe.


Muchembled, R. (1985). Popular culture and elite cul-ture in France, 1400–1750 (L. Cochrane, Trans.). Ba-ton Rouge: Louisiana State University Press.

Nelson, B. (1969). The idea of usury: From tribal broth-erhood to universal brotherhood. Princeton, NJ:Princeton University Press.

Pfister, J., & Schnog, N. (Eds.). (1997). Inventing thepsychological: Toward a cultural history of emo-tional life in America. New Haven, CT: Yale Univer-sity Press.

Phillips, K. L. (1998). “Stand by me”: Sacred quartetmusic and emotionology of African American audi-ences, 1900–1930. In P. N. Stearns & J. Lewis (Eds.),Emotional history of the United States (pp. 241–258). New York: New York University Press.

Pollock, L. A. (1983). Forgotten children: Parent–childrelations from 1500 to 1900. Cambridge, UK: Cam-bridge University Press.

Reddy, W. (2001). Navigation of feeling: A frameworkfor the history of emotions. Cambridge, UK: Cam-bridge University Press.

Rosenblatt, P. (1983). Bitter, bitter tears: Nineteenth-century diarists and twentieth-century grief theories.Minneapolis: University of Minnesota Press.

Rosenwein, B. H. (2006). Emotional communities in theearly middle ages. Ithaca, NY: Cornell UniversityPress.

Rotundo, A. (1989). Romantic friendship: Male inti-macy and middle-class youth in the northern UnitedStates, 1800–1900. Journal of Social History, 23, 1–25.

Sabean, D. (1984). Power in the blood: Popular cultureand village discourse in early modern Germany.Cambridge, UK: Cambridge University Press.

Salovey, P. (Ed.). (1991). The psychology of jealousyand envy. New York: Guilford Press.

Seidman, S. (1991). Romantic longings: Love in Amer-ica, 1830–1980. New York: Routledge.

Shields, S. A., & Koster, B. A. (1989). Emotional stereo-typing of parent in child rearing manuals, 1915–1980. Social Psychology Quarterly, 52(1), 44–55.

Shorter, E. (1975). The making of the modern family.New York: Basic Books.

Shumway, D. R. (1998). Something old, something new:Romance and marital advice in the 1920s. In P. N.Stearns & J. Lewis (Eds.), Emotional history of theUnited States (pp. 305–318). New York: New YorkUniversity Press.

Solomon, R. H. (1971). Mao’s revolution and Chinesepolitical culture. Berkeley: University of CaliforniaPress.

Stearns, C. Z., & Stearns, P. N. (1986). Anger: Thestruggle for emotional control in America’s history.Chicago: University of Chicago Press.

Stearns, C. Z., & Stearns, P. N. (Eds.). (1988). Emotion

and social change: Toward a new psychohistory.New York: Holmes & Meier.

Stearns, P. N. (1989). Jealousy: The evolution of anemotion in American history. New York: New YorkUniversity Press.

Stearns, P. N. (1994). American cool: Constructing atwentieth century emotional style. New York: NewYork University Press.

Stearns, P. N. (1997). Emotional change and politicaldisengagement in the 20th-century United States. In-novation—The European Journal of Social Sciences,10(4), 361–380.

Stearns, P. N. (1998). Consumerism and childhood:New targets for American emotions. In P. N. Stearns& J. Lewis (Eds.), Emotional history of the UnitedStates (pp. 396–416). New York: New York Univer-sity Press.

Stearns, P. N. (1999). Perceptions of death in the KoreanWar. War in History, 6, 72–87.

Stearns, P. N. (2005). American behavioral history: Anintroduction. New York: New York University Press.

Stearns, P. N. (2006). American fear: The causes andconsequences of high anxiety. New York: Routledge.

Stearns, P. N., with Stearns, C. Z. (1985). Emotion-ology: Clarifying the history of emotions and emo-tional standards. American Historical Review, 90(4),813–836.

Stone, L. (1977). The family, sex and marriage in Eng-land, 1500–1800. New York: Harper & Row.

Trumbach, R. (1978). The rise of the egalitarian family:Aristocratic kinship and domestic relations in eigh-teenth century England. New York: Academic Press.

Walthall, A. (1990). The family ideology of the rural en-trepreneurs in nineteenth-century Japan. Journal ofSocial History, 23, 463–484.

White, K. (1993). The first sexual revolution: The emer-gence of male heterosexuality in modern America.New York: New York University Press.

Wouters, C. (1991). On status competition and emotionmanagement. Journal of Social History, 24(4), 690–717.

Wouters, C. (1992). On status competition and emotionmanagement: The study of emotions as a new field.Theory Culture and Society, 9, 229–252.

Wouters, C. (1995). Etiquette books and emotion man-agement in the 20th century. Journal of Social His-tory, 29, 107–124, 325–340.

Wouters, C. (1996). Changing patterns of social con-trols and self-controls: On the rise of crime since the1950s. Paper presented at the “Causes of Crisis”symposium, Edinburgh, Scotland.

Wouters, C. (2004). Sex and manners: Female emanci-pation in the west. London: Sage.

Zelizer, V. (1985). Pricing the priceless child. New York:Basic Books.


C H A P T E R 3

The Sociology of Emotions

JAN E. STETS and JONATHAN H. TURNER

Within sociology, the study of emotions is acomparatively recent field of specialization,first emerging in its modern form in the 1970s(Heise, 1979; Hochschild, 1979; Kemper,1978). This late arrival of a field so essential tounderstanding human interaction and socialorganization is partly the result of weak foun-dational statements by the first sociologists,who, with few exceptions, did not offer de-tailed analyses of human emotional arousal.For example Karl Marx (1867/1990) empha-sized the negative arousal of alienation and an-ger that would push the proletariat to revolu-tionary class conflict; Emile Durkheim (1912/1965) identified the negative emotions of ego-ism and anomie while detailing the positiveemotions that come with the copresence of in-dividuals and rituals directed at totems sym-bolizing the group; and Charles HortonCooley (1902/1964) recognized that pride andshame are the background emotions for self-appraisals in the “looking glass” provided byothers’ responses to a person’s behaviors. How-ever, none of these early theorists’ analyses ofemotions provided a strong program for so-ciological work. Even George Herbert Mead

(1934), whose ideas formed the foundation forsociological analysis of micro-level social pro-cesses for the entire 20th century, failed to con-ceptualize emotions. Consequently, the sociol-ogy of emotions remained a recessive mode ofinquiry until its resurgence in the 1970s. Sincethen, the study of emotions has become one ofthe leading edges of theory and research in so-ciology, particularly sociological social psy-chology.

A sociological analysis of emotions beginswith the view that human behavior and inter-action are constrained by individuals’ locationin social structures guided by culture. Individ-uals are seen as incumbents in positions withina set of positions (i.e., social structures) that areregulated by systems of cultural symbols. Bothcognitive appraisal (people’s internal represen-tation of themselves, others, and situations)and emotional arousal are constrained in inter-action by culture and social structure. Figure3.1 outlines the essential elements of a socio-logical analysis of emotions.

“Culture” is defined as systems of symbolsthat humans create and use to regulate their be-haviors and interactions, with the key elements

32

of culture including emotion ideologies (appro-priate feelings and emotional responses in dif-ferent situations), emotion stocks of knowledge(emotional experiences that build up over timeand become available for use in interaction),emotion vocabularies, and feeling and displayrules (Turner & Stets, 2005). These elementsare invoked and used to guide social structureand individuals’ cognitions.

“Social structure” is conceptualized as eithera node in a network revealing varying proper-ties such as density, centrality, and power, or asa set of status positions carrying varying levelsof prestige and other resources (Turner & Stets,2005). Structures distribute resources, oftenunequally, while being embedded inside of eachother. For example, encounters are inside ofgroups, which are inside organizations, com-munities, and social categories, which are in-side institutional domains and stratificationsystems (Turner, 2002).

“Cognitive appraisals” revolve around thefollowing: general definitions of self, other(s),and situations; attributions for the causes be-hind situational outcomes; awareness of one’splace in the social structure; and recognitionof cultural guidelines. “Emotional arousal” isgenerally conceived to flow along a positive–negative polarity, with specific emotions

emerging as a result of the responses to self andothers, a cognitive assessment of one’s place insocial structure, and knowledge of the relevantcultural guidelines (Turner & Stets, 2005).“Interaction” is the process whereby the be-haviors of one or more persons influence thebehaviors of one or more others. For sociolo-gists, behavior is only relevant to the extentthat it takes into consideration the actual or an-ticipated responses of others. Interaction in-volves role playing and the presentation of selfthrough forms of talk, the use of props, and ex-pressive gestures.

As is evident from the reverse causal arrowsconnecting elements in Figure 3.1 (i.e., the ar-rows going from right to left), sociologicalanalysis always emphasizes the recursive na-ture of interaction, which always feeds back toand affects emotional arousal, cognitions, so-cial structure, and culture. Culture and so-cial structure are reproduced or potentiallychanged by the cognitions and emotions thatemerge during the course of interactionsamong individuals.

As we review in this chapter, there are nowat least five theoretical research traditions inthe sociology of emotions, which correspond tothe essential elements of a sociological analysisof emotions outlined in Figure 3.1. Moving

3. The Sociology of Emotions 33

FIGURE 3.1. A sociological analysis of emotions.

from left to right in the model, we reviewthe dramaturgical (culture), structural (socialstructure), symbolic-interactionist (cognitiveappraisal), and ritual and exchange (interac-tion) perspectives on emotion. Each traditionexamines only some of the causal connectionsportrayed in the figure, but together the theo-ries develop concepts, propositions, and empir-ical assessments for all of the direct, indirect,and recursive causal relations outlined in thefigure. As we will review at the end of thischapter, there are several critical issues beyondvarying points of emphasis that divide socio-logical theorizing and research. However, be-fore exploring these contentious issues, we re-view each of the traditions that have emergedin the sociology of emotions over the last 30years.

DRAMATURGICAL APPROACHES

Dramaturgical approaches to the sociology ofemotions emphasize the importance of cul-ture in providing emotion ideologies, stocks ofknowledge, vocabularies, and feeling rules.These elements of culture operate as cognitiveguidelines to what emotions should be experi-enced and expressed in a situation, as well as towhat vocabularies are to be used in adjustingemotional responses to the situation (Gordon,1981; Peterson, 2006). Some good examples ofresearch focusing on the role of culture onemotions include work on grief (Charmaz &Milligan, 2006) and sympathy (Clark, 1997;Schmitt & Clark, 2006). These studies showhow emotions are culturally scripted as to“when” to feel and “how” to express thesefeelings. Through the socialization process, in-dividuals learn how to associate particularemotion vocabularies with specific eliciting sit-uations, internal sensations, and expressive dis-plays. For example, a study by Pollak andThoits (1989) of a therapeutic nursery revealshow adult caretakers teach children how toidentify and express their emotions.

What makes this approach dramaturgical isthat individuals are viewed as acting on a stageconfigured by social structure in front of an au-dience (of others). Actors not only use the cul-tural script to orchestrate their presentations ofself to an audience; they also use stagingprops provided by social structure. However,individuals are more than dramatic actors ma-nipulating emotions through expressive controlin accordance with the feeling and display rules

of culture. They are also strategists who pres-ent themselves to others, manipulating theirforms of talk, role cues, bodies, staging props,and expressive display to their audience to real-ize their goals (Goffman, 1967). Individuals of-ten use their emotional displays strategically tocon and/or to gain favor, power, or status overothers (Clark, 1997).

When individuals violate the cultural script,however, they experience embarrassment andperhaps shame, which leads them to engage inrepair rituals that restore the normal order,while signaling to others that the breach of thisorder was only temporary (Goffman, 1967).Recent work in dramaturgy has emphasizedthat social structures and cultural rules oftendemand that individuals display emotions thatthey do not feel, and that in fact are systemati-cally generated by social structures. Underthese conditions, individuals must engage in“emotion work” or “emotion management” tocontrol their emotional display so that it is inaccordance with the feeling rules of the sit-uation (Hochschild, 1983; Rosenberg, 1990;Thoits, 1990).

Hochschild’s (1983) classic research on air-line attendants and bill collectors is one of thebest illustrations of how dramaturgy can be in-corporated into the study of emotions. Sheadopts elements of Goffman’s analysis to showhow individuals engage in emotion manage-ment to maintain a presentation of self thatconforms to emotion ideologies, cultural feel-ing rules, and cultural display rules of a society.When such emotion management is done atwork, for a wage, it constitutes emotional la-bor. Another classic piece in this genre isPierce’s (1995) study of lawyers and paralegalsin law firms. Her analysis highlights how emo-tional labor is gendered either because of thedifferent positions men and women hold in lawfirms (men are chiefly the lawyers and are ex-pected to be aggressive, while women are typi-cally the paralegals and are expected to be nur-turing), or because of the double standard onthe emotion norms for the same job (men canbe aggressive lawyers, but aggressive womenlawyers are considered unfeminine and domi-neering).

Various cognitive and behavioral strategiesare available for managing emotions (Gross,Richards, & John, 2006; Hochschild, 1983;Rosenberg, 1990; Thoits, 1990). Cognitivestrategies include invoking thoughts associatedwith the emotions demanded in the situation toshore up the emotions, using meditation to


arouse the emotions dictated by the culture, orpsychologically withdrawing from the situa-tion to mask the incongruence between actualfeelings and feelings expected in the situa-tion. Behavioral strategies include evoking ex-pressive gestures consistent with the emotionnorms of the situation to arouse the culturallyappropriate emotions, seeking help and advicefrom others on how to manage one’s emotions,or physically leaving the situation that gener-ates the discrepancy between actual and ex-pected feelings.

Using a national sample of respondents,Stets and Tsushima (2001) have examined howindividuals use these cognitive and behavioralstrategies to cope with anger at work and athome. Indeed, to remain in a continual state ofanger is too disruptive for relationships; angermanagement is needed. They find that individ-uals have a tendency to use behavioral copingstrategies, such as seeking social support fromothers to help them manage their anger, to dealwith work-related problems. In contrast, per-sons have a tendency to use cognitive strate-gies, such as praying to God, for family-relatedproblems. Other researchers have more closelyexamined emotion management among selectgroups of individuals, such as wheelchair users(Cahill & Eggleston, 1994), medical students(Smith & Kleinman, 1989), and animal shelterworkers (Arluke, 1994).

The emphasis on culture and self-presentationin dramaturgical approaches helps inform ushow emotional responses are learned and man-aged in situations. To advance this approachfurther, however, several issues should be ex-amined. There needs to be a more precise con-ceptualization of the elements of emotion cul-tures, a clearer description of how they arelearned, and a better explanation of how theyare used strategically in interaction. The con-cept of emotion work also needs more develop-ment, including the conditions under which itis facilitated or impeded, its positive and nega-tive consequences for the self and others ininteraction, and the impact of these conse-quences on social-structural arrangements andsocial change.

STRUCTURAL APPROACHES

All sociological approaches to the study ofemotions include social structure as part of theanalysis. Indeed, emphasis on the effects of so-cial structure and culture on behavior and

interaction, and vice versa, is the defining char-acteristic of sociology. It is not the analysis ofstructure per se, but the relative emphasis onthe properties of social structure, that is impor-tant. In the sociology of emotions, theorizingand research tend to emphasize a relativelysmall range of social-structural properties. Themost prominent set of approaches, workingfrom diverse theoretical traditions, examinemicro-level power (authority) and status (pres-tige) differences among individuals as these af-fect their cognitive assessments and emotionsduring the course of interaction, which in turneither reproduce or change the power and sta-tus order. In contrast, relatively few approachesfocus on macro-level processes, including insti-tutional spheres and stratification systems asthey determine the distribution of resources,and in turn the power and status order at themicro level.

Microstructural Analysis of Emotions

One set of approaches conceptualize the micro-level order of interaction as revealing differ-ences in the relative power and status amongindividuals or among the positions occupied byindividuals. Sometimes these differences areimposed by an existing social structure, and atother times they emerge during the course ofinteraction. Once in place, status differences es-tablish expectation states for the relative com-petence and performances of those who arehigh and low in the status order (Berger, Fisek,Norman, & Zelditch, 1977; Berger & Webster,2006). Moreover, these differences can becomecodified into cultural beliefs about the relativequalities of those in high and low positions.Such beliefs are most likely to emerge for dif-fuse status characteristics, such as gender, eth-nicity, age, and other status differences that arewidely known in a society and that generate ex-pectation states for competence and perfor-mances. In general, when individuals act in ac-cordance with expectation states and culturalbeliefs, those in higher-ranking positions expe-rience positive emotions such as happiness andpride, while those in lower-ranking positionsexperience negative emotions like anger andfear.

Attribution processes are brought into mostof these power and status theories, because thevalence and direction of emotional arousal de-pends on the attributions made by actors forconfirmation or disconfirmation of their rela-tive status (Houser & Lovaglia, 2002; Lovaglia


& Houser, 1996). Typically, higher-ranking in-dividuals will make self-attributions for theirrank, and thus experience satisfaction, happi-ness, and pride when their performances con-firm their rank. At times, however, higher-ranking individuals make external attributions,seeing their rank as due to broader structuralconditions rather than their performances. As aresult, they will experience status insecurityand will be vulnerable to negative emotionalarousal, particularly variants of fear and anxi-ety.

Lower-ranking persons’ emotional arousalwill also be influenced by attribution dynamics.If individuals perceive that their low rank istheir own fault (i.e., that it is due to lack ofability and/or performance), they will experi-ence low levels of positive emotions such as sat-isfaction. If they have difficulty accepting theirlow rank as their own fault, they will experi-ence mild negative emotions such as sadnessand alienation. If lower-ranking persons do notaccept their situation as the result of their ownlack of ability, but rather as a consequenceof structural arrangements that work unfairlyagainst them, they will experience anger—and,if possible, will direct this anger outward to-ward higher-ranking persons and the larger so-cial structure.

Different theories stress varying facets ofthe general approach summarized above. InKemper’s interactional theory (Kemper, 1978;Kemper & Collins, 1990) or power–status the-ory of emotions (Kemper, 2006), emotions arerelated to individuals’ relative power (what helabels “involuntary compliance”) and status(“voluntary compliance”), changes in relativepower and status, and expectations for gains orlosses of power and status. Thus Kemper con-siders both dimensions of social relations—power and status—in his theory, rather thanjust status.

According to Kemper, when individuals gainpower, they will experience satisfaction, confi-dence, and security; when they lose power andwhen they do not experience an expected gainin power, they will experience fear, anxiety, andloss of confidence. When they possess status orgain expected status, they will experience satis-faction and well-being, and they will expresspositive sentiments to others. When individualslose prestige, the emotional dynamics are influ-enced by attributions. When individuals blamethemselves for a loss of status, they will experi-ence shame and embarrassment, and if the lossis great, they will also feel sad and depressed.

If, however, they blame others for their loss ofstatus, they will experience anger and seek toforce others to honor their status claims.

In a direct test of the power–status ap-proach, Kemper (1991) analyzed data gatheredfrom an eight-nation study (Scherer, Wallbott,& Summerfield, 1986). Of particular interestto him were respondents’ reports of fear, joy,sadness, and anger. He investigated whetherthe power–status antecedents could be found inrespondents’ accounts of the situational cir-cumstances leading to the four emotions. Forexample, would fear emerge from power lossby the self or power gain by another? Wouldjoy emerge from status gain by the self?Coders’ evaluated respondents’ open-ended re-sponses regarding the events leading up toemotions as instances of power gain or loss andstatus gain or loss. The results indicated thatthe power–status conditions did lead to the ex-pected emotions at a rate higher than would beexpected by chance.

Many sociological social psychologists workwithin the expectation states theory researchprogram, which, as its name suggests, empha-sizes the effects of expectations on emotionalarousal. The focus is on the expectations asso-ciated with status. Unlike Kemper’s theory,power is not considered. Furthermore, in ex-pectation states theory, emotion is only oneoutcome of the status process in interaction, soemotion does not play a major role in the the-ory as it does in Kemper’s power–status theory.

In expectation states theory, when individu-als meet performance expectations associatedwith their status (e.g., when high-status actorsoutperform low-status actors), positive emo-tions will be felt by both sets of actors. In thisway, status expectations shape emotion. How-ever, emotions may themselves shape the devel-opment of a status hierarchy (Ridgeway, 2006).They may act as an additional factor thatshapes performance expectations (e.g., if welike a person, we will evaluate his or her perfor-mance more positively, thereby giving that per-son higher status). Alternatively, it may be anadditional factor that influences performanceexpectations only in certain situations, whenthere is not much at stake in a group decision—as opposed to a crisis situation, where perfor-mance is important rather than sentimentsabout each other. To date, the results are notdefinite as to which emotions may influence thestatus hierarchy.

Ridgeway (1994, 2006; Ridgeway & John-son, 1990) has shown how the evaluation of


task suggestions in a group (either agreeing ordisagreeing with the task suggestions) not onlyis important in the establishment of a status hi-erarchy in the group, but also is an elicitor ofpositive and negative emotions, thereby settinginto motion affective dynamics in a group.Ridgeway has emphasized the emotions thatensue when there are disagreements betweenhigher- and lower-status individuals in taskgroups. If a person is at the same rank as, or ata higher rank than, another with whom a dis-agreement occurs, this person will experiencemild anger or annoyance toward the other; ifthis individual is at a lower rank, he or she willexperience sadness and potentially, depression.Moreover, lower-status members of the groupwill often negatively sanction a lower-rankingindividual who disagrees with a higher-statusperson, thus increasing the negative emotionalarousal. In contrast, agreements work to in-crease the flow of positive emotions in thegroup. With agreements between higher- andlower-status persons, the higher-status individ-ual will experience satisfaction, and as a conse-quence will be disposed to give off positiveemotions to lower-status persons, who will ex-perience satisfaction and potentially happiness.Similarly, agreements among persons at thesame status will produce pleasure and evengratitude. Because positive emotional arousalleads individuals to positively sanction eachother, thus ratcheting up the flow of positiveemotions, groups tend to develop biases for be-haviors that maintain the status order. To chal-lenge the order works to arouse negative emo-tions, which lead to negative sanctioning andnegative emotional arousal.

Macrostructural Analysis of Emotions

Relatively few researchers in the sociology ofemotions examine social structure from a moremacro-level perspective. Particular subfieldswithin sociology, such as the study of collectivebehavior, social movements, political legiti-mization, terrorism, or revolutions, employimplicit conceptions of collective emotionalarousal, but these approaches typically do notdevelop a robust analysis of emotions. One ex-ception to this conclusion is work by Barbalet(1998). Although Barbalet does not develop asystematic theory of emotions, he does explorethe relationship among certain aspects of socialstructure—especially those having to do withinequality, power, and specific emotions(namely, resentment, fear, confidence, vengeful-

ness, and shame)—from a more macro-levelperspective.

For Barbalet, emotions are distributed differ-entially across segments of a society, most typi-cally corresponding to each segment’s socioeco-nomic status. Individuals react emotionally totheir respective shares of money, power, andprestige in ways that cannot be fully capturedin small-group experiments. Resentment is anemotion that arises when subpopulations per-ceive that others gain power and prestige that isnot deserved or that violates cultural rules ofjustice. For example, in his conceptual reexam-ination of Sennett and Cobb’s (1972) classicstudy of the hidden injuries of class, Barbaletmaintains that much of the resentment ofworking-class males against welfare recipientsis sublimated resentment about their lack of re-wards for their hard work. Rather than resentthe class system and the forces that havepushed them at the lower rungs of the system,working-class males sublimate their resentmentand transfer it to those who are perceived togain without doing the work.

Fear is another emotion that Barbalet concep-tualizes to be differentially distributed across asociety. Fear comes from a lack of power, and ifindividuals attribute this fear to their own short-comings, then fear leads to withdrawal andflight responses; if individuals make external at-tributions, fear turns into anger, aggression, andfight responses. Confidence, another emotionBarbalet discusses, arises among those membersof a society who perceive that their future isunder their control and is predictable. Vengeful-ness occurs among subpopulations, usually theless powerful, who often see the powerful as de-nying them the right to occupy high-status posi-tions and play roles in society that allow them toform meaningful and cooperative social rela-tionships. Finally, shame emerges among indi-viduals who claim or are given status that theydo not deserve.

Thus Barbalet’s preliminary and speculativeanalysis points to a form of social-structuralanalysis: the view that emotions are like mostresources, and are thereby distributed un-equally. Specific kinds of emotion states emergeamong subpopulations under predictable con-ditions, and sociological theory will need tospecify these conditions and their potential ef-fects on collective action in societies.

Most structural theories on emotions are de-cidedly micro-level in focus, examining pro-cesses in face-to-face interaction such as powerand status. However, micro-level processes oc-


cur within a larger macrostructure in whichpower and resources are unequally distributed.For a structural theory of emotions to develop,it needs to make connections between micro-and macrostructures. Whole categories of indi-viduals based on gender, class, or race, forexample, can experience similar emotionsbecause they stand in the same place inthe stratification system. They are structurallyequivalent; they are likely to have the same ex-periences; and hence they are likely to havesimilar emotional reactions. Barbalet’s workcomes closet to linking macrostructural pro-cesses to emotions at the micro level, but muchmore of such inquiry is needed.

SYMBOLIC INTERACTIONISTAPPROACHES

All symbolic interactionist theories and re-search programs emphasize the central place ofself and identities in the arousal of emotions.Two types of cognitive appraisals revolvearound individuals: (1) transsituational andgeneral conceptions of themselves as particularkinds of individuals, and (2) situational identi-ties about themselves in particular roles in thesocial structure. Individuals formulate self-meanings/identity standards, obtain self-perceptions of who they are in situations, as-sess the degree of congruence between self-meanings and situational self-meanings, andexperience emotional reactions (Stets, 2006).The emotional responses will be positive whensituational self-meanings are congruent withone’s own self-meanings/identity standards,and negative when they are incongruent. In thelatter case, individuals are motivated to bringinto congruence the self-meanings and situa-tional self-meanings. When individuals habitu-ally cannot confirm an identity, they will en-gage in cognitive or behavioral strategies orboth (discussed below) to derive positive emo-tions.

The gestalt-like processes described abovelead toward consistency among cognitionsabout the self, others, and (by extension) one’splace in the social structure and in reference tocultural standards. There are also more psy-choanalytic versions of symbolic interactionisttheory and research, which emphasize that neg-ative emotions like shame and guilt are painfuland hence are likely to activate defense mecha-nisms revolving around repression, thus dis-rupting the control system that seeks to sustain

consistency among cognitions about self, other,social structure, and culture (Turner, 2006). Be-low we examine representative approaches inboth the gestalt and psychoanalytic variants ofsymbolic interactionism.

Control Theoretical Variantsof Symbolic Interactionism

Symbolic interactionist approaches to emo-tions generally view the emotional reactions toeither verification or nonverification of self oridentity as part of a perceptual control system(Burke, 1991; Heise, 1979). Negative emo-tional arousal signals to a person that a generalself-conception or more situational identityis not confirmed, and this distress, anxiety,or sadness motivates individuals to employ anumber of cognitive and behavioral strategies.For example, cognitive strategies might includeselectively perceiving or interpreting the ges-tures of others so as to verify one’s identity ormaking external attributions that blame others,the situation, or the social structure for failingto verify the self. Behavioral strategies mightinvolve changing one’s behavior to obtain veri-fication or leaving the situation.

In several symbolic interactionist theories, ahierarchy of identities is conceptualized, withthose high in the hierarchy of “salience”(Stryker, 1980/2002, 2004) or “prominence”(McCall & Simmons, 1978) more likely to bepresented in situations than those low in the hi-erarchy. When a highly salient identity is dis-confirmed, individuals will feel negative emo-tions, and the identity may move down thehierarchy. Research appears to support thisidea. For example, Ellestad and Stets (1998)have found that when the mother identity ismore prominent for women, they are morelikely to report the negative emotion of jeal-ousy when fathers intrude into caretaking ac-tivities traditionally reserved for mothers. Un-less the women reassert their nurturing role inthe situation, such as spending more time withtheir children and not allowing their spouses tointrude, the mother identity may not be main-tained and thus may move down the hierarchyof identities. Indeed, Stryker theorizes thatwhen an identity is not being confirmed in a sit-uation, an alternative identity that is being con-firmed and that produces positive emotionssuch as satisfaction and happiness may moveup the identity hierarchy and guide behavior.

In the perceptual control system of identities,particularly identity control theory, there is an


“identity standard,” which is the stored set ofself-meanings attached to an identity thatguides behavior in situations. Individuals com-pare the self-meanings of who they are in a sit-uation, which are based on reflected appraisals(feedback from others) and self-appraisals(their own assessment), with their identity stan-dard (Burke, 1991, 1996). When self-meaningsin the situation correspond with identity stan-dard meanings, positive emotions will be expe-rienced. A lack of correspondence in meaningswill lead to negative emotions, and in response,the cognitive and behavioral strategies men-tioned previously will be used to generate con-gruency in meanings.

In a series of laboratory experiments, Stets(2003, 2005) has found that a lack of corre-spondence between self-meanings in a situationand identity standard meanings does not al-ways lead to negative emotions; in fact, it leadsto positive emotions when the identity non-verification is in a positive direction (receivingfeedback that is more positive or higher thanone’s identity standard), compared to a nega-tive direction. Stets argues that this unexpectedfinding may have been obtained because partic-ipants in her studies were not given sufficienttime to compare the positive feedback withtheir self-views, as well as because participantswere not invested in the identity that was in-voked in the laboratory. Having the mental re-sources to access one’s self-views, and being in-vested in one’s self-views, are both importantconditions in the verification process (Swann,Rentfrow, & Guinn, 2003); when they are ab-sent, the emotional response following feed-back from others may be different from what isexpected. Indeed, other research reveals thatwhen, for example, investment in one’s self-views is present in a situation, nonverificationin a positive direction does produce negativeemotions (Burke & Harrod, 2005).

Another perceptual control theory of identi-ties is affect control theory (Heise, 1979). Af-fect control theorists array emotions alonglines of “evaluation” (goodness–badness), “po-tency” (powerful–powerless), and “activity”(active–passive). This is known as the EPA di-mension. The theorists emphasize cognitionsabout the identities of self and others, the be-haviors of others, and the situation in whichidentities emerge. The key idea is that individu-als seek to keep their fundamental sentiments(culturally established affective meanings[along the EPA dimension] about identities andbehaviors that transcend particular situations)

in line with their transient impressions (feelingsabout what is actually occurring in a particularsituation in terms of people’s identities and be-haviors) (Robinson, Smith-Lovin, & Wisecup,2006). For example, if the mother identity is onaverage evaluated as good, somewhat power-ful, and somewhat active, but in a particularsituation one finds a mother scolding a child,this person will view the mother identity morenegatively because she is evaluated as badrather than good (for scolding) and as some-what less powerful (for losing control of herchild). There would be no change in the degreeof activity of the mother identity. When a dis-crepancy between fundamental sentiments andtransient impressions occurs, emotions willemerge, but their valence will depend upon thetransient meaning and its direction of changefrom the original fundamental sentiments. Forexample, if transient meanings are more posi-tive than fundamental meanings, one will feelmore positive, while transient meanings thatare more negative than fundamental meaningswill lead to negative emotions. Empirical testshave supported the role of emotions in affectcontrol theory (Heise & Weir, 1999; Robinson,Smith-Lovin, & Tsoudis, 1994).

Psychoanalytic Variantsof Symbolic Interactionism

Psychoanalytic approaches in the sociology ofemotions try to bring the Freudian legacy, par-ticularly as it was incorporated into the seminalanalysis of Helen Block Lewis (1971), into theperceptual control models presented by main-stream symbolic interactionism (Scheff, 1988,1990; Turner, 1999, 2002, 2006). Individualswill often engage in defensive strategies to pro-tect the self from negative emotions. Spe-cifically, when individuals behave incompe-tently in front of others and/or breach themicro-level social order, they experienceshame; when they act in ways violating culturalvalues, they experience guilt (Turner, 2002).

For shame in particular, persons will repressthis negative feeling to varying degrees (Turner,2006), because shame attacks the self andmakes a person feel small and unworthy(Tangney & Dearing, 2002). Turner (2006) ar-gues that once negative emotions are repressed,they will often intensify and be transmuted intonew emotions. For example, he maintains thatshame often emerges from the cortical sensorsas periodic spikes of anger, which, ironically,lead to more shame that is repressed. This is


consistent with Lewis’s (1971) idea that indi-viduals can become locked into shame–anger–shame cycles, with the successive repression ofthe shame intensifying the mix of unconsciousemotions that periodically erupt as anger.

For psychoanalytic versions of symbolicinteractionism, individuals are often unawareof the disjuncture between meanings given offby their self-presentations and meanings as tohow others see them in a particular situation,because repression has pushed below the levelof consciousness the negative emotions that re-sult and that are important to the perceptualcontrol system posited by mainstream symbolicinteractionism. Psychoanalytic theories thusadd a corrective to the gestalt dynamics of sym-bolic interactionism, arguing that when nega-tive emotions are intense, defense mechanismsmay change the dynamics of the control sys-tem, sometimes with severe consequences forthe individual.

Such consequences can extend to moremacro-level social structures. For example,Scheff and Retzinger (1991) have argued thatthe repression of shame in societies can leadto collective violence at the societal and evenintersocietal levels when the repressed shameis transmuted into anger and, through mobili-zation by leaders and the mass media, is di-rected at “enemies” within a society. In theiranalysis of the origins of World War II, theycontend that the unacknowledged shame ex-perienced by Germans as a result of theirblame and humiliation as the sole instigatorsof World War I, coupled with Hitler’s repres-sion of shame and rage toward his father andobsession with Jews, provided a lethal combi-nation: The shame and anger of both theGerman people and Hitler spiraled out ofcontrol and caused the murder of millions ofpeople.

The strength of the symbolic interactionistapproach is its emphasis on self and identity inunderstanding emotions. However, futurework on integrating the perceptual control andpsychoanalytic views of self is required.Whereas the former emphasizes the self-verification process, the latter emphasizes re-pression and defense mechanisms that mayemerge when the self is not verified. The condi-tions under which repression and defensivestrategies emerge in the perceptual control sys-tem, as well as the consequences of this emer-gence for future self-verification and one’s emo-tional experiences, need to be studied.

RITUAL APPROACHES

Interaction ritual theories borrow fromDurkheim’s (1912/1965) secondary analysis of“collective effervescence” among Australianaboriginals. When aboriginals would periodi-cally gather, their copresence would lead toanimated interaction, which in turn wouldgenerate a collective effervescence of positiveemotions. This effervescence led individuals toperceive that there was a power or mana exter-nal to their actions and, indeed, guiding theiractions. As a result, they would begin to repre-sent this power with totems. When rituals weredirected to these totems to symbolize the powerof the gods, they would rekindle the sense of ef-fervescence and mana. Durkheim concludedthat the aboriginals’ religion was nothing morethan the personification of their group solidar-ity, which they saw as all-powerful and in needof representation with totems and reverencethrough rituals.

This ritual approach has recently been ex-panded by Collins (2004) and Summers-Effler(2004b, 2006) into a general theory of micro-dynamic processes that operate as the under-pinnings of more macro-level social structures.Social structures bring individuals intocopresence, often allocating varying degrees ofsymbolic, material, and power resources topersons occupying different positions in thestructure. Individuals are seen as motivated toaugment their positive emotional energy (ini-tiative and enthusiasm) and cultural capital inall situations. Those with high levels of re-sources and power (higher-status persons) willgive orders in situations, and they will gener-ally act in ways that allow them to increasetheir emotional energy and cultural capital. Bycontrast, lower-status persons will have moredifficulty augmenting their emotional energyand cultural capital, because they must take or-ders.

These more recent versions of the theory em-phasize that interaction rituals are of two kinds(Collins, 2004). One kind consists of the tran-sient rituals that open and close interactionsand that arouse mild positive emotions (e.g.,handshakes, greetings). Another kind of ritual,such as a church service, is more inclusive andincorporates the larger sequence of interactionwhereby individuals become copresent, reveala common focus of attention, and develop ashared mood, which in turn causes a rhythmicsynchronization of talk and body alignment,


leading to a collective effervescence and in-creased level of emotional energy.

In general, the escalation of positive emo-tional energy in interaction rituals of the sec-ond kind increases group solidarity, whichleads individuals to symbolize the group withwords, emblems, phrases, and other “totems.”The more the group is symbolized in theseterms, the more likely it is that iterated interac-tions will reinforce the symbols of the group.And as symbols circulate among members, theydevelop particularized cultural capital uniqueto the group. Even individuals’ internal conver-sations or thoughts about the symbols of thegroup have the capacity to arouse positiveemotions, all of which increase group solidaritywhile enhancing cultural capital. Similarly, iter-ations of this more inclusive interaction ritualwill charge up group symbols and, in so doing,enhance not only particularized cultural capitalbut also social structure. Thus such interactionrituals provide a firm micro-level basis formacrostructures. When initial ritual openingsdo not cause even mild transient emotions, orwhen any other element of the early phases ofthe encounter (such as mutual focus of atten-tion, shared mood, and rhythmic synchron-ization) do not ensue, then emotional effer-vescence and positive emotional energy willbe low and perhaps even become negative,thereby reducing group solidarity. In turn, cul-tural capital will not be produced and solidar-ity will be low, thus undermining macro-levelsocial structures.

Empirical work that uses concepts from Col-lins’s theory includes Summers-Effler’s (2004a)analysis of two altruistic social movement or-ganizations: a Catholic Worker house seekingto establish a communal lifestyle, and an anti-death-penalty organization. She shows how in-teractions among group members generatehigh degrees of group solidarity through thearousal of positive emotional energy and thecirculation of particularized cultural capital.

The reconceptualization of Durkheim’s workinto a more robust conception of the dynamicsof interaction and emotions provided by cur-rent ritual theories confronts several problems.One is the portrayal of emotions in terms onlyof negative and positive valence states. There isa rich array of complex specific emotions thatneeds to be investigated beyond their posi-tive or negative valences. Another problem isthat ritual approaches posit a need for peopleto maximize emotional energy, and with the

exception of Summers-Effler’s (2004c) recentwork, more research on instances where indi-viduals stay in relationships arousing negativeemotional energy needs to be conducted. WhileSummers-Effler discusses how individuals de-velop strategies that minimize the loss of emo-tional energy when they cannot leave interac-tions that produce negative emotional energy,the ways in which individuals defend them-selves from situations in which negative emo-tional energy prevails have not been adequatelyconceptualized by ritual theories. Part of theproblem is that unlike symbolic interactionism,ritual theories do not include an adequate con-ceptualization of self, which is needed in under-standing the strategies for managing negativeemotional arousal.

EXCHANGE APPROACHES

Exchange theories view individuals as moti-vated to seek valued resources/rewards in ex-cess of their costs (alternative sources of re-wards forgone) and investments (accumulatedcosts over time). Actors give resources to oth-ers with the expectation that they will receivemore valuable resources in return, therebymaking a profit. Payoffs are assessed not onlyin terms of the level of profit, but alsoagainst normative standards of justice andfair exchange. Cognitive calculations of jus-tice and fairness also involve a comparisonprocess of assessing one’s own payoffs rela-tive to the payoffs (including costs and invest-ments) of others.

In general, individuals experience positiveemotions such as happiness when they receive aprofit that is defined as “just,” and when oth-ers’ profits are proportionate to their costs andinvestments as compared to the costs and in-vestments of the persons making the compari-sons. When payoffs are not proportionate toindividuals’ own costs and investments, whenothers receive higher payoffs disproportionateto their costs and investments, and when normsof fair exchange and justice are violated, indi-viduals will experience negative emotions (pri-marily anger). The nature of intensity of emo-tional arousal is, however, governed by severaladditional considerations: the type of ex-change, the relative power and dependence ofindividuals on each other for valued resources,the expectations for payoffs, the particularstandards of justice invoked, and the attribu-


tions that individuals make for profits andloses.

With respect to the nature of exchanges,there are four basic types (Lawler, 2001): “pro-ductive,” where close coordination of activitiesoccurs and where each person’s relative contri-bution to payoffs is difficult to separate fromthat of others; “negotiated,” where individualsbargain, making offers and counteroffers, inseeking resources from each other; “recipro-cal,” where a person provides resources to an-other with assurances of receiving resources inreturn at a later point in time; and “general-ized,” where an individual gives resources toanother who in turn gives them to another whodoes the same, in a chain that eventually circlesback to the original individual. Conceptually,productive exchanges will generate the most in-tense emotions, whether positive or negative,followed by negotiated, reciprocal, and gener-alized (Lawler, 2001). Molm’s (1997) researchreveals how negotiated exchanges often intro-duce conflict into interaction, and hence arouseanger and other conflict-oriented emotions.Reciprocal exchanges generate trust and posi-tive emotions revolving around the expectationthat the giving of resources will be recipro-cated.

The relative power of individuals in an ex-change influences payoffs. The power of actorA over actor B is a function of the dependenceof B on A for valued resources (Emerson,1962). Power-advantaged actors will typicallyimpose additional costs on dependent actorsfor resources, thus increasing the negative emo-tions (such as anger) among dependent actors.In a series of experimental studies, Lawler andYoon have shown how actors who are equallydependent on each other for resources (high to-tal power) and who thus possess equal levels ofpower will have a high rate of exchange amongthemselves, and how this high rate of exchangewill in turn generate more positive emotions(pleasure, satisfaction, interest, excitement),more relational cohesion, and more commit-ment behaviors during exchanges (Lawler &Yoon, 1996, 1998; Thye, Yoon, & Lawler,2002). Lawler and Yoon’s work, known as thetheory of “relational cohesion,” operates muchlike the interaction ritual theories discussedabove. In the theory of relational cohesion,new resources—positive emotions, cohesion,and commitments—are introduced into the ex-change and generate a kind of mild “collectiveeffervescence” reminiscent of ritual theories.

Moreover, commitments to exchanges operateto produce positive emotional arousal, becausethey reduce the fear and anxiety that comeswith uncertainty over exchange payoffs fromothers (Kollock, 1994).

The greater the expectations for payoffs at agiven level, the more intense the negative emo-tional arousal (anger, fear, frustration) whenthey are not forthcoming—and, conversely, thegreater the satisfaction and happiness whenthese payoffs are forthcoming (Jasso, 2006).Moreover, it takes much less underreward be-low expectations to generate anger than it doesoverreward to cause a person to feel guilty(Hegtvedt, 1990). In fact, individuals will typi-cally feel guilty only when their overrewardis perceived to cause underreward to others(Hegtvedt & Killian, 1999).

The nature of the rules of justice invokedwill also influence emotional arousal. In gen-eral, violation of any rule of justice causes aperson to experience and express anger. Normsof equity (payoffs are to be proportionate torelative costs and investments of individuals)will typically trump those of procedures (fair-ness of the procedures used in determining pay-offs). When payoffs are equitable but proce-dures are perceived as unfair, individuals willstill experience positive emotions, whereaswhen norms of equity are violated, individualswill almost always see procedures as unfair andexperience anger.

Emotional arousal is further influenced bythe attributions that individuals make for theirpayoffs. Receipt of a profit in exchanges re-veals a proximal bias (Lawler, 2001), causingindividuals to experience positive emotions to-ward self, such as pride and self-satisfaction;failure to make a profit will arouse negativeemotions that reveal a distal bias, whereby theindividual blames others or the social structure.When others are blamed, the dominant emo-tion is anger, while when social structure isblamed, alienation is the more likely emotion.When others are seen to facilitate profitable ex-change payoffs, gratitude will be experiencedand expressed, whereas when social structure isseen to facilitate profitable payoffs, individualswill experience commitments and attachmentto this structure.

The strength of exchange approaches toemotions is the link between the arousal of spe-cific emotions to the structure of the situationand to the dynamics of attribution. However,like ritual theories, exchange theories do not


include a conceptualization of self; thus we donot know how actors protect the self when theyexperience negative outcomes in a situation be-yond the notion of external attributions. An-other problem with exchange theories is thatthey are based on experimental research wherethe level of emotional arousal is low. Althoughit is remarkable that simple tasks and contrivedexchanges can generate emotions, thus indicat-ing that the dynamics under study are impor-tant for generating emotions, more intenseemotions need to be examined, and these arelikely to take us out of the experimental setting.

CONCLUSION: REMAINING PROBLEMSAND FUTURE PROSPECTS

Despite the late start, the sociology of emotionshas made considerable progress over the last 30years, as this cursory review of approachesdocuments (for more comprehensive reviews,see Stets & Turner, 2006; Turner & Stets,2005). There are now many difference theoriesand research programs testing these theories; asa consequence, considerable accumulation ofknowledge is now evident. There are, however,several issues that need to be addressed if thesociology of emotions is to continue makingprogress.

The Number and Types of Emotions

Sociologists view happiness, fear, anger,sadness, disgust, and surprise as the primaryemotions, with the presumption that theseemotions are hard-wired in the humanneuroanatomy (Turner, 2000). Even the mostextreme social constructivist in sociologywould generally admit this, although theywould also emphasize that the expression ofthese emotions is socially constructed. Withinsociology, several efforts to classify emotionshave been made (Kemper, 1987; Thamm,2006; Turner, 2000). Beyond the classificationof a few emotions as primary, there is no con-sensus over the various classifications; nor doresearchers agree on which are hard-wired andwhich are purely social constructions. In the fu-ture, sociology will need to make some deci-sions on those emotions that are basic to hu-man organization, and, if possible, to resolvethe debate over those emotions that are hard-wired and those that are wholly cultural con-structions.

The Relative Neglect of Biology

For many decades, sociologists have beenhighly suspicious of any view of human behav-ior, interaction, or social organization as bio-logically based. Fears of “reductionism”—and,more implausibly, of a return to forms of rac-ism that accompanied early 20th-century ef-forts at incorporating biology—still haunt thediscipline. The study of emotions provides anopportunity to reject these unfounded fearsabout biology, because there can be little doubtthat many emotions have a strong neurologicalbasis that evolved among our hominid ances-tors. Sociologists need not buy into all of theassumptions of sociobiology or evolutionarypsychology to study the selection pressures thatworked on human neuroanatomy (Franks,2006). Indeed, sociologists are well positionedto provide insights into the biology of emo-tions, because the discipline is sensitive to thesociocultural forces that constrain emotionalarousal and that no doubt guided selection ofhominid neuroanatomy. With a better under-standing of the biology of emotions, it will bepossible to connect biology, cognition, emo-tions, behavior, interaction, social structure,and culture into more robust models of emo-tional dynamics.

Cognition and Emotions

Like psychology, the sociology of emotions re-veals a cognitive bias. Emotions are seen bymany approaches as signals to the self thatthere is incongruence among some set ofcognitions (about the self, behavior, other[s],culture, and social-structural location). In re-sponse to this incongruence, individuals be-come motivated to bring their cognitions intoline. With the exception of a few who workwith psychoanalytic theory, the dynamics of re-pression and other defense mechanisms are notemphasized in sociological approaches to emo-tions. But it is now clear that humans store un-conscious memories and unconsciously acti-vate subcortical emotion centers of the brain.People often display emotions to others thatthey do not feel. In the future, it will be neces-sary to incorporate unconscious emotions intotheories and research programs.

The Study of Weak Emotions

A consideration related to the one just dis-cussed is that much research on emotions in so-


ciology is conducted in experimental laborato-ries and typically focuses on task behaviors.There are, of course, strengths to this ap-proach, because measurements can be moreprecise and controls can be imposed. Emotionsare generally measured by pencil-and-papertests, reactions to vignettes, and interactionswith real persons (or, more often, unreal per-sons, through computer-mediated interactionwith “virtual others”). As a result, the emo-tions studied generally evidence rather weakvalences.

Though clear findings emerge from studiesarousing weak emotional valences, the inten-sity of emotions does matter in most human af-fairs, if only because intense negative emo-tional arousal increases the likelihood ofactivation of defense mechanisms or because itmobilizes people to collective action. Whenemotions are intense, they have greater effectsthat cannot be adequately mapped in the labo-ratory. Thus it will be necessary in the future toconduct careful studies of highly valenced emo-tions in more naturalistic settings—using, forexample, the experience-sampling method(Hektner, Schmidt, & Csikszentmihalyi, 2007)—to better understand the dynamics of strongemotional arousal.

The Micro-Level Bias in Conceptionsof Social Structure

By virtue of experimental designs in most re-search on emotions, the conception of socialstructure reveals a micro-level bias. Small num-bers of individuals in small social structureswith just a few positions or nodes in a networkexist in most experiments. Differences in rela-tive power or prestige in small groups domi-nate the research. But human societies revealmacro-level domains, such as society-wide dis-tributions of power and wealth, that have largeeffects on people’s emotions. Though these of-ten come into focus in micro-level encountersas status differences, the more macro-level dis-tribution of resources and the structure of insti-tutional domains have large effects that makelarge differences in how emotions are collec-tively experienced. And, from collective experi-ences, we obtain some sense as to how macro-level social structures are sustained or changed.With a few exceptions (Barbalet, 1998; Scheff,1994; Scheff & Retzinger, 1991), the moremacro-level consequences of emotional arousalamong subpopulations in a society have not

been studied by those in the sociology of emo-tions. They need to be.

The Emphasis on Culture

Along with the bias toward the micro level inconceptions of social structure, there is a cul-tural bias. Considerable emphasis is placed onemotion ideologies, vocabularies, and rulesas these influence social structure, cognitions,emotional arousal, and interaction. This em-phasis on culture often supports a social-constructionist view, because sociologists tendto view culture as a purely human creation thatis independent of biology (except for the recog-nition that big brains allow humans to createculture, thus obviating human biology as a ma-jor force in human behavior, interaction, andsocial organization). Despite this emphasis onculture, there is surprisingly little in the way ofdetailed analysis of specific emotion ideologies,vocabularies, and rules. These ideas are used ina highly metaphorical way without specifyingthe exact substantive content of these systemsof culture and their effects on the arousal ofemotions, or vice versa. If there is an emotionculture, its components need to be conceptual-ized in more precise terms—and the reciprocaleffects among biology, behavior, interaction,and social structure on the one side, and cul-ture on the other, need to be more adequatelyconceptualized.

Despite the problems described above, thesociology of emotions has made remarkableprogress over the last 30 years. With more the-oretical integration and more robust data, thefield will continue to be a leading edge of theo-rizing and research in sociology.

REFERENCES

Arluke, A. (1994). Managing emotions in an animalshelter. In A. Manning & J. Serpell (Eds.), Animals inhuman society: Changing perspectives (pp. 145–165). New York: Routledge.

Barbalet, J. M. (1998). Emotion, social theory, and so-cial structure: A macrosociological approach. Cam-bridge, UK: Cambridge University Press.

Berger, J., Fisek, M. H., Norman, R. Z., & Zelditch, M.,Jr. (1977). Status characteristics and social interac-tion: An expectation–states approach. New York:Elsevier.

Berger, J., & Webster, M., Jr. (2006). Expectations, sta-tus, and behavior. In P. J. Burke (Ed.), Contemporary


social psychological theories (pp. 268–300). Stan-ford, CA: Stanford University Press.

Burke, P. J. (1991). Identity processes and social stress.American Sociological Review, 56, 836–849.

Burke, P. J. (1996). Social identities and psychosocialstress. In H. B. Kaplan (Ed.), Psychosocial stress: Per-spectives on structure, theory, life course, and meth-ods (pp. 141–174). San Diego: Academic Press.

Burke, P. J., & Harrod, M. M. (2005). Too much of agood thing? Social Psychology Quarterly, 68, 359–374.

Cahill, S. E., & Eggleston, R. (1994). Managing emo-tions in public: The case of wheelchair users. SocialPsychology Quarterly, 57, 300–312.

Charmaz, K., & Milligan, M. J. (2006). Grief. In J. E.Stets & J. H. Turner (Eds.), Handbook of the sociol-ogy of emotions (pp. 516–543). New York: Springer.

Clark, C. (1997). Misery and company: Sympathy in ev-eryday life. Chicago: University of Chicago Press.

Collins, R. (2004). Interaction rituals. Princeton, NJ:Princeton University Press.

Cooley, C. H. (1964). Human nature and the social or-der. New York: Schocken Books. (Original workpublished 1902)

Durkheim, E. (1965). The elementary forms of the reli-gious life. New York: Free Press. (Original work pub-lished 1912)

Ellestad, J., & Stets, J. E. (1998). Jealousy and parent-ing: Predicting emotions from identity theory. Socio-logical Perspectives, 41, 639–668.

Emerson, R. M. (1962). Power–dependence relations.American Sociological Review, 27, 31–40.

Franks, D. (2006). The neuroscience of emotions. In J.E. Stets & J. H. Turner (Eds.), Handbook of the soci-ology of emotions (pp. 38–62). New York: Springer.

Goffman, E. (1967). Interaction ritual: Essays on face-to-face behavior. Garden City, NY: Anchor Books.

Gordon, S. L. (1981). The sociology of sentiments andemotion. In M. Rosenberg & R. H. Turner (Eds.), So-cial psychology: Sociological perspectives (pp. 562–592). New York: Basic Books.

Gross, J. J., Richards, J. M., & John, O. P. (2006). Emo-tional regulation in everyday life. In D. K. Snyder, J.A. Simpson, & J. N. Hughes (Eds.), Emotion regula-tion in couples and families: Pathways to dysfunctionand health (pp. 13–35). Washington, DC: AmericanPsychological Association.

Hegtvedt, K. A. (1990). The effects of relationship struc-ture on emotional responses to inequity. Social Psy-chology Quarterly, 53, 214–228.

Hegtvedt, K. A., & Killian, C. (1999). Fairness andemotions: Reactions to the process and outcomes ofnegotiations. Social Forces, 78, 269–303.

Heise, D. R. (1979). Understanding events: Affect andthe construction of social action. Cambridge, UK:Cambridge University Press.

Heise, D. R., & Weir, B. (1999). A test of symbolicinteractionist predictions about emotions in imag-ined situations. Symbolic Interaction, 22, 139–161.

Hektner, J. M., Schmidt, J. A., & Csikszentmihalyi, M.

(2007). Experience sampling method: Measuring thequality of everyday life. Thousand Oaks, CA: Sage.

Hochschild, A. R. (1979). Emotion work, feeling rules,and social structure. American Journal of Sociology,85, 551–575.

Hochschild, A. R. (1983). The managed heart: Com-mercialization of human feeling. Berkeley: Universityof California Press.

Houser, J. A., & Lovaglia, M. J. (2002). Status, emo-tion, and the development of solidarity in stratifiedtask groups. Advances in Group Processes, 19, 109–137.

Jasso, G. (2006). Emotion in justice processes. In J. E.Stets & J. H. Turner (Eds.), Handbook of the sociol-ogy of emotions (pp. 321–346). New York: Springer.

Kemper, T. D. (1978). A social interactional theory ofemotions. New York: Wiley.

Kemper, T. D. (1987). How many emotions are there?:Wedding the social and autonomic components.American Journal of Sociology, 93, 263–289.

Kemper, T. D. (1991). Predicting emotions from socialrelations. Social Psychology Quarterly, 54, 330–342.

Kemper, T. D. (2006). Power and status and the power-status theory of emotions. In J. E. Stets & J. H.Turner (Eds.), Handbook of the sociology of emo-tions (pp. 87–113). New York: Springer.

Kemper, T. D., & Collins, R. (1990). Dimensions ofmicrointeraction. American Journal of Sociology, 96,32–68.

Kollock, P. (1994). The emergence of exchange struc-tures: An experimental study of uncertainty, commit-ment, and trust. American Journal of Sociology, 100,313–345.

Lawler, E. J. (2001). An affect theory of social exchange.American Journal of Sociology, 107, 321–352.

Lawler, E. J., & Yoon, J. (1996). Commitment in ex-change relations: A test of a theory of relational cohe-sion. American Sociological Review, 61, 89–108.

Lawler, E. J., & Yoon, J. (1998). Network structure andemotion in exchange relations. American Sociologi-cal Review, 63, 871–894.

Lewis, H. B. (1971). Shame and guilt in neurosis. NewYork: International Universities Press.

Lovaglia, M. J., & Houser, J. A. (1996). Emotional reac-tions and status in groups. American Sociological Re-view, 61, 867–883.

Marx, K. (1990). Capital. New York: Penguin. (Origi-nal work published 1867)

McCall, G. J., & Simmons, J. L. (1978). Identities andinteractions. New York: Free Press.

Mead, G. H. (1934). Mind, self, and society. Chicago:University of Chicago Press.

Molm, L. D. (1997). Coercive power in exchange. Cam-bridge, UK: Cambridge University Press.

Peterson, G. (2006). Cultural theory and emotions. In J.E. Stets & J. H. Turner (Eds.), Handbook of the sociol-ogy of emotions (pp. 114–134). New York: Springer.

Pierce, J. L. (1995). Gender trials: Emotional lives incontemporary law firms. Berkeley: University of Cal-ifornia Press.


Pollak, L. H., & Thoits, P. A. (1989). Processes in emo-tional socialization. Social Psychology Quarterly, 52,22–34.

Ridgeway, C. L. (1994). Affect. In M. Foschi & E. J.Lawler (Eds.), Group processes: Sociological analy-ses (pp. 205–230). Chicago: Nelson-Hall.

Ridgeway, C. L. (2006). Expectation states theory andemotions. In J. E. Stets & J. H. Turner (Eds.), Hand-book of the sociology of emotions (pp. 347–367).New York: Springer.

Ridgeway, C. L., & Johnson, C. (1990). What is the re-lationship between socioemotional behavior and sta-tus in task groups? American Journal of Sociology,95, 1189–1212.

Robinson, D. T., Smith-Lovin, L., & Tsoudis, O. (1994).Heinous crime or unfortunate accident?: The effectsof remorse on responses to mock criminal confes-sions. Social Forces, 73, 175–190.

Robinson, D. T., Smith-Lovin, L., & Wisecup, A. K.(2006). Affect control theory. In J. E. Stets & J. H.Turner (Eds.), Handbook of the sociology of emo-tions (pp. 179–202). New York: Springer.

Rosenberg, M. (1990). Reflexivity and emotions. SocialPsychology Quarterly, 53, 3–12.

Scheff, T. J. (1988). Shame and conformity: Thedeference–emotion system. American SociologicalReview, 53, 395–406.

Scheff, T. J. (1990). Socialization of emotion: Pride andshame as causal agents. In T. D. Kemper (Ed.), Re-search agendas in the sociology of emotions (pp. 281–304). Albany: State University of New York Press.

Scheff, T. J. (1994). Bloody revenge: Emotions, nation-alism, and war. Boulder, CO: Westview Press.

Scheff, T. J., & Retzinger, S. M. (1991). Emotions andviolence: Shame and rage in destructive conflicts.Lexington, MA: Lexington Books.

Scherer, K. R., Wallbott, H. G., & Summerfield, A. B.(Eds.). (1986). Experiencing emotion: A cross-cultural study. Cambridge, UK: Cambridge Univer-sity Press.

Schmitt, C. S., & Clark, C. (2006). Sympathy. In J. E.Stets & J. H. Turner (Eds.), Handbook of the sociol-ogy of emotions (pp. 467–492). New York: Springer.

Sennett, R., & Cobb, J. (1972). The hidden injuries ofclass. New York: Knopf.

Smith, A. C., & Kleinman, S. (1989). Managing emo-tions in medical school: Students’ contacts with theliving and the dead. Social Psychology Quarterly, 52,56–69.

Stets, J. E. (2003). Justice, emotion, and identity theory.In P. J. Burke, T. J. Owens, R. T. Serpe, & P. A. Thoits(Eds.), Advances in identity theory and research(pp. 105–122). New York: Kluwer Academic/Plenum.

Stets, J. E. (2005). Examining emotions in identity the-ory. Social Psychology Quarterly, 68, 39–56.

Stets, J. E. (2006). Identity theory and emotions. In J. E.Stets & J. H. Turner (Eds.), Handbook of the sociol-ogy of emotions (pp. 203–223). New York: Springer.

Stets, J. E., & Tsushima, T. (2001). Negative emotionand coping responses within identity control theory.Social Psychology Quarterly, 64, 283–295.

Stets, J. E., & Turner, J. H. (Eds.). (2006). Handbook ofthe sociology of emotions. New York: Springer.

Stryker, S. (2002). Symbolic interactionism: A socialstructural version. Caldwell, NJ: Blackburn Press.(Original work published 1980)

Stryker, S. (2004). Integrating emotion into identity the-ory. Advances in Group Processes, 21, 1–23.

Summers-Effler, E. (2004a). Radical saints and reformerheroes: Culture, interaction, and emotional intensityin altruistic social movement organizations. Unpub-lished doctoral dissertation, University of Pennsylva-nia.

Summers-Effler, E. (2004b). A theory of the self, emo-tion, and culture. Advances in Group Processes, 21,273–308.

Summers-Effler, E. (2004c). Defensive strategies: Theformation and social interactions of patterned self-destructive behavior. Advances in Group Processes,21, 309–325.

Summers-Effler, E. (2006). Ritual theory. In J. E. Stets &J. H. Turner (Eds.), Handbook of the sociology ofemotions (pp. 135–154). New York: Springer.

Swann, W. B., Jr., Rentfrow, P. J., & Guinn, J. S. (2003).Self-verification: The search for coherence. In M. R.Leary & J. P. Tangney (Eds.), Handbook of self andidentity (pp. 367–383). New York: Guilford Press.

Tangney, J. P., & Dearing, R. L. (2002). Shame andguilt. New York: Guilford Press.

Thamm, R. A. (2006). The classification of emotions. InJ. E. Stets & J. H. Turner (Eds.), Handbook ofthe sociology of emotions (pp. 11–37). New York:Springer.

Thoits, P. A. (1990). Emotional deviance: Researchagendas. In T. D. Kemper (Ed.), Research agendas inthe sociology of emotions (pp. 180–203). Albany:State University of New York Press.

Thye, S. R., Yoon, J., & Lawler, E. J. (2002). The theoryof relational cohesion: Review of a research program.Advances in Group Processes, 19, 139–166.

Turner, J. H. (1999). Toward a general sociological the-ory of emotions. Journal for the Theory of SocialBehavior, 29, 132–162.

Turner, J. H. (2000). On the origins of human emo-tions: A sociological inquiry into the evolution ofhuman affect. Stanford, CA: Stanford UniversityPress.

Turner, J. H. (2002). Face-to-face: Towards a sociologi-cal theory of interpersonal behavior. Stanford, CA:Stanford University Press.

Turner, J. H. (2006). Psychoanalytic sociological theo-ries and emotions. In J. E. Stets & J. H. Turner (Eds.),Handbook of the sociology of emotions (pp. 276–294). New York: Springer.

Turner, J. H., & Stets, J. E. (2005). The sociology ofemotions. New York: Cambridge University Press.


C H A P T E R 4

The Affective Brainand Core Consciousness

How Does Neural ActivityGenerate Emotional Feelings?

JAAK PANKSEPP

Joy & Woe are woven fine,A Clothing for the Soul divine;Under every grief & pineRuns a joy with silken twine.

—WILLIAM BLAKE,“Auguries of Innocence”

How does neural activity create affective expe-rience? How feelings are created may be themost important question in basic emotion re-search; although it has barely been addressedempirically, it is the source of heated philo-sophical and psychological debate. Without af-fect, we might not feel alive. Without positiveaffect, there are few reasons for living, and peo-ple in depressive despair often choose deathover life. Without affect, there is neither funnor pain. Affect is the source of all intimacy—the profound interpersonal giving of oneselfexhibited by people who are in love with lifeand others. Affect encourages people to digdeeply into their biological “souls”—to findempathy, to communicate their major concernssincerely, and to hope that their depth of feelingis reciprocated.

Without affect, we humans would have littleto talk about and no special reason to reach outto others. Affect motivates our urge to play andto speak: When one of the highest brain regionsthat encodes sadness, grief, and social bonding,the anterior cingulate, is damaged, people fallinto akinetic mutism. Such unfortunates retainthe physical capacity to speak, but they have nourge or wish to communicate (Devinsky, Mor-rell, & Vogt, 1995).

Although affect and cognition are com-pletely blended in mature psychological experi-ence, brain research and great art provideunique ways to distinguish cognition fromprimary-process affect. All artistic creations re-quire great cognitive skills, but works that donot stir our affects communicate little. In art asin life, affect motivates cognitive flow much as

47

flames (now batteries) allow torches to illumi-nate the darkness. But we must differentiateneuroscientifically between cognition and af-fect, since they reflect substantially different,albeit interactive, aspects of brain organiza-tion (Ciompi & Panksepp, 2004; Panksepp,2003b). Cognition involves the neocorticalprocessing of information gleaned largely fromenvironmental inputs via exteroceptive senses.Affects are not encoded as information. Theyare diffuse global states generated by deepsubcortical brain structures, interacting withprimitive viscerosomatic body (core self) repre-sentations that remain poorly mapped(Panksepp, 1998a, 1998b). There is presentlymore scientific work on the cognitive aspectsthan on the affective nature of emotions, eventhough, down through history, various scholarshave noted that cognitions are handmaidens tothe passions. Cognitions also resolve raw af-fects into higher emotions.

Many questions about emotions remain em-pirically unanswered, but also some remainlargely unasked in this modern era, as psychol-ogy remains timid of engaging with the powerand importance of cross-species neuroscience(Panksepp, 2005b, 2007a). The manner inwhich the brain generates emotional feelings isthe biggest of them all. Frijda states in Chapter5 of this volume: “As far as I know . . . thereexist no detailed hypotheses at the functionallevel of how innate affective stimuli evoke af-fect” (p. 84). How then might we understandaffects in a way that is scientifically credible?Presumably through psychologically guidedbrain research.

TOWARD A TAXONOMYAND NEUROBIOLOGY OF AFFECTS

Affects fill the mind with a large variety ofdesirable and undesirable experienced statesthat are hard to define objectively or to talkabout clearly. Partly this is because raw af-fects are pre-propositional forms of con-sciousness comprising brain and bodily pro-cesses of kaleidoscopic complexity. But thereseem to be several distinct types. Some ac-company major bodily disturbances (e.g.,pain and fatigue); some reflect sensory plea-sures and displeasures (ranging from tasty de-lights to disgust); still others gauge bodilyneed states (e.g., hunger and thirst); andperhaps most mysteriously, certain intrinsic

brain–body arousal states are stronglyvalenced—the emotional affects.

We do not yet agree upon a generallyaccepted taxonomy of primary-process af-fects,1 but the preceding description contains atleast three distinct types: (1) sensory affects,(2) bodily-homeostatic ones, and (3) brainemotional ones. Each is initially expressed atprimary-process levels, but during develop-ment they come to include learned object rela-tions (secondary-process affects) and get linkedto thoughts and other cognitive activities(tertiary-process affects). Until some interna-tional congress cobbles together an agreement,there are bound to be conceptual disagree-ments concerning where certain entities belong.For instance, in a recent commentary I have ar-gued that disgust should not be considered aprimary-process emotion; it fits better in othercategories, and only through learning might itbe deemed a secondary- or tertiary-processemotion (see Panksepp, 2007a, discussingToronchuk & Ellis, 2007b). In other words,when the affective power of primary sensoryor homeostatic disgust/nausea is cognitivelyresymbolized, social disgust may emerge as asocially constructed emotion.

The basic emotional affects are primarybrain/mind processes, similar to seeing a color.One can use a word, like “red,” as a label for acolor, but this word does not explain the expe-rience of seeing red. If someone is blind, theword “red” is meaningless. In order to explainseeing red, one must discover the neurophysio-logical and neurochemical causes of visual ex-perience. Similarly, one cannot use words to ex-plain primary-process raw emotions. Wordscan only be used as second-order symbols todiscuss affective experiences, but they do notadequately capture the fundamental causes offeelings. Like first-order sensory experiences,primary-process affects are best understood ifwe clarify the attending brain functions. Thusneural criteria are needed to define core emo-tional processes, as summarized in Figure 4.1(adapted from Panksepp, 1998a).

The aim of this chapter is to address howsubjectively experienced emotional feelingsmay arise from the activities of neural circuits.As we begin to understand the neurology of af-fect, we can avoid the circularity that usuallyattends verbal descriptions of basic processes.If we understand the neurology of affect, weare beginning to define primary-process experi-ential concepts, even when dealing with nonhu-


man species that have homologous brain sys-tems (Panksepp, 1982, 1998a). When suchbasic phenomenal experiences are defined interms of brain “mechanisms,” we can use af-fective concepts noncircularly in our explana-tory endeavors—something we cannot do sci-entifically as long as affects are defined onlyverbally.

EVIDENCE FOR EMOTIONALAFFECTS IN OTHER ANIMALS

There is substantial experimental evidence toindicate that other mammals are affective crea-tures. Probably all vertebrates that exhibitstrong instinctual emotional behaviors experi-ence affects (Panksepp, 1990a, 2005b, 2007c).The issue of affects among invertebrates re-mains a more open question, but preliminaryevidence for evaluative processes exists for afew species (Panksepp & Huber, 2004).

What is the best neural evidence for affect?First, one can elicit powerful emotional re-sponses by localized electrical stimulationsof specific neural systems (LESSNS). Indeed,

across all mammalian species, specific sub-cortical brain regions are where LESSNS canevoke the most powerful emotional reactions(Hess, 1957). Second, these subcortical struc-tures are homologous throughout the mamma-lian kingdom. So if one stimulates FEAR2 cir-cuits in rats, cats, or primates, all exhibitsimilar fear responses with differences inspecies-typical details (Panksepp, 1990a).Third, comparably localized stimulations ofthe human brain yield congruent affectiveexperiences—felt emotional arousals that typi-cally appear without reason (Heath, 1996).Fourth, human beings report basic emotionalfeelings that are commensurate with animalemotional behaviors (Panksepp, 1985); this,along with the accruing evidence that boththe feeling and behavioral responses emanatefrom homologous areas of the brain, stronglysuggests that other animals also experienceprimary-process emotions, even though theyobviously cannot talk about them. Fifth, by allmeasures imaginable, one can demonstrate thatanimals like and dislike stimulation of brain re-gions that evoke instinctual emotional behav-iors (Panksepp, 1998a, 2005a). Their affective

4. The Affective Brain and Core Consciousness 49

FIGURE 4.1. Proposed neural definition of an emotional system. These seven neural interactions arepostulated to be characteristics of all major primary-process emotional systems of the brain. (1) A fewsensory stimuli can unconditionally access emotional systems, but most inputs are developmentallylearned. (2) Emotional systems can promote coherent instinctual action outputs, as well as (3) modulatesensory inputs. (4) Emotional systems have positive-feedback components that can sustain emotionalarousal after precipitating events have passed. Also, (5) these systems can be modulated by cognitive in-puts and (6) can modify and channel cognitive activities. The important criterion that emotional systemscreate affective states is not included, but it is assumed that arousal of the whole executive circuitry foreach emotion is essential for elaborating emotional feelings within the brain, perhaps by interacting withother subneocortical brain circuits for organismic visceral self-representation. Adapted from Panksepp(1998a). Copyright 1998 by Oxford University Press. Adapted by permission.

preferences are well indexed by conditionedplace preferences and place aversions, as wellas by the tendency to voluntarily turn LESSNSon and off. Sixth, basic instinctual emotionalurges, even ones as complex as social play, re-main intact after radical neodecortication earlyin life (Panksepp, Normansell, Cox, & Siviy,1994); thus the neocortex is not essential forthe generation of primary-process emotional-ity.

Abundant evidence now supports the medialsubneocortical locus of control for instinctualemotional behaviors that are accompanied byraw affective experiences in all mammals thathave been studied with LESSNS. Modern brainimaging of self-referential processes in humans(including affective ownership) also highlightshigher medial limbic regions of the brain(Northoff, Heinzel, de Greck, Bermpohl, &Panksepp, 2006), long implicated in the regu-lation of emotional feelings in humans(MacLean, 1990). Thus the weight of evidenceindicates not only that all other mammals haveaffective lives, but also that their basic emo-tions are homologous across mammalian spe-cies. Many still disagree, but this debate typi-cally continues, because discussants fail toconsider all the evidence and fear even evolu-tionarily justified anthropomorphism. Con-cerned scientists need to situate themselves in atruth diagram (see Figure 4.2) based on theirreading of the evidence (e.g., Panksepp,2005a), as opposed to clinging to belief systemspassed down through generations. I encouragecultivation of the upper left quadrant, which ismore empirically defensible than situating one-self in the traditionally mandated lower rightquadrant.

FROM LOGICAL POSITIVISMTO MODERN REDUCTIONISM:THE SCIENTIFIC DEATH OF AFFECTIN BEHAVIORAL NEUROSCIENCE

In spite of robust evidence, many behavioralneuroscientists do not acknowledge the affec-tive lives of other animals. This bias has deeproots in Judeo-Christian and other dualisticphilosophies, as well as in the 19th-centurybattle between vitalism and materialism. Vital-ism, a prevailing mode of thought prior to themiddle of the 19th century, postulated nonma-terial “vital principles”—life forces such asélan vital. Such beliefs had long preventedpowerful materialistic scientific modes ofthought from gaining the influence they de-served. In particular, physiologists who soughtto study the body as a biophysical entity, and tobase the medical curriculum on solid anatomy,physiology, and biochemistry, wished to dis-card spooky “life forces” from their science.Descendants of these wise revolutionaries even-tually decided to dualistically conflate emo-tional feelings with other nonmaterial “lifeforces,” and the neuroaffective baby wastossed out with the vitalistic bathwater. In fact,affects are thoroughly biological.

The transition to hard-core materialism waseffected when influential continental scientistsjoined forces to form what is now called theBerlin Biophysics Club, led by “heavy hitters”such as Carl Ludwig (1816–1895), Emildu Bois-Reymond (1818–1896), Hermann vonHelmholtz (1821–1894), and Ernst Brücke(1819–1892). These scholars erased nonobjec-tive entities from the landscape of physiologyand medicine (see Greenspan & Baars, 2005).


The True Nature of the World

Our Judgmentsabout the World

Animals ExperienceEmotional Feelings

Animals Have NoEmotional Feelings

Animals ExperienceEmotional Feelings

ValidAnthropomorphism

Type I Error

Animals Have NoEmotional Feelings

Type II Error ValidAnthropodenial

FIGURE 4.2. Truth diagram concerning the validity of anthropomorphism in the study of emotionalfeelings in animals. Perhaps the appropriate project for the 20th century was to avoid Type I errors. Withthe advent of neuroscience, and our search for the actual mechanisms of emotional feelings, we should beequally careful to avoid Type II errors. The weight of evidence currently indicates that the quadrant thatis most concordant with research evidence is the upper left quadrant. Surprisingly, most behavioralneuroscientists seem happy to reside in the lower right quadrant, as they were trained as students.

They succeeded in establishing a rigorous med-ical curriculum on a solid scientific base. Theirvictory led to an evidence-based medicine thatremains the foundation of modern medical ed-ucation. Members of this coterie were not con-cerned with emotional matters, but some oftheir followers were. Pavlov studied underLudwig, and Freud under Brücke. Pavlov nevermarginalized affect in his studies of autonomicreflexes in dogs, and even discovered neuroticbehaviors in his animals after Neva Riverfloods destroyed his lab. Freud eventually de-veloped an emotion-based “metapsychology,”but psychoanalysis eventually lacked the hardstamp of science that exists when one is con-cerned with the evolved mechanisms of thebrain.

Our current, lingering problems arose whenexcessive antivitalistic forces spilled over intoexperimental psychology, where mental con-structs were eventually deemed too spookyto be accepted as scientific. The physiologistJacques Loeb, a peripheral member of theBerlin Biophysics movement, brought antimen-talistic biases to the United States. At the Uni-versity of Chicago, he enthralled John B. Wat-son, the eventual “father of behaviorism.” AtHarvard, B. F. Skinner came under the spell ofanother Loeb protégé, William Crozier. To-gether, Watson and Skinner inaugurated amethodologically rigorous, and eventually doc-trinaire, radical behaviorism. As they brought anew level of sophistication to the environmen-tally controlled analysis of learning, they ac-tively discouraged instinct- and brain-basedpsychologies. Through their influence, radicalbehaviorism, with no room for emotionality,became the focus of inquiry.

Watson (1919) was initially interested inemotions. Skinner (1953) famously claimed“The ‘emotions’ are excellent examples of thefictional causes to which we commonly attrib-ute behavior” (p. 160). Their important contri-butions to methodologically rigorous behavior-al analysis resulted in narrow, nonaffectiveviews of animal behavior (Panksepp, 1990b).As they marginalized instincts (inbuilt tools forbehaving and feeling) in preference to learning,they impoverished our understanding of theevolved emotional and motivational underpin-nings of mind.

Watson and Skinner’s radical “black box” ofbehaviorism overwhelmed academic psychol-ogy until the 1970s, whereupon an oversimpli-fied computer-inspired mentality—information

processing that could be rigorously moni-tored by reaction times and other objectivemeasures—moved to center stage as cognitivescience. This transition did not happen in ani-mal research, where behaviorism continued torule. In the 1970s dramatic shifts in funding(partly resulting from Vietnam War expendi-tures) coaxed many American behaviorists tobecome neuroscientists, with little interest inemotional processes that control animal behav-ior.

To this day, neurobehaviorists are generallyunwilling to discuss affective processes withinanimal brains. This inhibits linkages betweenbiological psychiatry and neuroscientific psy-chology. Most behavioral neuroscientists con-sider it unrealistic to consider that affectivefeelings actually exist in animal brains; thusthey position themselves in the lower rightquadrant of Figure 4.2. Most subscribe to ruth-less neural reductionism, where psychologicalprocesses play no role in the control of behav-ior.

In this context, it is noteworthy that the first“law” of behavior—Thorndike’s (1911) “lawof effect”—envisioned how “satisfying” and“annoying” events could mediate learning.These affective concepts were eventually trans-formed by behaviorists into “positive rein-forcements” and “punishments,” more by self-appointed decisions than through evidence-based discussions. Thus the diverse affective as-pects of the “law of effect” were discarded be-cause they raised the specter of scientificallyunobservable internal processes—apparentnonmaterial principles—within the brain.

Now that we can link affects to observablebrain events (Berridge, 2000; Panksepp, 1998a,2005a; Peciña, Smith, & Berridge, 2006), wemust reconsider that a “law of affect” doescontrol behavior: Animals do seek brain/mindaffective comfort zones and avoid discomforts.Such visions of animal minds could nourish thecurrent affective revolution in human psychol-ogy. Primary brain/mind processes can only bewell studied in animal models. Indeed, the pro-cess of “reinforcement” may arise from fluctu-ations of affective states. Reinforcers may onlybe potent to the degree to which they arouseor diminish affective processes (Panksepp,2005b).

Modern cognitive science has become fasci-nated, even enchanted, with the topic of af-fects, and some researchers recognize that suchprocesses are prereflective (Lambie & Marcel,


2002). Some social constructivists are begin-ning to postulate that raw positive and negativeaffect may be birthrights of the human brain(Barrett, 2006; Russell, 2003). Despite thisturn toward nativism, they rarely consider thepossibility that many positive and negative af-fective birthrights exist in mammalian brains(Panksepp, in press-b).

Thus cognitive scientists have difficulty gen-erating realistic psychobiological hypotheses ofhow emotional feelings arise from brain net-works, above and beyond highlighting brainregions revealed by modern brain imaging—which are correlative, not causal, approaches(Panksepp, 2007c). The potentially misleadinguse of correlative data to generate causal con-clusions also remains common in behavioralneuroscience. For instance, recording of activ-ity from brain dopamine neurons (namely, lis-tening to what these cells are listening to) hasbeen used to generate causal “reward error pre-diction”/teaching-signal hypotheses for whatdopamine neurons do in the brain (Schultz,2006). In fact, such neurophysiological studiestell us much about what dopamine neurons arelistening to, but little about what dopamine re-lease is doing upstream within the brain. Abun-dant evidence that dopamine arousal of brain“reward” circuits actually promote affectiveSEEKING urges, based on many causal experi-ments (Alcaro, Huber, & Panksepp, 2007;Ikemoto & Panksepp, 1999; Panksepp,1998a), is commonly ignored—because suchviews imply that neuroaffective processes lie atthe heart of animal foraging behaviors.

A “dual-aspect monism” approach suggeststhat certain complex neurodynamics createemotional feelings. Accordingly, instinctualbehaviors may provide public evidence for in-ternal affective states.3 This yields many test-able neurochemical hypotheses, derived fromanimal research, for the underpinning ofprimary-process emotional feelings in hu-mans (Panksepp, 1999; Panksepp & Harro,2004).

THE ABC’S OF AFFECT:AFFECTIVE, BEHAVIORAL,AND COGNITIVE NEUROSCIENCEVIEWS OF AFFECTIVE CONTROL

Many modern brain researchers (e.g., Berridge& Robinson, 2003; Craig, 2003; Damasio,2003; Dolan, 2002) have asserted that human

beings experience affects, whereas animals dis-play only emotional behaviors. They suggestvarious “readout” hypotheses of how higherbrain functions in humans create emotionalfeelings from presumably unconscious emo-tional processes of animal brains.

There currently exist three major neurallybased schools of thought regarding how affec-tive experience emerges from neural activities.First, an ultrareductionistic behavioral neuro-science view denies that affective capacities ex-ist as inherent aspects of mammalian life; suchscholars exhibit little tolerance for discussionsof affective control in the governance of animalbehaviors. Affects, if they exist, are assumed tobe epiphenomenal aspects of higher cognitiveactivities in human brains. In contrast, cogni-tive neuroscience envisions affects as inherentlycoupled to higher human cognitive functions;hence other animals, because of their compara-tively modest cognitive abilities, may not becapable of experiencing affects. Along withneurobehaviorists, cognitive neuroscientists of-ten assert that the emotional behaviors of otheranimals are unconscious. Finally, affective neu-roscience proposes that affects emanate fromdeep subcortical structures found in the brainsof all mammals. Let us consider these views inmore detail.

The Behavioral Neuroscience View

Just as traditional black-box behaviorism mar-ginalized the study of feelings, for neurobehav-iorists affects remain scientific nonstarters, be-cause they cannot be directly monitored withobjective neurophysiological measures. Also,animals cannot provide verbal feedback abouttheir feelings. The worrisome fact that most as-pects of nature are as invisible to direct obser-vation as subatomic particles has not con-vinced behaviorists to utilize indirect measures.Still, two behavioral neuroscientists, Gray andRolls, have envisioned how emotional feelingsare created by reinforcement contingencies.

I have critiqued Gray’s (1990) debatable as-sertion that emotional feelings are created,rather than just triggered, by reinforcementcontingencies; alternatively, I have suggestedthat the process of reinforcement is mediatedby affective change within the brain (Panksepp,1990c). Rolls (2005) has further advanced aposition similar to Gray’s, suggesting that rein-forcement is a more fundamental neurobiologi-cal construct than emotional feelings. Rolls


proposes that our human affective experienceemerges from the capacity of our cognitive cor-tex to re-represent various unconscious pro-cesses, shared by animals, in linguistic terms. Itis merely a concept, not a process. This makesaffect phlogiston-like in our attempts to under-stand animal behavior.

The Cognitive Neuroscience View

Damasio (1994, 1999), and LeDoux (1996)have offered two well-resolved cognitive neu-roscience views. Damasio discovered that cog-nitive decision making is impaired when the af-fective structures of orbito-frontal cortex aredamaged. LeDoux helped refine our under-standing of how classical conditioning of fear iselaborated within the amygdala. Both proposemodernized James–Lange “cortical readout”views of affective experience.

Damasio’s “somatic marker” hypothesis, al-though neuroscientifically more sophisticatedthan the classic James–Lange view, essentiallyenvisions affect as the emotional turmoil ofbodily states within the somatosensory repre-sentation areas of the neocortex. As originallyhighlighted by MacLean (1990), most investi-gators, including Damasio, are recognizing theimportance of insular cortex in the genesis ofvarious affective experiences, ranging from dis-gust (Toronchuk & Ellis, 2007) to nicotinecraving (Naqvi, Rudrauf, Damasio, &Bechara, 2007).

Damasio’s superb neuroimaging of the pri-mal affective states of anger, fear, sadness, andjoy (Damasio et al., 2000) has led him to incor-porate animalian subcortical emotional sys-tems as contributors to emotional behaviors(Damasio, 1999), but not, apparently, to emo-tional feelings; he repeatedly states that ani-mals exhibit emotional behaviors, while hu-mans have emotional feelings (Damasio, 2003;critiqued by Panksepp, 2003a). Few cognitiv-ists believe that subneocortical systems cangenerate affective mentality on their own, butonly the physiological and behavioral emo-tional turmoil that gets translated into psycho-logical experience by the higher brain systemsmediating reflective awareness. This view doesnot seem to distinguish adequately betweenraw phenomenal experience and reflectiveawareness, and, like the behavioral neurosci-ence view, it envisions feelings as arising fromcognitive readout of the unconscious emotionalcommotions of the body. Such views discour-

age effective utilization of abundant animaldata on the functions of subcortical emotionalsystems for understanding primary-processhomeostatic and emotional affects in humans(Denton, 2006; Panksepp, 2003b, 2003c).

Some investigators also make the assump-tion that emotional experience is irrelevant forunderstanding animal emotions. For instance,LeDoux (1996) has suggested that human emo-tional feelings arise when subcortical informa-tion about emotional bodily changes reacheshigher working memory and cognitive con-sciousness generators in the dorsolateral fron-tal cortex.

Compounding misplaced confidence in mod-ern cortical readout theories of emotions arethe prevailing claims about the overriding im-portance of the amygdala in emotional feelings.The central nuclei of amygdala are clearly partof the unconditional FEAR circuit (Panksepp,1991), but their capacity to promote fearfulfeelings probably relies on intact hypothalamicand midbrain FEAR circuitry (Panksepp, 1998a,Fig. 10.4).

The Affective Neuroscience View

Abundant animal emotion research indicatesthat emotional feelings arise from long swathsof subcortical tissues that control visceral pro-cesses, long known as the “limbic system.” Theemotional–visceral brain extends from medialfrontal regions throughout the core of mid-brain, and all brain areas in between. Evidencefor medial subcortical emotional behavioralcircuits in the generation of emotional feelingsremains robust for humans (Heath, 1996;Panksepp, 1985) and other animals (Panksepp,1998a, 2005a). To the best of our knowledge,raw emotional feelings arise from ancient pre-propositional subneocortical substrates thatare homologous in all mammals. If so, animalbrain research becomes a critically importantstrategy for deriving detailed causal knowledgeabout the major sources of the emotional af-fects within the human brain. The major repre-sentatives of this school of thought areMacLean (1990) and myself (Panksepp, 1982,1998a).

Denton (2006) has also highlighted how thebasic motivations, such as thirst, air, hunger,and various other needs of the body, are repre-sented deep in the brain. Even though he callssuch homeostatic feelings “emotions” (perhapsa category error), clearly these affective states


correspond to arousals of the primitive brainareas that animal research has revealed as criti-cally important for generating the instinctualbehaviors associated with those bodily drives—a principle central to the affective neuroscienceview of basic feelings (Panksepp, 1998a). If oneimages brain arousals during intense affectivefeelings such as orgasms (Holstege et al.,2003a), as well as feelings like anger, sadness,and fear (Damasio et al., 2000), animaliansubcortical circuits light up. If anything, higherneocortical systems exhibit diminished arousal(Liotti & Panksepp, 2004; Panksepp, 2003b).

As already noted, we can evoke a variety ofspecific instinctual emotional action tendenciesby activating very specific brain regions in allmammals. Animals clearly enjoy some of thesetypes of stimulation and dislike others. Compa-rable LESSNS of human brains yield commen-surate affective experiences that people canverbalize. Also, decorticate animals display thesame basic emotions as neurologically intactanimals (Kolb & Tees, 2000), even ones ascomplex as play (Panksepp et al., 1994). Com-parable evidence is available for human chil-dren (Shewmon, Holmse, & Byrne, 1999).

Solid evidence exists for at least seven basicemotional operating systems in mammalianbrains (Panksepp, 1998a, 2005a). There is in-sufficient space to cover all these systems here,so I discuss two of special importance for socialneuroscience—PANICky separation distress andPLAYful joy. These systems had not receivedneuroscientific attention before our work in the1970s.

THE BRAIN SUBSTRATESOF SADNESS AND PANIC

Our research into the nature of separation dis-tress was initiated in 1972 when the first neu-rotransmitter receptor, for brain opioids, wasdiscovered. We suspected that separation dis-tress had evolved from ancient pain mecha-nisms of the brain, and thus evaluated whetherbrain opioids mediate social bonding. We pre-dicted that endogenous opioids released duringpositive social interactions would facilitate so-cial bonding by alleviating psychic pain in thesame way that exogenous opiates can reducephysical pain. This idea provided an evolu-tionary rationale for why opiates are habit-forming: Social bonding is a naturally addictiveprocess. Social attachments and addictive de-

pendencies share three key attributes: (1) aninitial intense positive affect (“loving”) phase,followed by (2) a tolerance phase with dimin-ished positive feelings, which sets up (3) a pow-erful separation distress phase of opiate with-drawal in drug addiction and physiologies ofgrief following social loss (Panksepp, 1998a,Fig. 13.5). Such hypotheses have been extendedto other high-incentive rewards, including foodtreats (Colantuoni et al., 2002).

Such causal relations between drug addic-tions and social rewards clarify why lonely, dis-enfranchised individuals are powerfully at-tracted to opiate drugs. Addicts commonly lackenough positive human contact to sustain hap-piness and satisfaction, and opiate intake ispartly self-medication (Khantzian, 2003 withfive commentaries). Indeed, the pleasureof touch is partly opioid-mediated (Keverne,Nevison, & Martel, 1997; Panksepp, Bean,Bishop, Vilberg, & Sahley, 1980; Panksepp &Bishop, 1981; Roth & Sullivan, 2006).

The social implications of such ideas are far-reaching. Why waste money putting more andmore people in jails and fighting perpetualwars on drugs, rather than seeking to build so-cial structures that support human connec-tions?

Now we know that the affective qualities ofearly social bonds leave lasting marks on hu-man and animal development (Champagne,Francis, Mar, & Meaney, 2003; Meaney, 2001;Panksepp, 2001). Early attachment failurespromote depression (Heim & Nemeroff,1999), and opiates can exert strong antidepres-sant effects. Indeed, relatively nonaddictiveopiates such as buprenorphine are excellent an-tidepressants (Bodkin, Zornberg, Lukas, &Cole, 1995).

When we initiated this work, discrete brainsystems for social affect were not being enter-tained in neuroscience, although behavioral evi-dence had long been suggestive (Bowlby, 1969).The prevailing behaviorist assumption was thatsocial attachments arise indirectly from thelearned association of social “objects” with pri-mary rewards, especially nourishment andwarmth. Our approach revealed the first lawfulpsychobiological relationship in socialneuroscience—namely, that brain opioids regu-late the intensity of separation distress. This ef-fect was evident in puppies (Panksepp, Herman,Conner, Bishop, & Scott, 1978), young guineapigs (Herman & Panksepp, 1978), chicks(Panksepp, Vilberg, Bean, Coy, & Kastin, 1978),


and subsequently all other mammals tested.Hence human sadness and grief were envisionedin part as opioid withdrawal states—a proposi-tion affirmed in humans a quarter of a centurylater (Zubieta et al., 2003).

Many others have confirmed and extendedour animal findings (e.g., Kalin, Shelton, &Barksdale, 1988; Kehoe & Blass, 1986; Keverneet al., 1997; Moles, Kieffer, & D’Amato, 2004).It is now certain that social-emotional processesare regulated by opioid and oxytocinergic dy-namics (Nelson & Panksepp, 1998), but in re-cent years the oxytocinergic extensions of thework to adult bonding have garnered more pop-ular attention (Carter, Lederhendler, &Kirkpatrick, 1999; Insel, 2003; Young & Wang,2004). Still, mother–infant attachments may bemore closely related to dynamics of PANIC/sepa-ration distress and CARE/maternal-nurturancesystems than any other primary social system,while those of older animals have strong contri-butions from LUST/sexuality and PLAY systems.

Accordingly, in early work we used localizedLESSNS to map out the separation distress sys-tems in guinea pigs and chicks (Bishop, 1984;

Herman, 1979). We were pleased that the trajec-tories of these systems were remarkably similarin these divergent species, as well as in primates(for summaries, see Newman, in press;Panksepp, Normansell, Herman, Bishop, &Crepeau, 1988). Functional brain imaging withpositron emission tomography (PET) has high-lighted these same brain regions during humansadness (Damasio et al., 2000; see Figure 4.3).

During our initial work, no other psychoac-tive drugs commonly used in psychiatry, exceptclonidine, had effects as robust as those ofopioids (Rossi, Sahley, & Panksepp, 1983). Ofthe many other neuropeptides we evaluatedacross the years (a total of more than 30, eachinjected into brain ventricles, since neuro-peptides do not penetrate the brain readily fol-lowing peripheral administration), we foundthat only oxytocin and prolactin had equallyrobust calming effects. Conversely, other pep-tides, especially corticotropin-releasing factor(CRF) and the excitatory amino acid gluta-mate, could robustly promote separation calls(Panksepp, 1998a). Along the way we foundhow many other social behaviors, from simple


FIGURE 4.3. There are remarkable similarities between regions of the guinea pig brain where LESSNSprovoke separation distress calls (data from Herman, 1979; Panksepp et al., 1988) and areas of the hu-man brain that are activated during feelings of sadness (data from Damasio et al., 2000). Areas of subhu-man mammalian brains where separation calls are evoked are the anterior cingulate (AC), the ventralseptal (VS) and dorsal preoptic areas (dPOA), the bed nucleus of the stria terminalis (BN), thedorsomedial thalamus (DMT), and the periaqueductal central gray area of the brainstem (PAG). Bothhuman sadness and separation distress are low-brain-opioid states (Panksepp, 1998a; Zubieta et al.,2003). The correspondence between the brain regions activated during human sadness and those acti-vated during animal separation distress suggests that human feelings may arise from the instinctual emo-tional action systems of ancient regions of the mammalian brain. OB, olfactory bulb; CC, corpus callo-sum; CB, cerebellum. Adapted from Panksepp (2003c). Copyright 2003 by the Association for theAdvancement of Science. Adapted by permission.

gregariousness to playfulness, are regulated bybrain opioids, confirming and extending theopioid theory of social affect (Panksepp, Bean,Bishop, Vilbers, & Sahlberg, 1980; also see ap-plication to autism, Panksepp, Lensing, Le-boyer, & Pouvard, 1991).

Although many details remain to be workedout, these animal findings have provided thefirst entry points for understanding the neuro-chemical and evolutionary bases of pro-socialfeelings in humans. When the PANIC (separa-tion distress) system is quiescent, individualsexperience a “secure emotional base” for theirlife activities. Other animal studies have indi-cated that high opioid tone makes animalsmore confident and dominant (Panksepp,Jalowiec, DeEskinazi, & Bishop, 1985), andoxytocin facilitates feelings of trust and socialsensitivity (Kosfeld, Heinrichs, Zak, Fisch-bacher, & Fehr, 2005). It is important to notethat some of the effects of oxytocin may be in-directly mediated by opioids, since oxytocincan strongly reduce accruing tolerance to en-dogenous opioids (Kovacs, Sarnyai, & Szabo,1998). In contrast, diminished chemistriesof bonding are expected to promote sadness,grief, and ultimately chronic feelings of emo-tional emptiness and depressive despair(Panksepp, Yates, Ikemoto, & Nelson,1991). Indeed, depression may be rapidly re-versed with nonaddictive opioids (Bodkin etal., 1995).

In summary, those areas of the brain wherewe can evoke separation calls with LESSNS(e.g., Herman & Panksepp, 1981; Panksepp etal., 1988) may contain emotional “command”circuitry that generates feelings of social loss inhumans—from sadness to panic, depending onthe intensity of arousal. Opioids, along withoxytocin and prolactin, are brain neurochemi-stries that powerfully promote feelings of con-fidence and the social comfort-enhancingeffects of friendly companionship. However,because of habituation processes (perhaps re-flecting decreased receptor sensitivity), suchfeelings may typically recede into the back-ground of consciousness.

Only when there is a loss of an attachment“object” (a loved one) do feelings of emotionaldependency reemerge as the separation dis-tress/PANIC system becomes aroused, perhapsby rapidly escalating glutamatergic and CRFtone, along with diminished activity of neuro-peptides that engender feelings of social com-fort. Thus the power of companionate love to

promote happy minds arises in part from theseneurochemistries of social attachment.

THE BRAIN SUBSTRATESOF PLAYFUL JOY AND LAUGHTER

To gain a further understanding of positivesocial-emotional systems, we initiated a paral-lel research program on social play in the late1970s. It provided the first working hypothesisof how social joy is created in the brain. Whenone asks young children what they like to domore than anything else, the universal reply is“To play!” This urge is widespread throughoutthe animal kingdom (Burghardt, 2005). Threedecades of research on the nature of rough-and-tumble play in juvenile rats has clarifiedthat PLAY urges arise from primary-processemotional systems of the mammalian brain.This PLAY system probably evolved in mamma-lian brains to help young animals learn aboutthe social structures into which they are born,and to provide them with a host of experience-expectant social-emotional skills that, throughthe power of positive affect, allow them to be-come well-integrated members of their commu-nities. We think that the implications for childrearing practices are profound (Panksepp,2001; Sunderland, 2006).

The available evidence (see Panksepp, Siviy,& Normansell, 1984; Panksepp, 1998a; Pellis& Pellis, 1998; Siviy, 1998; Vanderschuren,Niesink, & Van Ree, 1997) indicates thatrough-and-tumble play circuitry, like allprimary-process emotional systems, is concen-trated in subcortical regions of the brain(Panksepp et al., 1994), especially diencephalicregions. In particular, dorsomedial parafas-cicular regions of the thalamus are importantfor processing the somatosensory stimuli (Siviy& Panksepp, 1987). Within the hypothala-mus, play-induced brain dopamine release(Panksepp, 1993) certainly facilitates the goodfeelings of play (Burgdorf & Panksepp, 2006).However, if one simply pushes this systempharmacologically into chronic overactivity,play invariably diminishes (Beatty, Dodge,Doge, Whike, & Panksepp, 1982). This mayindicate that dopamine must operate in a stateof dynamic, as opposed to static, facilitation inorder for play to occur. Furthermore, play vo-calizations in rats, 50-kHz ultrasonic chirps,are strongly controlled by brain dopamine(Burgdorf, Knutson, Panksepp, & Shippen-


berg, 2001; Burgdorf, wood, Kroes, Moskal, &Panksepp, in press).

Many other neurochemical modulators ofplay have been identified (Panksepp,Normansell, Cox, Crepeau, & Sacks, 1987;Vanderschuren et al., 1997), but nothing asclear-cut as the neuropeptides that regulateseparation distress. Small doses of opiates canconsistently elevate play (Panksepp, Jalowiec,et al., 1985). We have also obtained modestfacilitations of play with peripheral injectionsof the nicotinic acetylcholine receptor antago-nist mecamylamine (Panksepp et al., 1984)and with intracerebral neurotensin andthyrotropin release hormone (Panksepp,1998). Developmentally, it is likely that ac-cess to play has positive effects, such as facili-tating the construction of a fully social brainthrough the experience-dependent maturationof prosocial brain circuits, perhaps in partthrough the genetic activation of neuronalgrowth factors (Gordon, Burke, Akil, Watson,& Panksepp, 2003) and many other gene ex-pressions (Kroes, Panksepp, Burgdorf, Otto,& Moskal, 2006).

We believe that the relative lack of real playamong children of our culture has implicationsfor understanding childhood problemssuch as attention-deficit/hyperactivity disorder,which are routinely treated with dopamine-promoting psychostimulants that may havelong-term effects on brain maturation(Panksepp, 2001). Such drugs may sensitize thebrain, thereby chronically amplifying SEEKINGurges that may provide fertile breeding groundsfor addictive tendencies (Nocjar & Panksepp,2002). With both “pro” and “con” data onthis, we question the wisdom of substitutingplay-inhibiting drug treatment for the natural,neurochemically mediated brain benefits andjoys of physical play (Panksepp, 1998a, 1998b,2007d).

After many years of research, we encoun-tered an intriguing sound, a ~50-kHz ultra-sonic play vocalization, that had previouslybeen heard primarily in the context of sexualsolicitation in adult rats. To a lesser extent, thissound is emitted during the anticipation of allmajor rewards, and we focused on the idea thatthis might be a way to monitor appetitivemotivational intensity (Knutson, Burgdorf, &Panksepp, 2002). Indeed, this response may bethe first easily conditioned, natural measure ofdrug desire (Burgdorf et al., 2001; Panksepp,Knutson, & Burgdorf, 2002). However, more

than anything, the sound is a superb indicatorof social joy that accompanies rough-and-tumble play.

Juvenile play arouses this vocalization morethan any other social activity. We now believethat it may be evolutionarily related to the in-stinctual fixed-action patterns of laughter,which is so common during human play (Scott& Panksepp, 2003). Accordingly, we started totickle juvenile rats, and the 50-kHz play chirpsrose dramatically (Panksepp & Burgdorf,1999, 2003). After years of focused work, weknow much about the underlying circuitry. Itcourses along the ascending mesolimbic dopa-mine systems that innervate the nucleus accum-bens and medial frontal cortex (Burgdorf et al.,2001, 2007).

Perhaps primordial forms of joyful laughterare more prevalent among the playful juvenilesof our fellow animals than scientists eversuspected. Traditionally, laughter has beendeemed a unique trait of humans and perhapschimpanzees (Provine, 2000). Our work bringsthat anthropocentric supposition into question.Tickling (or, more formally, heterospecific handplay) robustly promotes 50-kHz ultrasonicchirping, with characteristics suggestive of an-cestral laughter (Panksepp & Burgdorf, 2003;Panksepp, 2007d). We have evaluated thisidea abundantly without disconfirmations.Accordingly, we propose that this is a neuro-scientifically workable model for understand-ing the neural details of social joy in mamma-lian brains.

Here are a dozen reasons to consider rela-tions between childhood laughter in our speciesand play chirping in young rats (more fullydiscussed in Panksepp & Burgdorf, 2003;Burgdorf & Panksepp, 2006):

1. The 50-kHz ultrasonic chirping is pro-voked robustly by positive social inter-change (namely, play), and even morepowerfully by tickling.

2. All negative affective stimuli, includingfearful foot shock, cat smell, scary places,bright lights, and unpleasant handling, re-duce the chirping response.

3. Juvenile rats chirp more than mature rats.4. Rats that chirp most abundantly during

tickling also play the most with each other.5. Animals approach hands that have tickled

them more than those that have only pet-ted them.

6. Young rats readily learn to run down alleys


and to press levers to get tickled, but not tobe picked up and petted.

7. Rats show conditioned place preferencesfor tickling and seek out stimuli that havebeen associated with tickling (e.g., if tick-led by a hand that smelled of coffeegrounds, they are attracted to that smell inother situations).

8. The tickling response classically conditionsrapidly to cues that predict tickling.

9. Children are more ticklish in certain bodilyregions (e.g., ribs), and young rats have“tickle skin” concentrated at the nape ofthe neck, where they typically direct theirown play activities.

10. Juvenile rats prefer to spend more timewith adults that still chirp a lot than thosethat do not.

11. Tickle-induced chirping is hard-wired inthe brain and is a temperamental charac-teristic of rats, for it can be successfullyincreased or decreased within four gen-erations of selective breeding (Burgdorf,Panksepp, Brudzynski, Kroes, & Moskal,2005).

12. Tickling circuitry courses along thedopamine-modulated self-stimulation cir-cuitry, and such brain regions are arousedwhen humans laugh. (For a more detailedsummary of evidence relevant for under-standing the sources of human and ratlaughter, see Panksepp, in press-d).

This model system offers a simplified way tounderstand the primary-process social joy sys-tems of the mammalian brain. We are currentlyin the 17th generation of selectively breedinganimals for high and low chirping in responseto tickling. Behavioral phenotyping of theselines has been initiated (see Burgdorf &Panksepp, 2006; Burgdorf et al., 2007;Harmon et al., in press). Our microarray-basedscreening of how play modifies gene expressionpatterns in the brain is already indicating thatthe genetic “orchestra” is dramatically modi-fied by play (a third of 1,200 brain genes signif-icantly changed activity).

We anticipate that a fuller understanding ofthe neurochemistries of joy will yield new in-sights and new molecular targets for treatinghuman depression. Comparable possibilitiesexist for all the other basic emotional systems,and some relevant clinical issues are discussedelsewhere (Panksepp, 2004, 2006).

AFFECTIVE NEUROSCIENCEONTOLOGY

All basic psychological processes are thor-oughly dependent on brain biophysical pro-cesses, working in concert with body, environ-ment, and culture. Within our current scientificunderstanding of mind, there are no emotionalfeelings that are independent of neuronal activ-ities, operating in the complexities of bodieswithin environments. In spite of abundant evi-dence, however, the neuroscientific and psy-chological communities remain unconvincedthat other animals experience affects of greatrelevance for understanding primary-processhuman feelings. Many in behavioral neurosci-ence consider affects to be epiphenomenal—conceptual flotsam of idle anthropomorphism(Figure 4.2).

Much philosophical ink has been spilledconcerning the existence and causal efficacyof mental processes. Besides battalions of re-ductionist neuroscientists, some philosophersdo not accept mentality as “really” doing anycausal work in our physiochemical universe.Others advocate a foundational “panpsychic”or “panexperiential” view of the physical uni-verse. For example, they appear to accept theso-far empirically inconsequential “mind-dust”as a property of matter—a possibility thatWilliam James (1890) entertained, tongue incheek, in his Principles of Psychology. Withinour dust-to-dust existence, no “seeds” of men-tality have yet been detected in brute matter.Still, some brilliant people are willing to arguethat if no fundamental panpsychic “mind-dust” stuff exists in the physical universe—ifno kind of protoconsciousness exists at thequantum level—then surely our own experi-ences of mentality must be epiphenomenal.Such arguments hold little water. Emergenceis a pervasive aspect of nature, and new func-tions do arise from interacting levels of organi-zation.

Even if the physical universe has no intrin-sic qualities of mentality, no intrinsic values,evolved brains most certainly do. Affectiveexperience is an emergent function of com-plex network-level neurodynamics, intimatelyconnected to body and world. Primary-process affects arise from complex neuronalnetwork interactions (see Figure 4.1), yieldingintrinsic brain/mind tools that help animalssurvive.


ON THE EVOLUTIONOF AFFECTIVE ADAPTATIONS

Organisms have competed for resources sincethe beginning of life. Initially the competitionmay have been largely reflexive, with no men-tality. However, experiential specializationspresumably led to increasingly successful com-petition in unpredictable environments. With-out interanimal competition, neural networksfor generating affects probably would not haveevolved; reflexive zombie-like behavioral con-trols might have sufficed. With ever-increasinglevels of competition, there were new choicesto be made, and affects may have been idealheuristics for effective decision making. Affectspresent survival “compass” bearings for thecognitive apparatus. The most competitivebrains developed affective heuristics—whichare completely biological, largely neuronal, butwith strong bodily and cultural connections—to facilitate rapid decision making for individ-ual benefits, as well as empathy and groupsharing of survival-related resources. In short,affects are real brain functions. Their most im-portant infrastructure is neuronal.

Accordingly, a great deal of cognitive activitymay be less conscious than the affects. Con-sider the dispersion of dopamine cells at themesodiencephalic junction. The more medialventral tegmental area (VTA) neurons project-ing to nucleus accumbens mediate urges to seekresources, and thereby to learn about the con-tingencies of the world (Ikemoto & Panksepp,1999; Panksepp, 1998a). The more lateral neu-rons in substantia nigra (SN) have similarelectrophysiological properties during learning(Schultz, 2006), but they are more important inlaying down habitual ways of behaving aslearning is consolidated into well-worn paths(habit learning), especially within the dorsalstriatum (in the caudate–putamen complex).We may ask this question: Which of these tworegions sustains experienced feelings more? Bythe measure of self-stimulation, animals showvigorous self-stimulation of the medial VTA re-gions, and little of SN regions. Thus arousal ofmedial, more ancient brain regions is valuedmore than that of evolutionarily more recentlateral regions.

Instinctual feelings (basic tools for living)presumably present survival knowledge to thecognitive apparatus. Increasing positive feel-ings, of diverse kinds, inform the cognitive ap-

paratus that organisms are behaving with ahigh probability of facilitating survival. Vari-ous bad feelings inform the cognitive apparatusthat ongoing actions may harm survival(Cabanac, 1992). For instance, disgust is an af-fective response that helps prevent severe nau-sea once sickness has taken hold. Both feelingsguard against death.

Affects, although genetically ingrained, arelargely intrinsic anticipatory states that facili-tate learning by interacting with information-processing abilities of the cognitive apparatus.Perhaps the primal feeling is pain. For instance,when I feel excruciating, throbbing pain in myoccasionally gouty toe, it controls my behaviorunambiguously. When I take pain-dissolvingmedicines, I can again walk with relative ease.Pain rarely informs the rest of our brain thatwe are at death’s door. It typically tells us toprotect an injured body part, optimizing heal-ing. Without painful feelings, all humanswould die prematurely.

Evolutionarily, the primal capacity to experi-ence pain may have promoted the possibility ofgenerating other negative feelings (such as sep-aration distress), which promote grief andloneliness (Panksepp, 2005c, 2005d). After theemergence of maternal nurturance, theoxytocin-driven sexual affects of females mayhave been recrafted to generate the pleasures ofmaternal interactions (Panksepp, 1998a).

Affects control behavior because they arepart and parcel of brain network functions,which are increasingly accepted in psychiatryand psychoanalysis (Panksepp, 1999, 2004,2006). The discarding of affective concepts inbehavioral neuroscience may be tantamount tocrippling our ability to discuss the nature oflarge-scale adaptive network functions of thebrain, and aborting our ability to envision howan understanding of affective mentality pro-vides many avenues to reconceptualizing psy-chiatric disorders (Johnson-Laird, Mancini, &Gangemi, 2006; Panksepp, 2006). If so, abetter neural understanding of affect is a foun-dational issue for all of psychology and psychi-atry, as well as for anyone interested in thefunctional organization of the mammalianbrain. This is not to deny that an enormousamount of autonomic and cognitive brain ac-tivity can proceed unconsciously and habitu-ally (Wegner, 2002), but to assert, in agreementwith Freud, that affects are quintessentially ex-perienced and real, even in dreams.


Although discussions about affect are be-coming more common in psychology (e.g.,Lambie & Marcel, 2002; Russell, 2003), rele-vant fact finding remains sparse. That awaitsmore intellectual commerce (joint causal re-search) between psychologists and neuro-scientists. Indeed, understanding the nature ofaffect neuroscientifically may become a majorgateway for understanding consciousness(Watt & Pincus, 2004). To do that well, weneed new neurophilosophical perspectives.

DUAL-ASPECT MONISM

We have identified emotional systems throughLESSNS that arouse distinct emotional behav-iors in animals. People experience affects whenthese same brain networks are stimulated, andanimals are never neutral about such arousals.They either seek to avoid such states or toreexperience them. Because these affective andemotional responses emerge from homologousbrain areas, and because the choice behaviorsof animals are commensurate with the affectsthat human beings verbalize, the underlyingcircuits may be envisioned in terms of aSpinozan “dual-aspect monism.” The “dual-aspect” refers to the fact that certain emotionalarousals concurrently generate instinctual emo-tional urges as well as complementary emo-tional feelings. The “monism” rejects Cartesiandualism by asserting that both behavior andexperience are manifestations of the physicalbrain. This view coaxes psychology to acceptneuroscience as an essential partner for clarify-ing primary-process affective issues (Panksepp,2005a). This view also restores many other ani-mals back into the circle of affective mentallife, from which they were excluded by the rad-ical behaviorists of the 20th century.

If we share homologous instinctual circuitswith our animal cousins, as subcortical ana-tomical, neurochemical, and functional stud-ies suggest, we mammals should have similarprimary-process affective feelings—albeit notsecondary- and tertiary-process cognitiveabilities, which arise from our differential neo-cortical complexities. We humans have muchmore random-access-type neocortical information-processing and memory space for re-representing primary-process experiences.These capacities can parse affective space intocomplex conceptual structures, many of which

may never be illuminated by research on otheranimals. Our cortical abilities can constructcultural differences and subtleties that exist inno other species. Many emotional complexitiesemerge from these epigenetic cortical re-representations of raw experience.

Dual-aspect monism is a robust workingstrategy for understanding primary-process af-fective experience, but not necessarily associ-ated cognitive processes. Dual-aspect monismmaintains, as did William James, that a sub-stantial part of the affectively experienced sideof life is completely entwined with neuronalprocesses that generate certain instinctualbehavior patterns. For example, a cat that ap-pears angry because of LESSNS-induced in-stinctual displays of rage actually experiencesanger within the brain. Such large-scale net-work activities do their work well because theyare experienced. Thus classical behavioral rein-forcers may need to promote affective-experiential states within the brain. Accordingto this approach, the instinctual behavioralneurodynamics are isomorphic with theprimary-process affective dynamics, but areonly preconditions for associated cognitivemoldings of experience that ultimately serve af-fective needs. In this view, there is no logicalproblem in asserting that affects, because theyare completely neuronal, do regulate behavior.

In summary, affects can be conceived of as aprimordial form of consciousness that arose asa way for the brain to represent biological sur-vival values to the evolutionarily emergent cog-nitive deliberations of the brain (Denton, 2006;Panksepp, 1998a, 2005a; Merker, 2007). Howcognitive activities reciprocally mold and parseaffective experience remains an open empiricalquestion. If the model described above is on theright track, then the facilitation of learnedbehavioral change may be dependent in part onthe emergence of brain affective changes whenanimals are rewarded and punished. Howmight this be tested?

If affective change is pharmacologically pre-vented, the progression of learning should bechanged. The likelihood that the reinforcementis fundamentally affective helped us interpreta strange behavior we discovered almost 30years ago—the inhibition of extinction by lowdoses of morphine in juvenile rats trainedto run for social rewards (Panksepp &DeEskinazi, 1980). Since social reward ispartly due to brain opioid release (Panksepp,


1981), it was reasonable that a reduced braincapacity to experience affective change duringextinction would drastically retard extinctionof learned habits. Indeed, low doses of mor-phine that dramatically diminish separationdistress blocked extinction almost com-pletely. In similar work with food reward,antiextinction effects were half the size as seenwith social reward (Panksepp & DeEskinazi,1980).

AFFECTIVE FOUNDATIONS OF CORECONSCIOUSNESS AND CORE SELF

The foundation of phenomenal experience, in-cluding primordial affective mentality, is called“core consciousness” (Panksepp, 2007). Theunderstanding of conscious awareness—a re-flective higher brain function—may require theprior clarification of the neural mechanisms ofphenomenal consciousness, which may come intwo primal varieties: (1) the here-and-now per-ceptual fields (the movie in the head, so tospeak) and (2) affective experiences (the manynonpropositional valenced ways we feel alive).Our fledgling attempts to understand phenom-enal affective experience may require somekind of a “core self” conceptualization—a co-herent organismic representation of the organ-ism within the brain, to provide a body-feelingfoundation for affective states (Damasio, 1999;Panksepp, 1998a, 1998b, 2005a).

My preferred theoretical vision is that thevisceral nervous system has a coherent repre-sentation of the inside of the body laid out inprimitive action coordinates of the soma-tic musculoskeletal system. In contrast toDamasio’s view of the core self as closely linkedto the somatosensory cortex, I envision a con-siderably more ancient neurosymbolic repre-sentation of organisms within the ancient me-dial subcortical regions that abound in visceralorgan representations (Hess, 1957). Both ap-proaches seek to specify the brain circuits,neurochemistries, and neurodynamics that helpconstruct core feelings, but I see the visceralhomunculus as highly coordinated with in-stinctual emotional action coordinates(Panksepp, 1998b, 2000a).

The neurodynamics of such a core SELF mayvary as a function of basic emotional and moti-vational homeostatic and sensory states, allgenerating distinct affective feelings (see

above). Such a vision needs to be given sub-stance with a huge amount of research, butexisting electrophysiological approaches maybe currently beset by seemingly insurmount-able challenges (Panksepp, 2000b). At present,the easiest way to untangle this Gordian knot isto focus on the likelihood that there existssome kind of neurochemical, especially neuro-peptide, coding of emotional affects (Panksepp& Harro, 2004), as well as of certain sensoryaffects that have clear behavioral indicators(Peciña et al., 2006; Steiner, Glaser, Hawilo, &Berridge, 2001), within this complex homun-cular representation of the SELF.

The neurochemical aspects are presentlycomparatively easy to address empirically,because of the discovery of hundreds ofneuropeptides concentrated within visceral,subcortical brain regions (Panksepp, 1993;Tohyama & Takatsuji, 1998). Many of theseneuropeptides may establish distinct neurody-namics, which are psychologically experiencedas distinct affective states, within the sub-cortical trajectory of the core SELF (Panksepp& Harro, 2004). The concurrent abundancesof neuropeptides in peripheral visceral organsand in visceral parts of the brain suggest conti-nuities between the neurochemical regulationof certain bodily functions and brain emotionalfunctions. A striking example is the fact thatboth oxytocin and prolactin, which regulatethe metabolic nourishment of infants by regu-lating the production and dispensation of milk,are important for the infant’s social-emotionalnourishment. Both neuropeptides are powerfulinhibitors of separation distress (Panksepp,1998a) and facilitators of a variety of maternal,sexual, and other prosocial behaviors (Carter,2003; Insel, 2003; Young & Wang, 2004). Asalready noted, maternal circuits may be evolu-tionary outgrowths of brain systems that pro-mote female sexuality, helping explain theirshared neurochemical controls. Separation dis-tress may be an outgrowth of pain systems(Panksepp, 2003c)

Similar theoretical linkages could be madewith many other central and peripheral neuro-peptides, allowing us to envision harmoniousevolutionary continuities between peripheralbodily and central affective functions(Panksepp, 1993). Thus the regional localiza-tions of brain neuropeptide systems—theirabilities to regulate specific classes of emo-tional behaviors—again affirm that many


varieties of positive and negative affects arecritically linked to the dynamics of ancientsubcortical regions of the brain (Panksepp,1998a, 2005a). The recent social-constructivistidea that primal positive and negative affectsare brain substrates from which the complexityof emotional life is created (Barrett, 2006; Rus-sell, 2003) awaits neurobiological support. Nocausal tests have been proposed, while abun-dant predictions are available for a more re-solved vision of primary-process emotional sys-tems (Panksepp, 2005a, 2007). Why is it notgood news for social constructivists that at theprimary-process level there are various coreaffective tools from which the complexities ofindividual lives and cultures could be con-structed?

The possibility that core emotional affectsreflect the neurodynamics of various brain sys-tems that control instinctual emotional actionsis eminently testable in humans via psy-chopharmacological approaches (Depue &Collins, 1999; Panksepp, 1999; Panksepp &Harro, 2004). The core SELF hypothesis is notas easily evaluated, and may first require abun-dant research characterizing the primitiveviscerosomatic homuncular topography thatexists in subcortical regions of the brain (Hess,1957). Because of modern functional magneticresonance imaging (fMRI) techniques, this featis already achievable for higher reaches of self-representation (Northoff et al., 2006)—inbrain regions where neural firing rates aremuch higher than in the depths of the visceralnervous system. However, PET studies havehad some success in monitoring brain dynam-ics during diverse forms of emotional and moti-vational arousal (Damasio et al., 2000;Denton, 2006).

CONCLUDING REMARKS

Can we ever scientifically understand how af-fective feelings are generated by neural activi-ties? Given the wealth of supportive data, it isevident that mammals and many other verte-brates do experience emotional, sensory, andhomeostatic affects, and hence provide work-able model systems via which the infrastructurecan be objectively characterized. Past LESSNSdemonstrate that powerful emotional re-sponses emanate from specific subcorticalbrain systems that are homologous across spe-cies. Arousals of homologous sites in the hu-

man brain evoke affective feelings that arecommensurate with the instinctual emotionalactions evoked in other animals (Heath, 1996;Panksepp, 1985). Also, animals that have beensurgically deprived of their neocortices con-tinue to display spontaneous primary-processemotional behaviors with apparent affectivecontents, whether they are rats (Kolb & Tees,2000; Panksepp et al., 1994), cats (Villablanca& Marcus, 1972), or human beings (Shewmonet al., 1999).

I have summarized historical issues that per-suaded psychologists to reject affective consid-erations in favor of behavior-only and even-tually cognition-only research strategies. Incontrast, affective neuroscience approaches cannow envision how subcortical networks con-stitute the affective crucible of the brain(Panksepp, 1998a, 2005a). I have focused ontwo emotional systems, SADNESS/PANIC andPLAY/JOY, that highlight how social isolation/loss and the urge to enjoy positive social inter-actions are fundamental needs of the mamma-lian brain.

At birth, the basic emotional systems, andaccompanying raw affects, have few intrinsic“object relations.” However, life experiencesrapidly promote development linkage to worldevents and objects. Such cognitivized affectsenhance behavioral guidance, allowing animalsto compete and cooperate with each other suc-cessfully in order to survive. All raw affectsinterpenetrate the cognitive apparatus in thehigher reaches of the limbic system—corticessuch as cingulate, insula, temporal, and hippo-campal paleocortices—allowing animals togenerate more flexible adaptive responses topreviously neutral stimuli. The neocortex co-mes to reflect on such experiences, providingultimately conscious awareness.

Finally, I have considered the hypothesis thataffective feelings arise when primary-processemotional systems interact with a shared coreSELF that is laid out in primitive visceral and so-matic coordinates. Clarification of such ancientneural networks may help illuminate the bio-logical nature of the “soul.”

Even if this core affective SELF model re-mains imprecise, it does not diminish the per-suasive power of the evidence supporting vari-ous basic emotional systems in the mammalianbrain. These systems are homologous acrossspecies; they are aroused by homologousneurochemicals, and they produce homologousbehaviors and commensurate affects. These


data, coupled with the abundance of converg-ing positive evidence suggesting that animalsare affective creatures (from Darwin, 1872/1998, to Panksepp, 1998a, 2005a), highlightthe great value that animal research can pro-vide to the science of mind and its emotionaldisorders.

ACKNOWLEDGMENTS

I appreciate comments on an earlier draft of thischapter by Luce Bivins and Terry McGuire. Thiswork was partially supported by the Hope for De-pression Research Foundation.

NOTES

1. “Primary processes” reflect brain/mind functionsthat are intrinsic in the normal biological organiza-tion of the brain. These are contrasted with “second-ary processes,” which reflect learned resymboliza-tions and other developmental landscapes of eachorganism (such as thoughts about the world), andhigher “tertiary processes,” which may be uniquelyhuman, constituting thoughts about thoughts andemotions. As basic affects come to be molded bylearned individual and cultural influences, the shadesand subtlety of emotions become enormous. An un-derstanding of primary-process emotions, best stud-ied in animal models where the necessary neuro-scientific work can be done, is essential for makingsense of higher-order derivative processes, which canonly be well studied through the analysis of humanexperience. This chapter focuses on primary-processemotions.

2. To highlight the fact that we are talking about thenecessary neural substrates for various emotionalconstructs rather than any comprehensive sets of at-tributes of such systems operating in the real world,the labels for the various systems are given in smallcapital letters. In other words, the labels refer to spe-cific neural systems that remain poorly understood.

3. It is noteworthy that John Dewey held the view thatemotional feelings occur prior to instinctual actions,and dissipate as a consequence of instinctual dis-charge. This conjecture may reflect the fact that hu-mans often actively inhibit their instinctual displays,and that this inhibition leaves the instinctual urgewithin the brain active. It may be the inhibited activ-ity of such urges, which is commonly evident inenculturated humans, that prolongs emotional affectin the absence of overt instinctual displays duringwhich covert instinctual arousal may still be present.At the same time, each emotion is indeed self-limiting, and hence the display gradually does dissi-pate the underlying, neurochemically based emo-tional “energy.” In this way, Dewey’s view can beseen to be a harmonious variant of the present view.

REFERENCES

Alcaro, A., Huber, R., & Panksepp, J. (2007). Behavior-al functions of the mesolimbic dopaminergic system:An affective neuroethological perspective. Brain Re-search Reviews, 56, 283–321.

Barrett, L. F. (2006). Are emotions natural kinds? Per-spectives on Psychological Science, 1, 28–58.

Beatty, W. W., Dodge, A. M., Dodge, L. J., Whike, K., &Panksepp, J. (1982). Psychomotor stimulants, socialdeprivation and play in juvenile rats. Pharmacology,Biochemistry and Behavior, 16, 417–422.

Berridge, K. C. (2000). Measuring hedonic impact inanimals and infants: Microstructure of affective tastereactivity patterns. Neuroscience and BiobehavioralReviews, 24, 173–198.

Berridge, K. C., & Robinson, T. E. (2003). Parsing re-ward. Trends in Neurosciences, 9, 507–513.

Bishop, P. (1984). Brain and opiate modulation of avianaffective vocalizations. Unpublished doctoral disser-tation, Bowling Green State University.

Bodkin, J. A., Zornberg, C. L., Lukas, S. E., & Cole, J.O. (1995). Buprenorphine treatment of refractory de-pression. Journal of Clinical Psychopharmacology,15, 49–57.

Bowlby, J. (1969). Attachment: Vol. 1. Attachment andloss (2nd ed.). New York: Basic Books.

Burgdorf, J., Knutson, B., Panksepp, J., & Shippenberg,T. (2001). Evaluation of rat ultrasonic vocalizationsas predictors of the conditioned aversive effects ofdrugs. Psychopharmacology, 155, 35–42.

Burgdorf, J., & Panksepp, J. (2006). The neurobiologyof positive emotions. Neuroscience and Biobe-havioral Reviews, 30, 173–187.

Burgdorf, J., Wood, P. L., Kroes, R. A., Moskal, J. R., &Panksepp, J. (2007). Neurobiology of 50-kHz ultra-sonic vocalizations in rats: Electrode mapping, le-sion, and pharmacological studies. BehaviouralBrain Research, 182, 274–283.

Burghardt, G. M. (2005). The genesis of animal play.Cambridge, MA: MIT Press.

Cabanac, M. (1992). Pleasure: The common currency.Journal of Theoretical Biology, 155, 173–200.

Carter, C. S. (2003). Developmental consequences ofoxytocin. Physiology and Behavior, 79, 383–397.

Carter, C. S., Lederhendler, I., & Kirkpatrick, B. (Eds.).(1999). The integrative neurobiology of affiliation.Cambridge, MA: MIT Press.

Champagne, F. A., Francis, D. D., Mar, A., & Meaney,M. J. (2003). Variations in maternal care in the rat asa mediating influence for the effects of environmenton development. Physiology and Behavior, 79, 359–371.

Ciompi, L., & Panksepp, J. (2004). Energetic effects ofemotions on cognitions: Complementary psychobio-logical and psychosocial finding. In R. Ellis & N.Newton (Eds.), Consciousness and emotions (Vol. 1,pp. 23–55). Amsterdam: Benjamins.

Colantuoni, C., Rada, P., McCarthy, J., Patten, C.,Avena, N. M., Chadeayne, A., et al. (2002). Evidence


that intermittent, excessive sugar intake causes en-dogenous opioid dependence. Obesity Research 10,478–488.

Craig, A. D. (2003). How do you feel? Interoception:The sense of the physiological condition of the body.Nature Reviews Neuroscience, 3, 655–666.

Damasio, A. R. (1994). Descartes’ error. New York:Putnam.

Damasio, A. R. (1999). The feeling of what happens.New York: Harcourt, Brace.

Damasio, A. R. (2003). Looking for Spinoza. Orlando,FL: Harcourt, Brace.

Damasio, A. R., Grabowski, T. J., Bechara, A.,Damasio, H., Ponto, L. L. B., Parvizi, J., et al. (2000).Subcortical and cortical brain activity during the feel-ing of self-generated emotions. Nature Neuroscience,3, 1049–1056.

Darwin, C. (1998). The expression of the emotions inman and animals (3rd ed.). New York: Oxford Uni-versity Press. (Original work published 1872)

Denton, D. (2006). The primordial emotions: Thedawning of consciousness. New York: Oxford Uni-versity Press.

Depue, R. A., & Collins, P. F. (1999). Neurobiology ofthe structure of personality: Dopamine, facilitationof incentive motivation, and extraversion. Behavioraland Brain Sciences, 22, 491–517.

Devinsky, O., Morrell, M. J., & Vogt, B. A. (1995).Contributions of anterior cingulate cortex to behav-ior. Brain, 118, 279–306.

Dolan, R. J. (2002). Emotion, cognition, and behavior.Science, 298, 1191–1194.

Gordon, N. S., Burke, S., Akil, H., Watson, J., &Panksepp, J. (2003). Socially induced brain fertiliza-tion: Play promotes brain derived neurotrophic fac-tor expression. Neuroscience Letters, 341, 17–20.

Gray, J. A. (1990). Brain systems that medicate bothemotion and cognition. Cognition and Emotion, 4,269–288.

Greenspan, R. J., & Baars, B. J. (2005). Consciousnesseclipsed: Jacques Loeb, Ivan P. Pavlov and the rise ofreductionistic biology after 1900. Consciousness andCognition, 14, 219–230.

Harmon, K. M., Cromwell, H. C., Burgdorf, J., Moskal,J. R., Brudzynski, S. M., Kroes, R. A., & Panksepp, J.(in press). Rats selectively bred for low levels of50 kHz ultrasonic vocalizations exhibit alterationsin early social motivation. Developmental Psycho-biology.

Heath, R. G. (1996). Exploring the mind–body relation-ship. Baton Rouge, LA: Moran.

Heim, C., & Nemeroff, C. B. (1999). The impact ofearly adverse experiences on brain systems involvedin the pathophysiology of anxiety and affective disor-ders. Biological Psychiatry, 46, 1509–1522.

Herman, B. H. (1979). An exploration of brain socialattachment substrates in guinea pigs. Unpublisheddoctoral dissertation, Bowling Green State Univer-sity.

Herman, B. H., & Panksepp, J. (1978). Effects of mor-phine and naloxone on separation distress and ap-proach attachment: Evidence for opiate mediation ofsocial affect. Pharmacology, Biochemistry andBehavior, 9, 213–220.

Herman, B. H., & Panksepp, J. (1981). Ascendingendorphinergic inhibition of distress vocalization.Science, 211, 1060–1062.

Hess, W. R. (1957). The functional organization of thediencephalon. New York: Grune & Stratton.

Holstege, G., Georgiadis, J. R., Paans, A. M., Meiners,L. C., van der Graaf, F. H., & Reinders, A. A. (2003).Brain activation during human male ejaculation.Journal of Neuroscience, 23, 9185–9193.

Ikemoto, S., & Panksepp, J. (1999). The role of nucleusaccumbens DA in motivated behavior: A unifying in-terpretation with special reference to reward-seeking.Brain Research Reviews, 31, 6–41.

Insel, T. R. (2003). Is social attachment an addictive dis-order? Physiology and Behavior, 79, 351–357.

James, W. (1890). Principles of psychology. New York:Holt.

Johnson-Laird, P. N., Mancini, F., & Gangemi, A.(2006). A hyper-emotion theory of psychological ill-nesses. Psychological Review, 113, 822–841.

Kalin, N. H., Shelton, S. E., & Barksdale, C. M. (1988).Opiate modulation of separation-induced distress innon-human primates. Brain Research, 440, 285–292.

Kehoe, P., & Blass, E. M. (1986). Opioid-mediation ofseparation distress in 10-day-old rats: Reversal ofstress with maternal stimuli. Developmental Psycho-biology, 19, 385–398.

Keverne, E. B., Nevison, C. M., & Martel, F. L. (1997).Early learning and the social bond. Annals of theNew York Academy of Sciences, 807, 329–339.

Khantzian, E. J. (2003). Understanding addictive vul-nerability: An evolving psychodynamic perspective.Neuro-Psychoanalysis, 5, 5–56.

Knutson, B., Burgdorf, J., & Panksepp, J. (2002). Ultra-sonic vocalizations as indices of affective states in rat.Psychological Bulletin, 128, 961–977.

Kolb, B., & Tees, C. (Eds.). (2000). The cerebral cortexof the rat. Cambridge, MA: MIT Press.

Kosfeld, M., Heinrichs, M., Zak, P. J., Fischbacher, U.,& Fehr, E. (2005). Oxytocin increases trust in hu-mans. Nature, 435, 673–676.

Kovacs, G. L., Sarnyai, Z., & Szabo, G. (1998).Oxytocin and addiction: A review. Psychoneuro-endicronol, 23, 945–962.

Kroes, R. A., Panksepp, J., Burgdorf, J., Otto, N. J., &Moskal, J. R. (2006). Social dominance–submissiongene expression patterns in rat neocortex. Neurosci-ence, 137, 37–49.

Lambie, J. A., & Marcel, A. J. (2002). Consciousnessand the varieties of emotion experience: A theoreticalframework. Psychological Review, 109, 219–259.

LeDoux, J. E. (1996). The emotional brain. New York:Simon & Schuster.

Liotti, M., & Panksepp, J. (2004). On the neural nature


of human emotions and implications for biologicalpsychiatry. In J. Panksepp (Ed.), Textbook of biologi-cal psychiatry (pp. 33–74). Hoboken, NJ: Wiley-Liss.

MacLean, P. (1990). The triune brain. New York: Ple-num Press.

Meaney, M. J. (2001). Maternal care, gene expression,and the transmission of individual differences instress reactivity across generations. Annual Reviewof Neuroscience, 24, 1161–1192.

Merker, B. (2007). Consciousness without a cerebralcortex: A challenge for neuroscience and medicine.Behavioral and Brain Sciences, 30, 63–134.

Moles, A., Kieffer, B. L., & D’Amato, F. R. (2004). Defi-cit in attachment behavior in mice lacking the mu-opioid receptor gene. Science, 304, 1983–1986.

Naqvi, N. H., Rudrauf, D., Damasio, H., & Bechara, A.(2007). Damage to the insula disrupts addiction tocigarette smoking. Science, 315, 531–534.

Nelson, E. E., & Panksepp, J. (1998). Brain substratesof infant–mother attachment: Contributions ofopioids, oxytocin, and norepinephrine. Neuroscienceand Biobehavioral Reviews, 22, 437–452.

Newman, J. D. (2007). Neural circuits underlying cry-ing and cry responding in mammals. BehaviouralBrain Research, 182, 155–165.

Nocjar, C., & Panksepp, J. (2002). Chronic intermittentamphetamine pretreatment enhances future ap-petitive behavior for drug- and natural-reward: Inter-action with environmental variables. BehaviouralBrain Research, 128, 189–203.

Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F.,& Panksepp, J. (2006). Our brain and its self: Thecentral role of cortical midline structures. Neuro-Image, 15, 440–457.

Panksepp, J. (1981). Brain opioids: A neurochemicalsubstrate for narcotic and social dependence. In S.Cooper (Ed.), Progress in theory in psychopharma-cology (pp. 149–175). London: Academic Press.

Panksepp, J. (1982). Toward a general psychobiologicaltheory of emotions. Behavioral and Brain Sciences, 5,407–467.

Panksepp, J. (1985). Mood changes. In P. Vinken, G.Bruyn, & H. Klawans (Eds.), Handbook of clinicalneurology (pp. 271–285). Amsterdam: Elsevier.

Panksepp, J. (1990a). The psychoneurology of fear:Evolutionary perspectives and the role of animalmodels in understanding human anxiety. In G. D.Burrows, M. Roth, & R. Noyes, Jr. (Eds.), Hand-book of anxiety (pp. 3–58). Amsterdam: Elsevier/North-Holland.

Panksepp, J. (1990b). Can “mind” and behavior be un-derstood without understanding the brain?: A re-sponse to Bunge. New Ideas in Psychology, 8, 139–149.

Panksepp, J. (1990c). Gray zones at the emotion–cognition interface: A commentary. Cognition andEmotion, 4, 289–302.

Panksepp, J. (1991). Affective neuroscience: A concep-tual framework for the neurobiological study of emo-

tions. In K. T. Strongman (Ed.), International reviewof emotion (Vol. 1, pp. 59–99). Chichester, UK:Wiley.

Panksepp, J. (1993). Neurochemical control of moodsand emotions: Amino acids to neuropeptides. In M.Lewis & J. Haviland (Eds.), Handbook of emotions(pp. 87–107). New York: Guilford Press.

Panksepp, J. (1998a). Affective neuroscience: The foun-dations of human and animal emotions. New York:Oxford University Press.

Panksepp, J. (1998b). The periconscious substrates ofconsciousness: Affective states and the evolutionaryorigins of the SELF. Journal of ConsciousnessStudies, 5, 566–582.

Panksepp, J. (1999). Emotions as viewed by psycho-analysis and neuroscience, An exercise in consilience.Neuro-Psychoanalysis, 1, 15–38.

Panksepp, J. (2000a). Affective consciousness and theinstinctual motor system: The neural sources of sad-ness and joy. In R. Ellis, & N. Newton (Eds.), Thecaldron of consciousness: Motivation, affect and self-organization (pp. 27–54). Amsterdam: Benjamins.

Panksepp, J. (2000b). The neurodynamics of emotions:An evolutionary–neurodevelopmental view. In M. D.Lewis & I. Granic (Eds.), Emotion, self-organization,and development (pp. 236–264). New York: Cam-bridge University Press.

Panksepp, J. (2001). The long-term psychobiologicalconsequences of infant emotions: Prescriptions forthe 21st century. Infant Mental Health Journal, 22,132–173.

Panksepp, J. (2003a). Damasio’s error. Consciousnessand Emotion, 4, 111–134.

Panksepp, J. (2003b). At the interface of affective,behavioral and cognitive neurosciences: Decodingthe emotional feelings of the brain. Brain and Cogni-tion, 52, 4–14.

Panksepp, J. (2003c). Feeling the pain of social loss. Sci-ence, 302, 237–239.

Panksepp, J. (Ed.). (2004). Textbook of biological psy-chiatry. Hoboken, NJ: Wiley-Liss.

Panksepp, J. (2005a). Affective consciousness: Coreemotional feelings in animals and humans. Con-sciousness and Cognition, 14, 30–80.

Panksepp, J. (2005b). On the embodied neural nature ofthe core emotional affects. Journal of ConsciousnessStudies, 5, 158–184.

Panksepp, J. (2005c). Social support and pain: Howdoes the brain feel the ache of a broke heart? Journalof Cancer Pain and Symptom Palliation, 1, 59–65.

Panksepp, J. (2005d). Feelings of social loss: The evolu-tion of pain and the ache of a broken heart. In R. Ellis& N. Newton (Eds.), Consciousness and emotions(Vol. 1, pp. 23–55). Amsterdam: Benjamins.

Panksepp, J. (2006). Emotional endophenotypes in evo-lutionary psychiatry. Progress in Neuro-Psycho-pharmacology and Biological Psychiatry, 30, 774–784.

Panksepp, J. (2007a). The neuroevolutionary and


neuroaffective psychobiology of the prosocial brain.In C. Crawford & D. Krebs (Eds.), Handbook ofevolutionary psychology (pp. 145–162). New York:Erlbaum/Taylor & Francis.

Panksepp, J. (2007b). Criteria for basic emotions: IsDISGUST a primary “emotion”? Cognition andEmotion, 21, 1819–1828.

Panksepp, J. (2007c). Neurologizing the psychology ofaffects: How appraisal-based constructivism and ba-sic emotion theory can co-exist. Perspectives on Psy-chological Science, 2, 281–296.

Panksepp, J. (2007d). Neuroevolutionary sources oflaughter and social joy: Modeling primal humanlaughter in laboratory rats. Behavioural Brain Re-search, 182, 231–244.

Panksepp, J., Bean, N. J., Bishop, P., Vilberg, T., &Sahley, T. L. (1980). Opioid blockade and socialcomfort in chicks. Pharmacology, Biochemistry andBehavior, 13, 673–683.

Panksepp, J., & Bishop, P. (1981). An autoradiographicmap of 3H diprenorphine binding in the rat brain: Ef-fects of social interaction. Brain Research Bulletin, 7,405–410.

Panksepp, J., & Burgdorf, J. (1999). Laughing rats?:Playful tickling arouses high frequency ultrasonicchirping in young rodents. In S. Hameroff, D.Chalmers, & A. Kazniak (Eds.), Toward a science ofconsciousness III (pp. 231–244). Cambridge, MA:MIT Press.

Panksepp, J., & Burgdorf, J. (2003). “Laughing” ratsand the evolutionary antecedents of human joy?Physiology and Behavior, 79, 533–547.

Panksepp, J., & DeEskinazi, F. G. (1980). Opiates andhoming. Journal of Comparative and PhysiologicalPsychology, 94, 650–663.

Panksepp, J., & Harro, J. (2004). The future ofneuropeptides in biological psychiatry and emotionalpsychopharmacology: Goals and strategies. In J.Panksepp (Ed.), Textbook of biological psychiatry(pp. 627–660). Hoboken, NJ: Wiley-Liss.

Panksepp, J., Herman, B., Conner, R., Bishop, P., &Scott, J. P. (1978). The biology of social attachments:Opiates alleviate separation distress. Biological Psy-chiatry, 9, 213–220.

Panksepp, J., Herman, B. H., Vilberg, T., Bishop, P., &DeEskinazi, F. G. (1980). Endogenous opioids andsocial behavior. Neuroscience and Biobehavioral Re-views, 4, 473–487.

Panksepp, J., & Huber, R. (2004). Ethological analysesof crayfish behavior: A new invertebrate system formeasuring the rewarding properties of psychostim-ulants. Behavioural Brain Research, 153, 171–180.

Panksepp, J., Jalowiec, J., DeEskinazi, F. G., & Bishop,P. (1985). Opiates and play dominance in juvenilerats. Behavioral Neuroscience, 99, 441–453.

Panksepp, J., Knutson, B., & Burgdorf, J. (2002). Therole of emotional brain systems in addictions: Aneuro-evolutionary perspective. Addiction, 97, 459–469.

Panksepp, J., Lensing, P., Leboyer, M., & Bouvard, M.P. (1991). Naltrexone and other potential new phar-macological treatments of autism. Brain Dysfunc-tion, 4, 281–300.

Panksepp, J., Normansell, L. A., Cox, J. F., Crepeau, L.,& Sacks, D. S. (1987). Psychopharmacology of socialplay. In B. Olivier, J. Mos, & B. F. Brain (Eds.),Ethnopharmacology of social behavior (pp. 132–144). Dordrecht, The Netherlands: Nijhoff.

Panksepp, J., Normansell, L. A., Cox, J. F., & Siviy, S.(1994). Effects of neonatal decortication on the so-cial play of juvenile rats. Physiology and Behavior,56, 429–443.

Panksepp, J., Normansell, L. A., Herman, B., Bishop, P.,& Crepeau, L. (1988). Neural and neurochemicalcontrol of the separation distress call. In J. D.Newman (Ed.), The physiological control of mam-malian vocalizations (pp. 263–300). New York: Ple-num Press.

Panksepp, J., Siviy, S., & Normansell, L. A. (1984). Thepsychobiology of play: Theoretical and methodologi-cal perspectives. Neuroscience and Biobehavioral Re-views, 8, 465–492.

Panksepp, J., Vilberg, T., Bean, N. J., Coy, D. H., &Kastin, A. J. (1978). Reduction of distress vocaliza-tion in chicks by opiate-like peptides. Brain ResearchBulletin, 3, 663–667.

Panksepp, J., Yates, G., Ikemoto, S., & Nelson, E.(1991). Simple ethological models of depression:Social-isolation induced “despair” in chicks andmice. In B. Olivier, J. Mos, & J. L. Slangen (Eds.),Animal models in psychopharmacology (pp. 161–181). Basel: Birkhauser-Verlag.

Peciña, S., Smith, K., & Berridge, K. C. (2006). Hedonichot spots in the brain. The Neuroscientist, 12, 500–511.

Pellis, S. M., & Pellis, V. C. (1998). Structure–functioninterface in the analysis of play fighting. In M. Bekoff& J. A. Byers (Eds.), Animal play: Evolutionary,comparative, and ecological perspectives (pp. 115–140). Cambridge, UK: Cambridge University Press.

Provine, R. R. (2000). Laughter: A scientific investiga-tion. New York: Viking.

Rolls, E. T. (2005). Emotions explained. Oxford: Ox-ford University Press.

Rossi, J., III, Sahley, T. L., & Panksepp, J. (1983). Therole of brain norepinephrine in clonidine suppressionof isolation-induced distress in the domestic chick.Psychopharmacology, 79, 338–342.

Roth, T. L., & Sullivan, R. M. (2006). Examining therole of endogenous opioids in learned odor-stroke as-sociations in infant rats. Developmental Psycho-biology, 48, 71–78.

Russell, J. A. (2003). Core affect and the psychologicalconstruction of emotion. Psychological Review, 110,145–173.

Schultz, W. (2006). Behavioral theories and the neuro-physiology of reward. Annual Review of Psychology,57, 87–115.


Scott, E., & Panksepp, J. (2003). Rough-and-tumbleplay in human children. Aggressive Behaviour, 29,539–551.

Shewmon, D. A., Holmse, D. A., & Byrne, P. A. (1999).Consciousness in congenitally decorticate children:Developmental vegetative state as self-fulfillingprophecy. Developmental Medicine and Child Neu-rology, 41, 364–374.

Siviy, S. M., & Panksepp, J. (1987). Juvenile play in therat: Thalamic and brain stem involvement. Physiol-ogy and Behavior, 41, 39–55.

Skinner, B. F. (1953). Science and human behavior. NewYork: Macmillan.

Steiner, J. E., Glaser, D., Hawilo, M. E., & Berridge, K.C. (2001). Comparative expression of hedonic im-pact: Affective reactions to taste by human infantsand other primates. Neuroscience and BiobehavioralReviews, 25, 53–74.

Sunderland, M. (2006). Science of parenting. London,UK: Doring Kindersley Books.

Thorndike, E. L. (1911). Animal intelligence. NewYork: Macmillan.

Tohyama, M., & Takatsuji, K. (Eds.). (1998). Atlas ofneuroactive substances and their receptors. NewYork: Oxford University Press.

Toronchuk, J. A., & Ellis, G. F. R. (2007). Disgust: Sen-

sory affect or primary emotional system? Cognitionand Emotion, 21, 1799–1818.

Vanderschuren, L. J., Niesink, R. J., & Van Ree, J. M.(1997). The neurobiology of social play behavior inrats. Neuroscience and Biobehavioral Reviews, 21,309–326.

Villablanca, J., & Marcus, R. (1972). Sleep-wakefulness, EEG and behavioral studies of chroniccats without neocortex and striatum: The ‘dien-cephalic’ cat. Archives Italiennes de Biologie, 110,348–382.

Watson, J. B. (1919). A schematic outline of the emo-tions. Psychological Review, 26, 165–196.

Watt, D. F., & Pincus, D. I. (2004). Neural substrates ofconsciousness: Implications for clinical psychiatry. InJ. Panksepp (Ed.), Textbook of biological psychiatry(pp. 627–660). Hoboken, NJ: Wiley-Liss.

Wegner, D. M. (2002). The illusion of conscious will.Cambridge, MA: MIT Press.

Young, L. J., & Wang, Z. (2004). The neurobiology ofpair bonding. Nature Neuroscience, 7, 1048–1054.

Zubieta, J. K., Ketter, T. A., Bueller, J. A., Xu, Y.,Kilbourn, M. R., Young, E. A., et al. (2003). Regu-lation of human affective responses by anteriorcingulate and limbic mu-opioid neurotransmission.Archives of General Psychiatry, 60, 1145–1153.


C H A P T E R 5

The Psychologists’ Point of View

NICO H. FRIJDA

The psychological point of view is that “emo-tion” represents a meaningful and necessaryconcept. This statement is not a truism. Onecan hold that “emotion” is no more than achapter heading (Bentley, 1928), or a folk con-ception based on preconceptions rather thanfacts. As Magda Arnold wrote in 1970, thereare a number of “perennial problems in thefield of emotion.” The definition of emotion isone of them, and so is the question of whetherit represents a coherent domain of study.

Indeed, the word “emotion” itself may givereasons for doubt. It may not represent a natu-ral class of phenomena. It also is of relativelyrecent origin. In its present sense, it dates fromabout 1649, the publication date of Descartes’sPassions de l’Âme. Not all languages possess amore or less equivalent concept (Wierzbicka,1995). Nevertheless, most languages do, andfor good reason. The word fills a need in point-ing to particular phenomena of feeling andbehavior. Which phenomena these are tran-spires from the term itself, and from its equiva-lents in other languages. Around the time thatDescartes (1649/1970) first used “emotions”

for les emotions de l’âme, it meant “uproar” or“social unrest,” with implications of unrulinessand vehemence (Cayrou, 1924). Ancient Greekused the term pathèma, and early French andEnglish used passion—that is, mental events in-volving passivity. “Passions” differ from “ac-tions,” in that the individual feels the actions orinclinations for them passively coming overhim or her, rather than flowing from their ini-tiative. The Latin near-equivalent, affectus, hada similar connotation: an event or experiencethat one is affected by. Sanskrit employedbhava, meaning something like a “state ofmind that becomes”—that is, one that move-ments flow from (Shweder & Haidt, 2000).The notions of passivity, being affected, anduproar all refer to the phenomenon that certainfeelings and behavioral inclinations tend to in-trude upon ongoing thought and behavior.They seek to assume control, tend to persistover time, and may do so even when prevailingconditions make it advisable for them not to doso.

These control shifts appear largely due tofeelings. Pathèma, affectus, emotion, and

68

bhava all are felt. People profess to feelings,and perhaps do so everywhere. Feelings indeedloom large in definitions of emotion (seeKleinginna & Kleinginna, 1981). Humans ap-pear to “express” them in movements andsounds that do not appear to serve instrumen-tal purposes, but still function in communicat-ing states of mind to others. Both feelings andbehavior pivot around acceptance and rejec-tion of people (including oneself), objects, orevents. They pivot around inclinations and dis-inclinations to deal with them, around pro-fessed feelings of like and dislike or of pleasureand pain, and around judgments of good andbad (Nussbaum, 2001). The distinction be-tween good and bad appears to be a universalcomponent of language (Wierzbicka, 1995).

People, objects, and events, and the feelingsthey evoke, moreover, do not leave one cold.They affect one’s body and one’s cognitivefunctioning. One may tremble, become con-fused, or believe what one knows to be untrue.

The psychologists’ point of view thus pointsto a domain of phenomena of feelings, behav-iors, and bodily reactions. These phenomenarequire explanations different from those re-quired for explaining habit, voluntary action,and sensory impressions and thought as such.They appear to demand explanatory conceptssuch as pleasure and pain, evaluation, controlpriorities, preferences, and desires.

This psychological perspective has two inter-connected implications. First, its focus is onphenomena manifested or felt by individuals.Second, the explanations for these phenomenarequire hypotheses about intrapersonal causalprocesses. “Emotion,” first of all, serves as ashorthand for, or pointer to, intrapersonal pro-cesses and mechanisms. It thereby also pointsto a human and animal system architecturethat enables such mechanisms (Sloman, 1987).

The notion of emotion thereby serves to re-solve discrepancies between what people do orfeel and the events surrounding them; betweenthe immediate cues for why they do what theydo and what they actually do; between whatthey do and what they say; between what theydo and what seems most appropriate, mostuseful, most reasonable, and best organized;and between what they do and what they pro-fess to know they should do. It serves to helpour understanding that different people mayreact in different ways to the same situations,and that one given person may react differentlyto one given situation on different occasions.

THE TASK OF THE PSYCHOLOGYOF EMOTION

The psychology of emotion has the task of de-scribing features and patterns of phenomenathat qualify as “emotional,” and of explainingthem in terms of underlying processes and thespecies’ and individuals’ process repertoires.Explanations include operating characteristicssuch as sensitivities and thresholds, processingand response repertoires, and attentional andenergetic resources. They also include analysisof the various kinds of information that theseprocesses have to work with. Some of that in-formation is received from the environment;some is generated by an individual’s bodily pro-cesses; and some comes from the individ-ual’s store of representations of facts, cognitiveschemas, and behavioral skills. The psychologi-cal point of view focuses on intraindividualprocesses, even if these may represent conver-gence points for influences of sociocultural ori-gin.

Explanations can be formulated at differentlevels of description. Dennett (1987) has distin-guished an intentional or phenomenologicallevel (description in terms of feelings, aims, de-sires, and expectations), a functional or prop-erly psychological level (description in terms ofhabits, programs, information-processing pro-cedures, and memory stores), and a hardwareor neural and neurohumoral level. Within eachlevel, there are gradations of integration of ele-ments, and corresponding levels of description.There are neurons, neuron circuits, and neuro-humoral systems. There are muscle twitches,movements, and actions at many scales of hier-archical complexity. There are momentary feel-ings of discontent, and feelings that the worldis heading toward its ruin. The descriptionsmay not immediately reflect the underlying in-tegration and complexity. Suffering from fore-seeing the world’s doom is not necessarily amore complex feeling than suffering from painin one’s toe. Only the foundational processesdiffer in complexity and, of course, so do theexplications of the feelings and their implica-tions for behavior.

These relationships complicate the task ofpsychology. Operation, properties, and effectsof more integrative phenomena usually cannotbe meaningfully reduced to some combinationof those of the founding elements, or be pre-dicted from these. They may possess emergentproperties, such as a sense of agency or of being

5. The Psychologists’ Point of View 69

affected (Metzinger, 2003). Founding elementsmay change their operation within the integrat-ed whole, in what has been called “circularcausation” (Lewis, 1996). For instance, the ef-fects of emotional feelings on thought maydrastically differ from the mere effects of soma-tic feelings, even if the latter are among thefoundations of the former.

Among its explanatory tools, the psychologyof emotion includes the dynamic interactionsof the individual with his or her environment.These interactions bring sensory stimuli infrom environment and body, produce effects onhow smoothly the individual’s faculties andprocesses function, and produce effects on theenvironment as well as feedback from that en-vironment.

Psychological explanations of emotion phe-nomena thus are composed of three terms: thestructure (properties, capacities, propensities)of the individual; incoming and stored infor-mation; and online dynamic interaction withthe environment. How the emotional phenom-ena emerge from what corresponds to thosethree terms raises several of the other perennialproblems that Arnold (1970) alluded to.

The three terms can be combined in an indef-inite number of ways. Theoretical orientationsdiffer in emphasizing one or another of them.Classical behaviorism favored structural sim-plicity, together with simple rules of informa-tion acquisition; McDougall’s (1923) innate in-stincts theory formed a contrast. In emotiontheory proper, Mandler’s (1984) view explainsemotions with few structural provisions: Emo-tions are viewed as arousal responses triggeredby goal interruption. Russell’s (2003) core af-fect theory likewise requires only pleasure–painand activation provisions as basic elements.Both theories contrast with Ekman’s (1992)highly structured neurocultural theory of basicemotions, and with that of Öhman and Mineka(2001)—which, for instance, posits an evolvedfear module, innately sensitive to particularsets of stimuli. Striving for structural simplicityis reductionist. It seeks to follow the principleof Occam’s razor: Be as stingy with structure aspossible. Complexity is motivated by the effortto account as fully as possible for phenomenalvariety. Dynamic interaction seeks to combinestructural simplicity and richness of phenom-ena.

All efforts at description and explanationaim to find regularities among responses, and

in the links among antecedents, responses,and further outcomes. Ideally, the regularitiesindicate causal laws, such as “All emotional re-actions are instigated by appraisal processes,”or “Anger is the inevitable consequence of frus-tration.” However, it may well be that truelaws cannot be found at any integrative level ofdescription. Elster (1999a) has argued that atsuch levels, explanatory rules must and do havea limited scope, and are subject to unspecifiedrestrictions. They are rules, not laws; Elstercalls them “mental mechanisms.” Their limita-tions are due to the fact that the conditionsunder which each rule applies cannot be strictlyspecified because of the chaotic nature of psy-chological causation. Elster illustrates the argu-ment with the following pair of rules: Rule 1,“Tyranny decreases the likelihood of rebel-lion,” and Rule 2, “Tyranny increases the like-lihood of rebellion.” Both are true; however,their respective conditions cannot be exhaus-tively specified because of their unboundedcontext dependence, which can reverse the bal-ances.

WHAT ARE EMOTIONS?

Emotions thus do exist. This does not imply,however, that these phenomena justify a singleconcept or form a natural class. The underlyingprocesses or provisions may not possess suffi-cient unity and specificity. In that sense, onemay repeat the question: Do emotions exist?Do the phenomena justify assuming a distinctfunction of “Emotion,” separate from “Cogni-tion” and “Conation,” as an older psychologywould have it?

Specificity and unity of emotions are com-monly assumed; the assumptions are not neces-sarily correct, however. Nor are they univer-sally held. James (1884) did not adhere to theone about specificity. Emotional behavior, inhis view, does not differ from other behaviorcalled forth by key stimuli. James supposed itto originate in the cerebral cortex, just like allother behavior; it took Cannon’s work onsubcortical mechanisms to prove him wrong.Neither did he give emotional feeling any spe-cial status; he viewed it as just consisting ofbody sensations resulting from feedback frommovements and other bodily responses. Otherinvestigators also argued that there is nothingspecific in emotion experience. For instance,


“emotional experience is a highly variable state[and] often partakes of the complicated natureof a judgment” (Landis & Hunt, 1932, quotedby Hebb, 1949, p. 237). Duffy (1941) took adifferent nonspecificity position. In her view,what are called “emotions” are just the highand low levels of activation, with no qualitativeproperty separating them from the emotionallyneutral middle range.

One may also deny the unity assumption.Dumas (1948) and Hebb (1949) denied thatgoal-directed reactions such as angry attack orfearful flight have anything in common withthe diffuse ones of mere upset or excitement, orof emotional shock and behavioral disorgani-zation. More recently, LeDoux (1996) has sug-gested that the various emotions may not in-volve shared mechanisms.

There are good reasons to raise the unity andspecificity issues. The central problem is themodest agreement between investigators aboutwhich features so-called emotions might share.Bentley (1928), in the paper mentioned earlier,concluded: “Well, emotion is at least a topic! Itis something to talk about and to disagreeupon” (p. 21). The main problem is that thefeatures mentioned in my first section (passiv-ity, feeling, driving force, etc.), as well asthe various behavioral, experiential, and physi-ological response aspects, do not stronglycovary. They each define overlapping but non-identical sets of responses. Theorists select dif-ferent features for their definitions, and as aconsequence, these definitions are numerousand may diverge widely (Kleinginna &Kleinginna, 1981).

In some definitions, the essence of emotionsis feeling, notably that of pleasure or pain. Af-fective valence is commonly regarded as acriterial aspect. Emotions, say Ortony, Clore,and Collins (1988), are valenced reactions, orconscious experiences of events with valencedmeanings (e.g., Elster, 1999a). Affects, pleasureand pain, certainly set the experiences in whichthey occur apart from all other kinds ofexperience—if only because, as feelings, theycannot be readily reduced to something else,such as cognitions or body sensations (Frijda,1986, 2007a). Yet, conversely, many valencedreactions are not usually classed as emotions.Tasting sweet substances merely produces apleasant sensation that usually is not regardedas an emotion. By contrast, some reactions areoften considered emotions even though they do

not involve pleasure or pain. Surprise and won-der are among them. But, precisely for thatreason, Spinoza (1677/1989) did not includethem among the emotions; he saw them asmerely cognitive reactions.

Even feelings may not be considered criterialfor emotions, though. Emotions occur even inlower animals, according to behavioral criteria.Valenced reactions—acceptance and rejection,pursuit and avoidance—occur without aware-ness. They occur in insects and fish. They occurin humans in response to pleasant and unpleas-ant stimuli that are not consciously perceived.Backward masking of such stimuli may lead tosubsequent affect ratings (e.g., Zajonc, 1980),or even to enhanced consumption of likeddrinks (Berridge, 2004).

But the latter findings point to a level ofanalysis that might enlarge the domain of re-sponses that share pleasure or pain: viewing thefeelings as but one kind of expression of the un-derlying processes that Berridge (2004) hascalled “core pleasure” and “core pain.” Theseprocesses are also expressed in readiness forevent acceptance or rejection.

Pleasure and pain, in turn, result from a stilllarger process domain: that of “appraisal.”Appraisal processes are defined as intrasub-jective processes that turn incoming stimulusevents into ones with affective value and fur-ther meaning for well-being and motive satis-faction (Lazarus, 1991). Event appraisal can beconsidered one of the basic abilities of humanand animal systems; emotions can therefore bedefined as processes that involve appraisal.Appraisal processes vary in kind and complex-ity, ranging from immediate, automatic affectarousal (stimuli may directly evoke pleasure orpain because one has innate or acquired dispo-sitions for them to do so; LeDoux, 1996) tocomplex integration of cues for promotion orobstruction in achieving goals and safeguard-ing concerns. The latter are often designated as“cognitive appraisals” (e.g., Lazarus, 1991;Oatley, 1992; Scherer, 2001; Stein & Trabasso,1992). Because of the role of appraisals in emo-tional feelings and action instigations (Elster,1999a; Scherer, 2001), emotions haven beendefined as kinds of judgments (e.g., Nussbaum,2001). However, regardless of complexity,appraisal processes mostly proceed automati-cally and nonconsciously, even if one is oftenaware of their outcomes (Bargh, 1997; Zajonc,1980). Obviously, the extent to which felt and


overt responses are taken to share the processesof appraisal depends on one’s level of analysisand theoretical interpretation.

Another domain that looms large in effortsto define emotions is that of motivation.Emotional behavior as well as felt emotionsstrongly suggest a certain prominence of urgesto act, desires, or motive states (including thoseof loss or decay of motivation, as observed inemotional confusion and depressive apathy).Motivation was central in older emotiontheories, such as those of Thomas Aquinas,McDougall (1923), Tomkins (1962), andWallon (1942), and more recently that ofMagda Arnold (1960). The motivational sideof emotions has been lost from view in more re-cent theoretical attempts.

Yet motive states as occurring in what wecall emotions are rather specific for them. Theyare triggered by events as appraised. They arefelt as urges and lead to impulsive actions,which means that they do not result from plan-ning and are little controlled by foresight. Theycommand actions that are not premeditated.Impulses to act represent a mode of action in-stigation and control differing from that of au-tomatic and habitual behavior, as well as fromthat of planned behavior. Impulsive control ofaction was indeed prominent among the phe-nomena that gave rise to the emotion-like no-tions.

Emotions can therefore be regarded aspassions—as defined as event-instigated orobject-instigated states of action readiness withcontrol precedence. “States of action readiness”are states of readiness for actions to maintain ormodify one’s relationship with the world or one-self, including loss or decay of motivation to re-late (Frijda, 1986, 2007a). “Readiness” impliesbeing set for action if and when appropriate con-ditions arise, and if relevant actions are availablein one’s action repertoire. Some states of actionreadiness are diffuse and have no aim other thanto relate or not to relate in general; they arecalled “activation states.” Besides apathy, dif-fuse excitement and unfocused receptivity, as insome states of meditation, are examples. Otherstates of action readiness have the aim of achiev-ing, maintaining, or modifying one’s relation-ship to a particular object or event in a particularway—by seeking proximity, by moving away orprotecting oneself, or by moving against andblocking interference. They are called “actiontendencies,” and command actions that can ful-fill their aim.

Action readiness is conspicuous in event-triggered actions that a definition of emotionsin terms of only pleasure and pain would leaveout. Desire is the clearest instance, since manydesires are not guided by foresight of pleasure,or by a wish to escape from pain other thanthat of unfulfilled desire. Surprise and wonderare further instances that I have already men-tioned. Conversely, emotions defined by actionreadiness do include instances of pleasure andpain that produce some change in action readi-ness: hunger that leads to restlessness and theurge to find food, and craving for something ir-resistibly attractive.

Many reactions that are considered “emo-tional” do not include much overt action.States of action readiness may remain just that:states of readiness. However, the actions forwhich action readiness is readiness include cog-nitive actions: changes in beliefs. Emotionshave been defined by Aristotle as well as bySpinoza as inclinations to think one way ratherthan another.

Such cognitive actions share with overt ac-tions one of the defining aspects of passion andaction readiness: They are not initiated by vol-untary planning, nor can they be readily modi-fied at will. By contrast, they possess the fea-ture that I have termed “control precedence”(Frijda, 1986, 2007). They tend to take con-trol and overrule other actions or actioninstigations—the passivity aspect of passion.They not only command actions to fulfill theiraim, but also are set toward achieving that aimin the face of delays, obstacles, and other diffi-culties, turning action into persistent striving.They seek precedence over ongoing behavior orinterference from other sources.

Perhaps the major feature of passions, theiremotionally driven actions, and their beliefchanges is their being reward-insensitive(Elster, 1999b). One tends to do certainthings—in love, in hate, in greed, and inaddiction—even when one knows that onlybad things will come from it. One shrugs one’sshoulders over such things. By contrast, pas-sions are stimulus-governed; one feels irresist-ibly drawn to good-smelling food, or to un-attainable but attractive sexual targets, inproportion to their perceptual salience (Frijda,2007a). The smell of alcohol, the cues for aheroin shot, and the sight of a syringe may fillan addicted individual with frenzy (Berridge,2004). This aspect of impulsive action controlis lawful: Feeling and action instigation are in-


verse hyperbolic functions of temporal distanceto the target in time. Ainslie’s (2001) law oftime discounting describes the “breakdown ofwill” when a person is facing temptations, andthe null effects of mere warnings of future mis-ery. Hume said that reason is emotion’s slave.That may not be entirely true (Solomon, 2004):One can to some extent control one’s emotions;one can even to some extent choose them. Buteven if one is not truly passion’s slave, one cer-tainly is not their master.

Viewing the role of motive states as centralin emotion resembles the view that definesemotions by activation of largely innate actionsystems, each with a particular adaptivefunction (e.g., Bradley, Codispoti, Cuthbert,& Lang, 2001; Buck, 1999; Damasio, 2003;Öhman, Chapter 44, this volume; Plutchik,2003). Both views presumably cover many ofthe same behavioral phenomena, except thatthe action system view is silent on the behavior-al and motivational dynamics, and the actionreadiness view is silent on the provenance ofmodes of action readiness and the action reper-toires.

So far, these are various possible bases fordefining emotions. What specifies and unifiesthe reactions with the phenomena discussedmay not, however, be one or the other of thevarious components, or even a particularcombination of them. “Emotion” can perhapsbest be taken to designate any process inwhich the various components are intimatelyconnected. The concept properly fits states ofsynchronization of several components(Scherer, 2000, 2007). “Synchronization”here does not mean that fixed linkages existbetween particular components, but ratherthat the various components flexibly influ-ence each other. In this flexible fashion, theyrepresent a unitary reaction of the entire sys-tem. Synchronization in this sense is involvedwhenever appraisal and action readiness areevoked, and when control precedence is oper-ative. The three tend to be coupled and to in-volve a number of further components, suchas autonomic arousal and cognitive orienta-tion (Frijda, 2007b).

THE EMOTION CONCEPT

Just as there are arguments to restrict the do-main of emotion (smaller than all impulsivemotor reactions, such as startle; smaller than

all affective responses), there also are argu-ments to enlarge it.

There is a large class of emotional phenom-ena called “emotional attitudes” or “senti-ments” (Arnold, 1960; Shand, 1920). They areusually treated as distinct from emotions. Beingfrightened by a nearby dog and afraid of dogsare not the same sort of thing. The distinctionis that between an occurrent state and a dispo-sition. Emotions have a limited duration; senti-ments may persist over a lifetime. But neverthe-less, occurrent emotions and sentiments are nottotally separate. Emotions and sentiments havethe same structure. They can both be character-ized by an object, its appraisal, and a particularpropensity to act in relation to the object—a la-tent, dispositional propensity in the case of sen-timents, and an acute, occurrent propensity inthe case of emotions proper. Also, sentimentsare not all that dispositional. One can feel thatone fears dogs and that one loves one’s be-loved. One can know these sentiments and actaccordingly—by avoiding places where oneknows a dog to live, or by going upstairs to em-brace the beloved. One can also know that thesentiment may turn into an emotion at theslightest provocation. One may, in otherwords, join sentiments and occurrent emotionstogether in one emotion category, contrary towhat Kenny (1963) proposed. Some authors infact do this.

The preceding section has been mainly con-cerned with the definition of emotions. So ismuch debate in emotion theory. Such debatecan be considered an unprofitable undertaking.Natural processes of any complexity are nottailored into discrete categories. The questionof whether or not someone has an emotionnevertheless is sometimes meaningful. It comesup when wondering whether a given reaction isa “false” or faked emotion. This has a moralor legal side: Has some act been committed“coolly” or “with emotion”? There also existimagined and empathic emotions, as well asanticipated emotions (e.g., anticipated guilt,shame, or regret), which exert powerful socialcontrol (Harré & Parrott, 1996) and whichmay be absent in some individuals like psycho-paths.

However, the multicomponent nature ofemotions entails a looseness of structure thatleads to considering a discrete category labeled“emotion” as only a shorthand. The same maybe true of the distinctions between emotionsand sentiments and between feelings and


moods. Distinguishing categories reflects thegeneral problem of using “substance concepts”(Cassirer, 1908) instead of function concepts,when organizing continuous and continuouslyvarying phenomena. Emotions indeed are oftentreated as thing-like states. Language sustainsthis; emotions are usually referred to by nouns.This may be useful in social communication.For psychological analysis, however, it mightbe better to treat emotions as the observable re-sults of processes that are better denoted byverbs. “She is angering” might not be a bad ex-pression, which neatly matches “She has beenangered,” as well as “She is loving.”

Instead of talking about emotions, one mightinstead describe streams of concurrent and in-teracting ongoing processes: appraisals thatlast and change, that activate processes of ac-tion readiness that generate action prepara-tions and overt actions, which in turn act backupon appraisals, and that all vary in degree ofactivation, each having different time coursesand different moments at which they die down.This is the picture sketched by multicom-ponential emotion theory (e.g., Scherer, 2000),according to which matching categories to on-going processes—“an emotion,” “a mood,”and even “anger!” or “excitement!”—can onlybe sloppy, often a bit arbitrary, and not veryconsequential except for verbal communica-tion.

The usefulness of category labels for verbalinteraction does not imply, incidentally, thatthey are needed for nonverbal communication.States of action readiness, degrees of activa-tion, and degrees of control precedence of feel-ings and action urges can be grasped directly byan observer, and can probably be copied di-rectly in motor empathy (Gallese, 2005). Inpsychological analysis, too, the various phe-nomena can be described directly in terms ofthe ongoing processes. All these processes aregraded in strength, and making cutoffs at cer-tain levels of strength is arbitrary; that, ofcourse, was the point of Duffy’s (1941) attackon the emotion concept. Her analysis was lim-ited by its focus on activation. The same ap-plies to other emotional processes, such asaction readiness, appraisals, and control prece-dence. Degree of articulation of an intentionalobject, for instance, can be considered a pa-rameter. Its introduction would turn the dis-tinction between emotions and moods into acontinuum of “emotionness” versus “mood-ness.”

EMOTION INSTANCES

So far, I have referred to emotions as modes ofacting and reacting, or as modes of feeling anddoing. This is decidedly vaguer than callingthem “responses” or “reactions.” There is,however, good reason to do so. What we usewords like “emotion” for are usually complexresponse patterns spread out over time. Theircomponents show variations in duration andtime course. Brief facial expressions occuralongside longer-lasting changes in heart rateor respiration, interspersed with flashes ofstrong feeling against a background of feelingthat waxes and wanes.

These observable and experienced phenom-ena can in fact be described at very differentlevels of analysis and integration. One can de-scribe single response components, such as asmile or an angry facial expression. One candescribe more or less coherent expressive andbehavioral sequences, such as a series of fearfulmovements when a person is facing an ongoingfrightening event, or of various angry expres-sions (a scowl, a glare, a deep frown, a clench-ing of fists) during a hostile interchange. Onecan observe a sequence of behavior and feelingmodes that all center around dealing with aparticular event, such as viewing somethingwith apprehension that grows into alarm andturns into anger, then calms down to merewatchfulness. Such sequences form “emotionepisodes.”

Emotion episodes are in fact what subjectsusually recount when asked to recall some in-stance of an emotion (Frijda, Mesquita,Sonnemans, & Van Goozen, 1991). Episodesusually include several emotions (Oatley &Duncan, 1994). All this is obviously relevantfor the question of what represents an emotioninstance, and how long emotions last. The du-ration depends on what one focuses on. It maybe 5 seconds at most, for an individual facialresponse. It may be an hour for a hostile orfearful interchange. It is up to days or longerfor emotion episodes that continue over restlessdream-ridden sleep. It is up to a lifetime whenthe notion of emotions is extended to senti-ments and their latent readiness for emotions inthe acute, occurrent sense (Frijda, 2007a).

How long one considers an emotion to last isto some extent, as emotions of any personalimportance tend to entail extensions. They mayinstigate the individual to ruminate, to walkaround pondering revenge or ways of restoring


self-esteem, or to engage in social sharing ofone’s emotion (Rimé, 2005). It is not immedi-ately transparent when an emotion terminates,since all components may do so at differentmoments; lowered thresholds (e.g., irritability,rapid tears) may outlast all more manifest com-ponents. It is also a matter of taste which of thephenomena discussed are considered as belong-ing to the emotion itself, and which to its ante-cedents or consequents (e.g., Dumas, 1948).Conscious appraisal is as much a part of theemotion as an antecedent for action impulse.These questions clearly lose much of their sensewhen “emotion” is understood to refer to acollection of interacting processes, and not to asingle or a solidly integrated entity.

Emotion instances thus can be viewed atvery different levels. That is why “response” or“reaction” is not the best designation for anemotion instance. Most instances are complex.Actions and feelings in a particular emotionform part of interactions and interchanges. Anemotion thus can thus be viewed as an intra-personal state of feeling, arousal, activation,and so forth, but also as the individual’s sharein an interactive occurrence that involves deal-ing with another person and one’s mutual rela-tionship. Which level of analysis is selected canhave appreciable consequences. A focus onfeelings, for instance, readily loses sight of theintentional nature of emotion—its being, andbeing experienced as, an event between the in-dividual and the other (Lambie & Marcel,2002). A focus on motor responses may, in ex-perience and in analysis, turn an action into amere movement.

Analyses at different levels are compatible;they coexist. However, incompatibilities mayarise when emotions are categorized on differ-ent levels. One cannot, as a rule, reduce de-scriptions at some higher, more integrative levelto a combination of more elementary constitu-ents without loss. Jealousy, as pain caused byrivalry, is not just pain, nor is it just anger, noris it just despair; it is an interpersonal pain thatcan change its face at any moment. Likewise,higher-level categorizations cannot always bebuilt up from the lower-level phenomena.Higher-level categorizations often (or perhapsusually) include more phenomena, as well asmore interactions between the lower-level phe-nomena, or feedback from them. Indignation ismore than anger with a moral overtone, since itis felt as carried by one’s moral values ratherthan by one’s personal interests.

Noticing that emotions are streams of inde-pendent but interacting processes raises ques-tions about the nature of these interactions.Process analyses suggest that processes thatlogically follow from certain other processesmay still, by their feedback, influence those lat-ter. Facial expressions that respond to apprais-als in turn influence appraisals. The expectedeffect of one’s emotion on others influences theactual occurrence of that emotion. Processesmay produce a higher-order outcome (say, acategorization of one’s state) that then controlsand modifies logically prior processes (say,appraisal). Such processes have been called“processes of vertical causality.” Emotion pro-cesses thus are not linearly organized. A non-linear, dynamic model is called for (Lewis,1996, 2005); such a model accounts better forhow emotions actually proceed and how theyobtain and modify their shape.

The interactions suggest that givensubprocesses do not follow each other in a veryregular manner, even if the links are lawful.The actual linkages are sensitive to a host offeedback influences, and to influences fromvarious levels. Whether certain stimuli actuallyelicit certain responses, or certain appraisalselicit certain emotions, may strongly depend onsecondary conditions such as personality,mood, the state of the organism, and coinci-dences in the physical and social situation.These secondary conditions may be so impor-tant that a model involving chaotic determina-tion may be more satisfactory than the usuallinear model. It also is what makes Elster’s(1999a) mental mechanisms better explanatorytools than “laws” in a strict sense.

HOW ARE WE TO DISTINGUISHDIFFERENT EMOTIONS?

How are we to distinguish different emotionsproperly or profitably? Verbal categories havebeen used as the starting point, since some ofthem suggest elementary entities. Some emo-tion labels have indeed been interpreted asreflecting elementary, irreducible feelings orqualia, with other labels representing theirblends or subspecies (Izard, 1977; Oatley,1992). The feelings may reflect (or be part of)basic organized component structures, includ-ing motivational states (e.g., Oatley, 1992), ac-tion and expression patterns (e.g., Ekman,1992; Izard, 1977; Tomkins, 1962), and pat-


terns of physiological reactions (e.g., James,1884). The feelings may not be irreduciblequalia, but themselves patterns of values ontwo feeling dimensions of hedonic tone andactivation—variants of “core affect” (Barrett,2005; Russell, 2003).

There is no a priori reason why verbal labelsshould offer a good clue to distinct emotions.Languages differ in their taxonomies, and ver-bal labels may reflect eliciting contingenciesrather than modes of feeling and doing (e.g.,likelihood of certain social sanctions or partic-ular rewards).

Several hypotheses on how the various com-ponents are related and organized have ori-ented research. In one major hypothesis, thevarious components form solidly coherentpackets, each based on a common neural andneurohumoral disposition. Activation of thedispositions by events jointly activates the vari-ous components. This basic-emotions hypo-thesis exists in several variants (Buck, 1999;Ekman, 1992; Izard, 1977; Öhman, Chapter44, this volume; Panksepp, Chapter 4, this vol-ume; Tomkins, 1962). Such a hypothesis findssupport in indications for the existence ofdedicated brain circuits and neurohumors(Panksepp, Chapter 4, this volume), apart fromthe unstable evidence for label-specific facialand other expressions (Ekman, 1992; Izard,1977); however, correlations in the occurrenceof various components tend to be moderate orweak (Scherer, 2005).

A second hypothesis starts from this lastfinding. The “multicomponential view”(Scherer, 2000) views emotions as more or lessunordered collections of components, jointlyactivated by how an emotional event is ap-praised and by individual component propensi-ties. Emotions can occupy almost any positionin a multidimensional space, with as many di-mensions as there exist components. Each com-ponent is activated by separate external condi-tions or aspects of the event–subject interactionas appraised (Ortony & Turner, 1990; Scherer,2001).

This second hypothesis leads to abandoningthe very notion of distinct emotion types. Eachemotion instance stands on is own. Emotionsare just bundles of component processes. Emo-tion labels just loosely and fuzzily cover subre-gions of the multicomponential space, or de-limit such subregions themselves by culturally,linguistically, or ecologically determined proto-types or scripts linked to them (Russell, 1991).

Several theorists have taken this second option.Emotion labels are viewed as arbitrary distinc-tions in a more or less unstructured domain(Mandler, 1984), reflecting ecologically fre-quent or socially important patterns of compo-nents (as in Scherer’s [1992] conception of“modal emotions”), or reflecting cultural con-cerns and values that give emphasis to some,and not to other, sets of feelings and modes ofbehavior (as in social-constructivist theory; see,e.g., Harré & Parrott, 1996).

The multicomponential view appears betterable to deal with cultural differences in emo-tion taxonomies (e.g., Lutz, 1988), as well aswith differences in the precise semantic contentof similar categories in different languages(such as “anger” in English and ikari in Japa-nese, or “shame” in English and hassam inArabic). On the other hand, a basic-emotionsview more readily handles evidence suggestingthat certain emotion categories are very com-mon (Mesquita, Frijda, & Scherer, 1997) andmay even occur in most or all languages(Hupka, Lenton, & Hutchinson, 1999; Shaver,Wu, & Schwartz, 1992; but see Russell, 1991).

The two hypotheses may not be as dissimilaras they seem, though. The basic-emotions viewhas ample room for cultural differences. Basicemotions can be considered to represent func-tionally defined classes (Ekman, 1992). Withineach class, the precise antecedents, nature ofthe objects, full gamut of appraisal compo-nents, precise type of action goal or action todeal with the appraised contingency, and signif-icance of the emotion (see below) all may vary.They all provide the leeway that culturaland individual differences require. More perti-nent still is the fact that any component—facial expression, autonomic reaction, actiontendency—is the joint outcome of the emotiondisposition as such, and of facilitation by situa-tional and other momentary factors. Respira-tion rate is influenced by excitement as well asby moving fast. Smiling flows from pleasureand from desire to please. The moderate corre-lations between components can be readily ac-commodated within the basic-emotions view.

Conversely, even within the multicompo-nential hypothesis, the componential space isnot at all fully unstructured. Many componentsmutually constrain or entrain each other. Thereexist coordinate structures. Vigorous action re-quires sympathetic energy arousal, and readi-ness for it prepares it; feeling consists in part offeedback from autonomic and skeletal re-


sponses and from states of action readiness;motor relaxation and high autonomic arousalare incompatible; and so forth.

In a third, hierarchical approach, compo-nents differ in their organizational power.Some are more central than others. Pleasant-ness and unpleasantness may each cascade intocollections of separate subsystems and compo-nent streams (Bradley et al., 2001). Action ten-dency may well organize all other componentsfor implementing its aim. This third approachwould fit the functional interpretation of dif-ferent emotions as different, specifically fo-cused adaptive provisions (Damasio, 2003;Frijda, 1986, 2007; Plutchik, 2003).

The multicomponential and hierarchical ap-proaches follow the shift from categorical toprocess conceptualizations. This shift extendsfrom the phenomena of emotions as such to theemotion components themselves. Many compo-nents are best broken down into smaller compo-nents and their variable conjunctions. This hasappeared necessary for “autonomic arousal”;no valid, consistent index has been found, be-cause indices do not strongly covary (Levenson,2003; Stemmler, 2003). The same has been ar-gued for facial expressions. Most patterns, eventhose presumably typical for particular emo-tions, are best understood as built up from indi-vidual facial action units that each can be consid-ered functional units, such as serving protective,attentional, or force-assembling functions (e.g.,eye narrowing, eye widening, or vertical frown-ing, respectively, in these three examples)(Camras, 2000; Ortony & Turner, 1990; Smith& Scott, 1997; Scherer, 1992). In any case, shift-ing from category to process approaches turnsthe relationships between components into asubject for unprejudiced research on several ba-sic questions. Which processes are linked towhich other processes, and to what degree?Which linkages are due to joint response to thesame antecedent contingencies, and which totheir forming coordinative motor structures?Which linkages represent functional depend-ence? Which represent the effects of a sharedcommand system (Lewis, 2005)?

THE RELATIONS BETWEENEMOTION AND MOTIVATION

The relations between motivation and emotionconstitute another of the perennial problems.Many emotions form motivational states, but

many motivational states (e.g., need for food)are not emotions. This has led some investi-gators (e.g., Bindra, 1959) to abandon theemotion–motivation distinction. This does notappear justified, however, because there is areal distinction that counts.

The term “motivation” suffers from a poly-semy that is similar to that for “emotion.” It,too, has an occurrent reading (referring toemotions proper) as well as a dispositional one(referring to sentiments). When saying thatone’s social motivation causes one to be un-happy when alone, a dispositional reading ismeant. Saying that felt loneliness motivates oneto seek company implies an occurrent reading;loneliness is an emotion. “Motivation” thuscan label a cause, a consequence, or an aspectof an emotion. It is a cause in its dispositionalsense, and a consequence and/or aspect in theoccurrent one.

This type of distinction can generally be con-sidered fruitful. McDougall (1923) distin-guished dispositional “instincts” from emo-tions as their actualizations. Buck (1999) callsemotions the “readouts” of motivation. Oatley(1992) distinguishes “goals” from emotions,with the latter identified as the responses tocontingencies in achieving or not achievinggoals. Goals, in this context, function as dispo-sitions that drive the occurrent emotional moti-vations to reach them.

Such a distinction is important. Disposi-tional motives (I call them “concerns”; “inter-ests” is a useful alternative) belong to the majorexplanatory notions for emotions (with “ac-tion tendencies” as the latter’s occurrent moti-vational component). Concerns are what ren-der events and objects emotionally relevant inthe first place. The psychology of motivationforms the cornerstone of emotion analysis.Why do we seek and enjoy company, and whydoes loneliness makes us sad? Because beingwith and interacting with others are our con-cerns. Concerns of course also form the corner-stone of dynamic exploration of emotionalconflict: Conflict flows from incompatible con-cerns, such as desiring wealth and estimatingpersonal decency.

The two senses of “motivation” are not al-ways easy to tease apart, though. Is “love” adispositional or an occurrent motivation, aconcern or an emotion? And what about beingin love? Both obviously can be both—the for-mer more continuously, the latter more inter-mittently.


EMOTION ELICITATION

What elicits emotions? The simplest type of an-swer is that different emotions are evoked bydifferent kinds of events—each emotion by adifferent one. In behaviorist parlance, emotionsare different reflex-like response patterns todifferent unconditioned stimuli. The range ofelicitors then is augmented by the usual con-ditioning constellations. This was Watson’s(1929) answer. It is still dominant in emotionexplanations from evolutionary psychologyand evolutionary orientation generally (e.g.,Bradley et al., 2001; Tooby & Cosmides,Chapter 8, this volume; Öhman, Chapter 44,this volume).

Later behaviorism and more recent analyseshave refined the type of explanation a bit. Stim-ulus events are pleasant or unpleasant, or rep-resent rewards or punishments. They may havecome to do so by evolutionary selection or byactual need or concern satisfaction or frustra-tion. But different emotions are aroused whenrewards are embedded in constellations involv-ing the temporal context: their advent, de-crease, or omission (Bouton, 2005; Gray, 1987;Mowrer, 1960; Rolls, 2005). This refinementhas been extended to include further contextaspects, such as previous reinforcement ornonreinforcement of particular responses todeal with the event (Gray, 1987). In appraisalapproaches, the context variables are formu-lated in cognitive rather than contingencyterms. They are designated as “appraisal com-ponents” (e.g., Ellsworth & Scherer, 2003).“Appraisal of coping resources,” as a compo-nent of emotion antecedents, thus is largelyequivalent to previous reinforcement ornonreinforcement of coping responses. Whatdiffers between behaviorist and cognitive ap-proaches is the process by which one or theother emotion is supposedly evoked: It is linkedmore strongly to environmental events in be-haviorist theory, but more strongly to subject-bound processing variables in appraisal theory.

A major, somewhat different conceptualiza-tion starts from the previously mentioned cen-tral place of dispositional motivation in be-havior, rather than that of stimulus–responseconnections. This conceptualization posits pro-cesses of appraisal of the goal or concern rele-vance of events—or, for short, of events’ rele-vance to well-being (Lazarus, 1991; Oatley,1992; Stein & Trabasso, 1992). Those eventselicit emotions that are appraised as beneficial

or harmful for achieving the individual’s well-being. This conceptualization seeks to accountfor positive and negative reward values, orpleasantness and unpleasantness of events,from a general perspective. It implies a differ-ent account of the basic mode of operation ofhuman and animal systems. In the end, this ap-proach and the behavioral approach may turnout not to be as different as they seem, since thedefinitions of “unconditioned stimuli” and“states of satisfaction of concerns” may not al-ways be so drastically different, as evidencedby Hull’s (1953) notion of “drive reduction.”On the other hand, their emphasis differs. Sen-sitivities for the various concerns are, by theirnature, more general than can be formulated assensitivities for sets of discrete stimuli. “Novel”stimuli (that trigger curiosity) present a case inpoint: Novelty is not a stimulus attribute. Thesame is true of interpersonal warmth and ofrelevance to self-esteem.

The emotional efficacy of eliciting eventscannot always be meaningfully reduced to “un-conditioned” stimuli contained in or signaledby them. Reduction is often problematic. Sev-eral elementary event types may converge ona given higher-order sensitivity. Personal losselicits sadness because it implies loss of per-sonal warmth, of protection, or of soothingstimuli. It may elicit sadness by any mixture ofthose. The stimuli may have nothing more incommon than that they all engage an interper-sonal intimacy-achieving system.

Moreover, events are not always emotion-ally efficacious because of the stimuli theycontain or signal. Some are efficacious by en-abling or obstructing exerting particular skillsand competences. Novel stimuli may be plea-surable (when they are) because and whenthey enable cognitive assimilation skills(Frijda, 2007).

Emotion elicitors can be more profitably un-derstood at some intermediary level betweenspecific stimuli such as painful stimuli, muti-lated bodies, snakes, or spiders, and the generallevel of negative and positive rewards. “Con-cern relevance” points at emergent propertiesthat reside not in the positive or negative stim-uli as such, but in a more involved interaction.Exploring why certain stimuli or actions are re-warding or aversive—receiving a gift or givingone (Mauss, 1914/1957), being sensitive to amagic curse, shrinking back from seeing muti-lated bodies—forms a challenging task foremotion psychology.


THE ORIGINS OF EMOTIONS

Nature or nurture? Here is another of the pe-rennial problems in emotion. Specifically, howmuch, and what, in the phenomena that sug-gest “emotion”—and in the precise sets of phe-nomena that are labeled as “joy” or “fear”—are due to provisions and constraints laid downin the mechanisms with which humans are bynature equipped? How much, and what, arethe results of individual learning, social shap-ing, and social pressure?

Nobody, I think, contests that emotions havebiological bases. There is compelling evidenceof neural and neurochemical mechanisms (seeLeDoux & Phelps, Chapter 10, this volume;Panksepp, Chapter 4, this volume). Basic phe-nomena and underlying processes—unlearnedaction systems, appraisal processes, pleasureand pain, control shift provisions—extend be-yond the human species over many other ani-mal species. Some may extend over inverte-brates.

Indications for several different emotions areshared by most or all birds and mammals. Theindications include evidence for dedicated neu-ral and neurohumoral circuits (Buck, 1999;Panksepp, Chapter 4, this volume), and for ac-tion systems with particular functions: for ap-proach, for establishing close nonhostile inter-action, for hostility, for escape and avoidance,for active rejection, for behavioral inhibition.As I have mentioned, verbal emotion labelsthat imply references to these occur in almostall languages.

Cross-species and intrahuman generalitiesoffer strong support for arguments that someor most of the preceding phenomena and pro-cesses have an evolutionary basis (e.g., Ekman,1994). These arguments are strengthened bythe adaptive utility of the behavior systemslinked to the emotions: protecting oneself fromintruders, predators, and rivals; warding theseoff and threatening them; obtaining food anddrink; detecting and contacting mates; seekingshelter; protecting the young; submitting to themore powerful. For all of them, utility in pro-moting reproductive fitness is obvious or canbe readily constructed (Buss, 1994; Tooby &Cosmides, Chapter 8, this volume).

Nevertheless, reservations are in order. Sofar, it is rather unclear what precisely the vari-ous neural circuits do in functional terms. Doextended amygdala circuits link affectiveappraisal to particular stimuli or contingencies,

or to sets of prewired motor programs, or tomotivational reactions activating motor pro-grams that have been shaped in other than spe-cific emotional contexts? The last of these isnot unlikely.

By itself, universality in fact does not provebiological origin. Major emotions correspondto sensitivities to universal contingencies orcore relational themes, such as threat, loss,competition, or success (Lazarus, 1991). Uni-versal contingencies themselves present univer-sal occasions for learning, and for universallysimilar modes of problem solving and dynamiccompilation of action patterns. Take anger, forexample. Its behavior repertoire could be ex-plained by the facts that harm is universallypainful, and that one is equipped with severalsorts of actions that can produce externalchanges in general and social changes in partic-ular. Kicking, scratching, throwing objects, andshouting are useful within emotional contexts(e.g., when chasing intruders), as well as out-side such contexts (e.g., when cracking nuts,breaking branches, and throwing over one’scompanions in play). Instrumental behaviorshave an applicability range. The efficacy ofsuch actions can be discovered in the dynamiccontexts of hostile, playful, and instrumentalinteractions, in the same way that a baby dis-covers the possibility of walking when bodyweight and muscle strength have reached theright relationship (Thelen, 1985). In otherwords, universality can be explained in morethan one way. Biological roots are still in-volved, but may well exist at deeper or more el-ementary levels than being roots of emotionalresponse patterns as such.

And then, even if biological roots are in-volved, it may not be easy to assess what theyare. Mother–child and child–mother attach-ment no doubt have such roots, but which arethey, and what is their role? Western attach-ment theory and findings, on close scrutiny,may reflect more cultural influences than primafacie likelihood suggests. Independence maynot be a general component of the attachmentsystem, but one more characteristic of theWestern world (Rothbaum, Weisz, Pott,Miyake, & Morelli, 2000).

Biological dispositions, moreover, need innerand outer environments to take behavioral andexperiential shape. It may be useful to stressthat the role of cultural differences in emo-tional phenomena depends to an important de-gree upon one’s level of analysis: The deeper


the level, the larger the generality (Mesquita etal., 1997). What is usually translated into Eng-lish as “shame” differs strongly in social roleand behavioral implications between Westernand Arab cultures; yet both flow from a similarsensitivity to being socially accepted and in-volve a similar motivation to correct one’sdeviations from acceptance. But again, theystrongly diverge in what represents such devia-tions and how to correct for them. Culturalprescriptions and models provide such shaping,and it may even shape the motivation to suchan extent that the emotions are not recogniz-ably similar (Mesquita, 2003). Both symboliccapacities and social interactions penetrate ev-ery phenomenon, its occasions for appearing,and its duration. It largely remains to explorehow far these penetrations can go, and whatthe implications of cultural differences are.

THE FUNCTIONS OF EMOTIONS

The negative aspects of emotions dominatedearlier theorizing in both philosophy and psy-chology. For Aquinas, most emotions belongedto the capital sins. For Kant, emotions repre-sented illnesses of the mind (“Krankheiten desGemüts”). Several early 20th-century authorsconsidered emotions to be “states of functionaldecortication” or of neural disorganization(e.g., Claparède, 1928; Hebb, 1949).

In contemporary theorizing, the tides haveturned, mainly under the influence of evolu-tionary interpretations. Emotions are largelyviewed as adaptively useful, or at least as hav-ing been so in the evolutionary past. The func-tional perspective now dominates.

Considering the evidence for phylogeneticorigin and continuity, this functional perspec-tive on emotions is plausible. It is also plausiblebecause of the wide range of likely functions ofemotions—wider than dealing very directlywith individual or species survival or reproduc-tive fitness. Joy can be understood as motivat-ing readiness for novel exploits and expandingcompetences (Fredrickson & Cohn, Chapter48, this volume). Anger, shame, guilt-feeling,and sympathy are powerful regulators of socialinteraction (Hoffman, Chapter 27, this vol-ume; Fischer & Manstead, Chapter 28, thisvolume). Sadness may serve disengage-ment from attachments after personal loss.Irrational-seeming emotions like compassion

and desire for revenge serve adaptive purposes:They represent one’s commitment, and signalothers that one may act upon them; theythereby outweigh occasional costs in short-term interaction (Frank, 1988).

Evolutionary origins may mean that emo-tions are now mere obsolete remnants, sincethe original adaptive problems have waned(e.g., nonhuman predators) or rationally de-vised methods for dealing with such problemsare now available (e.g., precision bombing).Energy mobilization by sympathetic arousalmay be wasteful, now that rationally guidedmental actions can be instigated. However, byand large, emotions and emotional actions arestill generally adaptive in about the originalsense. Love still drives sexual reproduction; in-timacy is still profitable for child care andsocial support; fear still motivates prudence;anger still promotes our interests and may dis-courage rivals. Expressive behaviors, it canlikewise be argued, still serve to strengthen re-lations (e.g., smiling), to discourage offenses(e.g., bracing, shouting, readying for a fight),and to protect one’s body (e.g., ducking andclosing one’s eyes). Considerable amounts ofwork have gone into generating plausible func-tional hypotheses and assembling evidence(Frijda & Tcherkassof, 1997; Scherer, 1992;Keltner & Busswell, 1997).

Many emotions and emotional behaviorsare, moreover, functional in a somewhat differ-ent sense. They may not be adaptive in promot-ing survival, and this may not have been whatmade them come into existence—but they dohave effects on oneself and on others that keepthem going. Joy is an obvious example; it maynot help, but it is nice. Similarly, positive ex-citement and curiosity keep boredom at bayand keep one’s competences active. Grief maynot have any direct advantage, but its anticipa-tion keeps people together. It helps to knowthat one will be missed when the time hascome.

One needs to be cautious with functional in-terpretations, though. They are easily found,and the evolutionary perspective almost forcessearching for them. Anger? Small wonder if it isinnate, since it helps in protecting one’s terri-tory and offspring. Grief? It may well serve todetach oneself from lost attachments. Apathyin grief? Small wonder again; it saves expend-ing energy that would be useless after the loss.Heart rate increase in fear? It was obviously


useful when the emotion developed underthreat of rivals and predators, and when onehad to be always ready to climb a tree; al-though its net profit even then can be doubted(Arnold, 1960). Aversion to pictures of bloodand mutilation? It obviously serves prudence,and thus survival (Bradley et al., 2001). Butdoes it? It might hinder the process of keepingone’s enemies off.

In fact, in the ancient adaptational environ-ment, no observer was around to gauge thevarious benefits against the costs that anger,apathy, the wear caused by heart rate increase,and fear of the sight of blood might have en-tailed. Evolutionary emotion hypotheses rarelyexamine such implications or possible alterna-tive, more dynamic explanations. Emotionalphenomena may indeed have originated asmere “spandrels”—chance offshoots of quitedifferent potentials (Gould & Lewontin,1979). Anger, as I have indicated, might largelyhave emerged as a by-product of provisions forpower deployment that developed for crackingnuts; moral disgust may stem from a pre-adaptation for ejecting foul substances (Rozin,1999).

Functional advantages of emotions canreadily be thought of; that does not make theseadvantages actually occur, nor do they actuallyexplain the emotions’ origins. Grief may nothelp anything at all. Grief of loss may be simi-lar to pain in a phantom limb: Pain generally isuseful, but not all pain is. Depression may bean offshoot when all objects worth striving forhave dropped away, and exhaustion has set inbecause of fruitless efforts to regain them;Weiss, Glazer, and Poherecky (1976) opposedthis hypothesis to Seligman’s learned helpless-ness hypothesis. Emotional shock may just rep-resent disorganization due to sudden impactthat cannot be managed; that is how Dumas(1984) and Hebb (1949) interpreted it. There isno doubt that emotions, by and large, fulfillprofitable functions, but one should not fallprey to the Panglossian fallacy (Gould &Lewontin, 1979). Pangloss, as readers may re-call, was the philosopher in Voltaire’s Candidewho, upon every misfortune, echoed Leibniz’sdictum that we live in the best of all possibleworlds. By now we know that we do not.Many emotional events are simply beyond hu-man and animal coping resources. Such re-sources are of necessity limited in an 80-kilogram organism that has to be operational

after only a 9-month gestation. Disturbance ofoptimal functioning can just be dysfunctional.

And, as noted, emotions that in principlerepresent functional provisions are often disad-vantageous or outright harmful. Emotions cancause suboptimal action. In panic, people pressthrough narrow exits; stage fright spoils per-formance; nervousness spoils precision ofmovement; rage may lead to childish behaviorand destroy social harmony (De Waal, 1996).Parrott (2001) has convincingly shown thatwhether or not emotions are functional de-pends critically on the adequacy of the apprais-als that led to these emotions, on the choiceand control of behavior that is motivated bythe emotions, and on adequately evaluating theimpact on others of both the behaviors andone’s feelings.

In short, a functional perspective on emo-tions should not lose sight of the limits withinwhich this perspective holds.

EMOTION AND COGNITION

Traditionally, emotion and cognition have beenconsidered different faculties. They have beenput in opposition; so have feeling and thinking.Pondering and raging would seem to constituteopposite ends of a continuum.

Contemporary psychology tends to considerthese oppositions problematic. The contraststend to dissolve upon analysis of how inti-mately information processing and acting, re-acting emotionally and appraising emotionalmeanings, and goal setting and pursuing impul-sive aims and desires are intertwined. Opposi-tions dissolve at lower levels of analysis, whenscrutiny of processes replaces categorical dis-tinctions.

The emotion–cognition distinction has fre-quently been confounded with that betweenconscious and nonconscious informationprocessing. The distinctions are, however, or-thogonal (Clore, Storbeck, Robinson, &Centerbar, 2005); I return to the latter one be-low. But the problems in giving an account ofinformation processing in emotion are consid-erable, because most of this processing cannotbe analyzed in the way that the processingof propositional, discursive information can.Cognitive psychology has invented representa-tional tools such as schemas, cognitive net-works, and symbolic representations, and


ultimately abstract representations in “mental-ese.” These tools do not appear appropriate forinformation that operates in emotion pro-cesses. The term “intuition” is often used (Ar-nold, 1960). “Appraisal” is contrasted with“knowledge” (Lazarus, 1991)—a distinctionneeded because of the central problem that somuch information that is in principle relevantto emotional arousal often fails to arouse, andknowledge that shows events to be neutral stilldoes not prevent emotional arousal. Tools forunderstanding the processes involved in ren-dering information emotionally efficacioushave begun emerging only slowly and recently.One consists of recourse to concrete and mode-specific representations and their networks(Barsalou, 1999). Another is recognizing theimportant role of motor representations, bothas underlying identification of emotionalmeanings in the behavior of other individuals,and as underlying action programs, action rep-resentations, and action foresight (e.g., Gallese,2005; Meltzoff, 2002). This is one of the areasthat gave rise to extensive recognition of whathas been termed “embodied cognition”(Niedenthal, Barsalou, Ric, & Krauth-Gruber,2005). Still another tool, not used yet in under-standing emotions, is representation of pro-cesses over time, as is being developed in thedescription and analysis of human movement.

Besides the problems of representing emo-tional processes and information are those thatconcern the structures of the involved informa-tion. These structures range from simplestimulus–response transductions (which implyrepresentations of stimulus “keys” that fit rep-resentations of particular response “locks”), toelaborate multimodal networks of associa-tions, and on to abstract meanings and infer-ences. Massive research over the last two orthree decades has demonstrated how very ele-mentary information can on occasion be emo-tionally effective (e.g., Zajonc, 1980). Researchhas also shown how complex emotional infor-mation often is (Barsalou, 1999), even when itdoes not immediately appear to be so because itoperates nonconsciously (Clore et al., 2005).

Close to the contrast between emotion andthinking has been that between emotion andreason. “Reason” has at least two differentmeanings: that of using complex thought pro-cesses such as logical inference, and that of us-ing means–end relationships to reach optimalproblem solutions. The first meaning contrasts

reason with the intuitive and impulsive pro-cesses; the second contrasts it with the oftenharmful and disorganizing aspects of emotions,and their command of primitive responses.Both contrasts have been attenuated in moderntheory. The first is weakened by recognition ofthe complex appraisals that underlie most emo-tions. These appraisals derive from standardsof comparison, cultural values, and thwartingor fulfillment of expectancies and concerns. Allare involved in elicitation of social emotionssuch as shame and regret. The second contrastis weakened by recognition of the “rationalityof emotions” (de Sousa, 1987): recognition ofemotions as in principle reasonable and func-tionally appropriate responses to events as ap-praised. Emotions also appear as conditions forrational choice (Damasio, 1994; de Sousa,1987; Frank, 1988; Solomon, 1993, 2004), be-cause they respond to a wider array of relevantvariables.

And nevertheless, both contrasts betweenemotion and rationality remain. The emotionsaroused by simple stimuli may override furthercognitions, notably those relevant to conse-quences that are more remote in time. Emo-tionally effective beliefs may disagree with si-multaneously held, more solidly foundedbeliefs. People with spider phobias often recog-nize that spiders are harmless, and still fearthem. Dysfunctional emotions have been dis-cussed previously. As de Sousa (1987) has ar-gued, emotions possess only “local rational-ity.” But the contrast is perhaps not reallybetween emotions and rationality, but betweenwhat does elicit emotions and what could elicitemotions if and when “rational” informationwere to acquire emotional appeal.

EMOTION EXPERIENCE

One major problem has hovered over the fieldof emotion study, and it still does: that of con-scious feelings. What are they, and what role dothey play with regard to emotional behaviorand the conduct of life?

It was not too long ago that “emotions”were defined as particular states of conscious-ness that causally determine bodily and behav-ioral responses. That time has changed. In thecurrent literature, one frequently comes acrossthe view that feelings are mere epiphenomenaof neural and bodily processes. Neural and


bodily processes come first, and can proceedentirely or largely without the intervention offeelings. This view has solid support in the con-viction that all conscious phenomena find theirroots in neural processes, reflect neural pro-cesses, and depend on neural processes. Also,there is the fairly general conviction that physi-cal causal sequences are closed. Neural pro-cesses may give rise to conscious experience,but it is hard to conceive that conscious experi-ence might influence neural processes in turn.

Yet this view contradicts everyday ex-perience. It does so outside the domainof emotions. Although there is appreciable“blindsight”—the name given to perceptionwithout conscious awareness—the blind-sighted person is severely handicapped(Weiskrantz, 1997). It also contradicts experi-ence within the emotion domain. One refusesto enter some dark alleyway because it looksthreatening, breaks off a friendship with some-one who has made one very angry, and goes togreat lengths to please someone who provideswarmth and joy. One decides to go to the mov-ies because that spells fun.

Much of such everyday experience may wellbe illusionary. One often acts for other reasonsthan the feelings one is aware of or constructs.Psychoanalytic exploration found evidencethat this is so; contemporary experiments haveconfirmed it. So, everyday experience notwith-standing, feelings might still be useless. Thegeneral theoretical argument for this conclu-sion has been mentioned: How could feelingspossibly influence neurons? One can state thisas a problem, rather than as an argument. Ithas been called “the” hard problem of con-sciousness (Chalmers, 1996): how to conceiveof the relationship between body and con-sciousness. It is the problem that comes clearestto the fore in emotions, where feelings, but alsocognitions, appear to influence the body andget information back from the latter. It is theproblem that Descartes was the first to trulyrecognize, in his Passions de l’Âme. It is still en-tirely unsolved, recourse to Spinoza (Damasio,2003) notwithstanding.

There is good reason, other than everydayexperience, not to be satisfied with consideringfeelings epiphenomenal and irrelevant forbehavior. That view makes feeling into an evo-lutionary oddity. It came about—when? withthe advent of humans? with the advent of pri-mates? with the advent of vertebrates?—with-

out any adaptive advantage. It is unlikely. Thismeans that research has to continue on the pos-sible effects of conscious feelings, and on howsuch effects can occur if they exist.

What created the doubts about the functionof feelings has been twofold. First, there hasbeen the force of James’s argument that feelingsare the “readouts” from bodily emotional reac-tions, which therefore must have been presentbefore the feelings. Second, there is the massiveevidence that affective reactions can occurwithout the subject’s being aware of their ob-jects, of the emotionally relevant attributes, oreven of the reactions’ having taken place(Bargh, 1997; Berridge, 2004; Moors & DeHouwer, 2006; Zajonc, 1980).

Yet the range of emotional reactions shownto occur nonconsciously is as yet limited. It in-cludes nonconsciously induced liking for neu-tral stimuli, enhanced preferences for liked sub-stances or stimulus objects, nonconsciousimitation of behaviors, and delayed extinctionof conditioned physiological arousal. To whatextent affective influences may unwittingly de-termine strivings, goal-directed behavior, andthe conduct of life remains to be demonstrated.

Examining the role of conscious feelings inemotions is rendered difficult because of thevarious ways in which feelings can occur. Con-scious experience occurs in various modes: re-portable or not reportable, diffuse and globalor articulate and amenable to verbal descrip-tion. The several modes have given rise tosharp disagreements about the nature of emo-tional feelings: irreducible qualia, sensory bodysensations, central body state and position rep-resentations, action representations, or percep-tions of external events with emotional mean-ings (Frijda, 2005, 2007a; Lambie & Marcel,2003). It can be argued that all can occur, de-pending on circumstances and attentional atti-tude; if so, this has perhaps given rise to need-less disagreements.

CONCLUDING REMARKS

Will the perennial problems in the psychologyof emotion remain? Such problems in variousfields often cannot be resolved, since they re-flect either different world views or our limitedhuman capacity for conceptualization. Thewave–particle dilemma in physics would seemto be something of the latter sort; the contrast


between social constructivism and explanationby laws and mechanisms something of the for-mer. But perhaps the scope of the problems inemotion psychology can be narrowed some-what by insight into the relationships betweenproposed solutions for the various problems.

As I have remarked, psychological explan-ations of emotional phenomena are beingsought at different levels. Answers to severalquestions may seem incompatible when theyare answers to questions being sought at differ-ent levels of explanation or at different levels ofthe phenomena.

The study of emotion may advance, I think,when Dennett’s (1987) level distinctions arebetter heeded, and when his functional or psy-chological level receives more attention. In cur-rent analyses of emotion antecedents, little isbeing said about what constitutes a reinforce-ment and why; Schroeder (2004) has coura-geously presented a single effort. Only prelimi-nary endeavors are being made at constructingmodels of appraisal processes and of the innerstructure of concerns, both of which play sucha pivotal role in explaining emotions. As far asI know, for instance, there exist no detailed hy-potheses at the functional level of how innateaffective stimuli evoke affect. I have no ideahow sugar evokes not only the sensation ofsweetness but also the experience of pleasant-ness, even granted that somehow opoids maybecome active. Ultimately, intentional phenom-ena such as experiences, desires, and goalsshould be clarified in terms of subpersonal,functionally defined processes. But such clarifi-cations are scarce (however, see Metzinger,2003). As a consequence, jumps are beingmade from the intentional level to the hard-ware or neurophysiological level, and viceversa. Fear arousal is mediated by theamygdalae, but how exactly do the amygdalaedo that? We sure like certain things, but howdoes that proceed?

All this is important for advances in emotionresearch, because there is no guarantee thatcategories of analysis at one level will projectonto coherent categories at another level.There is no guarantee that emotions as definedexperientially or behaviorally will all involveone mechanism, or one coherent set of mecha-nisms. The mechanisms of fear of failure mayhave little in common with the mechanisms offear of the unknown or of spiders, except thatthey all share the final common pathways of ef-

forts to escape or behavioral inhibition. Thatall stimuli evoking emotions are in some wayappraised does not imply the existence of onesingle coherent appraisal mechanism. And soon.

How explanations at different levels are re-lated depends on the findings at different levels,of course. It would be profitable if researchersin different areas and on different levels talkedand listened more to each other. It would beprofitable if they knew better what is happen-ing at other areas and levels. Experimental in-vestigators of emotions often know little of thesocial and cultural psychology of emotions,and vice versa. Such limitations of knowledgepainfully restrict the range of emotion elicitorsconsidered in the explanatory hypotheses. Stu-dents of the neuropsychology of emotion oftenknow little of the contemporary psychology ofemotion. They sometimes write as if the para-digm of what causes emotions is an electricshock, and as if the paradigm of motivation isthirst or hunger, or something weird like “sur-vival.” To most psychological researchers, thelimbic area is merely somewhere in the brain,and the amygdala is an amorphous blob of tis-sue. There is no real reason why all of thisshould remain so. To facilitate communicationacross disciplines, of course, is one of the mainpurposes of this handbook.

REFERENCES

Ainslie, G. (2001). Breakdown of will. Cambridge, UK:Cambridge University Press.

Arnold, M. B. (1960). Emotion and personality (Vol. I).New York: Columbia University Press.

Arnold, M. B. (1970). Perennial problems in the field ofemotion. In M. B. Arnold (Ed.), Feelings and emo-tions: The Loyola Symposium (pp. 169–186). NewYork: Academic Press.

Bargh, J. A. (1997). The automaticity of everyday life. InR. S. Wyer (Ed.), Advances in social cognition (Vol.10, pp. 1–61). Mahwah, NJ: Erlbaum.

Barrett, L. F. (2005). Feeling is perceiving: Core affectand conceptualization in the experience of emotion.In L. F. Barrett, P. M. Niedenthal, & P. Winkielman(Eds.), Emotion and consciousness (pp. 255–286).New York: Guilford Press.

Barsalou, L. W. (1999). Perceptual symbol systems.Behavioral and Brain Sciences, 22, 577–660.

Bentley, M. (1928). Is “emotion” more than a chapterheading? In M. L. Reymert (Ed.), Feelings andemotions: The Wittenberg Symposium (pp. 17–23).Worcester, MA: Clark University Press.


Berridge, K. C. (2004). Unfelt affect and irrational de-sire: A view from the brain. In A. R. S. Manstead, N.H. Frijda, & A. Fischer (Eds.), Feelings and emo-tions: The Amsterdam Symposium (pp. 243–262).Cambridge, UK: Cambridge University Press.

Bindra, D. (1959). Motivation: A systematic reinterpre-tation. New York: Ronald Press.

Bouton, M. E. (2005). Behavior systems and the contex-tual control of anxiety, fear, and panic. In L. F.Barrett, P. M. Niedenthal, & P. Winkielman (Eds.),Emotion and consciousness (pp. 205–230). NewYork: Guilford Press.

Bradley, M. M., Codispoti, M., Cuthbert, B. N., &Lang, P. J. (2001). Emotion and motivation I: Defen-sive and appetitive reactions in picture processing.Emotion, 1, 276–298.

Buck, R. (1999). Biological affects: A typology. Psycho-logical Review, 106, 301–336.

Buss, D. M. (1994). The evolution of desire. New York:Basic Books.

Camras, L. A. (2000). Surprise!: Facial expressions canbe coordinative motor structures. In M. D. Lewis &I. Granic (Eds.), Emotion, development, and self-organization (pp. 100–124). New York: CambridgeUniversity Press.

Cassirer, E. (1908). Substanzbegriff und funcktions-begriff [Substance concepts and function concepts].Leipzig, Germany: B. Cassirer.

Cayrou, G. (1924). Le français classique. Lexique de lalangue française du dix-septième siècle (2nd ed.).Paris: Dedier.

Chalmers, D. (1996). The conscious mind: In search ofa fundamental theory. New York: Oxford UniversityPress.

Claparède, E. (1928). Emotions and feelings. In M. L.Reymert (Ed.), Feelings and emotions: TheWittenberg Symposium (pp. 124–139). Worcester,MA: Clark University Press.

Clore, G. L., Storbeck, J., Robinson, M. D., &Centerbar, D. (2005). The seven deadly sins of re-search on affect. In L. F. Barrett, P. M. Niedenthal, &P. Winkielman (Eds.), Emotion and consciousness(pp. 384–408). New York: Guilford Press.

Damasio, A. (1994). Descartes’ error: Emotion, reason,and the human brain. New York: Putnam.

Damasio, A. (2003). Looking for Spinoza: Joy, sorrow,and the feeling brain. London: Heinemann.

Dennett, D. C. (1987). The intentional stance. Cam-bridge, MA: MIT Press.

de Sousa, R. (1987). The rationality of emotion. Cam-bridge, MA: MIT Press.

Descartes, R. (1970). Les passions de l’âme [The pas-sions of the soul]. Paris: Vrin. (Original work pub-lished 1649)

De Waal, F. B. M. (1996). Good natured. Cambridge,MA: Harvard University Press.

Duffy, E. (1941). An explanation of “emotional” phe-nomena without the use of the concept “emotion.”Journal of General Psychology, 25, 283–293.

Dumas, G. (1948). La vie affective. Paris: PressesUniversitaires de France.

Ekman, P. (1992). An argument for basic emotions.Cognition and Emotion, 6, 169–200.

Ekman, P. (1994). Strong evidence for universals in fa-cial expression: A reply to Russell’s mistaken cri-tique. Psychological Bulletin, 115, 268–287.

Ellsworth, P. C., & Scherer, K. R. (2003). Appraisal pro-cesses in emotion. In R. Davidson, K. R. Scherer, &H. H. Goldsmith (Eds.), Handbook of the affectivesciences (pp. 572–596). Mahwah, NJ: Erlbaum.

Elster, J. (1999a). Alchemies of the mind. Cambridge,UK: Cambridge University Press.

Elster, J. (1999b). Strong feelings. Cambridge, MA: MITPress.

Frank, R. H. (1988). Passions within reason: The strate-gic role of the emotions. New York: Norton.

Frijda, N. H. (1986). The emotions. Cambridge, UK:Cambridge University Press.

Frijda, N. H. (2005). Emotion experience. Cognitionand Emotion, 19, 473–498.

Frijda, N. H. (2007a). The laws of emotion. Mahwah,NJ: Erlbaum.

Frijda, N. H. (Ed.). (2007b). Klaus Scherer’s article on“What are emotions?”: Comments. Social Science In-formation, 46, 381–443.

Frijda, N. H., Mesquita, B., Sonnemans, J., & VanGoozen, S. (1991). The duration of affective phe-nomena, or emotions, sentiments and passions. In K.T. Strongman (Ed.), International review of studieson emotion (Vol. 1, pp. 187–225). Chichester, UK:Wiley.

Frijda, N. H., & Tcherkassof, A. (1997). Facial expres-sions as modes of action readiness. In J. A. Russell &J. M. Fernández-Dols (Eds.), The psychology of fa-cial expression (pp. 78–102). Cambridge, UK: Cam-bridge University Press.

Gallese, V. (2005). Embodied simulation: From neuronsto phenomenal experience. Phenomenology and theCognitive Sciences, 4, 23–48.

Gould, S. J., & Lewontin, R. C. (1979). The spandrelsof San Marco and the Panglossian paradigm: A cri-tique of the adaptationist programme. Proceedings ofthe Royal Society of London, 205, 581–598.

Gray, J. A. (1987). The psychology of fear and stress(2nd ed.). Cambridge, UK: Cambridge UniversityPress.

Harré, R., & Parrott, W. G. (Eds.). (1996). The emo-tions: Social, cultural and biological dimensions.Thousand Oaks, CA: Sage.

Hebb, D. O. (1949). The organization of behavior. NewYork, Wiley.

Hull, C. L. (1953). A behavior system: An introductionto behavior theory concerning the individual organ-ism. New Haven, CT: Yale University Press.

Hupka, R. B., Lenton, A. P., & Hutchison, K. A. (1999).Universal development of emotion categories in natu-ral language. Journal of Personality and Social Psy-chology, 77, 247–278.


Izard, C. E. (1977). Human emotions. New York: Ple-num Press.


Keltner, D., & Buswell, B. N. (1997). Embarrassment:Its distinct form and appeasement functions. Psycho-logical Bulletin, 122, 250–270.

Kenny, A. (1963). Action, emotion and will. London:Routledge & Kegan Paul.

Kleinginna, P. R., & Kleinginna, A. M. (1981). A cate-gorized list of emotion definitions, with suggestionsfor a consensual definition. Motivation and Emotion,5, 345–379.

Lambie, J., & Marcel, A. (2002). Consciousness andemotion experience: A theoretical framework. Psy-chological Review, 109, 219–259.

Landis, C., & Hunt, W. A. (1932). Adrenalin and emo-tion. Psychological Review, 39, 467–485.


LeDoux, J. (1996). The emotional brain. New York: Si-mon & Schuster.

Levenson, R. W. (2003). Autonomic specificity andemotion. In R. J. Davidson, K. R. Scherer, & H. H.Goldsmith (Eds.), Handbook of affective sciences(pp. 212–224). New York: Oxford University Press.

Lewis, M. (1996). Self-organizing cognitive appraisals.Cognition and Emotion, 10, 1–26.

Lewis, M. D. (2005). Bridging emotion theory and neu-robiology through dynamic system modeling. Behav-ioral and Brain Sciences, 28, 105–131.

Lutz, C. (1988). Unnatural emotions: Everyday senti-ments on a Micronesian atoll and their challenge toWestern theory. Chicago: University of ChicagoPress.

Mandler, G. (1984). Mind and body: The psychology ofemotion and stress. New York: Norton.

Mauss, M. (1957). Essai sur le don [The gift]. London:Routledge & Kegan Paul. (Original work published1914)

McDougall, W. (1923). Outline of psychology. NewYork: Scribner.

Meltzoff, A. N. (2002). Imitation as a mechanism of so-cial cognition: Origins of empathy, theory of mind,and the representation of action. In U. Goswami(Ed.), Handbook of childhood cognitive develop-ment (pp. 6–25). Oxford: Blackwell.

Mesquita, B. (2003). Emotions as dynamic culturalphenomena. In R. J. Davidson, K. R. Scherer, &H. Goldsmith (Eds.), Handbook of affective sci-ences (pp. 871–890). New York: Oxford UniversityPress.

Mesquita, B., Frijda, N. H., & Scherer, K. R. (1997).Culture and emotion. In P. R. Dasen & T. S.Saraswathi (Eds.), Handbook of cross-cultural psy-chology (Vol. 2, pp. 255–298). Boston: Allyn &Bacon.

Metzinger, T. (2003). Being no one: The self-model the-ory of subjectivity. Cambridge, MA: MIT Press.

Moors, A., & De Houwer, J. (2006). Automaticity: A

theoretical and conceptual analysis. PsychologicalBulletin, 132, 297–326.

Mowrer, O. H. (1960). Learning theory and behavior.New York: Wiley.

Niedenthal, P. M., Barsalou, L. W., Ric, F., & Krauth-Gruber, S. (2005). Embodiment in the acquisitionand use of emotion knowledge. In L. F. Barrett, P. M.Niedenthal, & P. Winkielman (Eds.), Emotion andconsciousness (pp. 21–50). New York: GuilfordPress.

Nussbaum, M. C. (2001). Upheavals of thought: Theintelligence of emotions. New York: Cambridge Uni-versity Press.

Oatley, K. (1992). Best laid schemes: The psychology ofemotions. Cambridge, UK: Cambridge UniversityPress.

Oatley, K., & Duncan, E. (1994). The experience ofemotions in daily life. Cognition and Emotion, 8,369–382.

Öhman, A., & Mineka, S. (2001). Fears, phobias, andpreparedness: Toward an evolved module of fear andfear learning. Psychological Review, 108, 483–522.

Ortony, A., Clore, G., & Collins, A. (1988). The cogni-tive structure of emotions. Cambridge, UK: Cam-bridge University Press.

Ortony, A., & Turner, T. (1990). What’s basic about ba-sic emotions? Psychological Review, 97, 315–331.

Parrott, W. G. (2001). Implications of dysfunctionalemotions for understanding how emotions function.Review of General Psychology, 5, 180–186.

Plutchik, R. (2003). Emotions and life. Washington,DC: American Psychological Association.

Rimé, B. (2005). Emotion et expression. Paris: PressesUniversitaires de France.

Rolls, E. T. (2005). Emotion explained. Oxford: OxfordUniversity Press.

Rothbaum, F. M., Weisz, J. R., Pott, M., Miyake, K., &Morelli, G. (2000). Attachment and culture: Securityin the United States and Japan. American Psycholo-gist, 55, 1093–1104.

Rozin, P. (1999). Preadaptation and the puzzles andproperties of pleasure. In D. Kahneman, E. Diener, &N. Schwarz (Eds.), Well-being: The foundations ofhedonic psychology. Scientific perspectives on enjoy-ment and suffering (pp. 109–133). New York: Rus-sell Sage Foundation.

Russell, J. A. (1991). Culture and the categorization ofemotions. Psychological Bulletin, 110, 426–450.


Scherer, K. R. (1992). What does facial expression ex-press? In K. T. Strongman (Ed.), International reviewof studies on emotion (Vol. 2, pp. 139–165).Chichester, UK: Wiley.

Scherer, K. R. (2000). Emotions as episodes of subsys-tem synchronization driven by nonlinear appraisalprocesses. In M. D. Lewis & I. Granic (Eds.), Emo-tion, development, and self-organization (pp. 70–99). New York: Cambridge University Press.


Scherer, K. R. (2001). Appraisal considered as a processof multilevel sequential checking In K. R. Scherer, A.Schorr, & T. Johnstone (Eds.), Appraisal processes inemotion: Theory, methods, research (pp. 92–120).New York: Oxford University Press.

Scherer, K. R. (2005). What are emotions? And how canthey be measured? Social Science Information, 44,695–729.

Schroeder, T. (2004). Three faces of desire. Oxford: Ox-ford University Press.

Shand, A. F. (1920). The foundations of character: Astudy of the emotions and sentiments. London:Macmillan.

Shaver, P., Wu, S., & Schwartz, J. C. (1992). Cross-cultural similarities and differences in emotion and itsrepresentation: A prototype approach. In M. Clark(Ed.), Review of personality and social psychology(Vol. 13, pp. 175–212). Newbury Park, CA: Sage.

Shweder, R. A., & Haidt, J. (2000). The cultural psy-chology of the emotions: Ancient and new. In M.Lewis & J. M. Haviland-Jones (Eds.), Handbookof emotions (2nd ed., pp. 397–416). New York:Guilford Press.

Sloman, A. (1987). Motives, mechanisms and emotions.Cognition and Emotion, 1, 217–234.

Smith, C. A., & Scott, H. H. (1997). A componentialapproach to the meaning of facial expressions. In J.A. Russell & J.-M. Fernández-Dols (Eds.), The psy-chology of facial expression (pp. 229–254). Cam-bridge, UK: Cambridge University Press.

Solomon, R. C. (1993). The passions (2nd ed.). India-napolis, IN: Hackett.

Solomon, R. C. (2004). Not passion’s slave. New York:Oxford University Press.

Spinoza, B. (1989). Ethics. (G. H. R. Parkinson, Trans.).London: Dent. (Original work published 1677)

Stein, N., & Trabasso, T. (1992). The organization ofemotional experience: Creating links between emo-tion, thinking, and intentional action. Cognition andEmotion, 6, 225–244.

Stemmler, G. (2003). Methodological considerations inthe psychophysiological study of emotion. In R. J.Davidson, K. R. Scherer, & H. H. Goldsmith (Eds.),Handbook of affective sciences (pp. 225–255). NewYork: Oxford University Press.

Thelen, E. (1995). Motor development: A new synthe-sis. American Psychologist, 50, 79–95.

Tomkins, S. S. (1962). Affect, imagery and conscious-ness: Vol. 1. The positive affects. New York: Springer.

Wallon, H. (1942). De l’acte à la pensée. Paris:Flammarion.

Watson, J. B. (1929). Psychology from the stand-point of a behaviorist (3rd ed.). Philadelphia: Lip-pincott.

Weiskrantz, L. (1997). Consciousness lost and found:A neuropsychological exploration. Oxford: OxfordUniversity Press.

Weiss, J. M., Glazer, H. I., & Poherecky, L. A. (1976).Coping behavior and neurochemical changes: Analternative explanation for the original “learnedhelplessness” experiments. In A. Serban & A. Kling(Eds.), Animal models in human psychobiology(pp. 141–173). New York: Plenum Press.

Wierzbicka, A. (1995). Everyday conceptions of emo-tion: A semantic perspective. In J. A. Russell, J.-M.Fernández-Dols, A. S. R. Manstead, & J.Wellenkamp (Eds.), Everyday conceptions of emo-tion (pp. 17–48). Dordrecht, The Netherlands:Kluwer.

Zajonc, R. B. (1980). Thinking and feeling: Preferencesneed no inferences. American Psychologist, 35, 151–175.


C H A P T E R 6

The Clinical Applicationof Emotion in Psychotherapy

LESLIE S. GREENBERG

This chapter addresses the emerging focus onemotion in psychotherapy. Given that emotionis seen as information (i.e., as signaling the sig-nificance of a situation to a person’s well-being), and that affect regulation is seen as akey human motivation, it has become clear thatemotion needs to be focused on, validated, andworked with directly in therapy to promoteemotional change (Greenberg, 2002; Samoilov& Goldfried, 2000; Fosha, 2000). The ideathat accessing and exploring painful emotions,within the context of a secure therapeutic rela-tionship, makes one feel better has been widelyheld by several schools of psychotherapy(Freud, 1915/1957; Rogers, 1951; Perls, 1969),but it has been difficult to prove. However,over the last decade, newer therapeutic ap-proaches that treat affect as a primary target ofintervention within the context of an empathicrelationship have been developed and tested. Inthis chapter, results of evidence-based treat-ment studies that show evidence for effective-ness of emotion-focused treatment of mooddisorders, personality disorders, and traumaare reviewed briefly. Principles of emotion

assessment and emotional change are then dis-cussed. These provide a map for differentialintervention with emotion by showing that dif-ferent classes of emotions in therapy benefitfrom different types of interventions—rangingfrom awareness of adaptive emotions, throughregulation of dysregulated emotions, to trans-formation of maladaptive emotions. The chap-ter concludes with a discussion of differentmethods of activating new emotional responsesto change habitual problematic emotional re-sponses.

EVIDENCE-BASED TREATMENT

A number of treatment approaches that focuson painful emotions have been demonstratedin randomized clinical trials to be effective(Elliott, Greenberg, & Lietaer, 2004; Green-berg & Pascual-Leone, 2006; Whelton, 2004).A manualized form of emotion-focused ther-apy (EFT) for depression has been found to behighly effective in treating depression in threeseparate studies (Greenberg & Watson, 1998;

88

Goldman, Greenberg, & Angus, 2006; Watson,Gordon, Stermac, Kalogerakos, & Steckley,2003). In these studies, EFT was found to be aseffective as or more effective than either apurely relational empathic treatment or a formof cognitive-behavioral therapy (CBT). Boththe treatments with which it was comparedwere themselves also found to be highly effec-tive in reducing depression, but EFT was foundto be more effective in both reducing inter-personal problems and promoting change insymptoms than the purely relational treatment.It also was found to be highly effective in pre-venting relapse.

The objective of EFT is to access and restruc-ture the habitual maladaptive emotional statesthat are seen as the source of the depression.This involves accessing feelings of shame-basedworthlessness, anxious dependence, power-lessness, abandonment, and invalidation, andtransforming these through accessing adaptiveemotions (such as healthy grief and empower-ing anger), as well as reflecting on emotionalexperience to create new meaning and developnew narratives. This process of emotionalchange is aided by the use of specific therapeu-tic techniques that help stimulate both thearousal and processing of emotion. Various ex-periential interventions—such as two-chair di-alogue in response to in-session statements ofself-critical conflicts, empty-chair dialogue inresponse to in-session statements of unresolvedfeelings toward a significant other, and focus-ing on bodily felt meaning in response to an un-clear felt sense—are used to engage patients inemotion processing in session.

Short-term dynamic therapy, which workson overcoming defenses and treats affect pho-bia by exposure to dreaded emotion, has beenfound to be effective in treating personality dis-orders in two studies (Winston et al., 1994;Svartberg, Stiles, & Seltzer, 2004). EFT foradult survivors of childhood abuse, which usesempathy plus empty-chair work and involvesthe arousal and processing of painful emotions,has been found effective in treating abuse(Paivio & Nieuwenhuis, 2001). Emotionallyfocused couple therapy (Greenberg & Johnson,1988; Johnson, 1996), which involves part-ners’ revealing their underlying attachment-and identity-related vulnerable feelings to eachother, has been found to be effective in treat-ing couples’ distress (Johnson, Hunsley,Greenberg, & Schindler, 1999). In addition,versions of CBT based on exposure to imaginal

stimuli have been shown to be effective forposttraumatic stress disorder (PTSD) and otheranxiety disorders (Borkovec, Alcaine, & Behar,2004; Shapiro, 1999). Finally, therapy basedon an avoidance theory in which worry is un-derstood as a cognitive response that orients in-dividuals to a threat, while insulating themfrom the immediacy of their emotional experi-ence, has recently gained support (Borkovec etal., 2004).

A recent study examining the therapists’stance in interpersonal therapy (IPT) and CBTfor depression showed the importance of focus-ing on emotion, regardless of orientation. Thisstudy (Coombs, Coleman, & Jones, 2002)found that collaborative emotional exploration(which occurred significantly more frequentlyin IPT) was found to relate positively to out-come in both forms of therapy, whereas educa-tive/directive process (which was more fre-quent in CBT) had no relationship to outcome.Helping people overcome their avoidance ofemotions, focusing collaboratively on emo-tions, and exploring them in therapy thus ap-pears to be important in therapeutic change,whichever therapeutic orientation is employed.What is needed now is a more differentiatedunderstanding of how to work with emotion. Adifferential approach to assessment and inter-vention has recently emerged.

DIFFERENTIAL EMOTIONASSESSMENT AND INTERVENTION

In assessing and working with emotion in ther-apy, it is helpful to make certain distinctionsbetween different types of emotional experi-ences and expressions, which require differenttypes of in-session interventions. This involvesdifferentiating between “primary” emotionsand “secondary” or “reactive” emotions, andbetween “adaptive” and “maladaptive” emo-tional experiences (Greenberg, 2002). Primaryemotions are people’s first, gut-level emotionalresponses to situations. These responses needto be accessed and assessed for whether theyprovide adaptive information and the capacityto organize action, or whether they are mal-adaptive. Maladaptive emotions are learned re-sponses that are not appropriate to current sit-uations and thus are no longer adaptive; theseemotions need to be regulated and trans-formed. Secondary emotions are those re-sponses that are secondary to other, more pri-

Application of Emotion in Psychotherapy 89

mary internal processes and may be defensesagainst these, such as feeling ashamed of one’ssadness or hopeless when angry. Secondarymaladaptive emotions need to be explored toaccess their more primary cognitive or emo-tional generators. These distinctions betweendifferent types of emotions thus provide clini-cians with a map for differential interventionwith emotions.

PRINCIPLES OF WORKING WITHEMOTION

Outcome and process research findings pointtoward emotional processing as centrally im-portant to good therapy, but what good pro-cessing is actually remains to be elucidated.Emotional insight, catharsis, awareness, andexposure have all been put forward as explana-tions of the role of emotion in change, but thereis still not a comprehensive, empirically basedunderstanding of how emotion and its process-ing lead to change. The following five princi-ples provide an empirically based understand-ing of emotional change processes in clinicalchange: (1) increasing awareness of emotion,(2) expressing emotion, (3) enhancing emotionregulation, (4) reflecting on emotion, and (5)transforming emotion (Greenberg, 2002;Greenberg & Watson, 2006).

Emotion Awareness

The first and most general goal of emotionalchange is for clients to become aware of theiremotions in general and their primary adap-tive emotions in particular. Increased emotionawareness is therapeutic in a variety of ways.Becoming aware of core emotional experienceand symbolizing it in words provide access toboth the adaptive information and action ten-dency in the emotion. Labeling emotions is of-ten a first step in problem definition. It is im-portant to note that emotion awareness is notthinking about feeling; it involves feeling thefeeling in awareness. Only once an emotion isfelt does its articulation in language become animportant component of its awareness.

The therapist thus needs to help clients ap-proach, tolerate, and accept their emotions.Acceptance of emotional experience as op-posed to its avoidance is the first step in aware-ness work. Once clients have accepted theiremotions rather than avoided them, the thera-

pist then helps the clients in the utilization ofemotion to improve coping. That is, clients arehelped to make sense of what each emotion istelling them and to identify the goal/need/con-cern that it is organizing them to attain. Emo-tion is thus used both to inform and to move.

A measure of levels of emotion awarenesshas been developed by Lane and Schwartz(1992). Five such levels can be measured. In as-cending order, these are physical sensations, ac-tion tendencies, single emotions, blends ofemotions, and blends of blends of emotionalexperience (the capacity to appreciate com-plexity in the experiences of self and other).The dynamic interactions among phenomenalexperience, establishing a representation of it,elaborating that representation (e.g., identify-ing the source of the emotional response), andintegrating it with other cognitive processes arethe fundamental processes involved in the cog-nitive elaboration of emotion and addressed bythe levels of this measure. Scores on the mea-sure have been found to correlate significantlywith self-reported self-restraint and impulsecontrol. Individual differences in emotionawareness have also been found to predict re-covery of positive mood and decrements inruminative thoughts following a distressingstimulus (Salovey, Mayer, Golman, Turvey, &Palfai, 1995).

Awareness of emotion also involves over-coming the avoidance of emotional experience.There is a strong human tendency to avoid orinterrupt painful emotions. Normal cognitiveprocesses often deny, distort, or interrupt emo-tion and transform adaptive but unpleasantemotions into dysfunctional behaviors de-signed to avoid feeling. Leahy (2002) has de-fined “emotional schemas” as cognitive struc-tures that frame the interpretation of emotionalexperience and guide the strategies used in cop-ing with emotion. He has noted that there aretwo fundamental coping pathways for deal-ing with emotion. One involves attending toand labeling emotions in a manner that ac-cepts and normalizes them; the other pathwaypathologizes some emotional experiences, andthis leads to attempts to distort or avoid them(initiating guilt, frantic efforts at control, ob-sessive rumination, etc.). To overcome emotionavoidance, clients must first be helped to ap-proach emotions by attending to their emo-tional experience. This often involves changingthe cognitions governing their avoidance. Thenclients must allow and tolerate being in live


contact with their emotions. These two stepsare consistent with notions of exposure.

Emotional Arousal and Expression

Emotional expression has recently been shownto be a unique aspect of emotional processingthat predicts adjustment to breast cancer(Stanton et al., 2000). Women who coped withcancer through expressing emotion had fewermedical appointments, enhanced physicalhealth and vigor, and decreased distress thanthose low in emotional expression did. Expres-sive coping was also related to increased qual-ity of life for those who perceived their socialenvironment to be highly receptive. Analysesfurther suggested that expressive coping en-hanced the pursuit of goals, perhaps by helpingclients attend to and clarify central concerns,but that this relationship was mediated byhope. Emotional arousal and its expression intherapeutic contexts thus appear to constitute atherapeutic aspect of emotional processing. Ex-pressing emotion in therapy does not involvesimply venting emotion, but rather overcomingavoidance of, strongly experiencing, and ex-pressing previously constricted emotions.

There is a long line of evidence on the effec-tiveness of arousal of and exposure to pre-viously avoided feelings as a mechanism ofchange. Results from a variety of studies indi-cate that emotional engagement (fear expres-sion) with trauma memories during exposurein early sessions, and habituation (reduced dis-tress) during exposure over the course of ther-apy, predict better outcome (Foa & Jaycox,1999; Jaycox, Foa, & Morral, 1998). Emo-tional engagement with trauma memories earlyin therapy appears to be important in overcom-ing trauma. However, only a subgroup of indi-viduals are able to engage in the exposure tech-nique and therefore maximally benefit fromtherapy (Jaycox et al., 1998). Pretreatment se-verity of PTSD symptoms is also associatedwith engagement difficulties and poorer out-come. Foa, Zoellner, Feeny, Hembree, andAlvarez-Conrad (2002) have shown that al-though imaginal exposure, which arousesstrong emotion, can exacerbate symptoms insome clients, it does this in relatively few cases,and even then this does not impede a long-termpositive outcome. Overall, a chain of factorsbeginning with trauma symptom severity,through initial engagement in imaginal expo-sure, activation of the fear structure, and re-

peated exposure while providing new informa-tion, appears to predict outcome (Jaycox et al.,1998).

However, studies on exposure and arousaldo not take into account the importance to cli-ent change of the therapeutic relationship. Ina process–outcome study evaluating EFT foradult survivors of childhood abuse (Paivio,Hall, Holowaty, Jellis, & Tran, 2001), the ther-apeutic alliance, initial engagement in the pri-mary imaginal confrontation intervention, theoverall dosage of this intervention (quality ×frequency), and client predisposing variablesall contributed to reduced global and trauma-specific symptomatology and interpersonalproblems. The effect of emotional arousal intherapy also depends on the quality of theworking alliance. Beutler, Clarkin, and Bongar(2000) studied several therapies, in an attemptto match patient variables with treatments.Across modalities, emotional intensity of ses-sions was a strong predictor of outcome, butthis effect was mediated by the working alli-ance. Likewise, Iwakabe, Rogan, and Stalikas(2000) documented that high arousal predictedgood session outcome only when there was astrong alliance.

Learning to contain and regulate strongemotions is central to adaptive emotionalarousal and expression, and these skills are of-ten lacking in people seeking therapy. The abil-ity to regulate emotion is believed to emergefrom early attachment experiences of valida-tion, soothing, and safety, and involves attend-ing to emotions and dampening or expressingthem as appropriate. The emotional validationand empathy of the therapist seem to be partic-ularly important in allowing clients with dys-regulated emotions to learn to self-soothe andrestore emotional equilibrium (Greenberg,2002).

It is clear is that emotional arousal and ex-pression are not always helpful or appropriatein therapy or in life, and that for some clients,training in the capacity for emotion regulation(see next section) must precede or accompanyit (Greenberg, 2002). Any benefits believed toaccrue from the intense expression of emotionare generally predicated on the client’soverregulation (overcontrol) or suppression ofemotion (Gross, 1999), but it is apparent thatfor some individuals with certain psychologicaldisorders or in particular situations, emotionsare under- or dysregulated (Linehan, 1993;Gross, 1999). Some support has been found for


the cathartic expression of feeling in therapy,but only with certain people under specifiedcircumstantial conditions (Pierce, Nichols, &DuBrin, 1983). Catharsis is not helpful all thetime for all people. Process–outcome researchon EFT for depression, however, has shownthat higher expressed emotional arousal atmidtreatment predicted positive treatment out-comes (Warwar, 2003). This supports the im-portance of expressed arousal as a key changeprocess in these treatments. It is important alsoto note that this study measured expressed asopposed to experienced emotion. A follow-upstudy examining in-session client reports ofexperienced emotional intensity (Warwar,Greenberg, & Perepeluk, 2003) found that cli-ent reports of in-session experienced emotionwere not related to positive therapeutic change.A discrepancy was observed between clients’reports of in-session experienced emotions andthe emotions that were actually expressed,based on arousal ratings of videotaped therapysegments. For example, one client reported thatshe had experienced intense emotional painand anger in a session. Her level of expressedemotional arousal, however, was judged to bevery low by observers who rated emotionalarousal from videotaped therapy segments.

Exposure methods have established a basisfor understanding the emotional processing re-quired for therapeutic change in the treatmentof fear and anxiety. Hunt (1998), however,looked at emotional processing of depressiveevents and found that although greater short-term attention to negative feelings inducedshort-term emotional pain, those who wentthrough this pain felt better in the long runthan individuals who engaged in problem solv-ing or avoided processing their feelings afterthe depressive event. This benefit was mediatedby degree of emotional arousal, suggestingagain that emotions must be “up and running”and must be experienced for beneficial emo-tional processing to occur. Evidence also hasbeen found supporting the specific effectivenessof arousing and expressing anger in the treat-ment of depression and traumatic sexual abuse(Beutler et al., 1991; Van Velsor & Cox, 2001).Anger can be a means for survivors of sexualabuse to develop self-efficacy, heal memories,and correctly attribute blame. In these studies,arousal and expression of anger was related tothe development of agency, self-efficacy, andself-assertion. In a review of research literaturerelated to the benefits and dangers of

reexperiencing painful emotion in therapy,Littrell (1998) concluded that when therapy isdesigned so as to allow for the planned restruc-turing of painful memories, the reexperience ofpain in therapy has been demonstrated to bebeneficial and therapeutic.

Pierce et al. (1983) found that catharsis wastherapeutically useful only under very specificcircumstances and only for certain people.There can be no universal rule about the effec-tiveness of arousing emotion or evoking emo-tional expression. The role of arousal and thedegree to which it may be useful in therapy de-pend on what emotion is expressed and aboutwhat issue; how it is expressed, by whom, towhom, when, and under what conditions; andin what way the emotional expression is fol-lowed by other experiences of affect and mean-ing (Whelton, 2004). Nonetheless, the evidencesuggests that emotional processing is mediatedby arousal, so that for emotion processingto occur, the distressing affective experiencemust be activated and viscerally experienced bythe client. Arousal is necessary but not suffi-cient for therapeutic progress. Recently we(Greenberg, Auszra, & Herrmann, 2007)found that the productivity of processing ofaroused emotions, rather than arousal alone,distinguished good from poor outcomes.

Emotion Regulation

The third principle of emotional processing in-volves the regulation of emotion. Important is-sues in any treatment are what emotions are tobe regulated and how. Undercontrolled second-ary emotions and maladaptive emotions arewhat need to be regulated. Clients with thesetypes of underregulated affect have been shownto benefit both from validation and from thelearning of specific emotion regulation and dis-tress tolerance skills (Linehan, 1993).

The provision of a safe, validating, support-ive, and empathic environment is the first levelof intervention for automatically generatedunderregulated distress (Bohart & Greenberg,1997). Linehan et al. (2002) found evidence forthe effectiveness of emotional validation andsoothing as part of the treatment for borderlinepersonality disorder. Empathy from anotherperson seems to be particularly important inlearning to self-soothe, restore emotional equi-librium, and strengthen the self.

Emotion regulation skills—including suchthings as identifying and labeling emotions, al-


lowing and tolerating emotions, establishing aworking distance, increasing positive emotions,reducing vulnerability to negative emotions,self-soothing, breathing, and distraction—alsohave been found to help with high distress(Linehan, 1993). Particularly important amongthese skills are getting some distance fromoverwhelming shame, despair, hopelessness,and/or shaky vulnerability, and developing self-soothing capacities to calm and comfort coreanxieties and humiliation. Forms of meditativepractice and self-acceptance are often mosthelpful in achieving a working distance fromoverwhelming core emotions. The ability toregulate breathing, and to observe one’s emo-tions and let them come and go, are importantprocesses to help regulate emotional distress.Mindfulness treatments have been shown to beeffective in treating generalized anxiety disor-der and panic (Kabat-Zinn et al., 1992), treat-ing chronic pain (Kabat-Zinn, Lipworth,Burney, & Sellers, 1986), and preventing re-lapse in depression (Teasdale et al., 2000).

Another important aspect of regulation,however, is developing clients’ abilities to toler-ate emotion and to self-soothe automatically.Such abilities can be developed at various levelsof processing. Physiological soothing involvesteaching clients to activate the parasympatheticnervous system to regulate heart rate, breath-ing, and other sympathetic functions that speedup under stress. At the more deliberate behav-ioral and cognitive levels, promoting clients’abilities to receive and be compassionate totheir emerging painful emotional experience isthe first step toward helping them develop au-tomatic emotion tolerance and self-soothing.This form of self-soothing involves, amongother things, diaphragmatic breathing, relax-ation, and the development of self-empathyand self-compassion.

Soothing also comes interpersonally in theform of another’s empathic attunement to one’saffect and through acceptance and validationby another person. Internal security developsfrom feeling that one exists in the mind andheart of another, and the security of being ableto soothe the self develops through internal-izing the soothing functions of the protectiveother (Sroufe, 1996). It is also important tomake a distinction in emotion work betweenintensity of emotion per se and the depth ofprocessing of the emotion. It is the latter that isthe aim in EFT, not the former, and the regula-tion of overwhelming emotional intensity is vi-

tal in promoting the required depth ofprocessing of emotion.

Reflection on Emotion

The fourth principle of emotional change is re-lated to the first principle, emotion awareness,in that it involves making meaning of emotion.Reflection on emotion helps people make senseof their experience and promotes its assimila-tion into their ongoing self-narratives. Whatwe make of our emotional experience makes usall who we are. In addition to the informa-tional value of emotion awareness, symbolizingemotion in awareness promotes reflection onexperience to create new meaning, and thishelps people develop new narratives to explaintheir experience (Pennebaker, 1995; Greenberg& Angus, 2004). Understanding an emotionalexperience always involves putting it into nar-rative form. In therapy as well as in literature,all emotions occur in the context of significantstories, and all stories involve significant emo-tions (Greenberg & Angus, 2004). Therapythus involves both change in emotional experi-ences and change in the narratives in whichthey are embedded.

In particular, symbolizing traumatic emotionmemories in words helps promote their assimi-lation into people’s ongoing self-narratives (vander Kolk, 1995). This process of verbalizationallows previously unsymbolized experience inemotion memory to be assimilated into peo-ple’s conscious, conceptual understandings ofself and world, where it can be organized into acoherent story. Once such emotions are inwords, they allow people to reflect on whatthey are feeling, create new meanings, and eval-uate their own emotional experience. For ex-ample, reflecting on interpersonal difficultiesand understanding that one is prone to get an-gry at one’s partner because one is feeling aban-doned, and that this relates to one’s ownhistory of abandonment rather than to thewithholding nature of the partner, can be mosttherapeutic.

Pennebaker and colleagues have shown thepositive effects of writing about emotional ex-perience on autonomic nervous system activity,immune functioning, and physical and emo-tional health. Pennebaker (1995) concludesthat through language, individuals are able toorganize, structure, and ultimately assimilateboth their emotional experiences and theevents that may have provoked the emotions.


Both insight and reframing of emotional expe-rience have long been viewed as ways to changeemotion. The role in psychotherapy of humans’capacity for conscious awareness of the pro-cesses and contents of their own minds, and forreason and insight to shed light on unconsciousmotivations, has been substantial—from thebeginnings of psychoanalysis right up to thepresent day. In addition, many therapists havewritten on the importance of changing people’sassumptive frameworks in therapy (see, e.g.,Frank, 1961).

In a study of events in which problematic is-sues were resolved in session, Watson (1996)found that vivid descriptions, emotionalarousal, and cognitive meaning making inter-acted in complex yet orderly stages to producetherapeutic change. Theses stages allowed forclients to reflect on the emotions they were ex-periencing. Similarly, Stalikas and Fitzpatrick(1995) did an intensive analysis of “good clientmoments” and showed that in-session changewas related to the combination of strength offeeling and higher-order levels of reflection. Inaddition, computer-assisted studies of verbalpatterns in psychodynamic and other therapieshave shown that in the key moments in therapyin which substantial shifts happened, there wasa frequent co-occurrence of high emotion tone(emotional arousal) and high abstraction (a re-flection on this emotional process)—a benefi-cial co-occurrence that was called “making aconnection” (Mergenthaler, 1996). It seems tobe the timely conjunction of emotional arousaland a thoughtful exploration of the emotion’smeaning that generates change.

Thus, as well as having emotions, we alsolive in a constant process of making sense ofour emotions. A dialectical-constructivist viewof human functioning has been offered toexplain this process (Greenberg & Pascual-Leone, 1995, 2001; Neimeyer & Mahoney,1995). In this view, personal meaning emergesfrom the self-organization and explication ofone’s own emotional experience, and optimaladaptation involves an integration of reasonand emotion. This integration is achieved by anongoing circular process of making sense of ex-perience by symbolizing bodily felt sensationsin awareness and articulating them in lan-guage, thereby constructing new ones.

In this dialectical-constructivist view, symboland bodily felt referents are viewed as interact-ing to carry meaning forward, and newly sym-bolized experience is organized in different

ways to construct new views. Attendingto and discovering preconceptual elements ofemotional experience influence the process ofmeaning construction. New experiential ele-ments from many sources from within, andsometimes from without, can be integrated intothis process. People are then viewed as con-stantly striving toward making sense of theirpreconceptual emotional experience by sym-bolizing it, explaining it, and putting it intonarrative form. Preconceptual tacit meaningcarries implications and acts to constrain, butdoes not fully determine, meaning. Rather, it issynthesized with conceptual, explicit meaningto form explanations constrained by experienc-ing (Greenberg & Pascual-Leone, 1995, 2001).This provides the ongoing narrative of a per-son’s life.

Thus, although the recipe for emotional pro-cessing from the perspective of behavioral ther-apies and CBT is that arousal plus habituationto the distressing stimulus produces change,approach, arousal, acceptance, and toleranceof emotional experience are necessary but notsufficient from an EFT perspective. Optimumemotional processing involves in addition theintegration of cognition and affect and thecreation of new meaning (Greenberg, 2002;Greenberg & Pascual-Leone, 1995; Pos,Greenberg, Goldman, & Korman, 2003). Oncecontact with emotional experience is achieved,clients must also cognitively orient to that ex-perience as information, and explore, reflecton, and make sense of it.

EFT appears to work by enhancing emo-tional processing, and this involves helpingpeople both accept their emotions and makesense of them (Pos et al., 2003; Goldman,Greenberg, & Pos, 2005). Deepening of experi-ence over therapy as measured by the Experi-encing Scales (Klein, Mathieu-Coughlan, &Kiesler, 1986), which measure clients’ ability tofocus on feelings and use them to solve prob-lems and create new meaning, has been shownto be a specific change process that predictsoutcome over and above the change predictedby the therapeutic alliance. Past studies alsoshow a strong relationship between in-sessionemotional experiencing and therapeutic gainin dynamic, cognitive, and experiential thera-pies (Castonguay, Goldfried, Wiser, Raue, &Hayes, 1996; Silberschatz, Fretter, & Curtis,1986). This suggests that this variable may bea common factor that helps explain changeacross approaches.


In addition, it has been shown that thera-pists’ depth of experiential focus influencedclients’ depth of experiencing in the nextmoment, and that this predicted outcome.Moreover, the effect of early emotional pro-cessing on outcome was found to be mediatedby late emotional processing, where “emo-tional processing” was defined as depth of ex-periencing emotion episodes (Pos et al., 2003).Early capacity for emotional processing alonethus did not guarantee good outcome. Nor didentering therapy without this capacity guaran-tee poor outcome. Therefore, although earlyemotional processing skill was probably an ad-vantage, it appeared not as critical as the abilityto acquire and/or increase depth of experienc-ing throughout therapy. In this study, late emo-tional processing independently added 21% tothe explained variance in reduction in symp-toms over and above early alliance and earlyemotional processing level.

In another study (Warwar, 2003), not onlydid midtherapy expressed emotional arousalpredict outcome; a client’s ability to use inter-nal experience to make meaning and solveproblems as measured by the ExperiencingScales, particularly in the late phase of treat-ment, added to the outcome variance over andabove middle-phase emotional arousal. Thisstudy thus showed that a combination of emo-tional arousal and reflection was a better pre-dictor of outcome than either index alone. Re-flection on aroused emotion thus appears tobe an important change process (Missirlian,Toukmanian, Warwar, & Greenberg, 2005).

Emotion Transformation

The final and probably most fundamental prin-ciple of emotional change is the transformationof one emotion into another. This applies mostspecifically to transforming primary maladap-tive emotions—those old familiar bad feelingsthat occur repeatedly but do not change bycontact with more adaptive emotions. Al-though the more traditional ways of transform-ing emotions (either exposure through experi-ence, expression, and completion, or reflectionon them) can occur with primary maladaptiveemotions, another process appears to be moreimportant. This is a process of changing emo-tion with emotion (Greenberg, 2002). In otherwords, a maladaptive emotional state can betransformed best by undoing it with another,more adaptive emotion. In time, the activation

of the more adaptive emotion along with or inresponse to the maladaptive emotion helpstransform the maladaptive emotion. Whilethinking usually changes thoughts, only feelingcan change emotions.

Spinoza was the first to note that emotion isneeded to change emotion; he proposed that“an emotion cannot be restrained nor removedunless by an opposed and stronger emotion”(1677/1967, p. 195). Reason clearly is seldomsufficient to change automatic emergency-based emotional responses. Darwin (1872/1998), on jumping back from the strike of aglassed-in snake, noted that despite his havingapproached it with the determination not tostart back, his will and reason were powerlessagainst the imagination of a danger that he hadnever even experienced. Rather than reasoningwith an emotion, one needs to transform theemotion by accessing another emotion.

In an interesting line of investigation, posi-tive emotions have been found to undo linger-ing negative emotions (Frederickson, 2001).The basic observation is that key componentsof positive emotions are incompatible withnegative emotions. Frederickson (2001) sug-gests that by broadening a person’s momentarythought–action repertoire, a positive emotionmay loosen the hold that a negative emotionhas on the person’s mind. The experiencesof joy and contentment were found to pro-duce faster cardiovascular recovery from nega-tive emotions than a neutral experience.Frederickson, Mancuso, Branigan, and Tugade(2000) also found that resilient individualscoped by recruiting positive emotions to reg-ulate negative emotional experiences. Theyfound that these individuals manifested a phys-iological “bounce-back” that helped them toreturn to cardiovascular baseline more quickly.

It thus seems possible to replace bad feelingswith happy feelings—not in a simple manner,by trying to look on the bright side, but byevoking meaningfully embodied alternative ex-periences to undo the negative feelings. For ex-ample, in grief, laughter has been found to be apredictor of recovery; thus being able to re-member the happy times, to experience joy,serves as an antidote to sadness (Bonanno &Keltner, 1997). Similarly, warmth and affectionare often antidotes to anxiety. In depression, aprotest-filled, submissive sense of worthless-ness can be transformed therapeutically byguiding people to the desire that drives theirprotest—a desire to be free of their cages and to


access their feelings of joy and excitement forlife. Isen (1999) notes that at least some of thepositive effects of happy feelings have been hy-pothesized to depend on the effects of theneurotransmitters involved in the emotion ofjoy on specific parts of the brain that influencepurposive thinking. Mild positive affect hasbeen found to facilitate problem solving.

In a study of self-criticism, Whelton andGreenberg (2004) found that people who weremore vulnerable to depression showed morecontempt but also less resilience in response toself-criticism than people less vulnerable to de-pression. The less vulnerable people were ableto recruit positive emotional resources, suchas self-assertive pride and anger, to combatthe depressogenic contempt and negativecognitions. In other words after a distressingexperience, resilient people appear to generatea positive feeling (often through imagery ormemory) in order to soothe themselves, andthey can combat negative feelings and views ofself in this more resilient state. Accessing a pos-itive emotional state therefore helps themcounteract the effect of a negative emotionalstate. These studies together indicate that posi-tive affect can be used to regulate negative feel-ings.

Davidson (2000) suggests that the right-hemispheric, withdrawal-related negative af-fect system can be transformed by activationof the approach-related system in the leftprefrontal cortex. He defines “resilience” asthe maintenance of high levels of positive affectand well-being in the face of adversity; he high-lights that it is not that resilient people do netfeel negative affect, but that their negative af-fect does not persist. Levenson (1992) has re-viewed research indicating that specific emo-tions are associated with specific patterns ofautonomic nervous system activity, providingevidence that different emotions change one’sphysiology differentially. Emotion also hasbeen shown to be differentially transformed bypeople’s differing capacity to self-generate im-agery to replace unwanted, automatically gen-erated emotions with more desirable imageryscripts (Derryberry & Reed, 1996); this findingsuggests the importance of individual differ-ences in this domain.

It is important to note that the process ofchanging emotion with emotion goes beyondideas of catharsis, completion, exposure, ex-tinction, or habituation, in that the maladap-tive feeling is not purged, nor is it simply atten-uated by the person feeling it. Rather, another

feeling is used to transform or undo it. Al-though exposure to emotion at times may behelpful to overcome affect phobia in many situ-ations in therapy, change also occurs becauseone emotion is transformed by another emo-tion rather than simply attenuated. In these in-stances emotional change occurs by the acti-vation of an incompatible, more adaptiveexperience that undoes or reverses the old re-sponse.

Clinical observation and research suggestthat emotional transformation occurs througha process of dialectical synthesis of opposingschemes. When opposing schemes are co-activated, they synthesize compatible elementsfrom the coactivated schemes to form newhigher-level schemes, just as in developmentwhen schemes for standing and falling, in atoddler, are dynamically synthesized into ahigher-level scheme for walking (Greenberg &Pascual-Leone, 1995; Pascual-Leone, 1991).Schemes of different emotional states are simi-larly synthesized to form new higher-levelstates. Thus, in therapy, maladaptive fear, oncearoused, can be transformed into security byevoking the more boundary-establishing emo-tions of adaptive anger or disgust, or the softerfeelings of compassion or forgiveness. Simi-larly, maladaptive anger can be undone byadaptive sadness. Maladaptive shame can betransformed by accessing anger at violation,self-comforting feelings, and pride and self-worth. For example, the tendency to shrinkinto the ground in shame is transformed by thethrusting-forward tendency in newly accessedanger at violation. Withdrawal emotions fromone side of the brain are replaced with ap-proach emotions from another part of thebrain, or vice versa (Davidson, 2000). Once thealternative emotion has been accessed, it trans-forms or undoes the original state, and a newstate is forged.

Given the importance of accessing new, moreadaptive emotions to transform old maladaptiveemotions, this question arises: “How then arenew emotions accessed?” How does a therapisthelp people in the midst of maladaptive experi-ences access emotions that will help them trans-form their maladaptive feelings and beliefs?Some different ways are listed below(Greenberg,2002;Greenberg&Watson,2006).

1. The therapeutic relationship. A goodtherapeutic relationship provides an ongoingsource of new emotions by providing a secureempathic environment that soothes and calms.


2. Shifting attention. Shifting people’s focusof attention so that they pay attention to abackground feeling is a key method of helpingthem change their emotional states. On theedge of awareness or in the background, be-hind their current dominant emotion, often liesa subdominant emotion that can be found if at-tended to or searched for. Another feeling isthere, but not yet in focal awareness. Behindanger is sadness, love, or forgiveness; at theedge of sadness is anger; within hurt or fear isanger; behind shame are pride and self-esteem.The subdominant emotion is often present inthe room nonverbally, in tone of voice or man-ner of expression.

3. Accessing needs/goals. A more process-directive way in which a therapist can help cli-ents access their healthy healing emotions andinternal resources is by asking them when theyare in a maladaptive state, such as shame, whatthey need to resolve their pain. People usuallyknow what they need when they are sufferingtheir pain. Once they know what they need in asituation, they often begin to feel as if they havesome control over it. Raising a need or a goalto a self-organizing system has a number of ef-fects. At the conscious, intentional level, itopens a problem space to search for a solution.At an affective level, it conjures up a feeling ofwhat it is like to reach the goal, and opens upneural pathways to both the feeling and thegoal. Most important, raising a need that is un-met helps a new feeling such as anger or sad-ness to emerge.

4. Positive imagery. Another way to activatealternative feelings is to use imagery. Imagina-tion is a means of bringing about an emotionalresponse. This involves helping clients use theirconscious capacities to generate new experi-ence. People can use their imaginations to cre-ate scenes that they know will help them feel anemotion, and can use this emotion as an anti-dote to a maladaptive feeling they want tochange. They can change what they feel, not bychanging feelings with reason, but by usingimagination to evoke new emotions. Withpractice, people can learn how to generate op-posing emotions through imagery and use theseas antidotes to negative emotions.

5. Expressive enactment of the emotion.Yet another way to access alternate emotionsis to have people enact a feeling that is notcurrently being experienced. This goes backto William James’s idea that we feel afraidbecause we run. The therapist asks clients toadopt certain emotional stances, and helps

them to deliberately assume the expressiveposture of that feeling and then to intensifyit. Thus a therapist might use a psychodra-matic enactment and instruct a client, “Trytelling him, ‘I’m angry.’ Say it again—yes,louder. Can you put your feet on the floorand sit up straight? Yes, do it some more.”Here the therapist coaches the person in ex-pressing an emotion until the emotion actu-ally begins to be experienced. This is notencouraging phony expression, but trying tofacilitate access to a suppressed, disallowedexperience. Instructions to take on expressivepostures are always balanced by asking peo-ple what they experience after doing this.

6. Remembering another emotion. Remem-bering a situation in which an emotion oc-curred can bring the emotion memory alive inthe present. This technique is related to the im-aging process described above. Rememberingpast emotional scenes clearly produces emo-tion. To help people change what they feel, atherapist has to help them access and restruc-ture their emotion memories. One importantway of changing emotion memories involvesaccessing the emotion memory to be changedand then transforming it with another emotionmemory. Once another emotion memory isevoked, either the new memory dominates andthe old one recedes into the background andbecomes less accessible, or the new one eventu-ally transforms the old memory.

7. Cognitive change. A therapist can alsohelp people access a new emotion by talkingwith them about the more desirable emotion.This is using cognitive meaning to generatenew feelings. Talking about an emotional epi-sode helps people reexperience the feelings theyhad in that episode.

8. Expressing the emotion for the client. Atherapist might express outrage or sadness fora client that the client is not yet able to express.This gives the client permission to begin experi-encing this emotion. We all often see ourselvesin the reflections of ourselves that we get backfrom others. Seeing that our stories have an im-pact on others and that they are moved canalso move us.

9. Using the therapy relationship to gener-ate a new emotion. A new emotion can beevoked in response to a new interaction with atherapist. The therapist can evoke a particularemotion in a client by taking a particular posi-tion in the interaction. For example, a therapistwho is comforting will evoke soft feelings; onewho is confrontive will evoke anger.


EMOTION COACHING

A view of the therapist as an “emotioncoach”—a view that encompasses the impor-tance of both the therapeutic relationship andemotional processing—has been proposed as amodel of the therapist’s role and functionin working with emotion. Emotion coaching(Gottman, 1997; Greenberg, 2002) is essen-tially aimed at helping people become aware ofand make sense of their emotional experience.The effects of a good therapeutic relationshipon outcome is widely recognized (Norcross,2002), and there is good reason to believe thata good alliance is also a prerequisite to produc-tive emotional processing, as noted throughoutthis chapter. An accepting, empathic relationalenvironment provides people with the experi-ence of emotional soothing and support theyneed to pay attention to their bodily felt experi-ence. This type of relational environment helpspeople to sort out their feelings, develop self-empathy, and find alternative inner resourcesfrom which new responses can be constructed.Within this relational context, emotion coach-ing aims to help clients be informed by theiremotions, regulate them, transform them, anduse them intelligently to solve problems in liv-ing.

In addition to following where the client ismoment by moment, the therapist acting as anemotion coach guides the client in new ways ofprocessing experiential information. Emotioncoaching thus involves a style that combinesleading and following, and embodies the ideathat it is possible to influence the constructionof people’s subjective experience. Change andnovelty can be introduced in the emotional do-main by training people to become aware oftheir emotional processes, and by guiding theirattention and meaning construction processes.

Based on a major principle that one cannotleave an emotional place until one has arrived,two phases of emotion coaching—the “arriv-ing” phase and the “leaving” phase—havebeen proposed. Each phase includes four stepsdesigned to help people experience their emo-tions more skillfully. The first phase, focusedon awareness of emotion, is designed to helppeople arrive at what they feel and involves thefollowing steps:

1. The coach helps people attend to their emo-tions.

2. The coach encourages people to welcometheir emotional experience and allow it (this

does not necessarily mean they must ex-press everything they feel to other people;rather, they must acknowledge it them-selves). People also need to be coached inskills of emotion regulation if these areneeded.

3. The coach helps people to describe theirfeelings in words, in order to aid them insolving problems.

4. The coach guides people to become awareof whether their emotional reactions aretheir primary feelings in this situation. Ifnot, they need help in discovering whattheir primary feelings are.

The second phase focuses on emotion utiliza-tion or transformation and is designed to helpclients leave the place where they have arrived.This stage involves moving on and transform-ing core feelings when necessary. It is here thatthe coaching aspect is more central.

5. When a person has been helped to experi-ence a primary emotion, the coach and per-son together need to evaluate whether theemotion is a healthy, adaptive emotion oran unhealthy, maladaptive response to thecurrent situation. If it is healthy, it shouldbe used as a guide to action. If it is un-healthy, it needs to be changed.

6. If the accessed primary emotion is un-healthy, the person has to be helped to iden-tify the negative cognition associated withthis emotion.

7. Alternative, adaptive emotional responsesand needs are now processed and devel-oped.

8. People are coached to challenge the destruc-tive thoughts in their maladaptive emotionsfrom their new inner voice, based on theiradaptive primary emotions and needs, andto regulate maladaptive emotions whennecessary.

Coaching in the emotional domain thus in-volves helping clients verbally label each emo-tion they are feeling, accept the emotion, talkabout what it is like to experience the emotion,develop new ways of processing the emotion,and learn ways of soothing or regulating theemotion. It is important to note that people of-ten cannot simply be explicitly taught newstrategies for dealing with difficult emotions;they have to be assisted to engage in the newprocess experientially, and only later explicitlytaught what to do. For example, accessing a


need or goal may be very helpful in overcominga sense of passivity or defeat or to help a pain-ful feeling. However, explicitly teaching peoplethat this is what they should do is not nearly ashelpful as interpersonally facilitating this byasking them at the right time what it is theyneed. For example, it is through experiencing aprocess of shifting from negative to more posi-tive emotional states that the experiential linksbetween states are best forged. This then is con-solidated only later by explicit knowledge ofthe process.

CONCLUSION

EFT emphasizes becoming aware of and re-flecting on primary emotions, as well as regu-lating and transforming maladaptive emotions.Both the utilization and the transformation ofemotion are seen as therapeutic. A two-steptherapeutic process is recommended when thecore emotion accessed is adaptive. First, thesymptomatic secondary emotion (such as feel-ing upset, despairing, or hopeless) is evoked intherapy; then the core primary adaptive emo-tion that is being interrupted (such as sadness,grief, or empowering anger) is accessed andvalidated and utilized to promote adaptive ac-tion. A three-step sequence is required to trans-form a core maladaptive emotion. In this se-quence, the secondary emotion is first evoked;then the core maladaptive emotion beingavoided (such as shame, fear, or anger) is ac-cessed. This latter emotion is then transformedby accessing a more adaptive emotion (such asanger, sadness, or compassion). When adaptiveemotions finally are evoked, they are incorpo-rated into new views of self and used to trans-form personal narratives.

REFERENCES

Beutler, L. E., Clarkin, J. F., & Bongar, B. (2000). Guide-lines for the systematic treatment of the depressed pa-tient. New York: Oxford University Press.

Beutler, L., Engle, D., Mohr, D., Daldrup, R., Bergan,M., & Merry, W. (1991). Predictors of differential re-sponse to cognitive, experiential, and self-directedpsychotherapeutic procedures. Journal of Consultingand Clinical Psychology, 59, 333–340.

Bohart, A. C., & Greenberg, L. S. (Eds.). (1997). Empa-thy reconsidered: New directions in psychotherapy.Washington, DC: American Psychological Associa-tion.

Bonanno, G. A., & Keltner, D. (1997). Facial expres-

sions of emotion and the course of conjugal bereave-ment. Journal of Abnormal Psychology, 106, 126–137.

Borkovec, T. D., Alcaine, O., & Behar, E. (2004). Avoid-ance theory of worry and generalized anxiety disor-der. In R. G. Heimberg, C. L. Turk, & D. S. Mennin(Eds.), Generalized anxiety disorder: Advances in re-search and practice (pp. 77–108). New York:Guilford Press.

Castonguay, L. G., Goldfried, M. R., Wiser, S., Raue, P.J., & Hayes, A. M. (1996). Predicting the effect ofcognitive therapy for depression: A study of uniqueand common factors. Journal of Consulting andClinical Psychology, 64, 497–504.

Coombs, M. M., Coleman, D., & Jones, E. E. (2002).Working with feelings: The importance of emotion inboth cognitive-behavioral and interpersonal therapyin the NIMH Treatment of Depression CollaborativeResearch Program. Psychotherapy: Theory, Re-search, Practice, Training, 39, 233–244.

Darwin, C. (1998). Expression of the emotions in manand animals (3rd ed.). London: HarperCollins.

Davidson, R. J. (2000). Affective style, psychopatholo-gy, and resilience: Brain mechanisms and plasticity.American Psychologist, 5(11), 1193–1196.

Derryberry, D., & Reed, M. A. (1996). Regulatory pro-cesses and the development of cognitive representa-tions. Development and Psychopathology, 8, 215–234.

Elliott, R., Greenberg, L., & Lietaer, G. (2004). Re-search on experiential psychotherapy. In M. Lambert(Ed.), Bergin and Garfield’s handbook of psychother-apy and behavior change (pp. 493–539). New York:Wiley.

Foa, E. B., & Jaycox, L. H. (1999). Cognitive-behavioral theory and treatment of posttraumaticstress disorder. In D. Spiegel (Ed.), Efficacy and cost-effectiveness of psychotherapy (pp. 23–61). Washing-ton, DC: American Psychiatric Press.

Foa, E. B., Zoellner, L. A., Feeny, N. C., Hembree, E. A.,& Alvarez-Conrad, J. (2002). Does imaginal expo-sure exacerbate PTSD symptoms? Journal of Con-sulting and Clinical Psychology, 70, 1022–1028.

Fosha, D. (2000). The transforming power of affect: Amodel of accelerated change. New York: BasicBooks.

Frank, J. (1961). Persuasion and healing. Baltimore:Johns Hopkins University Press.

Frederickson, B. (2001). The role of positive emotionsin positive psychology: The broaden-and-build the-ory of positive emotions. American Psychologist,56(3), 218–226.

Frederickson, B., Mancuso, R., Branigan, C., & Tugade,M. (2000). The undoing effects of positive emotion.Motivation and Emotion, 24, 237–258.

Freud, S. (1957). Instincts and their vicissitudes. In J.Strachey (Ed. & Trans.), The standard edition of thecomplete psychological works of Sigmund Freud(Vol. 14, pp. 109–140). London: Hogarth Press.(Original work published 1915)

Goldman, R., Greenberg, L. S., & Angus, L. (2006).


The effects of adding emotion-focused interventionsto the therapeutic relationship in the treatment of de-pression. Psychotherapy Research, 16, 537–549.

Goldman, R., Greenberg, L., & Pos, A. (2005). Depthof emotional experience and outcome. Psychothera-py Research, 15, 248–260.

Gottman, J. (1997). The heart of parenting: How toraise an emotionally intelligent child. New York: Si-mon & Schuster.

Greenberg, L. S. (2002). Emotion-focused therapy:Coaching clients to work through their feelings.Washington, DC: American Psychological Associa-tion.

Greenberg, L. S., & Angus, L. (2004). The contributionsof emotion processes to narrative change in psycho-therapy: A dialectical constructivist approach. In L.Angus & J. McLeod (Eds.), Handbook of narrativepsychotherapy (pp. 331–352). Thousand Oaks, CA:Sage.

Greenberg, L. S., Auszra, L., & Herrmann, I. (2007).The relationship between emotional productivity,emotional arousal and outcome in experiential ther-apy of depression. Psychotherapy Research, 2, 57–66.

Greenberg, L. S., & Johnson, S. M. (1988). Emotionallyfocused therapy for couples. New York: GuilfordPress.

Greenberg, L. S., & Pascual-Leone, A. (2006). Emotionin psychotherapy: A practice-friendly research re-view. In Session: Psychotherapy in Practice, 62, 611–630.

Greenberg, L. S., & Pascual-Leone, J. (1995). A dialecti-cal constructivist approach to experiential change.In R. A. Neimeyer & M. Mahoney (Eds.), Con-structivism in psychotherapy (pp. 169–191). Wash-ington, DC: American Psychological Association.

Greenberg, L. S., & Pascual-Leone, J. (2001). A dialecti-cal constructivist view of the creation of personalmeaning. Journal of Constructivist Psychology,14(3), 165–186.

Greenberg, L. S., & Watson, J. (1998). Experientialtherapy of depression: Differential effects of client-centered relationship conditions and process experi-ential interventions. Psychotherapy Research, 8,210–224.

Greenberg, L. S., & Watson, J. C. (2006). Emotion-focused therapy for depression. Washington, DC:American Psychological Association.

Gross, J. J. (1999). Emotion and emotion regulation. InL. A. Pervin & O. P. John (Eds.), Handbook of per-sonality: Theory and research (2nd ed., pp. 525–552). New York: Guilford Press.

Hunt, M. G. (1998). The only way out is through: Emo-tional processing and recovery after a depressing lifeevent. Behaviour Research and Therapy, 36, 361–384.

Isen, A. (1999). Positive affect. In T. Dalgleish & M.Power (Eds.), Handbook of cognition and emotion.Chichester, UK: Wiley.

Iwakabe, S., Rogan, K., & Stalikas, A. (2000). The rela-

tionship between client emotional expressions,therapist interventions, and the working alliance: Anexploration of eight emotional expression events.Journal of Psychotherapy Integration, 10, 375–402.

Jaycox, L. H., Foa, E. B., & Morral, A. R. (1998). Influ-ence of emotional engagement and habituation onexposure therapy for PTSD. Journal of Consultingand Clinical Psychology, 66, 185–192.

Johnson, S. (1996). The practice of emotionally focusedcouples therapy: Creating connections. New York:Brunner-Routledge.

Johnson, S., Hunsley, J., Greenberg, L., & Schlindler, D.(1999). Emotionally focused couples therapy: Statusand challenges. Clinical Psychology: Science andPractice, 6, 67–79.

Kabat-Zinn, J., Lipworth, L., Burney, R., & Sellers, W.(1986). Four year follow-up of a meditation-basedprogram for the self-regulation of chronic pain:Treatment outcomes and compliance. Clinical Jour-nal of Pain, 2, 159–173.

Kabat-Zinn, J., Massion, A. O., Kristeller, J., Peterson,L. G., Fletcher, K. E., Pbert, L., et al. (1992). Effec-tiveness of a meditation-based stress reduction pro-gram in the treatment of anxiety disorders. AmericanJournal of Psychiatry, 149, 936–943.

Klein, M. H., Mathieu-Coughlan, P., & Kiesler, D. J.(1986). The Experiencing Scales. In L. S. Greenberg& W. M. Pinsof (Eds.), The psychotherapeutic pro-cess: A research handbook (pp. 21–71). New York:Guilford Press.

Lane, R. D., & Schwartz, G. E. (1992). Levels of emo-tional awareness: Implications for psychotherapeuticintegration. Journal of Psychotherapy Integration, 2,1–18.

Leahy, R. L. (2002). A model of emotional schemas.Cognitive and Behavioral Practice, 9, 177–191.

Levenson, R. W. (1992). Autonomic nervous system dif-ferences among emotions. Psychological Science, 3,23–27.

Linehan, M. M. (1993). Cognitive-behavioral treatmentof borderline personality disorder. New York:Guilford Press.

Linehan, M. M., Dimeff, L. A., Reynolds, S. K.,Comtois, K. A., Shaw Welch, S., Heagerty, P., et al.(2002). Dialectical behavior therapy versus compre-hensive validation plus 12 step for the treatment ofopioid dependent women meeting criteria for border-line personality disorder. Drug and Alcohol Depend-ence, 67, 13–26.

Littrell, J. (1998). Is the reexperience of painful emotiontherapeutic? Clinical Psychology Review, 18, 71–102.

Mergenthaler, E. (1996). Emotion–abstraction patternsin verbatim protocols: A new way of describing psy-chotherapeutic processes. Journal of Consulting andClinical Psychology, 64, 1306–1315.

Missirlian, T., Toukmanian, S., Warwar, S., &Greenberg, L. (2005). Emotional arousal, client per-petual processing, and the working alliance in experi-ential psychotherapy for depression. Journal of Con-sulting and Clinical Psychology, 73(5), 861–871.


Neimeyer, R. A., & Mahoney, M. (1995).Constructivism in psychotherapy. Washington, DC:American Psychological Association.

Norcross, J. C. (2002). Empirically supported therapyrelationships. In J. C. Norcross (Ed.), Psychotherapyrelationships that work: Therapist contributions andresponsiveness to patients (pp. 3–16). London: Ox-ford University Press.

Paivio, S. C., Hall, I. E., Holowaty, K. A. M., Jellis, &Tran, (2001). Imaginal confrontation for resolvingchild abuse issues. Psychotherapy Research, 11, 56–68.

Paivio, S. C., & Nieuwenhuis, J. A. (2001). Efficacy ofemotionally focused therapy for adult survivors ofchild abuse: A preliminary study. Journal of Trau-matic Stress, 14, 115–134.

Pascual-Leone, J. (1991). Emotions, development andpsychotherapy: A dialectical constructivist perspec-tive. In J. Safran & L. S. Greenberg (Eds.), Emotion,psychotherapy, and change (pp. 302–335). NewYork: Guilford Press.

Pennebaker, J. W. (Ed.). (1995). Emotion, disclosure,and health. Washington, DC: American Psychologi-cal Association.

Perls, F. S. (1969). Gestalt therapy verbatim. Lafayette,CA: Real People Press.

Pierce, R. A., Nichols, M. P., & DuBrin, J. R. (1983).Emotional expression in psychotherapy. New York:Gardner Press.

Pos, A. E., Greenberg, L. S., Goldman, R. N., &Korman, L. M. (2003). Emotional processing duringexperiential treatment of depression. Journal of Con-sulting and Clinical Psychology, 71, 1007–1016.

Rogers, C. R. (1951). Client-centered therapy. Boston:Houghton Mifflin.

Salovey, P., Mayer, J. D., Golman, S. L., Turvey, C., &Palfai, T. P. (1995). Emotional attention, clarity, andrepair: Exploring emotional intelligence using theTrait Meta-Mood Scale. In J. W. Pennebaker (Ed.),Emotion, disclosure, and health (pp. 125–154).Washington, DC: American Psychological Associa-tion.

Samoilov, A., & Goldfried, M. (2000). Role of emotionin cognitive behavior therapy. Clinical Psychology:Science and Practice, 7, 373–385.

Shapiro, F. (1999). Eye-movement desensitization andreprocessing (EMDR) and the anxiety disorders:Clinical and research implications of an integratedpsychotherapy treatment. Journal of Anxiety Disor-ders, 13, 35–67.

Silberschatz, G., Fretter, P. B., & Curtis, J. T. (1986).How do interpretations influence the process of psy-chotherapy? Journal of Consulting and Clinical Psy-chology, 54, 646–652.

Spinoza, B. (1967). Ethics (Part IV). New York: Hafner.(Original work published 1677)

Sroufe, L. A. (1996). Emotional development: The orga-nization of emotional life in the early years. NewYork: Cambridge University Press.

Stalikas, A., & Fitzpatrick, M. (1995). Client good mo-ments: An intensive analysis of a single session. Ca-nadian Journal of Counselling, 29, 160–175.

Stanton, A., Danoff-Burg, S., Cameron, C., Bishop, M.,Collins, C., Kirk, S. B., et al. (2000). Emotionally ex-pressive coping predicts psychological and physicaladjustment to breast cancer. Journal of Consultingand Clinical Psychology, 68, 875–882.

Svartberg, M., Stiles, T. C., & Seltzer, M. H. (2004).Randomized, controlled trial of the effectiveness ofshort-term dynamic psychotherapy and cognitivetherapy for Cluster C personality disorders. Ameri-can Journal of Psychiatry, 161, 810–817.

Teasdale, J. D., Segal, Z. V., Williams, J. M. G.,Ridgeway, V. A., Soulsby, J. M., & Lau, M. A.(2000). Prevention of relapse/recurrence in major de-pression by mindfulness-based cognitive therapy.Journal of Consulting and Clinical Psychology, 68,615–623.

van der Kolk, B. A. (1995). The body, memory, and thepsychobiology of trauma. In J. L. Alpert (Ed.). Sexualabuse recalled: Treating trauma in the era of the re-covered memory debate (pp. 29–60). Northvale, NJ:Aronson.

Van Velsor, P., & Cox, D. L. (2001). Anger as a vehiclein the treatment of women who are sexual abuse sur-vivors: Reattributing responsibility and accessingpersonal power. Professional Psychology: Researchand Practice, 32, 618–625.

Warwar, N. (2003). Relating emotional processes tooutcome in experiential psychotherapy of depression.Unpublished doctoral dissertation, York University,Toronto.

Warwar, N., Greenberg, L. S., & Perepeluk, (2003,June). Reported in-session emotional experience intherapy. Paper presented at the annual meeting of theInternational Society for Psychotherapy Research,Weimar, Germany.

Watson, J. C. (1996). The relationship between vivid de-scription, emotional arousal, and in-session resolu-tion of problematic reactions. Journal of Consultingand Clinical Psychology, 64, 459–464.

Watson, J. C., Gordon, L. B., Stermac, L., Kalogerakos,F., & Steckley, P. (2003). Comparing the effectivenessof process-experiential with cognitive-behavioralpsychotherapy in the treatment of depression. Jour-nal of Consulting and Clinical Psychology, 71, 773–781.

Whelton, W. J. (2004). Emotional processing in psycho-therapy: Evidence across therapeutic modalities.Clinical Psychology and Psychotherapy, 11, 58–71.

Whelton, W., & Greenberg, L. (2005). Emotion in self-criticism. Personality and Individual Differences, 38,1583–1595.

Winston, A., Laikin, M., Pollack, J., Samstag, L.,McCullough, L., & Muran, C. (1994). Short-termpsychotherapy of personality disorders: 2 yearfollow-up. American Journal of Psychiatry, 151,190–194.


C H A P T E R 7

Emotions, Music, and Literature

P. N. JOHNSON-LAIRD and KEITH OATLEY

Most human activities have emotional conse-quences. And what we humans do in our lei-sure is often designed to elicit emotions. Theenjoyment of works of art is no exception.Novels, plays, and movies can all prompt realemotions about unreal events. The mechanismswe use to understand fiction are the same asthose we use to understand the everyday world(Gerrig, 1993); part of the result is that we maylaugh or weep about what we know is imagi-nary. Music is still more mysterious. Whyshould a piece of pure music—Beethoven’s latepiano sonata, the “Hammerklavier,” say—have any emotional impact on us? It can moveus even though it refers to nothing at all. If wefeel sad as a result of listening to a piece of mu-sic, then the music isn’t the object of our emo-tion, but its cause. Usually we cannot say whythis piece of music makes us happy, whereasthat piece makes us sad. Philosophers havestruggled with these problems, but they havereached no consensus (see, e.g., Budd, 1985;Robinson, 1997; Nussbaum, 2001). Psycholo-gists and others have addressed emotions andthe arts (see, e.g., Hjort & Laver, 1997). But no

one appears to have a complete explanation ofthe causal link between the perception of awork of art and the ensuing emotional experi-ence. Our goal in the present chapter is both toreview the psychology of emotions in relationto music and literature, as befits a chapter in ahandbook, and at the same time to take somesteps toward a solution to this mystery.

Music is older than writing: The earliest mu-sical instrument, a flute found in Slovenia, ismore than 43,000 years old (Huron, 2003),whereas the invention of writing is only about5,000 years old (Coulmas, 1996). Painting andsculpture started with the creation of objectsthat are models or metaphors for somethingelse—marks on a cave wall that depict a bison.They survive from 30,000 years ago (Mithen,1996), about the same time as the start ofburial practices, and presumably of mythsabout the afterlife. Coming at the beginningand end of this sequence of emergence, musicand literature are our topics.

Pure music is without propositional content.By “pure” music, we mean music that is not asetting for words—that does not follow a

102

program such as Richard Strauss’s TilEulenspiegel’s Merry Pranks, or that does notdepict a series of events such as Claude De-bussy’s La Mer, with its three movements enti-tled, “From dawn till noon on the ocean,”“Play of the waves,” and “Dialogue of windand sea.”

In contrast, literature does have proposi-tional content. It relies on understanding thepropositions that the writer expresses. If youcannot understand “It is a truth universally ac-knowledged, that a single man in possession ofa good fortune, must be in want of a wife,”then you will not make much of Jane Austen’snovel Pride and Prejudice (1813/1906), ofwhich it is the opening sentence. The appeal ofAusten’s novel occurs because many of us aremoved by the love that develops between Eliza-beth Bennet and Mr. Darcy. The propositionalcontent of the novel, and the ways in whichAusten expresses it, lead us to experience a se-quence of emotional states, even though weknow that the characters and their actionsare fictitious. Other arts—plays, movies,operas, and representational paintings andsculptures—can also convey propositional con-tents that have an emotional impact on us. Be-fore children learn to read, these sorts of artcan affect them. Hence written literature is anideal case study, because it depends more thanany other art on the mental work that we putinto its interpretation.

Our plan for this chapter is to begin with arepresentative cognitive theory of emotions,which makes the contrasts that we need tosolve our puzzle. We then consider the natureof music, and as we review the area, we pro-pose a theory of how music evokes emotions inlisteners. We next describe relevant aspects ofliterature, and extend our theory to cope withthe effects of propositional content on emo-tions. Finally, we draw some conclusions aboutrelations between the arts in general and hu-man emotions.

A COMMUNICATIVE THEORYOF EMOTIONS

The vital problems that social mammals, suchas human beings, must solve are those createdin their internal environment, such as fatigueand pain; those created in their physical envi-ronment, such as the need for shelter; and thosecreated in their social environment, such as the

desire for a mate and for offspring. Accordingto a “communicative” theory of emotions(Oatley & Johnson-Laird, 1987, 1996), emo-tions are communications, in which a small setof signals conveys an individual’s emotionalstates to others. Although the idea of basicemotions and their expressions remains some-what controversial (see, e.g., Oatley, Keltner,& Jenkins, 2006; Russell, Bachorowski, &Fernandez-Dols, 2003) it is scarcely controver-sial that a smile is recognized in all cultures asa signal of happiness and social welcome,whereas weeping is a signal of sadness that of-ten elicits sympathy. Such nonverbal signals arehuman universals (see, e.g., Keltner, Ekman,Gonzaga, & Beer, 2003) although the cognitiveappraisals that elicit them may differ amongcultures (Johnson-Laird & Oatley, 2000). Theway in which such signals are recognized isvery different from the interpretation of propo-sitions expressed in language. Human facial ex-pressions, gestures, and vocal tone can all con-tribute to distinctive signals of emotions, akinto signals of enjoyment, alarm, threat, or sub-mission made by other species of social mam-mals. Their interpretation is carried out in spe-cialized regions of the brain. In contrast, themeaning of a sentence is composed from themeanings of its parts according to the gram-matical relations among them. It depends onworking memory, because grammatical struc-tures are often recursive (Fitch, Hauser, &Chomsky, 2005), and because the interpreta-tion of discourse calls for readers to determinethe referents of expressions.

Many messages that travel through the braincarry specific information of a propositionalnature. The communicative theory, however,postulates an evolutionarily older, and cruder,form of internal communication by means ofsimple signals that are the internal equivalentof the nonverbal expressions we have just dis-cussed. They do not require working memoryfor their interpretation, because their meaningsare not composed grammatically from themeanings of their parts. One set of these signalsconcerns bodily feelings such as hunger andthirst, which arise from the monitoring of theinternal environment. Another set of signalsconcerns basic emotions, which direct atten-tion, mobilize innate bodily resources, and pre-pare appropriate suites of behaviors. Theydon’t carry propositional information, but setthe brain into specific states to coordinate ourmultiple goals and plans, given the constraints

Emotions, Music, and Literature 103

of time pressure and of our limited intellectualresources.

An emotional signal begins with an apprais-al, which may or may not be conscious. But,the transition from the appraisal to the emo-tion is always unconscious. We cannot switchthe emotion on or off. As Aristotle (1984,Nichomachean Ethics, line 1106a3) wrote, wecannot choose to feel anger or fear. Hence ouremotions are a primitive sort of unconsciousreasoning that manifests itself not in proposi-tions, but in simple signals. Typically, we areaware of both the signal and of the appraisal.Introspection can tell us that we are angry be-cause someone insulted us, but it cannot revealthe transition from the appraisal (that an insulthas occurred) to the emotion.

The communicative theory postulates basicemotions and complex emotions (Oatley &Johnson-Laird, 1987, 1996). Basic emotionsare innate and have their own distinctive sig-nals in the brain (see, e.g., Panksepp, 1998,2005) and in universal nonverbal expressions.They include happiness, sadness, anger, andfear. Basic emotions can arise as a result of ru-dimentary appraisals, and they can be ex-perienced for no known reason (Oatley &Johnson-Laird, 1996). Basic emotions are thebiological foundation of the complex emotionsthat appear to be unique to humans. Complexemotions depend on conscious appraisals thatrelate to our models of ourselves, and often tocomparisons between alternative possibilitiesor between actual events and possibilities thatwe imagine in alternative histories. They there-fore can be experienced only for known rea-sons. They include such emotions as empathy,jealousy, pride, and embarrassment. We feelempathy when we imagine ourselves in some-one else’s position and feel that person’s sad-ness. Thus complex emotions integrate a basicemotional signal and a conscious cognitiveappraisal. Complex emotions appear to dependon a region in the prefrontal lobes of the brain.If this region is damaged, individuals suffer im-pairments in their experience of these emo-tions, and cease to be able to plan their lives orto make sensible decisions (see, e.g., Damasio,1994).

MUSIC AND EMOTIONS

Music appears to be universal to all cultures.Some evolutionary psychologists have arguedthat it serves no useful purpose (e.g., Pinker,

1997); others have suggested to the contrarythat it may serve an adaptive role in sexual se-lection (e.g., Miller, 2000)—an idea that goesback to Darwin (1872/1965). The issue doesnot appear amenable to empirical testing. Nev-ertheless, music does depend on some innatepredispositions. For example, 2-day-old hear-ing infants of congenitally deaf parents prefersinging that is intended for infants, which ismore emotional, than singing that is intendedfor adults (Masataka, 1999). Likewise, infantsrecognize melodies that are transposed to anew key or played at a different tempo(Trehub, 2003). At 2 months of age, they preferconsonance to dissonance (Trainor, Tsang, &Cheung, 2002); at 4 months, they prefer a con-sonant melody to a dissonant one (Zentner &Kagan, 1998). So why is music universal?What is its appeal?

Most of us respond to this question with theanswer that music stirs our emotions. But noteveryone agrees. The great 19th-century musiccritic Eduard Hanslick argued that the appreci-ation of music does not depend on the emo-tions that it creates. He wrote (Hanslick, 1854/1957, p. 11): “An art aims, above all, at pro-ducing something beautiful which affects notour feelings but the organ of pure contempla-tion, our imagination” (italics Hanslick’s). Andthe great 20th-century composer Igor Stravin-sky (1936, p. 91) wrote: “I consider that musicis, by its very nature, powerless to express any-thing at all, whether a feeling, an attitude ofmind, a psychological mood, etc.” A commonargument in defense of this position is thatemotions are about something, but pure musicdoes not have any propositional content thatenables it to be about anything (see, e.g.,Nussbaum, 2001, and Sloboda & Juslin, 2001,for discussion). According to the communica-tive theory, however, basic emotions can be ex-perienced in the absence of propositional con-tent. They needn’t be about anything. Indeed,from the 1890s onward, many experimentshave shown that individuals are emotion-ally moved by music (Gilman, 1891, 1892;Downey, 1897), that it is a reliable way to in-duce moods, and that it elicits activity in re-gions of the brain known to mediate emotions.(For reviews of these three topics, see, re-spectively, Gabrielsson & Lindström, 2001;Västfjäll, 2002; and Peretz & Zatorre, 2003.)

A subtle distinction occurs between the emo-tions that listeners experience from listening toa piece of music and their judgments aboutwhat emotion the music expresses. Listeners


can tell that a piece is intended to convey hap-piness, though they themselves are irritated byit. The same distinction occurs for readersof imaginative literature. As Oscar Wilde re-marked about Charles Dickens’s The Old Curi-osity Shop, “One would have to have a heart ofstone to read the death of Little Nell withoutlaughing.” Granted this distinction, studiesthat evoke emotions by way of music are es-sentially over when the experimenter hasselected the materials. That is, their selec-tion shows that at least one individual—theexperimenter—is affected by the music.

Anecdotal evidence also supports the ideathat music evokes emotions. In the 46th minuteof Alfred Hitchcock’s 47th movie, Psycho, thedirector did something shocking: He killed offhis leading lady, Janet Leigh. He had plannedto have no music during her murder in theshower, just sounds of the shower and herstruggle with her assailant. Unbeknownst tohim, the composer of the film’s music, BernardHerrmann, had put together a sequence ofscary music—high-pitched shrieks on the vio-lins, which anyone who has seen the movie islikely to recall. Hitchcock viewed both versionsof the movie, one with natural sounds and theother with Herrmann’s music, and declaredthat he had been wrong. The montage wasmuch more frightening with the music (see alsoOatley, 2004).

The standard answer is therefore correct:Music moves us. But it raises a mystery: Howdoes it do so? It also raises another mystery:Why do we like to have our emotions stirred bymusic and other forms of art? We have only alittle to say about this second mystery in ourconclusions, and so we turn to the fundamen-tals of music pertinent to emotion.

Music is a social activity in most cultures. Inthe West, the single most popular piece of musicis “Happy Birthday to You,” which serves a so-cial function. Readers may be surprised to learnthat it is not a folk tune, but was composed byMildred and Patti Hill and published in 1893.Most music in Western culture consists of a mel-ody and an accompaniment, where the accom-paniment is a simultaneous performance ofother pitches in the form of chords.

A “melody” is a rhythmical sequence ofpitches in a metrical framework. “Meter” is aregular pulse that provides the framework forrhythm (see, e.g., Lerdahl & Jackendoff, 1983;Longuet-Higgins & Lee, 1984; Johnson-Laird,2002). It makes prediction easier, especially formusicians who have to synchronize their per-

formances. But meter is more than the numberof beats in the measure: Beats and their subdi-visions can be grouped in different ways. Agood example is the contrast between twobeats to the measure that are each subdividedinto three, and waltz time, with three beats tothe measure each subdivided into two. Onemeasure in these two meters contains the samenumber of units, but their structures differ:One has two main pulses, and the other hasthree main pulses. Even individuals who arenot musicians perceive meter, as shown in anunpublished experiment that Johnson-Lairdcarried out with Jung-Min Lee and MalcolmBauer. The experimenter counts, “1, 2, 3, 4,” ina regular way to establish a meter, and thenclaps the following rhythm in the same tempo:

� � � � ��

Listeners judge all four notes to be of the sameduration. If, instead, the experimenter claps:

� � � � � ��

then the listeners judge the last note to beshorter than the others. Of course, the claps areall of the same brief duration. So why is the lastnote in the first case judged to be longer thanthe last note in the second case? The answermust be that listeners perceive both rhythms ashaving a meter of four beats to the measure,and that they tacitly infer that there will be anote on the first beat of the next measure. Theinterval from the onset of the last clap to theonset of this imagined clap is indeed longer inthe first case than in the second. Hence listenersinfer unconsciously that the onset of the nextclap, even though it is imaginary, will occur onthe first beat of the next measure. These judg-ments demonstrate the cognitive reality of met-rical structure even for nonmusicians.

As many theorists have pointed out (e.g.,Povel, 1984), the critical feature of a rhythm isthe sequence of onsets of its notes. Hence if youclap the rhythm of a familiar piece, then listen-ers will be able to identify it. Clapping, ofcourse, provides information only about onsettimes. If you play the sequence of pitches in amelody with each pitch having the same dura-tion, it is hard for listeners to identify the mel-ody. You might therefore suppose that more in-formation is conveyed by the rhythm than bythe sequence of pitches. The conclusion is un-warranted. The problem is that the equal dura-tion of the pitches produces a new, albeit uni-


form, rhythm, which masks the real rhythm ofthe melody.

A simple demonstration showing that a se-quence of pitches can identify a melody is Par-sons’s (1975) Directory of Tunes and MusicalThemes. This directory represents any melodymerely by its contour. It represents, for exam-ple, the famous opening of Beethoven’s FifthSymphony as follows:

* R R D U R R D . . .

where * denotes the first note, R a repeat of theprevious note, U an upward step, and D adownward step. As the directory shows, theseeight symbols are common to five otherthemes, including one from Gilbert andSullivan’s HMS Pinafore. But once the first 15notes of any theme in the classical repertoryhave been encoded in the notation, it is almostalways identified uniquely.

In the Western tradition, melodies are“tonal” from the earliest music that survivesthrough the great classical composers, the Ro-mantics, and most 20th-century music (includ-ing popular songs, jazz, and rock and roll).Tonal melodies are made up from a subset ofthe possible pitches in the Western scale, and inthis subset one note, the “tonic,” is more im-portant than the others. This note also gives itsname to the key of the piece. For example, amelody in the key of C major is made up fromthe following 7 notes (from the standard scaleof 12 notes): C D E F G A B. The chorus of TheBeatles’s “Yellow Submarine” starts with thefollowing sequence of pitches, which we’vetransposed to the key of C major:

| G G G G A | D D D D D D | D D D D D | C C C C C |

The vertical lines demarcate the measures. Thekey of C minor, which is slightly more disso-nant, is made up from the following notes: C DE� F G A B C. Only one note makes the key aminor one, E� (i.e., E flat).

Western music has a number of global vari-ables, whose values are normally held constantfor some length of time during a piece. Theseven most important of these variables are asfollows:

1. Tempo, which varies from slow to fast.2. Volume, which varies from soft to loud.3. Register, which may be high or low (e.g., a

melody may be a sequence of low pitches).

4. Dissonance, which includes whether the keyis major or minor.

5. Timbre, which depends on the particular in-strument that is played.

6. Range of melody, which may include onlynotes close to each other or may includelarge jumps in pitch.

7. Meter, which is the number of beats permeasure.

Our theory of how music creates emotionsis, like Aristotle’s (1984, Politics, line 1340a11et seq.), a mimetic one. Its first hypothesis isthat music in itself creates only the basic emo-tions of happiness, sadness, anxiety, and anger.Music can also manipulate our level of arousal,and thus our general level of excitement. Ofcourse, factors outside music, such as lyrics ora drama that it accompanies, can elicit complexemotions. Similarly, individuals can acquire as-sociations between certain sorts of music andemotions, such as the association betweenwaltzes and dancing, and these too can colorthe emotions that music creates. But what thepresent hypothesis rules out is that music alonecan create object-oriented emotions of love orhate, or complex emotions. Evidence corrobo-rates this account. Children of ages 4–6 areable to discriminate music expressing basicemotions (Cunningham & Sterling, 1988).Adults can recognize these emotions in music(Krumhansl, 1997). Music therapists can im-provise music to convey these emotions to oth-ers (Bunt & Pavlicevic, 2001).

Meyer (1956), in one of the earliest and best-known theories of emotion and music, arguedthat music sets up expectations, which may bedelayed in fulfillment: “Emotion or affect isaroused when a tendency to respond is arrestedor inhibited” (p. 14). The longer resolution ispostponed, the greater the affect. Hence emo-tion depends on the structure of music. This ac-count is not incompatible with our hypothesis,and it may explain how music affects arousal.But it has little to say about the different emo-tions, such as happiness or sadness, that musicevokes (Budd, 1985).

Arguments to the contrary of our hypothesisalso exist. The music critic Deryck Cooke(1959) argued that music can express attitudes;that is, it does have propositional content.Hence he allocated to the different notes of thescale complex emotions, such as “pleasurablelonging in the context of flux.” And he as-signed different emotions to various sequences


of pitches in melodies. Gabriel (1978) failed tocorroborate their occurrence in an experimen-tal investigation (cf. Sloboda, 1985, p. 63).Scherer (2004) also argues that basic emotionsare far too impoverished to do justice to ourexperiences of listening to music. One aspect ofthe perception of music, however, is that it iseasy to “project” a program onto pure music—a propensity that musicians, critics, andnonmusicians all possess (see, e.g., Downey,1897). This propensity can in turn lead us toascribe emotions other than basic ones to apiece of music. Ives (1962, p. 36), for example,does so when he describes the opening of Bee-thoven’s Fifth Symphony as “the soul of hu-manity at the door of divine mysteries, radiantin the faith that it will be opened—and the hu-man become divine.” McClary (1991, p. 128)goes a step further but in the opposite directionwhen she writes about Beethoven’s Ninth Sym-phony, “The point of recapitulation in the firstmovement in the Ninth is one of the most hor-rifying moments in music, as the carefully pre-pared cadence is frustrated, damming up en-ergy which finally explodes in the throttling,murderous rage of a rapist incapable of attain-ing release.” When listeners indulge in thispractice, they are imagining a program forwhich the music might make a suitable accom-paniment, rather than responding to the musicin itself. Pure music has no propositional con-tent, and so it cannot express anything otherthan basic emotions. One person’s program isanother person’s parody.

The second hypothesis of our theory is thatmusic creates emotions in a mimetic way. Itmimics the main characteristics of emotionalbehavior, speech, and thought (Scherer, 1986;Davies, 1994; Juslin & Laukka, 2003). It doesso, in part, with settings of global parametersthat create basic emotions. A simple illustra-tion is that when people are sad, they moveslowly and speak softly, with an intonationcontour that does not make great leaps. Whenpeople are happy, however, they move morerapidly and speak loudly, with an intonationcontour that can make great leaps for empha-sis. The theory accordingly postulates the fol-lowing sorts of settings:

Happiness—medium tempo, loud, concor-dant

Sadness—slow, muted, slightly discordantAnxiety—scurrying, low pitch, discordantAnger—fast, loud, high pitch, discordant

Bunt and Pavlicevic (2001) report a similar listof settings in the music that therapists impro-vise to convey emotions. Juslin (2001) also de-scribes such a list, which he relates to innateprograms for the vocal expression of emotions.

The third hypothesis of our theory is thatmusic creates basic emotions because they de-pend on rudimentary cognitive appraisals call-ing for only minimal computational power.“Computational power” concerns what a sys-tem can compute, and minimal computationalpower depends on a system that has only a fi-nite number of states; it makes no use of work-ing memory for the results of intermediatecomputations (Hopcroft & Ullman, 1979).One corroboration of this hypothesis is thespeed with which music conveys emotions. In-dividuals need less than a quarter of a second—a chord or a few notes of melody—to identifywhether a musical excerpt is happy or sad(Peretz, Gagnon, & Bouchard, 1998).

Another corroboration comes from a com-puter program that Johnson-Laird devised tocreate melodies. A common view among com-posers is that melodies are a result of uncon-scious inspiration. Aaron Copland (1957,p. 102) wrote: “The composer starts with histheme; and the theme is a gift from heaven. Hedoesn’t know where it comes from—has nocontrol over it. It comes almost like automaticwriting.” The process of creating a melody isindeed akin to improvisation. Beethoven was agreat improviser, and had no time for musi-cians who could not improvise:

It has always been known that the greatest piano-forte players were also the greatest composers; buthow did they play? Not like the pianists of to-day,who prance up and down the keyboard with pas-sages that they have practised—putsch, putsch,putsch; what does that mean? Nothing! (Beetho-ven, quoted in Kinderman, 2000, p. 106)

Melodies are often the results of unconsciousprocesses, and so they too are created—at leastin their initial form—using minimal computa-tional power, because unconscious processeshave no access to working memory. But the cre-ation of melodies can rely on long-term memo-ries of other melodies, and of musical structuressuch as chord sequences. A minimal-finite-statesystem that is capable of infinitely many differ-ent outputs is one based on a matrix of transi-tions. Many theorists have analyzed the transi-tion probabilities from one pitch to another in a


set of melodies (e.g., Simonton, 2001). But amelody also has a rhythm, which needs to betaken into account. One solution, which thecomputer makes possible, is to construct an ar-ray of transitions from one pitch and its dura-tion to the interval of the next pitch and its dura-tion. The program takes as input a corpus ofmelodies and constructs such an array fromthem. The user then provides a chord sequenceand specifies the number of beats in the bar. Theprogram uses the probabilities in its array oftransitions to generate a novel melody. Each callto the program generates a new melody, and thechances of its creating the same melody twiceare negligible. Johnson-Laird used the programto construct arrays of transitions from corporaof happy, sad, and anxiety-provoking melodies,and then to generate melodies from these arrays.Figure 7.1 presents examples of each sort ofmelody. Audiences to whom they were playedconcurred that they conveyed the appropriateemotion, even though the melodies exploitedonly some of the relevant global variables.

Melodies may have an improvisational qual-ity, at least when they first come to a composer’smind, but the process of composition may leadto modifications. Even though Beethoven was agreat improviser, his process of composition waslaborious; at least 8,000 pages of his notebookssurvive. Composed music has a large-scale struc-ture, and Beethoven began his career by using“models” of these structures from Haydn andMozart (Sisman, 2000). Such structures, and thesequences of chords, call for more computa-tional power than unconscious processes canmuster. They need a working memory for the re-sults of intermediate computations (seeJohnson-Laird, 2002), and musical notation can

act as a substitute for such a memory. Musicalstructure, however, is what Hanslick (1854/1957) considered as crucial for beauty. Our per-ception of this structure may in turn lead to an“esthetic” emotion (Scherer, 2004)—an emo-tion that we regard as akin to awe, a complexemotion rooted in anxiety.

LITERATURE AND EMOTIONS

Just as pure music is nonpropositional, fic-tional literature is propositional. When we readfiction, we expect to experience emotions. Thequestion is this: How can such propositionalcontent prompt these emotions?

As with music, we distinguish the estheticemotion that occurs in response to literaturefrom the emotions of involvement in a story(Oatley, 1994). The esthetic emotion dependson an appreciation of beauty in the use of lan-guage. It calls for readers to achieve an estheticdistance from the work (Cupchik, 2002) andfor them to have had a sufficient experiencewith literature to develop an appreciation of itsskillful use. Consider, for instance, the firststanza of Coleridge’s “Kubla Khan,” publishedin 1816 (Coleridge, 1977):

In Xanadu did Kubla KhanA stately pleasure dome decree:Where Alph, the sacred river, ranThrough caverns measureless to manDown to a sunless sea.

The iambic rhythm, the assonance on thesound of “ah,” the alliterations (e.g., “KublaKhan”), the rhymes (e.g., “Khan, ran, man”),


FIGURE 7.1. Three melodies (happy in the key of C major, sad in the key of C minor, and anxious in thekey of C minor) created by a computer program using a finite-state system of transitions derived fromcorpora of real melodies.

and the exoticism (“Xanadu”) may all move usin a way akin to music. However, our under-standing of the propositional content of litera-ture also elicits emotions. In “Kubla Khan,”the content is difficult. We come to understandit either from several readings and muchcontemplation, or from knowing some ofColeridge’s preoccupations, or from havingsomeone knowledgeable explain it to us (e.g.,Paglia, 2005). The poem is a metaphor for theplace of art and the artist in society. Hence thepleasure dome is a metonymic figure (a“synecdoche”) for art in general. The sacredriver is a metaphor for the stream of artisticcreativity that flows through society. The poemevokes further images as it continues: thesource of creativity in nature and sexuality (“asif this earth in fast thick pants were breath-ing”), the potential enmity between the artistand society (“ancestral voices prophesyingwar”), an image of the artist (“A damsel with adulcimer”), and the status of the artist as demi-god (“for he on honey dew hath fed / anddrunk the milk of paradise”). Some poetry,such as Mallarmé’s, makes its emotional effectalmost entirely in terms of its music rather thanits propositional content. It was Mallarmé whoreminded his painter friend, Degas, that poetrywas made from words, not ideas.

Certain novelists, notably Flaubert, have in-sisted that the arrangements of words in theirworks should be as delicate and precise asthose of poetry (see, e.g., Williams, 2004). Yet,in prose, the emphasis shifts toward a primacyof the propositional. Consider the opening lineof L. P. Hartley’s (1953) novel The Go-Between: “The past is another country: they dothings differently there.” “Another country” isa wonderful metaphor, but it does not have therhythmic and alliterative music of “A damselwith a dulcimer.” Above all in literature, prop-ositional content prompts our emotions. Theseemotions can be basic, but they can be complexemotions too, depending on appraisals of theself in relation to others. Consider jealousy. Itsprings from our suspicion that a third personmight displace us in a relationship with some-one we love. The basic emotion may be fear oranger, but jealousy depends on inferencesabout the relations among three individuals,and these inferences in turn hinge on proposi-tional content. A beautiful depiction of jeal-ousy occurs in Frank O’Connor’s short story“My Oedipus Complex” (O’Connor, 1963).The protagonist is a young boy, and the storyfollows his growing understanding of the im-

pact of his father’s return from the war and hisown displacement from the center of hismother’s affections. We enter his world andempathize with his anger and distress as he re-alizes that he has been displaced. The emo-tional effect on us is powerful and fascinating,because it resonates with emotions that wehave experienced too. But something mysteri-ous and ambiguous remains: The boy has notquite understood his own emotion.

T. S. Eliot (1919/1953) argued that artists donot describe emotions subjectively, but offer in-stead an external, objective pattern of events,which he referred to as the “objective correla-tive.” Our appraisal of these events leads us toexperience emotions. This view foreshadowsthose theories of emotions that hinge on ourcognitive appraisals. We use our knowledge ofthe language to assemble the meanings of sen-tences from the meanings of the words and thegrammatical relations among them. And, as ex-periments have shown, we use these meaningsto construct a mental model of the individualsand events to which the discourse refers (see,e.g., Johnson-Laird, 1983, Ch. 14; Garnham,2001). But how does such a model—the objec-tive correlative—suggest emotions?

We propose three hypotheses intended to an-swer this question (see Oatley, 1994). The firsthypothesis is that literature can suggest emo-tions because we identify with the protagonistand resist antagonists. The propositional con-tent of a story affects us most powerfully when,like Alice through the looking glass, we enterthrough its surface into its interior. The typicalmode of narrative, as Bruner (1986) haspointed out, concerns human plans and theirvicissitudes. The author provides us with thecontent that allows us to construct a dynamicmodel that simulates such a world, its charac-ters, and their interactions (Oatley, 1999). Andwe enter the simulation when we identify witha character in the story (see Freud, 1905–1906/1985). Flaubert is supposed to have said, “Ma-dame Bovary, c’est moi.” But we too can be-come Madame Bovary as we read the novel.We are caught up in her experience (see alsoMiall & Kuiken, 2002; Zillmann, 1994). Werun her plans on our own planning processors.When these plans meet vicissitudes, we experi-ence emotions. They occur within our simula-tion, and they are our emotions. Identificationis empathy, as shown in the emotions elicitedfrom movies (Trabasso & Chung, 2004). Butthe empathy is not quite as it is in real life, be-cause we empathize with a fiction—a nonexis-


tent individual whom we have created in oursimulation of the story (see also Lipps, 1962;Kreitler & Kreitler, 1962). Although the effectsare familiar, they remain surprising. As Hamletsays, after witnessing an actor affected by emo-tion as he plays a part, “What’s Hecuba to him,or he to Hecuba?” (Shakespeare, c. 1600/1981). It may be that actors, during rehearsalthough not necessarily during performance(Konijn, 2000), use Stanislavski’s (1936)method of drawing on their own autobiogra-phies to reexperience certain emotions.

Our second hypothesis is that literature canprompt us to feel an emotion about a character.Tan and Frijda have proposed a theory that is aversion of this idea, but intended to cover thesame ground as our first hypothesis about iden-tification with characters (Tan, 1996; Tan &Frijda, 1999). They argue that an author pro-vides appraisal patterns and that we pick upthese patterns as they apply to characters. As aresult, we may feel sympathetic emotions to-ward these characters. Tan (1996) calls these“witness emotions.” We argue that when weanalyze events in our simulation, we infer howthey would strike a character in it so that wecan feel sympathy for the character, but in ourview the process coexists with identification.Indeed, we may like or admire a protagonist, orfear or loathe a villain. When Dorothea, theheroine of George Eliot’s novel Middlemarch(1872/1991), agrees to marry the aged scholarCasaubon, she is full of enthusiasm to help himwith his work, and we share some of her enthu-siasm. At the same time, our hearts sink. We re-spond this way in part because Eliot has con-veyed Casaubon’s character to us in a way thatgoes beyond Dorothea’s understanding of him.One of the skills of great writers is to make usfeel the emotions of a character with whom weidentify, and quite different emotions towardthe same character.

This distinction was important for medievalIndian literary theorists such as Abhinavagupta(see, e.g., Ingalls, Massson, & Patwardhan,1990). They described the emotions depictedby an actor in a play, using facial expressions,gestures, tone of voice, and the content of ut-terances, and the corresponding emotions(rasas in Sanskrit) that occur in audiences (seeOatley et al., 2006, p. 112). For example, whenan actor depicts amusement or anger, the audi-ence identifies and feels amusement or anger.But when an actor depicts sorrow, the audiencefeels compassion for the character; when an ac-

tor depicts something disgusting, the audiencefeels loathing. According to these theorists,rasas are literary emotions and subtly distinctfrom the emotions of daily life. Each well-constructed work should be based on a singlerasa, which is the basis for a genre (e.g., a lovestory, a comedy, a tragedy). Within a story,other rasas may also occur, but in a supportingand transient way. Hogan (2003) has reviewedstories worldwide and reports that the twomost common genres are the love story and thestory of an angry conflict.

The third hypothesis of our theory is thatemotions can also depend on personal memo-ries (Larsen & Seilman, 1998). Events in a sim-ulation may elicit only our memory of an emo-tion, while the events themselves remainimplicit. Hence, according to Scheff (1979), wemay cry at the fate of the protagonists in Ro-meo and Juliet because we are reexperiencingearlier losses of our own, which are reacti-vated, though not necessarily consciously, bythe play’s events.

Following these hypotheses we can proposea generalization. Along with the cues in a textthat enable the reader or audience member toconstruct a simulation, any literary work hasalso what Oatley (1999) has called a “sugges-tion structure.” It is principally this structurethat prompts emotions in the ways indicated byour three hypotheses: by suggesting (1)empathetic identifications, (2) sympathies andantagonisms, and (3) scenes that might promptmemories. Although figurative language is notnecessary to this structure (see, e.g., Oatley,2004), tropes such as metaphor and metonymy(see Lodge, 1977) prompt emotions principallyby suggestion. An individual may readHartley’s (1953) metaphor, “The past is an-other country,” and think of a visit to anothercountry, Kashmir—which triggers thoughts ofbeing in a 450-year-old painting by Bruegel,with chickens walking in the road, and peoplecarrying bundles of firewood. This thoughtadds emotional potency to the metaphor. West-ern theories of poetics neglect the role of sug-gestion, but in Eastern poetics it is stressed, andthere is a Sanskrit word for it, dhvani.Abhinavagupta argued that dhvani is the heartof poetry (see Ingalls et al., 1990).

Experiments have shown that stories doelicit emotions in people who read them (Miall& Kuiken, 2002; Oatley, 2002). Furthermore,a study by Nundy (1996) has corroborated therole of personal suggestion. The participants


read a short story by Russell Banks in which aman cruelly severs his relationship with awoman. The participants experienced strongemotions as a result, but their nature differedfrom one person to another: Some readers wereangry, others were sad, and a few were dis-gusted. Another corroboration of suggestion isthat readers of a narrative piece had morememories that were personal (as comparedwith generic memories) than readers of a non-narrative piece of the same length, proposi-tional content, and reading difficulty (Mar,Oatley, & Eng, 2003).

CONCLUSIONS

Emotions arise from unconscious transitions,and so we are often puzzled by our emotionalreactions. Sometimes they are so aberrant in in-tensity and so prolonged in their effects thatthey create a psychological illness (Johnson-Laird, Mancini, & Gangemi, 2006). Art mayhelp us to understand them better. One of theprincipal theories of emotion’s relation to art isthat of Collingwood (1938): Art is the expres-sion of an emotion in a particular language—words, sculpture, paintings—so that we cometo understood the emotion better (Oatley,2003). Music is mystifying in its emotional ef-fects, but we enjoy the emotions it creates. Incontrast, literature helps us to understand therelation between propositional content andsubjective feelings—to understand the causesof an emotion, and why a particular individualin a particular circumstance feels a particularemotion. The Indian theorists said that we failto understand our emotions because a thickcrust of egotism obscures our vision. Rasas,however, allow us to see more deeply into theirnature, because of their literary context.Nussbaum (1986) similarly translates Aris-totle’s katharsis as “clarification” or “illumina-tion.” In literary fiction, it becomes possible tounderstand emotions as they occur in a widerset of circumstances than we would encounterin our ordinary lives.

REFERENCES

Aristotle. (1984). The complete works of Aristotle(Vols. 1 and 2, rev. Oxford trans., J. Barnes, Ed.).Princeton, NJ: Princeton University Press.

Austen, J. (1906). Pride and prejudice. London: Dent.(Original work published 1813)

Bruner, J. (1986). Actual minds, possible worlds. Cam-bridge, MA: Harvard University Press.

Budd, M. (1985). Music and the emotions: The philo-sophical theories. London: Routledge.

Bunt, L., & Pavlicevic, M. (2001). Music and emotion:Perspectives from music therapy. In P. N. Juslin & J.A. Sloboda (Eds.), Music and emotion: Theory andresearch (pp. 181–201). Oxford: Oxford UniversityPress.

Coleridge, S. T. (1977). The portable Coleridge.Harmondsworth, UK: Penguin.

Collingwood, R. G. (1938). The principles of art. Ox-ford: Oxford University Press.

Cooke, D. (1959). The language of music. Oxford: Ox-ford University Press.

Copland, A. (1957). What to listen for in music. NewYork: McGraw-Hill.

Coulmas, F. (1996). The Blackwell encyclopedia of writ-ing systems. Oxford, UK: Blackwell.

Cunningham, J. G., & Sterling, R. S. (1988). Develop-mental change in the understanding of affectivemeaning in music. Motivation and Emotion, 12,399–413.

Cupchik, G. C. (2002). The evolution of psychical dis-tance as an aesthetic concept. Culture and Psychol-ogy, 8, 155–187.

Damasio, A. R. (1994). Descartes’ error: Emotion, rea-son, and the human brain. New York: Putnam.

Darwin, C. (1965). The expression of the emotions inman and animals. Chicago: University of ChicagoPress. (Original work published 1872)

Davies, S. (1994). Musical meaning and expression.Ithaca, NY: Cornell University Press.

Downey, J. E. (1897). A musical experiment. AmericanJournal of Psychology, 9, 63–69.

Eliot, G. (1991). Middlemarch. New York: Knopf.(Original work published 1872)

Eliot, T. S. (1953). Hamlet. In J. Hayward (Ed.), T. S.Eliot: Selected prose (pp. 104–110). Harmonds-worth, UK: Penguin. (Original work published 1919)

Fitch, W. T., Hauser, M., & Chomsky, N. (2005). Theevolution of the language faculty: Clarifications andimplications. Cognition, 97, 179–210.

Freud, S. (1985). Psychopathic characters on the stage.In A. Dickson (Ed.), Pelican Freud library: Vol. 14:Art and literature (pp. 119–127). London: Penguin.(Original work published 1905–1906)

Gabriel, C. (1978). An experimental study of DeryckCooke’s theory of music and meaning. Psychology ofMusic, 6, 13–20.

Gabrielsson, A., & Lindström, E. (2001). The influenceof musical structure on emotional expression. In P.N. Juslin & J. A. Sloboda (Eds.), Music and emotion:Theory and research (pp. 223–248). Oxford: OxfordUniversity Press.

Garnham, A. (2001). Mental models and the representa-tion of anaphora. Hove, UK: Psychology Press.

Gerrig, R. J. (1993). Experiencing narrative worlds: On


the psychological activities of reading. New Haven,CT: Yale University Press.

Gilman, B. I. (1891). Report of an experimental test ofmusical expressiveness. American Journal of Psy-chology, 4, 558–576.

Gilman, B. I. (1892). Report of an experimental test ofmusical expressiveness (continued). American Jour-nal of Psychology, 5, 42–73.

Hanslick, E. (1957). The beautiful in music (7th ed., G.Cohen, Trans.). New York: Liberal Arts Press. (Origi-nal work published 1854; this ed. originally pub-lished 1885)

Hartley, L. P. (1953). The go-between. London: Hamil-ton.

Hjort, M., & Laver, S. (Eds.). (1997). Emotion and thearts. Oxford: Oxford University Press.

Hogan, P. C. (2003). The mind and its stories: Narrativeuniversals and human emotion. Cambridge, UK:Cambridge University Press.

Hopcroft, J. E., & Ullman, J. D. (1979). Formal lan-guages and their relation to automata. Reading, MA:Addison-Wesley.

Huron, D. (2003). Is music an evolutionary adaptation?In I. Peretz & R. Zatorre (Eds.), The cognitive neuro-science of music (pp. 57–75). Oxford: Oxford Uni-versity Press.

Ingalls, D. H., Masson, J. M., & Patwardhan, M. V.(1990). The Dhvanyaloka of Anandavardana withthe Locana of Abhinavagupta. Cambridge, MA:Harvard University Press.

Ives, C. (1962). Essays before a sonata and other writ-ings. (H. Boatwright, Ed.). New York: Norton.

Johnson-Laird, P. N. (1983). Mental models: Towards acognitive science of language, inference, and con-sciousness. Cambridge, MA: Harvard UniversityPress.

Johnson-Laird, P. N. (2002). How jazz musicians im-provise. Music Perception, 19, 415–442.

Johnson-Laird, P. N., Mancini, F., & Gangemi, A.(2006). A hyper-emotion theory of psychological ill-nesses. Psychological Review, 113, 822–841.

Johnson-Laird, P. N., & Oatley, K. (2000). The cogni-tive and social construction of emotions. In M. Lewis& J. M. Haviland-Jones (Eds.), Handbook of emo-tions (2nd ed., pp. 458–475). New York: GuilfordPress.

Juslin, P. N. (2001). Communicating emotion in musicperformance: A review and theoretical framework. InP. N. Juslin & J. A. Sloboda (Eds.), Music and emo-tion: Theory and research (pp. 309–337). Oxford:Oxford University Press.

Juslin, P. N., & Laukka, P. (2003). Communication ofemotions in vocal expression and music perfor-mance: Different channels, same code? PsychologicalBulletin, 129, 770–814.

Keltner, D., Ekman, P., Gonzaga, G. C., & Beer, J.(2003). Facial expression of emotion. In R. J.Davidson, K. R. Scherer, & H. H. Goldsmith (Eds.),Handbook of affective sciences (pp. 415–432). NewYork: Oxford University Press.

Kinderman, W. (2000). The piano music: Concertos,sonatas, variations, small forms. In G. Stanley(Ed.), The Cambridge companion to Beethoven(pp. 105–126). Cambridge, UK: Cambridge Univer-sity Press.

Konijn, E. (2000). Acting emotions: Shaping emotionson stage. Amsterdam: Amsterdam University Press.

Kreitler, H., & Kreitler, S. (1972). Psychology and thearts. Durham, NC: Duke University Press.

Krumhansl, C. L. (1997). An exploratory study of musi-cal emotions and psychophysiology. Canadian Jour-nal of Experimental Psychology, 51, 336–352.

Larsen, S. F., & Seilman, U. (1988). Personal meaningswhile reading literature. Text, 8, 411–429.

Lerdahl, F., & Jackendoff, R. (1983). A generative the-ory of tonal music. Cambridge, MA: MIT Press.

Lipps, T. (1962). Empathy, inner imitation, and sensefeeling. In M. Rader (Ed.), A modern book on esthet-ics: An anthology (3rd ed., pp. 374–382). New York:Holt, Rinehart & Winston.

Lodge, D. (1977). The modes of modern writing: Meta-phor, metonymy, and the typology of modern fiction.Ithaca, NY: Cornell University Press.

Longuet-Higgins, H. C., & Lee, C. S. (1984). The rhyth-mic interpretation of monophonic music. Music Per-ception, 1, 424–441.

Mar, R., Oatley, K., & Eng, A. (2003, August). Abstrac-tion and the vividness of details in fiction. Paper pre-sented at the annual convention of the American Psy-chological Association, Toronto.

Masataka, N. (1999). Preference for infant-directedsinging in 2-day-old hearing infants of deaf parents.Developmental Psychology, 35, 1001–1005.

McClary, S. (1991). Feminine endings: Music, gender,and sexuality. Minneapolis: University of Minneapo-lis Press.

Meyer, L. B. (1956). Emotion and meaning in music.Chicago: University of Chicago Press.

Miall, D. S., & Kuiken, D. (2002). A feeling for fiction:Becoming what we behold. Poetics, 30, 221–241.

Miller, G. (2000). Evolution of human music throughsexual selection. In N. L. Wallis, B. Merker, & S.Brown (Eds.), The origins of music (pp. 319–360).Cambridge, MA: MIT Press.

Mithen, S. (1996). The prehistory of the mind: The cog-nitive origins of art and science. London: Thames &Hudson.

Nundy, S. (1996). The effects of emotion on humaninference: Towards a computational model. Un-published doctoral dissertation, University of To-ronto.

Nussbaum, M. C. (1986). The fragility of goodness:Luck and ethics in Greek tragedy and philosophy.Cambridge, UK: Cambridge University Press.

Nussbaum, M. C. (2001). Upheavals of thought: Theintelligence of emotions. Cambridge, UK: CambridgeUniversity Press.

Oatley, K. (1994). A taxonomy of the emotions of liter-ary response and a theory of identification in fic-tional narrative. Poetics, 23, 53–74.


Oatley, K. (1999). Why fiction may be twice as true asfact: Fiction as cognitive and emotional simulation.Review of General Psychology, 3, 101–117.

Oatley, K. (2002). Emotions and the story worlds of fic-tion. In M. C. Green, J. J. Strange, & T. C. Brock(Eds.), Narrative impact: Social and cognitive foun-dations (pp. 39–69). Mahwah, NJ: Erlbaum.

Oatley, K. (2003). Creative expression and communica-tion of emotion in the visual and narrative arts. In R.J. Davidson, K. R. Scherer, & H. H. Goldsmith(Eds.), Handbook of affective sciences (pp. 481–502). New York: Oxford University Press.

Oatley, K. (2004). Scripts, transformations, and sugges-tiveness, of emotions in Shakespeare and Chekhov.Review of General Psychology, 8, 323–340.

Oatley, K., & Johnson-Laird, P. N. (1987). Towards acognitive theory of emotion. Cognition and Emotion,1, 29–50.

Oatley, K., & Johnson-Laird, P. N. (1996). The commu-nicative theory of emotions: Empirical tests, mentalmodels, and implications for social interaction. In L.L. Martin & A. Tesser (Eds.), Striving and feeling: In-teractions among goals, affect, and self-regulation(pp. 363–393). Mahwah, NJ: Erlbaum.

Oatley, K., Keltner, D., & Jenkins, J. M. (2006). Under-standing emotions (2nd ed.). Malden, MA: Blackwell.

O’Connor, F. (1963). My Oedipus complex, and otherstories. Harmondsworth, UK: Penguin.

Paglia, C. (2005). Break, blow, burn. New York: Pan-theon.

Panksepp, J. (1998). Affective neuroscience: The foun-dations of human and animal emotions. Oxford: Ox-ford University Press.

Panksepp, J. (2005). Affective consciousness: Core emo-tional feelings in animals and humans. Consciousnessand Cognition, 14, 30–80.

Parsons, D. (1975). The directory of tunes and musicalthemes. Cambridge, UK: Brown.

Peretz, I., Gagnon, L., & Bouchard, B. (1998). Musicand emotion: Perceptual determinants, immediacy,and isolation after brain damage. Cognition, 68,111–141.

Peretz, I., & Zattore, R. J. (Eds.). (2003). The cognitiveneuroscience of music. New York: Oxford UniversityPress.

Pinker, S. (1997). How the mind works. New York:Norton.

Povel, D.-J. (1984). A theoretical framework for rhythmperception. Psychological Research, 45, 315–337.

Robinson, J. (Ed.). (1997). Music and meaning. Ithaca,NY: Cornell University Press.

Russell, J. A., Bachorowski, J. A., & Fernandez-Dols, J.M. (2003). Facial and vocal expression of emotion.Annual Review of Psychology, 54, 329–349.

Scheff, T. J. (1979). Catharsis in healing, ritual, anddrama. Berkeley: University of California Press.

Scherer, K. R. (1986). Vocal affect expression: A review

and a model for future research. Psychological Bulle-tin, 99, 143–165.

Scherer, K. R. (2004). Which emotions can be inducedby music? What are the underlying mechanisms? Andhow can we measure them? Journal of New MusicResearch, 33, 239–251.

Shakespeare, W. (1981). Hamlet (H. Jenkins, Ed.). Lon-don: Methuen. (Original work performed c. 1600)

Simonton, D. K. (2001). Emotion and composition inclassical music: Historiometric perspectives. In P. N.Juslin & J. A. Sloboda (Eds.), Music and emotion:Theory and research (pp. 205–222). Oxford: OxfordUniversity Press.

Sisman, E. (2000). “The spirit of Mozart from Haydn’shands”: Beethoven’s musical inheritance. In G. Stan-ley (Ed.), The Cambridge companion to Beethoven(pp. 45–63). Cambridge, UK: Cambridge UniversityPress.

Sloboda, J. A. (1985). The musical mind: The cognitivepsychology of music. Oxford: Clarendon Press.

Sloboda, J. A., & Juslin, P. N. (2001). Music and emo-tion: Commentary. In P. N. Juslin & J. A. Sloboda(Eds.), Music and emotion: Theory and research(pp. 453–462). Oxford: Oxford University Press.

Stanislavski, C. (1936). An actor prepares (E. R.Habgood, Trans.). New York: Routledge.

Stravinsky, I. (1936). Chronicles of my life. New York:Simon & Schuster.

Tan, E. S. (1996). Emotion and the structure of film:Film as an emotion machine. Mahwah, NJ: Erlbaum.

Tan, E. S., & Frijda, N. H. (1999). Sentiment in filmviewing. In C. Plantinga & G. M. Smith (Eds.), Pas-sionate views: Film, cognition, and emotion (pp. 48–64). Baltimore: Johns Hopkins University Press.

Trabasso, T., & Chung, J. (2004, January). Empathy:Tracking characters and monitoring their concerns infilm. Paper presented at the Winter Text Conference,Jackson Hole, WY.

Trainor, L. J., Tsang, C. D., & Cheung, V. H. W. (2002).Preference for consonance in 2- and 4-month-old in-fants. Music Perception, 20, 187–194.

Trehub, S. E. (2003). Musical predispositions in infancy:An update. In I. Peretz & R. Zatorre (Eds.), The cog-nitive neuroscience of music (pp. 3–20). Oxford: Ox-ford University Press.

Västfjäll, D. (2002). Emotion induction through music:A review of the musical mood induction procedure.Musicae Scientiae, 173–211.

Williams, T. (2004). The writing process: Scenarios,sketches and rough drafts. In T. Unwin (Ed.), TheCambridge companion to Flaubert (pp. 165–179).Cambridge, UK: Cambridge University Press.

Zentner, M. R., & Kagan, J. (1998). Infants’ perceptionof consonance and dissonance in music. InfantBehavior and Development, 21, 483–492.

Zillmann, D. (1994). Mechanisms of emotional involve-ment with drama. Poetics, 23, 33–51.


C H A P T E R 8

The Evolutionary Psychologyof the Emotions andTheir Relationship to

Internal Regulatory Variables

JOHN TOOBY and LEDA COSMIDES

Evolutionary psychology is an attempt to unifythe psychological, social, and behavioral sci-ences theoretically and empirically within asingle, mutually consistent, seamless scientificframework. The core of this enterprise is the in-tegration of principles and findings drawnfrom evolutionary biology, cognitive science,anthropology, economics, and neurosciencewith psychology in order to produce high-resolution maps of human nature. By “humannature,” evolutionary psychologists mean theevolved, reliably developing, species-typicalcomputational architecture of the humanmind, together with the physical structures andprocesses (in the brain, in development, and ingenetics) that give rise to this information-processing architecture. For evolutionary psy-chologists, all forms of knowledge about brainsand behavior are relevant, but the pivotal stepis using these facts to form accurate models ofthe information-processing structure of psy-chological mechanisms.

The discovery of a correct information-processing description of a psychologicalmechanism is the fundamental clarifying scien-tific step, because each mechanism came intoexistence and was organized by natural selec-tion in order to carry out its particular set ofinformation-processing functions. It is not ametaphor but a reality that the brain is acomputer—a physical system that came intoexistence to carry out computations. The com-putations were needed to solve the adaptiveproblem of regulating behavior successfully.Hence the brain (and its subsystems) evolved tocarry out specific varieties of computation inorder to regulate behavior so that it was bio-logically successful—that is, to assemble the in-dividual somatically and neurally, to preventprereproductive death, to increase the proba-bility of achieving conditions (social and physi-cal) that would have led to successful reproduc-tion in the ancestral world, to reproducesuccessfully, and to assist genetic relatives (in-

114

cluding children) to achieve and maintain con-ditions for their own successful reproduction.

In short, the functional subcomponents (pro-grams) that constitute our psychological archi-tecture were designed by natural selection tosolve adaptive problems faced by our hunter–gatherer ancestors by regulating behavior inways that increased genetic propagation—what biologists call “fitness.” Against the oth-erwise disordering forces of entropy that per-vade all of physical reality, natural selection isthe only process that introduces functional or-ganization into the designs of organisms(Tooby, Cosmides, & Barrett, 2003). So, to theextent that there is functional organization inthe human psychological architecture, it wascreated by, reflects, and is explained by theoperation of natural selection among our an-cestors. This is why evolutionary psychology isnot a specific subfield of psychology, such asthe study of vision, reasoning, or social behav-ior. It is a way of approaching the science ofpsychology that produces (or is intended toproduce) stable functional descriptions of theelements of the mind. (Detailed arguments forthese positions can be found in Tooby &Cosmides, 1990a, 1990b, 1992a, 2005, and inCosmides & Tooby, 1987, 1992, 1997.)

Researchers less familiar with evolutionarypsychology often equate adaptive problems ex-clusively with short-run threats to physical sur-vival. However, survival per se is not central toevolution: All individual organisms die sooneror later. In contrast, genes—which can bethought of as particles of design—are poten-tially immortal, and design features spread bypromoting the reproduction of the genes thatparticipate in building them. Survival is signifi-cant only insofar as it promotes the reproduc-tion of design features into subsequent genera-tions. Survival is no more significant thananything else that promotes reproduction, andis often advantageously risked or sacrificed inthe process of promoting reproduction in self,children, or other relatives. Nearly every kindof event or condition has the potential to havesome impact on the prospect of reproductionfor individuals, their children, and their rela-tives. Consequently, selection on neural designsfor functional behavior reaches out to encom-pass, in a network of cause and effect linkages,virtually all of human life, from the subtletiesof facial expression to attributions of responsi-bility to the intrinsic rewards of play. The realmof adaptive information-processing problems is

not limited to one area of human life, such assex, violence, or resource acquisition. Instead,it is a dimension cross-cutting all areas ofhuman life, as weighted by the strange,nonintuitive metric of their cross-generationalstatistical effects on direct and kin reproduc-tion.

By “computation,” evolutionary psycholo-gists simply mean the organized causation ofpatterned information input–output relations.Natural selection poses adaptive problems ofbehavior regulation, and the mechanisms of thebrain evolved to engineer solutions in the formof these regulatory input–output relations. Ofcourse, these computational relations must beembodied physically in neural tissue, and mustbe designed to develop reliably. Adaptationsare not just the products of the genes, but arethe products of the coordinated interaction of astable genetic inheritance and the evolution-arily long-enduring features of the environ-ment.

A model of an evolved neurocomputationalmechanism or program would answer ques-tions such as these: What information does theprogram take as input? How is this infor-mation encoded, formatted, and representedas data structures? What operations are per-formed on these data structures to transformthem into new representations or regulatory el-ements? And how do these procedures anddata structures interact to generate and regu-late behavior? In short, how does each pro-gram work in cause-and-effect terms?

AN EVOLUTIONARY-PSYCHOLOGICAL APPROACHTO THE EMOTIONS

Although an evolutionary-psychological ap-proach can be applied to any topic in psychol-ogy, it is especially illuminating when appliedto the emotions. To the extent that there isfunctional order to be found in the mechanismsresponsible for the emotions, it was forged overevolutionary time by natural selection actingon our ancestors. The analysis of adaptiveproblems that arose ancestrally has led evolu-tionary psychologists to apply the concepts andmethods of the evolutionary sciences to scoresof topics that are relevant to the study of emo-tion. These include anger, cooperation, sexualattraction, jealousy, aggression, parental love,friendship, romantic love, the aesthetics of

8. The Evolutionary Psychology of the Emotions 115

landscape preferences, coalitional aggression,incest avoidance, disgust, predator avoidance,kinship, and family relations (for reviews, seeBarkow, Cosmides, & Tooby, 1992; Buss,2005; Crawford & Krebs, 1998; Daly & Wil-son, 1988; Pinker, 1997).

Indeed, a rich theory of the emotions natu-rally emerges out of the core principles of evo-lutionary psychology (Tooby, 1985; Tooby &Cosmides, 1990a; Cosmides & Tooby, 2000;see also Nesse, 1991). In this chapter, we (1)briefly state what we think emotions are andwhat adaptive problem they were designed tosolve; (2) explain the evolutionary and compu-tational principles that led us to this view; (3)identify how the emotions relate to motiva-tional and other underlying regulatory vari-ables the human brain is designed to generateand access; and (4) using this background, ex-plicate in a more detailed way the design ofemotion programs and the states they create.

It may strike some as odd to speak aboutlove, jealousy, or disgust in computationalterms. “Computation” has an affectless, fla-vorless connotation. But if the brain evolved asa system of information-processing relations,then emotions are, in an evolutionary sense,best understood as information-processingrelations—that is, programs—with naturallyselected functions. Initially, the commitment toexploring the underlying computational archi-tecture of the emotions may seem infelicitous,but viewing them as programs leads to a largenumber of scientific payoffs. In particular, theclaim that emotion is computational does notmean that an evolutionary-psychological ap-proach misconstrues human experience asbloodless, affectless, disembodied ratiocina-tion. It is simply the claim that one can describethe underlying set of informational relation-ships that explain emotional phenomena, in-cluding the nature of emotional experience. Ev-ery mechanism in the brain—whether it doessomething categorizable as “cold cognition”(such as reasoning, inducing a rule of grammar,or judging a probability) or as “hot cognition”(such as computing the intensity of parentalfear, the imperative to strike an adversary, or anescalation in infatuation)—depends on an un-derlying computational organization to give itsoperation its patterned structure, as well as aset of neural circuits to implement it physically.In these terms, an evolutionary and comp-utational view of emotion can open up for

exploration new empirical and theoretical pos-sibilities obscured by other frameworks.

AN EVOLUTIONARY-PSYCHOLOGICAL THEORYOF THE EMOTIONS

Both deductions from theoretical evolutionarypsychology and a large supporting body of em-pirical findings in psychology, biology, andneuroscience support the view that the humanmental architecture is crowded with evolved,functionally specialized programs. Each is tai-lored to solve a different adaptive problem thatarose during human evolutionary history (orbefore), such as face recognition, foraging,mate choice, heart rate regulation, sleep man-agement, or predator vigilance, and each is ac-tivated by a different set of cues from the envi-ronment.

But the existence of all these diverse programsitself creates an adaptive problem: Programsthat are individually designed to solve specificadaptive problems could, if simultaneously acti-vated, deliver outputs that conflict with one an-other, interfering with or nullifying each other’sfunctional products. For example, sleep andflight from a predator require mutually inconsis-tent actions, computations, and physiologicalstates. It is difficult to sleep when your heart andmind are racing with fear, and this is no accident:Disastrous consequences would ensue ifproprioceptive cues were activating sleep pro-grams at the same time that the sight of a stalkinglion was activating ones designed for predatorevasion. To avoid such consequences, the mindmust be equipped with superordinate programsthat override and deactivate some programswhen others are activated (e.g., a program thatdeactivates sleep programs when predator eva-sion subroutines are activated). Reciprocally,many adaptive problems are best solved by thecoordinated activation of a specific subset ofprograms, with each program being entrainedinto the computational settings most appropri-ate for the particular adaptive problem beingfaced. For example, predator avoidance may re-quire simultaneous shifts in both heart rate andauditory acuity (see below). To do this, a specialtype of program is required that manages andharmonizes other programs, aligning each ofthem into the proper configuration at the righttime.


In general, to behave functionally accordingto evolutionary standards, the mind’s manysubprograms need to be orchestrated so thattheir joint product at any given time is coordi-nated to deal with the adaptive challenge beingfaced, rather than operating in a self-defeating,discoordinated, and cacophonous fashion. Weargue that such coordination is accomplishedby a special class of programs: the emotionsthat evolved to solve these superordinate de-mands. In this view, the best way to understandwhat the emotions are, what they do, and howthey operate is to recognize that mechanism or-chestration is the function that defines theemotions, and explains in detail their designfeatures. They are the neurocomputational ad-aptations that have evolved in response to theadaptive problem of matching arrays of mecha-nism activation to the specific adaptive de-mands imposed by alternative situations(Tooby & Cosmides, 1990a; Tooby, 1985;Cosmides & Tooby, 2000; Nesse, 1991).

Thus each emotion evolved to deal with aparticular, evolutionarily recurrent situationtype. The design features of the emotion pro-gram, when the emotion is activated, presumethe presence of an ancestrally structured situa-tion type (regardless of the actual structure ofthe modern world). Hence the exploration ofthe statistical structure of ancestral situationsand their relationship to the mind’s battery offunctionally specialized programs is central tomapping the emotions. This is because themost useful (or least harmful) deployment ofprograms at any given time will depend criti-cally on the exact nature of the situation beingencountered. The abstract, distilled, recurrentcharacteristics of the situation are reflected inthe architecture of the emotion. For example,because sexual rivals could be advantageouslydriven off by violence or its threat in a substan-tial fraction of the trillions of ancestral cases ofmate competition, sexual jealousy is engineeredto prepare the body physiologically for com-bat, and (when the rival is weak or unwary)motivates the individual to behave violently. Inmodern situations of potential or actual infidel-ity, police and prisons create additional conse-quences, and so violence against a sexual rivalis likely to lead to maladaptive outcomes now.However, the design features of jealousy weredesigned to mesh with the long-enduring struc-ture of the ancestral world, and not the modernworld—so the emotion program continues to

execute its own ancestral functional logic evenunder modern conditions.

How did emotions arise and assume theirdistinctive structures? Fighting, falling in love,escaping predators, confronting sexual infidel-ity, experiencing a failure-driven loss in status,responding to the death of a family member,and so on each involved conditions, contingen-cies, situations, or event types that recurred in-numerable times in hominid evolutionary his-tory. Repeated encounters with each kind ofsituation selected for adaptations that guidedinformation processing, behavior, and the bodyadaptively through the clusters of conditions,demands, and contingencies characterizing thatparticular class of situation.

The payoffs accruing to alternative mutantdesigns for program activation, in interactionwith recurrent classes of situations, engineeredprograms each of which jointly mobilizes asubset of the psychological architecture’s otherprograms in a particular configuration. Eachconfiguration was selected to deploy computa-tional and physiological mechanisms in a waythat, when averaged over individuals and gen-erations, would have led to the most fitness-promoting subsequent lifetime outcome, giventhat ancestral situation type. Thus an emotionis a bet placed under conditions of uncertainty:It is the evolved mind’s bet about what internaldeployment is likely to lead to the best averagelong-term set of payoffs, given the structureand statistical contingencies present in the an-cestral world when a particular situation wasencountered. Running away in terror, vomitingin disgust, or attacking in rage are bets that areplaced because these responses had the highestaverage payoffs for our ancestors, given theeliciting conditions.

This coordinated adjustment and entrain-ment of mechanisms constitutes a mode ofoperation for the entire psychological architec-ture, and serves as the basis for a precise com-putational and functional definition of eachemotion state (Tooby & Cosmides, 1990a;Tooby, 1985; Cosmides & Tooby, 2000). Eachemotion entrains various other adaptiveprograms—deactivating some, activating oth-ers, and adjusting the modifiable parameters ofstill others—so that the whole system operatesin a particularly harmonious and efficaciousway when the individual is confronting certainkinds of triggering conditions or situations.The conditions or situations relevant to the


emotions are those that (1) recurred ances-trally; (2) could not be negotiated successfullyunless there was a superordinate level of pro-gram coordination (i.e., circumstances inwhich the independent operation of programscaused no conflicts would not have selected foran emotion program, and would lead to emo-tionally neutral states of mind); (3) had a richand reliable repeated structure; (4) had recog-nizable cues signaling their presence; and (5)would have resulted in large fitness costs if anerror had occurred (Tooby & Cosmides,1990a; Tooby, 1985; Cosmides & Tooby,2000). When a condition or situation of anevolutionarily recognizable kind is detected, asignal is sent out from the emotion programthat activates the specific constellation ofsubprograms appropriate to solving the type ofadaptive problems that were regularly embed-ded in that situation, and deactivates programswhose operation might interfere with solvingthose types of adaptive problems. Programs di-rected to remain active may be cued to entersubroutines that are specific to that emotionmode, and that were tailored by natural selec-tion to solve the problems inherent in the trig-gering situation with special efficiency. (Wherethere was no repeated structure, or there wereno cues to signal the presence of a repeatedstructure, then selection could not build an ad-aptation to address the situation.)

According to this theoretical framework, anemotion is a superordinate program whosefunction is to direct the activities and interac-tions of the subprograms governing perception;attention; inference; learning; memory; goalchoice; motivational priorities; categorizationand conceptual frameworks; physiological re-actions (such as heart rate, endocrine function,immune function, gamete release); reflexes;behavioral decision rules; motor systems; com-munication processes; energy level and effortallocation; affective coloration of events andstimuli; recalibration of probability estimates,situation assessments, values, and regulatoryvariables (e.g., self-esteem, estimations of rela-tive formidability, relative value of alternativegoal states, efficacy discount rate); and so on.An emotion is not reducible to any one cate-gory of effects, such as effects on physiology,behavioral inclinations, cognitive appraisals, orfeeling states, because it involves evolved in-structions for all of them together, as well asother mechanisms distributed throughout thehuman mental and physical architecture.

FEAR AS A MODE OF OPERATION

Consider the following example. The ances-trally recurrent situation is being alone atnight, and a situation detector circuit perceivescues that indicate the possible presence of a hu-man or animal predator. The emotion mode isa fear of being stalked. (In this conceptualiza-tion of emotion, there might be several distinctemotion modes that are lumped together underthe folk category “fear,” but that are at leastpartially distinguishable, computationally andempirically, by the overlapping but nonidenti-cal constellation of programs each entrains.)When the situation detector signals that onehas entered the situation of “possible stalkingand ambush,” the following kinds of mentalprograms are entrained or modified:

1. There are shifts in perception and atten-tion. You may suddenly hear with far greaterclarity sounds that bear on the hypothesis thatyou are being stalked, but that ordinarily youwould not perceive or attend to, such as creaksor rustling. Are the creaks footsteps? Is the rus-tling caused by something moving stealthilythrough the bushes? Signal detection thresh-olds shift: Less evidence is required before yourespond as if there were a threat, and more truepositives will be perceived at the cost of ahigher rate of false alarms.

2. Goals and motivational weightingschange. Safety becomes a far higher priority.Other goals and the computational systemsthat subserve them are deactivated: You are nolonger hungry; you cease to think about how tocharm a potential mate; practicing a new skillno longer seems rewarding. Your planning fo-cus narrows to the present; worries about yes-terday and tomorrow temporarily vanish.Hunger, thirst, and pain are suppressed.

3. Information-gathering programs are redi-rected: Where is my child? Where are otherswho can protect me? Is there somewhere I cango where I can see and hear what is going onbetter?

4. Conceptual frames shift, with the auto-matic imposition of categories such as “danger-ous” or “safe.” Walking a familiar and usuallycomfortable route may now be mentally taggedas “dangerous.” Odd places that you normallywould not occupy—a hallway closet, thebranches of a tree—suddenly may become sa-lient as instances of the category “safe” or“hiding place.”


5. Memory processes are directed to new re-trieval tasks: Where was that tree I climbed be-fore? Did my adversary and his friend look atme furtively the last time I saw them?

6. Communication processes change. De-pending on the circumstances, decision rulesmay cause you to emit an alarm cry, or be para-lyzed and unable to speak. Your face may auto-matically assume a species-typical fear expres-sion.

7. Specialized inference systems are acti-vated. Information about a lion’s trajectory oreye direction may be fed into systems for infer-ring whether the lion saw you. If the inferenceis yes, then a program automatically infers thatthe lion knows where you are; if no, then thelion does not know where you are (the “seeing-is-knowing” circuit identified by Baron-Cohen,1995, as impaired in persons with autism).This variable may automatically governwhether you freeze in terror or bolt. Are therecues in the lion’s behavior that indicate whetherit has eaten recently, and so is unlikely to bepredatory in the near future? (Savanna-dwelling ungulates, such as zebras and wilde-beests, commonly make this kind of judgment;Marks, 1987.)

8. Specialized learning systems are acti-vated, as the large literature on fear condition-ing indicates (e.g., LeDoux, 1995; Mineka &Cook, 1993; Pitman & Orr, 1995). If the threatis real, and the ambush occurs, you may experi-ence an amygdala-mediated recalibration (as inposttraumatic stress disorder) that can last forthe remainder of your life (Pitman & Orr,1995).

9. Physiology changes. Gastric mucosa turnwhite as blood leaves the digestive tract (an-other concomitant of motivational prioritieschanging from feeding to emergency motor ac-tivity in pursuit of safety); adrenalin spikes;heart rate may go up or down (depending onwhether the situation calls for flight or immo-bility), blood rushes to the periphery, and so on(Cannon, 1929; Tomaka, Blascovich, Kibler, &Ernst, 1997); instructions to the musculature(face and elsewhere) are sent (Ekman, 1982).Indeed, the nature of the physiological re-sponse can depend in detailed ways on the na-ture of the threat and the best response option(see, e.g., Marks, 1987).

10. Behavioral decision rules are activated:Depending on the nature of the potentialthreat, different courses of action will be poten-tiated: hiding, flight, self-defense, or even tonic

immobility (the last of these is a common re-sponse to actual attacks, both in other animalsand in humans1). Some of these responses maybe experienced as automatic or involuntary.

From the point of view of avoiding danger,these computational changes are crucial: Theyare what allowed the adaptive problem to besolved with high probability, on average overevolutionary time. Of course, in any single casethey may fail, because they are only theevolutionarily computed best bet, based on an-cestrally summed outcomes; they are not a surebet, based on an unattainable perfect knowl-edge of the present.

Whether individuals report consciously ex-periencing fear is a separate question fromwhether their mechanisms assumed the charac-teristic configuration that, according to thistheoretical approach, defines the fear emotionstate. Individuals often behave as if they are inthe grip of an emotion, while denying they arefeeling that emotion. We think it is perfectlypossible that individuals sometimes remain un-aware of (or lose conscious access to) theiremotion states, which is one reason we do notuse subjective experience as the sine qua non ofemotion. At present, both the function of con-scious awareness, and the principles that regu-late conscious access to emotion states andother mental programs, are complex and unre-solved questions (but see Tooby, Cosmides,Sell, Lieberman, & Sznycer, in press). Mappingthe design features of emotion programs canproceed independently of their resolution, atleast for the present.

ADAPTATIONIST FOUNDATIONS

Adaptations, By-Products, and Noise

Because of the different roles played by chanceand selection, the evolutionary process buildsthree different types of outcomes into organ-isms: (1) adaptations—that is, functional ma-chinery built by selection, and usually species-typical (see Tooby & Cosmides, 1990b, for de-tails and exceptions); (2) by-products of adap-tations, which are present in the design of or-ganisms because they are causally coupled totraits that were selected for (usually species-typical); and (3) random noise, injected by mu-tation and other random processes (often notspecies-typical) (Tooby & Cosmides, 1990a,1990b, 1992a; Williams, 1966). The emotion


of sexual jealousy is an adaptation (Daly, Wil-son, & Weghorst, 1982; Buss, 1994); stress-induced physical deterioration is arguably a by-product of the flight–fight system; and herita-ble personality variation in emotional function-ing (e.g., extreme shyness, morbid jealousy, bi-polar disorder) is probably noise (Tooby &Cosmides, 1990b). Evidence of the presence (orabsence) of high degrees of coordination be-tween adaptive problems and the design fea-tures of putative adaptations allows research-ers to distinguish adaptations, by-products,and noise from one another (Williams, 1966;Cosmides & Tooby, 1997).

The emotions are often thought of as crude,but we expect emotions to be very well-designed computational adaptations. Biologistshave found that selection has routinely pro-duced exquisitely engineered biological ma-chines of the highest order at all scales, fromgenetic error correction and quality control inprotein assembly to photosynthetic pigments,the immune system, efficient bee foraging algo-rithms, echolocation, and color constancy sys-tems. Indeed, the best-studied psychologicaladaptation—the eye and visual system—hasbeen held up for centuries as the apotheosis ofengineering excellence, as yet unrivaled by anyhuman engineer. There is no principled reasonto expect other neurocomputational (i.e., psy-chological) adaptations to be less well engi-neered than the eye. Although Stephen JayGould (1997) and his followers have energeti-cally argued in the popular science literaturethat natural selection is a weak evolutionaryforce, evolutionary biologists, familiar with theprimary literature, have found it difficult totake these arguments seriously (Tooby &Cosmides, 1997). So although adaptations arein some abstract sense undoubtedly far fromoptimal (and there is genetic noise in all sys-tems), the empirical evidence falsifies the claimthat evolved computational adaptations tendto be crude or primitive in design, and insteadsupports the opposite view: that our mentalmachinery—including the emotions—is likelyto be very well designed to carry out evolvedfunctions. For emotion researchers, this meansthat their working hypotheses (which are al-ways open to empirical revision) should beginwith the expectation of high levels of evolu-tionary functionality, and their research meth-ods should be sensitive enough to detect suchorganization. This does not mean that emo-tions are well designed for the modern world—

only that their functional logic is likely to besophisticated and well engineered to solve an-cestral adaptive problems.

The Environmentof Evolutionary Adaptedness

Behavior in the present is generated by evolvedinformation-processing mechanisms that wereconstructed in the past. They were constructedin the past because they solved adaptive prob-lems that were recurrently present in the ances-tral environments in which the human lineevolved. For this reason, evolutionary psychol-ogy is both environment-oriented and past-oriented in its functionalist orientation. Adap-tations become increasingly effective as selec-tion makes their design features more and morecomplementary to the long-enduring structureof the world. The articulated features of the ad-aptation are designed to mesh with the featuresof the environment that were stable during theadaptation’s evolution, so that their interactionproduced functional outcomes. The regulationof breathing assumes the presence of certainlong-enduring properties of the atmosphereand the respiratory system. Vision assumes thepresence of certain evolutionarily stable prop-erties of surfaces, objects, and terrestrial spec-tral distributions. The digestive enzyme lactasepresupposes an infant diet of milk with lactose.Fear presupposes dangers in the environment,and even presupposes higher probabilities ofspecific kinds of dangers, given certain cues:darkness, spiders, snakes, heights, predators,open spaces, and so on (Marks, 1987). That is,each emotion program presupposes that cer-tain cues signal the presence of a structure ofevents and conditions that held true during theevolution of that emotion. Disgust circuits pre-sume a world in which rotten smells signal tox-ins or microbial contamination, for example.

Accordingly, to understand an adaptation asa problem solver, one needs to model the en-during properties of the task environment thatconstituted the problem and provided materi-als that could be exploited for its solution: the“environment of evolutionary adaptedness,”or (EEA). Although the human line is thoughtto have first differentiated itself from the chim-panzee lineage on the African savannahs andwoodlands, the EEA is not a place or time. It isthe statistical composite of selection pressuresthat caused the genes underlying the design ofan adaptation to increase in frequency until


they became species-typical or stably persistent(Tooby & Cosmides, 1990a). Thus statisticalregularities define the EEA for any given adap-tation. The conditions that characterize theEEA are usefully decomposed into a constella-tion of specific environmental regularities thathad a systematic (though not necessarily un-varying) impact on reproduction, and thatendured long enough to work evolutionarychange on the design of an adaptation. Some ofthese regularities are extremely simple: Dis-tance from a predator is protection from thepredator. Sex with an opposite-sex adult ismore likely to produce offspring than sex witha child or a nonhuman. These regularities canequally well include complex conditionals (e.g.,if one is a male hunter–gatherer and one is hav-ing a sexual liaison with someone else’s mateand that liaison is discovered, then one is thetarget of lethal retributory violence 14% of thetime). Descriptions of these regularities are es-sential parts of the construction of a task analy-sis of the adaptive problem a hypothesized ad-aptation evolved to solve (Tooby & Cosmides,1990a). Conceptualizing the EEA in probabil-istic terms is fundamental to the functional def-inition of emotion that we have presentedabove and will elucidate below.

Each adaptive problem recurred billions ortrillions of times in the EEA, and so manifesteda statistical and causal structure whose ele-ments were available for specialized exploita-tion by design features of the evolving adapta-tion. For example, predators use darkness andcover to ambush (Marks, 1987). Physicalappearance varies with fertility and health(Symons, 1979). Among hunter–gatherers, in-fants that a mother primarily cares for are al-most invariably genetic siblings (Lieberman,Tooby, & Cosmides, 2007). Specializedprograms—for predator fear, sexual attraction,and kin detection, respectively—could evolvewhose configuration of design features embod-ied and/or exploited these statistical regulari-ties, allowing these adaptive problems to besolved economically, reliably, and effectively.Such specializations, by embodying “innateknowledge” about the problem space, operatebetter than any general learning strategy could.Children did not have to wait to experience be-ing ambushed and killed in the dark to pru-dently modulate their activities. Adults did notneed to observe the negative effects of incest,because the human kin detection system mobi-lizes disgust toward having sex with individu-

als the mind has tagged as siblings (Liebermanet al., 2007).

The Functional Structureof an Emotion Program Evolved to Matchthe Evolutionarily Summed Structureof Its Target Situation

Each emotion program was constructed by aselective regimen consisting of repeated en-counters with a particular kind of evolution-arily recurrent situation. By an “evolutionarilyrecurrent situation,” we mean a cluster of re-peated probabilistic relationships amongevents, conditions, actions, and choice conse-quences that endured over a sufficient stretchof evolutionary time to have favored some vari-ant designs over others. Many of these relation-ships were probabilistically associated withcues detectable by humans, allowing psycho-physical triggers to activate the task-appropriate program.

For example, the condition of having a mateplus the condition of the mate’s copulating withsomeone else constitutes a situation of sexualinfidelity—a situation that has recurred overevolutionary time, even though it has not hap-pened to every individual. Associated with thissituation were cues reliable enough to allow theevolution of a “situation detector” (e.g., ob-serving a sexual act, flirtation, or even the re-peated simultaneous absence of the suspectedlovers were cues that could trigger the categori-zation of a situation as one of infidelity). Evenmore importantly, there were many necessarilyor probabilistically associated elements thattended to be present in the situation of infidel-ity as encountered among our hunter–gathererancestors. These additional elements included(1) a sexual rival with a capacity for social ac-tion and violence, as well as allies of the rival;(2) a discrete probability that one’s mate hadconceived with the sexual rival; (3) changes inthe net lifetime reproductive returns of invest-ing further in the mating relationship; (4) aprobable decrease in the degree to which theunfaithful mate’s mechanisms would value thevictim of infidelity (the presence of an alterna-tive mate would lower replacement costs); (5) acue that the victim of the infidelity was likely tohave been deceived about a range of pastevents, leading the victim to confront the likeli-hood that his or her memory was permeatedwith false information; and (6) a likelihoodthat the victim’s status and reputation for being


effective at defending his or her interests in gen-eral would plummet, inviting challenges inother arenas. These are just a few of the manyfactors that would constitute a list of elementsassociated in a probabilistic cluster, and thatwould constitute the evolutionary recurrentstructure of a situation of sexual infidelity. Theemotion of sexual jealousy evolved in responseto these properties of the world, and thereshould be evidence of this in its computationaldesign.

Emotion programs have evolved to take suchelements into account, whether they can beperceived or not. Thus not only do cues of a sit-uation trigger an emotion mode, but embeddedin that emotion mode is a way of seeing theworld and feeling about the world related tothe ancestral cluster of associated elements. De-pending on the intensity of the jealousy evoked,less and less evidence will be required for indi-viduals to believe that these conditions apply totheir personal situation. Individuals with mor-bid jealousy, for example, may hallucinatecounterfactual but evolutionarily thematic con-tents, such as seeing their mates having sexwith someone else (Mowat, 1966; Shepherd,1961). This leads many to consider emotions“irrational,” but this property was selected forbecause it allows emotional computation to gobeyond the evidence given, producing correctresponses (when averaged over evolutionarytime).

To the extent that situations exhibited astructure repeated over evolutionary time, theirstatistical properties would be used as the basisfor natural selection to build an emotion pro-gram whose detailed design features were tai-lored for that situation. This would be accom-plished by selection acting over evolutionarytime, differentially incorporating programcomponents that dovetailed with individualitems on the list of properties probabilisticallyassociated with the situation.

For example, ancestrally a male’s ability toinflict costs through violence (his “formid-ability”) was associated with his status andreputation for defending his interests. More-over, the fitness consequences of being cuck-olded are great, and males have become moti-vated by design to resist this outcome. If amale’s mate is sexually unfaithful and this infi-delity becomes public, this advertises a weak-ness previously unappreciated by those whoknow him best. This decrease in perceivedformidability decreases his value to his male al-

lies and increases the probability that he will bechallenged by competitors in other domains oflife. The sexual jealousy, anger, and shame sys-tems have been shaped by the distillation ofthese (and other) payoff probabilities. Each ofthese recurrent subelements in a situation ofsexual infidelity, and the adaptive circuits theyrequire, can be added together to form a gen-eral theory of sexual jealousy, as well as a the-ory of the functional coactivation of linkedprograms (such as anger and shame).

Hence the emotion of sexual jealousy consti-tutes an organized mode of operation specifi-cally designed to deploy the programs govern-ing each psychological mechanism, so that eachis poised to deal with the exposed infidelity.Physiological processes are prepared for suchthings as violence, sperm competition, and thewithdrawal of investment; the goal of deter-ring, injuring, or murdering the rival emerges;the goal of punishing, deterring, or desertingthe mate appears; the desire to make oneselfmore competitively attractive to alternativemates emerges; memory is activated toreanalyze the past; confident assessments of thepast are transformed into doubts; the generalestimate of the reliability and trustworthinessof the opposite sex (or indeed everyone) maydecline; associated shame programs may betriggered to search for situations in which theindividual can publicly demonstrate acts of vio-lence or punishment that work to counteractan (imagined or real) social perception ofweakness; and so on.

It is the relationship between the summeddetails of the ancestral condition and the de-tailed structure of the resulting emotion pro-gram that makes this approach so useful foremotion researchers. Each functionally distinctemotion state—fear of predators, guilt, sexualjealousy, rage, grief, and so on—will corre-spond to an integrated mode of operation thatfunctions as a solution designed to take advan-tage of the particular structure of the recurrentsituation or triggering condition to which thatemotion corresponds. This approach can beused to create theories of each individual emo-tion, through four steps: (1) Reconstruct theclusters of properties of ancestral situations; (2)analyze what behavioral and somatic alter-ations would solve the adaptive problem posedby the recurrent situation (or minimize thedamage it causes); (3) construct a provisionalmodel of the program architecture of the emo-tion that could generate the necessary


mechanism-, body-, and behavior-regulatingoutputs, including the cues used, the regulatoryvariables the emotion needs to track, and soon; and (4) design and conduct experimentsand other investigations to test each hypothe-sized design feature of the proposed emotionprogram, revising them as necessary.

It is also important to understand that evo-lutionarily recurrent situations can be arrayedalong a spectrum in terms of how rich or skele-tal the set of probabilistically associated ele-ments defining the recurrent situation is. Richlystructured situations—such as sexual infidelity,exposure to potential disease vectors, or preda-tor ambush—will support a richly substruc-tured emotion program in response to the nu-merous ancestrally correlated features eachmanifests: Many detailed adjustments will bemade to many psychological mechanisms as in-structions for the mode of operation. In con-trast, some recurrent situations have less struc-ture (i.e., they share fewer properties), and sothe emotion mode makes fewer highly special-ized adjustments, imposes fewer specializedand compelling interpretations and behavioralinclinations, and so on. For example, surges ofhappiness or joy are an emotion program thatevolved to respond to the recurrent situation ofencountering unexpected positive events. Theclass of events captured by “unexpectedly posi-tive” is extremely broad and general, and suchevents have only a few additional properties incommon. Emotion programs at the most gen-eral and skeletal end of this spectrum corre-spond to what some call “mood” (happiness,sadness, excitement, anxiety, playfulness,homesickness, etc.).

HOW TO CHARACTERIZEAN EMOTION

To characterize an emotion adaptation, onemust identify the following properties of envi-ronments and of mechanisms.

1. An evolutionarily recurrent situation orcondition. A “situation” is a repeated structureof environmental and organismic properties,characterized as a complex statistical compos-ite of how such properties covaried in the envi-ronment of evolutionary adaptedness. Exam-ples of these situations are being in a depletednutritional state, competing for maternal atten-tion, being chased by a predator, being about to

ambush an enemy, having few friends, experi-encing the death of a spouse, being sick, havingexperienced a public success, having others actin a way that damages you without regard foryour welfare, having injured a valued otherthrough insufficient consideration of self–otherbehavioral tradeoffs, and having a baby.

2. The adaptive problem. Identifying theadaptive problem means identifying which or-ganismic states and behavioral sequences willlead to the best average functional outcome forthe remainder of the lifespan, given the situa-tion or condition. For example, what is the bestcourse of action when others take the productsof your labor without your consent? What isthe best course of action when you are in a de-pleted nutritional state? What is the best courseof action when a sibling makes a sexual ap-proach?

3. Cues that signal the presence of the situa-tion. For example, low blood sugar signals adepleted nutritional state; the looming ap-proach of a large, fanged animal signals thepresence of a predator; seeing your mate hav-ing sex with another signals sexual infidelity;finding yourself often alone, rarely the recipi-ent of beneficent acts, or actively avoided byothers signals that you have few friends.

4. Situation-detecting algorithms. A multi-modular mind must be full of “demons”—al-gorithms that detect situations. The NewHacker’s Dictionary defines a “demon” as a“portion of a program that is not invoked ex-plicitly, but that lies dormant waiting for somecondition(s) to occur” (Raymond, 1991,p. 124). Situation-detecting subprograms liedormant until they are activated by a specificconstellation of cues that precipitates the anal-ysis of whether a particular ancestral situationhas arisen. If the assessment is positive, it sendsthe signal that activates the associated emotionprogram. Emotion demons need two kinds ofsubroutines:

a. Algorithms that monitor for situation-defining cues. These programs include percep-tual mechanisms, proprioceptive mechanisms,and situation-representing mechanisms. Theytake the cues in point 3 above as input.

b. Algorithms that detect situations. Theseprograms take the output of the monitoring al-gorithms and targeted memory registers inpoint a as input, and through integration,probabilistic weighting, and other decision cri-teria, identify situations as absent or presentwith some probability and with some index of


the magnitude of the fitness consequences in-herent in the situation.

The assignment of a situation interpretationto present circumstances involves a problem insignal detection theory (Tooby & Cosmides,1990a; Swets, Tanner, & Birdsall, 1961; seealso Gigerenzer & Murray, 1987). Animalsshould be designed to detect what situationthey are in on the basis of cues, stored regula-tory variables, and specialized interpretationalgorithms. Selection will not shape decisionrules so that they act solely on the basis of whatis most likely to be true, but rather on the basisof the weighted consequences of acts, giventhat something is held to be true. Should youwalk under a tree that might conceal a preda-tor? Even if the algorithms assign a 51% (oreven 98%) probability to the tree’s beingleopard-free, under most circumstances anevolutionarily well-engineered decision ruleshould cause you to avoid the tree—to act as ifthe leopard were in it. The benefits of caloriessaved via a shortcut, scaled by the probabilitythat there is no leopard in the tree, must beweighed against the benefits of avoiding be-coming catfood, scaled by the probability thatthere is a leopard in the tree. Because the costsand benefits of false alarms, misses, hits, andcorrect rejections are often unequal, the deci-sion rules may still treat as true situations thatare unlikely to be true. In the modern world,this behavior may look “irrational” (as is thecase with many phobias), but we do it becausesuch decision biases were adaptive under an-cestral conditions, given ancestral payoff asym-metries. That is, they were “ecologically ratio-nal” (Tooby & Cosmides, 1990a; Haselton &Buss, 2003).

Situation-detecting algorithms can be of anydegree of complexity, from demons that moni-tor single cues (e.g., “snake present”) to algo-rithms that carry out more complex cognitiveassessments of situations and conditions(LeDoux, 1995; Lazarus & Lazarus, 1994;Tooby & Cosmides, 1990a). Inherent in thisapproach is the expectation that the humanmind has a series of evolved subsystemsdesigned to represent events in terms ofevolutionarily recurrent situations and situa-tional subcomponents. The operations of theserepresentational systems are not necessarilyconsciously accessible. By their structure, theyimpose an evolutionary organization on repre-sentational spaces that are updated by data in-

puts. When the representational space assumescertain configurations, an interpretation is trig-gered that activates the associated emotionprogram—corresponding approximately towhat others have called a “cognitive appraisal”(see, e.g., Lazarus & Lazarus, 1994). It is im-portant to recognize that the evolutionary pastframes the experienced present, because thesesituation-detecting algorithms provide the di-mensions and core elements out of which manycross-culturally recurring representations ofthe world are built. To some extent, the worldwe inhabit is shaped by the continuous inter-pretive background commentary provided bythese mechanisms.

5. Algorithms that assign priorities. A givenworld state may correspond to more than onesituation at a time; for example, you may benutritionally depleted and in the presence of apredator. The prioritizing algorithms definewhich emotion modes are compatible (e.g.,hunger2 and boredom) and which are mutuallyexclusive (e.g., feeding and predator escape).Depending on the relative importance of thesituations and the reliability of the cues, theprioritizing algorithms decide which emotionmodes to activate and deactivate, and to whatdegree. Selection, through ancestral mutant ex-periments, would have sorted emotions basedon the average importance of the consequencesstemming from each, and the extent to whichjoint activation was mutually incompatible orfacilitating. (Prioritizing algorithms can bethought of as a supervisory system operatingover all of the emotions.)

6. An internal communication system.Given that a situation has been detected, the in-ternal communication system sends a situation-specific signal to all relevant programs andmechanisms; the signal switches them into theappropriate adaptive emotion mode. In addi-tion, information is fed back into the emotionprogram from other programs and systems thatassess body states and other regulatory vari-ables, which may govern the intensity, trajec-tory, supplantation, or termination of theemotion. Along with the sensorium and moti-vational systems, the emotions are embeddedin and partly responsible for what might becalled “feeling computation.” In this view, therichly textured representations we experienceas feeling constitute our conscious access to ahigh-bandwidth system of computational de-vices and program interfaces that amalgamate


valuation information with other representa-tions to guide decision making and torecalibrate decisions in an ongoing way (see,e.g., Tooby et al., 2003).

Some modes of activation of the psychologi-cal architecture are accompanied by a charac-teristic feeling state, a certain quality of experi-ence. The fact that we are capable of becomingaware of certain physiological states—ourhearts thumping, bowels evacuating, stomachstightening—is surely responsible for some ofthe qualia evoked by emotion states that en-train such responses. The fact that we arecapable of becoming aware of certain mentalstates—such as the magnitude of certain regu-latory variables or the retrieved memories ofpast events—is probably responsible for otherqualia. In our view, the characteristic feelingstate that accompanies an emotion mode re-sults (in part) from mechanisms that allow usto sense the signal activating and deactivatingthe relevant programs, as well as signals com-municating necessary parameters and variablemagnitudes to the various programs. Suchinternal sensory mechanisms—analogous toproprioception—can be selected for if there aremechanisms requiring as input the informationthat a particular emotion mode has been acti-vated. (This might be true, for example, ofmechanisms designed to inhibit certainstimulus-driven actions when the conditionsare not auspicious.)

7. Each program and physiological mecha-nism entrained by an emotion program musthave associated algorithms that regulate how itresponds to each emotion signal. These algo-rithms determine whether the mechanismshould switch on or switch off, and if on, whatemotion-specialized performance it will imple-ment. For example, there should be algorithmsin the auditory system that, upon detecting thefear signal (see point 6), reset signal detectionthresholds, increasing acuity for predator-relevant sounds.

WHAT KINDS OF PROGRAMSCAN EMOTIONS MOBILIZE?

Any controllable biological or neurocomputa-tional process that, by shifting its performancein a specifiable way, would lead to enhancedaverage fitness outcomes should have come tobe partially governed by emotional state (see

point 7 above). Some such processes are dis-cussed in this section.

Goals

The cognitive mechanisms that define goalstates and choose among goals in a planningprocess should be influenced by emotions. Forexample, vindictiveness—a specialized subcate-gory of anger—may define “injuring the of-fending party” as a goal state to be achieved.(Although the evolved functional logic of thisprocess is deterrence, this function need not berepresented, either consciously or uncon-sciously, by the mechanisms that generate thevindictive behavior.)

Motivational Priorities

Mechanisms involved in hierarchically rank-ing goals or calibrating other kinds of moti-vational and reward systems should beemotion-dependent. What may be extremelyunpleasant in one state, such as harming an-other, may seem satisfying in another state(e.g., aggressive competition may facilitatecounterempathy). Different evolutionarily re-current situations predict the presence—visible or invisible—of different opportunities,risks, and payoffs, so motivational thresholdsand valences should be entrained. For exam-ple, a loss of face should increase the motiva-tion to take advantage of opportunities forstatus advancement, and should decrease at-tention to attendant costs.

Information-Gathering Motivations

Because establishing which situation one is inhas enormous consequences for the appropri-ateness of behavior, the process of detectionshould in fact involve specialized inference pro-cedures and specialized motivations to discoverwhether certain suspected facts are true orfalse. What one is curious about, what onefinds interesting, and what one is obsessed withdiscovering should all be emotion-specific.

Imposed Conceptual Frameworks

Emotions should prompt construals of theworld in terms of concepts that are appropriateto the decisions that must be made. When oneis angry, domain-specific concepts such as so-


cial agency, fault, responsibility, and punish-ment will be assigned to elements in the situa-tion. When one is hungry, the food–nonfooddistinction will seem salient. When one is en-dangered, safety categorization frames will ap-pear. The world will be carved up into catego-ries based partly on what emotional state anindividual is in.

Perceptual Mechanisms

Perceptual systems may enter emotion-specificmodes of operation. When one is fearful, acuityof hearing may increase. Specialized perceptualinference systems may be mobilized as well: Ifyou’ve heard rustling in the bushes at night, hu-man and predator figure detection may be par-ticularly boosted, and not simply visual acuityin general. In fact, nonthreat interpretationsmay be depressed, and the same set of shadowswill “look threatening”—that is, given a spe-cific threatening interpretation such as “a manwith a knife”—or not, depending on emotionstate.

Memory

The ability to call up particularly appropriatekinds of information out of long-term memoryought to be influenced. A woman who has justfound strong evidence that her husband hasbeen unfaithful may find herself flooded by atorrent of memories about small details thatseemed meaningless at the time but that now fitinto an interpretation of covert activity. Wealso expect that what is stored about presentexperience will also be differentially regulated.Important or shocking events, for example,may be stored in great detail (as has beenclaimed about “flashbulb memories”), butother, more moderate emotion-specific effectsmay occur as well.

Attention

The entire structure of attention, from percep-tual systems to the contents of high-level rea-soning processes, should be regulated by emo-tional state. If you are worried that your spouseis late and might have been injured, it ishard to concentrate on other ongoing tasks(Derryberry & Tucker, 1994), but easy to con-centrate on danger scenarios. Positive emotionsmay broaden attentional focus (Fredrickson,1998).

Physiology

Each organ system, tissue, or process is a po-tential candidate for emotion-specific regula-tion, and “arousal” is insufficiently specific tocapture the detailed coordination involved.Each emotion program should send out a dif-ferent pattern of instructions (to the face andlimb muscles, the autonomic system, etc.), tothe extent that the problems embedded in theassociated situations differ. This leads to an ex-pectation that different constellations of effectswill be diagnostic of different emotion states(Ekman, Levenson, & Friesen, 1983). Changesin circulatory, respiratory, and gastrointestinalfunctioning are well known and documented,as are changes in endocrinological function. Weexpect thresholds regulating the contraction ofvarious muscle groups to change with certainemotion states, reflecting the probability thatthey will need to be employed. Similarly, im-mune allocation and targeting may vary withdisgust, with the potential for injury, or withthe demands of extreme physical exertion.

Communication and Emotional Expressions

Emotion programs are expected to mobilizemany emotion-specific effects on thesubcomponents of the human psychological ar-chitecture relevant to communication. Mostnotably, many emotion programs producecharacteristic species-typical displays thatbroadcast to others the emotion state of an in-dividual (Ekman, 1982). Ekman and his col-leagues have established in a careful series oflandmark studies that many emotional expres-sions are human universals, both generated andrecognized reliably by humans everywhere theyhave been tested (Ekman, 1994). Indeed, manyemotional expressions appear to be designed tobe informative, and these have been so reliablyinformative that humans have coevolved auto-mated interpreters of facial displays of emo-tion, which decode these public displays intoknowledge of others’ mental states.

Two things are communicated by an authen-tic emotional expression:3 (1) that the associ-ated emotion program has been activated in anindividual, providing observers with informa-tion about the state of that individual’s mentalprograms and physiology (e.g., “I am afraid”);and (2) the identity of the evolutionarily recur-rent situation being faced, in the estimation ofthe signaler (e.g., the local world holds a dan-


ger). Both are highly informative, and emo-tional expressions provide a continuous com-mentary on the underlying meaning of thingsto companions. This provokes the question:Why did selection build facial, vocal, and pos-tural expressions at all? More puzzlingly, whyare they often experienced as automatic and in-voluntary? The apparent selective disadvan-tages of honestly and automatically broadcast-ing one’s emotional state have led Fridlund(1994), for example, to argue that expressionsmust be voluntary and intentional communica-tions largely unconnected to emotion state. Buteven when people deliberately lie, microex-pressions of face and voice often leak out(Ekman, 1985), suggesting that certain emo-tion programs do in fact create involuntarilyemitted signals that reliably broadcast the per-son’s emotion state and that are difficult tooverride. Why?

First, natural selection has shaped emotionprograms to signal their activation, or not, onan emotion-by-emotion basis. For each emo-tion program considered by itself (jealousy,loneliness, disgust, predatoriness, parentallove, sexual attraction, gratitude, fear), therewas a net benefit or cost to having others knowthat mental state, averaged across individualsover evolutionary time. For those recurrent sit-uations in which, on average, it was beneficialto share one’s emotion state (and hence assess-ment of the situation) with those one was with,species-typical facial and other expressions ofemotion were constructed by selection. For ex-ample, fear was plausibly beneficial to signal,because it signaled the presence of a dangerthat might menace one’s kin and cooperators aswell, and it also informed others in a way thatmight recruit assistance. Guilt was not selectedto cause a presentation with an unambiguous,distinctive signal.

Nevertheless, averaged over evolutionarytime, it was functional for the organism to sig-nal the activation of only some emotion states.The conditions favoring signaling an emotionare hard to meet (for conditions and discus-sion, see Tooby & Cosmides, 1996b; Cosmides& Tooby, 2000a). Consequently, only someemotions out of the total species-typical set areassociated with distinctive, species-typical fa-cial expressions. There should be a larger set ofemotions that have no automatic display.Moreover, emotions that lack a display are notnecessarily less fundamental or less anchored inthe evolved architecture of the human mind.

For this reason, the existence of a distinctiveexpression is not a necessary aspect of an emo-tion, nor should it be part of its definition. Jeal-ousy and guilt are both genuine emotions lack-ing distinctive signals.

Precisely because we are designed to monitorbroadcast emotions, our attention goes dispro-portionately to the subset of emotions that docome equipped with emotional expressions. Wethink it likely that this has had an impact onthe history of emotion research—specifically,that the emotions associated with distinctiveexpressions have been unnecessarily considered“primary” or “fundamental.”

Finally, many features of facial expressionsmay not just be arbitrary, but may be reliableindicators of an emotion state. Many seem tobe functional concomitants of the activity asso-ciated with the emotion (such as eyes wideningor hyperventilation). Others may be signalsthat are nonarbitrary; that is, they remove bar-riers to the correct assessment of aspects of thephenotype that the organism benefits by dem-onstrating. For example, the anger expressionmay be designed to maximize the perception ofstrength—an advertisement of a property rele-vant to the negotiation, and not just an arbi-trary signal to others that one is angry (Sell,Tooby, & Cosmides, in press-b). That is, theremay be functional reasons why the anger facehas the characteristics it does, rather than con-sisting of ear flapping or nose twitching. Simi-larly, the baring of teeth may be combat prepa-ration and advertisement (Archer, 1988); thenarrowing of the pupil may be preparation forthe detection of fast motion; and so on.

Behavior

All psychological mechanisms are involved inthe generation and regulation of behavior, soobviously behavior will be regulated by emo-tion state. More specifically, however, mecha-nisms proximately involved in the generationof actions (as opposed to such processes as facerecognition, which are only distally regulatory)should be very sensitive to emotion state. Notonly may highly stereotyped behaviors of cer-tain kinds be released (as during sexual arousalor rage, or as with species-typical facial expres-sions and body language), but more complexaction generation mechanisms should be regu-lated as well. Specific acts and courses of actionwill be more available as responses in somestates than in others, and more likely to be im-


plemented. Emotion mode should govern theconstruction of organized behavioral sequencesthat solve adaptive problems.

Biologists, psychologists, and economistswho adopt an evolutionary perspective haverecognized that game theory can be used tomodel many forms of social interactions(Maynard Smith, 1982). If the EEA imposescertain evolutionarily repeated games, then the“strategies” (the evolved cognitive programsthat govern behavior in those contexts) shouldevolve in the direction of choices that lead tothe best expected fitness payoffs. The strategyactivated in the individual should match thegame (e.g., exchange) and the state of play inthe game (e.g., having just been cheated)—aprocess that requires the system of cues, situa-tion detection, and so on, already discussed. Sodifferent emotion and inference programs orsubprograms may have evolved to correspondto various evolved games, including zero-sumcompetitive games, positive-sum exchangegames, coalitional lottery games, games of ag-gressive competition corresponding to“chicken,” and so on (for exchange, seeCosmides, 1989; Cosmides & Tooby, 1992).Corresponding emotion programs guide the in-dividual into the appropriate interactive strat-egy for the social “game” being played, giventhe state of play. Surprisingly, for some games,rigid obligatory adherence to a prior strategythroughout the game is better than the abilityto revise and change strategies (“voluntarily”)in the light of events. If an individual contem-plating a course of action detrimental to youknows you will take revenge, regardless of themagnitude of the punishment to you that thismight unleash, then that individual will be lesslikely to take such harmful action. This maytranslate into emotion programs in which thedesire to attempt certain actions should beoverwhelming, to the point where the actionsare experienced as compulsory. In the grip ofsuch programs, competing programs, includingthe normal integration of prudential concernsand social consequences, are muted or termi-nated. For example, the desire to avenge a mur-der or an infidelity is often experienced in thisway, and crimes resulting from this desire areeven culturally recognized as “crimes of pas-sion” (Daly & Wilson, 1988). In modern statesocieties, where there are police who are paidto punish and otherwise enforce agreements, itis easy to underestimate the importance thatdeterrence based on the actions of oneself and

one’s coalition had in the Pleistocene(Chagnon, 1983). Hirshleifer (1987) and Frank(1988) are evolutionary economists who havepursued this logic the furthest, arguing thatmany social behaviors are the result of such“commitment problems.”

Specialized Inference

Research in evolutionary psychology hasshown that “thinking” or reasoning is not aunitary category, but is carried out by a varietyof specialized mechanisms. So, instead of emo-tion’s activating or depressing “thinking” ingeneral, the specific emotion program activatedshould selectively activate appropriate special-ized inferential systems, such as cheater detec-tion (Cosmides, 1989; Cosmides & Tooby,1989, 1992), bluff detection (Tooby & Cosmides,1989), precaution detection (Fiddick,Cosmides, & Tooby, 2000), attributions ofblame and responsibility, and so on. For exam-ple, fear could influence precautionary reason-ing (Boyer & Liénard, 2006), competitive losscould regulate bluff detection, and so on.

Reflexes

Muscular coordination, tendency to blink,threshold for vomiting, shaking, and manyother reflexes are expected to be regulated byemotion programs to reflect the demands of theevolved situation.

Learning

Emotion mode is expected to regulate learningmechanisms. What someone learns from stim-uli will be greatly altered by emotion mode, be-cause of attentional allocation, motivation,situation-specific inferential algorithms, and ahost of other factors. Emotion mode will causethe present context to be divided up intosituation-specific, functionally appropriate cat-egories so that the same stimuli and the sameenvironment may be interpreted in radicallydifferent ways, depending on emotion state.For example, which stimuli are considered sim-ilar should be different in different emotionstates, distorting the shape of the individual’spsychological “similarity space” (Shepard,1987). Highly specialized learning mechanismsmay be activated, such as those that controlfood aversions (Garcia, 1990), predator learn-ing (Mineka & Cook, 1993), or fear condition-


ing (LeDoux, 1995). Happiness is expected tosignal the energetic opportunity for play, and toallow other exploratory agendas to be ex-pressed (Frederickson, 1998).

Mood, Energy Level, Effort Allocation,and Depression

Overall metabolic budget will be regulated byemotion programs, as will specific allocationsto various processes and facilitation or inhibi-tion of specific activities. The effort that ittakes to perform given tasks will shift accord-ingly, with things being easier or more effortful,depending on how appropriate they are to thesituation reflected by the emotion (Tooby &Cosmides, 1990a). Thus fear will make it moredifficult to attack an antagonist, whereas angerwill make it easier. The confidence with whicha situation has been identified (i.e., emotionalclarity) should itself regulate the effortfulnessof situation-appropriate activities. Confusion(itself an emotional state) should inhibit the ex-penditure of energy on costly behavioral re-sponses and should motivate more informationgathering and information analysis. Nesse(1991) has suggested that the function of moodis to reflect the propitiousness of the presentenvironment for action—a hypothesis withmany merits. We have hypothesized (Tooby &Cosmides, 1990a) a similar function of mood,based on recognizing that the action–rewardratio of the environment is not a function of theenvironment alone, but an interaction betweenthe structure of the environment and the indi-vidual’s present understanding of it. (By “un-derstanding,” we mean the correspondence be-tween the structure of the environment, thestructure of the algorithms, and the weightingsand other information they use as parameters.)The phenomenon that should regulate this as-pect of mood is a perceived discrepancy be-tween expected and actual payoff. The suspen-sion of behavioral activity accompanied byvery intense cognitive activity in depressed peo-ple looks like an effort to reconstruct models ofthe world so that future action can lead to pay-offs, in part through stripping away previousvaluations that led to unwelcome outcomes.Depression should be precipitated by (1) aheavy investment in a behavioral enterprisethat was expected to lead to large payoffs thateither failed to materialize or were not largeenough to justify the investment; or (2) insuffi-cient investment in maintaining a highly valued

person or condition that was subsequently lost(possibly as a consequence); or (3) gradual rec-ognition by situation detectors that one’s long-term pattern of effort and time expenditure hasnot led to a sufficient level of evolutionarilymeaningful reward, when implicitly comparedto alternative life paths (the condition of Dick-ens’s Scrooge). Discrepancies between expectedand actual payoff can occur in the other direc-tion as well: Joy, or a precipitated surge of hap-piness, is an emotion program that evolved torespond to the condition of an unexpectedlygood outcome. It functions to recalibrate previ-ous value states that led to underinvestment inor underexpectation for the successful activitiesor choices. Moreover, energy reserves that werebeing sequestered under one assumption aboutfuture prospects can be released, given new,more accurate expectations about a more plen-tiful or advantageous future. Similarly, one canbe informed of bad outcomes to choices notmade: For example, one may find out that acompany one almost invested in went bank-rupt, or that the highway one almost took wassnowed in. Information of this kind leads to astrengthening of the decision variables used(experienced as pleasure), which is sometimesmistaken for pleasure in the misfortune of oth-ers. Reciprocally, one can be informed of goodoutcomes to choices not made, which will beexperienced as unpleasant.

Moreover, the functional definition of emo-tion given here invites the possibility that manywell-known mental states should be recognizedas emotion states—such as the malaise engen-dered by infectious illness, coma, shock, the ap-preciation of beauty, homesickness, sexualarousal, confusion, nausea, and so on. For ex-ample, when you are sick, initiating actionsand going about your daily activities is moreeffortful than usual; your impulse is to stayhome and lie still. Although you feel as if yourenergy reserves are depleted, at a physical levelthe same fat reserves and digestively deliveredglucose are available. Malaise is a computa-tional state, not a physical one, and is designedto cope with the adaptive problem of illness: Itshunts energy from behavior to the immunesystem, and possibly signals the need for aid.Similarly, when situation-detecting algorithmsdetect the presence of a very grave internal in-jury, or the potential for one as indicated by amajor blow, these may trigger a mode of opera-tion of the psychological architecture that is de-signed to prevent any discretionary movement:


coma. The function of coma, in a world beforehospitals, was to prevent further injury frombeing done, minimize blood loss and internalhemorrhaging, and allow the mobilization ofthe body’s resources toward repair of immedi-ate threats to life. Note that a coma is not aphysically mandated state of paralysis; it is acomputational state—technically, “a state ofunconsciousness from which the patient cannotbe roused” (Miller, 1976, p. 46), or “un-arousable unresponsiveness” (Berkow, 1992, p.1398). It can occur even when there has beenno damage to the motor system.

INTERNAL REGULATORY VARIABLESAND FEELING COMPUTATION

We expect that the architecture of the humanmind, by design, is full of registers for evolvedvariables whose function is to store summarymagnitudes that are useful for regulatingbehavior and making inferences involving valu-ation. These are not explicit concepts, repre-sentations, goal states, beliefs, or desires, butrather indices that acquire their meaning viathe evolved behavior-controlling andcomputation-controlling procedures that ac-cess them. That is, each has a location embed-ded in the input–output relations of ourevolved programs, and their function inheres inthe role they play in the decision flow of thesethe programs.

For example, in our recent mapping of thearchitecture of the human kin detection system,we have identified a series of regulatory vari-ables needed to make the system work func-tionally and to explain the data (Lieberman etal., 2007). For example, for each familiar indi-vidual i, the system computes and updates acontinuous variable, the “kinship index” (Ki),which corresponds to the system’s pairwise es-timate of genetic relatedness between self and i.When the kinship index is computed or up-dated for a given individual, the magnitude istaken as input to procedures that are designedto regulate kin-relevant behaviors in a fitness-promoting way. For the case of altruism, thekinship index is fed as one of many inputs tothe “welfare tradeoff ratio estimator,” whosefunction is to compute a magnitude, the “in-trinsic welfare tradeoff ratio” (intWTRi), whichregulates the extent to which the actor is intrin-sically disposed to trade off his or her own wel-fare against that of individual i. A high kinship

index up-regulates the weight put on i‘s wel-fare, while a low kinship index has little effecton the disposition to treat i altruistically. This isone element that up-regulates the emotion oflove, attachment, or caring. Independently, thekinship index is fed as one of many inputs intothe “sexual value estimator.” Its function is tocompute a magnitude, “sexual value” (SVi),which regulates the extent to which the actor ismotivated to value or disvalue sexual contactwith individual i. As with altruism, many fac-tors (e.g., health, age, symmetry) affect sexualvalue, but a high kinship index renders sexualvaluation strongly negative, while a low kin-ship index is expected to have little effect onsexual valuation. The system takes as inputtwo cues, whose values must themselves bestored and updated as regulatory variables.The first is maternal perinatal association (i.e.,whether an older sibling observes his or hermother caring for a younger sibling as an in-fant), and the second is duration of coresidencebetween birth and the end of the period of pa-rental investment. These two cues are pro-cessed to set the value of the kinship index foreach familiar childhood companion. This sys-tem was designed by natural selection to detectwhich familiar others were close genetic rela-tives; to create a magnitude corresponding tothe degree of genetic relatedness; and then todeploy this information to motivate both a sex-ual aversion between brothers and sisters, anda disposition to behave altruistically towardsiblings.

An internal regulatory variable like the kin-ship index or the welfare tradeoff ratio acquiresits meaning and functional properties from itsrelationship to the programs that compute it,and from the downstream decisions or processesthat it regulates. The claim is not that such com-putations and their embedded variables are de-liberate or consciously accessible. We think thatthey are usually nonconscious or implicit. Out-puts of processes that access these variables maybe consciously experienced—as disgust (at theprospect of sex with a sibling), affection forthem, fear (on their behalf), grief (at their loss),and so on. Indeed, we think that it may be possi-ble eventually to arrive at a precise description ofcomputational understructure subserving theworld of feeling, by considering feeling to be aspecial form of computation that evolved to dealwith the world of valuation.

Because the computational mapping of moti-vational systems and emotion programs is a


new enterprise, at present it is difficult to knowthe full range of internal regulatory variablesthat our psychological architecture is designedto compute and access. On adaptationistgrounds, we suspect that the full set may in-clude a surprising variety of registers for spe-cialized magnitudes, corresponding to suchthings as these: how valuable a mate is, a childis, one’s own life is, and so on; how stable orvariable the food supply is over the long term;the distribution of condition-independent mor-tality in the habitat; one’s expected future life-span or period of efficacy; how good a friendsomeone has been to you; the extent of one’ssocial support; the aggressive formidability forself or others (i.e., the ability to inflict costs);the sexual value of self and others; one’s status,as well as the status of the coalition one be-longs to; present energy stores; one’s presenthealth; the degree to which subsistence requirescollective action; and so on. However, even fo-cusing on one small set of internal regulatoryvariables, welfare tradeoff ratios, offers to clar-ify the functional architecture of several emo-tions, including anger, guilt, and gratitude.

ANGER AS AN EVOLVEDREGULATORY PROGRAM

Consistent with the views of many other re-searchers, we have hypothesized that anger isan evolved emotion program with a special re-lationship to aggression. However, we thinkthat it has an equal relationship to cooperation.In the evolutionary-psychological approach tothe emotions, anger (in addition to being an ex-perienced psychological state) is the expressionof a functionally structured neurocomputa-tional system whose design features andsubcomponents evolved to regulate thinking,motivation, and behavior in the context of re-solving conflicts of interest in favor of the an-gry individual (Sell, 2005; Sell, Tooby, &Cosmides, in press-a, in press-b). Two negotiat-ing tools regulated by this system are the threatof inflicting costs (aggression) and the threat ofwithdrawing benefits (the down-regulation ofcooperation). Humans differ from most otherspecies in the number, intensity, and durationof close cooperative relationships, so tradi-tional models of animal conflict must be modi-fied to integrate the cooperative dimensionmore fully. Given its apparent functional logic,its universality across individuals and cultures

(Ekman, 1973; Brown, 1991), and its earlyontogenetic development (Stenberg, Campos,& Emde, 1983; Stenberg & Campos, 1990), itseems likely that anger is an adaptation de-signed by natural selection. If so, then its com-putational structure (i.e., what variables causeanger, what behavioral patterns are enacted byit, and what variables cause it to subside) mightbe usefully illuminated by testing predictionsderived by reference to the selection pressuresthat designed them.

Humans evolved embedded in small-scalesocial networks involving both cooperationand conflict. In many situations, each individ-ual has open to him or her a range of al-ternative behaviors that embody—as onedimension—a spectrum of possible tradeoffsbetween the individual’s own welfare and thewelfare of one or more others. By choosing onecourse of conduct, the individual is intention-ally or unintentionally expressing what can betermed a welfare tradeoff ratio with respect tothe affected party or parties. For example, anindividual might act in a way that weights thewelfare of another person slightly or not at all(e.g., being late, theft, marital abandonment,rape, burning down someone’s house for thefun of it), in a way that balances the two, or ina way that minimizes one’s own welfare by sac-rificing one’s life for the other party. In thisview, humans have a system that, in each indi-vidual, computes the welfare tradeoff ratio ex-pressed in the actions of one person toward an-other (individual i to j), and stores it as asummary characterization of i‘s disposition to-ward j in the form of a regulatory variable—thewelfare tradeoff ratio of i to j (WTRij). Indeed,there are at least two parallel, independent wel-fare tradeoff ratios: the intrinsic one (intWTR),which guides an individual’s behavior towardanother, regardless of whether his or her ac-tions are being observed; and the public one(publicWTR), which guides an individual’sbehavior when the recipient (or others) can ob-serve the behavior. Some altruism is motivatedthrough love, and some through fear, shame, orhope of reward—and the mechanisms involvedare different.

If the human mind really contains welfaretradeoff ratios as regulatory variables that con-trol how well one individual treats another,then evolution can build emotions whose func-tion is to alter welfare tradeoff ratios in otherstoward oneself. Anger is conceptualized as amechanism whose functional product is the


recalibration in the mind of another of thisother person’s welfare tradeoff ratio with re-spect to oneself. That is, the goal of the system(rather than a conscious intention) is to changethe targeted person’s disposition to make wel-fare tradeoffs so that he or she more stronglyfavors the angered individual in the present andthe future. As in animal contests, the target ofanger may relinquish a contested resource, ormay simply in the future be more careful tohelp or to avoid harming the angered individ-ual. In cooperative relationships, where there isthe expectation that the cooperative partnerwill spontaneously take the welfare of the indi-vidual into account, the primary threat fromthe angered person that potentially inducesrecalibration in the targeted individual is thesignaled possibility of the withdrawal of futurehelp and cooperation if the welfare tradeoff ra-tio is not modified. If the withdrawal of this co-operation would be more costly to the target ofthe anger than the burden of placing greaterweight on the welfare of the angry individual,then the target should increase his or her wel-fare tradeoff ratio toward the angry individual,and so treat him or her better in the future.

Reciprocally, the program is designed torecalibrate the angry individual’s own welfaretradeoff ratio toward the target of the anger fortwo functional reasons. This first is that it cur-tails the wasteful investment of cooperative ef-fort in individuals who do not respond with asufficient level of cooperation in return. Thesecond is that the potential for this downwardrecalibration functions as leverage to increasethe welfare tradeoff ratio of the target towardthe angry individual. In the absence of coopera-tion, the primary threat is the infliction of dam-age. In the presence of cooperation, the pri-mary threat is the withdrawal of cooperation.Concepts that are anchored in the internal reg-ulatory variable publicWTR include respect, con-sideration, deference, status, rank, and so on.

For example, ancestrally, one major cue thatan individual would have been able to inflictcosts to enforce welfare tradeoff ratios in his orher favor was the individual’s physical strength(as noted earlier, we call the ability to inflictcosts “formidability”). Consistent with this, inmany species the degree to which an organismvalues a nonrelative is determined primarily bythe relative strength of the two; thus animalswith higher relative strength will, when otherfactors are held constant, fight more effectivelyfor resources and have a higher expectation of

gaining a larger share of disputed resources orsocial rank. Because strength was consistentlyone factor (out of several) relevant under an-cestral conditions, and the nervous system hadreliable access to the body, it seems plausibleand worth investigating that the mind is de-signed to compute a strength self-assessmentautomatically and nonconsciously, and to usethis self-assessment as an input regulatingbehavior. Thus the human brain should haveevolved a set of programs that (1) evaluatesone’s own and others’ formidabilities; (2) trans-forms each of these evaluations into a magni-tude (a “formidability index”) associated witheach person; and, in situations where coopera-tion is not presumed, (3) implicitly expects oraccords some level of deference based on rela-tive formidability.

The approach briefly sketched above can beunpacked into a large number of empirical pre-dictions derived from this analysis of the designfeatures of the program regulating anger. Forexample, it is predicted that in humans, physi-cal strength should be a partial cause of indi-vidual differences in the likelihood of experi-encing and expressing anger. Other thingsbeing equal, stronger individuals are predictedto be more likely to experience anger and ex-press anger; they should feel more entitled; theyshould expect others to give greater weight totheir welfare, and become angrier when theydo not. Although physical strength by nomeans exhausts the set of relevant variables, itoffers an easily operationalizable and measur-able gateway into a series of tests of this gen-eral model of the logic underlying the regula-tion of anger. Arguments precipitated by angershould reflect the underlying logic of the wel-fare tradeoff ratio: The complainant will em-phasize the cost of the other’s transgression tohim or her, as well as the value of the complain-ant’s cooperation to the transgressor, and willfeel more aggrieved if the benefit the transgres-sor received (the justification) is small com-pared to the cost inflicted. A series of empiricalstudies supports both sets of predictions of thistheory about the design of anger (Sell, 2005;Sell et al., in press-a, in press-b).

RECALIBRATIONAL EMOTIONSSUCH AS GUILT AND GRATITUDE

The EEA was full of event relationships (e.g.,“Mother is dead”) and psychophysical regular-


ities (e.g., “Blood indicates injury”) that cuedreliable information about the functionalmeanings and properties of things, events, per-sons, and regulatory variables to the psycho-logical architecture. For example, certain bodyproportions and motions indicated immaturityand need, activating the emotion program ofexperiencing cuteness (see Eibl-Ebesfeldt,1970). Others indicated sexual attractiveness(Symons, 1979; Buss, 1994). To be moved withgratitude, to be glad to be home, to see some-one desperately pleading, to hold one’s new-born baby in one’s arms for the first time, to seea family member leave on a long trip, to en-counter someone desperate with hunger, tohear one’s baby cry with distress, to be warmwhile it is storming outside—these all meansomething to us. How does this happen? In ad-dition to the situation-detecting algorithms as-sociated with major emotion programs such asfear, anger, or jealousy, we believe that humanshave a far larger set of evolved specializations,which we call “recalibrational releasing en-gines.” These are activated by situation-detecting algorithms, and their function is totrigger appropriate recalibrations, includingaffective recalibrations, when certainevolutionarily recognizable situations are en-countered.

We believe that the psychophysical or inter-pretive “front ends” of emotion programs usethese cues not only to trigger the appropriateemotion, but to alter the weightings of regula-tory variables embedded in decision rules. (Forexample, if you experience someone treatingyou disrespectfully, it makes you angry.) In-deed, most evolutionarily recurrent situationsthat selected for corresponding emotion pro-grams bristle with information that allows therecomputation of one or more variables.Recalibration (which, when consciously acces-sible, appears to produce rich and distinct feel-ing states) is therefore a major functional com-ponent of most emotion programs. Jealousy,for example, involves several sets ofrecalibrations (e.g., diminution in estimate ofone’s own mate value, diminution of trust, low-ering of the welfare tradeoff ratio toward themate).

Indeed, from an evolutionary-psychologicalperspective, recalibrational emotion programsappear to be the dominant (but not the only)components of such emotions as guilt, grief,depression, shame, and gratitude. Their pri-mary function is not to orchestrate any short-

run behavioral response (as fear or anger do),but instead to carry out valuation recom-putations in the light of the new informationrelevant to evolved regulatory variables that isprovided by external or internal environments(Tooby & Cosmides, 1990a). An evolutionaryviewpoint is a utilitarian one, which suggeststhat the time humans spend simply feeling—attending inwardly not to factual representa-tions, but to something else—is doing some-thing useful that will be reflected eventually inbehavior. The hypothesis is that feeling is aform of computational activity that takes timeand attention, that can compete with or pre-empt motivation to engage in other activities,and whose function is to recalculate andreweight the regulatory variables implicated bythe newly encountered information. This ap-proach has the potential to provide an accountof the characteristics of emotions such as guiltor depression, which appear otherwise puz-zling from a functional perspective. The feel-ings these emotion programs engender interferewith short-term utilitarian action that an activeorganism might be expected to engage in. Ifthey were not useful, the capacity to feel themwould have been selected out.

Consider guilt: We believe that guilt func-tions as an emotion mode specialized forrecalibration of regulatory variables that con-trol tradeoffs in welfare between self and other(Tooby & Cosmides, 1990a). Three importantreasons why humans evolved to take the wel-fare of others into account are genetic related-ness toward relatives (Hamilton, 1964), thepositive externalities others emit (Tooby &Cosmides, 1996a), and the maintenance of co-operative relationships (Trivers, 1971; Tooby& Cosmides, 1996a). The regulatory variableapproach provides a clear framework for un-derstanding why guilt evolved and what its un-derlying logic is. In this view, guilt involves therecalibration of regulatory variables consideredwhen one is making decisions about tradeoffsin welfare between the self and others, basedon new information about actual or potentialharm arising from having placed too littleweight on the other person’s welfare in past ac-tions. Kin selection would favor a mechanismdesigned to effect such recalibration towardthose the kin detection mechanism identifies asclose genetic relatives. Similarly, individualshave an intrinsic interest in the welfare of thosewhose existence benefits them, and with whomthey share deep engagement relationships


(Tooby & Cosmides, 1996a). Third, reciprocal,exchange, or cooperative relationships need tobe proximately motivated, so that benefit flowsare appropriately titrated. Individuals who ex-perienced guilt (and the associated modifica-tion of decision rules) would have been lesslikely to injure relationship partners repeatedly,and they would have had more success in main-taining beneficial cooperative relationships.

In the case of kin selection, we now have anempirical map of the architecture of theneurocomputational program that detects ge-netic relatedness and passes this information tothe welfare tradeoff system (Lieberman et al.,2007). The theory of kin selection says noth-ing, however, about the procedures by which amechanism could estimate the value of, say, aparticular piece of food to oneself and one’skin. The fitness payoffs of such acts of assis-tance vary with circumstances. Consequently,each decision about where to allocate assis-tance depends on inferences about the relativeweights of these variables. These nonconsciouscomputations (however they are carried out)must be subject to error, selecting for feedbacksystems of correction.

Imagine a hunter–gatherer woman with asister. The mechanisms in the woman’s brainhave been using the best information availableto her to weight the relative values of the meatshe has been acquiring to herself and her sister,leaving her reassured that it is safe to leave hersister for a while without provisioning her. Thesudden discovery that her sister, since she waslast contacted, has been starving and hasbecome desperately sick functions as aninformation-dense situation allowing therecalibration of the algorithms weighting therelative values of the meat to self and sister(among other things). The sister’s sicknessfunctions as a cue that the previous allocationweighting was in error and that the variablesneed to be reweighted—including all of theweightings embedded in habitual action se-quences that might be relevant to the sister’swelfare. Guilt should be triggered when the in-dividual receives (1) unanticipated informationabout the welfare of a valued other (or the in-creased value of the other), indicating that (2)the actor’s actions or omissions caused or al-lowed the welfare of the valued individual to bedamaged in a way that is inconsistent with theactor’s ideal welfare tradeoff ratio, given (3) theactor’s resources and potential for action.

When guilt is triggered, the welfare tradeoff ra-tio is adjusted, as well as a variety of subsidiaryvariables expressing this ratio in action. As aresult of this recalibration, the guilty individ-ual’s behavior should reflect this higher valua-tion. In cases where the effects were inten-tional and anticipated, there should be littlerecalibration.

Existing findings substantiate these predic-tions and explain some of their otherwise puz-zling features. Unsurprisingly, when the valuedother is negatively affected unexpectedly, sub-jects feel guiltier (Baumeister, Stillwell, &Heatherton, 1995, Kubany & Watson, 2003).More surprisingly, individuals feel guiltierwhen the harm was caused accidentally ratherthan anticipated, even though individuals areusually considered less responsible and culpa-ble for the harm when it occurs accidentally(McGraw, 1987; Baumeister et al., 1995). If thefunction of guilt is, however, to recalibrate animproperly set welfare tradeoff ratio, then in-formation that merely confirms the evaluationpresent in the decision requires no recal-ibration. If the effect was foreseen and chosenanyway in the light of the existing ratio, thenno adjustment is necessary.

Gratitude is a recalibrational emotion pro-gram that is complementary to guilt. Guiltturns up the welfare tradeoff ratio toward anindividual when one has evidence that one’sown actions have expressed too low a valua-tion of the other. Gratitude is triggered by newinformation indicating that another places ahigher value on one’s welfare than one’s systemhad previously estimated—again leading to anup-regulation of the WTR toward that person.Anger, guilt, and gratitude all play differentroles in cooperation, and their computationalstructure reflects their recalibrational functionswith respect to welfare tradeoff ratios and thechoice points they involve.

The evolutionary-psychological stance moti-vating the investigation of the program archi-tecture of the emotions suggests that the emo-tions are intricate, functionally organized, andsensitively related to the detailed structure ofancestral problems. In this view, the emotionsare likely to be far more sophisticated engineer-ing achievements than previously appreciated,and there are many decades of work ahead foremotion researchers before they are compre-hensively mapped.


NOTES

1. Marks (1987, pp. 68–69) vividly conveys how manyaspects of behavior and physiology may be entrainedby certain kinds of fear:

During extreme fear humans may be “scared stiff” or“frozen with fear.” A paralyzed conscious state withabrupt onset and termination is reported by survivorsof attacks by wild animals, by shell-shocked soldiers,and by more than 50% of rape victims (Suarez & Gal-lup, 1979). Similarities between tonic immobility andrape-induced paralysis were listed by Suarez & Gallup(features noted by rape victims are in parentheses): (1)profound motor inhibition (inability to move); (2)Parkinsonian-like tremors (body-shaking); (3) silence(inability to call out or scream); (4) no loss of con-sciousness testified by retention of conditioned reac-tions acquired during the immobility (recall of detailsof the attack); (5) apparent analgesia (numbness andinsensitivity to pain); (6) reduced core temperature(sensation of feeling cold); (7) abrupt onset and termi-nation (sudden onset and remission of paralysis); (8)aggressive reactions at termination (attack of the rap-ist after recovery); (9) frequent inhibition of attack bya predator . . .

2. We think that some emotion programs evolved in re-sponse to the situation cue provided by a strong drivestate, such as hunger, when the motivational intensityreached a point that other mechanisms became domi-nated and entrained by the magnitude of the motiva-tion. We see no principled reason for distinguishingstrong drive states from other emotion programs,and suspect that this practice originated from out-dated notions of natural selection that separated“survival-related” functions (hunger, thirst) fromother functions, such as mate acquisition or reciproc-ity. Thus we propose that it is useful to model special-ized motivational states as emotion programs, just asone would disgust, anger, or fear.

3. The evolutionary purpose of deceitful emotional ex-pressions is to (falsely) communicate the same twothings.

REFERENCES

Archer, J. (1988). The behavioural biology of aggres-sion. Cambridge, UK: Cambridge University Press.

Barkow, J., Cosmides, L., & Tooby, J. (Eds.). (1992).The adapted mind: Evolutionary psychology and thegeneration of culture. New York: Oxford UniversityPress.

Baron-Cohen, S. (1995). Mindblindness: An essay onautism and theory of mind. Cambridge, MA: MITPress.

Baumeister, R., Stillwell, A., & Heatherton, T. (1995).Personal narratives about guilt. Basic and AppliedSocial Psychology, 17(1–2), 173–198.

Berkow, R. (Ed.). (1992). The Merck manual of diagno-sis and therapy (16th ed.). Rahway, NJ: Merck.

Brown, D. (1991). Human universals. New York:McGraw-Hill.

Boyer, P., & Liénard, P. (2006). Why ritualized behav-ior? Precaution systems and action parsing in devel-opmental, pathological, and cultural rituals. Behav-ioral and Brain Sciences, 12(6), 595–613.

Buss, D. M. (1994). The evolution of desire. New York:Basic Books.

Buss, D. M. (Ed.). (2005). The handbook of evolution-ary psychology. Hoboken, NJ: Wiley.

Cannon, W. (1929). Bodily changes in pain, hunger, fearand rage: Researches into the function of emotionalexcitement. New York: Harper & Row.

Chagnon, N. (1983). Yanomamo: The fierce people (3rded.). New York: Holt, Rinehart & Winston.

Cosmides, L. (1989). The logic of social exchange: Hasnatural selection shaped how humans reason?Studies with the Wason selection task. Cognition, 31,187–276.

Cosmides, L., & Tooby, J. (1987). From evolution tobehavior: Evolutionary psychology as the missinglink. In J. Dupre (Ed.), The latest on the best: Essayson evolution and optimality (pp. 276–306). Cam-bridge, MA: MIT Press.

Cosmides, L., & Tooby, J. (1989). Evolutionary psy-chology and the generation of culture: Part II. Casestudy: A computational theory of social exchange.Ethology and Sociobiology, 10, 51–97.

Cosmides, L., & Tooby, J. (1992). Cognitive adapta-tions for social exchange. In J. Barkow, L. Cosmides,& J. Tooby (Eds.), The adapted mind: Evolutionarypsychology and the generation of culture (pp. 163–228). New York: Oxford University Press.

Cosmides, L., & Tooby, J. (1997). Dissecting the com-putational architecture of social inference mecha-nisms. In G. Bock & G. Cardeco (Eds.), Character-izing human psychological adaptations (CibaSymposium No. 208, pp. 132–156). Chichester, UK:Wiley.

Cosmides, L., & Tooby, J. (2000). Evolutionary psy-chology and the emotions. In M. Lewis & J. M.Haviland-Jones (Eds.), Handbook of emotions (2nded., pp. 91–115). New York: Guilford Press.

Crawford, C., & Krebs, D. (Eds.). (1998). Handbook ofevolutionary psychology. Mahwah, NJ: Erlbaum.

Daly, M., & Wilson, M. (1988). Homicide. New York:Aldine.

Daly, M., Wilson, M., & Weghorst, S. J. (1982). Malesexual jealousy. Ethology and Sociobiology, 3, 11–27.

Derryberry, D., & Tucker, D. (1994). Motivating the fo-cus of attention. In P. M. Neidenthal & S. Kitayama(Eds.), The heart’s eye: Emotional influences in per-ception and attention (pp. 167–196). San Diego, CA:Academic Press.

Eibl-Ebesfeldt, I. (1970). Ethology: The biology ofbehavior. New York: Holt, Rinehart & Winston.

Ekman, P. (1973). Cross-cultural studies of facial ex-pression. In P. Ekman (Ed.), Darwin and facial ex-


pression: A century of research in review (pp. 169–222). New York: Academic Press.

Ekman, P. (Ed.). (1982). Emotion in the human face(2nd ed.). Cambridge, UK: Cambridge UniversityPress.

Ekman, P. (1985). Telling lies. New York: Norton.Ekman, P. (1994). Strong evidence for universals in fa-

cial expressions. Psychological Bulletin, 115, 268–287.

Ekman, P., Levenson, R., & Friesen, W. (1983). Auto-nomic nervous system activities distinguishes be-tween emotions. Science, 221, 1208–1210.

Fiddick, L., Cosmides, L., & Tooby, J. (2000). No inter-pretation without representation: The role ofdomain-specific representations and inferences in theWason selection task. Cognition, 77, 1–79.

Frank, R. (1988). Passions within reason: The strategicrole of the emotions. New York: Norton.

Fredrickson, B. (1998). What good are positive emo-tions? Review of General Psychology, 2, 300–319.

Fridlund, A. (1994). Human facial expression: An evo-lutionary view. San Diego, CA: Academic Press.

Garcia, J. (1990). Learning without memory. Journal ofCognitive Neuroscience, 2, 287–305.

Gigerenzer, G., & Murray, D. (1987). Cognition as intu-itive statistics. Hillsdale, NJ: Erlbaum.

Gould, S. J. (1997, June 12). Darwinian fundamental-ism. New York Review of Books, pp. 34–38.

Hamilton, W. D. (1964). The genetical evolution of so-cial behaviour. Journal of Theoretical Biology, 7, 1–52.

Haselton, M. G., & Buss, D. M. (2003). Biases in socialjudgment: Design flaws or design features? In J.Forgas, K. Williams, & B. von Hippel (Eds.), Re-sponding to the social world: Implicit and explicitprocesses in social judgments and decisions (pp. 23–43). New York: Cambridge University Press.

Hirshleifer, J. (1987). On the emotions as guarantors ofthreats and promises. In J. Dupre (Ed.), The latest onthe best: Essays on evolution and optimality (pp.307–326). Cambridge, MA: MIT Press.

Kubany, E., & Watson, S. (2003). Guilt: Elaboration ofa multidimensional model. Psychological Record, 53,51–90.

Lazarus, R., & Lazarus, B. (1994). Passion and reason.New York: Oxford University Press.

LeDoux, J. (1995). In search of an emotional system inthe brain: Leaping from fear to emotion to conscious-ness. In M. S. Gazzaniga (Ed.), The cognitive neuro-sciences (pp. 1049–1061). Cambridge, MA: MITPress.

Lieberman, D., Tooby, J., & Cosmides, L. (2007). Thearchitecture of human kin detection. Nature, 445,727–731.

Marks, I. (1987). Fears, phobias, and rituals. New York:Oxford University Press.

Maynard Smith, J. (1982). Evolution and the theory ofgames. Cambridge, UK: Cambridge University Press.

McGraw, K. (1987). Guilt following transgression: an

attribution of responsibility approach. Journal ofPersonality and Social Psychology, 53(2), 247–256.

Miller, S. (Ed.). (1976). Symptoms: The complete homemedical encyclopedia. New York: Crowell.

Mineka, S., & Cook, M. (1993). Mechanisms involvedin the observational conditioning of fear. Journal ofExperimental Psychology: General, 122, 23–38.

Mowat, R. R. (1966). Morbid jealousy and murder: Apsychiatric study of morbidly jealous murders atBroadmoor. London: Tavistock.

Nesse, R. (1991). Evolutionary explanations of emo-tions. Human Nature, 1, 261–289.


Pitman, R., & Orr, S. (1995). Psychophysiology of emo-tional and memory networks in posttraumatic stressdisorder. In J. McGaugh, N. Weinberger, & G. Lynch(Eds.), Brain and memory: Modulation and media-tion of neuroplasticity (pp. 75–83). New York: Ox-ford University Press.

Raymond, E. S. (1991). The new hacker’s dictionary.Cambridge, MA: MIT Press.

Sell, A. (2005). Regulating welfare trade-off ratios:Three tests of an evolutionary–computational modelof human anger. Doctoral dissertation, Departmentof Psychology, University of California, SantaBarbara.

Sell, A., Tooby, J., & Cosmides, L. (in press-a). The logicof anger: Formidability regulates human conflict.

Sell, A., Tooby, J., & Cosmides, L. (in press-b). Compu-tation and the logic of anger.

Shepard, R. N. (1987). Evolution of a mesh betweenprinciples of the mind and regularities of the world.In J. Dupre (Ed.), The latest on the best: Essays onevolution and optimality (pp. 251–275). Cambridge,MA: MIT Press.

Shepherd, M. (1961). Morbid jealousy: Some clinicaland social aspects of a psychiatric symptom. Journalof Mental Science, 107, 687–753.

Stenberg, C. R., & Campos, J. J. (1990). The develop-ment of anger expression in infancy. In N. Stein, B.Leventhal, & T. Trabasso (Eds.), Psychological andbiological approaches to emotion (pp. 247–282).Hillsdale, NJ: Erlbaum.

Stenberg, C. R, Campos, J. J., & Emde, R. N. (1983).The facial expression of anger in seven-month-old in-fants. Child Development, 54(1), 178–184.

Suarez, S. D., & Gallup, G. G. (1979). Tonic immobilityas a response to rage in humans: A theoretical note.Psychological Record, 29, 315–320.

Swets, J. A., Tanner, W. D., & Birdsall, T. G. (1961). De-cision processes in perception. Psychological Review,68, 301–340.

Symons, D. (1979). The evolution of human sexuality.Oxford: Oxford University Press.

Tomaka, J., Blascovich, J., Kibler, J., & Ernst, J. (1997).Cognitive and physiological antecedents of threatand challenge appraisal. Journal of Personality andSocial Psychology, 73, 63–72.


Tooby, J. (1985). The emergence of evolutionary psy-chology. In D. Pines (Ed.), Emerging syntheses in sci-ence: Proceedings of the founding workshops of theSanta Fe Institute (pp. 1–6). Santa Fe, NM: Santa FeInstitute.

Tooby, J., & Cosmides, L. (1989, August). The logic ofthreat. Paper presented at the meeting of the HumanBehavior and Evolution Society, Evanston, IL.

Tooby, J., & Cosmides, L. (1990a). The past explainsthe present: emotional adaptations and the structureof ancestral environments. Ethology and Socio-biology, 11, 375–424.

Tooby, J., & Cosmides, L. (1990b). On the universalityof human nature and the uniqueness of the individ-ual: The role of genetics and adaptation. Journal ofPersonality, 58, 17–67.

Tooby, J., & Cosmides, L. (1992). The psychologicalfoundations of culture. In J. Barkow, L. Cosmides, &J. Tooby (Eds.), The adapted mind: Evolutionarypsychology and the generation of culture (pp. 19–136). New York: Oxford University Press.

Tooby, J., & Cosmides, L. (1996a). Friendship and thebanker’s paradox: Other pathways to the evolutionof adaptations for altruism. In W. G. Runciman,J. M. Smith, & R. I. M. Dunbar (Eds.), Evolu-tion of Social Behaviour Patterns in Primates andMan. Proceedings of the British Academy, 88, 119–143.

Tooby, J., & Cosmides, L. (1996b). The computa-tionalist theory of communication. Paper presentedat the meeting of the Human Behavior and Evolu-

tion Society Meetings, University of Arizona,Tucson.

Tooby, J., & Cosmides, L. (1997). On evolutionary psy-chology and modern adaptationism: A reply to Ste-phen Jay Gould. Available online at www.psy-ch.ucsb.edu/research/cep/critical_eye.htm

Tooby, J., & Cosmides, L. (2005). Conceptual founda-tions of evolutionary psychology. In D. M. Buss(Ed.), The handbook of evolutionary psychology (pp.5–67). Hoboken, NJ: Wiley.

Tooby, J., & Cosmides, L. (in press). Internal regulatoryvariables and the design of human motivation: Anevolutionary and computational approach. In A.Elliot (Ed.), Handbook of approach and avoidancemotivation. Mahwah, NJ: Erlbaum.

Tooby, J., Cosmides, L., & Barrett, H. C. (2003). Thesecond law of thermodynamics is the first law of psy-chology: Evolutionary developmental psychologyand the theory of tandem, coordinated inheritances.Psychological Bulletin, 129(6), 858–865.

Tooby, J., Cosmides, L., Sell, A., Lieberman, D., &Sznycer, D. (in press). Internal regulatory variablesand the design of human motivation: An evolution-ary and computational approach. In A. Elliot (Ed.),Handbook of approach and avoidance motivation.Mahwah, NJ: Erlbaum.

Trivers, R. L. (1971). The evolution of reciprocal altru-ism. The Quarterly Review of Biology, 46, 35–37.

Williams, G. C. (1966). Adaptation and natural selec-tion: A critique of some current evolutionarythought. Princeton, NJ: Princeton University Press.


C H A P T E R 9

The Role of Emotionin Economic Behavior

SCOTT RICK and GEORGE LOEWENSTEIN

IMMEDIATE ANDEXPECTED EMOTIONS

Consequentialist Modelsof Decision Making

Economic models of decision making areconsequentialist in nature; they assume that de-cision makers choose between alternativecourses of action by assessing the desirabilityand likelihood of their consequences, and inte-grating this information through some type ofexpectation-based calculus. Economists referto the desirability of an outcome as its “utility,”and decision making is depicted as a matter ofmaximizing utility.

This does not, however, imply that con-sequentialist decision makers are devoid ofemotion or immune to its influence. To seewhy, it is useful to draw a distinction between“expected” and “immediate” emotions(Loewenstein, Weber, Hsee, & Welch, 2001;Loewenstein & Lerner, 2003). Expected emo-tions are those that are anticipated to occur asa result of the outcomes associated with differ-ent possible courses of action. For example, ifLaura, a potential investor, were deciding

whether to purchase a stock, she might imaginethe disappointment she would feel if shebought it and it declined in price, the elationshe would experience if it increased in price,and possibly emotions such as regret and reliefthat she might experience if she did not pur-chase the stock and its price either rose or fell.The key feature of expected emotions is thatthey are experienced when the outcomes of adecision materialize, but not at the moment ofchoice, at the moment of choice they are onlycognitions about future emotions.

Immediate emotions, by contrast, are experi-enced at the moment of choice and fall into oneof two categories. “Integral” emotions, like ex-pected emotions, arise from thinking about theconsequences of one’s decision, but integralemotions, unlike expected emotions, are expe-rienced at the moment of choice. For example,in the process of deciding whether to purchasethe stock, Laura might experience immediatefear at the thought of the stock’s losing value.“Incidental” emotions are also experienced atthe moment of choice, but arise fromdispositional or situational sources objectivelyunrelated to the task at hand (e.g., the TV pro-

138

gram playing in the background as Lauracalled her brokerage house).1

The notion of expected emotions is perfectlyconsistent with the consequentialist perspectiveof economics. Nothing in the notion of utilitymaximization rules out the idea that the utilityan individual associates with an outcome mightarise from a prediction of emotions; for exam-ple, one might assign higher utility to an Italianrestaurant dinner than a French restaurant din-ner because one anticipates being happier atthe former. While not explicitly denying theidea that utilities might depend on expectedemotions, however, most economists until re-cently viewed detailed accounts of such emo-tions as outside the purview of their discipline.

Integral immediate emotions can also be in-corporated into a consequentialist framework,although it takes one farther afield from con-ventional economics. Integral emotions, it canand in fact has been argued, might provide de-cision makers with information about theirown tastes—for instance, to help inform Lauraof how she would actually feel if she purchasedthe stock and it rose or declined in value. How-ever, this assumes, contrary to the usual as-sumption in economics, that people have animperfect understanding of their own tastes.

An influence of incidental immediate emo-tions on decision making would pose a muchmore fundamental challenge to the consequen-tialist perspective, because such emotions, bydefinition, are irrelevant to the decision athand. Any influence of incidental emotionswould suggest that decisions are influenced byfactors unrelated to the utility of their conse-quences.

Figure 9.1 presents a schematic representa-tion of the traditional perspective of econom-ics. Although immediate emotions are repre-sented in the figure, they would not be part ofany traditional economist’s representation oftheir framework, because they play no role indecision making; they are “epiphenomenal”by-products of, but not determinants of, deci-sions.

However, a great deal of market activity canbe understood in terms of both expected andimmediate emotions. Much advertising at-tempts to inform consumers, whether accu-rately or not, about emotions that they can ex-pect to feel if they do or do not buy a particulargood. “One-day-only” sales, for example, areprobably effective because they make consum-ers think that they will regret not seizing the

opportunity. Marketers also attempt to cap-italize on immediate emotions—for example,charitable organizations that make potentialdonors feel guilty about what they squandertheir money on while less fortunate peoplestarve.

The food industry is particularly motivatedto capitalize on immediate emotions. Mrs.Field’s Cookies, for example, has been knownto pump enticing cookie smells into the atmo-sphere of shopping malls to stimulate hunger(Hoch & Loewenstein, 1991). A companynamed “ScentAir” sells similar odors (e.g.,“Glazed Donut,” “Iced Cinnamon Pretzel,”“Blue Cotton Candy”) to businesses looking tostimulate hunger.2 By contrast, the dieting in-dustry often attempts to market its services byfocusing people on the positive emotions theycan anticipate experiencing once they are fi-nally able to fit into the perfect pair of jeans.

Enter Behavioral Economics

Fortunately, many economists would view thesnapshot of their discipline presented above asoutdated. This is largely attributable to theadvent of “behavioral economics,” a subdis-cipline of economics that incorporates morepsychologically realistic assumptions to in-crease the explanatory and predictive power ofeconomic theory. The field first achieved prom-inence in the 1980s and has been gaining influ-ence since then. And much of the thrust ofbehavioral economics has involved, or at leastcould be construed as involving, an enhancedunderstanding of emotions.

The first, and less controversial, interactionof behavioral economics with emotions was toquestion the neglect of the topic and to begin toexamine exactly how utility depended on out-

9. The Role of Emotion in Economic Behavior 139

FIGURE 9.1. Consequentialist model of decisionmaking.

comes. For example, whereas conventionaleconomics assumes that the utility of an out-come depends only on the outcome itself, someeconomists showed how counterfactual emo-tions (e.g., regret), which arise from consider-ing alternative outcomes that could have oc-curred, can influence decision making. Notethat these analyses focus on expected emotionsand hence help to elaborate the connectionamong outcomes, emotions, and utility, but donot challenge the consequentialist perspective.

More recently, economists as well as psy-chologists who are specifically interested in de-cision making have begun to take greater ac-count of immediate emotions. Some of theresearch has shown that immediate integralemotions play a critical role in decision mak-ing. However, other research has shown thatimmediate emotions, and especially but not ex-clusively incidental emotions, often propel de-cisions in different directions from expectedemotions—that is, in directions that run con-trary to the predictions of a consequentialistperspective. The new research thus suggeststhat the consequentialist perspective is muchtoo simple to be a descriptively valid account ofactual behavior.

In this chapter, we review some of the critical(consequentialist) assumptions and predictionsof the dominant economic models of risky deci-sion making, intertemporal choice, and socialpreferences. For each of these areas, we firstdiscuss behavioral phenomena that are anoma-lous from the consequentialist perspective, butthat are rectified once the role of expected emo-tions is taken into account. Next, we discussphenomena that can potentially be illuminatedby taking account of immediate emotions, bothintegral and incidental. We conclude by pro-posing directions for future research on the roleof emotion in decision making.

DECISION MAKING UNDER RISK

Most decisions, including decisions of eco-nomic importance, entail an element of risk,because the consequences of alternative coursesof action are rarely known with certainty. Thusdecision making under risk is a central topic ineconomics.

Since first proposed by Daniel Bernoulli(1738/1954), the “expected utility” (EU)model has served as the normative benchmarkfor decision making under risk in economics.

EU assumes that people choose between alter-native courses of action by assessing thedesirability or “utility” of each action’s possi-ble outcomes and linearly weighting those utili-ties by their probability of occurring. The nor-mative status of the EU model was enhanced byvon Neumann and Morgenstern’s (1944) dem-onstration that it could be derived from a prim-itive, intuitively appealing set of axioms—forexample, that preferences are transitive (if A ispreferred to B, and B is preferred to C, then Ashould be preferred to C). In addition to itsnormative appeal, this model’s assumption thatdecisions are based on EU, rather than ex-pected value, gives it descriptive appeal as well.For instance, it assumes that the difference inhappiness (i.e., utility) between winning $1 andwinning $2 is not necessarily equal to the dif-ference in happiness between winning $101and winning $102 (though the difference invalue is equal).

However, empirical research has docu-mented many behavioral phenomena that areinconsistent with the basic axioms, and thus in-consistent with the predictions of the EUmodel, and many of these anomalies can be at-tributed to unrealistic assumptions about thedeterminants of expected emotions and the in-fluence of immediate emotions. Several modelshave accounted for some of these anomalies bymaking more realistic assumptions about thedeterminants of expected emotions. We nextreview some of these theoretical innovations.We then discuss anomalies that can potentiallybe explained by taking account of the influenceof immediate emotions.

Innovations to the EU ModelInvolving Expected Emotions

Relaxing the Asset Integration Assumption

In its original form, the EU model assumes thatpeople do not narrowly focus on potential out-comes when making a decision, but rather onhow those outcomes affect their overall wealth.Thus the utility of a particular outcome is notsimply based on that outcome, but instead onthe integration of that outcome with all assetsaccumulated to that point. However, as origi-nally noted by Markowitz (1952) and devel-oped more fully by Kahneman and Tversky(1979), people typically make decisions with anarrower focus. When evaluating the potentialoutcomes of a decision, people tend to think in


terms of incremental gains and losses, ratherthan in terms of changes in overall welfare.

Suppose, for example, that Bob must decidewhether to accept or reject a gamble that offersa 50% chance of winning $20 and a 50%chance of losing $10. If Bob currently possesses$1 million in wealth, then the EU model as-sumes that he views the gamble as offering a50% chance of experiencing the utility of$1,000,020 and a 50% chance of experiencingthe utility of $999,990. Markowitz (1952) ar-gued, however, that most people would insteadprocess the gamble as it was presented, namelyas offering a 50% chance of experiencing theutility of winning $20 and a 50% chance of ex-periencing the disutility of losing $10.3

Relaxing the Assumption That Utility Is StrictlyDefined over Realized Outcomes

Another problematic assumption of the EUmodel is that unrealized outcomes do not influ-ence how we feel about realized outcomes. Forexample, suppose you anticipate a pay raise of$10,000 and subsequently receive a $5,000raise. Although the raise is a gain relative to thestatus quo, you will likely code it as a loss,since it fails to meet expectations. Indeed,Koszegi and Rabin (2006) have recently pro-posed a model assuming that gains and lossesare defined relative to expectations, rather thanthe status quo.

Additionally, several modifications of the EUmodel incorporate the tendency to comparewhat happens to what was expected to happen(e.g., Loomes & Sugden, 1986; Mellers,Schwartz, Ho, & Ritov, 1997). Other theoriesattempt to account for regret, a counterfactualemotion that arises from a comparison be-tween the outcome one experiences as a conse-quence of one’s decision and the outcome onecould have experienced as a consequence ofmaking a different choice. Early versions of re-gret theory (e.g., Loomes & Sugden, 1982) pre-dicted that regret aversion could lead to viola-tions of fundamental axioms of the EU model,such as monotonicity (i.e., stochastically domi-nating gambles are preferred to the gamblesthey dominate).

Regret can also lead to violations of transi-tivity. Consider, for example, the three gamblesbelow. Assume that there are three equallylikely states of nature; the table lists what eachgamble pays if a particular state of nature is re-alized. If people care more about one big regret

than they do about two smaller ones, as as-sumed in Loomes and Sugden (1982), thenGamble A will be preferred to Gamble B. Simi-larly, B is likely to be preferable to C. Since A ispreferred to B, and B is preferred to C, thentransitivity requires that A is preferred to C.However, in fact C is preferred to A, sincechoosing A over C exposes one to the risk ofone large regret instead of two small ones.

State 1 State 2 State 3Gamble A $10 $20 $30Gamble B $20 $30 $10Gamble C $30 $10 $20

Disappointment aversion and regret aver-sion theories have only met with modest empir-ical support. One problem with the predictivevalidity of regret aversion theories may be thatanticipated regret only influences decisionmaking when the possibility of regret is salient(Zeelenberg & Beattie, 1997; Zeelenberg, Beat-tie, van der Plight, & De Vries, 1996). Con-sider, for example, the following gambles, inwhich one of four colors can be drawn withvarying probability:

Gamble A Gamble B90% chance of White,

which pays $090% chance of White,

which pays $06% chance of Red,


which pays $451% chance of Green,


which pays –$103% chance of Yellow,

which pays -$152% chance of Yellow,

which pays –$15

Since Green wins $30 in Gamble A and loses$10 in Gamble B, choosing B could produce re-gret if Green is drawn. This very salient poten-tial for regret could lead to a preference for Aover B, even though such a preference violatesmonotonicity. However, the gambles can be re-written to make the possibility of regret less sa-lient:

Gamble A Gamble B90% chance of White,

which pays $090% chance of White,





which pays $451% chance of Blue,

which pays –$151% chance of Blue,



which pays –$15


Note that Gambles A and B′ are equivalent toGambles A and B, respectively; A and A′ bothhave an expected value of $2.55, and B and B′both have an expected value of $2.75. How-ever, the potential for regret is no longer sa-lient. Rather, B′ pays at least as much as A′ foreach possible color. Thus, even though A andA′ are equivalent, A′ is likely to be less attrac-tive than A, only because the way A′ and B′ areframed obfuscates the potential for regret.4

However, note that regret is often moresalient in prospect than in retrospect.5

Consider, for example, a study by Gilbert,Morewedge, Risen, and Wilson (2004) that ex-amined the extent to which subway passengersregretted missing their train. Passengers whoentered a subway station within 6 minutes ofmissing the train (experiencers) were told thatthey missed their train by either 1 minute or 5minutes. They were then asked to report howmuch regret they felt. These ratings were com-pared to the ratings of passengers leaving thestation (forecasters), who were asked to imag-ine how much regret they would feel if theymissed their train by 1 or 5 minutes. Fore-casters anticipated feeling greater regret if theymissed their train by 1 minute than by 5 min-utes, though actual regret did not depend onhow close experiencers came to catching thetrain. A subsequent study suggested that the ef-fect was driven by forecasters’ inability to real-ize how quickly they would absolve themselvesof responsibility for the disappointing out-come.

Although work remains to be done to incor-porate more determinants of expected emo-tions into consequentialist models of decisionmaking under risk, great progress has beenmade. We now discuss risky choice phenomenadriven by immediate emotions.

Innovations to the EU ModelInvolving Immediate Emotions

Integral Emotions Influence Risky Decision Making

When sufficiently strong, immediate emotionscan directly influence behavior, completelyprecluding cognitive decision making(Loewenstein, 1996). Ariely and Loewenstein(2005) experimentally examined the influenceof sexual arousal on (hypothetical) risky deci-sion making (see also Loewenstein, Nagin, &Paternoster, 1997). Male participants weregiven a laptop computer and asked to answer a

series of questions. In the control treatment,participants answered the questions while intheir natural (presumably not highly aroused)state. In the arousal treatment, participantswere first asked to self-stimulate (masturbate)while viewing erotic photographs, and werepresented with the same questions only afterthey had achieved a high but suborgasmic levelof arousal. When asked about their intention touse birth control in the future, aroused partici-pants were less likely to report intending to usea condom. Although arousal affected partici-pants’ risk attitudes, it did not affect their riskperception. For example, aroused participantswere no less likely to endorse this statement:“If you pull out before you ejaculate, a womancan still get pregnant.” Although the authorsdid not ask questions that would permitmediational analyses, the preliminary resultssuggest that immediate emotions had a directeffect on (predicted) behavior.

When experienced at more moderate levels,however, affect can mediate the relationshipbetween cognition and behavior. AntonioDamasio and his colleagues (Damasio, 1994;Bechara, Damasio, Tranel, & Damasio, 1997)have argued that decision makers encode theconsequences of alternative courses of actionaffectively, and that such “somatic markers”critically influence decision making. Damasioand colleagues have further argued that theventromedial prefrontal cortex (VMPFC) playsa critical role in this affective encoding process.Bechara et al. (1997) investigated the proposedrole of the VMPFC in an experiment in whichpatients suffering damage to the VMPFC andnon-brain-damaged individuals played a gamein which the objective was to win as muchmoney as possible. Players earned hypotheticalmoney by turning over cards that yielded eithermonetary gains or losses. On any given turn,players could draw from one of four decks, twoof which included $100 gains and two of whichcontained $50 gains. The high-paying decksalso included a small number of substantiallosses, resulting in a net negative expectedvalue for these decks. Bechara et al. (1997)found that both nonpatients and those withVMPFC damage avoided the high-payingdecks immediately after incurring substantiallosses. However, individuals with VMPFCdamage resumed sampling from the high-paying decks more quickly than nonpatientsdid after encountering a substantial loss. Thus,even though patients understood the game and


wanted to win, they often went “bankrupt.”Bechara et al. (1997) reasoned that patients“knew” the high-paying decks were risky, butthat their failure to experience fear when con-templating sampling from these decks maderisky draws more palatable.6 While the Becharaet al. (1997) study has not been immune to crit-icism (see Maia & McClelland, 2004, for aparticularly compelling critique, and Dunn,Dalgleish, and Lawrence, 2005, for a review ofseveral critiques), the somatic marker hypothe-sis remains intuitively appealing.

Other evidence suggesting that integral emo-tion influences decision making comes fromstudies of consumers’ willingness to insureagainst a variety of risks. Johnson, Hershey,Meszaros, and Kunreuther (1993), for exam-ple, asked participants how much they wouldbe willing to pay for flight insurance that pro-tected against death due to “any act of terror-ism” or “any reason.” Since terrorism is onlyone of many reasons why a plane might crash,consequentialist models of decision makingpredict that participants will pay more for in-surance covering all types of crashes than forinsurance just covering terrorism. However,Johnson et al. (1993) found that participantswere willing to pay slightly more for insuranceprotecting against terrorism.7

Additional evidence of integral emotions’impact on risky decision making comes fromstudies of probability weighting. The EU modelassumes that the weight an outcome’s probabil-ity receives in decision making is independentof the outcome; in fact, the model assumes lin-ear probability weighting (i.e., that outcomesare weighted in exact proportion to their likeli-hood of occurring). However, more recentmodels of decision making under risk havechallenged this assumption, suggesting insteadthat probabilities are weighted nonlinearly, asin Figure 9.2 (Kahneman & Tversky, 1979).Kahneman and Tversky’s (1979) proposed“probability-weighting function” suggests thatsmall probabilities are overweighted and largeprobabilities are underweighted.

Despite the innovation, models such asKahneman and Tversky’s (1979) still assumethat probability weights are independent ofoutcomes. This suggests, for example, that a1% chance of losing $1 has the same psycho-logical impact as a 1% chance of losing yourlife. Rottenstreich and Hsee (2001) suggest thatthe probability-weighting function is flatter foraffect-rich outcomes than for affect-poor out-

comes. They speculate that affect-rich prizeselicit greater degrees of hope and fear, and thusan extreme overweighting of small probabili-ties and an extreme underweighting of largeprobabilities. Indeed, Rottenstreich and Hsee(2001) found that participants’ willingness topay to avoid an electric shock was insensitiveto the probability of the shock, whereas will-ingness to pay to avoid losing $20 was ex-tremely sensitive to the probability of the loss.8

Incidental Emotion InfluencesRisky Decision Making

In a study of market index returns across 26countries from 1982 to 1997, Hirshleifer andShumway (2003) found that the amount ofsunshine (relative to expected amount of sun-shine for a given time of year) was positivelyand significantly correlated with market re-turns. The authors speculate that the phenome-non may be driven by incorrect attributions ofgood mood to positive economic prospectsrather than correct attributions to the sunshine(cf. Schwarz & Clore, 1983). Similarly,Edmans, García, and Norli (2007) have foundthat stock market returns plummet when acountry’s soccer team is eliminated from an im-portant tournament (e.g., the World Cup).They also document a dip in market returnsfollowing important losses in other sports (e.g.,


FIGURE 9.2. Kahneman and Tversky’s prob-ability-weighting function. From Kahneman andTversky (1979). Copyright 1979 by the Econo-metric Society. Reprinted by permission.

cricket, rugby, and hockey) in countries wherethose sports are popular.9

INTERTEMPORAL CHOICE

Models of intertemporal choice address howdecision makers choose between alternativesinvolving costs and benefits that are distributedover time. The “discounted utility” (DU)model is the dominant model of intertemporalchoice in economics (Samuelson, 1937). Struc-turally, this model is closely parallel to the EUmodel—and, like the EU model, has been de-rived from a series of intuitively compelling ax-ioms (Koopmans, 1960). However, a numberof anomalies have been identified that call intoquestion the descriptive validity of these axi-oms, and thus the predictions of the DU model(Loewenstein & Prelec, 1992). We next reviewanomalies that can be reconciled with thismodel once more realistic assumptions aremade about the determinants of expected emo-tions; we then discuss anomalies that can be ex-plained by taking account of immediate emo-tions.

Innovations to the DU ModelInvolving Expected Emotions

Relaxing the Assumption That UtilityIs Strictly Defined over Realized Outcomes

Like the EU model, the DU model assumesthat utility (and thus expected emotion) isonly a function of realized outcomes. If peo-ple devalue future emotions, they shouldwant to experience pleasurable outcomes im-mediately and postpone painful outcomeswhenever possible. However, contrary to thisbasic assumption, in many situations peopleprefer to get unpleasant outcomes over withquickly, or to “leave the best for last.” In anearly study documenting this phenomenon,Loewenstein (1987) asked 30 undergraduateshow much they would be willing to pay im-mediately to obtain a kiss from the moviestar of their choice and to avoid receiving a(nonlethal) 110-volt shock, after several timedelays. Contrary to the predictions of the DUmodel, respondents were willing to pay moreto experience a kiss delayed by 3 days thanan immediate kiss or one delayed by 3 hoursor 1 day, and were also willing to pay moreto avoid a shock that was delayed for 1 year

or 10 years than to avoid a shock experi-enced within the next 3 days.

These anomalies can be reconciled with theDU model if one takes account of the observa-tion that utility is not strictly a function of real-ized outcomes, but also of emotions experi-enced while waiting for those outcomes tooccur. Loewenstein (1987) proposes that peo-ple derive utility from “savoring” future goodoutcomes and disutility from dreading bad out-comes.10 Indeed, in a brain imaging study inwhich participants were confronted with theprospect of a real impending shock, Berns et al.(2006) found that components of the brain’s“pain matrix” (a cluster of regions that are ac-tivated during the experience of pain) are alsoactive in anticipation of shock. Furthermore,providing support for the idea that utility fromanticipation plays a causal role in the desire toexpedite negative outcomes, individual differ-ences in activation in response to anticipatorypain predict individual tendencies to expediteshocks.11

Incorporating Affective Forecasting Errors

For the DU model to be descriptively valid, peo-ple must be able to forecast accurately how theywill react emotionally to future outcomes.However, there is by now substantial evidencethat people have difficulty making such fore-casts. Consider, for instance, a study byBrickman, Coates, and Janoff-Bulman (1978) inwhich lottery winners, persons with paraplegiaor quadriplegia, and a control group were askedto report their current happiness on a 5-pointscale. The lottery group (n = 22) consisted ofpeople who had recently won at least $50,000 inthe Illinois state lottery. The paraplegic andquadriplegic participants (n = 29) had becomeparalyzed within the past year. Lottery winnersreported a mean level of happiness virtuallyidentical to that of the control group (4.00 vs.3.82), whose happiness was significantly differ-ent from, but surprisingly close to, the meanhappiness level among paraplegic and quadri-plegic participants (2.96). Although the lotterywinners and the paraplegic and quadriplegicparticipants were not prospectively asked topredict their future happiness (since they couldnot be identified beforehand), it seems likelythat both groups would have overestimated thehedonic impact of their future circumstances.12

Loewenstein and Adler (1995) examinedwhether people could predict falling subject to


the “endowment effect” (Thaler, 1980), whichrefers to the tendency for people to value anobject more highly if they possess it than theywould value the same object if they did not. Inthe typical demonstration of the effect (see,e.g., Kahneman, Knetsch, & Thaler, 1990),some participants (sellers) are endowed with anobject and given the option of trading it forvarious amounts of cash; other participants(choosers) are not given the object, but aregiven a series of choices between receiving theobject and receiving various amounts of cash.Although the objective wealth position of thetwo groups is identical, as are the choices theyface, endowed participants hold out for signifi-cantly more money than those who are not en-dowed. Loewenstein and Adler (1995) in-formed some participants that they would beendowed with an object (a mug engraved withtheir school logo) and asked them to predictthe price at which they would sell the objectback to the experimenter once they were en-dowed. These participants, and others who didnot make a prediction, were then endowedwith the object and given the opportunity tosell it back to the experimenter. Participantswho were not yet endowed substantiallyunderpredicted their own postendowment sell-ing prices. In a second study, selling prices wereelicited from participants who were actuallyendowed with an object and from others whowere told they had a 50% chance of getting theobject. Selling prices were substantially higherfor the former group, and the valuations ofparticipants who were not sure of getting theobject were indistinguishable from the buyingprices of participants who did not have the ob-ject.

Loewenstein and Adler’s (1995) results sug-gest that participants who were not endowedwith an object failed to predict how painful itwould be to part with the object once they pos-sessed it. That is, non-endowed participantsmade “affective forecasting” errors when pre-dicting their future attachment to the object.However, a recent study by Kermer, Driver-Linn, Wilson, and Gilbert (2006) suggests thatit may be the sellers who are making the affec-tive forecasting error (see also Galanter, 1992).Kermer et al. (2006) first asked participants toreport their baseline affect. Participants thenreceived a $5 show-up fee and were told that acoin would be flipped to determine whetherthey would win an additional $3 or lose $2.Next, they predicted how they would feel im-

mediately after the coin toss. The experimenterthen flipped a coin and paid participants ac-cordingly. Participants then rated how they feltat that moment. Some participants were alsoasked to report what they would think after thecoin toss, and once the coin had actually beentossed, they were asked to report their actualthoughts. Kermer et al. (2006) found that peo-ple expected losing $3 to diminish their happi-ness (relative to happiness reported at thebeginning of the experiment) more than it actu-ally did.13 This suggests that the predictors inLoewenstein and Adler (1995) may have accu-rately based their predicted selling prices onhow they would actually feel after losing an ob-ject. Sellers, by contrast, may have based theirselling prices on unrealistically negative fore-casts of how they would feel after losing an ob-ject.14

In a behavioral economic model of inter-temporal choice that incorporates affectiveforecasting errors, Loewenstein, O’Donoghue,and Rabin (2003) propose that people exagger-ate the degree to which their future tasteswill resemble their current tastes. Conlin,O’Donoghue, and Vogelsang (2007) find evi-dence of such “projection bias” in catalog or-ders of cold-weather-related clothing items andsports equipment. People are overinfluenced bythe weather at the time they make decisions, asmeasured by their likelihood of returning theitem: A decline of 30° F on the date an item isordered increases the probability of a return by3.95%.

Economists have incorporated more realisticassumptions about expected emotions intomodels of intertemporal choice. However, somephenomena, driven by immediate emotions, re-main anomalous from the perspective of suchmodels. We now turn to these phenomena.

Innovations to the DU ModelInvolving Immediate Emotions

Relaxing the Assumption of Exponential Discounting

The DU model assumes that people discountfuture flows of utility at a fixed discount ratebased on when the utility will be experienced.Discounting at a fixed rate (i.e., “exponential”discounting) means that a given time delayleads to the same amount of discounting re-gardless of when it occurs. According to theDU model, delaying the delivery of a good by 1day leads to the same degree of time discount-


ing whether that delay makes the difference be-tween consuming the good tomorrow ratherthan today or in 101 days rather than 100days. However, an overwhelming amount ofempirical work suggests that people (as well asanimals) do not discount the future exponen-tially (Kirby & Herrnstein, 1995; Rachlin &Raineri, 1992). Rather, people care more aboutthe same time delay if it is proximal rather thandistal—a general pattern that has been refereedto as “hyperbolic time discounting” (Ainslie,1975). For example, delaying consumption of apleasurable good from today to tomorrow ismore distressing than delaying consumptionfrom 100 days from now to 101 days fromnow. Hyperbolic time discounting predicts thatpeople will behave farsightedly when the con-sequences of their decision are delayed. In suchsituations, decision makers will place greatweight on long-term costs and benefits. How-ever, when consequences are immediate, hyper-bolic time discounting will produce behaviorthat appears impulsive.15

Consider, for example, an experiment byRead, Loewenstein, and Kalyanaraman (1999)in which participants were asked to select 1 of24 movies to watch. Some of the movies were“highbrow” (e.g., Schindler’s List), and somewere “lowbrow” (e.g., The Mask). Some par-ticipants were asked to choose a movie towatch that night, whereas others were asked tochoose a movie to watch in the future. Consis-tent with hyperbolic discounting, “lowbrow”movies (ones that are high in short-run bene-fits, but low in long-run benefits) were mostpopular among participants selecting a moviefor immediate viewing.16

Behavioral economists have made greatprogress in modeling hyperbolic discounting(e.g., Laibson, 1997). Such models implicitlyassume that discounting leads to impulsivebehavior by diminishing the importance of ex-pected emotions. However, when the timing ofconsumption is held constant, various other sit-uational factors can also lead to impulsivity.Walter Mischel (1974) and colleagues, for ex-ample, have extensively studied the impact ofphysical proximity of rewards on the impulsivi-ty of children. Children faced with the choicebetween a small immediate reward (e.g., onemarshmallow immediately or two marshmal-lows in 15 minutes) and a larger delayed re-ward (two marshmallows) tend to behave moreimpatiently when the immediate reward is visi-ble.

Thus impulsivity may reflect factors otherthan a devaluation of expected emotions. Im-mediate emotions may also produce non-exponential discounting. To examine the influ-ence of immediate emotions on impulsivity,McClure, Laibson, Loewenstein, and Cohen(2004) measured the brain activity of partici-pants with functional magnetic resonance im-aging (fMRI) while they made a series ofintertemporal choices between small proximalrewards ($R available at delay d) and larger de-layed rewards ($R′ available at delay d′), where$R < $R′ and d < d′. Rewards ranged from $5to $40 Amazon.com gift certificates, and thedelay ranged from the day of the experiment to6 weeks later. McClure et al. (2004) investi-gated whether there were brain regions thatshowed elevated activation (relative to a restingstate benchmark) only when immediacy was anoption (i.e., activation when d = 0, but no acti-vation when d > 0), and whether there were re-gions that showed elevated activation whenparticipants were making any intertemporaldecision irrespective of delay. McClure et al.(2004) found that time discounting is associ-ated with the engagement of two neural sys-tems. Limbic and paralimbic cortical struc-tures, which are known to be rich indopaminergic innervation, were preferentiallyrecruited for choices involving immediatelyavailable rewards. In contrast, fronto-parietalregions, which support higher cognitive func-tions, were recruited for all intertemporalchoices. Moreover, the authors found thatwhen choices involved an opportunity for im-mediate reward, thus engaging both systems,greater activity in fronto-parietal regions thanin limbic regions was associated with choosinglarger delayed rewards.17 These results suggestthat the experience of immediate emotionrather than the devaluation of expected emo-tion may, at least in some situations, driveimpulsivity.18

Integral Emotions Influence Intertemporal Choice

Suppose you are deciding whether or not tobuy a CD for $10. The DU model predicts thatyou will buy the CD if the anticipated pleasureof listening to it exceeds its “opportunity cost”(i.e., the forgone pleasure that could have beenpurchased with the $10). However, Frederick,Novemsky, Wang, Dhar, and Nowlis (2006)suggest that people do not spontaneously con-sider opportunity costs when deciding whether


or not to purchase goods. Frederick et al.(2006) asked participants whether they would(hypothetically) be willing to purchase a desir-able video for $14.99. They simply variedwhether the decision not to buy it was framedas “not buy this entertaining video” or “keepthe $14.99 for other purchases.” Although thetwo phrases described objectively equivalentactions, the latter highlighted the pleasure thatwould be forgone by purchasing the video.Frederick et al. (2006) found that drawing at-tention to opportunity costs significantly re-duced the proportion of participants willing topurchase the video, suggesting that many par-ticipants were not spontaneously consideringopportunity costs (cf. Jones, Frisch, Yurak, &Kim, 1998).

If many people do not take opportunity costsinto account when deciding whether or not topurchase goods, then how do they make suchdecisions? In a project wiht Brian Knutson,Elliott Wimmer, and Drazen Prelec (Knutson,Rick, Wimmer, Prelec, & Loewenstein, 2007),we investigated this question in an experimentin which participants chose whether or not topurchase a series of discounted consumergoods while having their brains scanned withfMRI. The goods ranged in retail price from$10 to $80, and were offered at a 75% dis-count to encourage spending. Participants weregiven $20 to spend and were told that one oftheir decisions would be randomly selected tocount for real. At the conclusion of the experi-ment, participants indicated how much theyliked each product and how much they wouldbe willing to pay for it.

We found that the extent to which partici-pants reported liking the products correlatedpositively with activation in nucleusaccumbens, a target of dopaminergic projec-tions that has previously been associated withanticipation of gains and self-reported happi-ness (Knutson, Adams, Fong, & Hommer,2001). Moreover, consumer surplus (i.e., thedifference between self-reported willingness topay for the good and its price) correlated posi-tively with activation in medial prefrontal cor-tex, a region previously associated with the re-ceipt of unexpectedly large gains (e.g.,Knutson, Fong, Bennett, Adams, & Hommer,2003). We also found that activation in bothregions correlated positively with purchasingdecisions. However, we found that activationin insula during the period when subjects firstsaw the price correlated negatively with pur-

chasing decisions. Insula activation haspreviously been observed in connection withaversive stimuli such as disgusting odors(Wicker et al., 2003), unfairness (Sanfey,Rilling, Aronson, Nystrom, & Cohen, 2003),and social exclusion (Eisenberger, Lieberman,& Williams, 2003). Thus when the delayedcosts of immediage indulgence are not explic-itly represented (as in, e.g., McClure et al.,2004), but rather implicitly captured by prices,participants appear to rely on an anticipatory“pain of paying” (Prelec & Loewenstein, 1998)to curtail their spending.

Incidental Emotions Influence Consumer Choice

In other research conducted wiht CynthiaCryder (Rick, Cryder, & Loewenstein, 2008)investigated whether individuals chronicallydiffered in their tendency to experience antici-patory pain when making purchasing deci-sions. We hypothesized that individuals whotypically experience an intense pain of payingmay generally spend less than they would ide-ally like to spend, whereas individuals who ex-perience minimal pain of paying may typicallyspend more than they would ideally like tospend. We developed a “Spendthrift–Tight-wad” scale to measure individual differences inthe pain of paying and found that tightwadsoutnumbered spendthrifts by a 3:2 ratio in asample of more than 13,000 people. Rick(2007) hypothesized that incidental sadnesscould help both tightwads and spendthriftsovercome their prepotent affective responses tospending. The hypothesis was based on previ-ous experimental work suggesting that sadnessdeepens deliberation (e.g., Tiedens & Linton,2001) and motivates people to change their cir-cumstances (e.g., Lerner et al., 2004). Rick(2007) tested the hypothesis in an experimentin which tightwads and spendthrifts decidedwhether or not to purchase a variety of goodswhile listening to neutral or sad music. As pre-dicted, tightwads spent more when sad thanwhen in a neutral state, and spendthrifts spentless when sad than when in a neutral state.

SOCIAL PREFERENCES

Although there are widely accepted normativebenchmarks for risky decision making andintertemporal choice, no such benchmarks existfor how people should behave toward others.


However, many economic models make the sim-plifying, but unrealistic, assumption that peopleare strictly self-interested. Below we reviewbehavioral economic models of social prefer-ences that have incorporated more realistic as-sumptions about the determinants of expectedemotions in social interactions. We then reviewanomalies driven by immediate emotions.

Expected Emotion: Relaxingthe Pure Self-Interest Assumption

Economists frequently study social preferencesin the context of the “ultimatum game” (Guth,Schmittberger, & Schwarze, 1982). In the typi-cal ultimatum game, a “proposer” offers someportion of an endowment to a “responder,”who can either accept the offer or reject it. Ifthe responder accepts the offer, the money is di-vided according to the proposed split. If the re-sponder rejects the offer, both players leavewith nothing. Since purely self-interested re-sponders should accept any positive offer, self-interested proposers should offer no more thanthe smallest positive amount possible. How-ever, average offers typically exceed 30% of thepie, and offers of less than 20% are frequentlyrejected (see Camerer, 2003).

Several behavioral economic models haveemerged that incorporate a taste for fairness.19

Rabin (1993) proposes a model in which peo-ple derive utility from reciprocating intentional(un)kindness with (un)kindness (see alsoDufwenberg & Kirchsteiger, 2004). Blount(1995) conducted an interesting variant of theultimatum game to investigate the role of inten-tions in social behavior. Some responders weretold that the proposer with whom they werepaired would make an offer, as in the standardultimatum game. Other responders were toldthat the offer would be randomly generated.Blount (1995) found that responders were will-ing to accept significantly less when the offerwas generated randomly than when it camefrom the proposer.

Sanfey, Rilling, Aronson, Nystrom, and Co-hen (2003) conducted a similar study in whichparticipants played the ultimatum game whilehaving their brains scanned with fMRI. Partici-pants, all responders, were told they wouldplay the ultimatum game with 10 different hu-man proposers (though offers were actually de-termined by the experimenters). Responders re-ceived five “fair” offers ($5 for proposer, $5 forrespondent) and five unfair offers. In 10 other

trials, responders received the same offer, butthis time from a computer. As in Blount (1995),participants were more willing to accept lowoffers from computer proposers than from hu-man proposers. Moreover, activation in ante-rior insula, a region commonly implicated inthe experience of pain (e.g., Knutson et al.,2007), was greater for unfair offers from hu-man proposers than for fair offers from humanproposers. Insula activation was also signifi-cantly greater in response to unfair offers fromhuman proposers than in response to unfair of-fers from computer proposers. In fact, whetherplayers rejected unfair offers from human pro-posers could be predicted reliably by the levelof their insula activity. Thus it appears that in-tegral emotions influence responders’ behav-ior in the ultimatum game (cf. Pillutla &Murninghan, 1996).

Behavioral economists have created moredescriptively valid models of social prefer-ences by relaxing the assumption of pure self-interest. However, some phenomena driven byimmediate emotion cannot be explained bysuch models. We review such anomalies below.

Integral Emotions InfluenceSocial Preferences

Recent work on the “identifiable-victim effect”(Small & Loewenstein, 2003), which refers tothe tendency to give more to identifiable vic-tims than to statistical victims, suggests that in-tegral emotions play a role in generosity to-ward others (see also Schelling, 1968; Kogut &Ritov, 2005). Subsequent research has demon-strated that people are also more punitive to-ward identifiable wrongdoers than towardequivalent but unidentified wrongdoers, andthat anger mediates the effect of identifiabilityon punishing behavior (Small & Loewenstein,2005). To capture these phenomena, as well asa variety of experimental findings, Loewensteinand Small (2007) have proposed a dual-processmodel of helping behavior in which a sympa-thetic but highly immature emotional systeminteracts with a more mature but uncaring de-liberative system.

Incidental Emotions InfluenceSocial Preferences

Andrade and Ariely (2006) investigated the im-pact of incidental emotions on behavior in theultimatum game. They induced either inciden-


tal happiness or anger, and then had partici-pants play the role of responder in an ultima-tum game in which they were offered $4 of a$20 endowment. After deciding whether to ac-cept or reject the offer, participants then playedthe role of proposer in a second ultimatumgame, with a presumably different partner thanin the first game. Andrade and Ariely (2006)found that happy responders were less likelythan angry responders to reject unfair offers inthe initial ultimatum game. Surprisingly, how-ever, proposers who were initially induced tofeel happy made more selfish proposals in thesecond ultimatum game. The authors reasonedthat angry individuals, who were more likely toreject unfair offers than happy individuals inthe initial ultimatum game, misattributed theirbehavior to stable preferences rather than to in-cidental affect. Later, due to a “false-consensuseffect” (Ross, Greene, & House, 1977; but seeDawes & Mulford, 1996), the previously angryindividuals inferred that others would also belikely to reject unfair offers and therefore, asproposers, made very generous offers. By con-trast, the authors reasoned that happy individ-uals, who were less likely to reject unfair offersthan angry individuals in the initial ultimatumgame, also misattributed their behavior to sta-ble preferences. Accordingly, previously happyindividuals inferred that others would also beunlikely to reject unfair offers and therefore, asproposers, made very selfish offers.20

CONCLUSION

As the foregoing review indicates, emotionsinfluence economic behavior in two distinctways. First, people anticipate, and take into ac-count, how they are likely to feel about the po-tential consequences of alternative courses ofaction. As discussed, such a role for expectedemotions is entirely consistent with consequen-tialist economic accounts of decision making.Research on the role of expected emotions indecision making has taken a variety of direc-tions. It has assessed the types of emotions thatpeople actually experience when different out-comes are realized, with a special focus oncounterfactual emotions. It has examined peo-ple’s predictions of what emotions they will ex-perience, and the accuracy of such predictions.And, it has sought to determine the degree towhich decisions are in fact guided by predictedemotions.

Second, substantial research supports theidea that immediate emotions also play an im-portant role in decision making. Integral imme-diate emotions arise from contemplating thepotential outcomes of a decision. In somecases, these emotions seem to play a beneficialrole in decision making, informing decisionmakers about their own values. But in othercases, such as the disproportionate fear com-monly associated with flying as opposed todriving, integral emotions may cause people toact contrary to their own material interests. Incontrast to the mixed role played by integralemotions, it is much more difficult to justify thewell-documented role of incidental emotions,which by definition are unrelated to the deci-sion at hand.

In general, research on expected emotions isfar more advanced than that on immediateemotions. As a result, there is a pressing needfor more research to examine the causes andconsequences of immediate emotions, and tounderstand the complex interplay of immediateand expected emotions in the production ofbehavior. In some cases, immediate and ex-pected emotions seem to complement one an-other. This is true, for example, when immedi-ate emotions provide decision makers with abetter understanding of their own values—anunderstanding that may help them to betterpredict their own future feelings. For instance,the experience of anticipatory guilt may helpstudents who are contemplating cheating on anexam to appreciate the guilt they would experi-ence after doing so. In other cases, however,immediate and expected emotions come intoconflict. For example, the immediate effect of apositive mood may be to make decision makersmore inclined to take risks—but, by a different,consequentialist path, a positive mood mightalso make decision makers more risk-avoidant,with the goal of not risking a disturbance to thepositive feelings (Isen, Nygren, & Ashby, 1988;Kahn & Isen, 1993).

The clash between immediate and expectedemotions is also a major cause of self-controlproblems. For example, people are often drivenby immediate emotions to eat, drink, and makemerry, but in some of these situations, contem-plation of expected emotional consequencesmay discourage indulgence. Psychologists havefor decades been developing “dual-process”models that can be interpreted in such terms(see Evans, 2008, for a review), and in recentyears economists have begun to follow their


lead. Thaler and Shefrin (1981) were the firsteconomists to do so; their model adopts aprincipal–agent framework, in which a far-sighted planner (the principal) attempts to rec-oncile the competing demands of a series ofmyopic doers (the agents). More recently, manydual-process models have focused on the prob-lem of self-control (Brocas & Carrillo, 2006;Fudenberg & Levine, 2006; Benhabib & Bisin,2005; Loewenstein & O’Donoghue, 2004;Bernheim & Rangel, 2004).

Although most of the dual-process modelsproposed by economists have sought to adhereto the standard consequentialist perspective,introducing a role for immediate emotionsshould raise questions about whether such aperspective is “up to the job” of providing auseful account of human behavior. Behaviorunder the control of immediate emotions bearslittle resemblance to the reflective weighing ofcosts and benefits that is the prototype of ratio-nal economic decision making. Instead, it is amuch more reflexive process that often drivesbehavior in exactly the opposite direction fromthat suggested by a weighing of costs and bene-fits. Whether behavior driven by immediateemotions even warrants the label of “decisionmaking” seems questionable.

In closing, we note two potential (and, webelieve, fruitful) directions for future researchon the role of emotion in decision making. Thefirst is the need to study stronger emotions thanhave generally been examined in the empiricalliterature. Many vitally important decisions aremade “in the heat of the moment,” and indeedimportant economic decisions such as majorpurchases often evoke powerful emotions. Butstudying the impact of such emotions isdifficult—in part because it is difficult if notimpossible to manipulate such strong emo-tional states experimentally, and in part be-cause people generally do not like to be studiedwhen they are in heightened emotional states.Gaining a better understanding of the roleof immediate emotions in economic decisionmaking, therefore, is going to require research-ers who are willing to extend themselves into“hot” situations and creative enough to findnatural experiments in which people are natu-rally assigned to different emotional states be-fore they make important decisions.

The second pressing need is for economic re-search that takes fuller account of the range ofinsights that psychologists are developing intoemotions. Thus, for example, economists

studying the impact of weather on the stockmarket have generally taken a rather simplisticview—that bad weather should lead to nega-tive emotions, which should in turn lead tonegative price movements. But psychologistsstudying the impact of emotions on risk takingfind that different specific negative emotionscan have very different effects. More relevantto the central theme of this chapter, that theyhave found negative emotions can exert con-flicting effects on risk taking, depending onwhether the mechanism is consequentialist ormore reflexive.

Economists’ understanding of the role ofemotions in economic behavior has made enor-mous strides in recent decades. However, thereis still a long distance to go.

NOTES

1. Note that the distinction between expected and im-mediate emotions closely maps onto other com-monly discussed distinctions in economics and psy-chology, such as the broad distinction betweencognition and emotion, or Adam Smith’s (1759/1981) distinction between the “impartial spectator”and the “passions.”

2. On the surface, it seems somewhat unethical to arti-ficially induce visceral states in order to sell prod-ucts. However, food companies that failed to preyon the affective vulnerability of consumers wouldprobably be driven out of business by other compa-nies that did. Hence one could argue that food com-panies that pump artificial smells into the atmo-sphere to stimulate hunger are not evil, but ratherare doing what they must to stay afloat.

3. Note that narrowly focusing on gains and lossesrather than on changes in overall welfare suggeststhat all people, regardless of their current wealthposition, view gambles the same way. Indeed, such anarrow focus may explain why some extraordi-narily wealthy individuals take big risks to achievesmall gains and avoid small losses (e.g., MarthaStewart, worth hundreds of millions of dollars, en-gaged in insider trading to avoid a loss of less than$50,000).

4. As Sugden (1986) notes, another problem withregret-aversion models may be that it is recrimina-tion—regret accompanied by the feeling that oneshould have behaved differently—rather than regretthat one cares about and attempts to avoid. Sup-pose, for example, you take your car to your regularmechanic, Sue, for an oil change. You have neverhad a problem with this mechanic’s work, but thistime she uses the wrong type of oil, which causes thecar to break down. In this situation, you surely re-gret that the mistake was made, but you probably


do not blame yourself for taking it to Sue, since youhad no reason to anticipate such a mistake based onher past performance. Now suppose that you in-stead had decided to change your own oil. You havenever done so before, but you decide it is worth try-ing to save the money. Your inexperience leads youto use the wrong type of oil, causing the car to breakdown. As in the previous scenario, you regret thatthe mistake was made. However, now there is likelyto be recrimination as well: You think that youshould have known better than to try to changeyour own oil.

5. Interestingly, however, Kivetz and Keinan (2006)show that regret from choosing virtues over vices in-creases over time, whereas regret from choosingvices over virtues diminishes over time.

6. Note that the extent to which emotional deficits leadto poor decision making depends largely on situa-tional factors. In a similar study in which riskychoices had a higher expected value than risklesschoices, Shiv, Loewenstein, Bechara, Damasio, andDamasio (2005) found that patients with damage tobrain regions associated with processing emotionearned more than control participants.

7. One natural explanation for these results is that“unpacking” vivid subsets of a larger set provides amore effective retrieval cue when people are recall-ing past causes of plane crashes (e.g., Tversky &Koehler, 1994). Such an account would be consis-tent with a consequentialist model of decision mak-ing that allows for errors in judging probabilities.However, other work suggests that this resultshould not be interpreted in purely cognitive terms.Slovic, Fischhoff, and Lichtenstein (1980), for ex-ample, speculated that people’s willingness to insurethemselves against unlikely losses may be related tohow much these potential losses cause worry orconcern. Consistent with this view, a number ofstudies have shown that knowing someone who hasbeen in a flood or earthquake, or having been in oneoneself, greatly increases the likelihood of purchas-ing insurance (Browne & Hoyt, 2000). Althoughthis finding, like that of Johnson et al. (1993), couldbe explained in consequentialist terms as resultingfrom an increase in individuals’ expectations of ex-periencing a flood or earthquake in the future, theeffect remains significant even after subjective ex-pectations are controlled for (Kunreuther et al.,1978).

8. Similarly, Ditto, Pizarro, Epstein, Jacobson, andMacDonald (2006) conducted an experiment inwhich participants were given the opportunity toplay a game that would either result in winningchocolate chip cookies or being required to work ona boring task for an extra 30 minutes. Half of theparticipants were only told about the cookies,whereas for the other half the cookies were freshlybaked in the lab and placed in front of the partici-pants as they decided whether or not to play thegame. Consistent with Rottenstreich and Hsee

(2001), Ditto et al. (2006) found that participants’willingness to play the game was insensitive to theprobability of winning cookies when the cookieswere baked in the lab, whereas willingness to playwas very sensitive to the probability of winningwhen the cookies were merely described.

9. Also, Lerner and Keltner (2001) find thatdispositional (i.e., incidental) anger and fear haveopposing effects on risk preferences. Specifically, an-gry people tend to prefer risk (see also Fessler,Pillsworth, & Flamson, 2004), whereas fearful peo-ple tend to avoid it. The authors explain their resultsin terms of the cognitive appraisals generated by theemotions (Smith & Ellsworth, 1985). Anger is gen-erally associated with appraisals of certainty and in-dividual control, whereas fear is generally associ-ated with appraisals of uncertainty and situationalcontrol. These incidental emotions, through theirassociated appraisals, appear to influence partici-pants’ cognitive evaluations of the problem, thus in-fluencing their subsequent decisions.

10. Loewenstein’s model applies only to deterministicoutcomes (e.g., a guaranteed kiss from a movie starin the future). Caplin and Leahy (2001) note thatmany anticipatory emotions (e.g., suspense) aredriven by uncertainty about the future. They pro-pose a model that modifies the EU model to incor-porate such anticipatory emotions, and then showthat it can explain a variety of phenomena (e.g., theoverwhelming preference for riskless bonds overstocks).

11. In addition to savoring and dread, an entirely differ-ent type of anticipation may also drive inter-temporal choice: the anticipation of memories(Elster & Loewenstein, 1992). For example, peoplemay perform challenging but unpleasant activities(e.g., mountain climbing) partly because they savorthe pleasant memories of conquering the challenge(see also Keinan, 2006).

12. Addressing an important limitation of the Brickmanet al. (1978) study, Gilbert et al. (1998) conducted astudy in which affective forecasts could be elicitedprior to an important life event. Specifically, Gilbertet al. (1998) studied assistant professors’ forecastsof how they would feel after their tenure decisions;the investigators compared these forecasts to theself-reported well-being of others whose tenure de-cisions had been made in the past. The sample con-sisted of all assistant professors who were consid-ered for tenure in the liberal arts college of a majoruniversity over a 10-year period, and it was dividedinto three categories: current assistant professors,those whose decisions were made less than 5 yearsearlier, and those whose decisions were made morethan 5 years earlier. Current assistant professorspredicted that they would be much happier duringthe first 5 years after a positive decision, but thatthis difference would dissipate during the subse-quent 5 years. Thus they expected to adapt muchmore slowly than others actually did: There was no


significant difference in reported well-being be-tween those who had and had not received tenure ineither the first 5 or the next 5 years afterward.

13. By contrast, participants accurately predicted howmuch winning the coin flip would increase theirhappiness.

14. Why are people often unable to accurately predicttheir affective reactions to negative events? Kermeret al. (2006) suggest that people do not realize howcapable they are of finding “silver linings.” For ex-ample, participants who were asked to report theirthoughts before and after losing the coin flip weresignificantly more likely to think about their $2profit after losing the coin flip than before the coinwas flipped. Conversely, participants were morelikely to think they would focus on the $3 loss be-fore the coin was flipped than they actually did afterlosing the coin flip. Other researchers (e.g., Schkade& Kahneman, 1998) attribute affective forecastingerrors to “focusing illusions,” whereby people exag-gerate the impact of specific narrow changes in theircircumstances on well-being. Both are plausible ex-planations of the affective forecasting errors docu-mented in the studies discussed here.

15. However, Kivetz and Simonson (2002) suggest thatsome people have a hard time selecting luxuries(items that are presumably high in short-run bene-fits, but low in long-run benefits) over cash when ei-ther would be received shortly after the decision.They demonstrate that choosing luxuries over cashis easier when the consequences of the decision aredelayed.

16. Goldstein and Goldstein (2006) document a similarphenomenon among Netflix customers, who watchand return low-brow movies right away, but lethigh-brow movies sit around much longer beforewatching them.

17. However, note that since the rewards were gift cer-tificates, the consumption they afforded was not im-mediate in any conventional sense. To address thislimitation, McClure, Ericson, Laibson, Loewen-stein, and Cohen (2007) ran an experiment in whichparticipants were asked not to drink any liquidsduring the 3 hours preceding their session. Whilehaving their brains scanned with fMRI, participantsmade a series of choices between receiving a smallamount of juice or water immediately (by having itsquirted into their mouths) and receiving a largeramount of juice or water up to 20 minutes later.Like McClure et al. (2004), McClure et al. (2007)found that limbic regions were preferentially re-cruited for choices involving immediately availablejuice or water, whereas fronto-parietal regions wererecruited for all choices.

18. The results are consistent with earlier behavioral re-search by Shiv and Fedorikhin (1999), who foundthat cognitive load increases the likelihood ofchoosing cake over fruit salad. The McClure et al.(2004) results suggest that cognitive load interferedmore with activation in fronto-parietal regions than

with activation in limbic regions, making partici-pants’ visceral attraction to the cake moreinfluential.

19. But see Dana, Weber, and Kuang (2007) for evidencesuggesting that some actions that appear to reflect ataste for fairness may in fact reflect a desire to ap-pear to have a taste for fairness.

20. Incidental emotion also influences prosocial behav-ior. Darlington and Macker (1966), for example,found that incidental guilt increased participants’willingness to donate blood. Alice Isen and her col-leagues (e.g., Isen & Levin, 1972; Isen, Horn, &Rosenhan, 1973; Isen, Clark, & Schwartz, 1976)have found in a variety of settings that incidentalhappiness (induced, e.g., by finding a dime in aphone booth or receiving free cookies) increasespeople’s willingness to help others (e.g., by pickingup their dropped papers or by helping the experi-menter with a subsequent task; but see Isen &Simmonds, 1978). Incidental gratitude also in-creases people’s willingness to help others (Bartlett& DeSteno, 2006). Although the preceding studiesdid not deal directly with money, note that the help-ing behavior they documented did involve expendi-tures of costly resources (e.g., blood, effort, atten-tion).

REFERENCES

Ainslie, G. (1975). Specious reward: A behavioral the-ory of impulsiveness and impulse control. Psycholog-ical Bulletin, 82(4), 463–496.

Andrade, E. B., & Ariely, D. (2006). Short- and long-term consequences of emotions in decision making.Manuscript in preparation, University of California,Berkeley.

Ariely, D., & Loewenstein, G. (2005). The heat of themoment: The effect of sexual arousal on sexual deci-sion making. Journal of Behavioral DecisionMaking, 18(1), 1–12.

Bartlett, M. Y., & DeSteno, D. (2006). Gratitude andprosocial behavior: Helping when it costs you. Psy-chological Science, 17(4), 319–325.

Benhabib, J., & Bisin, A. (2005). Modeling internal com-mitment mechanisms and self-control: A neuroeco-nomics approach to consumption-saving decisions.Games and Economic Behavior, 52(2), 460–492.

Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R.(1997). Deciding advantageously before knowing theadvantageous strategy. Science, 275, 1293–1295.

Bernheim, B. D., & Rangel, A. (2004). Addiction andcue-triggered decision processes. American Eco-nomic Review, 94(5), 1558–1590.

Bernoulli, D. (1954). Specimen theoriae novae demensura sortis [Exposition of a new theory on themeasurement of risk]. Econometrica, 22(1), 23–36.(Original work published 1738)

Berns, G. S., Chappelow, J., Cekic, M., Zink, C. F.,Pagnoni, G., & Martin-Skurski, M. E. (2006). Neu-


robiological substrates of dread. Science, 312, 754–758.

Blount, S. (1995). When social outcomes aren’t fair: Theeffect of causal attributions on preferences. Organi-zational Behavior and Human Decision Processes,63(2), 131–144.

Brickman, P., Coates, D., & Janoff-Bulman, R. (1978).Lottery winners and accident victims: Is happinessrelative? Journal of Personality and Social Psychol-ogy, 36(8), 917–927.

Brocas, I., & Carrillo, J. D. (2006). The brain as a hier-archical organization. Manuscript in preparation,University of Southern California.

Browne, M. J., & Hoyt, R. E. (2000). The demand forflood insurance: empirical evidence. Journal of Riskand Uncertainty, 20(2), 271–289.

Buchel, C., & Dolan, R. J. (2000). Classical fear condi-tioning in functional neuroimaging. Current Opinionin Neurobiology, 10(2), 219–233.

Camerer, C. (2003). Behavioral game theory: Experi-ments in strategic interaction. Princeton, NJ: Prince-ton University Press.

Caplin, A., & Leahy, J. (2001). Psychological expectedutility theory and anticipatory feelings. QuarterlyJournal of Economics, 116(1), 55–80.

Conlin, M., O’Donoghue, T., & Vogelsang, T. J. (2007).Projection bias in catalog orders. American Eco-nomic Review, 97(4), 1217–1249.


Dana, J., Weber, R. A., & Kuang, J. X. (2007). Ex-ploiting moral wiggle room: Experiments demon-strating an illusory preference for fairness. EconomicTheory, 33(1), 67–80.

Darlington, R. B., & Macker, C. E. (1966). Displace-ment of guilt-produced altruistic behavior. Journal ofPersonality and Social Psychology, 4(4), 442–443.

Dawes, R. M., & Mulford, M. (1996). The false consen-sus effect and overconfidence: Flaws in judgment orflaws in how we study judgment? OrganizationalBehavior and Human Decision Processes, 65(3),201–211.

Ditto, P. H., Pizarro, D. A., Epstein, E. B., Jacobson, J.A., & MacDonald, T. K. (2006). Visceral influenceson risk-taking behavior. Journal of Behavioral Deci-sion Making, 19(2), 99–113.

Dufwenberg, M., & Kirchsteiger, G. (2004). A theory ofsequential reciprocity. Games and Economic Behav-ior, 47(2), 268–298.

Dunn, B. D., Dalgleish, T., & Lawrence, A. D. (2005).The somatic marker hypothesis: A critical evaluation.Neuroscience and Biobehavioral Reviews, 30(2), 1–33.

Edmans, A., García, D., & Norli, Ø. (2007). Sports sen-timent and stock returns. Journal of Finance, 62(4),1967–1998.

Eisenberger, N. I., Lieberman, M. D., & Williams, K. D.(2003). Does rejection hurt: An fMRI study of socialexclusion. Science, 302, 290–292.

Elster, J., & Loewenstein, G. (1992). Utility from mem-

ory and anticipation. In G. Loewenstein & J. Elster(Eds.), Choice over time. New York: Russell SageFoundation.

Evans, J. St. B. T. (2008). Dual-processing accounts ofreasoning, judgment, and social cognition. AnnualReview of Psychology, 59.[pp tk]

Fessler, D. M. T., Pillsworth, E. G., & Flamson, T. J.(2004). Angry men and disgusted women: An evolu-tionary approach to the influence of emotions on risktaking. Organizational Behavior and Human Deci-sion Processes, 95, 107–123.

Frederick, S., Novemsky, N., Wang, J., Dhar, R., &Nowlis, S. (2006). Opportunity costs and consumerdecisions. Manuscript in preparation, MassachusettsInstitute of Technology.

Fudenberg, D., & Levine, D. K. (2006). A dual-selfmodel of impulse control. American Economic Re-view, 96(5), 1449–1476.

Galanter, E. (1992). Utility functions for nonmonetaryevents. American Journal of Psychology, 65, 45–55.

Gilbert, D. T., Morewedge, C. K., Risen, J. L., & Wil-son, T. D. (2004). Looking forward to looking back-ward: The misprediction of regret. PsychologicalSci-ence, 15(5), 346–350.

Gilbert, D. T., Pinel, E. C., Wilson, T. D., Blumberg, S.J., & Wheatley, T. P. (1998). Immune neglect: Asource of durability bias in affective forecasting.Journal of Personality and Social Psychology, 75(3),617–638.

Goldstein, D. G., & Goldstein, D. C. (2006). Profitingfrom the long tail. Harvard Business Review, 84(6),24–28.

Guth, W., Schmittberger, R., & Schwarze, B. (1982). Anexperimental analysis of ultimatum bargaining. Jour-nal of Economic Behavior and Organization, 3(4),367–388.

Hirshleifer, D., & Shumway, T. (2003). Good day sun-shine: Stock returns and the weather. Journal of Fi-nance, 58(3), 1009–1032.

Hoch, S. J., & Loewenstein, G. (1991). Time-inconsistent preferences and consumer self-control.Journal of Consumer Research, 17, 492–507.

Isen, A. M., Clark, M., & Schwartz, M. F. (1976). Dura-tion of the effect of good mood on helping: Foot-prints on the sands of time. Journal of Personalityand Social Psychology, 34(3) 385–393.

Isen, A. M., Horn, N., & Rosenhan, D. L. (1973). Ef-fects of success and failure on children’s generosity.Journal of Personality and Social Psychology, 27(2),239–247.

Isen, A. M., & Levin, P. F. (1972). The effect of feelinggood on helping: Cookies and kindness. Journal ofPersonality and Social Psychology, 15(4), 294–301.

Isen, A. M., Nygren, T. E., & Ashby, F. G. (1988). Influ-ence of positive affect on social categorization. Moti-vation and Emotion, 16(1), 65–78.

Isen, A. M., & Simmonds, S. F. (1978). The effect offeeling good on a helping task that is incompatiblewith good mood. Social Psychology, 41(4), 346–349.

Johnson, E. J., Hershey, J., Meszaros, J., & Kunreuther,


H. (1993). Framing, probability distortions, and in-surance decisions. Journal of Risk and Uncertainty,7(1), 35–51.

Jones, S. K., Frisch, D., Yurak, T. J., & Kim, E. (1998).Choices and opportunities: Another effect of framingon decisions. Journal of Behavioral DecisionMaking, 11(3), 211–226.

Kahn, B. E., & Isen, A. M. (1993). The influence of pos-itive affect on variety seeking among safe, enjoyableproducts. Journal of Consumer Research, 20(2),257–270.

Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1990).Experimental tests of the endowment effect and theCoase theorem. Journal of Political Economy, 98(6),1325–1348.

Kahneman, D., & Tversky, A. (1979). Prospect theory:An analysis of decision under risk. Econometrica,47(2), 263–291.

Keinan, A. (2006). Productivity mindset and the con-sumption of collectible experiences. Manuscript inpreparation, Columbia University.

Kermer, D. A., Driver-Linn, E., Wilson, T. D., & Gilbert,D. T. (2006). Loss aversion is an affective forecastingerror. Psychological Science, 17(8), 649–653.

Kirby, K. N., & Herrnstein, R. J. (1995). Preference re-versals due to myopic discounting of delayed reward.Psychological Science, 6(2), 83–89.

Kivetz, R., & Keinan, A. (2006). Repenting hyperopia:An analysis of self-control regrets. Journal of Con-sumer Research, 33, 273–282.

Kivetz, R., & Simonson, I. (2002). Self-control for therighteous: Toward a theory of precommitment to in-dulgence. Journal of Consumer Research, 29, 199–217.

Knutson, B., Adams, C. M., Fong, G. W., & Hommer,D. (2001). Anticipation of increasing monetary re-ward selectively recruits nucleus accumbens. Journalof Neuroscience, 21, RC159.

Knutson, B., Fong, G. W., Bennett, S. M., Adams, C. M.,& Hommer, D. (2003). A region of mesial prefrontalcortex tracks monetarily rewarding outcomes: Char-acterization with rapid event-related fMRI.NeuroImage, 18, 263–272.

Knutson, B., Rick, S., Wimmer, G. E., Prelec, D., &Loewenstein, G. (2007). Neural predictors of pur-chases. Neuron, 53, 147–156.

Kogut, T., & Ritov, I. (2005). The “identified victim” ef-fect: An identified group, or just a single individual?Journal of Behavioral Decision Making, 18(3), 157–167.

Koopmans, T. C. (1960). Stationary ordinal utility andimpatience. Econometrica, 28(2), 287–309.

Köszegi, B., & Rabin, M. (2006). A model of reference-dependent preferences. Quarterly Journal of Eco-nomics, 121(4), 1133–1165.

Kunreuther, H., Ginsberg, R., Miller, L., Sagi, P., Slovic,P., Borkan, B., et al. (1978). Disaster insurance pro-tection: Public policy lessons. New York: Wiley.

Laibson, D. I. (1997). Golden eggs and hyperbolic dis-

counting. Quarterly Journal of Economics, 112(2),443–477.

Lerner, J. S., & Keltner, D. (2001). Fear, anger, and risk.Journal of Personality and Social Psychology, 81(1),146–159.

Lerner, J. S., Small, D. A., & Loewenstein, G. (2004).Heart strings and purse strings: Carryover effects ofemotions on economic decisions. Psychological Sci-ence, 15(5), 337–341.

Loewenstein, G. (1987). Anticipation and the valuationof delayed consumption. The Economic Journal,97(387), 666–684.

Loewenstein, G. (1996). Out of control: Visceral influ-ences on behavior. Organizational Behavior and Hu-man Decision Processes, 65(3), 272–292.

Loewenstein, G., & Adler, D. (1995). A bias in the pre-diction of tastes. The Economic Journal, 105(431),929–937.

Loewenstein, G., & Lerner, J. (2003). The role of affectin decision making. In R. J. Dawson, K. R. Scherer, &H. H. Goldsmith (Eds.), Handbook of affective sci-ence (pp. 619–642). Oxford: Oxford University Pres.

Loewenstein, G., Nagin, D., & Paternoster, R. (1997).The effect of sexual arousal on predictions of sexualforcefulness. Journal of Crime and Delinquency,34(4), 443–473.

Loewenstein, G., & O’Donoghue, T. (2004). Animalspirits: Affective and deliberative processes in eco-nomic behavior. Manuscript in preparation, CarnegieMellon University.

Loewenstein, G., O’Donoghue, T., & Rabin, M. (2003).Projection bias in predicting future utility. QuarterlyJournal of Economics, 118(4), 1209–1248.

Loewenstein, G., & Prelec, D. (1992). Anomalies inintertemporal choice. Quarterly Journal of Econom-ics, 107(2), 573–597.

Loewenstein, G., & Small, D. A. (2007). The scarecrowand the tin man: The vicissitudes of human sympathyand caring. Review of General Psychology, 11, 112–126.

Loewenstein, G., Weber, E. U., Hsee, C. K., & Welch, N.(2001). Risk as feelings. Psychological Bulletin,127(2), 267–286.

Loomes, G., & Sugden, R. (1982). Regret theory: An al-ternative theory of rational choice under uncertainty.The Economic Journal, 92(368), 805–824.

Loomes, G., & Sugden, R. (1986). Disappointment anddynamic consistency in choice under uncertainty. Re-view of Economic Studies, 53(2), 271–282.

Maia, T. V., & McClelland, J. L. (2004). A reexamina-tion of the evidence for the somatic marker hypothe-sis: What participants really know in the Iowa gam-bling task. Proceedings of the National Academy ofSciences USA, 101(45), 16075–16080.

Markowitz, H. M. (1952). The utility of wealth. Journalof Political Economy, 60(2), 151–158.

McClure, S. M., Ericson, K. M., Laibson, D. I.,Loewenstein, G., & Cohen, J. D. (2007). Time dis-


counting for primary rewards. Journal of Neurosci-ence, 27(21), 5796–5804.

McClure, S. M., Laibson, D. I., Loewenstein, G., & Co-hen J. D. (2004). Separate neural systems value im-mediate and delayed monetary rewards. Science,306, 503–507.

Mellers, B. A., Schwartz, A., Ho, K., & Ritov, I. (1997).Decision affect theory: Emotional reactions to theoutcomes of risky options. Psychological Science,8(6), 423–429.

Mischel, W. (1974). Processes in delay of gratification.In L. Berkowitz (Ed.), Advances in experimental so-cial psychology (Vol. 7). New York: Academic Press.

Pillutla, M. M., & Murnighan, J. K. (1996). Unfairness,anger, and spite: Emotional rejections of ultimatumoffers. Organizational Behavior and Human Deci-sion Processes, 68(3), 208–224.

Prelec, D., & Loewenstein, G. (1998). The red and theblack: Mental accounting of savings and debt. Mar-keting Science, 17, 4–28.

Rabin, M. (1993). Incorporating fairness into game the-ory and economics. American Economic Review, 83,1281–1302.

Rachlin, H., & Raineri, A. (1992). Irrationality, impul-siveness, and selfishness as discount reversal effects.In G. Loewenstein & J. Elster (Eds.), Choice overtime. New York: Russell Sage Foundation.

Read, D., Loewenstein, G., & Kalyanaraman, S. (1999).Mixing virtue and vice: Combining the immediacy ef-fect and the diversification heuristic. Journal ofBehavioral Decision Making, 12(4), 257–273.

Rick, S. (2007). The influence of anticipatory affect onconsumer choice. Unpublished doctoral dissertation,Department of Social and Decision Sciences, Carne-gie Mellon University.

Rick, S., Cryder, C., & Loewenstein, G. (2008). Tight-wads and spendthrifts. Journal of Consumer Re-search, 34.

Ross, L., Greene, D., & House, P. (1977). The false con-sensus effect: An egocentric bias in social perceptionand attribution processes. Journal of ExperimentalSocial Psychology, 13(2), 279–301.

Rottenstreich, Y., & Hsee, C. K. (2001). Money, kisses,and electric shocks: On the affective psychology ofrisk. Psychological Science, 12(3), 185–190.

Samuelson, P. (1937). A note on measurement of utility.Review of Economic Studies, 4(2), 155–161.

Sanfey, A. G., Rilling, J. K., Aronson, J. A., Nystrom, L.E., & Cohen, J. D. (2003). The neural basis of eco-nomic decision-making in the ultimatum game. Sci-ence, 300, 1755–1758.

Schelling, T. C. (1968). The life you save may be yourown. In S. B. Chase (Ed.), Problems in public expen-diture analysis. Washington, DC: Brookings Insti-tute.

Schkade, D., & Kahneman, D. (1998). Does living inCalifornia make people happy?: A focusing illusionin judgments of life satisfaction. Psychological Sci-ence, 9(6), 340–346.

Schwarz, N., & Clore, G. (1983). Moods, misattri-bution, and judgments of well-being: Informativeand directive functions of affective states. Journal ofPersonality and Social Psychology, 45(3), 513–523.

Shiv, B., & Fedorikhin, A. (1999). Heart and mind inconflict: The interplay of affect and cognition in con-sumer decision making. Journal of Consumer Re-search, 26, 278–292.

Shiv, B., Loewenstein, G., Bechara, A., Damasio, H., &Damasio, A. R. (2005). Investment behavior and thenegative side of emotion. Psychological Science,16(6), 435–439.

Slovic, P., Fischhoff, B., & Lichtenstein, S. (1980). Factsand fears: Understanding perceived risk. In R. C.Schwing & W. A. Albers, Jr. (Eds.), Societal risk as-sessment: How safe is safe enough? San Francisco:Jossey-Bass.

Small, D. A., & Loewenstein, G. (2003). Helping ‘a’ vic-tim or helping ‘THE’ victim: Altruism andidentifiability. Journal of Risk and Uncertainty,26(1), 5–16.

Small, D. A., & Loewenstein, G. (2005). The devil youknow: The effects of identifiability on punishment.Journal of Behavioral Decision Making, 18(5), 311–318.

Smith, A. (1981). The theory of moral sentiments (D. D.Raphael & A. L. Macfie, Eds.). Indianapolis, IN: Lib-erty Fund. (Original work published 1759)

Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cog-nitive appraisal in emotion. Journal of Personalityand Social Psychology, 48(4), 813–838.

Sugden, R. (1986). Regret, recrimination, and rational-ity. In L. Daboni, A. Montesano, & M. Lines (Eds.),Theory and decision library series: Vol. 47. Recentdevelopments in the foundations of utility and risktheory. Dordrecht, The Netherlands: Reidel.

Thaler, R. H. (1980). Toward a positive theory of con-sumer choice. Journal of Economic Behavior and Or-ganization, 1(1), 39–60.

Thaler, R. H., & Shefrin, H. M. (1981). An economictheory of self-control. Journal of Political Economy,89(2), 392–410.

Tiedens, L. Z., & Linton, S. (2001). Judgment underemotional certainty and uncertainty: The effects ofspecific emotions on information processing. Journalof Personality and Social Psychology, 81(6), 973–988.

Tversky, A., & Koehler, D. J. (1994). Support theory: Anonextensional representation of subjective probabil-ity. Psychological Review, 101(4), 547–567.

von Neumann, J., & Morgenstern, O. (1944). Theory ofgames and economic behavior. New York: Wiley.

Wicker, B., Keysers, C., Plailly, J., Royet, J-P., Gallese,V., & Rizzolatti, G. (2003). Both of us disgusted inmy insula: The common neural basis of seeing andfeeling disgust. Neuron, 40, 655–664.

Zeelenberg, M., & Beattie, J. (1997). Consequences ofregret aversion: 2. Additional evidence for effectsof feedback on decision making. Organizational


Behavior and Human Decision Processes, 72(1),63–78.

Zeelenberg, M., Beattie, J. van der Plight, J., & De Vries,N. K. (1996). Consequences of regret aversion: Ef-fects of expected feedback on risky decision making.Organizational Behavior and Human Decision Pro-cesses, 65(2), 148–158.


PA R T I I

BIOLOGICAL ANDNEUROPHYSIOLOGICAL

APPROACHES

C H A P T E R 1 0

Emotional Networks in the Brain

JOSEPH E. LEDOUX and ELIZABETH A. PHELPS

Contemporary neuroscientists have available avast arsenal of tools for understanding brainfunctions, from the level of anatomical systemsto the level of molecules. Localization of func-tion at the anatomical level is the oldest butalso the most basic approach. Until the func-tion in question can be localized to a specificset of structures and their connections, applica-tion of cellular and molecular approaches is theneurobiological equivalent of a search for aneedle in a haystack. Fortunately, considerableprogress has been made in understanding theanatomical organization of one emotion, fear,and this chapter focuses on this work. Whereasmost of the progress in the past came fromstudies of experimental animals, recent studiesin humans, some capitalizing on new tech-niques for imaging the human brain, have con-firmed and extended the animal work.

IN SEARCH OFTHE EMOTIONAL BRAIN

Our understanding of the brain mechanisms ofemotion has changed radically over the past

120 years. In the late 19th century, WilliamJames (1884) suggested that emotion is a func-tion of sensory and motor areas of the neocor-tex, and that the brain does not possess a spe-cial system devoted to emotional functions.This idea was laid low by studies showing thatemotional reactions require the integrity of thehypothalamus (Cannon, 1929; Bard, 1929).On the basis of such observations, Papez(1937) proposed a circuit theory of emotion in-volving the hypothalamus, anterior thalamus,cingulate gyrus, and hippocampus. MacLean(1949, 1952) then named the structures of thePapez circuit, together with several additionalregions (amygdala, septal nuclei, orbito-frontalcortex, portions of the basal ganglia), the“limbic system”; he viewed the limbic systemas a general-purpose system involved in the me-diation of functions required for the survival ofthe individual and the species.

MacLean’s writings were very persuasive,and for many years the problem of relatingemotion to brain mechanisms seemed solved atthe level of anatomical systems. However, thelimbic system concept came under fire begin-ning in the 1980s (see Brodal, 1982; Swanson,

159

1983; LeDoux, 1987, 1991; Kotter & Meyer,1992). It is now believed that the concept suf-fers from imprecision at both the structural andfunctional levels. For example, it has provenimpossible to provide unequivocal criteria fordefining which structures and pathways shouldbe included in the limbic system (Brodal, 1982;Swanson, 1983). A standard criterion, connec-tivity with the hypothalamus, extends thelimbic system to include structures at all levelsof the central nervous system, from the neocor-tex to the spinal cord. Furthermore, classiclimbic areas, such as the hippocampus andmammillary bodies, have proven to be far moreimportant for cognitive processes (such as de-clarative memory) than for emotional pro-cesses (e.g., Squire & Zola, 1996; Cohen &Eichenbaum, 1993).

Nevertheless, one limbic area that has beenconsistently implicated in emotional processesin a variety of situations is the amygdala (e.g.,Weiskrantz, 1956; Gloor, 1960; Goddard,1964; Mishkin & Aggleton, 1981; Aggleton &Mishkin, 1986; LeDoux, 1987, 1996; Rolls,1986, 1992; Halgren, 1992; Aggleton, 1992;Davis, 1992; Kapp, Whalen, Supple, & Pascoe,1992; Ono & Nishijo, 1992; Damasio, 1994;Everitt & Robbins, 1992; McGaugh et al.,1995). Interestingly, the amygdala was not partof the Papez circuit model and was clearly asecond-class citizen, relative to the hippocam-pus at least, in the limbic system hypothesis.However, the survival of the limbic system hy-pothesis for so long is in part due to the inclu-sion of the amygdala (LeDoux, 1992). Other-wise, the relation between emotional functionsand classic limbic areas would have been farless prominent over the years.

THE AMYGDALAAS AN EMOTIONAL COMPUTER

The contribution of the amygdala to emotionemerged from studies of the Kluver–Bucy syn-drome, a complex set of behavioral changesbrought about by damage to the temporal lobein primates (Kluver & Bucy, 1937). Followingsuch lesions, animals lose their fear of previ-ously threatening stimuli, attempt to copulatewith members of other species, and attempt toeat a variety of things that “normal” primatesfind unattractive (feces, meat, rocks). Studiesby Weiskrantz (1956) then determined that le-sions confined to the amygdala and sparing

other temporal lobe structures produce theemotional components of the syndrome.Weiskrantz proposed that amygdala lesions in-terfere with the ability to determine the motiva-tional significance of stimuli. A host of subse-quent studies have shown that the amygdala isa key structure in the assignment of rewardvalue to stimuli (Jones & Mishkin, 1972;Spiegler & Mishkin, 1981; Gaffan & Harrison,1987; Gaffan, Gaffan, & Harrison, 1988;Everitt & Robbins, 1992; Ono & Nishijo,1992; Rolls, 1992), in the conditioning of fearto novel stimuli (Blanchard & Blanchard,1972; Kapp et al., 1992; Davis, 1992, 1994;LeDoux, 1996; Maren & Fanselow, 1996), inthe self-administration of rewarding brainstimulation (Kane, Coulombe, & Miliaressis,1991; Olds, 1977), and in the elicitation bybrain stimulation of a host of behavioral andautonomic responses typical of emotional reac-tions (Hilton & Zbrozyna, 1963; Fernandez deMolina & Hunsperger, 1962; Kapp, Pascoe, &Bixler, 1984; Iwata, Chida, & LeDoux, 1987).These and other findings have led a number ofauthors to conclude that the amgydala plays animportant role in the assignment of affectivesignificance to sensory events. As important asthe various studies described above have beenin establishing that the amygdala plays a role inemotional processes, most of the studies weredone with little appreciation for the anatomicalorganization of the amygdala. It is generallybelieved that there are at least a dozen differentnuclei, and that each has several subdivisions,each with its own set of unique connections(e.g., Pitkänen, Savander, & LeDoux, 1997). Ifwe are to understand how the amygdala partic-ipates in computation of emotional signifi-cance, we need to take these distinctions intoaccount.

NEURAL PATHWAYS INVOLVEDIN FEAR PROCESSING

Much of our understanding of the role of dif-ferent regions of the amygdala in emotionalprocesses has come from studies of fear condi-tioning, where an auditory stimulus, a condi-tioned stimulus (CS), is paired with footshock,the unconditioned stimulus (US) (Figure 10.1).The reason this task has been so successful inmapping the pathways is, in large part, the sim-plicity of the task itself. It involves a discrete,well-defined CS and stereotyped autonomic

160 II. BIOLOGICAL AND NEUROPHYSIOLOGICAL APPROACHES

and behavioral conditioned responses (CRs).These two features—a well-defined stimulusand stereotyped responses—are very helpfulwhen trying to relate brain function to brainstructure. Also, fear conditioning can and hasbeen used similarly in animal and human stud-ies, which has allowed the establishment ofcommonalities in the underlying brain systems.Fear conditioning may not be able to tell us ev-erything we need to know about emotions andthe brain, or even about fear and the brain, butit has been an excellent starting point.

Basic Circuits

The pathways involved in conditioning of fearresponses to a single-tone CS are shown in Fig-ure 10.2. The CS is transmitted through the au-ditory system to the auditory thalamus, includ-ing the subregions of the medial geniculate

body (MGB) and adjacent regions of the poste-rior thalamus (LeDoux, Sakaguchi, & Reis,1984; LeDoux, Ruggiero, & Reis, 1985;LeDoux, Farb, & Ruggiero, 1990). The signalis then transmitted from all regions of the audi-tory thalamus to the auditory cortex, and froma subset of thalamic nuclei to the amygdala.The thalamo-amygdala pathway originates pri-marily in the medial division of the MGB andthe associated posterior intralaminar nucleus(LeDoux, Cicchetti, Xagoraris, & Romanski,1990). The auditory association cortex alsogives rise to a projection to the amygdala(Romanski & LeDoux, 1993; Mascagni, Mc-Donald, & Coleman, 1993). Both the thalamo-amygdala and thalamo-cortico-amygdala path-ways terminate in the sensory input region ofthe amygdala, the lateral nucleus (LA) (seeTurner & Herkenham, 1991; LeDoux,Cicchetti, et al., 1990; Romanski & LeDoux,

10. Emotional Networks in the Brain 161

FIGURE 10.1. Fear conditioning. Fear conditioning occurs in three phases. In habituation, the rat is ac-climated to the chamber. No stimuli are presented. During conditioning, the tone conditioned stimulus(CS) is paired with the footshock unconditioned stimulus (US). Testing then involves presentation of theCS without the US the next day. Typically, the rat exhibits freezing responses to the CS during the test. Ifthe rat received unpaired presentations of the CS and US during conditioning, it freezes little to the CS,indicating that it did not come to associate the CS with the US.

1993; Mascagni et al., 1993). In fact, the twopathways converge onto single neurons in LA(Li, Stutzman, & LeDoux, 1996). Damage toLA interferes with fear conditioning (LeDoux,Cicchetti, et al., 1990), which can be mediatedby either the thalamo-amygdala or thalamo-cortico-amygdala pathways (for discussions,see Romanski & LeDoux, 1993; Campeau &Davis, 1995; Corodimas & LeDoux, 1995).Temporary inactivation of LA and the adjacentbasal nucleus (Helmstetter & Bellgowan, 1994;Muller, Corodimas, Fridel, & LeDoux, 1997;Wilensky, Schafe, & LeDoux, 1999, 2000), orpharmacological blockade of excitatory aminoacid receptors in this region (Miserendino,Sananes, Melia, & Davis, 1990; Kim &Fanselow, 1992; Maren & Fanselow, 1996;Gewirtz & Davis, 1997; Rodrigues, Schafe, &LeDoux, 2004), also disrupts the acquisition of

conditioned fear, and facilitation of excitatoryamino acid transmission enhances the rate offear learning (Rogan, Staubli, & LeDoux,1997). A host of intracellular signaling cas-cades downstream of excitatory amino acidtransmission have been implicated in fear con-ditioning as well (for a review, see Rodrigues etal., 2004).

Although the auditory cortex is not requiredfor the acquisition of conditioned fear to a sim-ple auditory stimulus (Romanski & LeDoux,1992; Armony, Servan-Schreiber, Romanski,Cohen, & LeDoux, 1997), processing of the CSby cells in the auditory cortex is modified as aresult of its pairing with the US (Weinberger,1995; Quirk, Armony, & LeDoux, 1997). Insituations involving more complex stimuli thatmust be discriminated, recognized, and/or cate-gorized, the auditory cortex may be critical for


FIGURE 10.2. Neural circuit of fear conditioning. Fear conditioning occurs in the brain via the modifica-tion of the processing of the auditory CS by the somatosensory US. As shown, the CS and US converge inthe lateral amygdala (LA), which receives CS and US inputs from both thalamic and cortical areas. TheLA then communicates with the central amygdala (CE) both directly and by way of other amygdala ar-eas, including the basal nucleus (B) and the intercalated masses (ITC). The CE connects with brainstemand hypothalamic areas that control the expression of fear responses, including freezing behavior (medi-ated by the central gray, CG), autonomic nervous system (ANS) responses (mediated by the lateral hypo-thalamus, LH), and hormonal responses (mediated by the paraventricular hypothalamus).

aversive learning (e.g., Cranford & Igarashi,1977; Whitfield, 1980) and may provide an im-portant set of inputs to the amygdala.

What are the advantages of the parallel pro-cessing capabilities of this system? First, the ex-istence of a subcortical pathway allows theamygdala to detect threatening stimuli in theenvironment quickly, in the absence of a com-plete and time-consuming analysis of the stim-ulus. This “quick and dirty” processing routemay confer an evolutionary advantage to thespecies. Second, the rapid subcortical pathwaymay function to “prime” the amygdala to eval-uate subsequent information received along thecortical pathway (LeDoux, 1986a, 1986b; Li etal., 1996). For example, a loud noise may besufficient to alert the amygdala at the cellularlevel to prepare to respond to a dangerouspredator lurking nearby, but defensive reac-tions may not be fully mobilized until the audi-tory cortex analyzes the location, frequency,and intensity of the noise, to determine specifi-cally the nature and extent of this potentiallythreatening auditory signal. The convergenceof the subcortical and cortical pathways ontosingle neurons in the LA (Li et al., 1996) pro-vides a means by which the integration cantake place. Third, computational modelingstudies show that the subcortical pathway canfunction as an interrupting device that enablesthe cortex, by way of amygdalo-cortical projec-tions, to shift attention to dangerous stimulithat occur outside the focus of attention(Armony, Servan-Schreiber, Cohen, & LeDoux,1996). Modeling studies have also emphasizedthe evolutionary advantages of the dual-routehypothesis (den Dulk, Heerebout, & Phaf,2003).

As noted, sensory information from both thethalamus and the cortex enters the amygdalathrough the LA. Information processed by theLA is then transmitted via intra-amygdala con-nections (Pitkänen et al., 1997; Pare & Smith,1998) to the basal and accessory basal nuclei,where it is integrated with other incoming in-formation from other areas and further trans-mitted to the central nucleus. The central nu-cleus is the main output system of theamygdala (LeDoux, 1996; Davis, 1992;LeDoux, Iwata, Cicchetti, & Reis, 1988;Carrive, Lee, & Su, 2000; Fendt & Fanselow,1999; Davis & Whalen, 2001). Damage to thecentral amygdala or structures that project to itinterferes with the acquisition and expressionof all CRs, whereas lesions of areas to which

the central amygdala projects interfere with in-dividual responses, such as blood pressurechanges, freezing behavior, or hormone release.

With the neural circuits underlying fear con-ditioning well understood, studies have turnedto the molecular mechanisms that make emo-tional learning possible. These mechanisms(Figure 10.3) are understood in great detail, al-lowing novel approaches to the treatment offear-related disorders.

Contextualization of Fear

Whether a stimulus signals danger, and thuselicits fear reactions, often depends on the situ-ation (context) in which it occurs. For example,the sight of a bear in the zoo poses little threat,but the same bear seen while we are on a walkin the woods will probably make us run awayin fear. Furthermore, contexts may themselves


FIGURE 10.3. High road and low road to theamygdala. The amygdala receives inputs fromsensory processing regions in the thalamus andcortex. The thalamic pathway, the low road, pro-vides a rapid but crude representation, while thecortical pathway provides slower but more elabo-rate representation. The low-road inputs arethought to initiate amygdala processing, and thecortical inputs build upon this initial processing.Both pathways in all likelihood process emotionalstimuli unconsciously, with conscious awarenessof the stimulus requiring prefrontal areas (see Fig-ure 10.4).

acquire aversive value through prior experi-ences. If we are mugged, we are very likely tofeel “uneasy” when we return to the scene ofthe crime. The relationship between environ-mental situations and fear responses can beeninvestigated in the laboratory through contex-tual fear conditioning: When a rat is condi-tioned to expect a footshock in the presence ofa tone CS, it will also exhibit fear reactions tothe chamber where the conditioning tookplace, even in the absence of the CS (Kim &Fanselow, 1992; Phillips & LeDoux, 1992).Several studies have shown that the formationand consolidation of contextual fear associa-tions depend on the hippocampus. Lesions ofthe hippocampus made prior to training inter-fere with the acquisition of CRs to the context,without having any effect on the conditioningto the CS (Maren, Anagnostaras, & Fanselow,1998; Frankland, Cestari, Filipkowski, Mc-Donald, & Silva, 1998; Phillips & LeDoux,1992, 1994; Selden, Everitt, Jarrard, & Rob-bins, 1991). Furthermore, hippocampal lesionsmade after training interfere with the consoli-dation and retention of contextual fear associa-tions (Kim & Fanselow, 1992). Bidirectio-nal projections between the hippocampalformation and the amygdala (Amaral, Price,Pitkänen, & Carmichael, 1992; Canteras &Swanson, 1992; Ottersen, 1982) provide ana-tomical channels through which the attach-ment of emotional value to context may takeplace. The fibers from the hippocampus to theamygdala terminate extensively in the basaland accessory basal nuclei, and to a muchlesser extent in the LA, suggesting why lesionsof the LA have little effect on context condi-tioning, but lesions of the basal nucleus and/oraccessory basal nucleus seem to be disruptive(Maren, Aharonov, Stote, & Fanselow, 1996;Majidishad, Pelli, & LeDoux, 1996). The roleof the hippocampus in the evaluation of con-textual cues in fear conditioning is consistentwith current theories of spatial, configural,and/or relational processing in the hippocam-pus (Cohen & Eichenbaum, 1993; O’Keefe &Nadel, 1978; Sutherland & Rudy, 1989; Rudy,Huff, & Matus-Amat, 2004).

Getting Rid of Fear

Fear responses tend to be very persistent. Thishas obvious survival advantages, as it allows usto keep a record of previously encounteredthreatening experiences, and thus allows us to

respond quickly to similar situations in the fu-ture. Nonetheless, it is also important to beable to learn that a stimulus no longer signalsdanger. Otherwise, unnecessary fear responseswill be elicited by innocuous stimuli and maypotentially become liabilities, interfering withother important routine tasks. In humans, theinability to inhibit unwarranted fear responsescan have devastating consequences, as ob-served in phobias, posttraumatic stress disor-der, generalized anxiety disorder, and otheranxiety disorders. In laboratory experiments,learned fear responses can be reduced (extin-guished) by repeatedly presenting the CS with-out the US. It is important to note, however,that extinction of conditioned fear responses isnot a passive forgetting of the CS-US associa-tion, but an active process, often involving anew learning (Bouton & Swartzentruber,1991). In fact, CS-elicited responses can bespontaneously reinstated following an unre-lated traumatic experience (Pavlov, 1927/1960;Jacobs & Nadel, 1985; Rescorla & Heth,1975).

Experimental observations in fear condition-ing studies suggest that neocortical areas, par-ticularly areas of the medial prefrontal cortex,are involved in the extinction process. Lesionsof the medial prefrontal cortex lead to apotentiation of fear responses and a retarda-tion of the extinction (Morgan, Romanski, &LeDoux, 1993; Morgan & LeDoux, 1995;Quirk, Russo, Barron, & Lebron, 2000; Milad& Quirk, 2002; Quirk, Garcia, & Gonzalez-Lima, 2006; but see Gewirtz & Davis, 1997).These findings complement electrophysio-logical studies showing that neurons within theorbito-frontal cortex are particularly sensitiveto changes in stimulus–reward associations(Thorpe, Rolls, & Maddison, 1983; Rolls,1999). Lesions of sensory areas of the cortexalso retard extinction (LeDoux, Romanski, &Xagoraris, 1989; Teich et al., 1989), and neu-rons in auditory cortex exhibit extinction resis-tant changes to an auditory CS (Quirk et al.,1997). Thus the medial prefrontal cortex, pos-sibly in conjunction with other neocortical re-gions, may be involved in regulating amygdalaresponses to stimuli based on their current af-fective value. These findings suggest that feardisorders may be related to a malfunction ofthe prefrontal cortex that makes it difficult forpatients to extinguish fears they have acquired(Morgan et al., 1993; Morgan & LeDoux,1995; LeDoux, 1996). Imaging studies in hu-


mans have confirmed the role of medialprefrontal cortex in extinction and other as-pects of fear regulation, as described below.

Stress has deleterious effects on medialprefrontal cortex (Radley et al., 2006; Radley& Morrison, 2005) and also can have effectssimilar to lesions of the medial prefrontalcortex—that is, potentiation of fear responses(Corodimas, LeDoux, Gold, & Schulkin, 1994;Conrad, Margarinos, LeDoux, & McEwen,1997; Shors, 2006). Given that stress is a com-mon occurrence in psychiatric patients, andthat such patients often have functionalchanges in the prefrontal cortex (Drevets et al.,1997; Bremner et al., 1995), it is possible thatthe exaggeration of fear in anxiety disorders re-sults from stress-induced alterations in the me-dial prefrontal region.

Emotional Action

The defensive responses we have considered sofar are hard-wired reactions to danger signals.These are evolution’s gifts to us; they provide afirst line of defense against danger. Some ani-mals rely mainly on these. But mammals, espe-cially we humans, can do much more. We areable to take charge. Once we find ourselves in adangerous situation, we can think, plan, andmake decisions. We make the transition fromreaction to action. Considerably less is under-stood about the brain mechanisms of emo-tional action than reaction, due in part to thefact that emotional actions come in many vari-eties and are limited only by the ingenuity ofthe actor. For example, once we are freezingand expressing physiological responses to adangerous stimulus, the rest is up to us. On thebasis of our expectations about what is likelyto happen next and our past experiences insimilar situations, we make a plan about whatto do. We become instruments of action. In-strumental responses in situations of dangerare often studied by using avoidance condition-ing procedures.

Avoidance is a multistage learning process(Mowrer & Lamoreaux, 1946). First, fear CRsare acquired. Then the CS becomes a signalthat is used to initiate responses that preventencounters with the US. Finally, once avoid-ance responses are learned, animals no longershow the characteristic signs of fear (Rescorla& Solomon, 1967). They know what to do toavoid the danger, and simply perform the re-sponse in a habitual way. Consistent with this

is the fact that the amygdala is required foravoidance learning (for the fear conditioningpart), but not for the expression of well-trainedavoidance responses (the instrumental part)(Parent, Tomaz, & McGaugh, 1992). The in-volvement of an instrumental component tosome aversive learning tasks may explain whythese are not dependent on the amygdala forlong-term storage (Packard, Williams, Cahill,& McGaugh, 1995; McGaugh et al., 1995).Because avoidance learning involves fear con-ditioning, at least initially, it will be subject toall the factors that influence fear conditioningand conditioned fear responding. However, be-cause avoidance learning involves more thansimple fear conditioning, it is to be expectedthat avoidance will be subject to influences thathave little or no effect on conditioned fear.Much more work is needed to understand howfear and avoidance interact, and thus howemotional actions emerge out of emotional re-actions. From what we know so far, it appearsthat, as in other habit systems (Mishkin,Malamut, & Bachevalier, 1984), interactionsamong the amygdala, basal ganglia, and neo-cortex are important in avoidance (Everitt &Robbins, 1992; Gray, 1987; Killcross, Rob-bins, & Everitt, 1997).

The neural basis of emotional actions has re-cently been studied using a paradigm called“escape from fear” (Amorapanth, LeDoux, &Nader, 2000). In this paradigm, rats first un-dergo fear conditioning. They are then placedin a new context and exposed to the CS. Al-though they initially freeze, at some point theymove. When they move, the tone is terminated.Through careful shaping, behaviors that termi-nate the tone can be established as habits,which are learned because removal of the toneis a negative secondary reinforcer. Damage tothe lateral or central amygdala, but not thebasal amygdala, prevents learning in the firstphase (fear conditioning); damage to the lateralor the basal, but not the central, amygdala pre-vents learning of the escape-from-fear compo-nent. The transition from fear reaction to ac-tion thus requires that information processingin the amygdala be diverted from the reactionpathway (lateral to central amygdala) to the ac-tion pathway (lateral to basal amygdala)(LeDoux & Gorman, 2001).

The action circuits may be the meansthrough which emotional behaviors that areperformed initially as voluntary responses arethen converted into habits. Emotional habits


can be useful, but can also be quite detrimental.Successful avoidance is known to prevent ex-tinction of conditioned fear, since the opportu-nity to experience the CS in the absence of theUS is eliminated by the avoidance. In real life,this can perpetuate anxiety states. The patientwith panic disorder and agoraphobia whonever leaves home as a means of avoiding hav-ing a panic attack is but one example.

Emotional Influences on Cognition

Emotions influence cognitive processing in anumber of ways. The amygdala receives inputsfrom the sensory processing in the cortex foreach modality, and projects to these as well(Amaral et al., 1992; McDonald, 1998). Theprojections to the amygdala allow it to detectthe presence of danger, while the amygdala’sprojections to the cortex, which are widespreadand reach the initial stages of sensory process-ing, allow the amygdala to influence corticalprocessing very early in an emotional episode.Thus, once the amygdala is activated by a sen-sory event, it can begin to regulate the corticalareas that project to it, controlling the kinds ofinputs that these areas subsequently send to theamygdala. Through such connections, emo-tions can have direct influences on attentionand perception (see human studies below).

The amygdala also projects (directly or indi-rectly) to various modulatory networks, in-cluding the cholinergic systems in the basalforebrain and brainstem, and the dopaminergicand noradrenergic systems that innervate wide-spread areas of the cortex. For example,Weinberger and colleagues found that the pro-cessing of auditory signals—specifically, theregulation of auditory cortex during fearconditioning—is modulated by cholinergicmodulation (see Weinberger, 1995). They pro-posed that the amygdala plays an importantrole in activating the cholinergic system, whichthen modulates cortical arousal and condition-ing. Experimental studies have shown that thismay occur. Kapp et al. (1992), for example,found that stimulation of the amygdala canchange cortical arousal (as measured byelectroencephalographic patterns), and thatcholinergic blockade prevents this. Further-more, during the presentation of a CS, cells inthe auditory cortex become very active at thetime just preceding the occurrence of the US,and lesions of the amygdala prevent this antici-patory response (Quirk et al., 1997; Armony,

Quirk, & LeDoux, 1998). Through projectionsto the cortex, either directly (see Amaral et al.,1992) or by way of the cholinergic system, theamygdala may direct attentional resources to-ward environmental stimuli that are elicitingamygdala activity. Thus, once the amygdala de-tects danger, it can influence sensory and cogni-tive processing in the cortex by activating thesemodulatory systems.

The amygdala is also involved in triggeringthe release of so-called “stress hormones,” in-cluding epinephrine (adrenaline) and gluco-corticoids, from the adrenal gland. Theglucocorticoids circulate to the brain, wherethey bind to specific receptors in the hippocam-pus, medial prefrontal cortex, other cortical ar-eas, and amygdala. Epinephrine does not enterthe brain directly from the bloodstream, butstudies by McGaugh and colleagues suggestthat it nevertheless influences areas in the brainby changing the activity of the vagus nerve,which then, through several relays in the brain,affects the amygdala and other areas (seeMcGaugh et al., 1995; Cahill & McGaugh,1998; McGaugh, 2004). Their work in bothanimals and humans suggests that peripheralhormones, such as glucocorticoids and epi-nephrine, can modulate the strength of memo-ries (especially conscious or declarative memo-ries) formed in other brain regions (especiallythe hippocampus and neocortex), as discussedbelow when we consider human studies. Inbrief, intermediate levels of these hormonestend to facilitate declarative memory forevents, whereas high levels tend to impair it.This may account for why emotional arousalsometimes leads to an enhancement of memoryand sometimes a memory impairment for anevent. Glucocorticoids are particularly interest-ing for their contrasting effects on theamygdala and hippocampus. These hormonesare released when the amygdala detects danger-ous or otherwise threatening events. When theyreach the brain, they inhibit hippocampus-dependent processes (e.g., spatial memory), butenhance amygdala-dependent processes (e.g.,fear conditioning) (Corodimas et al., 1994;Conrad, Margarinos, et al., 1997; Conrad,Lupien, Thanosoulis, & McEwen, 1997; Kim& Diamond, 2002). Given that conscious ordeclarative memory is one of the prime func-tions of the hippocampus (Squire & Zola,1996; Cohen & Eichenbaum, 1993), we mayconclude that during periods of intense stressthe brain’s ability to form conscious memories


is impaired, but its ability to form unconsciousemotional memories is potentiated. This obser-vation has important implications for under-standing such processes as memory loss duringtrauma and stress, and may account for mem-ory disturbances in patients with depressionand posttraumatic stress disorder.

EMOTIONAL SYSTEMSIN THE HUMAN BRAIN

Investigations of emotion in the human brainhave used animal models of fear learning as astarting point. Findings from studies ofamygdala-damaged patients and from func-tional neuroimaging have demonstrated theimportance of the amygdala in fear condition-ing (Bechara et al., 1995; LaBar, LeDoux,Spencer, & Phelps, 1995; Buchel, Morris,Dolan, & Friston, 1998; LaBar, Gatenby, Gore,LeDoux, & Phelps, 1998). Furthermore, thehippocampus in humans, as in other animals(see above), plays a key role in the expressionof conditioned fear modulated by the learningcontext (LaBar & Phelps, 2005). In addition,the medial prefrontal cortex has been impli-cated in fear regulation in both humans andother animals (Phelps, Delgado, Nearing, &LeDoux, 2004). The existing evidence suggeststhat the neural mechanisms of fear condition-ing are preserved across species, although it isdifficult to study the unique contribution ofdifferent subnuclei of the amygdala in humans.

Fear Acquisition

The similarities in the neural circuitry of fearconditioning among different species providesome assurance when we are speculating abouthuman function on the basis of this animalmodel. However, fear learning can be signifi-cantly more complex than simple fear condi-tioning. In particular, humans and other pri-mates have developed social means ofacquiring fears, which permit learning aboutthe potential aversive properties of stimuli inthe environment without necessarily having toexperience an aversive event directly.

For example, one can learn that a neutralstimulus, such as a blue square, predicts a po-tentially painful shock because the blue squarewas paired with the delivery of a shock. This isan example of simple fear conditioning. How-ever, one could also learn that the blue square

predicts a potential shock by being told aboutthis relationship. In this case, the emotionalproperties of the blue square are communi-cated symbolically, through language. Thistype of instructed fear has been shown to leadto fear reactions to a neutral stimulus that pre-dicts a potentially aversive event (i.e., the“threat” stimulus) comparable to those ob-served with fear conditioning (Hugdahl &Öhman, 1977). In addition, one could learnabout the potential aversive properties of stim-uli in the environment by watching others. If aconspecific receives a shock that is paired witha blue square, this could lead to a fear responseto the blue square when it is presented in a sim-ilar context to the observer. This type of obser-vational fear learning has been shown to be ef-fective in producing robust learned fearresponses in both human and nonhuman pri-mates (Öhman & Mineka, 2001). These socialmeans of fear learning, instruction and obser-vation, allow us to acquire information aboutthe potential aversive properties of stimuli inthe environment without having to undergophysically painful experiences. They can resultin fears that are imagined and anticipated, butnever experienced; these may account for a sig-nificant portion of human fear learning. Thequestions remains: Do socially acquired fearsrely on the same neural circuitry as fear condi-tioning?

Instructed fear results in a symbolic repre-sentation of the emotional properties of a stim-ulus. The acquisition of this cognitive represen-tation probably does not rely on the amygdala.Patients with amygdala damage, whose hippo-campus is intact, can verbally report thesetypes of stimulus contingencies (e.g., a bluesquare predicts a shock), even when they fail toshow physiological evidence of conditionedfear (Bechara et al., 1995; LaBar et al., 1995).On the other hand, patients with hippocampaldamage, whose amygdalas are intact, show theopposite pattern—a normal CR, but an inabil-ity to report the stimulus contingencies verbally(Bechara et al., 1995). These findings suggestthat the retention of an abstract, cognitive rep-resentation of fear may depend on the hippo-campus. Nevertheless, damage to the leftamygdala impairs the physiological expressionof instructed fear (Funayama, Grillon, Davis,& Phelps, 2001), and functional magnetic reso-nance imaging (fMRI) shows enhanced activa-tion in the left amygdala to a “threat” stimulus,relative to one that is instructed to be “safe”


(Phelps et al., 2001). Unlike fear conditioning,where both the right and left amygdala seem tobe important (LaBar et al., 1995; LaBar &LeDoux, 1996), instructed fear depends pri-marily on the left amygdala, perhaps reflectingthe linguistic nature of the representation. Ininstructed fear, it is likely that a symbolic com-munication system (namely, language) is neces-sary for acquisition, that the hippocampalmemory system underlies the retention of thisrepresentation, and that the amygdala mediatesthe physiological expression of this learning.Although humans and other primates mayhave developed unique social means of com-municating fears, the expression of these fearsmay take advantage of the phylogeneticallyshared mechanisms underlying fear condition-ing.

Learning through social observation also de-pends on the amygdala. A recent fMRI study(Olsson & Phelps, 2007) found equally robustactivation of the bilateral amygdala both whena participant observed a conspecific receiving ashock paired with a blue square, and when theobserver was also presented with a blue squarein a similar context. The enhanced blood-oxygenation-level-dependent (BOLD) signal inthe amygdala during observation of a con-specific in a fear conditioning paradigm oc-curred primarily in response to watching theconspecific receive a shock, whereas theamygdala showed an anticipatory BOLD re-sponse to the blue square when the observerbelieved he or she might also receive a shock. Inthis study, the only knowledge of the contin-gency between the colored square and shockwas through social observation. These resultssuggest that learning fears by observing othersrecruits the amygdala to a similar degree as fearconditioning does. It may be that observingothers receive an aversive stimulus is aversiveto the observer as well, and may act as a US.Participants in this study showed a physiologi-cal arousal response both when they watched aconspecific receive a shock and when they be-lieved they might receive a shock themselves. Inthis way, observational fear learning may bemore similar to fear conditioning than learningby instruction. Consistent with this interpreta-tion, observational fear and fear conditioningare both expressed when the CS is presentedsubliminally without awareness. In contrast,instructed fear learning, which depends on asymbolic representation of the stimulus contin-gencies, requires awareness for expression

(Olsson & Phelps, 2004). Functional imagingresults showing amygdala activation in fearconditioning, instructed fear, and observationalfear learning are shown in Pate 10.1 (see colorinsert).

In summary, the neural mechanisms of fearacquisition and expression are similar acrossspecies. Humans and can acquire fears throughsocial means, including instruction and obser-vation. Other primates also have the capacityfor observational fear learning. Although theamygdala may only be necessary for the acqui-sition of some types of socially acquired fears,the amygdala plays a critical role in the physio-logical expression of fears acquired throughfear conditioning, verbal instruction, or socialobservation. Although social learning in pri-mates may be especially well developed, otherspecies do engage in social learning. For exam-ple, rodents use olfactory cues and birds useauditory cues in social learning.

Fear Regulation

Humans also have a range of means to reducefears once they are established. As in other ani-mals, acquired fears can be diminished throughextinction, in which the CS is no longer pairedwith the aversive US and the expression of theconditioned fear eventually diminishes. Asmentioned earlier, animal models of extinctionlearning indicate that the medial prefrontalcortex, particularly the infralimbic cortex, maybe critically involved in extinction. RecentfMRI studies have shown involvement of thisregion during extinction in humans (Plate 10.2;see color insert). Similar to electrophysiologicaldata from rats (Milad & Quirk, 2002), BOLDresponses in the medial prefrontal cortex in-crease as extinction training progresses(Knight, Smith, Cheng, Stein, & Helmstetter,2004; Phelps et al., 2004). This increase inBOLD in the medial prefrontal cortex predictsthe amount of extinction that has occurredwhen assessed after a delay and is inversely cor-related with the responses in the amygdala(Phelps et al., 2004). A recent anatomical MRIstudy demonstrated that the relative corticalsize of this region predicts the rate of extinctionlearning across individuals (Milad et al., 2005).Although it is difficult to specify the preciserole of the medial prefrontal cortex in extinc-tion learning by using correlational humanneuroimaging techniques, the similarities withanimal models suggest that this region might be


inhibiting the amygdala response. As men-tioned earlier, given this importance of the me-dial prefrontal cortex in extinction, it is possi-ble that dysfunction of this region may underliesome fear-related clinical disorders (Rauch,Shin, & Phelps, 2006).

Of course, we humans do not have to exposeourselves repeatedly to a fear-inducing stimulusto diminish a fear response. We can do thesame thing with our minds. Our ability to con-trol and regulate our emotions is critical in nor-mal social behavior and adaptive function. Re-cent studies of emotion regulation haveexplored the neural mechanisms underlying theuse of cognitive strategies to alter fear re-sponses (e.g., Ochsner, Bunge, Gross, &Gabrieli, 2002; Beauregard, Levesque, &Bourgouin, 2001). These studies, which oftenuse stimuli such as negative, arousing scenes,have shown that the successful use of a strategyto control an emotional response results in adecrease in BOLD signal in the amygdala andan increase in regions of the dorsolateralprefrontal cortex. The dorsolateral prefrontalcortex is a region that is anatomically ratherdifferent across species. It may be unique toprimates (Preuss & Goldman-Rakic, 1991),and it is thought to underlie higher cognitivefunctions, such as executive control (Smith &Jonides, 1999). However, this region has fewdirect projections to the amygdala (McDonald,Mascagni, & Guo, 1996; Stefanacci & Amaral,2002). As a result, any influence it may have onthe amygdala is likely to be mediated by itsconnections with more ventral and medial re-gions of the prefrontal cortex. Many studiesexamining the emotion regulation of scenesalso report increased BOLD responses in themedial prefrontal cortex when a negative emo-tional response is diminished through a cogni-tive strategy (e.g., Beauregard et al., 2001; Urryet al., 2006)

In order to explore whether the medialprefrontal cortex mediates the influence of thedorsolateral prefrontal cortex on the amygdalawhen humans are using a cognitive strategy tocontrol fears, a recent study examined the regu-lation of conditioned fear and its relation tofear extinction (Delgado et al., 2004). Muchlike extinction training, the use of a cognitivestrategy can diminish the physiological expres-sion of conditioned fear. Like extinction train-ing, the regulation of conditioned fear resultedin an increase in BOLD signal in the infralimbicregion of the medial prefrontal cortex and a

correlated decrease in the amygdala response.Much as in other studies of emotion regulation(e.g., Ochsner et al., 2002), the regulation ofconditioned fear resulted in increased BOLDsignal in the dorsolateral prefrontal cortex.This increase was more strongly correlatedwith the responses in the medial prefrontal cor-tex than the amygdala, indicating that the me-dial prefrontal cortex may be mediating any in-fluence of the dorsolateral prefrontal cortexand of executive processes on the amygdalaand the physiological expression of fear.

These findings suggest that much as in thecase of fear acquisition, humans may have de-veloped unique cognitive means for controllingfear responses, but probably did not developunique neural mechanisms for controllingfears. The use of complex cognitive strategiesto control fears may take advantage of phylo-genetically shared mechanisms of fear extinc-tion. As mentioned earlier, the medialprefrontal cortex may play a general role inregulating and inhibiting the amygdala re-sponse across a range of paradigms and tech-niques.

Emotional Influence on Perception,Attention, and Memory

Studies of the amygdala in animal models havegenerally focused on its role in fear condition-ing. However, studies in humans have alsohighlighted a role for the amygdala in mediat-ing emotion’s influence on a range of social andcognitive functions.

For example, studies of the perception of fa-cial expression in others have shown that theamygdala is critically involved in the normalperception of fear expressions. Patients with bi-lateral amygdala damage consistently have dif-ficulty indicating the intensity of fear expres-sions (Adolphs et al., 1999), and findings fromfunctional neuroimaging demonstrate thatfearful faces result in stronger activation of theamygdala than faces with other expressions do(Whalen et al., 1998). Two recent studies havespecified the component of fearful faces thatmay engage the amygdala. Using fMRI, thefirst study demonstrated that presenting theeyes of fearful versus happy faces results in ro-bust activation of the amygdala (Whalen et al.,2004). Consistent with this correlational evi-dence suggesting the importance of the eyes indetecting fear expressions, Adolphs et al.(2005) recently demonstrated that patients


with amygdala damage fail to focus on the eyesas much as non-brain-damaged control sub-jects do when viewing fearful faces. Interest-ingly, if amygdala-damaged patients are in-structed to focus on the eyes of a face, theirratings for the intensity of the fear expressionare normal. This only occurs, however, whenthe patients are instructed to focus on the eyes,whereas the control subjects do this spontane-ously. This finding suggests that the amygdalamay be mediating attentional behaviors thatenable the perception of fear from facial ex-pression. In addition to the perception of fearfrom facial expressions, the amygdala has beenshown to be involved in the perception of fearfrom bodily expressions (de Gelder, Snyder,Greve, Gerard, & Hadjikhani, 2004); the abil-ity to determine trustworthiness from facestimuli (Adolphs, Tranel, & Damasio, 1998;Winston, Strange, O’Doherty, & Dolan, 2002);the response to faces from other races (Phelpset al., 2001); and the ability to attribute social,motivational signals to nonsocial stimuli(Heberlein & Adolphs, 2004). These findingsindicate that the amygdala may play a complexrole in the perception of a range of social sig-nals.

The amygdala has also been shown be im-portant in mediating emotion’s influence on anumber of cognitive functions, most notablyattention and memory. Emotion influences at-tention by enhancing the perception of emo-tional events. This was first described byCherry (1953) as the “cocktail party effect,” inwhich an emotionally salient stimulus, suchas one’s name, breaks through a limitedattentional bottleneck to reach awareness andreceive priority in stimulus processing. Morerecently, the facilitation of attention with emo-tion was demonstrated by using the attentionalblink paradigm, which examines the temporallimitations of attention (Raymond, Shapiro, &Arnell, 1992). In this task, when two targetstimuli are presented in a short temporal win-dow, the ability to identify the second target isdiminished. However, if the second target is anarousing stimulus, such as a dirty word, it ismuch more likely to be correctly identified (An-derson, 2005). Patients with amygdala damagefail to show this attenuation of attentionalblink effect with emotion, suggesting that theamygdala is critically mediating emotion’s fa-cilitation of attention (Anderson & Phelps,2001).

Findings from functional neuroimaging sug-gest that one mechanism by which emotion caninfluence attention is through the modulationof visual processing regions (Plate 10.3; seecolor insert). The enhanced amygdala activa-tion observed to fearful versus neutral faces iscorrelated with a similar enhancement in visualcortex (Morris et al., 1998). This enhanced ac-tivation for fearful versus neutral expressionsin the visual cortex is diminished if there isdamage to the amygdala (Vuilleumier, Richard-son, Armony, Driver, & Dolan, 2004). Thesefindings suggest a model by which theamygdala receives information about the emo-tional significance of a stimulus early in stimu-lus processing, and then modulates furtherperceptual processing through its reciprocalconnections with visual cortical regions(Amaral, Behniea, & Kelly, 2003). Consistentwith the suggestion that the amygdala’s influ-ence on attention may be the result of its modu-lation of the visual cortex (Vuilleumier et al.,2004), it was recently demonstrated that thepresentation of a fearful face enhances the de-tection of early perceptual features known tobe coded by early visual cortex—specifically,sensitivity to contrast (Phelps, Ling, &Carrasco, 2006).

The amygdala’s influence on attention andperception ensures that stimuli that are arous-ing and emotionally salient receive priority ininitial stimulus processing. The amygdala’s in-fluence on memory ensures that emotionalevents are also more likely to be rememberedover time. Most memories for episodic eventsdepend on the hippocampal complex for long-term storage, and are not dependent on theamygdala (Squire & Kandel, 1999). However,when an event is emotional and arousing, theamygdala has a specific role in modulating thehippocampus and enhancing memory consoli-dation or storage (McGaugh, 2004). In thisway, emotional events are more likely to be re-tained. Animal models indicate that theamygdala’s modulation of hippocampal con-solidation is dependent on the neurohoro-monal changes that occur with arousal(McGaugh, 2004). In humans, it has been dem-onstrated that amygdala activation at encodingfor arousing stimuli predicts later memory(Canli, Zhao, Brewer, Gabrieli, & Cahill,2000; Hamann, Ely, Grafton, & Kilts, 1999).In addition, using a pharmacological agent thatblocks the effects of arousal on the amygdala


eliminates emotion’s influence on later memory(Cahill, Prins, Weber, & McGaugh, 1994).Finally, patients with amygdala damage fail toshow the normal enhancement of memory foremotional events (Cahill, Babinsky, Marko-witsch, & McGaugh, 1995), and this failure ismore pronounced over time (LaBar & Phelps,1998), consistent with the suggestion that theamygdala is modulating memory consolida-tion.

The human amygdala, which is critical forthe acquisition and expression of fear condi-tioning, also appears to modulate a range of so-cial and cognitive functions through its exten-sive connectivity with cortical regions (Young,Scannell, Burns, & Blakemore, 1994). Theamygdala mediates some forms of social per-ception and has a modulatory effect on a num-ber of cognitive functions, particularly atten-tion and memory. Through this combination ofroles, the amygdala helps ensure that stimulithat come to predict potential threat are alsomore likely to be noticed and remembered.

FEELINGS AND THE BRAIN

How do we become consciously aware that anemotion system in our brains is active? This isthe problem that most emotion studies and the-ories, especially psychological studies of andtheories about humans, have focused on (e.g.,Lewis & Michalson, 1983). However, we haveexamined the neural basis of emotion in thischapter without really mentioning feelings.Our hypothesis is that the mechanism of con-sciousness is the same for emotional andnonemotional subjective states, and that whatdistinguishes these states is the brain systemthat consciousness is aware of at the time. Thisis why we have focused on the nature of the un-derlying emotion systems throughout the chap-ter. Now, however, we consider feelings andtheir relation to consciousness.

Several theorists have proposed that con-sciousness has something to do with a mentalworkspace where things can be compared, con-trasted, and mentally manipulated (Johnson-Laird, 1988; Kihlstrom, 1984; Schacter, 1989;Shallice, 1988; Baars, 1997, 2005; Dehaene &Naccache, 2001; Dehaene, Kerszberg, &Changeux, 1998; Dehaene, Changeux,Naccache, Sackur, & Sergent, 2006). The oper-ation of this workspace is often described in

terms of working memory (Baddeley, 1986,1993, 2000), which includes the capacity tointegrate information across sensory modalitieswith long-term memory into unified represen-tations, to maintain the representation tempo-rarily in an active state, and to use the represen-tation in controlling mental activity andbehavior.

Various studies of humans and nonhumanprimates point to the prefrontal cortex, espe-cially the dorsolateral prefrontal areas, as beinginvolved in working memory processes(Goldman-Rakic, 1987; Fuster, 2000; Smith &Jonides, 1999; D’Esposito et al., 1995; Miller& Cohen, 2001; Ranganath, Johnson, &D’Esposito, 2003; Muller & Knight, 2006).Immediately present stimuli and stored repre-sentations are integrated in working memoryby way of interactions among prefrontal areas,sensory processing systems (which serve asshort-term memory buffers as well as percep-tual processors), and the long-term explicit ordeclarative memory system (involving the hip-pocampus and related areas of the temporallobe). Working memory may involve interac-tions among several prefrontal areas, includingthe anterior cingulate, insular, and orbital cor-tical regions, as well as dorsolateral prefrontalcortex (D’Esposito et al., 1995; Gaffan,Murray, & Fabre-Thorpe, 1993; Smith &Jonides, 1999; Muller & Knight, 2006;Ranganath et al., 2003; Rolls, 1999; LeDoux,2002; Posner & Dehaene, 1994; Pasternak &Greenlee, 2005; Constantinidis & Procyk,2004; Curtis, 2006). That prefrontal cortex isinvolved in conscious experiences is stronglysuggested by studies in humans showing thatsensory stimuli activate sensory cortex andprefrontal areas when the stimulus is con-sciously perceived, but only sensory cortexwhen conscious awareness is blocked by mask-ing or other procedures (Rees, Kreiman, &Koch, 2002; Beck, Rees, Frith, & Lavie, 2001;Frith, Perry, & Lumer, 1999; Frith & Dolan,1996; Cunningham, Raye, & Johnson, 2004;Carter, O’Doherty, Seymour, Koch, & Dolan,2006; Vuilleumier et al., 2002; Critchley,Mathias, & Dolan, 2002; Critchley, Wiens,Rotshtein, Öhman, & Dolan, 2004).

If a stimulus is affectively charged (say, atrigger of fear), the same sorts of processes willbe called upon as for stimuli without emotionalimplications; in addition, however, workingmemory will become aware of the fact that the


fear system of the brain has been activated.This further information, when added to per-ceptual and mnemonic information about theobject or event, may be the condition for thesubjective experience of an emotional state offear (Figure 10.4).

But what is the further information that isadded to working memory when the fear sys-tem is activated? As noted, the amygdala pro-jects to many cortical areas, even some fromwhich it does not receive inputs (Amaral et al.,1992). It can thus influence the operation ofperceptual and short-term memory processes,as well as processes in higher-order areas. Al-though the amygdala does not have extensiveconnections with the dorsolateral prefrontalcortex, it does communicate with the anteriorcingulate, insular, and orbital cortex—othercomponents of the working memory network.But in addition, the amygdala projects tobrainstem monoaminergic systems involved inthe regulation of cortical arousal. And theamygdala controls bodily responses (behavior-al, autonomic, endocrine), which then providefeedback that can influence cortical processingindirectly, such as the adrenal hormones men-

tioned above. Thus working memory receives agreater number of inputs, and receives inputsof a greater variety, in the presence of an emo-tional stimulus than in the presence of otherkinds of stimuli. These extra inputs may just bewhat are required to add affective charge toworking memory representations, and thus toturn subjective experiences into emotional ex-periences.

An alternative view, the somatic marker hy-pothesis, proposes that feelings arise whenbody sensing areas receive feedback aboutemotional arousal (Damasio, 1994). Somaticfeedback and processing by somatosensory ar-eas of the cortex are components of the work-ing memory model as well. However, theworking memory model proposes that somato-sensory cortex activity is not conscious until itis represented in working memory. The studiesdescribed above, showing that prefrontal cor-tex is active during the conscious processing ofexternal stimuli and not active when the stimuliare not consciously perceived, are consistentwith this idea.

Though described in terms of the fear sys-tem, the present hypothesis about feelings is ageneral one that applies to any emotion. Thatis, an emotional feeling results when workingmemory is occupied with the fact that an emo-tion system of the brain is active. The differ-ence between an emotional state and otherstates of consciousness, then, is not due to dif-ferent underlying mechanisms that give rise tothe qualitatively different subjective experi-ences. Instead, there is one mechanism of con-sciousness, and it can be occupied by eithermundane events or emotionally charged ones.

REFERENCES

Adolphs, R., Gosselin, F., Buchanan, T. W., Tranel, D.,Schyns, P., & Damasio, A. R. (2005). A mechanismfor impaired fear recognition after amygdala dam-age. Nature, 433, 68–72.

Adolphs, R., Tranel, D., & Damasio, A. R. (1998). Thehuman amygdala in social judgment. Nature, 393,470–474.

Adolphs, R., Tranel, D., Hamann, S., Young, A. W.,Calder, A. J., Phelps, E. A., et al. (1999). Recognitionof facial emotion in nine individuals with bilateralamygdala damage. Neuropsychologia, 37, 1111–1117.

Aggleton, J. P. (Ed.). (1992). The amygdala: Neurobio-logical aspects of emotion, memory, and mental dys-function. New York: Wiley-Liss.


FIGURE 10.4. A model of how the brain con-sciously experiences emotional stimuli. In thismodel, the conscious experience of the emotionalsignificance of a stimulus involves the integrationin working memory of information about the im-mediately present stimulus, long-term declarativememories about the stimulus, and the emotionalstate of the brain and body.

Aggleton, J. P., & Mishkin, M. (1986). The amygdala:Sensory gateway to the emotions. In R. Plutchik &H. Kellerman (Eds.), Emotion: Theory, research, andexperience (Vol. 3, pp. 281–299). Orlando, FL: Aca-demic Press.

Amaral, D. G., Behniea, H., & Kelly, J. L. (2003). Topo-graphic organization of projections from theamygdala to the visual cortex in the macaque mon-key. Neuroscience, 118, 1099–1120.

Amaral, D. G., Price, J. L., Pitkänen, A., & Carmichael,S. T. (1992). Anatomical organization of the primateamygdaloid complex. In J. P. Aggleton (Ed.), Theamygdala: Neurobiological aspects of emotion,memory, and mental dysfunction (pp. 1–66). NewYork: Wiley-Liss.

Amorapanth, P., LeDoux, J. E., & Nader, K. (2000).Different lateral amygdala outputs mediate reactionsand actions elicited by a fear-arousing stimulus. Na-ture Neuroscience, 3, 74–79.

Anderson, A. K. (2005). Affective influences on theattentional dynamics supporting awareness. Journalof Experimental Psychology: General, 134, 258–281.

Anderson, A. K., & Phelps, E. A. (2001). Lesions of thehuman amygdala impair enhanced perception ofemotionally salient events. Nature, 411, 305–309.

Armony, J. L., Quirk, G. J., & LeDoux, J. E. (1998).Differential effects of amygdala lesions on early andlate plastic components of auditory cortex spiketrains during fear conditioning. Journal of Neurosci-ence, 18, 2592–2601.

Armony, J. L., Servan-Schreiber, D., Cohen, J. C., &LeDoux, J. E. (1996). Emotion and cognition interac-tions in the thalalmo-cortico-amygdala network:Theory and model. Cognitive Neuroscience SocietyAbstracts, 3, 76.

Armony, J. L., Servan-Schreiber, D., Romanski, L. M.,Cohen, J. D., & LeDoux, J. E. (1997). Stimulus gen-eralization of fear responses: Effects of auditory cor-tex lesions in a computational model and in rats. Ce-rebral Cortex, 7, 157–165.

Baars, B. J. (1997). In the theatre of consciousness. Jour-nal of Consciousness Studies, 4, 292–309.

Baars, B. J. (2005). Global workspace theory of con-sciousness: Toward a cognitive neuroscience of hu-man experience. Progress in Brain Research, 150,45–53.

Baddeley, A. (1986). Modularity, mass-action and mem-ory. Quarterly Journal of Experimental PsychologyA, 38, 527–533.

Baddeley, A. (1993). Verbal and visual subsystems ofworking memory. Current Biology, 3, 563–565.

Baddeley, A. (2000). The episodic buffer: A new compo-nent of working memory? Trends in Cognitive Sci-ences, 4, 417–423.

Bard, P. (1929). The central representation of the sym-pathetic system: As indicated by certain physiologicalobservations. Archives of Neurology and Psychiatry,22, 230–246.

Beauregard, M., Levesque, J., & Bourgouin, P. (2001).

Neural correlates of conscious self-regulation ofemotion. Journal of Neuroscience, 21, RC165.

Bechara, A., Tranel, D., Damasio, H., Adolphs, R.,Rockland, C., & Damasio, A. R. (1995). Double dis-sociation of conditioning and declarative knowledgerelative to the amygdala and hippocampus in hu-mans. Science, 269, 1115–1118.

Beck, D. M., Rees, G., Frith, C. D., & Lavie, N. (2001).Neural correlates of change detection and changeblindness. Nature Neuroscience, 4, 645–650.

Blanchard, C. D., & Blanchard, R. J. (1972). Innate andconditioned reactions to threat in rats withamygdaloid lesions. Journal of Comparative andPhysiological Psychology, 81, 281–290.

Bouton, M. E., & Swartzentruber, D. (1991). Sources ofrelapse after extinction in Pavlovian and instrumen-tal learning. Clinical Psychology Review, 11, 123–140.

Bremner, J. D., Randall, T., Scott, T. M., Brunen, R. A.,Seibyl, J. P., Southwick, S. M., et al. (1995). MRI-based measurement of hippocampal volume in pa-tients with combat-related posttraumatic stress disor-der. American Journal of Psychiatry, 152, 973–981.

Brodal, A. (1982). Neurological anatomy. New York:Oxford University Press.

Buchel, C., Morris, J., Dolan, R. J., & Friston, K. J.(1998). Brain systems mediating aversive condition-ing: An event-related fMRI study. Neuron, 20, 947–957.

Cahill, L., Babinsky, R., Markowitsch, H. J., &McGaugh, J. L. (1995). The amygdala and emotionalmemory. Nature, 377, 295–296.

Cahill, L., & McGaugh, J. L. (1998). Mechanisms ofemotional arousal and lasting declarative memory.Trends in Neurosciences, 21, 294–299.

Cahill, L., Prins, B., Weber, M., & McGaugh, J. L.(1994). Beta-adrenergic activation and memory foremotional events. Nature, 371, 702–704.

Campeau, S., & Davis, M. (1995). Involvement ofsubcortical and cortical afferents to the lateral nu-cleus of the amygdala in fear conditioning measuredwith fear-potentiated startle in rats trained concur-rently with auditory and visual conditioned stimuli.Journal of Neuroscience, 15, 2312–2327.

Canli, T., Zhao, Z., Brewer, J., Gabrieli, J. D., & Cahill,L. (2000). Event-related activation in the humanamygdala associates with later memory for individ-ual emotional experience. Journal of Neuroscience,20, RC99.

Cannon, W. B. (1929). Bodily changes in pain, hunger,fear, and rage. New York: Appleton.

Canteras, N. S., & Swanson, L. W. (1992). Projectionsof the ventral subiculum to the amygdala, septum,and hypothalamus: A PHAL anterograde tract-tracing study in the rat. Journal of Comparative Neu-rology, 324, 180–194.

Carrive, P., Lee, J., & Su, A. (2000). Lidocaine blockadeof amygdala output in fear-conditioned rats reducesFos expression in the ventrolateral periaqueductalgray. Neuroscience, 95, 1071–1080.


Carter, R. M., O’Doherty, J. P., Seymour, B., Koch, C.,& Dolan, R. J. (2006). Contingency awareness in hu-man aversive conditioning involves the middle fron-tal gyrus. NeuroImage, 29, 1007–1012.

Cherry, E. C. (1953). Some experiments on the recogni-tion of speech, with one and two ears. Journal of theAcoustical Society of America, 25, 975–979.

Cohen, N. J., & Eichenbaum, H. (1993). Memory, am-nesia, and the hippocampal system. Cambridge, MA:MIT Press.

Conrad, C. D., LeDoux, J. E., Margarinos, A. M., &McEwen, B. S. (1997). Repeated restraint stress fearconditioning independently of causing hippocampalCA3 dendritic atrophyh. Behavioral Neuroscience,113, 902–913.

Conrad, C. D., Lupien, S. J., Thanosoulis, L. C., &McEwen, B. S. (1997). The effects of Type I and TypeII corticosteroid receptor agonists on exploratorybehavior and spatial memory in the Y-maze. BrainResearch, 759, 76–83.

Constantinidis, C., & Procyk, E. (2004). The primateworking memory networks. Cognitive, Affective, andBehavioral Neuroscience, 4, 444–465.

Corodimas, K. P., & LeDoux, J. E. (1995). Disruptiveeffects of posttraining perirhinal cortex lesions onconditioned fear: Contributions of contextual cues.Behavioral Neuroscience, 109, 613–619.

Corodimas, K. P., LeDoux, J. E., Gold, P. W., &Schulkin, J. (1994). Corticosterone potentiation ofconditioned fear in rats. Annals of the New YorkAcademy of Sciences, 746, 392–393.

Cranford, J. L., & Igarashi, M. (1977). Effects of audi-tory cortex lesions on temporal summation in cats.Brain Research, 136, 559–564.

Critchley, H. D., Mathias, C. J., & Dolan, R. J. (2002).Fear conditioning in humans: The influence ofawareness and autonomic arousal on functionalneuroanatomy. Neuron, 33, 653–663.

Critchley, H. D., Wiens, S., Rotshtein, P., Öhman, A., &Dolan, R. J. (2004). Neural systems supportinginteroceptive awareness. Nature Neuroscience, 7,189–195.

Cunningham, W. A., Raye, C. L., & Johnson, M. K.(2004). Implicit and explicit evaluation: FMRI corre-lates of valence, emotional intensity, and control inthe processing of attitudes. Journal of Cognitive Neu-roscience, 16, 1717–1729.

Curtis, C. E. (2006). Prefrontal and parietal contribu-tions to spatial working memory. Neuroscience, 139,173–180.

Damasio, A. (1994). Descartes’ error: Emotion, reason,and the human brain. New York: Putnam.

Davis, M. (1992). The role of the amygdala in condi-tioned fear. In J. P. Aggleton (Ed.), The amygdala:Neurobiological aspects of emotion, memory, andmental dysfunction (pp. 255–306). New York:Wiley-Liss.

Davis, M. (1994). The role of the amygdala in emo-tional learning. International Review of Neurobiolo-gy, 36, 225–266.

Davis, M., & Whalen, P. J. (2001). The amygdala: Vigi-lance and emotion. Molecular Psychiatry, 6, 13–34.

de Gelder, B., Snyder, J., Greve, D., Gerard, G., &Hadjikhani, N. (2004). Fear fosters flight: A mecha-nism for fear contagion when perceiving emotion ex-pressed by a whole body. Proceedings of the NationalAcademy of Sciences of the USA, 101, 16701–16706.

Dehaene, S., Changeux, J. P., Naccache, L., Sackur, J., &Sergent, C. (2006). Conscious, preconscious, andsubliminal processing: A testable taxonomy. Trendsin Cognitive Sciences, 10, 204–211.

Dehaene, S., Kerszberg, M., & Changeux, J. P. (1998).A neuronal model of a global workspace in effortfulcognitive tasks. Proceedings of the National Acad-emy of Sciences of the USA, 95, 14529–14534.

Dehaene, S., & Naccache, L. (2001). Towards a cogni-tive neuroscience of consciousness: Basic evidenceand a workspace framework. Cognition, 79, 1–37.

Delgado, M. R., Trujillo, J. L., Holmes, B., Nearing, K.I., LeDoux, J. E., & Phelps, E. A. (2004). Emotionregulation of conditioned fear: The contributions ofreappraisal. Paper presented at the 11th AnnualMeeting of the Cognitive Neuroscience Society, SanFrancisco.

den Dulk P., Heerebout, B. T., & Phaf, R. H. (2003). Acomputational study into the evolution of dual-routedynamics for affective processing. Journal of Cogni-tive Neuroscience, 15, 194–208.

D’Esposito, M., Detre, J., Alsop, D., Shin, R., Atlas, S.,& Grossman, M. (1995). The neural basis of the cen-tral executive system of working memory. Nature,378, 279–281.

Drevets, W. C., Price, J. L., Simpson, J. R., Jr., Todd, R.D., Reich, T., Vannier, M., et al. (1997). Subgenualprefrontal cortex abnormalities in mood disorders.Nature, 386, 824–827.

Everitt, B. J., & Robbins, T. W. (1992). Amygdala–ventral striatal interactions and reward-related pro-cesses. In J. P. Aggleton (Ed.), The amygdala: Neuro-biological aspects of emotion, memory, and mentaldysfunction (pp. 401–429). New York: Wiley-Liss.

Fendt, M., & Fanselow, M. S. (1999). The neuroana-tomical and neurochemical basis of conditioned fear.Neuroscience and Biobehavioral Reviews, 23, 743–760.

Fernandez de Molina, A., & Hunsperger, R. W. (1962).Organization of the subcortical system governing de-fense and flight reactions in the cat. Journal of Physi-ology, 160, 200–213.

Frankland, P. W., Cestari, V., Filipkowski, R. K., Mc-Donald, R. J., & Silva, A. J. (1998). The dorsal hip-pocampus is essential for context discrimination butnot for contextual conditioning. Behavioral Neuro-science, 112, 863–874.

Frith, C., & Dolan, R. (1996). The role of the prefrontalcortex in higher cognitive functions. Brain Research:Cognitive Brain Research, 5, 175–181.

Frith, C., Perry, R., & Lumer, E. (1999). The neural cor-relates of conscious experience: An experimentalframework. Trends in Cognitive Sciences, 3, 105–114.


Funayama, E. S., Grillon, C., Davis, M., & Phelps, E. A.(2001). A double dissociation in the affective modu-lation of startle in humans: Effects of unilateral tem-poral lobectomy. Journal of Cognitive Neuroscience,13, 721–729.

Fuster, J. M. (2000). The prefrontal cortex of the pri-mate: A synopsis. Psychobiology, 28, 125–131.

Gaffan, D., & Harrison, S. (1987). Amygdalectomy anddisconnection in visual learning for auditory second-ary reinforcement by monkeys. Journal of Neurosci-ence, 7, 2285–2292.

Gaffan, D., Murray, E. A., & Fabre-Thorpe, M. (1993).Interaction of the amygdala with the frontal lobe inreward memory. European Journal of Neuroscience,5, 968–975.

Gaffan, E. A., Gaffan, D., & Harrison, S. (1988). Dis-connection of the amygdala from visual associationcortex impairs visual reward-association learning inmonkeys. Journal of Neuroscience, 8, 3144–3150.

Gewirtz, J. C., & Davis, M. (1997). Second-order fearconditioning prevented by blocking NMDA recep-tors in amygdala. Nature, 388, 471–474.

Gloor, P. (1960). Amygdala. In J. Field, H. W. Magoun,& V. E. Hall (Eds.), Handbook of physiology: Vol. 2.Neurophysiology (pp. 1395–1420). Washington,DC: American Physiological Society.

Goddard, G. (1964). Functions of the amygdala. Psy-chological Review, 62, 89–109.

Goldman-Rakic, P. S. (1987). Circuitry of primateprefrontal cortex and regulation of behavior by rep-resentational memory. In F. Plum (Ed.), Handbook ofphysiology: Section 1. The nervous system. Vol. 5.Higher functions of the brain (pp. 373–418).Bethesda, MD: American Physiological Society.

Gray, J. A. (1987). The psychology of fear and stress.New York: Cambridge University Press.

Halgren, E. (1992). Emotional neurophysiology of theamygdala within the context of human cognition. InJ. P. Aggleton (Ed.), The amygdala: Neurobiologicalaspects of emotion, memory, and mental dysfunction(pp. 191–228). New York: Wiley-Liss.

Hamann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. D.(1999). Amygdala activity related to enhanced mem-ory for pleasant and aversive stimuli. Nature Neuro-science, 2, 289–293.

Heberlein, A. S., & Adolphs, R. (2004). Impaired spon-taneous anthropomorphizing despite intact percep-tion and social knowledge. Proceedings of the Na-tional Academy of Sciences of the USA, 101, 7487–7491.

Helmstetter, F. J., & Bellgowan, P. S. (1994). Effects ofmuscimol applied to the basolateral amygdala on ac-quisition and expression of contextual fear condi-tioning in rats. Behavioral Neuroscience, 108, 1005–1009.

Hilton, S. M., & Zbrozyna, A.W. (1963). Amygdaloidregion for defense reactions and its efferent pathwayto the brainstem. Journal of Physiology, 165, 160–173.

Hugdahl, K., & Öhman, A. (1977). Effects of instruc-

tion on acquisition and extinction of electrodermalresponses to fear-relevant stimuli. Journal of Experi-mental Psychology: Human Learning and Memory,3, 608–618.

Iwata, J., Chida, K., & LeDoux, J. E. (1987). Cardio-vascular responses elicited by stimulation of neuronsin the central amygdaloid nucleus in awake but notanesthetized rats resemble conditioned emotional re-sponses. Brain Research, 418, 183–188.

Jacobs, W. J., & Nadel, L. (1985). Stress-induced recov-ery of fears and phobias. Psychological Review, 92,512–531.


Johnson-Laird, P. N. (1988). The computer and themind: An introduction to cognitive science. Cam-bridge, MA: Harvard University Press.

Jones, B., & Mishkin, M. (1972). Limbic lesions and theproblem of stimulus-reinforcement associations. Ex-perimental Neurology, 36, 362–377.

Kane, F., Coulombe, D., & Miliaressis, E. (1991).Amygdaloid self-stimulation: A movable electrodemapping study. Behavioral Neuroscience, 105, 926–932.

Kapp, B. S., Pascoe, J. P., & Bixler, M. A. (1984). Theamygdala: A neuroanatomical systems approach toits contributions to aversive conditioning. In L. R.Squire & N. Butters (Eds.), Neuropsychology ofmemory (pp. 473–488). New York: Guilford Press.

Kapp, B. S., Whalen, P. J., Supple, W. F., & Pascoe, J. P.(1992). Amygdaloid contributions to conditionedarousal and sensory information processing. In J. P.Aggleton (Ed.), The amygdala: Neurobiological as-pects of emotion, memory, and mental dysfunction(pp. 229–254). New York: Wiley-Liss.

Kihlstrom, J. F. (1984). Conscious, subconscious, un-conscious: A cognitive perspective. In K. S. Bowers &D. Meichenbaum (Eds.), The unconscious reconsid-ered (pp. 149–211). New York: Wiley.

Killcross, S., Robbins, T. W., & Everitt, B. J. (1997). Dif-ferent types of fear-conditioned behaviour mediatedby separate nuclei within amygdala. Nature, 388,377–380.

Kim, J. J., & Diamond, D. M. (2002). The stressed hip-pocampus, synaptic plasticity and lost memories. Na-ture Reviews Neuroscience, 3, 453–462.

Kim, J. J., & Fanselow, M. S. (1992). Modality-specificretrograde amnesia of fear. Science, 256, 675–677.

Kluver, H., & Bucy, P. C. (1937). “Psychic blindness”and other symptoms following bilateral temporallobectomy in rhesus monkeys. American Journal ofPhysiology, 119, 352–353.

Knight, D. C., Smith, C. N., Cheng, D. T., Stein, E. A.,& Helmstetter, F. J. (2004). Amygdala and hippo-campal activity during acquisition and extinction ofhuman fear conditioning. Cognitive, Affective, andBehavioral Neuroscience, 4, 317–325.

Kotter, R., & Meyer, N. (1992). The limbic system: A re-view of its empirical foundation. Behavioural BrainResearch, 52, 105–127.


LaBar, K. S., Gatenby, J. C., Gore, J. C., LeDoux, J. E.,& Phelps, E. A. (1998). Human amygdala activationduring conditioned fear acquisition and extinction: Amixed-trial fMRI study. Neuron, 20, 937–945.

LaBar, K. S., & LeDoux, J. E. (1996). Partial disruptionof fear conditioning in rats with unilateral amygdaladamage: Correspondence with unilateral temporallobectomy in humans. Behavioral Neuroscience,110, 991–997.

LaBar, K. S., LeDoux, J. E., Spencer, D. D., & Phelps, E.A. (1995). Impaired fear conditioning following uni-lateral temporal lobectomy in humans. Journal ofNeuroscience, 15, 6846–6855.

LaBar, K. S., & Phelps, E. A. (1998). Arousal-mediatedmemory consolidation: Role of the medial temporallobe in humans. Psychological Science, 9, 490–493.

LaBar, K. S., & Phelps, E. A. (2005). Reinstatement ofconditioned fear in humans is context dependent andimpaired in amnesia. Behavioral Neuroscience, 119,677–686.

LeDoux, J. E. (1986a). Neurobiology of emotion. In J. E.LeDoux & W. Hirst (Eds.), Mind and brain (pp. 301–358). New York: Cambridge University Press.

LeDoux, J. E. (1986b). Sensory systems and emotion.Integrative Psychiatry, 4, 237–248.

LeDoux, J. E. (1987). Emotion. In F. Plum (Ed.), Hand-book of physiology: Section 1. The nervous system.Vol. 5. Higher functions of the brain (pp. 419–460).Bethesda, MD: American Physiological Society.

LeDoux, J. E. (1991). Emotion and the limbic systemconcept. Concepts in Neuroscience, 2, 169–199.

LeDoux, J. E. (1992). Emotion and the amygdala. In J.P. Aggleton (Ed.), The amygdala: Neurobiological as-pects of emotion, memory, and mental dysfunction(pp. 339–351). New York: Wiley-Liss.


LeDoux, J. E. (2002). Synaptic self: How our brains be-come who we are. New York: Viking.

LeDoux, J. E., Cicchetti, P., Xagoraris, A., & Romanski,L. M. (1990). The lateral amygdaloid nucleus: Sen-sory interface of the amygdala in fear conditioning.Journal of Neuroscience, 10, 1062–1069.

LeDoux, J. E., Farb, C., & Ruggiero, D. A. (1990). Top-ographic organization of neurons in the acousticthalamus that project to the amygdala. Journal ofNeuroscience, 10, 1043–1054.

LeDoux, J. E., & Gorman, J. M. (2001). A call to ac-tion: Overcoming anxiety through active coping.American Journal of Psychiatry, 158, 1953–1955.

LeDoux, J. E., Iwata, J., Cicchetti, P., & Reis, D. J.(1988). Different projections of the centralamygdaloid nucleus mediate autonomic and behav-ioral correlates of conditioned fear. Journal of Neu-roscience, 8, 2517–2529.

LeDoux, J. E., Romanski, L. M., & Xagoraris, A. E.(1989). Indelibility of subcortical emotional memo-ries. Journal of Cognitive Neuroscience, 1, 238–243.

LeDoux, J. E., Ruggiero, D. A., & Reis, D. J. (1985).Projections to the subcortical forebrain from ana-

tomically defined regions of the medial geniculatebody in the rat. Journal of Comparative Neurology,242, 182–213.

LeDoux, J. E., Sakaguchi, A., & Reis, D. J. (1984).Subcortical efferent projections of the medialgeniculate nucleus mediate emotional responses con-ditioned to acoustic stimuli. Journal of Neuroscience,4, 683–698.

Lewis, M., & Michalson, L. (1983). Children’s emo-tions and moods: Developmental theory and mea-surement. New York: Plenum Press.

Li, X. F., Stutzmann, G. E., & LeDoux, J. E. (1996).Convergent but temporally separated inputs to lat-eral amygdala neurons from the auditory thalamusand auditory cortex use different postsynaptic recep-tors: In vivo intracellular and extracellular record-ings in fear conditioning pathways. Learning andMemory, 3, 229–242.

MacLean, P. D. (1949). Psychosomatic disease and the“visceral brain”: Recent developments bearing on thePapez theory of emotion. Psychosomatic Medicine,11, 338–353.

MacLean, P. D. (1952). Some psychiatric implications ofphysiological studies on frontotemporal portion oflimbic system (visceral brain). Electroencephalogra-phy and Clinical Neurophysiology, 4, 407–418.

Majidishad, P., Pelli, D. G., & LeDoux, J. E. (1996).Disruption of fear conditioning to contextual stimulibut not to a tone by lesions of the accessory basal nu-cleus of the amygdala. Society for Neuroscience Ab-stracts, 22, 1116.

Maren, S., Aharonov, G., Stote, D. L., & Fanselow, M.S. (1996). N-methyl-D-aspartate receptors in thebasolateral amygdala are required for both acquisi-tion and expression of conditional fear in rats.Behavioral Neuroscience, 110, 1365–1374.

Maren, S., Anagnostaras, S. G., & Fanselow, M. S.(1998). The startled seahorse: Is the hippocampusnecessary for contextual fear conditioning? Trends inCognitive Sciences, 2, 39–41.

Maren, S., & Fanselow, M. S. (1996). The amygdalaand fear conditioning: Has the nut been cracked?Neuron, 16, 237–240.

Mascagni, F., McDonald, A. J., & Coleman, J. R.(1993). Corticoamygdaloid and corticocortical pro-jections of the rat temporal cortex: A Phaseolusvulgaris leucoagglutinin study. Neuroscience, 57,697–715.

McDonald, A. J. (1998). Cortical pathways to the mam-malian amygdala. Progress in Neurobiology, 55,257–332.

McDonald, A. J., Mascagni, F., & Guo, L. (1996). Pro-jections of the medial and lateral prefrontal corticesto the amygdala: A Phaseolus vulgaris leuco-agglutinin study in the rat. Neuroscience, 71, 55–75.

McGaugh, J. L. (2004). The amygdala modulates theconsolidation of memories of emotionally arousingexperiences. Annual Review of Neuroscience, 27, 1–28.

McGaugh, J. L., Mesches, M. H., Cahill, L., Parent,


M. B., Coleman-Mesches, K., & Salinas, J. A. (1995).Involvement of the amygdala in the regulation ofmemory storage. In J. L. McGaugh, F. Bermudez-Rattoni, & R. A. Prado-Alcala (Eds.), Plasticity inthe central nervous system (pp. 18–39). Mahwah,NJ: Erlbaum.

Milad, M. R., Quinn, B. T., Pitman, R. K., Orr, S. P.,Fischl, B., & Rauch, S. L. (2005). Thickness ofventromedial prefrontal cortex in humans is corre-lated with extinction memory. Proceedings of the Na-tional Academy of Sciences of the USA, 102, 10706–10711.

Milad, M. R., & Quirk, G. J. (2002). Neurons in medialprefrontal cortex signal memory for fear extinction.Nature, 420, 70–74.

Miller, E. K., & Cohen, J. D. (2001). An integrative the-ory of prefrontal cortex function. Annual Review ofNeuroscience, 24, 167–202.

Miserendino, M. J., Sananes, C. B., Melia, K. R., & Da-vis, M. (1990). Blocking of acquisition but not ex-pression of conditioned fear-potentiated startle byNMDA antagonists in the amygdala. Nature, 345,716–718.

Mishkin, M., & Aggleton, J. (1981). Multiple func-tional contributions of the amygdala in the monkey.In Y. Ben-Ari (Ed.), The amygdaloid complex (pp.409–420). Amsterdam: Elsevier/North-Holland.

Mishkin, M., Malamut, B., & Bachevalier, J. (1984).Memories and habits: Two neural systems. In G.Lynch, J. L. McGaugh, & N. M. Weinberger (Eds.),Neurobiology of learning and memory (pp. 65–77).New York: Guilford Press.

Morgan, M. A., & LeDoux, J. E. (1995). Differentialcontribution of dorsal and ventral medial prefrontalcortex to the acquisition and extinction of condi-tioned fear in rats. Behavioral Neuroscience, 109,681–688.

Morgan, M. A., Romanski, L. M., & LeDoux, J. E.(1993). Extinction of emotional learning: Contribu-tion of medial prefrontal cortex. Neuroscience Let-ters, 163, 109–113.

Morris, J. S., Friston, K. J., Buchel, C., Frith, C. D.,Young, A. W., Calder, A. J., et al. (1998). Aneuromodulatory role for the human amygdala inprocessing emotional facial expressions. Brain,121(Pt. 1), 47–57.

Mowrer, O. H., & Lamoreaux, R. R. (1946). Fear as anintervening variable in avoidance conditioning. Jour-nal of Comparative Psychology, 39, 29–50.

Muller, J., Corodimas, K. P., Fridel, Z., & LeDoux, J. E.(1997). Functional inactivation of the lateral andbasal nuclei of the amygdala by muscimol infusionprevents fear conditioning to an explicit conditionedstimulus and to contextual stimuli. Behavioral Neu-roscience, 111, 683–691.

Muller, N. G., & Knight, R. T. (2006). The functionalneuroanatomy of working memory: Contributions ofhuman brain lesion studies. Neuroscience, 139, 51–58.

Ochsner, K. N., Bunge, S. A., Gross, J. J., & Gabrieli, J.D. (2002). Rethinking feelings: An FMRI study of the

cognitive regulation of emotion. Journal of CognitiveNeuroscience, 14, 1215–1229.


O’Keefe, J., & Nadel, L. (1978). The hippocampus as acognitive map. Oxford: Clarendon Press.

Olds, J. (1977). Drives and reinforcement. New York:Raven Press.

Olsson, A., & Phelps, E. A. (2004). Learned fear of “un-seen” faces after Pavlovian, observational, and in-structed fear. Psychological Science, 15, 822–828.

Olsson, A., & Phelps, E. A. (2007). Social learning offear. Nature Neuroscience, 10, 1095–1102.

Ono, T., & Nishijo, H. (1992). Neurophysiological ba-sis of the Kluver–Bucy syndrome: Responses of mon-key amygdaloid neurons to biologically significantobjects. In J. P. Aggleton (Ed.), The amygdala: Neu-robiological aspects of emotion, memory, and mentaldysfunction (pp. 167–190). New York: Wiley-Liss.

Ottersen, O. P. (1982). Connections of the amygdala ofthe rat: IV. Corticoamygdaloid and intraamygdaloidconnections as studied with axonal transport ofhorseradish peroxidase. Journal of ComparativeNeurology, 205, 30–48.

Packard, M. G., Williams, C. L., Cahill, L., &McGaugh, J. L. (1995). The anatomy of a memorymodulatory system: From periphery to brain. In N.E. Spear, L. P. Spear, & M. L. Woodruff (Eds.),Neurobehavioral plasticity: Learning, development,and response to brain insults (pp. 149–150).Hillsdale, NJ: Erlbaum.

Papez, J. W. (1937). A proposed mechanism of emotion.Archives of Neurology and Psychiatry, 79, 217–224.

Pare, D., & Smith, Y. (1998). Intrinsic circuitry of theamygdaloid complex: Common principles of organi-zation in rats and cats. Trends in Neurosciences, 21,240–241.

Parent, M. B., Tomaz, C., & McGaugh, J. L. (1992). In-creased training in an aversively motivated taskattenuates the memory-impairing effects of post-training N-methyl-D-aspartate-induced amygdala le-sions. Behavioral Neuroscience, 106, 789–798.

Pasternak, T., & Greenlee, M. W. (2005). Workingmemory in primate sensory systems. Nature ReviewsNeuroscience, 6, 97–107.

Pavlov, I. P. (1960). Conditioned reflexes (G. V. Anrep,Trans.) New York: Dover. (Original work published1927)

Phelps, E. A., Delgado, M. R. Nearing, K. I., & LeDoux,J. E. (2004). Extinction learning in humans: Role ofthe amygdala and vmPFC. Neuron, 43, 897–905.

Phelps, E. A., Ling, S., & Carrasco, M. (2006). Emotionfacilitates perception and potentiates the perceptualbenefits of attention. Psychological Science, 17, 292–299.

Phelps, E. A., O’Connor, K. J., Gatenby, J. C., Gore, J.C., Grillon, C., & Davis, M. (2001). Activation ofthe left amygdala to a cognitive representation offear. Nature Neuroscience, 4, 437–441.


Phillips, R. G., & LeDoux, J. E. (1992). Differentialcontribution of amygdala and hippocampus to cuedand contextual fear conditioning. Behavioral Neuro-science, 106, 274–285.

Phillips, R. G., & LeDoux, J. E. (1994). Lesions of thedorsal hippocampal formation interfere with back-ground but not foreground contextual fear condi-tioning. Learning and Memory, 1, 34–44.

Pitkänen, A., Savander, V., & LeDoux, J. E. (1997). Or-ganization of intra-amygdaloid circuitries in the rat:An emerging framework for understanding functionsof the amygdala. Trends in Neurosciences, 20, 517–523.

Posner, M. I., & Dehaene, S. (1994). Attentional net-works. Trends in Neurosciences, 17, 75–79.

Preuss, T. M., & Goldman-Rakic, P. S. (1991).Ipsilateral cortical connections of granular frontalcortex in the strepsirhine primate Galago, with com-parative comments on anthropoid primates. Journalof Comparative Neurology, 310, 507–549.

Quirk, G. J., Armony, J. L., & LeDoux, J. E. (1997).Fear conditioning enhances different temporal com-ponents of tone-evoked spike trains in auditory cor-tex and lateral amygdala. Neuron, 19, 613–624.

Quirk, G. J., Garcia, R., & Gonzalez-Lima, F. (2006).Prefrontal mechanisms in extinction of conditionedfear. Biological Psychiatry, 60, 337–343.

Quirk, G. J., Russo, G. K., Barron, J. L., & Lebron, K.(2000). The role of ventromedial prefrontal cortex inthe recovery of extinguished fear. Journal of Neuro-science, 20, 6225–6231.

Radley, J. J., & Morrison, J. H. (2005). Repeated stressand structural plasticity in the brain. Ageing Re-search Reviews, 4, 271–287.

Radley, J. J., Rocher, A. B., Miller, M., Janssen, W. G.,Liston, C., Hof, P. R., et al. (2006). Repeated stressinduces dendritic spine loss in the rat medialprefrontal cortex. Cerebral Cortex, 16, 313–320.

Ranganath, C., Johnson, M. K., & D’Esposito, M.(2003). Prefrontal activity associated with workingmemory and episodic long-term memory.Neuropsychologia, 41, 378–389.

Rauch, S. L., Shin, L. M., & Phelps, E. A. (2006).Neurocircuitry models of posttraumatic stress disor-der and extinction: Human neuroimaging research—past, present, and future. Biological Psychiatry, 60,376–382.

Raymond, J. E., Shapiro, K. L., & Arnell, K. M. (1992).Temporary suppression of visual processing in anRSVP task: An attentional blink? Journal of Experi-mental Psychology: Human Perception and Perfor-mance, 18, 849–860.

Rees, G., Kreiman, G., & Koch, C. (2002). Neural cor-relates of consciousness in humans. Nature ReviewsNeuroscience, 3, 261–270.

Rescorla, R. A., & Heth, C. D. (1975). Reinstatement offear to an extinguished conditioned stimulus. Journalof Experimental Psychology: Animal Behavior Pro-cesses, 104, 88–96.

Rescorla, R. A., & Solomon, R. L. (1967). Two process

learning theory: Relationships between Pavlovianconditioning and instrumental learning. Psychologi-cal Review, 74, 151–182.

Rodrigues, S. M., Schafe, G. E., & LeDoux, J. E. (2004).Molecular mechanisms underlying emotional learn-ing and memory in the lateral amygdala. Neuron, 44,75–91.

Rogan, M. T., Staubli, U. V., & LeDoux, J. E. (1997).AMPA receptor facilitation accelerates fear learningwithout altering the level of conditioned fear ac-quired. Journal of Neuroscience, 17, 5928–5935.

Rolls, E. T. (1986). A theory of emotion, and its applica-tion to understanding the neural basis of emotion. InY. Oomur (Ed.), Emotions: Neural and chemical con-trol (pp. 325–344). Tokyo: Japan Scientific SocietiesPress.

Rolls, E. T. (1992). Neurophysiological mechanisms un-derlying face processing within and beyond the tem-poral cortical visual areas. Philosophical Transac-tions of the Royal Society of London: Series B.Biological Sciences, 335, 11–21.

Rolls, E. T. (1999). The brain and emotion. Oxford:Oxford University Press.

Romanski, L. M., & LeDoux, J. E. (1992). Bilateral de-struction of neocortical and perirhinal projection tar-gets of the acoustic thalamus does not disrupt audi-tory fear conditioning. Neuroscience Letters, 142,228–232.

Romanski, L. M., & LeDoux, J. E. (1993). Informationcascade from primary auditory cortex to theamygdala: Corticocortical and corticoamygdaloidprojections of temporal cortex in the rat. CerebralCortex, 3, 515–532.

Rudy, J. W., Huff, N. C., & Matus-Amat, P. (2004). Un-derstanding contextual fear conditioning: Insightsfrom a two-process model. Neuroscience andBiobehavioral Reviews, 28, 675–685.

Schacter, D. L. (1989). On the relation between memoryand consciousness: Dissociable interactions and con-scious experience. In H. L. I. Roediger & F. I. M.Craik (Eds.), Varieties of memory and consciousness:Essays in honour of Endel Tulving (pp. 355–389).Hillsdale, NJ: Erlbaum.

Selden, N. R., Everitt, B. J., Jarrard, L. E., & Robbins,T. W. (1991). Complementary roles for the amygdalaand hippocampus in aversive conditioning to explicitand contextual cues. Neuroscience, 42, 335–350.

Shallice, T. (1988). Information processing models ofconsciousness. In A. Marcel & E. Bisiach (Eds.),Consciousness in contemporary science (pp. 305–333). Oxford: Oxford University Press.

Shors, T. J. (2006). Stressful experience and learningacross the lifespan. Annual Review of Psychology,57, 55–85.

Smith, E. E., & Jonides, J. (1999). Storage and executiveprocesses in the frontal lobes. Science, 283, 1657–1661.

Spiegler, B. J., & Mishkin, M. (1981). Evidence for thesequential participation of inferior temporal cortexand amygdala in the acquisition of stimulus–reward


associations. Behavioural Brain Research, 3, 303–317.

Squire, L. R., & Kandel, E. R. (1999). Memory: Frommind to molecules. New York: Scientific AmericanLibrary.

Squire, L. R., & Zola, S. M. (1996). Structure and func-tion of declarative and nondeclarative memory sys-tems. Proceedings of the National Academy of Sci-ences of the USA, 93, 13515–13522.

Stefanacci, L., & Amaral, D. G. (2002). Some observa-tions on cortical inputs to the macaque monkeyamygdala: An anterograde tracing study. Journal ofComparative Neurology, 451, 301–323.

Sutherland, R. J., & Rudy, J. W. (1989). Configural as-sociation theory: The role of the hippocampal forma-tion in learning, memory, and amnesia.Psychobiology, 17, 129–144.

Swanson, L. W. (1983). The hippocampus and the con-cept of the limbic system. In W. Seifert (Ed.), Neuro-biology of the hippocampus (pp. 3–19). London: Ac-ademic Press.

Teich, A. H., McCabe, P. M., Gentile, C. C.,Schneiderman, L. S., Winters, R. W., Liskowsky, D.R., et al. (1989). Auditory cortex lesions prevent theextinction of Pavlovian differential heart rate condi-tioning to tonal stimuli in rabbits. Brain Research,480, 210–218.

Thorpe, S. J., Rolls, E. T., & Maddison, S. (1983). Theorbitofrontal cortex: Neuronal activity in the behav-ing monkey. Experimental Brain Research, 49, 93–115.

Turner, B., & Herkenham, M. (1991). Thalamo-amygdaloid projections in the rat: A test of theamygdala’s role in sensory processing. Journal ofComparative Neurology, 313, 295–325.

Urry, H. L., van Reekum, C. M., Johnstone, T., Kalin,N. H., Thurow, M. E., Schaefer, H. S., et al. (2006).Amygdala and ventromedial prefrontal cortex are in-versely coupled during regulation of negative affectand predict the diurnal pattern of cortisol secretionamong older adults. Journal of Neuroscience, 26,4415–4425.

Vuilleumier, P., Armony, J. L., Clarke, K., Husain, M.,Driver, J., & Dolan, R. J. (2002). Neural response toemotional faces with and without awareness: Event-

related fMRI in a parietal patient with visual extinc-tion and spatial neglect. Neuropsychologia, 40,2156–2166.

Vuilleumier, P., Richardson, M. P., Armony, J. L., Driver,J., & Dolan, R. J. (2004). Distant influences ofamygdala lesion on visual cortical activation duringemotional face processing. Nature Neuroscience, 7,1271–1278.

Weinberger, N. M. (1995). Retuning the brain by fearconditioning. In M. S. Gazzaniga (Ed.), The cognitiveneurosciences (pp. 1071–1090). Cambridge, MA:MIT Press.

Weiskrantz, L. (1956). Behavioral changes associatedwith ablation of the amygdaloid complex in mon-keys. Journal of Comparative and Physiological Psy-chology, 49, 381–391.

Whalen, P. J., Kagan, J., Cook, R. G., Davis, F. C., Kim,H., Polis, S., et al. (2004). Human amygdalaresponsivity to masked fearful eye whites. Science,306, 2061.

Whalen, P. J., Rauch, S. L., Etcoff, N. L., McInerney, S.C., Lee, M. B., & Jenike, M. A. (1998). Masked pre-sentations of emotional facial expressions modulateamygdala activity without explicit knowledge. Jour-nal of Neuroscience, 18, 411–418.

Whitfield, I. C. (1980). Auditory cortex and the pitch ofcomplex tones. Journal of the Acoustical Society ofAmerica, 67, 644–647.

Wilensky, A. E., Schafe, G. E., & LeDoux, J. E. (1999).Functional inactivation of the amygdala before butnot after auditory fear conditioning prevents memoryformation. Journal of Neuroscience, 19, RC48.

Wilensky, A. E., Schafe, G. E., & LeDoux, J. E. (2000).The amygdala modulates memory consolidation offear-motivated inhibitory avoidance learning but notclassical fear conditioning. Journal of Neuroscience,20, 7059–7066.

Winston, J. S., Strange, B. A., O’Doherty, J., & Dolan,R. J. (2002). Automatic and intentional brain re-sponses during evaluation of trustworthiness offaces. Nature Neuroscience, 5, 277–283.

Young, M. P., Scannell, J. W., Burns, G. A., &Blakemore, C. (1994). Analysis of connectivity: Neu-ral systems in the cerebral cortex. Reviews in theNeurosciences, 5, 227–250.


C H A P T E R 1 1

The Psychophysiologyof Emotion

JEFF T. LARSEN, GARY G. BERNTSON, KIRSTEN M. POEHLMANN,TIFFANY A. ITO, and JOHN T. CACIOPPO

You see me here, you gods, a poor old man,As full of grief as age; wretched in both!If it be you that stir these daughters’ heartsAgainst their father, fool me not so muchTo bear it tamely; touch me with noble anger.

—SHAKESPEARE, King Lear, II.5

Just as Shakespeare used words to give voice toKing Lear’s outrage at his daughters’ betrayal,psychology has often used the words that peo-ple use to describe their emotions to under-stand those emotions (e.g., Green, Salovey, &Truax, 1999; Osgood, Suci, & Tannenbaum,1957; cf. Cacioppo, Gardner, & Berntson,1999; LeDoux, 1996). Psychology’s reliance onself-reports has been remarkably successful interms of uncovering functional relationshipsbetween different variables. The words peopleuse to describe their current emotions typicallyshow consistent patterns (e.g., Russell, 1980).Those who report feeling happy, for instance,are unlikely to report feeling sad (Russell &Carroll, 1999; but see Larsen, McGraw, &Cacioppo, 2001). In addition, situational ma-nipulations reliably elicit different patterns ofself-reported emotional states. People generally

report feeling happy when hearing fast music ina major key, and sad when hearing slow musicin a minor key (Gagnon & Peretz, 2003).Moreover, children as young as 6 years of agecan find words that describe not only theiremotions, but the putative causes of their emo-tions (e.g., Stein & Levine, 1987).

One factor that allows people to use wordsto convey their emotions is the fact that theworld’s languages are rich in emotional terms(e.g., Clore, Ortony, & Foss, 1987; Russell,1978), which are supplemented with evenricher metaphors (Fainsilber & Ortony, 1987;Hoffman, Waggoner, & Palermo, 1991). Yetwords have their limitations. They are some-times used to deceive rather than inform (e.g.,DePaulo et al., 2003). Even when used in ear-nest, they are subject to contextual distortionsthat operate outside of awareness (e.g.,

180

Schwarz & Clore, 1983). One implication isthat emotions are not merely cerebral; they areembodied. Indeed, Lear’s emotions are mostpowerfully conveyed not when his lines areread, but when they are portrayed by an actoron the stage. It is perhaps not surprising, then,that many of the metaphors people use to ex-press emotion involve bodily sensations. Peopleexpress anxiety by remarking that they havebutterflies in their stomachs, or anger by re-marking that their blood is boiling. Of course,such metaphors are just that. They are notemotions; they merely symbolize emotions—albeit imperfectly.

Dating back to Freud, research from clinicalpsychology has revealed that some aspects ofemotional states can be reported and that oth-ers cannot (e.g., Bradley, 2000; Lang, 1971). Inaddition, research from the neurosciences (e.g.,Gazzaniga & LeDoux, 1978; Tranel &Damasio, 2000) and social psychology (e.g.,Winkielman, Berridge, & Wilbarger, 2005) hasshown that emotional processes can occur inthe absence of emotional experience, which is aprerequisite for the verbal expression of emo-tion. Thus self-reports of emotion tend to beonly modestly related to somatovisceral andbehavioral aspects of emotion (e.g., Bradley &Lang, 2000a; but see Mauss, Levenson,McCarter, Wilhelm, & Gross, 2005). One goalof the psychophysiology of emotion, then, is toinvestigate the physiological processes bywhich emotion is embodied, and thereby tocomplement and clarify insights gleaned fromratings of emotions that people are willing andable to report. A second goal is to usepsychophysiological methods to decomposethe component processes that result in an emo-tional experience or response (Cacioppo &Petty, 1987).

Most contemporary definitions of emotionshare several features. Most important is thenotion that emotions consist of affective,valenced (i.e., positive and/or negative) reac-tions to meaningful stimuli (Frijda, 1994). Yetnot all valenced reactions constitute emotions.Emotions typically are directed at particularobjects, unlike moods, which tend to be morediffuse (Frijda, 1994). Emotions also tend to beshort-lived, lasting on the order of seconds tominutes (Ekman, 1994). Other affective reac-tions, such as moods (Ekman, 1994; Frijda,1994) and especially attitudes (cf. sentiments;Frijda, 1994), tend to be more enduring (Eagly& Chaiken, 1993). Thus individuals who fear

dogs do not live in a constant state of fear, butthe appearance of a dog is likely to elicit fear(Frijda, Chapter 5, this volume).

Fear can be conceptualized as a discreteemotion, along with anger, sadness, happiness,relief, and other states that putatively differ interms of a variety of factors, including anteced-ent appraisals (e.g., Scherer, Schorr, & John-stone, 2001), facial expressions (Ekman,1994), and action tendencies (Frijda, 1986).Alternatively, emotions can be conceptualizedin terms of a small number of dimensions, suchas valence and arousal (Russell & Barrett,1999), positive and negative activation (Wat-son, Wiese, Vaidya, & Tellegen, 1999), orpositivity and negativity (Cacioppo et al.,1999). Though a review of the relative meritsof each approach is beyond the scope of thischapter, we have argued that discrete and di-mensional approaches are not incommensura-ble, and that both approaches can be useful(Cacioppo et al., 1999). Indeed, many of thefindings reviewed in this chapter indicate thatthe psychophysiological substrates of emotionare organized in terms of dimensions (e.g., va-lence), but other findings highlight the utility ofpostulating the existence of discrete emotions.

OVERVIEW

The nervous system is divided into two broadcomponents. The central nervous system com-prises the brain and spinal cord, which sendand receive inputs from the peripheral nervoussystem. In turn, the peripheral nervous systemcomprises the autonomic and somatic nervoussystems. The autonomic nervous systeminnervates smooth muscles (e.g., the heart) andglands, and is divided into the sympathetic andparasympathetic branches. Whereas the sym-pathetic branch generally prepares the body foraction (e.g., by stimulating heart rate), theparasympathetic branch aids restorative func-tions (e.g., by stimulating digestion). Finally,the somatic nervous system innervates skeletalmuscles, including those of the face.

Common to all definitions of psycho-physiology is that it involves the use of physio-logical signals to understand psychologicalprocesses (e.g., Cacioppo, Tassinary, &Berntson, 2007; Coles, Donchin, & Porges,1986; Stern, Ray, & Quigley, 2001). Cognitivepsychophysiology, for instance, is largely con-cerned with mental processes and has therefore

11. The Psychophysiology of Emotion 181

relied heavily on the electroencephalogram(EEG), which reflects electrical activity under-lying central nervous system processes (e.g.,Coles et al., 1986). So extensive is the embodi-ment of emotion, however, that surface record-ings of both central and peripheral nervoussystem activity have shed light on the physio-logical substrates of emotion. Early theory andresearch on the psychophysiology of emotionfocused on the autonomic portion of the pe-ripheral nervous system, so we begin there. Wethen turn to the somatic nervous system, whereresearch has investigated relationships betweenemotional states and activation of the facialmusculature. Finally, we turn to the central ner-vous system, where EEG research has helpeddelineate the neural mechanisms underlyingemotion.

THE PERIPHERAL NERVOUS SYSTEM

As mentioned above, many metaphors used todescribe emotions involve bodily states. Mostof those metaphors make reference to per-ipheral physiological processes in particular.Though most individuals in Western culturespresumably recognize that the heart is in realitya mere pump, it is often described as the seat ofemotion in general and love in particular. Thusour hearts go aflutter when we are in love, butbreak when our beloved departs. Though intu-ition about psychological phenomena is oftenimprecise at best and inaccurate at worst,psychophysiology has demonstrated that pe-ripheral processes in the form of autonomic(i.e., visceral) and somatic (e.g., motor, expres-sive) events are indeed associated with emo-tional processes.

The Autonomic Nervous System

The rich connections between the central andperipheral nervous systems have long fueleddebates over the precise role of peripheral ac-tivity in emotion. These connections are recip-rocal, such that the central nervous systemboth sends and receives input from the viscera.The fact that the brain sends efferents to the pe-riphery raises the possibility that the experienceof emotions (i.e., “feelings”) cause peripheralchanges. On the other hand, the fact that thecentral nervous receives afferents from the pe-riphery raises the perhaps more counter-intuitive possibility that the peripheral changes

contribute to the experience of emotion. Theformer possibility was set forth by WilliamJames (1884, 1890/1950) more than a centuryago. By this account, specific somatovisceralpatterns can not only precede but also generatethe experience of emotion. In an early critiqueof James’s position, however, Cannon (1927)argued that different autonomic patterns donot produce different emotions; rather, differ-ent emotions produce different autonomic pat-terns.

Schachter and Singer (1962) proposed an al-ternative to both James (1884, 1890/1950) andCannon (1927). Like James, Schachter andSinger contended that autonomic activity couldgenerate different emotions. Unlike James,however, they also suggested that the same pat-tern of autonomic activity could result in theexperience of different emotions (see alsoMandler, 1975). By this account, the percep-tion of neutral, unexplained physiologicalarousal creates an “evaluative need” that moti-vates the individual to come up with a cogni-tive label for the arousal state. Depending onsituational cues, the individual may come toexperience different emotional states (e.g., an-ger, euphoria; Schachter & Singer, 1962; cf.Reisenzein, 1983).

Thus theories have proposed three distinctrelationships between somatovisceral activityand emotional state. Whereas Cannon (1927)contended that different emotions produce dif-ferent patterns of somatovisceral activity,James (1884) contended that different patternsof somatovisceral activity can produce differ-ent emotions. Finally, Schachter and Singer(1962) contended that even undifferentiatedsomatovisceral activity can produce distinctand different emotions depending on the situa-tional context. Several subsequent theories ofemotion can be seen as falling within each ofthese camps. Damasio’s (1994) somatic markerhypothesis and Barrett’s (2006) somatic markerhypothesis and conceptual act model of emo-tion, for instance, bear resemblance to the con-ceptualizations of James (1884) and Schachterand Singer (1962), respectively. Researchershave marshaled empirical evidence both forand against each position. Vianna, Weinstock,Elliott, Summers, and Tranel (2006) studied in-dividuals with Crohn’s disease, which is associ-ated with excessive visceral outflow from thegut. As both James and Schachter and Singerwould expect, individuals with Crohn’s diseaseshowed greater gastric myoelectrical activity


during emotionally evocative films than didcomparison participants, which may help ex-plain why they also experienced more sub-jective arousal. (For a review of work oninteroception and emotion, see Craig, Chapter16, this volume.) Indirect evidence for James’sposition comes from Barrett, Quigley, Bliss-Moreau, and Aronson’s (2004) recent findingthat individuals who are more sensitive to theirheartbeats also tend to emphasize the arousal-related meanings of emotion words (see alsoCritchley, Wiens, Rotshtein, Öhman, & Dolan,2004).

On the other hand Cannon (1927) providedevidence from animal studies that autonomicevents are too slow, insensitive, and undifferen-tiated to influence emotions. More recently,neuropsychological research with humans sug-gests that spinal cord injuries have little impacton individuals’ experience of emotion(Chwalisz, Diener, & Gallagher, 1988; see alsoCobos, Sánchez, Garcia, Vera, & Vila, 2002).Perhaps most problematic for James’s hypothe-sis was that early research failed to demon-strate replicable and generalizable emotion-specific autonomic patterns. Ekman, Levenson,and Friesen (1983; see also Levenson, 1988) at-tributed such inconsistency to a variety ofmethodological limitations. For instance, pre-vious investigators made no attempt to equatethe intensity of different emotions, and theyfailed to synchronize physiological recordingswith the likely onset and offset of the emotion,or even to collect self-reports or behavioral ob-servations to confirm that the expected emo-tions had been elicited. Furthermore, Ekmanand colleagues argued that differentiation re-quires simultaneous examination of a numberof indices of autonomic nervous system activ-ity.

Ekman et al. (1983) overcame these limita-tions in a groundbreaking article published inScience that provided evidence for emotion-specific autonomic patterns. Ekman et al. mea-sured heart rate, finger temperature, skin resis-tance, and forearm flexor muscle tension asparticipants completed two sets of tasks de-signed to elicit anger, fear, sadness, happiness,surprise, and disgust. In one task participantswere asked to remember and relive past emo-tional episodes. Participants also completed adirected facial action task in which they wereasked to contract sets of muscles to produce fa-cial expressions associated with each emotion.During the anger condition, for instance, par-

ticipants were asked to pull their eyebrowsdown and together, raise their upper eyelids,push the lower lips up, and press their lips to-gether.

Ekman et al. (1983) found that in additionto differentiating positive from negative emo-tions, combinations of autonomic measurescould differentiate some negative emotions(e.g., fear) from others (e.g., anger). During thefacial action task, for instance, happiness wascharacterized by decreased heart rate, anger byincreased heart rate and increased skin temper-ature, and fear by increased heart rate and de-creased skin temperature. Based on such find-ings, Levenson, Ekman, and Friesen (1990)proposed that each discrete emotion is associ-ated with an innate affect program that coordi-nates changes in the organism’s biologicalstates. They further argued that such changessupport the behavioral reactions most often as-sociated with particular emotions (e.g., fleeing,in the case of fear).

Levenson and colleagues’ work sparked aflurry of research on emotion-specific auto-nomic patterning, which has been reviewed bynumerous researchers (e.g., Wagner, 1989).Different reviewers, however, have come tostrikingly different conclusions (e.g., Levenson,1992; Zajonc & McIntosh, 1992). Levensoncontended that the literature provided compel-ling evidence for emotion-specific autonomicresponding. For instance, he reviewed evidencethat sadness is associated with greater heartrate acceleration than anger and fear, which areboth associated with greater acceleration thanhappiness. Disgust, on the other hand, is asso-ciated with heart rate decelerations. Zajoncand McIntosh, however, contended that theevidence for emotion-specific autonomic re-sponding was inconsistent at best.

In an attempt to resolve the debate, we(Cacioppo, Berntson, Klein, & Poehlmann,1997; Cacioppo, Berntson, Larsen, Poehl-mann, & Ito, 2000) conducted meta-analysesof all published studies comparing the effects ofat least two discrete emotions on at least twoautonomic measures. Though meta-analysesallow one to examine statistically the differen-tiation of discrete emotions by individual mea-sures, it is possible that discrete emotions differin terms of the patterns of autonomic re-sponses, even if they do not differ in terms ofeach isolated autonomic response. Nonethe-less, the literature contains claims that emo-tions can be differentiated with individual au-


tonomic measures, and such hypotheses can berigorously tested with meta-analysis.

The results of our (Cacioppo et al., 2000)meta-analysis were consistent with some ofLevenson’s (1992) claims. For instance, anger,fear, and sadness were associated with greaterheart rate acceleration than disgust. Anger wasalso associated with higher diastolic bloodpressure than fear, and disgust was associatedwith larger increases in skin conductance levelthan happiness. Other claims were not sup-ported. For instance, fear was not associatedwith larger decreases in finger temperaturethan control conditions, nor was fear associ-ated with larger increases in skin conductancelevel. Several other reliable results emergedfrom our (Cacioppo et al., 2000) meta-analyses. In addition to being associated withhigher diastolic blood pressure than fear, angerwas associated with more nonspecific skin con-ductance responses, smaller heart rate accelera-tion, smaller increases in stroke volume andcardiac output, and larger increases in total pe-ripheral resistance, facial temperature, and fin-ger pulse volume than fear. Thus anger appearsto be more strongly associated with vascularactivity than is fear, but less strongly associatedwith cardiac activity.

Only a handful of studies were available formany of our (Cacioppo et al., 2000) compari-sons, so these results should only be acceptedtentatively and may need to be revisited, pend-ing more recent and future research. For in-stance, we found that anger was associatedwith greater total peripheral resistance andmore nonspecific skin conductance responsesthan fear, but only a few previous studies hadexamined these relationships. Thus it is not es-pecially surprising that whereas Pauls andStemmler (2003) more recently found that an-ger was associated with greater total peripheralresistance, they also found that anger was asso-ciated with fewer skin conductance responsesthan fear. By way of comparison, Pauls andStemmler’s results corroborated our findingthat fear was associated with larger increases inheart rate, which was based on 15 studies. Re-cent and future research will also allow meta-analytic comparisons that Cacioppo et al.(2000) were unable to complete because thenecessary data were simply unavailable.Britton, Taylor, Berridge, Mikels, and Iberzon(2006), for instance, measured participants’heart rate and skin conductance as they

watched films that elicited happiness, sadness,and disgust. They found that disgust was asso-ciated with greater skin conductance thansadness—a comparison that no studies in ourdatabase had reported.

As a research literature develops, meta-analysis becomes especially useful for identify-ing moderator variables. Several narrative re-views (e.g., Zajonc & McIntosh, 1992) havenoted that different emotion inductions appearto elicit different autonomic patterns. Skin re-sistance level, for instance, declines more dur-ing sadness than during other negative emo-tions (e.g., anger) when emotions are inducedwith imagery tasks; these patterns are not ob-tained, however, when emotions are inducedwith facial action tasks (Ekman et al., 1983).Similarly, whereas Cacioppo et al. (2000)found that fear is associated with greater heartrate increases than anger, Labouvie-Vief,Lumley, Jain, and Heinze (2003) found thatimagery-induced fear and anger elicited com-parably large increases in heart rate. As it turnsout, Labouvie-Vief et al.’s null effect is consis-tent with the results of a more recent meta-analysis by Stemmler (2004) that focused ex-clusively on the autonomic differentiation offear and anger. Stemmler’s meta-analysis re-vealed that whereas “real-life” fear inductions(e.g., unanticipated sudden darkness after hear-ing a frightening short story; Stemmler, 1989)elicited larger heart rate increases than real-lifeanger inductions (e.g., an experimenter’s ag-gressive demands; Pauls & Stemmler, 2003),fear imagery elicited no larger heart rate in-creases than anger imagery. Stemmler’s (2004)meta-analysis revealed the opposite pattern fordiastolic blood pressure: Fear imagery elicitedlarger increases than anger imagery, but real-life fear elicited no larger increases than real-life anger. Moreover, whereas sadness is typi-cally associated with increased heart rate, sadmusic seems to decrease heart rate (Etzel,Johnsen, Dickerson, Tranel, & Adolphs, 2006;Krumhansl, 1997). Such intriguing findings areproblematic for the notion of emotion-specificpatterning (e.g., Cacioppo et al., 1997; Zajonc& McIntosh, 1992).

Lang, Bradley, and Cuthbert’s (1990; seealso Bradley, 2000) distinction between “stra-tegic” and “tactical” aspects of emotions mayhelp explain the limited evidence for emotion-specific patterning. Tactics are the specific,context-bound patterns of action aimed at


achieving narrow goals. The same discreteemotion may call for different tactics in differ-ent situations. Lang et al. (1990), for instance,note that the behaviors associated with fear canrange from freezing to vigilance to flight, all ofwhich pose different metabolic demands. Suchtactical variability may account in part for thelimited evidence of emotion-specific autonomicpatterning (Zajonc & McIntosh, 1992).

In contrast to tactics, strategies direct actionsaimed at achieving broad end goals. Securingappetitive stimuli (e.g., food) and avoidingaversive stimuli (e.g., predators), for instance,represent strategic aspects of emotion (Lang etal., 1990). The ability of the autonomic ner-vous system to mobilize metabolic resources inresponse to hostile and hospitable stimuli iscrucial to survival, so valence-specific pattern-ing may be more pronounced than emotion-specific patterning. Among the ways in whichnegative and positive affective processes appearto differ, for instance, is the tendency for thechange in negative motivational output to belarger than the change in positive motivationaloutput per unit of activation. Though subjectto contextual factors (e.g., Smith et al., 2006),this “negativity bias” (Cacioppo & Berntson,1994; Cacioppo, Larsen, Smith, & Berntson,2004) has been observed in animal learning(Miller, 1961), risky decision making(Kahneman & Tversky, 1979), human affectivejudgments (Ito, Cacioppo, & Lang, 1998), andevent-related brain potentials to affective stim-uli (Ito, Larsen, Smith, & Cacioppo, 1998;Smith, Cacioppo, Larsen, & Chartrand, 2003).

A variety of theoretical and empirical worksuggests that, ceteris paribus, negative emo-tions may also elicit greater autonomic activitythan positive emotions (e.g., Taylor, 1991). Toexamine this hypothesis further, we (Cacioppoet al., 2000) conducted additional meta-analyses of autonomic activity in response toall negative emotions combined compared withall positive emotions combined. Diastolicblood pressure, blood volume, cardiac output,left ventricular ejection time, preejection pe-riod, pulse transit time, and heart rate allshowed significantly greater activation duringnegative than positive emotions, and no auto-nomic responses showed the opposite pattern.Thus, whereas Cacioppo et al.’s (2000) meta-analyses indicated that even a limited set of dis-crete emotions such as happiness, sadness, fear,anger, and disgust could not be fully differenti-

ated by visceral activity alone, they also indi-cated that negative emotions are associatedwith stronger autonomic responses than arepositive emotions.

Of course, Cacioppo et al.’s (1997, 2000)meta-analyses are mute with respect to the en-during theoretical question of whether auto-nomic activity contributes to or follows fromemotional experience. At first glance, the de-bate between James (1884) and Cannon (1927)presupposes the existence of a single invariantrelationship between emotional experience andperipheral physiological activity. Overlooked isthe fact that James actually viewed emotionsas being multiply determined. At the outset,James (1884) stated that “the only emotions Ipropose expressly to consider here are thosethat have a distinct bodily expression”(p. 189). Thus, in addition to contending thatperipheral patterns can produce emotions in abottom-up fashion, he acknowledged thatemotional experiences are sometimes governedby top-down processes that fail to producesomatovisceral activity (Ellsworth, 1994).Along these lines, one neurobiological modeldelineates the reciprocal relations between as-cending and descending systems that may con-tribute to anxiety (Berntson, Sarter, &Cacioppo, 1998). This model contends that af-fective states may be primed by top-down pro-cesses (e.g., appraisals) or bottom-up processes(e.g., visceral reactivity), and that the top-downand bottom-up mechanisms may mutually re-inforce one another (e.g., as in panic disorder).

We (Cacioppo, Berntson, & Klein, 1992)provided a general framework for conceptual-izing the multiple pathways by which periph-eral activity may shape emotional experience.At one end of the continuum, discrete emo-tional experiences result from the apperceptionof distinct somatovisceral patterns (e.g.,Ekman et al., 1983; James, 1884; Levenson,1988; Levenson et al., 1990). At the other endof the continuum, attributional processes insti-gated by the perception of undifferentiatedphysiological arousal generate discrete emo-tional experiences (e.g., Mandler, 1975;Schachter & Singer, 1962). Falling betweenthese extremes is yet another process by whichperipheral bodily reactions may contribute toemotional experience. By this account,“somatovisceral illusions” represent the resultof an active perceptual process by which am-biguous patterns of somatovisceral afference


are immediately and spontaneously disambigu-ated to produce distinct emotional experiences.

The essential feature of the proposition thatdiscrete emotions may result from somato-visceral illusions can be illustrated by analogy,using the ambiguous visual figure depicted inFigure 11.1 (see Cacioppo, Berntson, et al.,1992, for a more complete description of themodel). Even though the figure itself is un-changing, top-down processes allow viewers toperceive two very different images: a young

woman facing left or an elderly woman facingright (Leeper, 1935). Thus the same visualafference can lead to two different, discrete,and indubitable perceptual experiences, just asSchachter and Singer (1962) argued that thesame physiological afference may lead to twodifferent, discrete, and indubitable emotionalexperiences. Indeed, the architecture of thesomatovisceral apparatus may be better suitedto produce ambiguous afference than that ofthe visual system (Reed, Harver, & Katkin,1990). For instance, in the perception of am-biguous visual figures, the stimulus is a visualarray outside the body. In contrast, the centralnervous system serves both to create and to in-terpret the stimulus and the response tosomatovisceral information.

Cacioppo, Berntson, et al.’s (1992) model ar-gues against the tendency to view the psycho-physiological mechanisms underlying emotionin terms of a simple central–peripheral dichot-omy. It is also in accord with robust findings inthe literature that discrete emotional perceptscan occur even when the autonomic changes donot fully discriminate the emotions that are ex-perienced, and that autonomic activation canalter the intensity, if not the nature, of emo-tional experience. From this perspective, thequestion is not whether emotion-specific auto-nomic patterns occur, but under what condi-tions such patterns occur. We have emphasizedvisceral contributions to emotional experienceto this point, but, like Tomkins (1962),Cacioppo, Berntson, et al. (1992) also sug-gested that somatic processes can also affectemotional experience. In an ingenious test ofthis hypothesis, Strack, Martin, and Stepper(1988) reported that participants who had un-knowingly been induced to smile found comicstrips funnier than participants in a controlcondition. In the next section we examine therelationships between emotion and somatic ac-tivity, particularly facial expressions.

The Somatic Nervous System

Scientific studies of the link between facial ex-pressions and emotions originated with Dar-win (1890/1989), who noted, for instance, thatamong grief-stricken individuals, “the eyelidsdroop; . . . the lips, cheeks, and lower jaw allsink downwards from their own weight. Henceall the features are lengthened; and the face of aperson who hears bad news is said to fall”(p. 134). Investigators have since provided pro-


FIGURE 11.1. This ambiguous figure is called“My wife and mother-in-law” and is constructedfrom overlapping unambiguous elements. Theperceptual system tends to group like or relatedinformation together. Rather than presentingsome odd mixture of the two alternative pictures,partial identification of a young woman or an oldwoman in this figure supports a stable perceptionof a single coherent image. The identification ofwholes and of parts is reciprocally supportive,contributing further to the locking-in process. Ashift in gaze is not necessary for a perceptualchange to occur. In what may be analogous to dis-crete emotional feelings’ being spawned by thesame ambiguous pattern of somatovisceral infor-mation, ambiguous visual figures demonstratethat discrete images can derive from the same am-biguous pattern of visual information. From Bor-ing (1930).

vocative evidence that each of at least a subsetof discrete emotions is associated with distinctovert facial expressions (Ekman, 1973). Addi-tional evidence suggest that congenitally blindindividuals, members of non-Western cultures,and infants make emotional facial expressionssimilar to those of sighted Western adults,thereby raising the possibility that emotion-specific facial expressions are not merely theproduct of social learning (e.g., see Ekman,1973; Ekman & Friesen, 1978; Izard, 1977).

Although facial expressions may often revealthe nature of underlying emotions, many emo-tional reactions are not accompanied by visiblefacial actions (Cacioppo & Petty, 1981). Fur-thermore, although there is evidence that ob-servers across cultures recognize the facial ex-pressions of happiness, sadness, fear, anger,surprise, and disgust (Ekman, 1973), the dataare open to alternative interpretations (Russell,1994). In addition, individuals can invoke dis-play rules to mask or hide the emotion they arefeeling, and observers can confuse the meaningof expressions (e.g., fear and surprise; Ekman,1973; but see Cacioppo, Bush, & Tassinary,1992). As a result, the coding of overt facial ex-pressions can be a less than perfect measure ofemotion. An important complement to visualinspection of facial expression has been themeasurement of patterns of electrical activityassociated with contraction of the facial mus-cles. This technique—facial electromyography(EMG)—has made it possible to index muscleactivity even in the absence of visible facial ex-pressions (Cacioppo & Petty, 1981; Cacioppo,Tassinary, & Fridlund, 1990).

In pioneering research, Schwartz and col-leagues demonstrated that different types ofemotional imagery can elicit different patternsof EMG activity over the brow (corrugatorsupercilii), cheek (zygomaticus major), andperioral (depressor anguli oris) muscle regions.Schwartz, Fair, Salt, Mandel, and Klerman(1976), for instance, asked participants toimagine positive or negative events in theirlives. Results revealed that sad imagery elicitedmore EMG activity over the brow and less ac-tivity over the cheek than happy imagery. Thusfacial EMG patterns associated with emotionalstates appeared to represent “miniature repre-sentations” of those occurring during overt fa-cial expressions (Schwartz, Fair, Greenberg,Foran, & Klerman, 1975). Subsequent researchhas examined whether the facial EMG patternsobserved by Schwartz et al. (1975, 1976) with

emotional imagery tasks generalize to othertypes of emotion inductions. Hess, Banse, andKappas (1995), for instance, found that hu-morous films elicited greater activity overzygomaticus major and orbicularis oculi, par-ticularly when participants were accompaniedby friends. Patterns similar to those obtainedwith emotional imagery and films have alsobeen obtained with emotional pictures(Cacioppo, Petty, Losch, & Kim, 1986; Lang,Greenwald, Bradley, & Hamm, 1993; Larsen,Norris, & Cacioppo, 2003) and sounds(Bradley & Lang, 2000b; Larsen et al., 2003).

More recent research also indicates that af-fective stimuli can automatically elicit facialEMG reactions. Neumann, Hess, Schulz, andAlpers (2005) instructed participants to smileor frown whenever a word appeared on thescreen. Facial EMG recordings indicated thatparticipants smiled more quickly when pleas-ant words appeared and frowned more quicklywhen unpleasant words appeared. Dimberg,Thunberg, and Elmehed (2000) exposed partic-ipants to backwardly masked pictures of angryand happy faces for a mere 30 milliseconds.Results indicated that subliminal happy faceselicited more activity over zygomaticus majorand less activity over corrugator supercilii thandid angry faces. Taken together, these resultsindicate that facial reactions to affective stimulioccur spontaneously (Neumann et al., 2005),even when the eliciting stimulus is presentedoutside of awareness (Dimberg et al., 2000).An unresolved issue is whether the facial EMGpatterns observed by Dimberg et al. (2000) re-flect emotional reactions or mere mimicry (cf.Hess, Philippot, & Blairy, 1998). In any event,such mimicry may reflect the operation of cir-cuitry involving mirror neurons in theprefrontal cortex, which fire not only when anindividual performs some action (e.g., smiling;Leslie, Johnson-Frey, & Grafton, 2004), butalso when the individual observes someone elseperforming that action.

Though most research has investigated theeffects of valence on EMG activity, some re-search has focused on discrete emotions.Schwartz, Ahern, and Brown (1979) asked sub-jects to engage in thought and imagery tasksthat involved happiness, excitement, sadness,fear, and neutral emotional states. The only sig-nificant main effects for discrete emotionalstates indicated higher EMG activity over thecheek and lower EMG activity over the browduring positive than negative emotions. In a


similar study by Brown and Schwartz (1980),fear, anger, and sadness imagery elicited greateractivity over the brow than did happiness im-agery. Happiness imagery elicited the most ac-tivity over the cheek, but fear and anger alsoelicited elevations over the cheek. Whetherthese latter elevations reflect “miserablesmiles” (Ekman, Friesen, & Ancoli, 1980),cross-talk from other muscles of the middleand lower facial regions, or the putative phylo-genetic origin of smiling and laughter in primi-tive agonistic displays (Andrew, 1963; vanHooff, 1972) is unclear. More recently, Vrana(1993) found that disgust and anger imageryelicited similar levels of activity over the browand that disgust imagery also elicited greateractivity over the levator labii, which elevatesthe upper lip and wrinkles the nose. Thus theliterature provides some evidence for distinctcovert facial expressions associated with someemotions, but a great deal more evidence fordistinct covert facial expressions associatedwith broad positive and negative affectivestates (Cacioppo et al., 2000).

THE CENTRAL NERVOUS SYSTEM

EMG activity reflects the end result of centralnervous system activity and can therefore shedsome light on central processes involved inemotion. Early research by Schwartz et al.(1979) suggested that negative and positiveemotional imagery elicited greater activity overthe right and left sides of the face, respectively.The facial musculature is to some extentcontralaterally innervated, so Schwartz et al.’sfindings are consistent with neuropsychologi-cal evidence suggesting that the left and righthemispheres are differentially involved in nega-tive and positive emotional processing, respec-tively (Robinson & Downhill, 1995). Whereasleft anterior brain lesions are more likely toproduce major depression, for instance, rightanterior brain lesions are more likely to pro-duce mania. Subsequent research, however, hasprovided limited evidence for asymmetries inEMG responses (Hager & Ekman, 1985).

Unlike EMG activity, EEG activity directly re-flects central nervous system activity and hastherefore been more useful in exploring thehemispheric specialization of emotion. Much ofthis work has focused on measurement of EEGpower in the alpha band (8–13 Hz), which is in-

versely related to hemispheric activity (Lindsley& Wicke, 1974). A variety of EEG evidence hascorroborated the neuropsychological evidencethat the right hemisphere is more strongly asso-ciated with negative affect and the left hemi-sphere with positive affect. For instance, thosewith greater relative left frontal activity tend toreport higher levels of subjective well-being(Urry et al., 2004) and dispositional behavioralactivation (Coan & Allen, 2003; Sutton &Davidson, 1997). Earlier in life, toddlers withgreater relative left frontal activity tend to be lessbehaviorally inhibited (Davidson, 1993).

Davidson (1993, 2003) has integrated theseand other findings into a diathesis–stress modellinking individual differences in anterior corti-cal asymmetry to dispositional affective ten-dencies. According to this model, differences incerebral asymmetry have greater impact on af-fective reactions to stressors and other chal-lenges than on baseline affect (e.g., Davidson,1993; Davidson & Tomarken, 1989). Consis-tent with the diathesis–stress model, partici-pants with greater relative left frontal asymme-try reported stronger positive affectivereactions to pleasant film clips and weaker neg-ative affective reactions to unpleasant filmclips, but did not differ from those with greaterrelative right frontal asymmetry in terms ofbaseline mood (Tomarken, Davidson, &Henriques, 1990; see also Wheeler, Davidson,& Tomarken, 1993). Similarly, infants whocried when their mother left the room hadgreater relative right-hemisphere activity priorto the separation (Davidson & Fox, 1989).Some evidence also suggests that those withgreater relative left frontal activity experienceshorter bouts of negative affect in response tounpleasant pictures (Jackson et al., 2003).

To this point, EEG asymmetry has been re-lated to differences in affective traits, but thereis also substantial state-dependent variance inEEG asymmetry (Hagemann, Naumann,Thayer, & Bartussek, 2002), and some of thatthat variance may be due to differences in af-fective states. Davidson, Ekman, Saron,Senulis, and Friesen (1990) recorded EEG dur-ing a variety of evocative film clips. Partici-pants showed greater relative left frontal acti-vation during those moments when theydisplayed facial expressions of happiness thanwhen they displayed facial expressions of dis-gust. Analyses conducted across all artifact-freeEEG data (i.e., including those times in which a


facial expression was not present) failed to re-veal an effect of film valence on EEG asymme-try. Thus it may be that only emotional experi-ences strong enough to produce overt facialexpressions are associated with measurablehemispheric asymmetry.

Recent research has clarified whether frontalactivity better reflects approach- and avoidance-related behavioral motivation or positive andnegative affective valence. Motivation and va-lence tend to be correlated, such that positiveemotions are associated with approach and neg-ative emotions with avoidance. An interestingexception is anger, which is a negative emotionthat is associated with approach motivation.Harmon-Jones and Allen (1998) found thatthose who tend to experience anger show graterrelative left frontal, as opposed to right frontal,activity. Similarly, anger induced with facial ac-tion tasks is associated with greater relative leftfrontal activation (Coan, Allen, & Harmon-Jones, 2001). Such findings indicate that frontalactivity reflects approach–withdrawal motiva-tion more than it does positive–negative valence(Harmon-Jones, 2003).

Early conceptualizations of EEG asymmetryfocused on the relative difference between ac-tivity in the left and right hemispheres. In lightof neurophysiological and psychological evi-dence for separable approach- and avoidance-related motivational systems (e.g., Cacioppo &Berntson, 1994; Larsen et al., 2001), currentconceptualizations explicitly link activity in theleft hemisphere with approach motivation andactivity in the right hemisphere with avoidancemotivation. Individuals with depression or ahistory of depression show less left anterior ac-tivity than comparison participants, but do notdiffer in right anterior activity (Henriques &Davidson, 1990, 1991). These findings suggestthat depression reflects hypoactive approachmotivation rather than hyperactive avoidancemotivation (Henriques & Davidson, 1990,1991). More recently, Schmidt (1999) extendedprior evidence that shyness and sociability areseparable by demonstrating that women whoreport high levels of shyness and sociabilityshow greater right and left frontal activity.Schmidt’s findings, in particular, highlight theutility of considering left and right frontal ac-tivity separately. Indeed, difference scoreswould have failed to disambiguate individualswho are both shy and sociable from those whoare neither shy nor sociable.

EPILOGUE

As the varied perspectives represented in thisvolume suggest, a comprehensive understand-ing of emotions must encompass a wide rangeof viewpoints. One of the more interestingquestions concerning the psychophysiology ofemotions in particular is the role ofsomatovisceral afference in emotional experi-ence. Although it appears that negative emo-tions are associated with greater autonomic re-activity than positive emotions, the cumulativeevidence for emotion-specific patterns remainsinconclusive. In addition, facial EMG activityover the cheek (zygomaticus major) andperiocular (orbicularis oculi) muscle regions in-crease with positivity, whereas EMG activityover the brow (corrugator supercilii) muscle re-gion increases with negativity and decreaseswith positivity; research on EEG asymmetriessimilarly suggests that anterior brain regionsare differentially involved in approach-relatedversus avoidance-related behavioral processes.Thus EEG activity can differentiate approach-and avoidance-related negative emotions suchas anger and fear, respectively.

As detailed above, the patterns of autonomicactivity associated with emotion have not beenthoroughly delineated. Among the obstacles toidentifying emotion-specific autonomic pat-terns, particularly for dually and antagonisti-cally innervated organs such as the heart, arethe multiple causal pathways connecting neuralchanges and end-organ response (e.g., heartrate). Emotional stimuli do not invariablyevoke reciprocal activation of the sympatheticand parasympathetic branches of the auto-nomic nervous system. For instance, aversiveconditioned stimuli can produce coactivationof the sympathetic and parasympatheticbranches system, yielding accelerated, deceler-ated, or even unchanged heart rate, dependingon the relative strength of sympathetic versusparasympathetic activation (see Berntson,Cacioppo, & Quigley, 1991). Rainville,Bechara, Naqvi, and Damasio (2006) recentlyprovided initial evidence that measures of sym-pathetic and parasympathetic activity can dif-ferentiate discrete emotions even when end-or-gan responses cannot. Heart rate increasedduring both anger and fear imagery in theirstudy, but spectral analyses of heart rate vari-ability indicated that the effect of anger imag-ery was mediated by sympathetic activation


and the effect of fear imagery by parasympa-thetic deactivation. Whether Rainville et al.’sfindings will be replicated with larger samples,different types of emotion inductions (cf.Stemmler, 2004), and different discrete emo-tions remains to be seen.

Whether or not the conditions for and ele-ments of emotion-specific peripheral patternsof activity can be identified, what does seemclear from the extant research is that discreteemotions can be experienced even in the ab-sence of completely differentiated autonomicpatterns. Though Cannon (1927) took this asevidence that somatovisceral afference has noinfluence on the experience of emotion, we(Cacioppo, Berntson, et al., 1992) outlinedthree routes by which somatovisceralafferentiation may influence emotional experi-ence: emotion-specific autonomic patterns,somatovisceral illusions, or cognitive labelingof unexplained physiological arousal. Fromthis perspective, the traditional tendency toview the mechanisms underlying emotion interms of a simple central–peripheral dichotomyis untenable.

Emotions, particularly negative emotions,have also been linked to increases in healthproblems, including an enhanced susceptibilityto infection (e.g., Cohen, Doyle, Turner, Alper,& Skoner, 2003; see review by Herbert & Co-hen, 1993), poorer response to an influenzavaccine (Kiecolt-Glaser, Glaser, Gravenstein,Malarkey, & Sheridan, 1996), and impairedwound healing (Kiecolt-Glaser, Marucha, Ma-larkey, Mercado, & Glaser, 1995). Mediatorsof the relationship between emotion and healthhave not yet been fully delineated, but severalmechanisms are likely to be involved, some ofwhich imply autonomic differentiation of posi-tive from negative affective states. Health prob-lems increase with aging as well, with negativeemotions augmenting age-related declines inhealth and well-being (e.g., Kiecolt-Glaser,Dura, Speicher, Trask, & Glaser, 1991), andpositive emotions having less impact (Ewart,Taylor, Kraemer, & Agras, 1991).

Studies of the psychophysiology of emotionhave tended to treat autonomic, somatic, andcentral nervous system activity in isolation (cf.Davidson et al., 1990; Mauss et al., 2005). Re-cent work highlights the utility of studying therelationships among these systems and relatedsystems (e.g., the immune system). In a recentstudy investigating autonomic and immune re-

sponses to stress, for instance, individuals whoresponded to an experimenter’s demands withfear, as assessed by facial action coding,showed greater cardiovascular and cortisol re-activity than did those who responded with acombination of anger and disgust (Lerner,Gonzalez, Dahl, Hariri, & Taylor, 2005). In an-other study Waldstein et al. (2000) found thatthose who showed greater left frontal EEG ac-tivation during anger inductions also showedgreater cardiovascular activation. Hagemann,Waldstein, and Thayer (2003) have incorpo-rated such findings into a model that integratesthe roles of the central nervous system andautonomic nervous systems in emotion.Hagemann et al.’s (2003) model relied in parton evidence from positron emission tomogra-phy (PET) and functional magnetic resonanceimaging (fMRI), which now allow psycho-physiologists to visualize the human brain dur-ing emotional episodes with tremendous spa-tial resolution (Wager et al., Chapter 15, thisvolume). One must bear in mind, however, thatPET and fMRI have limited temporal resolu-tion, because they rely on relatively gradualchanges in blood flow through the brain. Emo-tional reactions tend to be fleeting (Ekman,1994), so a useful approach may be to conductstudies combining fMRI data with traditionalpsychophysiological recordings, many ofwhich offer superior temporal resolution (e.g.,Critchley et al., 2003). In the years to come,such integrative approaches promise to yieldever clearer insights into the physiological sub-strates of emotion.

REFERENCES

Andrew, R. J. (1963). The origin and evolution of thecalls and facial expressions of the primates. Behav-iour, 20, 1–109.

Barrett, L. F. (2006). Solving the emotion paradox: Cat-egorization and the experience of emotion. Personal-ity and Social Psychology Review, 10, 20–46.

Barrett, L. F., Quigley, K. S., Bliss-Moreau, E., &Aronson, K. R. (2004). Interoceptive sensitivity andself-reports of emotional experience. Journal of Per-sonality and Social Psychology, 87, 684–697.

Berntson, G. G., Cacioppo, J. T., & Quigley, K. S.(1991). Autonomic determinism: The modes of auto-nomic control, the doctrine of autonomic space, andthe laws of autonomic constraint. Psychological Re-view, 98, 459–487.

Berntson, G. G., Sarter, M., & Cacioppo, J. T. (1998).


Anxiety and cardiovascular reactivity: The basalforebrain cholinergic link. Behavioural Brain Re-search, 94, 225–248.

Boring, E. G. (1930). A new ambiguous figure. Ameri-can Journal of Psychology, 42, 444.

Bradley, M. M. (2000). Emotion and motivation. In J. T.Cacioppo, L. G. Tassinary, & G. G. Berntson (Eds.),Handbook of psychophysiology (2nd ed., pp. 602–641). New York: Cambridge University Press.

Bradley, M. M., & Lang, P. J. (2000a). Measuring emo-tion: Behavior, feeling, and physiology. In R. D. Lane& L. Nadel (Eds.), Cognitive neuroscience of emotion(pp. 242–276). New York: Oxford University Press.

Bradley, M. M., & Lang, P. J. (2000b). Affective reac-tions to acoustic stimuli. Psychophysiology, 37, 204–215.

Britton, J. C., Taylor, S., F., Berridge, K. C., Mikels, J.A., & Liberzon, I. (2006). Differential subjective andpsychophysiological responses to socially andnonsocially generated emotional stimuli. Emotion, 6,150–155.

Brown, S. L., & Schwartz, G. E. (1980). Relationshipsbetween facial electromyography and subjective ex-perience during affective imagery. Biological Psychol-ogy, 11, 49–62.

Cacioppo, J. T., & Berntson, G. G. (1994). Relationshipbetween attitudes and evaluative space: A critical re-view, with emphasis on the separability of positiveand negative substrates. Psychological Bulletin, 115,401–423.

Cacioppo, J. T., Berntson, G. G., & Klein, D. J. (1992).What is an emotion?: The role of somatovisceralafference, with special emphasis on somatovisceral“illusions.” Review of Personality and Social Psy-chology, 14, 63–98.

Cacioppo, J. T., Berntson, G. G., Klein, D. J., &Poehlmann, K. M. (1997). The psychophysiology ofemotion across the lifespan. Annual Review of Ger-ontology and Geriatrics, 17, 27–74.

Cacioppo, J. T., Berntson, G. G., Larsen, J. T.,Poehlmann, K. M., & Ito, T. A. (2000). Thepsychophysiology of emotion. In M. Lewis & J. M.Haviland-Jones (Eds.), Handbook of emotions (2nded., pp. 173–191). New York: Guilford Press.

Cacioppo, J. T., Bush, L. K., & Tassinary, L. G. (1992).Microexpressive facial actions as a function of affec-tive stimuli: Replication and extension. Personalityand Social Psychology Bulletin, 18, 515–526.

Cacioppo, J. T., Gardner, W. L., & Berntson, G. G.(1999). The affect system: Form follows function.Journal of Personality and Social Psychology, 76,839–855.

Cacioppo, J. T., Larsen, J. T., Smith, N. K., & Berntson,G. G. (2004). The affect system: What lurks belowthe surface of feelings? In A. S. R. Manstead, N. H.Frijda, & A. H. Fischer (Eds.), Feelings and emo-tions: The Amsterdam Conference (pp. 223–242).New York: Cambridge University Press.

Cacioppo, J. T., & Petty, R. E. (1981). Electromyograms

as measures of extent and affectivity of informationprocessing. American Psychologist, 36, 441–456.

Cacioppo, J. T., & Petty, R. E. (1986). Stalking rudi-mentary processes of social influence: Apsychophysiological approach. In M. P. Zanna, J. M.Olson, & C. P. Herman (Eds.), Social influence: TheOntario Symposium (Vol. 5, pp. 41–74). Hillsdale,NJ: Erlbaum.

Cacioppo, J. T., & Petty, R. E. (1987). Stalking rudi-mentary processes of social influence: Apsychophysiological approach. In M. P. Zanna, J. M.Olson, & C. P. Herman (Eds.), Social influence: TheOntario Symposium (Vol. 5, pp. 41–74). Hillsdale,NJ: Erlbaum.

Cacioppo, J. T., Petty, R. E., Losch, M. E., & Kim, H. S.(1986). Electromyographic activity over facial mus-cle regions can differentiate the valence and intensityof affective reactions. Journal of Personality and So-cial Psychology, 50, 260–268.

Cacioppo, J. T., Tassinary, L. G., & Berntson, G. G.(Eds.). (2007). Handbook of psychophysiology (3rded.). New York: Cambridge University Press.

Cacioppo, J. T., Tassinary, L. G., & Fridlund, A. J.(1990). The skeletomotor system. In J. T. Cacioppo& L. G. Tassinary (Eds.), Principles ofpsychophysiology: Physical, social, and inferential el-ements (pp. 325–384). New York: Cambridge Uni-versity Press.

Cannon, W. B. (1927). The James–Lange theory of emo-tions: A critical examination and an alternative the-ory. American Journal of Psychology, 39, 106–124.

Chwalisz, K., Diener, E., & Gallagher, D. (1988). Auto-nomic arousal feedback and emotional experience:Evidence from the spinal cord injured. Journal ofPersonality and Social Psychology, 54, 820–828.

Clore, G. L., Ortony, A., & Foss, M. A. (1987). The psy-chological foundations of the affective lexicon. Jour-nal of Personality and Social Psychology, 53, 751–766.

Coan, J. A., & Allen, J. J. B. (2003). Frontal EEG asym-metry and the behavioral activation and inhibitionsystems. Psychophysiology, 40, 106–114.

Coan, J. A., Allen, J. J. B., & Harmon-Jones, E. (2001).Voluntary facial expression and hemispheric asym-metry over the frontal cortex. Psychophysiology, 38,912–925.

Cobos, P., Sánchez, M., Garcia, C., Vera, M. N., & Vila,I. (2002). Revisiting the James versus Cannon debateon emotion: Startle and autonomic modulation in pa-tients with spinal cord injuries. Biological Psychol-ogy, 61, 251–269.

Cohen, S., Doyle, W. J., Turner, R. B., Alper, C. M., &Skoner, D. P. (2003). Emotional style and susceptibil-ity to the common cold. Psychosomatic Medicine,65, 652–657.

Coles, M. G. H., Donchin, E., & Porges, S. W. (Eds.).(1986). Psychophysiology: Systems, processes, andapplications. New York: Guilford Press.

Critchley, H. D., Mathias, C. J., Josephs, O., O’Doherty,


J., Zanini, S., Dewar, B., et al. (2003). Humancingulate cortex and autonomic control: Convergingneuroimaging and clinical evidence. Brain, 126,2139–2152.

Critchley, H. D., Wiens, S., Rotshtein, P., Öhman, A., &Dolan, R. J. (2004). Neural systems supportinginteroceptive awareness. Nature Neuroscience, 3,1049–1056.


Darwin, C. (1989). The expression of emotions in manand animals (2nd ed.). New York: New York Univer-sity Press. (Original work published 1890)

Davidson, R. J. (1993). Childhood temperament and ce-rebral asymmetry: A neurobiological substrate ofbehavioral inhibition. In K. H. Rubin & J. B.Asendorpf (Eds.), Social withdrawal, inhibition, andshyness in childhood (pp. 31–48). Hillsdale, NJ:Erlbaum.

Davidson, R. J. (2003). Affective neuroscience andpsychophysiology: Toward a synthesis. Psychophysi-ology, 40, 655–665.

Davidson, R. J., Ekman, P., Saron, C. D., Senulis, J. A.,& Friesen, W. V. (1990). Approach–withdrawal andcerebral asymmetry: Emotional expression and brainphysiology I. Journal of Personality and Social Psy-chology, 58, 330–341.

Davidson, R. J., & Fox, N. A. (1989). Frontal brainasymmetry predicts infants’ response to maternalseparation. Journal of Abnormal Psychology, 98,127–131.

Davidson, R. J., & Tomarken, A.J. (1989). Lateralityand emotion: An electrophysiological approach. In F.Boller & J. Grafman (Eds.), Handbook of neuropsy-chology (Vol. 3, pp. 419– 441). New York: Elsevier.

DePaulo, B. M., Lindsay, J. J., Malone, B. E.,Muhlenbruck, L., Charlton, K., & Cooper, H.(2003). Cues to deception. Psychological Bulletin,129, 74–112.

Dimberg, U., Thunberg, M., & Elmehed, K. (2000). Un-conscious facial reactions to emotional facial expres-sions. Psychological Science, 11, 86–89.

Eagly, A. H., & Chaiken, S. (1993). The psychology ofattitudes. Orlando, FL: Harcourt Brace Jovanovich.

Ekman, P. (1973). Cross-cultural studies of facial ex-pression. In P. Ekman (Ed.), Darwin and facial ex-pression: A century of research in review (pp. 1–83).New York: Academic Press.

Ekman, P. (1994). Moods, emotions, and traits. In P.Ekman & R. J. Davidson (Eds.), The nature of emo-tion (pp. 56–58). New York: Oxford UniversityPress.

Ekman, P., & Friesen, W. V. (1978). The Facial ActionCoding System: A technique for the measurement offacial movement. Palo Alto, CA: Consulting Psychol-ogists Press.

Ekman, P., Friesen, W. V., & Ancoli, S. (1980). Facialsigns of emotional experience. Journal of Personalityand Social Psychology, 39, 1125–1134.

Ekman, P., Levenson, R. W., & Friesen, W. V. (1983).Autonomic nervous system activity distinguishesamong emotions. Science, 221, 1208–1210.

Ellsworth, P. C. (1994). William James and emotion: Is acentury of fame worth a century of misunderstand-ing? Psychological Review, 101, 222–229.

Etzel, J. A., Johnsen, E. L., Dickerson, J., Tranel, D., &Adolphs, R. (2006). Cardiovascular and respiratoryresponses during musical mood induction. Interna-tional Journal of Psychophysiology, 61, 57–69.

Ewart, C. K., Taylor, C. B., Kraemer, H. C., & Agras, W.S. (1991). High blood pressure and marital discord:Not being nasty matters more than being nice.Health Psychology, 10, 155–163.

Fainsilber, L., & Ortony, A. (1987). Metaphorical usesof language in the expression of emotions. Metaphorand Symbolic Activity, 2, 239–250.


Frijda, N. H. (1994). Varieties of affect: Emotions andepisodes, moods, and sentiments. In P. Ekman & R.J. Davidson (Eds.), The nature of emotion (pp. 59–67). New York: Oxford University Press.

Gagnon, L., & Peretz, I. (2003). Mode and tempo rela-tive contributions to “happy–sad” judgements inequitone mequitone. Cognition and Emotion, 17,25–40.

Gazzaniga, M. S., & LeDoux, J. E. (1978). The integrat-ed mind. New York: Plenum Press.

Green, D. P., Salovey, P., & Truax, K. M. (1999). Static,dynamic, and causative bipolarity of affect. Journalof Personality and Social Psychology, 76, 856–867.

Hagemann, D., Naumann, E., Thayer, J. F., &Bartussek, D. (2002). Does resting electroencephalo-graph asymmetry reflect a trait?: An application oflatent state-trait theory. Journal of Personality andSocial Psychology, 82, 619–641.

Hagemann, D., Waldstein, S. R., & Thayer, J. F. (2003).Central and autonomic nervous system integration inemotion. Brain and Cognition, 52, 79–87.

Hager, J. C., & Ekman, P. (1985). The asymmetry of fa-cial actions is inconsistent with models of hemi-spheric specialization. Psychophysiology, 22, 307–318.

Harmon-Jones, E. (2003). Clarifying the emotive func-tions of asymmetrical frontal cortical activity.Psychophysiology, 40, 838–848.

Harmon-Jones, E., & Allen, J. J. B. (1998). Anger andfrontal brain activity: EEG asymmetry consistentwith approach motivation despite negative affectivevalence. Journal of Personality and Social Psychol-ogy, 74, 1310–1316.

Henriques, J. B., & Davidson, R. J. (1990). Regionalbrain electrical asymmetries discriminate betweenpreviously depressed and healthy control subjects.Journal of Abnormal Psychology, 99, 22–31.

Henriques, J. B., & Davidson, R. J. (1991). Left frontalhypoactivation in depression. Journal of AbnormalPsychology, 100, 535–545.


Herbert, T. B., & Cohen, S. (1993). Depression and im-munity: A meta-analytic review. Psychological Bulle-tin, 113, 472–486.

Hess, U., Banse, R., & Kappas, A. (1995). The intensityof facial expression is determined by underlying af-fective state and social situation. Journal of Personal-ity and Social Psychology, 69, 280–288.

Hess, U., Philippot, P., & Blairy, S. (1998). Facial reac-tions to emotional facial expressions: Affect or cogni-tion? Cognition and Emotion, 12, 509–531.

Hoffman, R. R., Waggoner, J. E., & Palermo, D. S.(1991). Metaphor and context in the language ofemotion. In R. R. Hoffman & D. S. Palermo (Eds.),Cognition and the symbolic processes: Applied andecological perspectives (pp. 163–185). Hillsdale, NJ:Erlbaum.

Ito, T. A., Cacioppo, J. T., & Lang, P. J. (1998). Elicitingaffect using the International Affective Picture Sys-tem: Bivariate evaluation and ambivalence. Personal-ity and Social Psychology Bulletin, 24, 855–879.

Ito, T. A., Larsen, J. T., Smith, N. K., & Cacioppo, J. T.(1998). Negative information weighs more heavilyon the brain: The negativity bias in evaluative catego-rizations. Journal of Personality and Social Psychol-ogy, 75, 887–900.

Izard, C. E. (1977). Human emotions. New York: Aca-demic Press.

Jackson, D. C., Mueller, C. J., Dolski, I., Dalton, K. M.,Nitschke, J. B., Urry, H. L., et al. (2003). Now youfeel it, now you don’t: Frontal brain electrical asym-metry and individual differences in emotion regula-tion. Psychological Science, 14, 612–617.


James, W. (1950). Principles of psychology (Vol. 1).New York: Dover. (Original work published 1890)

Kahneman, D., & Tversky, A. (1979). Prospect theory:An analysis of decisions under risk. Econometrica,47, 263–291.

Kiecolt-Glaser, J. K., Dura, J. R., Speicher, C. E.,Trask, O. J., & Glaser, R. G. (1991). Spousal care-givers of dementia victims: Longitudinal changes inimmunity and health. Psychosomatic Medicine, 53,345–362.

Kiecolt-Glaser, J. K., Glaser, R., Gravenstein, S., Malar-key, W. B., & Sheridan, J. (1996). Chronic stress al-ters the immune response to influenza virus vaccinein older adults. Proceedings of the National Academyof Sciences of the USA, 93, 3043–3047.

Kiecolt-Glaser, J. K., Marucha, P. T., Malarkey, W. B.,Mercado, A. M., & Glaser, R. (1995). Slowing ofwound healing by psychological stress. Lancet, 346,1194–1196.

Krumhansl, C. L. (1997). An exploratory study of musi-cal emotions and psychophysiology. Canadian Jour-nal of Experimental Psychology, 51, 336–353.

Labouvie-Vief, G., Lumley, M. A., Jain, E., & Heinze,H. (2003). Age and gender differences in cardiac re-activity and subjective emotion responses to emo-

tional autobiographical memories. Emotion, 3, 115–126.

Lang, P. J. (1971). The application ofpsychophysiological methods to the study of psycho-therapy and behavior change. In A. E. Bergin & S. L.Garfield (Eds.), Handbook of psychotherapy andbehavior change: An empirical analysis (pp. 75–125).New York: Wiley.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990).Emotion, attention, and the startle reflex. Psycholog-ical Review, 97, 377–395.

Lang, P. J., Greenwald, M. K., Bradley, M. M., &Hamm, A. O. (1993). Looking at pictures: Affective,facial, visceral, and behavioral reactions. Psycho-physiology, 30, 261–273.

Larsen, J. T., McGraw, A. P., & Cacioppo, J. T. (2001).Can people feel happy and sad at the same time?Journal of Personality and Social Psychology, 81,684–696.

Larsen, J. T., Norris, C. J., & Cacioppo, J. T. (2003). Ef-fects of positive affect and negative affect onelectromyographic activity over zygomaticus majorand corrugator supercilii. Psychophysiology, 40,776–785.

LeDoux, J. (1996). The emotional brain. New York: Si-mon & Schuster.

Leeper, R. (1935). A study of a neglected portion ofthe field of learning—the development of sensoryorganization. Journal of Genetic Psychology, 46,41–75.

Lerner, J. S., Gonzalez, R. M., Dahl, R. E., Hariri, A. R.,& Taylor, S. E. (2005). Facial expressions of emotionreveal neuroendocrine and cardiovascular stress re-sponses. Biological Psychiatry, 58, 743–750.

Leslie, K. R., Johnson-Frey, S. H., & Grafton, S. T.(2004). Functional imaging of the face and hand imi-tation: Towards a motor theory of empathy.NeuroImage, 21, 601–607.

Levenson, R. W. (1988). Emotion and the autonomicnervous system: A prospectus for research on auto-nomic specificity. In H. L. Wagner (Eds.), Socialpsychophysiology and emotion: Theory and clinicalapplications (pp. 17–42). Chichester, UK: Wiley.


Levenson, R. W., Ekman, P., & Friesen, W. V. (1990).Voluntary facial action generates emotion-specificautonomic nervous system activity. Psychophysi-ology, 27, 363–384.

Lindsley, D. B., & Wicke, J. D. (1974). The electroen-cephalogram: Autonomous electrical activity in manand animals. In R. Thompson & M. N. Patterson(Eds.), Bioelectric recording techniques (pp. 3–79).New York: Academic Press.

Mandler, G. (1975). Mind and emotion. New York:Wiley.

Mauss, I. B., Levenson, R. W., McCarter, L., Wilhelm, F.H., & Gross, J. J. (2005). The tie that binds?: Coher-


ence among emotion, experience, behavior, and phys-iology. Emotion, 5, 175–190.

Miller, N. E. (1961). Some recent studies on conflictbehavior and drugs. American Psychologist, 16, 12–24.

Neumann, R., Hess, M., Schulz, S., & Alpers, G. W.(2005). Automatic behavioural responses to valence:Evidence that facial action is facilitated by evaluativeprocessing. Cognition and Emotion, 19, 499–513.

Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957).The measurement of meaning. Urbana: University ofIllinois Press.

Pauls, C. A., & Stemmler, G. (2003). Repressive and de-fensive coping during fear and anger. Emotion, 3,284–302.

Rainville, P., Bechara, A., Naqvi, N., & Damasio, A. R.(2006). Basic emotions are associated with distinctpatterns of cardiovascular reactivity. InternationalJournal of Psychophysiology, 61, 5–18.

Reed, S. D., Harver, A., & Katkin, E. S. (1990).Interoception. In J. T. Cacioppo & L. G. Tassinary(Eds.), Principles of psychophysiology: Physical, so-cial, and inferential elements (pp. 253–294). NewYork: Cambridge University Press.

Reisenzein, R. (1983). The Schachter theory of emotion:Two decades later. Psychological Bulletin, 94, 239–264.

Robinson, R. G., & Downhill, J. E. (1995). Lateral-ization of psychopathology in response to focal braininjury. In R. J. Davidson & K. Hugdahl (Eds.), Brainasymmetry (pp. 693–711). Cambridge, MA: MITPress.

Russell, J. A. (1978). Evidence of convergent validity onthe dimensions of affect. Journal of Personality andSocial Psychology, 36, 1152–1168.

Russell, J. A. (1980). A circumplex model of affect.Journal of Personality and Social Psychology, 39,1161–1178.

Russell, J. A. (1994). Is there universal recognition ofemotion from facial expressions?: A review of thecross-cultural studies. Psychological Bulletin, 115,102–141.

Russell, J. A., & Barrett, L. F. (1999). Core affect, proto-typical emotional episodes, and other things calledemotion: Dissecting the elephant. Journal of Person-ality and Social Psychology, 76, 805–819.

Russell, J. A., & Carroll, J. M. (1999). On the bipolarityof positive and negative affect. Psychological Bulle-tin, 125, 3–30.

Schachter, S., & Singer, J. E. (1962). Cognitive, social,and physiological determinants of emotional state.Psychological Review, 69, 379–399.

Scherer, K. R., Schorr, A., & Johnstone, T. (Eds.).(2001). Appraisal processes in emotion. New York:Oxford University Press.

Schmidt, L. A. (1999). Frontal brain electrical activity inshyness and sociability. Psychological Science, 10,316–320.

Schwartz, G. E., Ahern, G. L., & Brown, S.-L. (1979).Lateralized facial muscle response to positive and

negative emotional stimuli. Psychophysiology, 16,561–571.

Schwartz, G. E., Fair, P. L., Greenberg, P. S., Foran, J.M., & Klerman, G. L. (1975). Self-generated affec-tive imagery elicits discrete patterns of facial muscleactivity. Psychophysiology, 12, 234. (Abstract)

Schwartz, G. E., Fair, P. L., Salt, P., Mandel, M. R., &Klerman, G. R. (1976). Facial muscle patterning toaffective imagery in depressed and nondepressed sub-jects. Science, 192, 489–491.

Schwarz, N., & Clore, G. L. (1983). Mood, mis-attribution, and judgments of well-being: Informa-tive and directive functions of affective states. Jour-nal of Personality and Social Psychology, 45, 513–523.

Smith, N. K., Cacioppo, J. T., Larsen, J. T., &Chartrand, T. L. (2003). May I have your attention,please: Electrocortical responses to positive and neg-ative stimuli. Neuropsychologia, 41, 171–183.

Smith, N. K., Larsen, J. T., Chartrand, T. L., Cacioppo,J. T., Katafiasz, H. A., & Moran, K. E. (2006). Beingbad isn’t always good: Evaluative context moderatesthe attention bias toward negative information. Jour-nal of Personality and Social Psychology, 90, 210–220.

Stein, N. L., & Levine, L. J. (1987). Thinking about feel-ings: The development and organization of emo-tional knowledge. In R. E. Snow & M. Farr (Eds.),Aptitude, learning, and instruction: Vol. 3. Cogni-tion, conation, and affect (pp. 165–197). Hillsdale,NJ: Erlbaum.

Stemmler, G. (1989). The autonomic differentiation ofemotions revisited: Convergent and discriminant val-idation. Psychophysiology, 26, 617–632.

Stemmler, G. (2004). Physiological processes duringemotion. In P. Phillippot & R. S. Feldman (Eds.), Theregulation of emotion (pp. 33–70). Mahwah, NJ:Erlbaum.

Stern, R. M., Ray, W. J., & Quigley, K. S. (2001).Psychophysiological recording. New York: OxfordUniversity Press.

Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibitingand facilitating conditions of the human smile: Anonobtrusive test of the facial feedback hypothesis.Journal of Personality and Social Psychology, 54,768–777.

Sutton, S. K., & Davidson, R. J. (1997). Prefrontal brainasymmetry: A biological substrate of the behavioralapproach and inhibition systems. Psychological Sci-ence, 8, 204–210.

Taylor, S. E. (1991). Asymmetrical effects of positiveand negative events: The mobilization-minimizationhypothesis. Psychological Bulletin, 110, 67–85.

Tomarken, A. J., Davidson, R. J., & Henriques, J. B.(1990). Resting frontal asymmetry predicts affectiveresponses to films. Journal of Personality and SocialPsychology, 59, 791–801.

Tomkins, S. S. (1962). Affect, imagery, and conscious-ness: Vol. 1. The positive affects. New York: Springer.

Tranel, D., & Damasio, A. (2000). Neuropsychology


and behavioral neurology. In J. T. Cacioppo, L. G.Tassinary, & G. G. Berntson (Eds.), Handbook ofpsychophysiology (2nd ed., pp. 119–141). NewYork: Cambridge University Press.

Urry, H. L., Nitschke, J. B., Dolski, I., Jackson, D. C.,Dalton, K. M., Mueller, C. J., et al. (2004). Making alife worth living: Neural correlates of well-being.Psychological Science, 15, 367–372.

van Hooff, J. A. R. A. M. (1972). A comparative ap-proach to the phylogeny of laughter and smiling. InR. Hinde (Ed.), Non-verbal communication(pp. 129–179). London: Royal Society & CambridgeUniversity Press.

Vianna, E. P. M., Weinstock, J., Elliott, D., Summers, R.,& Tranel, D. (2006). Increased feelings with in-creased body signals. Scan, 1, 37–48.

Vrana, S. R. (1993). The psychophysiology of disgust:Differentiating negative emotional contexts with fa-cial EMG. Psychophysiology, 30, 279–286.

Wagner, H. (1989). The physiological differentiation ofemotions. In H. Wagner & A. Manstead (Eds.),Handbook of social psychophysiology (pp. 77–89).New York: Wiley.

Waldstein, S. R., Kop, W. J., Schmidt, L. A., Haufler, A.J., Krantz, D. S., & Fox, N. A. (2000). Frontalelectrocortical and cardiovascular reactivity duringhappiness and anger. Biological Psychology, 55, 3–23.

Watson, D., Wiese, D., Vaidya, J., & Tellegen, A.(1999). The two general activation systems of affect:Structural findings, evolutionary considerations, andpsychobiological evidence. Journal of Personalityand Social Psychology, 76, 820–838.

Wheeler, R. E., Davidson, R. J., & Tomarken, A. J.(1993). Frontal brain asymmetry and emotional reac-tivity: A biological substrate of affective style.Psychophysiology, 30, 82–89.

Winkielman, P., Berridge, K. C., & Wilbarger, J. (2005).Unconscious affective reactions to masked happyversus angry faces influence consumption behaviorand judgments of value. Personality and Social Psy-chology Bulletin, 31, 121–135.

Zajonc, R. B., & McIntosh, D. N. (1992). Emotionsresearch: Some promising questions and somequestionable promises. Psychological Science, 3,70–74.


C H A P T E R 1 2

Vocal Expressions of Emotion

JO-ANNE BACHOROWSKI and MICHAEL J. OWREN

There is now a substantial body of work fo-cused on how emotion is conveyed and per-ceived from vocal acoustics. Although this re-search has arguably not enjoyed the samedegree of cumulative success as has work onthe facial channel, there is nonetheless a solidbody of evidence showing that specific vocalacoustic features are reliably associated withaffect-related arousal (or activation) on thepart of vocalizers, and that listeners in turn canreliably perceive arousal from vocal acoustics.Ongoing debates center on the more pointedquestion of whether vocal acoustics are rou-tinely associated with either discrete emotionalstates or affectively valenced experiences. Be-cause reliable associations with either are un-convincing, we suggest a different perspectivein which the primary function of vocal signal-ing is not so much to communicate emotion asto influence listener affect through vocal acous-tics, and thereby to influence behavior. In thischapter, we first review the two key theoreticalperspectives concerning the expression of affectvia the vocal channel, and introduce a third (af-

fect induction) possibility. We then brieflydescribe the acoustic measures that are mosttypically examined in empirical work. Next,we consider the results of research concerningthe expression and perception of affect via thevocal channel, with particular attention givento the interpretive claims that can reasonablybe made in the context of several paradigmaticlimitations. We then return to the affect induc-tion perspective, and show how that frame-work has been useful in understanding resultsfrom research on laughter. Finally, we morebroadly consider that an affect induction per-spective may help shape a more informed un-derstanding of emotion-related signaling viathe vocal channel.

THEORETICAL PERSPECTIVES

The notion that vocal acoustics are rich withcues to a vocalizer’s emotional state has a longhistory, with Cicero (55 B.C.E./2001) suggestingthat each emotion is associated with a distinc-

196

tive tone of voice. The first comprehensive ac-count of vocal emotion expression was pro-vided by Darwin (1872/1998), whose detailedcomparative descriptions led him to concludethat affective expressions, including those pro-duced via the vocal channel, are veridical. Inother words, he presumed that there is a directcorrespondence between particular signalerstates and the communicative display pro-duced. This position is largely consistent withthe best-known contemporary perspective. Ex-emplified by the ambitious work of Schererand colleagues (e.g., Scherer, 1986, 1989;Banse & Scherer, 1996; see reviews by John-stone & Scherer, 2000; Leinonen, Hiltunen,Linnankoski, & Laakso, 1997), this perspec-tive posits that discrete affective states experi-enced by the vocalizer are reflected in specificpatterns of acoustic cues in the speech beingproduced. Scherer has further proposed thatemotion-specific acoustic cues occur subse-quent to the outcomes of a vocalizer’s affectiveappraisal processes and associated physiologi-cal changes in vocal production anatomy (fordetails, see Scherer, 1986). The principal con-trasting perspective is that emotional cues inspeech acoustics instead reflect activationamong a small set of continuous dimensions. Inthis approach, a relatively common suggestionis that emotion-related acoustic effects are spe-cifically traceable to two orthogonal under-lying dimensions of arousal and pleasure(Bachorowski, 1999; Cowie, 2000; Frick,1985; Kappas, Hess, & Scherer, 1991; Pakosz,1983).

At this juncture, it is worth noting that thediscrete-emotion and dimensional perspectiveshave two important commonalities. First, bothexpect that affective states will be associatedwith differentiated acoustics, meaning that spe-cific patterns or configurations of acoustic cueswill be reliably associated with affective states.A second commonality is that both perspec-tives at least implicitly adopt a representationalperspective, meaning that information con-cerning emotional state is encoded in vocalacoustics, and subsequently decoded by listen-ers in order to make attributions about the vo-calizer’s state (e.g., Scherer, 1986, 2003; Juslin& Laukka, 2001). We discuss some implica-tions of a representational stance later in thischapter. For now, we note that outcomes fromempirical studies suggest that neither thediscrete-emotion nor the dimensional approach

is likely to be correct. Data from speech-relatedresearch provides only moderate support for adiscrete-emotions view, with emotion-relatedaspects of the acoustics seeming largely to re-flect to vocalizer arousal. However, links to acorresponding emotional valence dimensionhave also been difficult to demonstrate,suggesting a need for alternative interpreta-tions.

One such approach can arguably be tracedto the writing of Charles Darwin. AlthoughDarwin (1872/1998) generally emphasizedpossible veridical associations between vocalacoustics and vocalizer emotional state, he alsomade two key observations that are central tothe position taken here. The first was that vocalsignals can induce emotional responses in lis-teners, and the second was that these vocal sig-nals can elicit learned emotional responses.Thus he recognized that signalers can usesounds to influence listener affect, and that id-iosyncratic aspects of these signals can elicitlearned emotional responses in recipients whohear them in association with affect-inducingevents. These observations form the core of an“affect induction” view of vocal signaling,which began as a functional account of nonhu-man primate calling (Owren & Rendall, 1997,2001; Rendall & Owren, 2002), but may alsoapply to affect-related vocal signaling inhumans (Owren, Rendall, & Bachorowski,2003).

In the affect induction framework, the pri-mary function of signaling is not to expressemotion, but instead to influence listeners’ af-fect and thereby to shape their behavior. Al-though vocalizations may be triggered by a sig-naler’s emotions, signal function is not toconvey information about the nuances of thosestates, but rather to affect listeners’ arousal andaffect. In this view, inferences that listeners canmake concerning the signaler’s affect are a sec-ondary outcome, and reflect attributions thatlisteners base on their own affective responsesto the sounds, their past experience with suchsounds, and the context in which signal pro-duction is occurring. In sum, the affect induc-tion perspective argues that vocal expressionsof emotion are not displays of vocalizer statesas much as they are tools of social influence(see also Russell, Bachorowski, & Fernandez-Dols, 2003). We will return to this perspectivein the context of describing research outcomeson human laughter.

12. Vocal Expressions of Emotion 197

EMOTION-RELATEDVOCAL ACOUSTICS

Speech acoustics are imbued with “indexical”or “personal” cues, which are nonlinguistic as-pects of speech production that provide acous-tic correlates of the talker’s sex, individualidentity, age, and emotional state. Indexicalcues are also prominent in many nonlinguisticvocalizations, such as laughter. Regardless ofthe kind of sound produced, the source filtermodel of speech production (Fant, 1960;Stevens, 2000) has typically guided selection ofthe acoustic measurements to make (for details,see Owren & Bachorowski, 2007). In thismodel, vocal acoustics are treated as a linearcombination of (1) an underlying energysource, and (2) filtering effects that stem fromthe resonances of the supralaryngeal vocaltract. Source energy commonly has one of twoforms in vocal production, being either“voiced” or “unvoiced.” Voicing, also referredto as “phonation,” means that the vocal foldsare vibrating. Unvoiced sounds, in contrast, areproduced without vocal fold vibration, and aretypically noisy or breathy. The sounds of par-ticular interest here are those in which vocalfold vibration occurs in a quasi-periodic fash-ion, producing phonemes such as vowels. Thebasic rate of that vibration is called the “funda-mental frequency” (F0), and is primary in theperception of pitch. Measures of F0 and F0 vari-ability (including both overall range andmoment-to-moment perturbations) are the fo-cus of work seeking to identify acoustic corre-lates of emotional state. However, amplitude orperceived loudness has also been found to beimportant, particularly when listeners make in-ferences about a vocalizer’s emotional statefrom speech.

Supralaryngeal resonances (e.g., Johnson,1997; Lieberman & Blumstein, 1988) can alsobe measured, with the frequency of these“formants” being of greatest interest. Formantfiltering creates a shaping effect on the fre-quency content of voiced sounds, producinghigh-amplitude energy bands at resonance lo-cations and attenuating source energy else-where in the frequency spectrum. Measuringrapidly occurring formant changes has the po-tential to reveal the effects of momentary emo-tional states as they occur, although not muchpractical attention has yet been paid to thispromise. However, global frequency character-istics of emotion-laden voiced speech sounds

are readily characterized by calculating a“long-term average spectrum” (LTAS; e.g.,Pittam & Scherer, 1993), which usually in-volves computing mean energy over segmentsof 30 seconds or more. The LTAS is thus quickand robust, but does not track acoustic cues as-sociated with transient affective experiences.

EXPRESSION OF EMOTIONIN SPEECH

Overview of the Research

Although there has been no shortage of debatebetween those who advocate a discrete-emotions perspective and those who argue in-stead for a dimensional perspective, few (ifany) empirical studies have actually comparedthe two approaches directly. Direct compari-sons are in fact difficult to engineer, with moststudies therefore examining smaller sets of pre-dictions derived from one or the other of thetwo perspectives. Stemming from a discrete-emotions perspective, one prominent exampleis the work of Banse and Scherer (1996), whoanalyzed the vocalizations produced by 12 pro-fessional actors who each portrayed 14 emo-tions. This study is notable for both examininga large number of emotions and including verydetailed acoustic analyses. A central result wasthat of 29 acoustic features measured, F0 andmean amplitude clearly showed the strongestconnections to the emotions being portrayed.These findings are particularly important be-cause, regardless of one’s theoretical stance, in-vestigators widely agree that these features aremore likely to index talker arousal than specifi-cally valenced or differentiated states. Otheracoustic measures were also found to be linkedwith particular emotional portrayals, but thesemeasures accounted for much smaller propor-tions of the variance. For example, in examina-tions of the distribution of energy across 18 fre-quency bands extracted from LTAS analyses,the researchers found that on average, eachacted emotion accounted for only 10% of thevariance in each band.

Using acoustic outcomes, Banse and Scherer(1996) were able to demonstrate statisticalclassification accuracies of roughly 40% whenusing an empirically derived subset of 16acoustic measures to discriminate among thevarious emotions that were portrayed (resultsvaried depending on the particular statisticalmodel used, but were generally consistent). On


the one hand, these modest classification ratesmight be expected based on findings fromother kinds of indexical cues (e.g., talker iden-tity in speech segments; Bachorowski &Owren, 1999). On the other hand, the record-ings had been carefully screened prior to statis-tical classification; only 224 “high-quality”portrayals from the total of 1,344 recordedsamples were used. As a result, classificationrates reported in this study were probably in-flated relative to those that would have beenobtained with an unscreened sample. WhenBanse and Scherer tested 40 specific hypothesesconcerning the links between emotion and vo-cal acoustics (see Scherer, 1986), 23 predictionswere supported; the evidence was tenuous for 6others; and statistically significant deviationsfrom the expected magnitude and/or directionof effect occurred for 11 predictions. Naturally,any model of such scope will require adjust-ments in light of empirical findings. The moreimportant point here is that the results weremixed even for a small, carefully selectedsubsample of the recorded portrayals. Al-though these results are often cited as evidenceof differentiated vocal cues for discrete emo-tions, we suggest that they cannot be used tosubstantiate strong claims along these lines.

Other research has produced a similar mixof supportive and contrary outcomes (e.g.,Scherer, Banse, Wallbott, & Goldbeck, 1991;Sobin & Alpert, 1999), suggesting a need forfurther theoretical work or possibly even alter-native approaches (e.g., Kappas et al., 1991;Kappas & Hess, 1995; Scherer et al., 1991). Inaddition, we suggest that the use of actedrather than natural stimuli and carefullyscreened rather than representative samplesmay both have important, albeit unintended,consequences in studies of this kind (cf.Scherer, 2003). The obvious problem with rely-ing on acted samples is that these may not nec-essarily correspond to naturally produced vo-calizations. One counterargument is that muchof our verbal communication involves makingimpressions on others, and so having vocalizersact “as if” they were experiencing a particularstate is not markedly different from naturalcommunicative circumstances. To resolve thisimpasse, validation work could be performed,but for the most part has not (cf. Williams &Stevens, 1972). However, when this issue istaken together with evidence from naturalemotion-inducing circumstances showing thatindividual variability in vocalizer acoustics can

be quite substantial (e.g., Streeter, Macdonald,Apple, Krauss, & Galotti, 1983), it may be thecase that the careful analysis of acoustic cues toacted emotion is providing more informationabout emblematic portrayals of affective statesthan about naturally occurring cueing (cf.Scherer, 2003).

This problem is exacerbated by workingwith a nonrepresentative subset of the samplesin question. In Banse and Scherer’s (1996)work, selection was based largely on qualityevaluations made by 12 professional acting stu-dents, who found that some individual actors(especially females) were much more convinc-ing in their portrayals than others. One resultof the selection process, for instance, was thatone particular actor contributed a single sam-ple to the final set of 224, while another wasrepresented by 47 utterances. These kinds ofsex and talker differences in the “quality” ofportrayals have also been found by others (e.g.,Leinonen et al., 1997; Pell, 2001; Scherer et al.,1991; Schröder, 2000; Sobin & Alpert, 1999;Walbott & Scherer, 1986). Here again, the is-sue is whether the results of such studies can betaken at face value when the utterances beingclassified are not representative of the fullrange of variation observed in the originaldataset.

Comparing the Discrete-Emotionsand Dimensional Approaches

Overall, the most reliable empirical outcomesin testing the acoustics of emotional speechhave been arousal-related. Numerous studieshave shown, for example, that anger and joyare both associated with increased F0 andhigher amplitude in the sounds produced. Evenadvocates of discrete-emotions perspectivesagree that these have been the most prominenteffects (e.g., Johnstone & Scherer, 2000;Scherer, 1989). However, associations betweenvalence and vocal acoustics have been lessclear-cut (cf. Johnstone, van Reekum, Hird,Kirsner, & Scherer, 2005; Laukka, Juslin, &Bresin, 2005). For instance, we (Bachorowski& Owren, 1995) examined several acousticfeatures from speech samples produced by 120naïve participants, who were individually re-corded as they performed a lexical-decisiontask. As part of the task, each participant ut-tered a stock phrase just after receiving affect-inducing success or failure feedback. The threemost prominent acoustic changes from baseline


to on-task performance were F0-related andwere taken to reflect increases in vocalizerarousal. Valence-related differences were ob-served, but only emerged in more complex in-teractions among such variables as talker sex,the relative proportion of positive and negativefeedback each participant received, and traitdifferences in emotional intensity. Similar find-ings were obtained when a more exacting com-parison was conducted with recordings madeof 24 naïve participants who each described thethoughts and feelings evoked by affect-inducing slides (Bachorowski & Owren, 1996).Acoustic outcomes were strongly associatedwith self-reported arousal and to a much lesserextent with valence. Analyses testing whetheracoustic outcomes could be linked to discreteemotional states were largely nonsignificant.The results thus showed that arousal plays anoticeably more important role in shapingspeech acoustics than does valence, and thatready links to discrete emotional states are dif-ficult to demonstrate. Similar findings havebeen reported in other studies in which arousaland valence effects can be at least indirectlycompared (Hatfield, Hsee, Costello, Weisman,& Denney, 1995; Leinonen et al., 1997; Millot& Brand, 2001; Paeschke & Sendlmeier, 2000;Pell, 2001; Pereira, 2000; Pittam, Gallois, &Callan, 1990; Protopapas & Lieberman, 1997;Tolkmitt & Scherer, 1986; Trouvain & Barry,2000).

However, it would be premature to concludethat the acoustics of emotional speech only re-flect arousal (see also Johnstone et al., 2005),or that they can be better accounted for by a di-mensional than by a discrete-emotions ap-proach. Any such conclusion requires testingpredictions from the two perspectives withinthe same empirical framework, and we are notaware of any such comparisons. Furthermore,as Tassinary and Cacioppo (1992) indicated inthe context of facial expressions of emotion, af-fective intensity may moderate whether dis-crete or valenced effects are observed. In sup-port of this notion, Juslin and Laukka (2001;see also Banse & Scherer, 1996; Ladd,Silverman, Tolkmitt, Bergmann, & Scherer,1985; Scherer, Ladd, & Silverman, 1984)showed that the acoustics measured from thevocal portrayals of five emotions were differen-tially associated with the intensity (i.e., weak orstrong) with which each of these emotions wereproduced.

What is clearest at this point is that theacoustics of emotional speech are influenced bya variety of factors. Arousal and valence effectsdo not in and of themselves explain the avail-able data concerning vocal emotion effects.Other contributing factors include talker sex,individual talker identity, and emotional traits.Strategic shaping of vocal acoustics that atalker might show in pursuit of social goals,and the particular social context in which thespeech is being produced, may well also proveto be important (Bachorowski, 1999; Scherer,2003).

PERCEPTION OF EMOTIONFROM SPEECH

Overview of the Research

Theoretical approaches and empirical out-comes associated with the perception of emo-tion from speech acoustics generally parallelthose noted for production. When listeners areasked to identify the intended emotion in utter-ances produced by actors, accuracy is sig-nificantly better than chance—although at amoderate level overall, typically about 55%(reviewed by Johnstone & Scherer, 2000). Sim-ilar identification and misidentification ratesare observed across language groups, althougherror rates increase as vocalizer and listenerlanguages become more dissimilar (Scherer,Banse, & Wallbott, 2001; see also Tickle,2000).

Identification rates are usually best for anger,fear, and sadness. Performance is quite poor fordisgust, perhaps because this state is typicallyconveyed not through speech, but via vocal em-blems or exclamations. Results for positiveemotions have varied, but in an informativeway. Accuracy is typically quite high when lis-teners are given only one positive response op-tion, such as “happiness” (e.g., Johnson, Emde,Scherer, & Klinnert, 1986; Scherer et al.,1991). However, correct responses drop signifi-cantly when other positively toned options,such as “elation,” “contentment,” or “inter-est,” are tested (Banse & Scherer, 1996). A sim-ilar effect may contribute to the identificationof “sadness,” which is sometimes the only low-arousal option offered among the negativeemotions.

In an elegant set of studies, Laukka (2005)tested for evidence of categorical perception of


vocal emotion, using a set of synthetic stimuliderived from four prototypical expressions.These stimuli included expressions of anger,fear, happiness, and sadness produced by a pro-fessional actress. Synthesized continua werethen created for perceptual testing, using vocalsamples that differed by constant physicalamounts. Each continuum was anchored ateach end by one of two different emotionprototypes, and included five intermediatemorphs.

Outcomes from both identification and dis-crimination paradigms were taken as evidenceof categorical perception of vocal emotion—inother words, as supporting a discrete-emotionsperspective. Specifically, identification perfor-mance produced the distinct category bound-aries that are a hallmark feature of categoricalperception, and discrimination performancewas better for stimulus pairs that crossed cate-gory boundaries than for pairs drawn fromwithin a category. However, boundaries wereexpected to be defined by stimuli representingan equal mix of the two emotions involved.That outcome occurred for only two of the sixcontinua tested. In other cases, a 50–50 blendof emotions could be identified as representingone of the two emotion categories as much as70% of the time. In a second experiment, lis-teners were asked to rate each stimulus accord-ing to the extent to which each of the four emo-tions was represented. This more complex taskproduced fuzzier category boundaries. More-over, the disparities between the physical 50–50 mix and the listener-perceived 50–50 pointwere quite large for two of the four continua.For example, the continuum representing anequal mix of anger and fear was identified asbeing “angry” more than 90% of the time.Thus the categories produced by listeners werenot always aligned with the known physicalqualities of the stimulus, which suggests at theleast that listeners bring important perceptualbiases to this task. In addition, the results indi-cate that categorization can be critically shapedby procedural aspects of the identification taskthat in theory should have little effect.

As a package, Laukka’s (2005) outcomesleave open the possibility of categorical percep-tion of emotion from voice, but with several ca-veats. As Laukka notes, further work shouldinclude stimuli based on more than one talker.We would also add that in testing for categori-cal perceptual effects, Laukka in effect began

from the assumption that those categories existand could be represented in prototype form byusing single stimuli. In other words, there is adanger of circularity in this design, with theprocedures themselves encouraging listeners toconfirm categories that were built into the taskfrom the beginning. In emotions research, it isthe very existence of the categories themselvesthat is questioned (Barrett, 2006). Finally, it isquestionable whether these results can be usedto refute dimensional accounts, because thecontinua that Laukka tested are not the con-tinua that are proposed in dimensional ac-counts. Those accounts are typically based onvalence- and arousal-based dimensions, andeven discrete-emotions theorists who argue forthe occurrence of basic-emotion blends mightfind it hard to imagine speech that is an equalmix of happiness and fear.

The standard strategy for testing emotionperception from speech has not been to usesynthetic stimuli as in Laukka’s (2005) study,but instead to use acted speech samples in aforced-choice identification paradigm. Stimu-lus sets usually include only a small number oftalkers and emotions, and are often selected soas to include prototypical instances of the emo-tions in question. As mentioned earlier, Banseand Scherer (1996) used a subset of 224 sam-ples drawn from an original set of 1,344 re-cordings. Similarly, Leinonen and colleagues(1997) used 120 of their 480 samples to testparticipants; Scherer et al. (1991, 2001) usedeither 10 or 30 of 80 samples; and Sobin andAlpert (1999) used 152 of their 620 samples.This overall approach, which relies on actedemotional samples and then tests only a subsetselected on the basis of quality (cf. Juslin &Laukka, 2001), should necessarily produce atleast some evidence of differentiated perceptionof emotion. The outcomes described by Schereret al. (1991) speak directly to this point. Theseresearchers conducted four perception studies,with listeners in the first experiment hearingall the stimuli, but participants in subsequentstudies hearing only selected subsets. Notsurprisingly, higher performance rates wereachieved with the screened samples, regardlessof the emotion being tested.

Other factors also affect listener accuracy.Consistent with the literature on facial expres-sions of emotion, forced-choice proceduresproduce better performance than free-choicetests (e.g., Johnson et al., 1986; Pakosz, 1983).


Furthermore, Sobin and Alpert (1999) foundthat participants may be using quite variablecriteria when using a particular emotion label,and Scherer et al. (1991) have demonstratedthat the particular kind of participant beingtested can be important. Their study showed,for example, that college students and commu-nity volunteers differed in both accuracy ratesand error patterns. A final issue to note is thatstimulus duration can play an important role.For instance, Cauldwell (2000) found thatwords heard as conveying anger when heardin isolation could become less emotion-ladenor even neutral when embedded in carrierphrases. This effect occurred even when listen-ers were exposed to full-length versions beforehearing the words individually. Cauldwell’s an-swer to the rhetorical question “Where did theanger go?” was that the emotion was neverpresent in the first place. He suggested insteadthat the perception of emotion can be an arti-fact of the particular testing method used.

Comparing the Discrete-Emotionsand Dimensional Approaches

There is relatively good agreement among lis-teners asked to rate the perceived talker arousalof vocal stimuli. Pereira (2000; see also Green& Cliff, 1975; Juslin & Laukka, 2001), for ex-ample, had listeners rate acted vocal samples ofvarious discrete emotional states, using dimen-sional scales. The strongest associations be-tween these ratings and vocal acoustics wereclearly arousal-related, with F0 and amplitudeplaying crucial roles. The importance of thesefeatures in particular was also noted by Streeteret al. (1983), whose listeners rated talker stresslevels from speech. Participants reported thatvocalizers were stressed when talker F0 and am-plitude were significantly variable, but other-wise usually failed to perceive actually occur-ring stress.

Paralleling outcomes on the production side,the link between speech acoustics and arousalis generally stronger than the link betweenacoustics and perceived valence (cf. Laukka etal., 2005). Ladd et al. (1985), for instance, sys-tematically varied several acoustic parametersas participants rated vocalizer affect and atti-tude. Listeners in this study did not show cate-gorical response patterns in their attributionsof emotional states, but rather perceived vocal-izer arousal to be varying continuously in ac-cordance with similarly continuous changes in

F0. Other results have put dimensional and dis-crete accounts on more equal footing (e.g.,Scherer & Oshinsky, 1977; see also Hatfield etal., 1995). Of interest here is the pattern of mis-identifications that occurs when participantsare asked to identify discrete vocalizer states.Confusion matrices have shown that errors aremost likely to occur for emotions that are simi-lar in arousal (e.g., Pakosz, 1983; Pereira,2000) and between similarly valenced membersof emotion “families” (e.g., Banse & Scherer,1996; see also Breitenstein, Van Lancker, &Daum, 2001; Ladd et al., 1985).

There are thus at least two issues to considerwhen the discrete-emotions and dimensionalapproaches are contrasted from perceptual evi-dence. The first is that listener accuracy is inter-mediate, averaging about 55% across a num-ber of studies. Some investigators have arguedthat this result shows that vocalizer emotion isindeed associated with discrete patterns ofacoustic cues, and that these effects are percep-tible by listeners (e.g., Banse & Scherer, 1996;Johnstone & Scherer, 2000). We take a morecautious stance and argue that even better per-formance should be expected, given that thesamples used have typically been portrayals ofemotion produced by trained actors, and haveusually been culled for quality or proto-typicality by experimenters prior to beingtested with listeners. Both aspects may wellstrongly decrease acoustic variability relativeto naturally occurring emotional speech, andthereby artificially inflate listener scores. Giventhese circumstances, an outcome of 55% cor-rect seems low if emotion-specific acoustics arein fact present, and probably overestimateshow listeners would perform under less artifi-cial circumstances.

A second issue to consider is the extent towhich emotion-specific cues are actually pres-ent in naturally occurring speech, given that at-tributions about emotion also involve percep-tual processes, inferential capabilities, and evensimple biases that listeners bring to the situa-tion. For example, listeners tend to place dis-proportionate weight on particular acousticcues, whether or not these cues reliably covarywith vocalizer state (e.g., Johnstone & Scherer,2000; Streeter et al., 1983). Given that this sortof “stereotyping” is prominent in many auto-matic evaluations of personal characteristics inothers (Banaji & Hardin, 1996; Kawakami,Young, & Dovidio, 2002), it would not be sur-prising to find that listener inferences concern-


ing vocalizer affect from acoustics alone aresimply not very accurate. Under natural cir-cumstances, a listener would be hearing thespeech within a rich social context and oftenfrom a familiar vocalizer. Both factors couldsignificantly boost the accuracy of attribution(Bachorowski, 1999; Pakosz, 1983).

From a methodological point of view, ittherefore becomes critical to separate these fac-tors explicitly, given that perception of emotionfrom speech may rely as much or more on ac-tive interpretation by a listener as it does onacoustic cues provided by the vocalizer. Thispossibility is elaborated in the next section,which shifts the focus from hypothesized linksbetween emotional states and vocalizer acous-tics to the effects that sounds have on listeneraffect. Although compatible with various as-pects of both discrete and dimensional ac-counts of emotions and emotional expression,this perspective suggests that emotion-relatedcueing by vocalizers is actually a secondaryoutcome of the communication process. Theprimary function of emotional vocal acousticsis instead proposed to be influencing the lis-tener’s state and concomitant behavior.

AN AFFECT INDUCTION ACCOUNTOF VOCAL SIGNALING

Whether from a discrete-emotions or a dimen-sional perspective, a central theme of work onemotion-related aspects of vocal signals hasbeen that there are veridical links among vocal-izer affect, associated vocal acoustics, and lis-tener perception. This perspective assumes thatthe purpose of emotion-related signaling is forthe vocalizer to inform listeners that particu-lar affective states are being experienced. Ofcourse, simply informing listeners about suchstates is not necessarily beneficial. Instead, theeffect that signaling has on subsequent listenerbehavior must be what has shaped the sig-naling process over evolutionary time (seeDawkins, 1989; Dawkins & Krebs, 1978; seealso Owren & Bachorowski, 2001). It there-fore becomes important that there is no guar-antee that listeners will behave in ways thatbenefit vocalizers who are providing veridicalcues to their internal states. This observation inturn suggests that the most fundamental selec-tion pressure acting on signalers must be tomodulate others’ behavior in ways that arebeneficial to themselves. This logic leads to a

different way of thinking about affect-relatedvocal acoustics: From this point of view, theirfunction cannot be to inform as much as to in-fluence. We suggest that such influence can oc-cur either via the direct impact of signal acous-tics on listener affective systems or throughlearned associations between vocalizer acous-tics and listener state, thereby shaping the waylisteners behave toward the vocalizer. Afterbriefly describing this perspective, we specifi-cally apply it to the case of human laughter—aseemingly ubiquitous, affect-related vocal sig-nal.

Affect Induction through Sound

Anyone who has spent time with a crying in-fant can well appreciate that auditory stimulican readily induce emotion-related responses inlisteners. Sounds as varied as those of doorsopening and closing, sirens wailing, and thun-derstorms booming can all elicit attention,arousal, and valenced responses. Laughter andinfant crying are two of the more potent,affect-inducing vocal signals, but even speechacoustics can produce emotion in listeners(Neumann & Strack, 2000; Hatfield et al.,1995; Siegman & Boyle, 1993). More broadly,Bradley and Lang (2000) have shown that af-fective responses occur to an array of environ-mental and human sounds, and that these re-sponses can be readily organized along arousaland pleasure dimensions.

The fact that sounds can induce affect hasplayed little role in work on production andperception of vocal expression of emotion.Instead, as noted earlier, both the discrete-emotions and dimensional perspectives seem-ingly approach the problem from a representa-tional standpoint, in the sense that affect-related meaning is “encoded” by vocalizers andsubsequently “decoded” by listeners (e.g.,Juslin & Laukka, 2001; Scherer, 1988, 2003).Emotional expressions are thus treated as acode that vocalizers use to convey their emo-tional states—a perspective that at least implic-itly draws on a linguistically inspired interpre-tation.

The “affect induction” model has been pro-posed as an alternative to representational per-spectives. Originally developed in the contextof nonhuman primate vocalizations (Owren& Rendall, 1997, 2001; Rendall & Owren,2002), this approach may also have broad ap-plicability to human affective signaling (Owren


et al., 2003). Rather than treating emotionalcommunication as a process of information en-coding and decoding, the approach argues thatthe primary function of emotion-related vocalsignals is to influence listeners’ affect andthereby also to modulate their behavior.Whereas representational accounts of commu-nication implicitly implicate rather sophisti-cated but typically undescribed processes of in-formation encoding and decoding, the affectinduction approach argues that vocal signals“work” because they have an impact on listen-ers at comparatively low levels of neural orga-nization by eliciting emotional responses. Theeffects can be “direct,” meaning that signalacoustics themselves have an impact, or “in-direct,” meaning that listeners experience alearned affective response to sounds as a resultof previous experience. For the former, impactdepends on the signal energy itself, meaningthat such aspects as variability, amplitude, du-ration, and overall salience are of primary im-portance. In the latter, learning arises throughsocial interactions, and depends on instances inwhich individually distinctive sounds producedby a given vocalizer are repeatedly paired withthe affect being experienced by a listener.

Differences between the affect induction andrepresentational perspectives can be elucidatedin the context of the similarities that exist be-tween them. There are at least two importantpoints of contact, with the first being that bothapproaches assume some association betweenthe signaler’s internal state and the signal pro-duced. The representational approach arguesthat the function of signaling is to allow the lis-tener to infer that this particular state is occur-ring in the vocalizer. This view therefore pre-dicts that signals should be strongly associatedwith differentiated signaler states. The affectinduction approach instead proposes that thevocalizer’s internal state is important in themechanistic underpinnings of signal produc-tion, but that the function of signaling is to in-duce emotion in the listener. As a result, signalacoustics need not be strongly linked to vocal-izer states, because it may benefit the vocalizerto induce similar responses in the listeneracross a variety of situations. Conversely, a di-verse set of acoustic properties in sounds pro-duced in a given situation may serve a commonfunction in modulating listener arousal andvalenced emotion. The affect induction ap-proach therefore expects associations betweenvocalizer state and physical signal, but main-

tains that these associations will be probabilis-tic in nature.

A second point of contact is that listeners areprobably able to draw inferences about vocal-izer states and/or likely upcoming behaviors.From a representational perspective, such in-ferences are part and parcel of why communi-cation signals evolve: Signal recipients receiveencoded information about a signaler’s stateand act on that content. The affect inductionapproach, in contrast, views listener inferenceas a secondary outcome of vocal behavior,which has evolved first and foremost because itbenefits the vocalizer to influence listener affectand behavior. Substantial variability in vocaliz-ing behavior is thus to be expected. However,there will be probabilistic patterning involved,which listeners will inevitably benefit from at-tending to and learning about, to the extentthat they themselves benefit from drawinginferences about vocalizer states and likelybehavior.

Although affect-related signals do not haverepresentational value in this view, they are still“meaningful” in the sense that listeners makeinferences about their significance based on ahost of factors, which can include the acousticattributes of the sound, the listener’s affectivestate and familiarity with the sound, and theoverarching context in which the sound has oc-curred (see also Bachorowski & Owren, 2002;Hess & Kirouac, 2000; Kappas et al., 1991). Acorollary is that affective responses to highlysimilar sounds may be quite variable. For ex-ample, hearing a high-pitched shriek may bepleasurable when one is attending a party, butquite negative when one is walking down adark, isolated street. In both cases, the acous-tics of the shriek are likely to elicit orientation,increase listener arousal, and exacerbate what-ever affective state the listener is alreadyexperiencing—positive and negative, respec-tively, in these two situations. In addition, if thesound has distinctive acoustic features thathave previously been paired with either posi-tive or negative affect in the listener, the soundwill activate corresponding learned responses.Finally, the sound can have a larger and proba-bly more complex inferred significance in agiven context—for instance, if the shriek“means” that someone has had too much todrink and should be driven home.

Although it is thus fundamentally differentfrom representational approaches to emotionalexpression, the affect induction perspective


nonetheless has parallels with Scherer’s “ap-peal” and “pull” functions of vocal expressionof emotion (Scherer, 1988, 1992; see also John-stone & Scherer, 2000), as well as with the con-struct of “emotion contagion.” Hatfield,Cacioppo, and Rapson’s (1992) notion of“primitive” emotional contagion is particularlyrelevant, in that this process is described as ei-ther unconditioned or conditioned, and as oc-curring outside the realm of conscious aware-ness. However, the affect induction perspectivetakes mechanism a step further than the no-tions of “appeal,” “pull,” or “contagion” byspecifically emphasizing direct links betweenthe acoustics of affect-related vocalizations andlow-level neural responses in listeners.

A larger implication of the affect inductionapproach is that vocal emotion expressionsfunction most importantly as nonconsciousstrategies of social influence (see also Bargh &Chartrand, 1999; Zajonc, 1980). Rather thanacting as informative beacons to a vocalizer’sstate, these signals are held to sway or shapeperceiver affect and subsequent behavior or at-titude toward the vocalizer. Any informationvalue the signal has thus represents a combina-tion of inferences the perceiver may be able todraw, given the context at hand, previous gen-eral experience with such signals, and theunique history of interaction shared by the in-dividuals involved (Hess & Kirouac, 2000;Owren & Rendall, 2001).

Applying an Affect Induction Approachto Laughter

There is no shortage of hypotheses concerningthe associations between laughter and affect-related states on the part of vocalizers. On theone hand, laughter has obvious links with posi-tive states such as happiness (e.g., Darwin,1872/1998; Nwokah, Davies, Islam, Hsu, &Fogel, 1993; van Hooff, 1972) and with theenjoyment associated with humor (Deacon,1989; Edmonson, 1987; Weisfeld, 1993). Onthe other hand, laughter has also been noted toaccompany negative emotions such as guilt,shame, and nervousness (Darwin, 1872/1998;McComas, 1923). Other signaling values havealso been suggested, including sexual interestand sexual fitness (Dunbar, 1996; Grammer &Eibl-Eibesfeldt, 1990), self-deprecation (Glenn,1991–1992), appeasement or submission(Adams & Kirkevold, 1978; Deacon, 1997;Dovidio, Brown, Heltman, Ellyson, & Keating,

1988; Grammer & Eibl-Ebesfeldt, 1990), anddenigration (Eibl-Ebesfeldt, 1989).

These approaches to laughter have at leasttacitly adopted a representational stance inwhich laugh acoustics themselves are consid-ered to provide information about vocal state,with listeners being held to make inferencesabout that state from signal acoustics (and per-haps contextual cues). An implication is there-fore that if laughter is to represent vocalizer af-fect, it should exhibit differentiated acousticproperties that then allow listeners to make in-ferences about the laugher’s corresponding in-ternal states—in other words, that laugh acous-tics vary in accordance with affective state. Toour knowledge, there are no direct tests of thishypothesis.

As might be expected from the variety of hy-pothesized functions concerning laughter, theavailable data in fact show laugh acoustics tobe remarkably variable (Bachorowski, Smoski,& Owren, 2001; Grammer & Eibl-Ebesfeldt,1990; Vettin & Todt, 2004). However, thisvariability actually poses a challenge for repre-sentational perspectives, because it occurs bothwithin and among laughers reporting similararousal and positive affect states (see Owren &Bachorowski, 2003; Owren et al., 2003).Laughs in the Bachorowski et al. (2001) studywere recorded from 97 individuals as theywatched humorous film clips either alone orwith a same- or other-sex friend or stranger.The results of detailed acoustic analyses re-vealed that laughs could be reliably groupedinto voiced and unvoiced (noisy, breathy) cate-gories, with the latter being further separableinto grunt- and snort-like versions. There wasstriking variability in a number of F0-relatedmeasures in voiced laughs, suggesting thatmany of these sounds should have significant,direct impact on listener arousal and affect.Specifically, both males and females producedlaughs with very high mean and maximum F0’s,as well as considerable variability or F0 modu-lation both within single laugh sounds andacross bouts of laughter. A number of laughsalso had notable acoustic irregularities (i.e.,acoustic nonlinearities), which we suspect areespecially potent in tweaking listener responsesystems.

Further analyses (see Bachorowski, Smoski,Tomarken, & Owren, 2008) showed that laughacoustics were only loosely coupled with self-reported affect—an outcome that is inconsis-tent with a representational view of these


sounds. These acoustics were, however, signifi-cantly associated with the social circumstancesof the testing situation (see also Devereux &Ginsburg, 2001; Grammer & Eibl-Ebesfeldt,1990; Vettin & Todt, 2004). So, for example,males laughed the most and produced a highproportion of laughs likely to have potent di-rect effects on listeners when paired with afriend, especially a male friend. Females, on theother hand, were more likely to produce thesekinds of laugh sounds in the company of amale. Although a careful accounting of the in-teractions between laugher sex and social con-text is beyond the scope of this chapter, the gistis that both males and females produce moredirect-effect laughs when it would benefit themto elicit arousal and affect in listeners, and spe-cifically do not produce such sounds wheninducing arousal in listeners might only exa-cerbate negatively toned (e.g., wary) emo-tional states in these individuals (Owren &Bachorowski, 2003; Owren et al., 2003).

As on the production side, perceptual re-sponses to laughter have not received muchsystematic empirical attention. Nevertheless,the available data do show that laugh acousticselicit affective responses in listeners and canhave a modulating effect on a social partner’sattitude toward a vocalizer. As a first step to-ward testing hypotheses about the direct effectsthat laugh sounds may exert on listener states,we (Bachorowski & Owren, 2001) had listen-ers rate a sample of voiced and unvoicedlaughs. Outcomes were remarkably consistent,regardless of whether listeners rated their ownaffect in response to hearing a laugh, the likelyaffective response other listeners would experi-ence, or three affect-related attributes of thelaugher. Voiced laughs were given significantlymore positive evaluations than were unvoicedlaughs, supporting the idea that the acousticproperties of the former (e.g., high mean F0,marked F0 modulation, and high amplitude)activate listeners in a direct fashion. More re-cent work has shown that unvoiced laughs canactually elicit slightly negative affect in listeners(Owren, Trivedi, Schulman, & Bachorowski,2007). Grammer and Eibl-Ebesfeldt (1990; seealso Keltner & Bonanno, 1997) have furthershown that the occurrence of voiced laughter isassociated with listener interest in the laugher:Males reported being more interested in fe-males they had just met if those females pro-duced a larger number of voiced, but not un-voiced, laughs during a 10-minute interval.

Other research has begun to focus on indi-rect or learned effects. Here the expectation isthat learned emotional responses should accruein listeners experiencing positive affect, aslaughers are able to repeatedly produce indi-vidually distinctive sounds in close temporalassociation with the occurrence of those lis-tener states. We (Bachorowski et al., 2001)achieved moderate success in statistical classifi-cation of both the voiced and unvoiced laughsof individual vocalizers, suggesting that laughsare in fact individually distinctive. Initial be-havioral support for the notion of indirect ef-fects was obtained by finding that laughter isfar more likely to show coincident occurrencebetween friends than between strangers(Smoski & Bachorowski, 2003a), and that thistemporal link is more frequent for voiced thanfor unvoiced laughs (Smoski & Bachorowski,2003b).

The studies reviewed here are just the begin-ning, however, with much remaining to belearned about this affect-related vocal signal. Inthe current context, the most important pointto note is that there are some important con-nections between the major findings emergingfrom studies of emotional communicationthrough speech acoustics and laughter.

CONCLUSIONS:VOCAL EXPRESSIONS OF EMOTIONAS SOCIAL TOOLS

Laughter resembles speech in being a decidedlysocial event (Provine & Fischer, 1989) and acritical part of the communicative processesthat are central to human social relationships.There also appear to be some telling common-alities in how the acoustic features of bothspeech and laughter function in emotional as-pects of communication. A routine finding forspeech is that F0- and amplitude-related fea-tures play a primary role. On the one hand,speech produced while a vocalizer is experienc-ing salient emotion usually shows significantchanges in these particular characteristics,which also clearly have an impact on listenerperception and attributions of emotion. On theother hand, there is little indication that theseor other acoustic features of speech are tightlycoupled with either discrete emotional states orvalenced affect, and when listeners are asked toidentify vocalizer states, performance based onspeech acoustics alone is modest. Although less


information is available for laughter, outcomesso far have been similar. For example, the de-gree of variability occurring in laugh soundsproduced during positive states alone arguesagainst the possibility of emotion-specificacoustics. Laugh sounds nonetheless have un-ambiguous impact on listener emotion, withvoicing and F0-related features playing a cen-tral role.

It has thus been difficult to demonstrateclearly differentiated, state-specific acoustic ef-fects for either speech or laughter, but quiteeasy to show that acoustics nonetheless havegeneral and robust effects on listeners’ emo-tions and evaluation of vocalizers (see alsoScherer, 1986). To us, these results raise ques-tions about the adequacy of representationalapproaches, and are instead more compatiblewith the view that vocalizers are using rela-tively undifferentiated acoustics that have po-tent effects on listener affect. The evidence isfar from definitive, particularly as much of thework conducted on human vocal expression ofemotion has assumed rather than tested theproposition that representational signaling isinvolved. However, it is therefore also note-worthy that relevant research has largely failedto support this assumption either for pro-duction or for perception. Although listenersreadily hear emotional “content,” their judg-ments about a vocalizer’s states are constrainedand context-dependent. Our interpretation isthat these inferences primarily reflect the lis-teners’ own arousal and valenced responses,which are richly interpreted while also takingother information about the vocalizer and con-text into account. Affective communication inboth speech and laughter thus becomes a pro-cess of attributing emotion to vocalizers, ratherthan one of recovering encoded information.

A larger implication of this view is that emo-tional expressions in general function most im-portantly as nonconscious strategies of socialinfluence, rather than as veridical represen-tations of internal state (see also Bargh &Chartrand, 1999; Zajonc, 1980). Rather thaninforming per se, both kinds of signals functionto sway or shape the perceiver’s affect and atti-tude, promoting behavioral effects that ulti-mately benefit the signaler. The informationvalue of a signal is thus critically dependent ona perceiver’s previous general experience, par-ticular history with the signaler, and ability totake signaling context into account (Hess &Kirouac, 2000; Owren & Rendall, 2001).

These are difficult distinctions to make, how-ever, and illustrate the need to significantlyimprove on the data that are currently avail-able concerning vocal emotion expression(Douglas-Cowie, Campbell, Cowie, & Roach,2003). We suggest that it will be useful to focuson vocalizations acquired under controlled butnaturalistic circumstances, where vocalizers areexperiencing actual rather than simulated emo-tions. Furthermore, alternative interpretationsshould be contrasted as directly as possible, in-cluding discrete-emotions versus dimensionalapproaches and representational versus affectinduction perspectives within the same experi-mental framework. Finally, it will be importantto broaden the scope of inquiry, so that the in-ferential processes used by listeners to make at-tributions about vocalizer states can be uncov-ered. With these desiderata as a framework,a deeper understanding of how humans useemotion-related vocal acoustics in their com-municative endeavors should be within reach.

ACKNOWLEDGMENTS

During this work, Michael J. Owren was supportedin part by National Institute of Mental Health PrimeAward No. 1 R01 MH65317-01A2, Subaward No.8402-15235-X; by the Brains and Behavior Programof Georgia State University; and by the Center forBehavioral Neuroscience, STC Program of the Na-tional Science Foundation, under Agreement No.IBN-9876754.

REFERENCES

Adams, R. M., & Kirkevold, B. (1978). Looking, smil-ing, laughing, and moving in restaurants: Sex and agedifferences. Environmental Psychology and Nonver-bal Behavior, 3, 117–121.

Bachorowski, J.-A. (1999). Vocal expression and per-ception of emotion. Current Directions in Psycholog-ical Science, 8, 53–57.

Bachorowski, J.-A., & Owren, M. J. (1995). Vocal ex-pression of emotion: Acoustic properties of speechare associated with emotional intensity and context.Psychological Science, 6, 219–224.

Bachorowski, J.-A., & Owren, M. J. (1996). Vocal ex-pression of emotion is associated with vocal fold vi-bration and vocal tract resonance. Psychophysiology,33(Suppl. 1), S20.

Bachorowski, J.-A., & Owren, M. J. (1999). Acousticcorrelates of talker sex and individual talker identityare present in a short vowel segment produced inrunning speech. Journal of the Acoustical Society ofAmerica, 106, 1054–1063.


Bachorowski, J.-A., & Owren, M. J. (2001). Not alllaughs are alike: Voiced but not unvoiced laughterreadily elicits positive affect. Psychological Science,12, 252–257.

Bachorowski, J.-A, & Owren, M. J. (2002). The role ofvocal acoustics in emotional intelligence. In L. F.Barrett & P. Salovey (Eds.), The wisdom in feeling:Psychological processes in emotional intelligence(pp. 11–36). New York: Guilford Press.

Bachorowski, J.-A., Smoski, M. J., & Owren, M. J.(2001). The acoustic features of human laughter.Journal of the Acoustical Society of America, 110,1581–1597.

Bachorowski, J.-A., Smoski, M. J., Tomarken, A. J., &Owren, M. J. (2008). Laugh rate and acoustics areassociated with social context. Manuscript submittedfor publication.

Banaji, M. R., & Hardin, C. D. (1996). Automatic ste-reotyping. Psychological Science, 7, 136–141.

Banse, R., & Scherer, K. R. (1996). Acoustic profiles invocal emotion expression. Journal of Personality andSocial Psychology, 70, 614–636.

Bargh, J. A., & Chartrand, T. L. (1999). The unbearableautomaticity of being. American Psychologist, 54,462–479.

Barrett, L. F. (2006). Are emotions natural kinds? Per-spectives on Psychological Sciences, 1, 28–58.

Bradley, M. M., & Lang, P. J. (2000). Affective reactionsto acoustic stimuli. Psychophysiology, 37, 204–215.

Breitenstein, C., Van Lancker, D., & Daum, I. (2001).The contribution of speech rate and pitch variationto the perception of vocal emotions in a German andan American sample. Cognition and Emotion, 15,57–79.

Cauldwell, R. T. (2000). Where did the anger go?: Therole of context in interpreting emotion in speech. InR. Cowie, E. Douglas-Cowie, & M. Schröder (Eds.),Proceedings of the International Speech Communica-tion Association Workshop on Speech and Emotion(pp. 127–131). Belfast: Textflow.

Cicero, M. T. (1904). De oratore (2nd ed.). Translatedand with an introduction by E. N. P. Moor. London:Methuen.

Cowie, R. (2000). Describing the emotional states ex-pressed in speech. In R. Cowie, E. Douglas-Cowie, &M. Schröder (Eds.), Proceedings of the InternationalSpeech Communication Association Workshop onSpeech and Emotion (pp. 11–18). Belfast: Textflow.

Darwin, C. (1998). The expression of the emotions inman and animals. New York: Oxford University.(Original work published 1872)

Dawkins, R. (1989). The selfish gene (2nd ed.). Oxford:Oxford University Press.

Dawkins, R., & Krebs, J. R. (1978). Animal signals: Infor-mation or manipulation? In J. R. Krebs & N. B. Davies(Eds.), Behavioural ecology: An evolutionary approach(pp. 282–309). Oxford: Blackwell Scientific.

Deacon, T. W. (1989). The neural circuitry underlyingprimate calls and human language. Human Evolu-tion, 4, 367–401.

Deacon, T. W. (1997). The symbolic species. New York:Norton.

Devereux, P. G., & Ginsburg, G. P. (2001). Sociality ef-fects on the production of laughter. Journal of Gen-eral Psychology, 128, 227–240.

Douglas-Cowie, E., Campbell, N., Cowie, R., & Roach,P. (2003). Emotional speech: Towards a new genera-tion of databases. Speech Communication, 40, 33–60.

Dovidio, J. F., Brown, C. E., Heltman, K., Ellyson, S. L.,& Keating, C. F. (1988). Power displays betweenwomen and men in discussions of gender-linkedtasks: A multichannel study. Journal of Personalityand Social Psychology, 55, 580–587.

Dunbar, R. I. M. (1996). Grooming, gossip, and theevolution of language. London: Faber &Faber.

Edmonson, M. S. (1987). Notes on laughter. Anthropo-logical Linguistics, 29, 23–34.

Eibl-Ebesfeldt, I. (1989). Human ethology. New York:Aldine de Gruyler.

Fant, G. (1960). Acoustic theory of speech production.S-Gravenhage, The Netherlands: Mouton.

Frick, R. W. (1985). Communicating emotion: The roleof prosodic features. Psychological Bulletin, 97, 412–429.

Glenn, P. J. (1991–1992). Current speaker initiation oftwo-party shared laughter. Research on Languageand Social Interaction, 25, 139–162.

Grammer, K., & Eibl-Eibesfeldt, I. (1990). The ritualiza-tion of laughter. In W. Koch (Ed.), Naturlichkeit dersprache und der kultur: Acta colloquii (pp. 192–214). Bochum, Germany: Brockmeyer.

Green, R. S., & Cliff, N. (1975). Multidimensionalcomparisons of structures of vocally and facially ex-pressed emotion. Perception and Psychophysics, 17,429–438.

Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1992).Primitive emotional contagion. In M. S. Clark (Ed.),Emotion and social behavior (pp. 151–177). Lon-don: Sage.

Hatfield, E., Hsee, C. K., Costello, J., Weisman, M. S.,& Denney, C. (1995). The impact of vocal feedbackon emotional experience and expression. Journal ofSocial Behavior and Personality, 10, 293–312.

Hess, U., & Kirouac, G. (2000). Emotion expression ingroups. In M. Lewis & J. M. Haviland-Jones (Eds.),Handbook of emotions (2nd ed., pp. 368–381). NewYork: Guilford Press.

Johnson, K. (1997). Acoustic and auditory phonetics.Cambridge, MA: Blackwell.

Johnson, W. F., Emde, R. N., Scherer, K. R., & Klinnert,M. D. (1986). Recognition of emotion from vocalcues. Archives of General Psychiatry, 43, 280–283.

Johnstone, T., & Scherer, K. R. (2000). Vocal communi-cation of emotion. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook of emotions (2nd ed.,pp. 220–235). New York: Guilford Press.

Johnstone, T., van Reekum, C. M., Hird, K., Kirsner, K.,& Scherer, K. R. (2005). Affective speech elicitedwith a computer game. Emotion, 5, 513–518.


Juslin, P. N., & Laukka, P. (2001). Impact of intendedemotional intensity on cue utilization and decodingaccuracy in vocal expression of emotion. Emotion, 1,381–412.

Kappas, A., & Hess, U. (1995). Nonverbal aspects oforal communication. In D. U. M. Quasthoff (Ed.),Aspects of oral communication (pp. 169–180).Berlin: De Gruyter.

Kappas, A., Hess, U., & Scherer, K. R. (1991). Voiceand emotion. In B. Rime & R. Feldman (Eds.), Fun-damentals of nonverbal behavior (pp. 200–238).Cambridge, UK: Cambridge University Press.

Kawakami, K., Young, H., & Dovidio, J. F. (2002). Au-tomatic stereotyping: Category, trait, and behavioralactivations. Personality and Social Psychology Bulle-tin, 28, 3–15.

Keltner, D., & Bonanno, G. (1997). A study of laughterand dissociation: Distinct correlates of laughter andsmiling during bereavement. Journal of Personalityand Social Psychology, 73, 687–702.

Ladd, D. R., Silverman, K. E. A., Tolkmitt, F.,Bergmann, G., & Scherer, K. R. (1985). Evidence forthe independence of intonation contour type, voicequality, and F0 range in signaling speaker affect.Journal of the Acoustical Society of America, 78,435–444.

Laukka, P. (2005). Categorical perception of vocal emo-tion expressions. Emotion, 5, 277–295.

Laukka, P., Juslin, P. N., & Bresin, R. (2005). A dimen-sional approach to vocal expression of emotion. Cog-nition and Emotion, 19, 633–653.

Leinonen, L., Hiltunen, T., Linnankoski, I., & Laakso,M.-L. (1997). Expression of emotional–motivationalconnotations with a one-word utterance. Journal ofthe Acoustical Society of America, 102, 1853–1863.

Lieberman, P., & Blumstein, S. E. (1988). Speech physi-ology, speech perception, and acoustic phonetics.Cambridge, UK: Cambridge University Press.

McComas, H. C. (1923). The origin of laughter. Psy-chological Review, 30, 45–56.

Millot, J.-L., & Brand, G. (2001). Effects of pleasantand unpleasant ambient odors on human voice pitch.Neuroscience Letters, 297, 61–63.

Neumann, R., & Strack, F. (2000). “Mood contagion”:The automatic transfer of mood between persons.Journal of Personality and Social Psychology, 79,211–223.

Nwokah, E. E., Davies, P., Islam, A., Hsu, H. C., &Fogel, A. (1993). Vocal affect in three-year-olds: Aquantitative acoustic analysis of child laughter. Jour-nal of the Acoustical Society of America, 94, 3076–3090.

Owren, M. J., & Bachorowski, J.-A. (2001). The evolu-tion of emotional expression: A “selfish-gene” ac-count of smiling and laughter in early hominids andhumans. In T. Mayne & G. A. Bonanno (Eds.), Emo-tions: Current issues and future directions (pp. 152–191). New York: Guilford Press.

Owren, M. J., & Bachorowski, J.-A. (2003). Recon-sidering the evolution of nonlinguistic communica-

tion: The case of laughter. Journal of NonverbalBehavior, 27, 183–200.

Owren, M. J., & Bachorowski, J.-A. (2007). Measuringvocal acoustics. In J. A. Coan & J. J. B. Allen (Eds.),The handbook of emotion elicitation and assessment(pp. 239–266). New York: Oxford University Press.

Owren, M. J., & Rendall, D. (1997). An affect-conditioning model of nonhuman primate signaling.In D. H. Owings, M. D. Beecher, & N. S. Thompson(Eds.), Perspectives in ethology: Vol. 12. Communi-cation (pp. 299–346). New York: Plenum Press.

Owren, M. J., & Rendall, D. (2001). Sound on the re-bound: Bringing form and function back to the fore-front in understanding nonhuman primate vocal sig-naling. Evolutionary Anthropology, 10, 58–71.

Owren, M. J., Rendall, D., & Bachorowski, J.-A.(2003). Nonlinguistic vocal communication. In D.Maestripieri (Ed.), Primate psychology (pp. 359–394). Cambridge, MA: Harvard University Press.

Owren, M. J., Trivedi, N., Schulman, A., &Bachorowski, J.-A. (2007). An implicit associationtest (IAT) with nonlinguistic auditory stimuli: Lis-tener evaluations of human laughter. Manuscriptsubmitted for publication.

Paeschke, A., & Sendlmeier, W. F. (2000). Prosodiccharacteristics of emotional speech: Measurements offundamental frequency movements. In R. Cowie, E.Douglas-Cowie, & M. Schröder (Eds.), Proceedingsof the International Speech Communication Associa-tion Workshop on Speech and Emotion (pp. 75–80).Belfast: Textflow.

Pakosz, M. (1983). Attitudinal judgments in intonation:Some evidence for a theory. Journal of Psycho-linguistic Research, 12, 311–326.

Pell, M. D. (2001). Influence of emotion and focus loca-tion on prosody in matched statements and ques-tions. Journal of the Acoustical Society of America,109, 1668–1680.

Pereira, C. (2000). Dimensions of emotional meaning inspeech. In R. Cowie, E. Douglas-Cowie, & M.Schröder (Eds.), Proceedings of the InternationalSpeech Communication Association Workshop onSpeech and Emotion (pp. 25–28). Belfast: Textflow.

Pittam, J., Gallois, C., & Callan, V. (1990). The long-term spectrum and perceived emotion.Speech Com-munication, 9, 177–187.

Pittam, J., & Scherer, K. R. (1993). Vocal expressionand communication of emotion. In M. Lewis & J. M.Haviland (Eds.), Handbook of emotions (pp. 185–197). New York: Guilford Press.

Protopapas, A., & Lieberman, P. (1997). Fundamentalfrequency of phonation and perceived emotionalstress. Journal of the Acoustical Society of America,101, 2267–2277.

Provine, R. R., & Fischer, K. R. (1989). Laughing, smil-ing, and talking: Relation to sleeping and social con-text in humans. Ethology, 83, 295–305.

Rendall, D., & Owren, M. J. (2002). Animal vocal com-munication: Say what? In M. Bekoff, C.Allen, & G.


Burghardt (Eds.), The cognitive animal (pp. 307–314). Cambridge, MA: MIT Press.

Russell, J. A., Bachorowski, J.-A., & Fernandez-Dols,J.-M. (2003). Facial and vocal expressions of emo-tion. Annual Review of Psychology, 54, 329–349.

Scherer, K. R. (1986). Vocal affect expression: A reviewand model for future research. Psychological Bulle-tin, 99, 143–165.

Scherer, K. R. (1988). On the symbolic functions of vo-cal affect expression. Journal of Language and SocialPsychology, 7, 79–100.

Scherer, K. R. (1989). Vocal measurement of emotion.In R. Plutchik & H. Kellerman (Eds.), Emotion: The-ory, research, and experience Vol. 4. The measure-ment of emotions (pp. 233–259). New York: Aca-demic Press.

Scherer, K. R. (1992). Vocal affect expression as symp-tom, symbol, and appeal. In H. Papoušek, U.Jürgens, & M. Papoušek (Eds.), Nonverbal vocalcommunication: Comparative and developmentalapproaches (pp. 43–60). Cambridge, UK: CambridgeUniversity Press.

Scherer, K. R. (2003). Vocal communication of emotion:A review of research paradigms. Speech Communica-tion, 40, 227–256.

Scherer, K. R., Banse, R., & Wallbott, H. G. (2001).Emotion inferences from vocal expression correlateacross languages and cultures. Journal of Cross-Cultural Psychology, 32, 76–92.

Scherer, K. R., Banse, R., Wallbott, H. G., & Goldbeck,T. (1991). Vocal cues in emotion encoding and de-coding. Motivation and Emotion, 15, 123–148.

Scherer, K. R., Ladd, D. R., & Silverman, K. E. A.(1984). Vocal cues to speaker affect: Testing twomodels. Journal of the Acoustical Society of America,76, 1346–1356.

Scherer, K. R., & Oshinsky, J. S. (1977). Cue utilizationin emotion attribution from auditory stimuli. Moti-vation and Emotion, 1, 331–346.

Schröder, M. (2000). Experimental study of affectbursts. In R. Cowie, E. Douglas-Cowie, & M.Schröder (Eds.), Proceedings of the InternationalSpeech Communication Association Workshop onSpeech and Emotion (pp. 132–135). Belfast: Text-flow.

Siegman, A. W., & Boyle, S. (1993). Voices of fear andanxiety and sadness and depression: The effects ofspeech rate and loudness on fear and anxiety andsadness and depression. Journal of Abnormal Psy-chology, 102, 430–437.

Smoski, M. J., & Bachorowski, J.-A. (2003a). Antipho-nal laughter between friends and strangers. Cogni-tion and Emotion, 17, 327–340.

Smoski, M. J., & Bachorowski, J.-A. (2003b). Antipho-nal laughter in developing friendships. Annals of theNew York Academy of Sciences, 1000, 300–303.

Sobin, C., & Alpert, M. (1999). Emotion in speech: Theacoustic attributes of fear, anger, sadness, and joy.Journal of Psycholinguistic Research, 28, 347–365.

Stevens, K. N. (2000). Acoustic phonetics. Cambridge,MA: MIT Press.

Streeter, L. A., Macdonald, N. H., Apple, W., Krauss, R.M., & Galotti, K. M. (1983). Acoustic and percep-tual indicators of emotional stress. Journal of theAcoustical Society of America, 73, 1354–1360.

Tassinary, L. G., & Cacioppo, J. T. (1992). Un-observable facial actions and emotion. PsychologicalScience, 3, 28–33.

Tickle, A. (2000). English and Japanese speakers’ emo-tion vocalisation and recognition: A comparisonhighlighting vowel quality. In R. Cowie, E. Douglas-Cowie, & M. Schröder (Eds.), Proceedings of theInternational Speech Communication AssociationWorkshop on Speech and Emotion (pp. 104–109).Belfast: Textflow.

Tolkmitt, F. J., & Scherer, K. R. (1986). Effects of exper-imentally induced stress on vocal parameters. Journalof Experimental Psychology: Human Perception andPerformance, 12, 302–313.

Trouvain, J., & Barry, W. J. (2000). The prosody of ex-citement in horse race commentaries. In R. Cowie, E.Douglas-Cowie, & M. Schröder (Eds.), Proceedingsof the International Speech Communication Associa-tion Workshop on Speech and Emotion (pp. 86–91).Belfast: Textflow.

van Hooff, J. A. R. A. M. (1972). A comparative ap-proach to the phylogeny of laughter and smiling. InR. A. Hinde (Ed.), Non-verbal communication(pp. 209–241). Cambridge, UK: Cambridge Univer-sity Press.

Vettin, J., & Todt, D. (2004). Laughter in conversation:Features of occurrence and acoustic structure. Jour-nal of Nonverbal Behavior, 28, 93–115.

Wallbott, H. G., & Scherer, K. R. (1986). Cues andchannels in emotion recognition. Journal of Person-ality and Social Psychology, 51, 690–699.

Weisfeld, G. E. (1993). The adaptive value of humorand laughter. Ethology and Sociobiology, 14, 141–169.

Williams, C. E., & Stevens, K. N. (1972). Emotionsand speech: Some acoustical correlates.Journal ofthe Acoustical Society of America, 52, 1238–1250.

Zajonc, R. B. (1980). Feeling and thinking: Preferencesneed no inferences. American Psychologist, 35, 151–175.


C H A P T E R 1 3

Facial Expressions of Emotion

DAVID MATSUMOTO, DACHER KELTNER, MICHELLE N. SHIOTA,MAUREEN O’SULLIVAN, and MARK FRANK

Within the field of emotion, the study of facialexpressions has been notable both for empiri-cal advances and for theoretical controversy. Inthis chapter, we draw upon an “evolutionist”approach to emotion, inspired by Charles Dar-win, to draw together recent studies of facialexpression. This literature indicates that facialexpressions of emotion, as described by Dar-win over 135 years ago, (1) include universal,reliable markers of discrete emotions whenemotions are aroused and there is no reason tomodify or manage the expression; (2) covarywith distinct subjective experience; (3) are partof a coherent package of emotion responsesthat includes appraisals, physiological reac-tions, other nonverbal behaviors, and subse-quent actions, as well as individual differencesand mental and physical health; (4) are judgedas discrete categories; and (5) as such, servemany interpersonal and social regulatory func-tions.

PERSPECTIVE AND ASSUMPTIONS

An evolutionist approach to facial expressionhas its roots in the work of Darwin (1872/1998) and of those who have refined and elab-orated upon his evolutionist claims (Ekman,1992b; Izard, 1971). Darwin claimed, in hisprinciple of serviceable habits, that facial ex-pressions are the residual actions of morecomplete behavioral responses, and occur incombination with other bodily responses—vocalizations, postures, gestures, skeletal mus-cle movements, and physiological responses.For example, we express anger by furrowingthe brow and tightening the lips with teeth dis-played, because these actions are part of an at-tack response; we express disgust with an openmouth, nose wrinkle, and tongue protrusion aspart of a vomiting response. Facial expressions,then, are elements of a coordinated responseinvolving multiple response systems.

211

As part of our evolutionary heritage, accord-ing to Darwin, all people, regardless of race orculture, should express emotions in the face andbody in similar fashion. Darwin wrote The Ex-pression of the Emotions in Man and Animals torefute the claims of Sir Charles Bell, the leadingfacial anatomist of his time and a teacher of Dar-win’s, about how God designed humans withunique facial muscles to express uniquely hu-man emotions.1 Relying on advances in photog-raphy and anatomy (Duchenne de Boulogne,1862/1990), Darwin engaged in a detailed studyof the muscle actions involved in emotion (seeTable 13.1); he concluded that the muscle ac-tions are universal, and their precursors can beseen in the expressive behaviors of nonhumanprimates and other mammals.

Within the evolutionist framework thatguides our analysis, facial expressions shouldcovary with emotional experience, in large partbecause the signals that accompany invol-untary experience give additional credibilityto the display (Ekman, 1989; Ekman &O’Sullivan, 1991; although see Fridlund, 1994;Hauser, 1993; Krebs & Dawkins, 1984, for al-ternative perspectives). Those facial expres-sions that covary with emotion, it has beenfound, have certain properties, including briefduration, symmetry of muscle actions, and thepresence of involuntary muscle actions (Ekman& Friesen, 1982; Ekman & Rosenberg, 2005).Facial expressions that accompany actual emo-tional experience are more reliable signals; theyact as commitment devices to likely courses ofaction that are momentarily beyond the indi-vidual’s volitional control (R. H. Frank, 1988;Gonzaga, Keltner, & Londahl, 2001).

The evolutionist perspective also suggeststhat facial expressions are more than simplereadouts of internal states; they coordinatesocial interactions through their informative,evocative, and incentive functions (Keltner &Kring, 1998). They provide information toperceivers about the individual’s emotionalstate (Ekman, 1993; Scherer, 1986), behavioralintentions (Fridlund, 1994), relational statusvis-à-vis the target of the expression (Keltner,1995; Tiedens, Ellsworth, & Mesquita, 2000),and objects and events in the social environ-ment (Mineka & Cook, 1993). This view of fa-cial expressions emerged from developmentalstudies of emotional exchanges between par-ents and children (Hertenstein & Campos,2004; Klinnert, Campos, & Sorce, 1983;Klinnert, Emde, Butterfield, & Campos, 1986),

as well as from ethological studies of such so-cial behaviors as flirting, reconciliation,aggression, and play. It is consistent withclaims regarding the coevolution of signal andperceiver response to displays (Eibl-Ebesfeldt,1989; Hauser, 1993). Thus an individual’semotional expression serves as a “socialaffordance” that evokes “prepared” responsesin others (Esteves, Dimberg, & Öhman, 1994).Anger, for example, may have evolved to elicitfear-related responses and the inhibition of in-appropriate action (Dimberg & Öhman,1996); note that the Japanese often label an-other person’s angry expression as “scary”(Matsumoto, 2006). Distress calls may haveevolved to elicit sympathetic responses in ob-servers (Eisenberg et al., 1989). Through theseprocesses, emotional communication helps in-dividuals in relationships—parents and chil-dren, mates, bosses and subordinates—respondto the demands and opportunities of their so-cial environment. They are basic elements ofsocial interaction, from flirtatious exchanges togreeting rituals. This perspective provides acompelling rationale for the prediction thatpeople should be reliable judges of emotionaldisplays, and sets the stage for the claim thatdeficits in expression are associated with psy-chological disorders.

The evolutionist perspective that we have de-scribed thus far leads to the following fiveclaims, for which we review the most recentfindings. Specially, an evolutionist approach tofacial expression holds that discrete facial ex-pressions of emotion (1) occur universally inemotionally arousing situations; (2) are linkedwith subjective experience; (3) are part of a co-herent package of emotional responses; (4) arejudged universally and discretely; and (5) haveimportant social functions. In the discussionthat follows, we bring together recent evidencethat bears upon these claims.

THE CLAIMS OFTHE EVOLUTIONIST PERSPECTIVE:AN EMPIRICAL REVIEW

Universality of Facial EmotionalExpressions in EmotionallyArousing Situations

Claims concerning the universality of facial ex-pressions of emotion are rooted in the notionthat the facial anatomy is brought into servicein expressions to solve similar problems across


13. Facial Expressions of Emotion 213

TABLE 13.1. Descriptions of Facial Muscles and Other Nonverbal Behaviors Involvedin the Emotions Darwin Considered Universal

Emotion

Darwin’s description(nonfacial elements inparentheses)

Action units(AUs)associated withDarwin’sdescription

AUs found to beassociated with thisemotion in researchwith humans (optionalAUs in parentheses)

Homologousor analogousAUs found inchimpanzees

Anger Nostrils raised, mouthcompressed, furrowed brow,eyes wide open, head erect,(chest expanded, arms rigidby sides, stamping ground,body swaying backward/forward, trembling]

4; 5; 24; 38 4; 5 or 7; 22; 23; 24 4; 22; 23; 24

Contempt Lip protrusion, nose wrinkle,partial closure of eyelids,turning away eyes, upper lipraised, (snort, bodyexpiration, expiration)

9; 10; 22; 41;61 or 62

Unilateral 12;unilateral 14

9; 10

Disgust Lower lip turned down, upperlip raised, expiration, mouthopen, spitting, blowing out,protruding lips, throat-clearing sound, lower lip andtongue protruding

10; 16; 22; 25or 26

9 or 10; (25 or 26) 9; 10

Fear Eyes open, mouth open, lipsretracted, eyebrows raised,(crouching, paleness,perspiration, hair standing onend, muscles shivering,yawning, trembling)

1; 2; 5; 20 1; 2; 4; 5; 20; (25 or26)

1; 2; 4

Happiness Eyes sparkling, skin undereyes wrinkled, mouth drawnback at corners

6; 12 6; 12 6; 12

Joy Zygomatic and orbicularismuscles contracted, upper lipraised, nasolabial fold formed,(muscles trembling,purposeless movements,laughter, clapping hands,jumping, dancing about,stamping, chuckling/giggling)

6; 7; 12 6; 12 6; 12

Sadness Corners of mouth depressed,inner corner eyebrows raised,(low spirits)

1; 15 1; (4); 15; (17) 1; 4; 15; 17

Surprise Eyebrows raised, mouth open,eyes open, lips protruding,(expiration, blowing/hissing,open hands high above head,palms toward person withstraightened fingers, armsbackwards)

1; 2; 5; 25 or26

1; 2; 5; 25 or 26 1; 2

Note. The action unit (AU) numbers are those of Ekman and Friesen’s (1978) Facial Action Coding System (FACS).

cultures related to social living, such as restor-ing justice, attending to others in need, signal-ing danger, expressing sexual or affiliative in-terest, and so on. By implication, the facialmuscles themselves should be universal, and in-deed they are. All humans around the world,regardless of race or culture, have the same fa-cial anatomy (Gray & Goss, 1966). This uni-versal facial musculature, furthermore, appearsto be activated in emotion-specific ways acrosscultures.

Evidence from Adult Humans across Cultures

The strongest evidence for the universality offacial expressions of emotion comes from stud-ies that directly measure facial behaviors whenemotions are elicited. The first was Ekman’s(1972) well-known study involving Americanand Japanese participants who viewed neutraland stressful films, and whose facial behaviorswere recorded throughout the experiment (un-beknownst to them). Ekman coded the last 3minutes of facial behavior during the neutralfilms, and the entire 3 minutes of the last stressfilm clip, using a modified version of the FacialAffect Scoring Technique (FAST), a precursorto the Facial Action Coding System (FACS;Ekman & Friesen, 1978). The FAST identifiedfacial configurations of six emotions—anger,disgust, fear, happiness, sadness, and surprise—in different regions of the face. Two sets ofanalyses were performed on the facial codes:one involving separate facial areas, and one in-volving the whole face. The rank-order correla-tions on the facial behavior codes from the sep-arate areas between the American and Japaneseparticipants ranged from .72 for the eyes–lidsarea to .92 for the brows–forehead area. Whenthe codes were combined into emotion-relatedconfigurations, the correlations ranged from.86 in the brows–forehead region to .96 in thelower face. Disgust, sadness, anger, and sur-prise were the most frequently displayed emo-tions, but fear and happiness were also evident.When facial codes were combined for whole-face emotions, according to the theoretical ra-tionales of Darwin and of Tomkins (1962,1963), the correlation between the Americansand the Japanese on the frequencies of whole-face emotions expressed spontaneously was.88.

Subsequent research has yielded further evi-dence supportive of the notion that theoreti-cally relevant, universal facial expressions are

elicited by specific emotionally evocative stim-uli. There are at least 25 published studies inwhich the facial behaviors of individuals whoparticipated in emotionally arousing condi-tions were coded reliably with the FACS andmatched to the universal facial configurationsof emotion (see Table 13.2). These studies dem-onstrate that the facial configurations of atleast seven emotions, as postulated by Darwinand Tomkins, are produced when emotion isaroused and there is no reason to modify theexpression because of social circumstances.(See also Eibl-Ebesfeldt, 1989, for an ethologi-cal perspective.)

The range of cultures in the 26 studies (25 inTable 13.2 and Ekman, 1972) is extensive.Matsumoto and Willingham’s (2006) study, forinstance, involved 84 athletes from 35 coun-tries. Participants in other studies were Ameri-cans, Japanese, Germans, Canadians, andFrench. Collectively, these studies demonstratethat the facial expressions reported originallyby Ekman actually do occur when emotion isaroused in people of different cultures. Table13.1 contrasts the specific facial muscles origi-nally proposed by Darwin (1872/1998) withthe facial action units (AUs) that have beenshown to be related to various emotions ac-cording to Ekman and Friesen’s (1978) FACS.

Evidence from Nonhuman Primates

For years, ethologists (Chevalier-Skolnikoff,1973; Geen, 1992; Hauser, 1993; Snowdon,2003; Van Hoof, 1972) have noted the mor-phological similarities between human expres-sions of emotion and nonhuman primate ex-pressions displayed in similar contexts. VanHoof (1972) described the evolution of thesmile and laugh along two different evolution-ary tracks across early mammals, monkeys,apes, chimpanzees, and humans. Redican(1982) suggested that among nonhuman pri-mates, facial displays described as grimacesand open-mouth grimaces are akin to the hu-man emotions of fear and surprise; that thetense-mouth display is similar to anger; andthat grimaces and a tense mouth combinedform the often identified threat display.Redican also noted that nonhuman primatesshow a play face similar to the happy face ofhumans, and he suggested that the nonhumanpout serves a similar function to the human sadface. Ueno, Ueno, and Tomonaga (2004) dem-onstrated that both infant rhesus macaques



TABLE 13.2. Studies Examining Spontaneous Facial Expressions of Emotion

Citation ParticipantsEliciting stimuli orsituation

Measurementsystem

Emotionscorresponding to thefacial muscleconfigurations inuniversal expressions

Bonanno &Keltner(1997,2004);Keltner &Bonanno(1997)

Germanschizophrenic andpsychosomaticpatients, andhealthy controls

Engaging in a politicalconversation with apartner they had nevermet before

EMFACS Contempt, disgust,anger, sadness, fear,surprise, happiness

Bonanno etal. (2002)

Conjugallybereavedindividuals

Interviews about theirdeceased spouses orongoing importantrelationships

EMFACS Anger, contempt,disgust, fear, sadness,Duchenne smiles

Camras etal. (1992)

Individuals withexperience ofchildhood sexualabuse

Narrative interviewsabout the mostdistressing event or seriesof events in their lives

EMFACS Anger, disgust,sadness, fear,Duchenne smiles

Chesney etal. (1990)

American andJapanese infants

Arm restraint, whichproduces distress

FACS Anger, sadness, fear,and happiness

Ekman et al.(1980)

American salariedemployees inmanagerialpositions at anaerospace firm

Structured interviewdesigned to assess Type Abehavior

FACS Disgust, fear, sadness,happiness, anger,contempt, surprise

Ekman et al.(1988)

American collegestudents

Films designed to elicitpositive and negativeemotion

FACS Happiness, unspecifiednegative emotions

Ekman et al.(1990)

Student nurses Films designed to elicitstrong negative emotions

FACS Happiness

Ekman et al.(1997)


Two film clips designedto elicit positiveemotions, and twodesigned to elicit negativeemotions

FACS Happiness

Ellgring(1986)

Depressedinpatients

Intake and dischargeinterviews

FACS andEMFACS

Happiness, contempt,anger, disgust, fear,sadness

Frank et al.(1993)

German depressedpatients

Interviews FACS Happiness

Gosselin etal. (1995)

AmericanUniversity students

Films designed to elicitvarious emotions

FACS Happiness

Harris &Alvarado(2005)

Actors from theConservatory ofDramatic Arts inQuébec

Interpreting 2 of 24scenarios designed toelicit happiness, fear,anger, surprise, sadness,and disgust

FACS Happiness, fear, anger,surprise, sadness, anddisgust

(continued)


TABLE 13.2. (continued)

Citation ParticipantsEliciting stimuli orsituation

Measurementsystem

Emotionscorresponding to thefacial muscleconfigurations inuniversal expressions

Heller &Haynal(1994)


Either being tickled,listening to an audiotapeof jokes, or placing ahand in ice water

FACS Duchenne smiles

Keltner et al.(1995)

French depressedpatients

Interviews with thepatients’ psychiatrists

FACS andEMFACS

Contempt

Lerner et al.(2005)

Americanadolescents withbehavior problems

Administration of theWechsler IntelligenceScale for Children—Revised

EMFACS Anger, fear, andsadness

Matsumoto&Willingham(2005)


Induction of three kindsof stress

EMFACS Fear, anger, anddisgust

Mauss et al.(2005)

Olympic medalists Immediately afterwinning or losing amedal in competition,and on the podium

FACS Six different types ofsmiles, contempt,disgust, fear, andsadness

Gross &Levenson(1993)


Films designed to elicitamusement and sadness

EmotionExpressiveBehaviorCoding(Messingeret al., 2001)

Amusement andsadness

Rosenberg& Ekman(1994)

Infants and theirmothers

Play sessions betweeninfants and mothers

FACS andBaby FACS

Duchenne smiles ofhappiness

Ruch (1995) Americanuniversity students

Videos selected for theirability to elicit primarilydisgust and secondarilyfear

FACS Disgust, sadness, fear,happiness, contempt,and anger

Ruch (1993) German universitystudents

Slides of jokes andcartoons

FACS Happiness

Soto et al.(2005)

German universitystudents

Slides of jokes andcartoons

FACS Happiness

Gross &Levenson(1993)

Chinese Americanand MexicanAmerican collegestudents

Aversive acoustic startle EmotionExpressiveBehaviorCoding(Levenson,2003b)

Anger, anxiety,disgust, confusion,contempt, interest,embarrassment, fear,happiness, sadness,surprise, crying,laughter

and infant chimpanzees showed different facialexpressions to sweet and bitter tastes, but thatthe chimps’ facial expressions were more simi-lar to human facial expressions than to those ofthe macaques. However, even some of thesmaller apes, such as siamangs (Symphalangussyndactylus), noted for their limited facial ex-pression repertoire, have distinguishable facialexpressions accompanying sexuality, agonisticbehavior, grooming, and play (Liebal, Pika, &Tomasello, 2004). De Waal (2002) suggeststhat for some states a species less closely relatedto humans than chimpanzees, the bononos,may have more emotions in common with hu-mans.

The most recent research has gone beyonddemonstrating equivalence in morphologicaldescriptions of expressions to identifying theexact facial musculature involved in the displaybehavior.2 Indeed, the strongest support for thebiological bases of facial expression would be ademonstration of homology in facial expres-sions between humans and related species, suchas chimpanzees (Pan troglodytes). This wouldsuggest that as humans evolved during the 120million years of primate evolution, similar fa-cial muscles developed, presumably to servesimilar social functions. The newest work inthis area, by Waller and colleagues (2006), re-ported that the forehead musculature ofchimps is less well developed than that of hu-mans. (They speculate that the greater hairinessof chimps makes eyebrow movements less visi-ble, and hence less communicative.) But manyother facial muscles and expressions havehomologues and analogues comparable tothose defined in the human FACS (Ekman &Friesen, 1978). For example, Waller et al.(2006) report that many of the muscles codedon the human face have the same location andfunctional effect in humans and chimpanzees(see Table 13.1). Based on this work, there isnow a ChimpFACS that allows for identifica-tion of the specific AUs chimpanzees use inproducing facial expressions (see Vick, Waller,Parr, Pasqualini, & Bard, 2007).3

Linkages between Facial EmotionalExpressions and Subjective Experience

The evolutionist perspective suggests a linkagebetween each universal signal of emotion and asubjective experience, in particular when thereis no reason to manage or modify the expres-sion because of social circumstances. Eleven

studies reported in Table 13.2 report correla-tions between emotion-specific facial behaviorsand self-reports of the experience of the dis-crete emotion (these findings are summarizedin Table 13.3). It is noteworthy that linkagesbetween discrete facial expressions of emotionand self-reports of the same emotional statesare stronger in within-subject designs that in-volve precise, second-to-second measurementof both expression and experience, such asRosenberg and Ekman’s (1994) study and thatby Mauss, Levenson, McCarter, Wilhelm, andGross (2005). In the latter, cross-lag correla-tions indicated very high within-individual cor-relations between facial behavior and experi-ence intensity for both amusing and sadness-eliciting films (r’s = .73 and .74, respectively).When correlations were corrected fordisattenuation, they were even higher (r’s = .89and .97).

Moreover, important nonfindings not re-flected in Table 13.3 need to be considered. InBonanno and Keltner’s (1997) study, for in-stance, anger, contempt, and sadness were posi-tively correlated with reports of grief, but fearand disgust were not. In Ekman, Friesen, andAncoli’s (1980) study, expressions of disgustwere positively correlated with disgust but neg-atively correlated with anger and sadness. InHarris and Alvarado’s (2005) study, Duchennesmiles were correlated with happiness andamusement, but not with reports of feelinganxious, angry, or embarrassed.

Complementing the studies listed in Table13.3 are several lines of research that provideconvergent evidence. Matsumoto and Kupper-busch (2001) reported significant correlationsbetween judged expressions and self-reportedexperience. Duchenne smiles have been corre-lated with the experience of positive emotion inyoung and old adults (Frank, Ekman, &Friesen, 1993; Hess, Banse, & Kappas, 1995;Keltner & Bonanno, 1997; Smith, 1995). Du-chenne and non-Duchenne smiles distinguishednonharassed and harassed job applicants(Woodzicka & LaFrance, 2001), as well ashonest and deceptive interviewees (Ekman,Friesen, & O’Sullivan, 1988). Frank andEkman (1997) reported predicted differences infear between honest and deceptive men. The fa-cial signals related to embarrassment andamusement (e.g., gaze aversion and smile con-trols vs. the open-mouthed smile) were corre-lated with self-reports of these emotions(Keltner, 1995). Spontaneous laughter and



TABLE 13.3. Studies Reporting Significant Correlations between Spontaneous Facial Expressionsof Emotion and Self-Reports of Specific Emotions

CitationaFacial expressionsmeasured

Self-reportsobtained Correlation

Bonanno &Keltner (1997)

Duchenne laughing Grief (concurrent) –.39*Anger .36*Contempt .31*Sadness .34*

Keltner &Bonanno (1997)

Duchenne smiles Distress –.49**Fear –.31*Enjoyment .35*

Duchenne laughing Distress –.36*Enjoyment .34*

Bonanno &Keltner (2004)

Duchenne smiles Distress –.44*Duchenne laughing Anger –.51**

Ekman et al.(1980)

Smiling(frequency)

Happiness .60**

Smiling (duration) Happiness .35*Smiling (intensity) Happiness .34*Disgust(frequency)

Disgust .37*

Anger –.35*Sadness –.46**

Disgust (duration) Disgust .55**Fear .46*Pain .41*

Ekman et al.(1990)

Duchenne smiles Amusement .70*Happiness .59*Excitement .39***Interest .40***Anger –.38***Sadness –.44***

Harris &Alvarado (2005)

Duchenne smiles Happiness .19† (humor condition)Amusement .28**

Mauss et al.(2005)

Duchenne smiles Amusement .73***Sadness Sadness .74***

Rosenberg &Ekman (1994)b

Disgust and fear Disgust and fear .71 (rat film).90 (amputation film).83 (amputation film)

Ruch (1993) Duchenne smiles Positive affectivity .33* (experimental group).52* (control group).78* (rank order of cell means)

Verbal enjoyment .28*

Ruch (1995) Duchenne smiles Funniness .55 (between-subjects design, aggregate data).96 (within-subjects design, aggregate data).61 (between-subjects design, raw data).71 (within-subjects design, raw data).63 (across all stimuli and designs)

aGosselin et al. (1995) obtained self-reports but did not report correlations between the ratings and facial expressions.bThe statistics reported for this study are the probabilities of co-occurrence between the ratings of specific emotion categoriesand the corresponding facial expressions.*p < .05; **p < .01; ***p < .001; †p < .10

smiling were found to have some distinct expe-riential correlates (Keltner & Bonanno, 1997).The intensity of laughter or smiling correlatedwith self-reports of the funniness of the humor-ous stimuli (McGhee, 1977; Ruch, 1995).

Facial Expressions as Part of a CoherentPackage of Emotional Responses

Darwin (1872/1998) suggested, in his principleof serviceable habits, that facial expressions arethe residual actions of more complete behav-ioral responses. By implication, facial expres-sions not only should be related to emotionalexperience; they should also be coordinatedwith other components, such as autonomic orneuroendocrine changes, that enable the organ-ism to respond adaptively. Researchers refer tothis possibility in terms of “emotion pack-ages,” “emotion response system coherence,”or “response covariation” (Bonanno & Kelt-ner, 2004; Ekman, 1992a; Lazarus, 1991;Levenson, 1994). Distinct lines of evidence sug-gest that brief facial expressions of emotioncovary in systematic fashion with appraisalprocesses, physiological responses, specific ac-tions (e.g., aggressive behavior or cooperation),broad individual differences in emotionality,and measures of physical and mental health.

Co-occurrence with Distinct Appraisals

If facial expressions are part of a coherent re-sponse profile, they should covary withemotion-specific appraisal processes. Evidencesuggests that this is the case. For example, be-reaved adults’ facial expressions of anger andsadness while discussing their deceased spousesco-occurred with distinct appraisal themes (jus-tice and loss) coded from participants’ sponta-neous narratives that were contemporaneouswith the facial expressions (Bonanno &Keltner, 1997). Another study found that pos-ing facial expressions of anger was related tothe appraisal that others were responsible forsocial events, whereas posing facial sadnesswas associated with the appraisal that the sameevents were due to situational causes (Keltner,Ellsworth, & Edwards, 1993). Moreover,spontaneous facial expressions reliably differ-entiate whether Olympic athletes have won orlost a medal, and differences in their smilingbehavior differentiate what kind of medal theywon (Matsumoto & Willingham, 2006). As afinal example, expressions of anger, contempt,

and disgust are reliably associated with ap-praisals related to moral violations ofautonomy, community, and divinity, respec-tively (Rozin, Lowery, Imada, & Haidt, 1999).Taken together, these studies suggest that facialexpressions of distinct emotions covary withspecific appraisals.

Covariance with Distinct Physiological Responses

Facial expressions are also coordinated withphysiology. When emotions are aroused and fa-cial expressions are used as markers of thoseemotions, discrete physiological signatures oc-cur in both the autonomic nervous system andthe brain (Davidson, 2003; Ekman, Davidson,& Friesen, 1990; Ekman, Levenson, & Friesen,1983; Levenson, Carstensen, Friesen, &Ekman, 1991; Levenson & Ekman, 2002;Levenson, Ekman, & Friesen, 1990; Levenson,Ekman, Heider, & Friesen, 1992; Mauss et al.,2005; Tsai & Levenson, 1997). Table 13.4,adapted from Levenson (2003b), summarizesthe major findings in this area and highlightshow emotions signaled in facial expression areassociated with activity in other physiologicalsystems. These patterns have been found inpeople from cultures as widely divergent as theUnited States and the Minangkabau of WestSumatra, Indonesia.

Alongside these findings, it is important tonote that the coherence of facial expression andphysiology has not always been consistent.Some studies (Brown & Schwartz, 1980;Cacioppo, Martzke, Petty, & Tassinary, 1988)demonstrated only low correlations betweenexpression and physiological response, andsome found no relationship (Buck, 1977;Mauss, Wilhelm, & Gross, 2004). These nega-tive findings are the likely results of severalmethodological factors: (1) the fact that the oc-currence of an emotion is sometimes defined bythe attempt to manipulate it, instead of the in-dependent confirmation of its elicitation; (2)the type of emotion elicited; (3) the nature ofthe measures of emotional responding used; (4)the temporal resolution of the measurement(Mauss et al., 2005; Rosenberg & Ekman,1994); (5) the fact that the laboratory may notbe the optimal context in which to elicit adap-tive physiological responses; and (6) the differ-ence between between- and within-subjects de-signs. The Mauss et al. (2005) study describedearlier highlights the importance of the last is-sue. They measured facial behaviors, emotional


experience, and three types of physiological re-sponse (skin conductance, cardiovascular acti-vation, and somatic activity) with second-by-second precision while participants watchedfilms designed to elicit amusement and sadness.The results indicated clear, moderate-sized,within-individual correlations between facialbehavior and the various physiological re-sponse components.

Most recently, Lerner, Gonzalez, Dahl,Haririr, and Taylor (2005) demonstrated thatthe discrete facial expressions of fear, anger,and disgust were reliably linked not only tocardiovascular responses, but to neuroendo-crine activity as well. Participants were ex-posed to three different types of stressors dur-ing which they were videotaped, and their

cardiovascular and hypothalamic–pituitary–adrenocortical (cortisol) responses were mea-sured. Fear expressions were associated withelevated cardiovascular and cortisol levels; an-ger and disgust were linked with reduced re-sponses. Matsumoto, Nezlek, and Koopmann(2007) reported moderate-sized correlationsbetween self-reported expressive behavior andthree types of physiological sensations (ergo-tropic, trophotropic, and felt temperature) inapproximately 3,000 respondents from 27countries. They also reported consistent corre-lations between verbal and nonverbal expres-sions, as well as between emotion intensity andphysiological sensations, all of which suggestcoherence in an underlying neurophysiologicalreality.


TABLE 13.4. Changes in Appearance and Autonomic Nervous System (ANS) Activity Associatedwith the Discrete Emotional States Darwin Considered Universal

Emotion

AUs associated withthe physiologicalchanges reported Type of change Change ANS mediation

Anger 1; 4; 5; 17; 23 or 24 Coloration Reddening Vasodilation,increasedcontractability

Moisture andsecretions

Foaming Salivary glands

Protrusions Piloerection Muscle fibers at baseof hair follicles

Blood vessels bulging VasodilationEye appearance Constriction Pupils

Bulging Eyelid muscles

Disgust 9 or 10 Moisture andsecretions

Salivating, drooling Salivary glands

Fear 1; 2; 4; 5; 7; 20 Coloration Blanching VasoconstrictionMoisture and

secretionsSweating, clamminess Sweat glands

Protrusions Piloerection Muscle fibers at baseof hair follicles

Eye appearance Dilation PupilsBulging Eyelid muscles

Happiness 6; 12 Eye appearance Twinkling Lacrimal glands pluscontraction oforbicularis oculi

Sadness 1 (or 1 + 4); 15 Moisture andsecretions

Tearing, crying Lacrimal glands

Embarrassment Coloration Blushing Vasodilation

Note. Adapted with permission from Levenson (2003b). Changes from the original table include the reorganization of con-tents according to emotion, the addition of action units (AUs) associated with physiological changes, minor wording changes,and the removal of sexual arousal from the emotion category.

Covariance with Subsequent Behaviors

Another source of evidence supporting thelinks between expressions and emotional re-sponses comes from studies that demonstratecovariance between facial expressions of emo-tion and subsequent behaviors. Facial expres-sions of emotion can signal behavioral intent.In the first study to demonstrate this effect,Ekman, Liebert, et al. (1972) examined the re-lationship between facial expressions of emo-tion produced by children as they watched tele-vision and their subsequent hurtful behaviorsand aggressive play. Children were videotapedas they watched either a violent scene from amovie or competitive sports. Afterwards, theywere placed in a situation where they could ei-ther help or hurt another child, and then en-gaged in a free-play period. Boys who smiledduring the violent scenes engaged in more hurt-ful behavior and aggressive play afterwards;boys who showed sadness during the violentscenes engaged in more helpful behavior andless aggressive play when the video was fin-ished.

Matsumoto, Haan, Gary, Theodorou, andCooke-Carney (1986) videotaped the facial be-haviors of preschool dyads as they either coop-erated or competed in a Prisoner’s Dilemmagame, and used the Emotion Facial ActionCoding System (EMFACS) to code facial be-haviors. The children displayed varied emo-tional responses, and these were reliably linkedto the actions of the game. Cooperative be-haviors elicited decreased negative emotion,whereas children who were defected againstexpressed more non-Duchenne smiles andpositive–negative blends. Most importantly, fa-cial expressions that occurred after an action—cooperation or competition—predicted subse-quent behavior. Children who expressed Du-chenne smiles after their partners cooperatedwere more likely to cooperate also; those whoexpressed anything else after cooperation weremore likely to defect. Defections that were fol-lowed by non-Duchenne smiling were morelikely to lead to subsequent, repeated defec-tions; when defections were followed by Du-chenne smiling, however, the subsequent actwas likely to be cooperative.

More recently, Keltner, Moffitt, andStouthamer-Loeber (1995), in their study ofadolescent boys, examined relations betweenfacial expressions of emotion observed in a 2-

minute interaction and teacher reports of socialbehavior. Facial displays of anger observed inan interactive IQ testing context correlated sig-nificantly with teacher ratings of delinquentand aggressive behavior at school; facial dis-plays of fear correlated negatively with thesebehaviors, and positively with withdrawal-related behaviors.

Covariance with Broad Individual Differences

Further evidence consistent with the claim thatfacial expressions covary with multisystem re-sponses comes from studies linking facial ex-pressions to measures of individual differencesin emotionality. For example, as noted earlier,Bonanno and Keltner (1997) used the FACS tocode bereaved adults’ facial expressions as theytalked about their recently deceased spouses.Facial expressions of anger predicted increasedgrief severity 14 and 25 months after loss;laughing and smiling, however, predicted re-duced grief over time. Importantly, facial ex-pressions predicted long-term adjustment, in-dependently of initial levels of grief andindividual differences in the tendency to reporthigh levels of distress. These findings suggestthat brief measures of facial expressions (in thisstudy, 6 minutes of behavior were coded) pre-dict broad patterns of adaptation to importantlife events.

Harker and Keltner (2001) coded women’scollege yearbook photos with the FACS, andshowed that Duchenne smiling was positivelycorrelated with multiple measures of person-ality (i.e., affiliation, warmth, competence),personal well-being, and marital satisfactionat various times over the next 30 years. AndAbe and Izard (1999) measured discrete facialexpressions of emotion in 18-month-old in-fants during episodes of the Strange Situationprocedure, and correlated these with maternalratings of the five-factor model of personalitywhen the children were 3.5 years old. Nega-tive expressions were strongly correlated withNeuroticism and inversely related to Agree-ableness and Conscientiousness; full-face pos-itive expressions were positively correlatedwith Extraversion and Openness to Experi-ence. These findings suggest that facial ex-pressions of discrete emotions systematicallycovary with coherent patterns of thought,feeling, and action as captured in personalitymeasures.


Covariance with Measures of Mentaland Physical Health

If facial expressions are part of more coherentresponses to the environment, they should like-wise be related to measures of mental andphysical health, which capture maladaptivepatterns of thought, action, and feeling. Severalstudies indicate that fleeting facial expressionsof emotion are telling clues about personal ad-justment. Anger displays are related to the inci-dence of ischemia in patients with coronary ar-tery disease (Rosenberg et al., 2001). Theoblique eyebrows and pressed lips of sympathycorrelate with reduced heart rate, whereaswinces of pain are related to elevated heart rate(Eisenberg et al., 1989). Facial expressions dif-ferentiate among genuine pain, masked pain,and faked pain (Craig, Hyde, & Patrick, 1991;Prkachin, 1992), as well as between healthy in-dividuals and psychiatric patients (Ellgring,1986; Steimer-Krause, Krause, & Wagner,1990), schizophrenic and psychosomatic pa-tients (Steimer-Krause et al., 1990), schizo-phrenic and depressed patients (Berenbaum &Oltmanns, 1992; Ekman, Matsumoto, &Friesen, 1997; Ellgring, 1986), suicidal andnonsuicidal depressed patients (Heller &Haynal, 1994), and patients with major versusminor depression (Ekman et al., 1997). InEkman et al.’s (1997) study, for instance, pa-tients with major depression showed moresadness and disgust and fewer non-Duchennesmiles than those with minor depression.Manic patients showed more Duchenne andnon-Duchenne smiles, and less anger, disgust,or sadness than either group with depression.Schizophrenic patients showed more fear andfewer displays of all other emotions. Moreover,expressions of contempt and unfelt happinessmeasured during intake interviews of the de-pressed patients predicted improvement at dis-charge. In another study, facial expressions ofdisgust (and glaring) differentiated between in-dividuals with Type A and Type B personalities,and facial expressions of contempt, anger, anddisgust were all correlated with various speechindices of hostility, anger, competitiveness, anddespondency (Chesney et al., 1990). Moreover,Duchenne smiles differentiated whether de-pressed patients were improving as a result oftreatment (Ekman et al., 1997), and patientswith right-hemisphere damage were impairedin the production of facial expressions of emo-

tions, particularly positive emotion (Borod,Koff, Lorch, & Nicholas, 1986).

Pathologies also affect the ability to pose andrecognize emotional expressions. Abused chil-dren, for instance, have difficulties both posingand recognizing facial expressions (Camras etal., 1988; Pollak, Cicchetti, Hornung, & Reed,2000; Pollak & Sinha, 2002). Severely autisticchildren have broad and pervasive deficitsin recognizing emotions (Hobson, 1986;Ozonoff, Pennington, & Rogers, 1990).Children with high-functioning autism orAsperger’s syndrome have emotion-specificdeficits: They are generally able to recognizehappiness, sadness, fear, and anger (Capps,Yirmiya, & Sigman, 1992), but not facial sig-nals of embarrassment or shame (Heerey,Keltner, & Capps, 2003). Individuals withhigh trait anxiety recognize fearful faces betterthan those without such anxiety (Surcinelli,Codispoti, Montebarocci, Rossi, & Baldaro,2004), and individuals who are depressed aregenerally worse at recognizing all facial emo-tions (Persad & Polivy, 1993). Individuals withcurrent substance dependence and a history ofalcohol dependence are also generally worse atrecognizing facial emotions (Foisy et al., 2005).And patients with myotonic dystrophy Type 1have difficulty recognizing angry, disgusted,and fearful faces (Winblad, Hellstrom,Lindberg, & Hansen, 2006).

More than 135 years ago, Charles Darwinclaimed that brief facial expressions are reallytokens of more complex, multisystem re-sponses to specific environmental demands oropportunities. The literature we have reviewedin this section elaborates upon this early evolu-tionist claim, indicating that facial expressionsof distinct emotions covary systematically withdistinct appraisals, physiological response, so-cial behavior, stable individual differences, andmarkers of physical and mental health.

Recognition of FacialEmotional Expressions:Universality and Brain Involvement

Universal Recognition

Early studies supporting the universal recogni-tion of facial expressions of emotion were judg-ment studies, in which observers of differentcultures viewed facial stimuli and judged the


emotions portrayed in them. The earliest stud-ies by Ekman and Izard demonstrated the exis-tence of six universal expressions—anger, dis-gust, fear, happiness, sadness, and surprise—inliterate and preliterate cultures (Ekman, 1972,1973; Ekman & Friesen, 1971; Ekman,Sorenson, & Friesen, 1969; Izard, 1971). Evenwhen low-intensity expressions were used(Ekman et al., 1987; Matsumoto et al., 2002),there was strong agreement across culturesabout the emotion in the expression.

Since the original studies by Ekman andIzard, 27 studies examining judgments of facialexpressions have replicated the finding of uni-versal recognition of facial expressions of emo-tion (Matsumoto, 2001). In addition, a meta-analysis of 168 datasets examining judgmentsof emotion in the face and other nonverbalstimuli indicated universal emotion recognitionat well above chance levels (Elfenbein &Ambady, 2002b). Even after correction forchance guessing, this statistic was associatedwith a very large effect size, consistently sup-porting the findings that expressions are uni-versally recognized.

More recent research continues to comple-ment this conclusion. Horstmann (2003), forinstance, asked approximately 2,000 onlineparticipants whether the universal facial ex-pressions signaled feeling states, behavioral in-tentions, or action requests. The majority ofthe observers chose feeling states as the pri-mary message of the expressions of disgust,fear, sadness, surprise, and happiness. (Partici-pants did choose behavioral intentions or ac-tion requests as the message of angry ex-pressions.) Lawrence et al. (2005) asked 484children between 6 and 16 years of age to judgethe universal faces. All of these children wereable to judge the emotions accurately beyondchance levels, and the accuracy rates increasedlinearly across age for happiness, surprise, fear,and disgust; for sad and angry expressions, ac-curacy rates remained constant across age. AndParr (2003) has demonstrated that chimpan-zees are able to discriminate five emotional ex-pressions with similarities to human expres-sions (relaxed-lip face, pant-hoot, play face/relaxed open-mouth face, scream face, andbared-teeth face) and a neutral face. Moreover,chimpanzees are able to match different emo-tional faces with specific behaviors—a findingsuggesting that they understand specific mean-ings associated with facial expressions, much

as humans do in matching faces with emo-tional stories (Matsumoto & Ekman, 2004;Rosenberg & Ekman, 1995). These findingssuggest strongly that chimpanzees, like hu-mans, respond to different faces as categories.

There have been various methodological cri-tiques of the studies of face recognition (e.g.,Russell, 1991a, 1994; Russell, Bachorowski, &Fernandez-Dols, 2003). This body of criticismhas been addressed in a number of studies(e.g., Ekman et al., 1987; Matsumoto, 2005;Matsumoto & Ekman, 1989, 2004; Yrizarry,Matsumoto, & Wilson-Cohn, 1998; Frank &Stennett, 2001; Rosenberg & Ekman, 1995).These studies demonstrate that the originalfindings of universality in judgments of discreteemotion categories were not artifacts of theforced-choice judgment task and are, in fact,quite robust across different judgment tasksand cultures. The literature we review belowindicates that the recognition of distinct facialexpressions activates specific regions of thebrain.

Brain Activation Produced by PerceivingFacial Expressions

Electroencephalographic and brain imagingstudies demonstrate brain specificity in judg-ments of discrete emotions. The perception offearful faces activates regions in the leftamygdala (Breiter et al., 1996; Phillips et al.,1997; Whalen et al., 2004), even when the pre-sentation of a fearful face is masked by the pre-sentation of an immediately ensuing neutral ex-pression (Whalen et al., 1998), or by other,consciously perceived expressions in the face orvoice (de Gelder, Morris, & Dolan, 2005). Thisneural network is modulated by the rightprefrontal cortex (Hariri, Mattay, Tessitore,Fera, & Weinberger, 2003) and by the peptideoxytocin (Kirsch et al., 2005). The perceptionof sad faces activates the left amygdala andright temporal lobe (Blair, Morris, Frith,Perrett, & Dolan, 1998). The perception of an-gry faces activates the right orbito-frontal cor-tex and cingulate cortex (Blair et al., 1998;Sprengelmeyer et al., 1996). The perception ofdisgusted faces activates the basal ganglia, an-terior insula, and frontal lobes (Phillips et al.,1997; Sprengelmeyer et al., 1996). Duchennesmiles activate the left side of the lateral fron-tal, midfrontal, anterior temporal, and centralanterior scalp regions (Davidson et al., 1990;


Ekman & Davidson, 1993). Some evidencesuggests that the brain areas involved whenemotion is elicited are the same areas involvedin judging emotions in others (Calder, 2003;Calder, Keane, Manes, Antoun, & Young,2000).

Disease and lesion studies indicate that theperception of different emotions is associatedwith different brain regions. Bilateral lesions tothe amygdala impair the ability to recognizefearful faces and vocalizations, but not the abil-ity to recognize facial expressions of sadness,disgust, or happiness (Adolphs, Tranel,Damasio, & Damasio, 1994, 1995; Adolphs etal., 1999; Broks et al., 1998; Calder, Young, &Perrett, 1996; Sprengelmeyer et al., 1996;Young, Hellawell, Van de Wal, & Johnson,1996). Individuals suffering from Huntington’sdisease, which affects the basal ganglia, areunable to recognize disgusted expressionsaccurately but are accurate in judging facial ex-pressions of other negative emotions (Sprengel-meyer et al., 1996). Even carriers of Hunting-ton’s disease are unable to recognize facialexpressions of disgust (Gray, Young, Barker,Curtis, & Gibson, 1997).

The last 10 years, then, have seen the emer-gence of two robust literatures convergingupon the conclusion that evolution has shapedthe capacity, universal to humans, to reliablyrecognize distinct facial expressions of emo-tion. The first is the new wave of studies ofemotion recognition, which have dealt withconfounds and problems in interpretation ofthe influential Ekman and Izard studies, andfound consistently that the recognition of facialemotion expressions is universal. A second lit-erature has shown that different patterns of re-gional activation in the brain occur when indi-viduals perceive distinct facial expressions,raising the possibility that we humans haveevolved distinct emotion perception systems orcircuits.

Important Social Functionsof Facial Emotional Expressions

Central to an evolutionist analysis of emotionis the premise that the emotions evolved to helpsolve social problems (e.g., Ekman, 1992;Keltner, Haidt, & Shiota, in press; Tooby &Cosmides, 1992). Expressions are central tothese processes in three ways (Keltner, 2003).First, they provide information about theexpressor’s emotions, intentions, relationship

with the target, and relationship with the envi-ronment. Second, they evoke responses,particularly emotions, from others. Third, theyprovide incentives for desired social behavior.The research reviewed above supports the firstcontention. Research reviewed in this sectionsupports the second two.

Facilitating Specific Behaviors in Perceivers

Because facial expressions of emotion are uni-versal social signals, they contain meaning notonly about the expressor’s intent and subse-quent behavior, but also about what theperceiver is likely to do. Marsh, Ambady, andKleck (2005) showed observers fearful and an-gry faces, and asked them either to push or topull a lever when they saw them. Anger facili-tated avoidance-related behaviors, whereasfear facilitated approach-related behaviors.Winkielman, Berridge, and Wilbarger (2005)found that subliminal presentation of smilesproduced increases in how much of a beveragepeople poured and consumed, and how muchthey were willing to pay for it; presentation ofangry faces decreased these behaviors. Also,emotional displays evoke specific, complemen-tary emotional responses from observers. Forexample, anger has been found to evoke fear(Dimberg & Öhman, 1996; Esteves et al.,1994), whereas distress evokes sympathy andaid (Eisenberg et al., 1989).

Signaling the Nature of Interpersonal Relationships

Some of the more important and provocativeset of findings in the area of facial emotionalexpressions and interpersonal relationshipscome from Gottman and Levenson’s (Gottman& Levenson, 1992; Gottman, Levenson, &Woodin, 2001) studies involving married cou-ples. In their research, married couples visitedtheir laboratory after the spouses had not seeneach other for 24 hours, and then the spousesengaged in intimate conversations about dailyevents, issues of conflict, and so forth. Discreteexpressions of contempt, especially by the men,and disgust, especially by the women, predictedlater marital dissatisfaction and even divorce.

Regulating Social Interaction

Facial expressions of emotion, and other facialbehaviors, are important regulators of socialinteraction. In the developmental literature,


this concept has been investigated under the ru-bric of “social referencing” (Klinnert et al.,1983)—that is, the process whereby infantsseek out emotional information from others tointerpret ambiguous objects and events, andthen use that information to act (see alsoHertenstein & Campos, 2004).

The Importance of Judging Emotions Accurately

Because facial expressions are reliable markersof emotion, and because they serve importantsocial functions, the ability to judge them accu-rately may be linked to important intra- and in-terpersonal processes. In exploring this thesis,Matsumoto and his colleagues addressed thisproblem in the creation of the Japanese andCaucasian Brief Affect Recognition Test(JACBART), which was based on Ekman andFriesen’s (1969) observation of microex-pressions (described earlier). In the JACBART,seven universal facial expressions are presentedvery briefly (for 0.20 second), embedded with-in a 1-second presentation of the sameexpressor’s neutral face.4 Matsumoto et al.(2000) demonstrated the internal and temporalreliability of the emotion recognition scoresproduced; sufficient item discrimination andrange; convergent validity among emotions;and reliability across response alternatives.Construct validity was established by correla-tions between emotion recognition accuracyscores and the personality constructs Opennessand Conscientiousness in multiple measures ofpersonality. Moreover, emotion recognition ac-curacy scores were independent of visual acuityin judging general facial stimuli presented athigh speeds. The findings on Openness havebeen replicated by others (Terracciano, Merritt,Zonderman, & Evans, 2003).

Emotion recognition ability has been theo-retically linked to the concept of emotional in-telligence (Mayer, Salovey, Caruso, &Sitarenios, 2001; Salovey & Mayer, 1990). Infact, in a study using behavioral measures totest the predictive validity of the JACBART,emotion recognition accuracy was correlatedwith Problem Solving, Goal Setting, and totalEffectiveness as measured by an In-Basket task(Matsumoto, LeRoux, Bernhard, & Gray,2004). And Yoo, Matsumoto, and LeRoux(2006) used the JACBART to demonstrate thatemotion recognition abilities of internationalstudents measured at the beginning of theschool year were correlated with various ad-

justment indices at the beginning of the year(anxiety, homesickness, culture shock), as wellas at the end of the year (anxiety, contentment).Recognition ability was also correlated withstudents’ end-of-year grade point averages andwith ratings of their participation in the behav-ioral task obtained in the laboratory session.

The hypothesis that microexpressions are re-lated to deception, outlined above, suggeststhat the ability to recognize emotions frommicroexpressions should be related to the abil-ity to detect lies. In fact, using an early micro-expression test (the Brief Affect RecognitionTest), Ekman and Friesen (1974) did indeedfind that the ability to recognize microex-pressions was significantly correlated with liedetection accuracy. Ekman and O’Sullivan(1991) replicated this finding. Most recently,Ekman (2003) has developed the Micro Ex-pression Training Tool (METT), which allowsindividuals to test themselves on their ability toidentify correctly basic expressions of anger,fear, sadness, disgust, happiness, contempt, andsurprise. Training with this technique signifi-cantly increases the accurate recognition of de-ceptive items in which microexpressions occur(Ekman & Frank, 2005), and these results havebeen replicated across cultures (Frank, 2007).Recent research has indicated that schizo-phrenic individuals trained with the METT im-proved to a level that was not significantly dif-ferent from the performance of pretrainedcontrols (Russell, Chu, & Phillips, 2006).

In concluding our empirical review, we sug-gest that the studies brought together here, forthe first time, strongly support the five claimsmade from an evolutionist approach to facialexpression. The data suggest that facial expres-sions of emotion, as originally described byDarwin, are universally aroused in specific situ-ations (Claim 1); linked to subjective experi-ence (Claim 2); part of a coherent, multisystemresponse package (Claim 3); judged across dif-ferent cultures in similar fashion (Claim 4); andassociated with social functions (Claim 5).

A RESEARCH AGENDAFOR THE FUTURE

Evolutionist approaches to facial expression,and the methods and findings inspired by thisperspective, have been integral to the develop-ment of the field of emotion. This perspective


has inspired coherent lines of empirical inquiryon the universality, coherence, recognition, andsocial functions of facial expression. All remainfruitful areas of exploration. We close thischapter by referring briefly to questions wherethe need for empirical research is great, andwhere opportunities for discovery are clear.

When Do Discrete Facial Expressionsof Emotion Appear in Development?

There is continuing controversy about whethernewborns signal discrete emotions in the face.The biologically based programs that lead tothe regular occurrence of facial expressionsmay unfold later according to maturationalor developmental milestones, especially mile-stones in cognitive abilities. Few longitudinaldata in early infancy, however, shed light on thedevelopmental timetable for discrete facial ex-pressions, and future research is needed in thisarea.

What Other Emotions May Be Expressedin the Face?

Anger, disgust, fear, happiness, sadness, andsurprise are the only emotions for which evi-dence has been found to date demonstratingtheir universal signal characteristics, uniquephysiological signatures, universal appraisalmechanisms, and presence in other primates.But they may not be the only emotions withsuch qualities. Unfortunately, evidence for oth-ers is suggestive but incomplete. For instance,preliminary data suggest that candidates in-clude displays of contempt (Ekman & Friesen,1986; Ekman & Heider, 1988); embarrassment(Keltner & Buswell, 1997) and pride (Tracy &Robins, 2004) and positive emotions such asawe, desire, and love (Gonzaga, Keltner, &Londahl, 2001; Gonzaga et al., 2006; Shiota,Campos, & Keltner, 2003). Future studies willneed to address the possibility that other emo-tions are also displayed in the face, as well asvariants of the same emotion.

In What Other Channels Can EmotionsBe Expressed?

It is certain that emotions are expressed inchannels other than the face and voice, the twomost studied to date. Both the well-studiedemotions, such as anger and disgust, and theless investigated emotions, such as embarrass-

ment and love, are certain to involve expressivebehavior in other channels, such as gaze, headposition, posture, touch, or proximity. Recentresearch suggests such possibilities. New re-search in this area includes studies on embar-rassment (Keltner, 1995), shame (Halisch &Halisch, 1980; Lewis, Alessandri, & Sullivan,1992; Stipek & Gralinski, 1991), pride (Tracy& Matsumoto, 2005; Tracy & Robins, 2004),romantic love (Gonzaga et al., 2001), and sym-pathy (Eisenberg et al., 1989).

These new findings suggest that more emo-tions than previously thought can be communi-cated in brief expressive behaviors. These find-ings raise important questions to pursue. Towhat extent do the different channels—facialexpression, gaze, posture, touch—covary in co-herent fashion? To what extent to these chan-nels of emotion communication convey distinctemotion-relevant information (e.g., about feltexperience or behavioral intention), and towhat extent is the information they convey re-dundant?

What Factors Moderate Facial Expressionsand Their Linkageswith Other Emotion Responses?

Regulation of facial expressions in humans is acomplex neuropsychological phenomenon thatcan occur outside of conscious awareness(Matsumoto & Lee, 1993). Expression regula-tion via display rules provides an opportunityto understand how the linkage between expres-sion and experience can be systematically de-coupled. The coherence between emotion andexpression is moderated by context, and con-texts that require the modification of expres-sion may result in decoupling of the linkage be-tween experience and display. Research on thistopic, however, is still in its infancy.

What Factors Moderate the Productionof Facial Prototypesof the Universal Emotions?

The full-face, high-intensity facial prototypesof the basic emotions do not always occurwhen emotion is aroused. Facial expressionprototypes, we suggest, are more likely to beseen when strong emotions are aroused and thecontext allows for their expression, suchas when married couples fight (Gottman &Levenson, 1992), or athletes win or lose amedal at the Olympic Games (Matsumoto &


Willingham, 2006). When the full-face proto-types are not expressed, partial facial expres-sions may occur. In addition to the influence ofcontext and emotion intensity, partial expres-sions occur in part because of the different neu-ral wiring of the facial muscles (Matsumoto &Lee, 1993). Lower-face muscles are representedmore fully in the motor cortex than those of theupper face, allowing for more control of thelower face. Moreover, the number of bilateralversus contralateral fibers to the facial musclesdiffers depending on region, with the lowerface being primarily contralateral and bilateralfibers increasing in the upper face. And volun-tary and involuntary expressions are under thecontrol of different neural tracts. These factorsstrongly suggest that the lower-face compo-nents of facial expressions are more likely to bemoderated by various factors, including cultureand individual differences.

How Many Variants of Each FacialEmotional Expression Exist?

The original studies documenting universalityof facial expressions, as well as subsequent re-search, indicate that many of the emotionshave variants in expressions. With some excep-tions (Matsumoto, 1989; Rozin, Lowery, &Ebert, 1994), research has not examined thenature of these variants. Different variants maybe associated with different intentions; for in-stance, anger may lead to a joke, walking away,or physical aggression, and it is possible thatthese different behavioral consequences are sig-naled by different variants of anger expres-sions. Similar differences may exist for theother emotions as well, and future studies willneed to explore these possibilities.

The only emotion that appears to have a sin-gle facial muscle action is that of enjoyment,which is signaled by the Duchenne smile. Thereare, however, many kinds of enjoyment(Ekman, 2003), and it is likely that these differ-ent kinds of enjoyment are signaled throughdifferent expressive channels, such as posturalbehavior and gaze activity. This research is justgetting underway.

CONCLUSION

An evolutionist approach to facial expressionhas generated vibrant empirical literatures,which allow for the following conclusions:

1. Some facial expressions are universal, reli-able markers of discrete emotions whenemotions are aroused and there is no reasonto modify or manage the expressions.

2. Discrete facial expressions generally corre-spond to discrete underlying subjective ex-periences.

3. Discrete facial expressions are part of a co-herent package of emotion responses thatincludes appraisals, physiological reactions,other nonverbal behaviors, and subsequentactions; they are also reliable signs of indi-vidual differences and of mental and physi-cal health.

4. Discrete facial expressions are judged reli-ably in different cultures.

5. Discrete facial expressions serve many in-terpersonal and social regulatory functions.

Numerous questions of theoretical significanceremain, and await a next wave of empiricalstudies.

NOTES

1. To wit, Darwin penciled in the margin of Bell’s book,“He never looked at a monkey” (Darwin, 1872/1998).

2. The importance of specifying the exact muscles un-derlying facial expressions can be understood by rec-ognizing the difference in ethologists’ verbal descrip-tions of fear, play, and threat faces of apes. Anonfearful threat involves a tense mouth (AU 23)with no teeth showing. When fear is involved, as iscommon when subordinate animals are threateningthose superior to them in the social hierarchy, the lipsare stretched back (AU 20), thereby showing teeth.This seeming smile is not a smile at all, but a fear gri-mace. By contrast, in the commonly reported “playface” of apes, the corners of the lips go up (AU 12) asthey do in human smiles. The difference between AU20 (a tense lateral stretch) and AU 12 (a relaxed up-ward pull) is crucial.

3. In analyzing the facial musculature of human infants,Oster (2004) found that adjustments had to be madeto the human adult FACS to account for the greaterdegree of subcutaneous fat in infants. Similarly, theWaller et al. (2006) chimpanzee FACS has just beencompleted, and several adjustments had to be madeto this as well. With this tool, future researchers willbe able to ascertain for sure which muscles areinnervated in which types of expressions in chimpan-zees.

4. Many other tests have been developed over the yearsto assess related concepts, such as the Profile of Non-verbal Sensitivity (Rosenthal, Hall, DeMatteo, Rog-ers, & Archer, 1979), the Social Skills Inventory


(Riggio, 1986), the Social Interpretations Test (Ar-cher & Akert, 1977), and the Diagnostic Analysis ofNonverbal Accuracy Scale (Nowicki & Duke, 1994).These, however, do not focus on the recognition ofdiscrete emotional states. Other tests are moreemotion-focused, such as the Communication andReception of Affect Test (Buck, 1976), the Test ofEmotional Styles (Allen & Hamsher, 1974), the Un-derstanding our Feelings Test (Elmore, 1985), theFeldstein Affect Judgment Test (Wolitzky, 1973), theAffect Communication Test (Friedman, Prince,Riggio, & DiMatteo, 1980), the Contextual and Af-fective Sensitivity Test (Trimboli & Walker, 1993),and the Perception of Affect Task (Lane, Sechrest, &Reidel, 1996). These, however, suffer from question-able validity of the expressions used to portray emo-tion, the inability to produce scores on discreteemotions, and the lack of balance in encoder charac-teristics.

REFERENCES

Abe, J. A. A., & Izard, C. E. (1999). A longitudinalstudy of emotion expression and personality rela-tions in early development. Journal of Personalityand Social Psychology, 77(3), 566–577.

Adolphs, R., Tranel, D., Damasio, H., & Damasio, A.R. (1994). Impaired recognition of emotion in facialexpression following bilateral damage to the humanamygdala. Nature, 372, 669–672.

Adolphs, R., Tranel, D., Damasio, H., & Damasio, A.R. (1995). Fear and the human amygdala. Journal ofNeuroscience, 15, 5879–5891.

Adolphs, R., Tranel, D., Hamann, S., Young, A. W.,Calder, A. J., Phelps, E. A., et al. (1999). Recognitionof facial emotion in individuals with bilateralamygdala damage. Neuropsychologia, 37, 1111–1117.

Allen, J. R., & Hamsher, J. H. (1974). The developmentand validation of a test of emotional styles. Journalof Consulting and Clinical Psychology, 42, 663–668.

Archer, D., & Akert, R. (1977). Words and everythingelse: Verbal and nonverbal cues in social interaction.Journal of Personality and Social Psychology, 35,443–449.

Berenbaum, H., & Oltmanns, T. (1992). Emotional ex-perience and expression in schizophrenia and depres-sion. Journal of Abnormal Psychology, 101, 37–44.

Blair, R. J. R., Morris, J. S., Frith, C. D., Perrett, D. I., &Dolan, R. J. (1998). Dissociable neural responses tofacial expressions of sadness and anger. Brain,122(5), 883–893.

Bonanno, G. A., & Keltner, D. (1997). Facial expres-sions of emotion and the course of conjugal bereave-ment. Journal of Abnormal Psychology, 106, 126–137.

Bonanno, G. A., & Keltner, D. (2004). The coherence ofemotion systems: Comparing “on-line” measures of

appraisal and facial expressions, and self-report.Cognition and Emotion, 18(3), 431–444.

Bonanno, G. A., Keltner, D., Noll, J., G., Putnam, F. W.,Trickett, P. K., LeJeune, J., et al. (2002). When theface reveals what words do not: Facial expressions ofemotion, smiling, and the willingness to disclosechildhood sexual abuse. Journal of Personality andSocial Psychology, 83(1), 94–110.

Borod, J. C., Koff, E., Lorch, M. P., & Nicholas, M.(1986). The expression and perception of facial emo-tion in brain-damaged patients. Neuropsychologia,24, 169–180.

Breiter, H. C., Etcoff, N. L., Whalen, P. J., Kennedy, W.A., Rauch, S. L., Buckner, R. L., et al. (1996). Re-sponse and habituation of the human amygdala dur-ing visual processing of facial expression. Neuron,17, 875–887.

Broks, P., Young, A. W., Maratos, E. J., Coffey, P. J.,Calder, A. J., Isaac, C. I., et al. (1998). Face process-ing impairments after enchepalitis: Amygdala dam-age and recognition of fear. Neuropsychologia, 36,59–70.

Brown, A., & Schwartz, G. E. (1980). Relationship be-tween facial electromyography and subjective experi-ence during affective imagery. Biological Psychology,11, 49–62.

Buck, R. W. (1976). A test of nonverbal receiving abil-ity: Preliminary studies. Human Communication Re-search, 2, 162–171.

Buck, R. W. (1977). Nonverbal communication of affectin preschool children: Relationships with personalityand skin conductance. Journal of Personality and So-cial Psychology, 35, 225–236.

Cacioppo, J. T., Martzke, J. S., Petty, R. E., & Tassinary,L. G. (1988). Specific forms of facial emg responseindex emotions during an interview: From Darwin tothe continuous flow hypothesis of affect-laden infor-mation processing. Journal of Personality and SocialPsychology, 54, 592–604.

Calder, A. J. (2003). Disgust discussed. Annals of Neu-rology, 53(4), 427–428.

Calder, A. J., Keane, J., Manes, F., Antoun, N., &Young, A. W. (2000). Impaired recognition and expe-rience of disgust following brain injury. Nature Neu-roscience, 3(11), 1077–1078.

Calder, A. J., Young, A. W., & Perrett, D. I. (1996). Cat-egorical perception of morphed facial expressions.Visual Cognition, 3(2), 81–117.

Camras, L. A., Oster, H., Campos, J., Miyake, K., &Bradshaw, D. (1992). Japanese and American in-fants’ responses to arm restraint. Developmental Psy-chology, 28, 578–583.

Camras, L. A., Ribordy, S., Hill, J., Martino, S.,Spaccarelli, S., & Stefani, R. (1988). Recognition andposing of emotional expressions by abused childrenand their mothers. Developmental Psychology, 24(6),776–781.

Capps, L., Yirmiya, N., & Sigman, M. (1992). Under-standing of simple and complex emotions in non-


retarded children with autism. Journal of Child Psy-chology and Psychiatry, 33, 1169–1182.

Chesney, M. A., Ekman, P., Friesen, W. V., & Black, G.W. (1990). Type A behavior pattern: Facial behaviorand speech components. Psychosomatic Medicine,52(3), 307–319.

Chevalier-Skolnikoff, S. (1973). Facial expression ofemotion in nonhuman primates. In P. Ekman (Ed.),Darwin and facial expression (pp. 11–89). NewYork: Academic Press.

Craig, K., Hyde, S., & Patrick, C. (1991). Genuine, sup-pressed, and fake facial behavior during exacerbationof chronic low back pain. Pain, 46, 161–171.

Darwin, C. (1998). The expression of emotion in manand animals. New York: Oxford University Press.(Original work published 1872)

Davidson, R. J. (2003). Parsing the subcomponents ofemotion and disorders of emotion: Perspectives fromaffective neuroscience. In R. J. Davidson, K. R.Scherer, & H. H. Goldsmith (Eds.), Handbook of af-fective sciences (pp. 8–24). New York: Oxford Uni-versity Press.

de Gelder, B., Morris, J. S., & Dolan, R. J. (2005). Un-conscious fear influences emotional awareness offaces and voices. Proceedings of the National Acad-emy of Sciences USA, 102(51), 18682–18687.

De Waal, F. B. M. (2002). Apes from Venus: Bonobosand human social evolution. In F. B. M. De Waal(Ed.), Tree of origin: What primate behavior can tellus about human social evolution (pp. 39–68). Cam-bridge, MA: Harvard University Press.

Dimberg, U., & Öhman, A. (1996). Behold the wrath:Psychophysiological responses to facial stimuli. Mo-tivation and Emotion, 20(2), 149–182.

Duchenne de Boulogne, G. B. (1990). The mechanism ofhuman facial expression. New York: Cambridge Uni-versity Press. (Original work published 1862)

Eibl-Ebesfeldt, I. (1989). Human ethology. New York:Aldine de Gruyter Press.

Eisenberg, N., Fabes, R. A., Miller, P. A., Fultz, J., Shell,R., Mathy, R. M., et al. (1989). Relation of sympathyand distress to prosocial behavior: A multimethodstudy. Journal of Personality and Social Psychology,57, 55–66.

Ekman, P. (1972). Universal and cultural differences infacial expression of emotion. In J. R. Cole (Ed.),Nebraska symposium on motivation, 1971 (Vol.19, pp. 207–283). Lincoln: University of NebraskaPress.

Ekman, P. (Ed.). (1973). Darwin and facial expression.New York: Academic Press.

Ekman, P. (1989). Why lies fail and what behaviors be-tray a lie. In J. C. Yuille (Ed.), NATO Advanced Sci-ence Institutes Series. Series D: Behavioural and so-cial sciences. Vol. 47. Credibility assessment (pp. 71–81): Dordrecht, The Netherlands: Kluwer.

Ekman, P. (1992a). Are there basic emotions? Psycho-logical Review, 99(3), 550–553.

Ekman, P. (1992b). Facial expressions of emotion: New

findings, new questions. Psychological Science, 3(1),34–38.

Ekman, P. (1993). Facial expression and emotion.American Psychologist, 48(4), 384–392.

Ekman, P. (2003). Emotions revealed. New York: TimesBooks.

Ekman, P., & Davidson, R. J. (1993). Voluntary smilingchanges regional brain activity. Psychological Sci-ence, 4(5), 342–345.

Ekman, P., Davidson, R. J., & Friesen, W. V. (1990).The Duchenne smile: Emotional expression andbrain physiology: II. Journal of Personality and So-cial Psychology, 58(2), 342–353.

Ekman, P., & Frank, M. G. (2005). The ability to detectdeception improves with microexpression recogni-tion training. Manuscript submitted for publication.

Ekman, P., & Friesen, W. (1969). The repertoire of non-verbal behavior: Categories, origins, usage, and cod-ing. Semiotica, 1, 49–98.

Ekman, P., & Friesen, W. (1971). Constants across cul-ture in the face and emotion. Journal of Personalityand Social Psychology, 17, 124–129.

Ekman, P., Friesen, W., & Ancoli, S. (1980). Facial signsof emotional experience. Journal of Personality andSocial Psychology, 39, 1125–1134.

Ekman, P., & Friesen, W. V. (1974). Nonverbal behaviorand psychopathology. In R. J. Friedman & M. Katz(Eds.), The psychology of depression: Contemporarytheory and research (pp. 3–31). Washington, DC:Winston.

Ekman, P., & Friesen, W. V. (1978). Facial ActionCoding System: Investigator’s guide. Palo Alto, CA:Consulting Psychologists Press.

Ekman, P., & Friesen, W. V. (1982). Felt, false, and mis-erable smiles. Journal of Nonverbal Behavior, 6(4),238–258.

Ekman, P., & Friesen, W. V. (1986). A new pan-culturalfacial expression of emotion. Motivation and Emo-tion, 10(2), 159–168.

Ekman, P., Friesen, W. V., & O’Sullivan, M. (1988).Smiles when lying. Journal of Personality and SocialPsychology, 54(3), 414–420.

Ekman, P., Friesen, W. V., O’Sullivan, M., Chan, A.,Diacoyanni-Tarlatzis, I., Heider, K., et al. (1987).Universals and cultural differences in the judgmentsof facial expressions of emotion. Journal of Personal-ity and Social Psychology, 53(4), 712–717.

Ekman, P., & Heider, K. G. (1988). The universality of acontempt expression: A replication. Motivation andEmotion, 12(3), 303–308.

Ekman, P., Levenson, R. W., & Friesen, W. V. (1983).Autonomic nervous system activity distinguishesamong emotions. Science, 221(4616), 1208–1210.

Ekman, P., Liebert, R. M., Friesen, W., Harrison, R.,Zlatchin, C., Malmstrom, E. J., et al. (1972). Facialexpressions of emotion while watching televised vio-lence as predictors of subsequent aggression. In G. A.Comstock, E. A. Rubinstein, & J. P. Murray (Eds.),Television and social behavior: A technical report to


the Surgeon General’s Scientific Advisory Committeeon Television and Social Behavior. Vol. 1. Television’seffects: Further explorations (pp. 22–58). Washing-ton, DC: U.S. Government Printing Office.

Ekman, P., Matsumoto, D., & Friesen, W. (1997). Facialexpression in affective disorders. In P. Ekman & E. L.Rosenberg (Eds.), What the face reveals: Basic andapplied studies of spontaneous expression using theFacial Action Coding System (FACS) (pp. 331–341).New York: Oxford University Press.

Ekman, P., & O’Sullivan, M. (1991). Who can catch aliar? American Psychologist, 46(9), 913–920.

Ekman, P., & Rosenberg, E. L. (Eds.). (2005). What theface reveals: Basic and applied studies of spontane-ous expression using the Facial Action Coding Sys-tem (FACS) (2nd ed.). New York: Oxford UniversityPress.

Ekman, P., Sorenson, E. R., & Friesen, W. V. (1969).Pancultural elements in facial displays of emotion.Science, 164(3875), 86–88.

Elfenbein, H. A., & Ambady, N. (2002b). On the uni-versality and cultural specificity of emotion recogni-tion: A meta-analysis. Psychological Bulletin, 128(2),205–235.

Ellgring, H. (1986). Nonverbal expression of psycho-logical states in psychiatric patients. European Ar-chives of Psychiatry and Neurological Sciences, 236,31–34.

Elmore, B. C. (1985). Emotionally handicapped com-prehension of nonverbal communication. Journal ofHolistic Medicine, 7, 194–201.

Esteves, F., Dimberg, U., & Öhman, A. (1994). Auto-matically elicited fear: Conditioned skin conductanceresponses to masked facial expressions. Cognitionand Emotion, 8(5), 393–413.

Foisy, M.-L., Philippot, P., Verbanck, P., Pelc, I., VanDer Straten, G., & Kornreich, C. (2005). Emotionalfacial expression decoding impairment in persons de-pendent on multiple substances: Impact of a historyof alcohol dependence. Journal of Studies on Alco-hol, 66, 673–681.

Frank, M. G. (2007). Decoding deception and emotionby Americans and Australians. Manuscript in prepa-ration.

Frank, M. G., & Ekman, P. (1997). The ability to detectdeceit generalizes across different types of high-stakelies. Journal of Personality and Social Psychology,72, 1429–1439.

Frank, M. G., Ekman, P., & Friesen, W. V. (1993).Behavioral markers and recognizability of the smileof enjoyment. Journal of Personality and Social Psy-chology, 64(1), 83–93.

Frank, M. G., & Stennett, J. (2001). The forced-choiceparadigm and the perception of facial expressions ofemotion. Journal of Personality and Social Psychol-ogy, 80(1), 75–85.

Frank, R. H. (1988). Passions without reason. NewYork: Norton.

Fridlund, A. (1994). Human facial expression: An evo-lutionary view. San Diego: Academic Press.

Friedman, H. S., Prince, L. M., Riggio, R. E., &DiMatteo, M. R. (1980). Understanding and assess-ing nonverbal expressiveness: The Affective Commu-nication Test. Journal of Personality and Social Psy-chology, 39(2), 333–351.

Geen, T. (1992). Facial expressions in socially isolatednonhuman primates: Open and closed programs forexpressive behavior. Journal of Research in Personal-ity, 26, 273–280.

Gonzaga, G. C., Keltner, D., & Londahl, E. A. (2001).Love and the commitment problem in romantic rela-tionships and friendship. Journal of Personality andSocial Psychology, 81(2), 247–262.

Gonzaga, G. C., Turner, R. A., Keltner, D., Campos,B., & Altemus, M. (2006). Romantic love and sex-ual desire in close relationships. Emotion, 6(2),163–179.

Gosselin, P., Kirouac, G., & Dore, F. (1995). Compo-nents and recognition of facial expression in the com-munication of emotion by actors. Journal of Person-ality and Social Psychology, 68, 83–96.

Gottman, J. M., & Levenson, R. W. (1992). Maritalprocesses predictive of later dissolution: Behavior,physiology, and health. Journal of Personality andSocial Psychology, 63(2), 221–223.

Gottman, J. M., Levenson, R. W., & Woodin, E. (2001).Facial expressions during marital conflict. Journal ofFamily Communication, 1, 37–57.

Gray, H., & Goss, C. M. (1966). Anatomy of the humanbody (28th ed.). Philadelphia: Lea & Febiger.

Gray, J. M., Young, A. W., Barker, W. A., Curtis, A., &Gibson, D. (1997). Impaired recognition of disgust inhuntington’s disease gene carriers. Brain, 120, 2029–2038.

Gross, J. J., & Levenson, R. W. (1993). Emotional sup-pression: Physiology, self-report, and expressive be-havior. Journal of Personality and Social Psychology,64(6), 970–986.

Halisch, C., & Halisch, F. (1980). Cognitive assump-tions of young children’s self-evaluation reactions af-ter success and failure. Entwicklungspsychologie undPädagogische Psychologie, 12(3), 193–212.

Hariri, A. R., Mattay, V. S., Tessitore, A., Fera, F., &Weinberger, D. R. (2003). Neocortical modulation ofthe amygdala response to fearful stimuli. BiologicalPsychiatry, 53, 494–501.

Harker, L. A., & Keltner, D. (2001). Expressions of pos-itive emotion in women’s college yearbook picturesand their relationship to personality and life out-comes across adulthood. Journal of Personality andSocial Psychology, 80, 112–124.

Harris, C. R., & Alvarado, N. (2005). Facial expres-sions, smile types, and self-report during humous,tickle, and pain. Cognition and Emotion, 19(5), 655–669.

Hauser, M. (1993). Right hemisphere dominance for theproduction of facial expression in monkeys. Science,261, 475–477.

Heerey, E. A., Keltner, D., & Capps, L. M. (2003).Making sense of self-conscious emotion: Linking the-


ory of mind and emotion in children with autism.Emotion, 3(4), 394–400.

Heller, M., & Haynal, V. (1994). Depression and suicidefaces. Cahiers Psychiatriques Genevois, 16, 107–117.

Hertenstein, M. J., & Campos, J. J. (2004). The reten-tion effects of an adult’s emotional displays on infantbehavior. Child Development, 75(2), 595–613.

Hess, U., Banse, R., & Kappas, A. (1995). The intensityof facial expression is determined by underlying af-fective states and social situations. Journal of Person-ality and Social Psychology, 69, 280–288.

Hobson, P. R. (1986). The autistic children’s appraisalof expressions of emotion: A further study. Journal ofChild Psychology and Psychiatry, 27, 671–680.

Horstmann, G. (2003). What do facial expressions con-vey?: Feeling states, behavioral intentions, or actionrequests? Emotion, 3(2), 150–166.

Izard, C. E. (1971). The face of emotion. New York:Appleton-Century-Crofts.

Keltner, D. (1995). The signs of appeasement: Evidencefor the distinct displays of embarrassment, amuse-ment, and shame. Journal of Personality and SocialPsychology, 68, 441–454.

Keltner, D. (2003). Expression and the course of life:Studies of emotion, personality, and psychopatholo-gy from a social-functional perspective. Annals of theNew York Academy of Sciences, 1000, 222–243.

Keltner, D., & Bonanno, G. A. (1997). A study of laugh-ter and dissociation: The distinct correlates of laugh-ter and smiling during bereavement. Journal of Per-sonality and Social Psychology, 73, 687–702.

Keltner, D., & Buswell, B. N. (1997). Embarrassment:Its distinct form and appeasement functions. Psycho-logical Bulletin, 122(3), 250–270.

Keltner, D., Ellsworth, P. C., & Edwards, K. (1993). Be-yond simple pessimism: Effects of sadness and angeron social perception. Journal of Personality and So-cial Psychology, 64, 740–752.

Keltner, D., Haidt, J., & Shiota, M. N. (2006). Socialfunctionalism and the evolution of emotions. In M.Schaller, J. A. Simpson, & D. T. Kenrick (Eds.), Evo-lution and social psychology. New York: PsychologyPress.

Keltner, D., & Kring, A. M. (1998). Emotion, socialfunction, and psychopathology. Review of GeneralPsychology, 2(3), 320–342.

Keltner, D., Moffitt, T., & Stouthamer-Loeber, M.(1995). Facial expressions of emotion and psycho-pathology in adolescent boys. Journal of AbnormalPsychology, 104, 644–652.

Kirsch, P., Esslinger, C., Chen, Q., Mier, D., Lis, S.,Siddhanti, S., et al. (2005). Oxytocin modulates neu-ral circuitry for social cognition and fear in humans.Journal of Neuroscience, 25(49), 11489–11493.

Klinnert, M. D., Campos, J. J., & Sorce, J. F. (1983).Emotions as behavior regulators: Social referencingin infancy. In R. Plutchik & H. Kellerman (Eds.),Emotion: Theory, research, and experience (pp. 57–86). New York: Academic Press.

Klinnert, M. D., Emde, R. N., Butterfield, P., & Cam-pos, J. (1986). Social referencing: The infant’s use ofemotional signals from a friendly adult with a motherpresent. Developmental Psychology, 22(4), 427–432.

Krebs, J. R., & Dawkins, R. (1984). Animal signals:Mind reading and manipulation. In J. R. Krebs & R.Dawkins (Eds.), Behavioural ecology: An evolu-tionary approach (pp. 380–402). Sunderland, MA:Sinauer Associates.

Lane, R. D., Sechrest, L., & Reidel, R. (1996). Impairedverbal and nonverbal emotion recognition inalexithymia. Psychosomatic Medicine, 58, 203–210.

Lawrence, K., Bernstein, D., Pearson, R., Mandy, W.,Brand, S., Wade, A., et al. (2005). Age, gender andpuberty influence the development of facial emotionrecognition. Manuscript submitted for publication.

Lazarus, R. (1991). Emotion and adaptation. NewYork: Oxford University Press.

Lerner, J. S., Gonzalez, R. M., Dahl, R. E., Haririr, A.R., & Taylor, S. E. (2005). Facial expressions of emo-tion reveal neuroendocrine and cardiovascular stressresponses. Biological Psychiatry, 58, 743–750.

Levenson, R. W. (1994). Human emotions: A functionalview. In P. Ekman & R. J. Davidson (Eds.), The na-ture of emotion: Fundamental questions (pp. 17–42).New York: Oxford University Press.

Levenson, R. W. (2003b). Blood, sweat, and fears: Theautonomic architecture of emotion. Annals of theNew York Academy of Sciences, 1000, 348–366.

Levenson, R. W., Carstensen, L. L., Friesen, W. V., &Ekman, P. (1991). Emotion, physiology, and expres-sion in old age. Psychology and Aging, 6(1), 28–35.

Levenson, R. W., & Ekman, P. (2002). Difficulty doesnot account for emotion-specific heart rate changesin the directed facial action task. Psychophysiology,39(3), 397–405.

Levenson, R. W., Ekman, P., & Friesen, W. V. (1990).Voluntary facial action generates emotion-specificautonomic nervous system activity. Psychophysi-ology, 27(4), 363–384.

Levenson, R. W., Ekman, P., Heider, K., & Friesen, W. V.(1992). Emotion and autonomic nervous system ac-tivity in the Minangkabau of West Sumatra. Journalof Personality and Social Psychology, 62(6), 972–988.

Lewis, M., Alessandri, S. M., & Sullivan, M. W. (1992).Differences in shame and pride as a function of chil-dren’s gender and task difficulty. Child Development,63(3), 630–638.

Liebal, K., Pika, S., & Tomasello, M. (2004). Socialcommunication in siamangs (Symphalangus syndac-tylus): Use of gestures and facial expressions. Pri-mates, 45(1), 41–57.

Marsh, A. A., Ambady, N., & Kleck, R. E. (2005). Theeffects of fear and anger facial expressions onapproach- and avoidance-related behaviors. Emo-tion, 5(1), 119–124.

Matsumoto, D. (1989). Face, culture, and judgments ofanger and fear: Do the eyes have it? Journal of Non-verbal Behavior, 13(3), 171–188.


Matsumoto, D. (2001). Culture and emotion. In D.Matsumoto (Ed.), The handbook of culture and psy-chology (pp. 171–194). New York: Oxford Univer-sity Press.

Matsumoto, D. (2005). Scalar ratings of contempt ex-pressions. Journal of Nonverbal Behavior, 29(2), 91–104.

Matsumoto, D. (2006). Culture and cultural world-views: Do verbal descriptions of culture reflect any-thing other than verbal descriptions of culture? Cul-ture and Psychology, 12(1), 33–62.

Matsumoto, D., Consolacion, T., Yamada, H., Suzuki,R., Franklin, B., Paul, S., et al. (2002). American–Japanese cultural differences in judgments of emo-tional expressions of different intensities. Cognitionand Emotion, 16(6), 721–747.

Matsumoto, D., & Ekman, P. (1989). American–Japanese cultural differences in intensity ratings offacial expressions of emotion. Motivation and Emo-tion, 13(2), 143–157.

Matsumoto, D., & Ekman, P. (2004). The relationshipbetween expressions, labels, and descriptions of con-tempt. Journal of Personality and Social Psychology,87(4), 529–540.

Matsumoto, D., Haan, N., Gary, Y., Theodorou, P., &Cooke-Carney, C. (1986). Preschoolers’ moral ac-tions and emotions in Prisoner’s Dilemma. Develop-mental Psychology, 22(5), 663–670.

Matsumoto, D., & Kupperbusch, C. (2001). Idiocentricand allocentric differences in emotional expressionand experience. Asian Journal of Social Psychology,4, 113–131.

Matsumoto, D., & Lee, M. (1993). Consciousness, voli-tion, and the neuropsychology of facial expressionsof emotion. Consciousness and Cognition: An Inter-national Journal, 2(3), 237–254.

Matsumoto, D., LeRoux, J. A., Bernhard, R., & Gray,H. (2004). Personality and behavioral correlates ofintercultural adjustment potential. InternationalJournal of Intercultural Relations, 28(3–4), 281–309.

Matsumoto, D., LeRoux, J. A., Wilson-Cohn, C.,Raroque, J., Kooken, K., Ekman, P., et al. (2000). Anew test to measure emotion recognition ability:Matsumoto and Ekman’s Japanese and CaucasianBrief Affect Recognition Test (JACBART). Journal ofNonverbal Behavior, 24(3), 179–209.

Matsumoto, D., Nezlek, J., & Koopmann, B. (2007).Evidence for universality in phenomenological emo-tion response system coherence. Emotion, 7(1), 57–67.

Matsumoto, D., & Willingham, B. (2006). The thrill ofvictory and the agony of defeat: Spontaneous expres-sions of medal winners at the 2004 Athens OlympicGames. Journal of Personality and Social Psychology,91(3), 568–581.

Mauss, I. B., Levenson, R. W., McCarter, L., Wilhelm, F.L., & Gross, J. J. (2005). The tie that binds?: Coher-ence among emotion experience, behavior, and physi-ology. Emotion, 5(2), 175–190.

Mauss, I. B., Wilhelm, F. L., & Gross, J. J. (2004). Isthere less to social anxiety than meets the eye?: Emo-tion experience, expression, and bodily responding.Cognition and Emotion, 18(5), 631–662.

Mayer, J. D., Salovey, P., Caruso, D. R., & Sitarenios, G.(2001). Emotional intelligence as a standard intelli-gence. Emotion, 1(3), 232–242.

McGhee, P. E. (1977). Children’s humour: A review ofcurrent research trends. In A. J. Chapman & H. C.Foot (Eds.), It’s a funny thing, humour (pp. 199–209). Oxford: Pergamon Press.

Messinger, D. S., Fogel, A., & Dickson, K. L. (2001). Allsmiles are positive, but some smiles are more positivethan others. Developmental Psychology, 37(5), 642–653.

Mineka, S., & Cook, M. (1993). Mechanisms in-volved in the observational conditioning of fear.Journal of Experimental Psychology: General,122(1), 23–38.

Nowicki, S. J., & Duke, M. P. (1994). Individual differ-ences in the nonverbal communication of affect: TheDiagnostic Analysis of Nonverbal Accuracy Scale.Journal of Nonverbal Behavior, 18, 9–35.

Oster, H. (2004). Baby FACS: Facial Action Coding Sys-tem for infants and young children. Unpublishedmanuscript, New York University.

Ozonoff, S., Pennington, B. F., & Rogers, S. J. (1990).Are there emotion perception deficits in young autis-tic children? Journal of Child Psychology and Psychi-atry, 31, 343–361.

Parr, L. (2003). The discrimination of faces and theiremotional content by chimpanzees (Pan troglodytes).Annals of the New York Academy of Sciences, 100056–78.

Persad, S. M., & Polivy, J. (1993). Differences betweendepressed and non-depressed individuals in the rec-ognition of and response to facial emotional cues.Journal of Abnormal Psychology, 102(3), 358–368.

Phillips, M. L., Young, A. W., Senior, C., Brammer, M.,Andrew, C., Calder, A. J., et al. (1997). A specificneural substrate for perceiving facial expressions ofdisgust. Nature, 389, 495–498.

Pollak, S. D., Cichetti, D., Hornung, K., & Reed, A.(2000). Recognizing emotion in faces: Developmen-tal effects of child abuse and neglect. DevelopmentalPsychology, 36(5), 679–688.

Pollak, S. D., & Sinha, P. (2002). Effects of early experi-ence on children’s recognition of facial displays ofemotion. Developmental Psychology, 38(5), 784–791.

Prkachin, K. (1992). The consistency and facial expres-sions of pain: A comparison across modalities. Pain,51, 297–306.

Redican, W. K. (1982). An evolutionary perspective onhuman facial displays. In P. Ekman (Ed.), Emotion inthe human face (pp. 212–280). New York: Cam-bridge University Press.

Riggio, R. E. (1986). Assessment of social skill. Journalof Personality and Social Psychology, 51, 649–660.

Rosenberg, E. L., & Ekman, P. (1994). Coherence be-


tween expressive and experiential systems in emo-tion. Cognition and Emotion, 8(3), 201–229.

Rosenberg, E. L., & Ekman, P. (1995). Conceptual andmethodological issues in the judgment of facial ex-pressions of emotion. Motivation and Emotion,19(2), 111–138.

Rosenberg, E. L., Ekman, P., Jiang, W., Babyak, M.,Coleman, R. E., Hanson, M., et al. (2001). Linkagesbetween facial expressions of anger and transientmyocardial ischemia in men with coronary heart dis-ease. Emotion, 1(2), 107–115.

Rosenthal, R., Hall, J. A., DeMatteo, M. R., Rogers, P.L., & Archer, D. (1979). Sensitivity to nonverbalcommunication: The PONS test. Baltimore: JohnsHopkins University Press.

Rozin, P., Lowery, L., & Ebert, R. (1994). Varieties ofdisgust faces and the structure of disgust. Journal ofPersonality and Social Psychology, 66(5), 870–881.

Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). TheCAD triad hypothesis: A mapping between threemoral emotions (contempt, anger, disgust) and threemoral codes (community, autonomy, divinity). Jour-nal of Personality and Social Psychology, 75(4), 574–585.

Ruch, W. (1993). Extraversion, alcohol, and enjoyment.Personality and Individual Differences, 16, 89–102.

Ruch, W. (1995). Will the real relationship between fa-cial expression and affective experience stand up?:The case of exhilaration. Cognition and Emotion, 9,33–58.

Russell, J. A. (1991a). Culture and the categorization ofemotions. Psychological Bulletin, 110, 426–450.

Russell, J. A. (1994). Is there universal recognition ofemotion from facial expression?: A review of cross-cultural studies. Psychological Bulletin, 115, 102–141.

Russell, J. A., Bachorowski, J.-A., & Fernandez-Dols, J.M. (2003). Facial and vocal expressions of emotion.Annual Review of Psychology, 54, 329–349.

Russell, T. A., Chu, E., & Phillips, M. L. (2006). A pilotstudy to investigate the effectiveness of emotion rec-ognition mediation in schizophrenia using the micro-expression training tool. British Journal of ClinicalPsychology, 45, 579–583.

Salovey, P., & Mayer, J. D. (1990). Emotional intelli-gence. Imagination, Cognition, and Personality, 9(3),185–211.

Scherer, K. R. (1986). Vocal affect expression: Reviewand a model for future research. Psychological Bulle-tin, 99, 143–165.

Shiota, M. N., Campos, B., & Keltner, D. (2003). Thefaces of positive emotion: Prototype displays of awe,amusement, and pride. In P. Ekman, J. Campos, R. J.Davidson, & F. B. M. De Waal (Eds.), Emotions in-side out: 130 years after Darwin’s The Expression ofEmotions in Man and Animals (pp. 296–299). NewYork: The New York Academy of Sciences.

Smith, M. C. (1995). Facial expression in mild dementiaof the Alzheimer type. Behavioural Neurology, 8,149–156.

Snowdon, C. T. (2003). Expression of emotion in non-human animals. In R. J. Davidson, K. R. Scherer, &H. H. Goldsmith (Eds.), Handbook of affective sci-ences (pp. 457–480). New York: Oxford UniversityPress.

Soltysik, S., & Jelen, P. (2005). In rats, sighs correlatewith relief. Physiology and Behavior, 85(5), 598–602.

Soto, J. A., Levenson, R. W., & Ebling, R. (2005). Cul-tures of moderation and expression: Emotional expe-rience, behavior, and physiology in Chinese Ameri-cans and Mexican Americans. Emotion, 5(2), 154–165.

Sprengelmeyer, R., Young, A. W., Calder, A. J., Karnat,A., Lange, H., Homberg, V., et al. (1996). Loss of dis-gust: Perceptions of faces and emotions in Hunting-ton’s disease. Brain, 119, 1647–1665.

Steimer-Krause, E., Krause, R., & Wagner, G. (1990).Interaction regulations used by schizophrenic andpsychosomatic patients: Studies on facial behavior indyadic interactions. Psychiatry, 53, 209–228.

Stipek, D. J., & Gralinski, J. H. (1991). Gender differ-ences in children’s achievement-related beliefs andemotional responses to success and failure in mathe-matics. Journal of Educational Psychology, 83(3),361–371.

Surcinelli, P., Codispoti, M., Montebarocci, O., Rossi,N., & Baldaro, B. (2004). Facial emotion recognitionin trait anxiety. Anxiety Disorders, 20, 110–117.

Terracciano, A., Merritt, M., Zonderman, A. B., & Ev-ans, M. K. (2003). Personality traits and sex differ-ences in emotion recognition among African Ameri-cans and Caucasians. Annals of the New YorkAcademy of Sciences, 1000, 309–312.

Tiedens, L. Z., Ellsworth, P. C., & Mesquita, B. (2000).Stereotypies about sentiments and status: Emotionalexpectations for high- and low-status group mem-bers. Personality and Social Psychology Bulletin,26(5), 560–574.


Tomkins, S. S. (1963). Affect, imagery, and con-sciousness: Vol. 2. The negative affects. New York:Springer.

Tooby, J., & Cosmides, L. (1992). Psychological foun-dations of culture. In J. Barkow, L. Cosmides, & J.Tooby (Eds.), The adapted mind (pp. 19–136). NewYork: Oxford University Press.

Tracy, J. L., & Matsumoto, D. (2005). Unique expres-sions of pride and shame: Evidence from the 2004Athens Olympic games. Manuscript submitted forpublication.

Tracy, J. L., & Robins, R. W. (2004). Show your pride:Evidence for a discrete emotion expression. Psycho-logical Science, 15(3), 104–197.

Trimboli, A., & Walker, M. (1993). The CAST test ofnonverbal sensitivity. Journal of Language and SocialPsychology, 12, 49–65.

Tsai, J. L., & Levenson, R. W. (1997). Cultural influ-ences of emotional responding: Chinese American


and European American dating couples during inter-personal conflict. Journal of Cross-Cultural Psychol-ogy, 28, 600–625.

Ueno, A., Ueno, Y., & Tomonaga, M. (2004). Facial re-sponses to four basic tastes in newborn rhesus ma-caques (Macaca mulatta) and chimpanzees (Pan trog-lodytes). Behavioural Brain Research, 154(1), 261–271.

Van Hoof, J. A. R. A. M. (1972). A comparative ap-proach to the phylogeny of laughing and smiling. InR. A. Hinde (Ed.), Nonverbal communication(pp. 209–241). Cambridge, UK: Cambridge Univer-sity Press.

Vick, S.-J., Waller, B. M., Parr, L. A., Pasqualini, M. S.,& Bard, K. A. (2007). A cross species comparison offacial morphology and movement in humans andchimpanzees using the Facial Action Coding System(FACS). Journal of Nonverbal Behavior, 31, 1–20.

Waller, B. M., Vick, S.-J., Parr, L., Bard, K. A.,Pasqualini, M. S., Gothard, K. M., et al. (2006). In-tramuscular electrical stimulation of facial muscles inhumans and chimpanzees: Duchenne revisited andextended. Emotion, 6(3), 367–382.

Whalen, P. J., Kagan, J., Cook, R., Davis, F. C., Kim, H.,Polis, S., et al. (2004). Human amygdala responsivityto masked fearful eye whites. Science, 306, 2061.


Winblad, S., Hellstrom, P., Lindberg, C., & Hansen, S.(2006). Facial emotion recognition in myotonic dys-trophy type 1 correlates with CTG repeat expansion.Journal of Neurology, Neurosurgery and Psychiatry,77, 219–223.

Winkielman, P., Berridge, K. C., & Wilbarger, J. L.(2005). Unconscious affective reactions to maskedhappy versus angry faces influence consumptionbehavior and judgments of value. Personality and So-cial Psychology Bulletin, 31(1), 121–135.

Wolitzky, D. (1973). Cognitive control and person per-ception. Perceptual and Motor Skills, 36, 619–623.

Woodzicka, J., & LaFrance, M. (2001). Real versusimagined gender harassment. Journal of Social Is-sues, 57(1), 15–30.

Yoo, S. H., Matsumoto, D., & LeRoux, J. A. (2005).Emotion regulation, emotion recognition, and inter-cultural adjustment. Manuscript submitted for publi-cation.

Yoo, S. H., Matsumoto, D., & LeRoux, J. A. (2006).Emotion regulation, emotion recognition, and inter-cultural adjustment. International Journal of Inter-cultural Relations, 30(3), 345–363.

Young, A. W., Hellawell, D. J., Van de Wal, C., & John-son, M. (1996). Facial expression processing afteramygdalotomy. Neuropsychologia, 34, 31–39.

Yrizarry, N., Matsumoto, D., & Wilson-Cohn, C.(1998). American-Japanese differences inmultiscalar intensity ratings of universal facial ex-pressions of emotion. Motivation and Emotion,22(4), 315–327.


C H A P T E R 1 4

A “Nose” for EmotionEmotional Information and Challenges

in Odors and Semiochemicals

JEANNETTE M. HAVILAND-JONESand PATRICIA J. WILSON

Suppose you could walk into an empty theater,sniff the air, and know that it had staged a hor-ror production. Would the departed actors andaudience leave behind “mood chemicals” in theair? Would audiences coming to a later showfind the show more horrible because of linger-ing chemicals? What does the nose sense?

It is surprising that olfaction has waited solong to be considered an important sensory sys-tem for the field of emotion, since experts inolfaction largely regard the primary responseto odorants as “emotional” (see Van Toller &Dodd, 1988, 1992). This odd incongruity canprobably be attributed to the persistent beliefthat olfaction is a primitive and vestigial system(Schaal & Porter, 1991). As Cain (1978) noted30 years ago, many significant scientists in theWestern world have attended to odors and“semiochemicals” (message-bearing chemi-cals), but seldom for very long. The result is adistinguished historical literature largely un-known to modern psychologists, even thoughsome of the great historical psychologists orprepsychologists—such as Galen, Boring, Ellis,

and Titchener—made contributions. The last30 years have seen astonishing developments inthe research on smell, but much of what is inthe psychological realm remains devoted towhat Cain (1978) has called “industrial en-deavors”—perfumery, malodor control, andfood products.

In this chapter, we focus on the emerging re-search on humans, semiochemicals, and moodsor emotions. This means that we only use thevast literature on the influence of semio-chemicals on insects’ and other animals’ behav-ior to inform our presentation, but do not re-view it. We begin with a brief review of the newdevelopments in the physiology of olfaction,including the olfactory, trigeminal, andvomeronasal systems, because this opens upthe diversity of research related to odors andsemiochemicals. From there, we move to thegrowing body of work revealing the profoundinfluence that pheromones and odors (bothmanufactured and body odors) have beenfound to have on human moods and cognitiveprocesses. Next, we cover research showing theeffects that flowers and floral odors have on

235

human cognition and emotion, and we proposean evolutionary explanation for those effects.This part of the chapter is included because itcontributes to our hypotheses about howsemiochemicals function in human perfor-mance. We then strive to integrate the fullrange of sometimes quirky findings on humans,odors, and emotions into a system demonstrat-ing how humans might make sense of the ubiq-uitous presence of semiochemicals in their dailylives.

NEW DISCOVERIES ANDTHE HUMAN GENOME CHALLENGE

In 2004 the Nobel Prize for medicine wasawarded to Buck and Axel for their discoveryof the genes (accounting for 3% of the humangenome) that express for the receptors detect-ing odors in the nose (see Buck & Axel, 1991).Although 3% may seem small, the olfactorysystem is second in size only to the immune sys-tem. It can detect and discriminate many hun-dreds of different odors at remarkably low con-centrations (less than one part per trillion formercaptan, for example), due to the pattern ofactivation of some 1,000 odor receptors. It ap-pears that olfaction may have greater status inhuman perceptual experience than previous re-searchers ever suspected. Given this new per-spective on olfaction, we have an extraordi-narily limited understanding of the psychologyof olfaction compared with other sensory mo-dalities. Knowing what we now do, it is timefor psychology to expand its training and re-search programs in this area.

Odors can both attract and repel. The histor-ical record reveals an ancient human awarenessof odors and use of odorants to create healthyand attractive personal, physical, and social en-vironments, and to mask the unpleasant andthe foul. In ancient times, the best of Persianscientists worked with aromas both as physi-cians and as social managers. Lavender oil wasused to cleanse a sick room, while fragrancesused at a wedding had complex meanings forthe expected future of the couple. Perfumerswere prized enough in ancient Egypt to be fa-vored with their own god, Nefertum. Pharaohswere buried surrounded by flowers and per-fumes, the better to be accepted into the after-life (Steele, 1992). Cleopatra adorned herselfwith fragrant oils to seduce Marc Antony

(Stoddart, 1990), and lovers created wreaths ofscented flowers to adorn the beloved’s hair(Classen, Howes, & Synnott, 1994). Duringthe Middle Ages, it was believed that foul aircarried the plague, so people fumigated theirhomes with all manner of scents to counteractit. Odor, then, was seen as both potentiallydangerous and potentially beneficial. (The goldhead of a physician’s cane originally containedaromatics to cleanse and heal.) Fragrance as apersonal adornment was thought to stimulatethe mind as well as the emotions (Classen et al.,1994).

In the modern day, we know that environ-mental pollutants and malodors do pose healthrisks (Schiffman & Williams, 2005;Rosenkranz & Cunningham, 2003). Some ofthese risks are present in seemingly innocuoussettings. For example, a recent study found thatexposure to the solvents typically found in nailsalons can lead to a decline in cognitive andneurosensory performance over time (LoSasso,Rapport, Axelrod, & Whitman, 2002).LoSasso and her colleagues note that thesechanges are consistent with those found inworkers exposed to industrial levels of suchsolvents. Rotton (1983) found that exposure touncontrollable unpleasant odors decreased per-formance on complex tasks, increased negativejudgments on evaluative tasks, and decreasedtolerance levels after performing such tasks.Furthermore, the state of Pennsylvania has in-vested heavily in finding solutions to the odorsgenerated by mushroom and pig farming, be-cause of the negative impact on most peopleexposed to these odors (Rouhi, 2002).

What we already know suggests the perva-sive importance of olfactory processes for bothemotion and cognition. Yet, although testingfor vision and hearing is fairly routine (espe-cially among children), and although the firstolfactometer came into use in the late 1880s,olfactory tests are not normally used. This is apotentially important omission, as impairedsense of smell is associated in at least somecases with disease processes; it occurs early inthe onset of both Alzheimer’s and Parkinson’sdisease (Mesholam, Moberg, Mahr, & Doty,1998). People who lose their olfactory sensecommonly become profoundly depressed. Ol-factory loss also appears to be associated witha decline in verbal memory (Swan & Carmelli,2000) and may be predictive of general cogni-tive decline (Graves et al., 1999).


The possible associations between olfactionand emotional and cognitive processes hold im-plications for research environments, yet weare not careful about the quality of air or odorcompounds in those research environments.The importance of air control in research envi-ronments has not been self-evident outsidesmell and taste laboratories. We know thatmalodors adversely affect human responses,but what of the many omnipresent odors thatdrift innocently through a research environ-ment: coffee, microwaved lunches, a strongshampoo, perfumes or colognes? Other sorts of“air control” are less obvious but equally (per-haps even more) important—for example, anx-iety of researchers and participants, or contin-ued use of the same space for negative (orpositive) emotion investigations.

THE BIOLOGY OF SEMIOCHEMICALS

Not only the genetics of olfaction, but even thephysiology, has undergone a change. TheWoodworth and Schlosberg text ExperimentalPsychology (1955) simply stated that “the re-ceptors for smell are found in two smallpatches of yellowish olfactory epithelium”(p. 304). We now know that not one system,but three, respond to airborne chemical signals:(1) the olfactory system, (2) the trigeminal sys-tem, and (3) the vomeronasal system (VNS). Inthis section, we quickly highlight pertinent as-pects of each system that help to inform emo-tion research. We focus on the odor discrimina-tion feature of what is traditionally called theolfactory sensory system (e.g., is an odor ba-nanas or gardenias?); for the trigeminal system,we focus on somatic arousal detection (e.g., isan odor irritating?); for the still-controversialVNS, we focus on “putative” pheromones.

The Olfactory System

The popular wisdom that odor and emotionshare a privileged link is supported by evidencefrom neuroscience; they share common path-ways within the limbic system. Succinctly, odorinformation is detected by the olfactory epithe-lium in the nasal cavity and relayed to the mainolfactory bulb, which in turn projects directlyto the ipsilateral primary olfactory cortex(Brand, 2006). The olfactory tract projects to,among other brain regions, the amygdala and

the orbito-frontal cortex. Both of these havebeen implicated in emotion processing (Baas,Aleman, & Kahn, 2004; Britton et al., 2006;Rolls, 1996; Zald, 2003). It is particularlynoteworthy that the projections are bidirec-tional (Zald & Pardo, 1997). Recent brain im-aging studies of olfactory processes suggestthat olfactory information is processed both hi-erarchically and in a parallel fashion, as aredata from other sensory systems (Savic, 2005).However, Savic notes that immediate emo-tional response is commonly associated withodor perception. The amygdala, she notes, isactivated at the most primary level of odor per-ception. As the odor-related task becomes moresemantic (i.e., odor memory), additional brainstructures become engaged. Later in this chap-ter, we present a functional approach to the ol-factory system, which suggests that semanticand memory functions of odors or semio-chemicals may also occur after emotional ormood changes have begun. In effect, odors maypotentiate emotional responses even whenthere is no attentional focus on the odor or it isbelow the threshold for detection.

Olfaction, in contrast to the other sensorymodalities, does not affect the thalamus ini-tially, but instead innervates the central ner-vous system (CNS) (Stockhorst & Pietrowsky,2004). It is further distinguished by its exter-nally exposed neurons, which die and renewthemselves on a regular basis (Carleton,Petreanu, Lansford, Alvarez-Buylla, & Lledo,2003). The generative nature of the system hasled several researchers, including us, to wonderwhether “learning” to discriminate some air-borne chemicals might involve building a struc-ture to key onto previously undiscriminatedmolecules. This would account for reports thatpeople who are anosmic to certain odors be-come sensitive over a period of several weeksor months of exposure. These unique featuresof the olfactory system imply fundamental neu-ral and CNS differences from other sensorymodalities.

Another interesting feature of the olfactorysystem is that organisms may respond physio-logically and emotionally in the presence ofsemiochemicals even when they do not noticethe presence of any odor. Apparently, consciousawareness, much less the identification ofodors, is not a prerequisite for eliciting a physi-ological or behavioral emotional response (e.g.,Bensafi et al., 2003). Once again this observa-

A “Nose” for Emotion 237

tion lends itself to the hypothesis that semio-chemicals may influence emotional preatten-tive processes and affect both mind and body.

People seem largely unaware that they them-selves produce chemosignals. This is not to saythat people are not aware of body odors, butthis awareness seems to be limited to specialcases, such as bad odors on the negative side orthe odor of clean infants on the positive side.People seem equally unaware that they are de-tecting and processing these chemical messagescoming from others. Our review (below) showsthat some mood signals as well as sexual onesare communicated largely without awareness.

The Trigeminal System

The trigeminal system serves somatosensory(e.g., temperature or pain) and motor (e.g., bit-ing and chewing) functions for the face andhead (see Brand, 2006, for a review). Thetrigeminal nerve, the fifth cranial nerve, is com-posed of three branches—the ophthalmic, themaxillary, and the mandibular—each of whichinnervates different parts of the face. Thesenerves project eventually to the thalamus andthen to the cortex. Only the mandibular nervecarries motor information. The first twobranches exhibit complex interactions with theolfactory system as these branches of thetrigeminal nerve extend into the nasal lining.Among its functions is to detect chemical irrita-tion. The typical reaction of recoiling from thesmell of ammonia, for example, is due to acti-vation of the trigeminal system. The emotionalreaction to such irritation is usually negative,but it can be uplifting or exciting, as when asniff of strong peppermint is “fresh” and ap-pealing. It is noteworthy that the trigeminalnerve relays proprioceptive stimulation to thesensory cortex. This is part of the facial feed-back of emotional expression in the facial mus-culature. Emotional expression is part of thissystem as well.

The Vomeronasal System

A third system receptive to specific airbornesemiochemicals—“pheromones”—is the VNS.Present in a wide variety of species, neuronsfrom the vomeronasal organ (VNO) form thevomeronasal nerve, which synapses in the ac-cessory olfactory bulb (Halpern, 1987). Hereinlies controversy. Although a VNO is observablein the human fetus and in many adult humans,

its existence may not be universal; when it ispresent, it may not be functional (see Meredith,2001, for a review). A neural pathway has notbeen observed in humans. Since the purpose ofthe VNO in most animals seems to be detectionof pheromones, this suggests that humans maynot respond to pheromones, at least not withthe VNO.

Pheromones are specific subsets of semio-chemicals that are emitted by members of aspecies and elicit behavior change in recipientsof the same species (Karlson & Luscher, 1959).Although pheromones are most often consid-ered in terms of sexual attraction, they serveother functions as well. For example, the antthat has discovered a new food source leads itscolony to that source via a pheromone trail.Research has demonstrated that humans areinfluenced by chemical stimuli from otherhumans—findings we will return to later. Aresuch semiochemicals pheromones? The presentgeneral suspicion is that pheromone reaction isvery limited or nonexistent in humans; how-ever, there are many reasons to consider thisquestion open to new experimentation andtechniques. What about the lack of conclusionregarding a functional VNO? Animal researchsuggests that a VNO is not necessary forpheromone detection (Dorries, Adkins-Regan,& Halpern, 1997), diminishing the importanceof the VNO question and placing detectionback into the olfactory system. In any case,there has been only a modest research effort toexamine the effect of “putative” human phero-mones on emotional reactions.

Our early understanding of olfaction hasbeen replaced with a more nuanced apprecia-tion that expands the neural impact of olfac-tory responses, considers the impact of at leastthe trigeminal system in the olfactory experi-ence, and raises tantalizing questions when theVNS is considered as well. This begins to inter-sect with folk wisdom and naïve psychologybeliefs that smell matters in mood and emo-tional behavior. The question to be resolved isin what ways and to what extent it matters.

EMOTIONAL AND COGNITIVEPROCESSES: MEMORY

In the novel Swann’s Way (Proust, 1913/1957),the protagonist experiences a profound emo-tional experience while sipping tea and dunk-ing madeleines (French tea cakes):


. . . I raised to my lips a spoonful of the tea inwhich I had soaked a morsel of the cake. Nosooner had the warm liquid, and the crumbs withit, touched my palate than a shudder ran throughmy whole body, and I stopped, intent upon the ex-traordinary changes that were taking place. Anexquisite pleasure had invaded my senses, but in-dividual, detached, with no suggestion of its ori-gin. And at once the vicissitudes of life had be-come indifferent to me, its disasters innocuous, itsbrevity illusory—this new sensation having hadon me the effect which love had of filling me witha precious essence; or rather this essence was notin me, it was myself. I had ceased now to feel me-diocre, accidental, mortal. Whence would it havecome to me, this all-powerful joy? (p. 56)

After much struggle, a memory reaches “theclear surface of [his] consciousness” (p. 57),and his reminiscence begins. The phenomenonthat Proust describes resonates with many peo-ple (see Gilbert & Wysocki, 1987). (The expe-rience of Proust’s protagonist is primarily a fla-vor experience. Flavor is the result of theintegration of taste and smell, with olfactionsupplying about 70% of the sensory informa-tion.) Although this particular taste/odor pro-duces an idiosyncratic intense “joy” response,the Proust tale has led to a general hypothesisthat odors can stimulate the intense recall ofevents, particularly childhood events, betterthan other stimuli and with rich emotional con-tent. The underlying argument is that odorshave a privileged function in memory. Researchoffers support for this hypothesis, but it mustbe qualified.

Discussing her work investigating odor–memory links, Herz (2000) posits that odorstrigger memories that are more emotional thanmemories triggered by other sensory modali-ties. In one study (Herz & Cupchik, 1995),odors described as either positive or negative,or their verbal odor labels, were paired withpaintings described as either positive or nega-tive. Participants wrote a description of thepaintings after viewing them in the presence ofeither the actual odor or the verbal cue. Partici-pants were tested 2 days later for recall of thedescription of the paintings. Odor was associ-ated with greater emotional content in the rec-ollections, but not greater accuracy.

In another test of the Proust phenomenon,Aggleton and Waskett (1999) recruited partici-pants who had taken from one to three guidedtours of the Jorvik Viking Centre in York,Great Britain. The underground Centre recre-

ates 10th-century York, complete with distinc-tive smells of that period. The tours arescripted and the smells are consistent. In thestudy, participants were exposed to eitherJorvik’s distinctive smells, novel odors, or noodors while completing questionnaires thatprobed for things they might have rememberedfrom their visits. Those exposed to the Jorviksmells recalled significantly more details abouttheir previous visits to the Centre, compared tothose in both the novel-odor and no-odor con-ditions. The Jorvik smells elicited more detailedrecall of the memory of the visit. In this case,the odor association did improve accuracy.

The previously mentioned studies investi-gated recall as a result of the associations be-tween odor and an autobiographical memory.Odors may evoke more than learned associa-tions, however, and may frequently affectnonattentive or preattentive processes. In fact,one of the issues concerning odor-evoked mem-ories is that the full range of odor effects maybe masked by a tendency to focus only onlearned, semantic processes. Semantic pro-cesses, as higher-order, explicit processes, mayoverride other processes linking emotion andodors when explicit verbal responses are usedas a measure (see Savic, 2005). Implicit pro-cessing may be equally or more effective forodors evoking both emotion and behavior, andmay lead to responses that are not easily linkedto previous experience (Degel & Koster, 1999;Hudson, Wilson, Freyberg, & Haviland-Jones,2007).

As an example of odor exposure leading tosubconscious behavioral change, Holland,Hendriks, and Aarts (2005) showed that expo-sure to undetected citrus cleaner led partici-pants in a word recognition task to respondfaster to words related to cleanliness. They alsospontaneously cleaned up their own cookiecrumbs in the next phase of the study. This is afascinating example of unconscious searchesfor congruent information related to unde-tected odors. (Similar effects are described be-low in the section on pheromones.) This studydoes not address the question of learning—participants may have learned the link betweenthe odor of cleaner and cleanliness—but it doesspeak to the question of whether an odor mustbe identified before it can affect behavior. Inthis case, the odor was not identified. Thismakes us suspect with amusement and alarmthat our human moods and actions are contin-uously influenced by chemicals (both natural


and artificial) that evoke a search for congruentinformation and behavior.

Phillips and Cupchik’s (2004) work on odorsfacilitating memory for textual material alsodemonstrates that explicit semantic processingis not necessary to evoke congruent search andbehavior. They found that lower-intensityodors may be more influential in heighteningrecollection, hinting that the explicit semanticassociations may sometimes take precedencewhen attention is brought to them and thenmay block other aspects of odors. Phillips andCupchik found that pleasant paragraphs readin the presence of a pleasant floral odor weremore likely to be remembered for affiliativeand social content. Accuracy of recall was en-hanced if the participants perceived the odor tobe of low intensity. The authors suggested thatan odor that is too intense may distract fromthe inherent unity between odor and experi-ence. One of the important results of this studyis the link not just between valences (pleasantto pleasant), but also of the particular pleasantfloral odor to social memory. This associationraises the possibility that some odors may carryvestiges of communicative messages that arenot necessarily learned. As some flowers emitodors that mimic the social odors of a mamma-lian species in order to attract them (vonHelversen, Winkler, & Bestmann, 2000), it ispossible that odors produced by a variety ofplants and animals may have some element ofintra- or interspecies information.

ODORS AS COMMUNICATION

We know that odors are meaningful to us; wejust do not appreciate the breadth of theirmeanings. Chemical signals serve a wide vari-ety of communicative functions, ranging fromthe mundane to the profound. Some of thesewe hold in common wisdom: Holiday smellsmay evoke festivity, while medicinal smells mayspecify illness and (we hope) healing. Newmothers like infant body odors more thannonmothers do (Fleming et al., 1993). And in-fants sleep better if they have an object withtheir mothers’ odor on it (Goodlin-Jones,Eiben, & Anders, 1997). These observationsmake sense to us. Who would not associateholiday smells with specific emotional re-sponses? Or who would challenge the idea thatan infant is comforted by its mother’s smell?Such associations seem reasonable, even

though we would also not be surprised if theassociation were absent.

But there is evidence for other kinds ofmeaning in odors. For example, several studieshave demonstrated that relatives can identifythe clothing of other relatives: Mothers canidentify the clothing of their neonates (Russell,Mendelson, & Peeke, 1983), and siblings canidentify clothing of siblings (Porter, Balogh,Cernoch, & Franchi, 1986). At first glance, onemight say that these could very well be learnedassociations. But grandparents seem to be ableto find the hospital shirts of their newborngrandchildren before meeting the infants (Por-ter et al., 1986). These are “family” identifica-tions. Strangers also can match the worn t-shirts of mothers and their children, but not thet-shirts of husbands and wives (Porter,Cernoch, & Balogh, 1985). These studies sug-gest that body odors provide kinship informa-tion, and thus communication functions, at adeep genetic level.

Related to this idea of genetic information isthe research on the major histocompatibilitycomplex (MHC). The MHC is a group of genesthat code for the immune system, as well as forone’s unique odor signature. In humans, thissignature is expressed as human leukocyte anti-gen (HLA) type. Studies have demonstratedthat in closed populations, people are morelikely to prefer partners with an HLA type dif-ferent from their own, thus incurring the ge-netic advantages of a more diverse gene poolfor the immune systems of their offspring(Ober et al., 1997). Another study shows thatwomen prefer the scent of t-shirts worn by menwith an HLA type other than their own(Wedekind, Seebeck, Bettens, & Paepke, 1995).These areas are still in their infancy and remaincontroversial, but suggest more influences ofhuman semiochemicals.

Pheromones

As mentioned earlier, pheromones are chemicalcompounds secreted into the environment bymembers of a species and detected by othermembers of the same species, in whom theytrigger behavioral, physical, and/or neuro-endocrine responses (Karlson & Luscher,1959). The studies to date that present thestrongest evidence for pheromonal influencesin humans are menstrual synchrony studies(McClintock, 1971; Stern & McClintock,1998) and studies of the effects of male axillary


compounds on women’s cycles and reportedmood (Preti, Wysocki, Barnhart, Sondheimer,& Leyden, 2003). Women exposed to theaxillary secretions of other women (Stern &McClintock, 1998) or even of men (Preti et al.,2003) will have changes in their menstrual cy-cle. Our basic biological regulatory cycles areinfluenced by exposure to the body chemicalsof others, even when the persons themselvesare absent.

In behavioral research, studies have focusedmost often on the 16 androstene steroids aspheromone candidates. Between the mid-1970sand the mid-1990s, androstenone andandrostenol were targets of research becausethe quantities of these steroids present in adultmales are roughly equivalent to those producedby boars, and these compounds reliably func-tion as pheromones in pigs (Dorries, Adkins-Regan, & Halpern, 1995). The results of thisbody of work are contradictory. For example,although women who wore necklaces lacedwith androstenol while they slept reportedmore social encounters with men the next day(Cowley & Brooksbank, 1991), other womenexposed to androstenol did not report an in-creased sexual response to erotic passages(Benton & Wastell, 1986). Women exposed toandrostenone rated themselves as less sexythan women exposed to other compounds(Filsinger, Braun, Monte, & Linder, 1984)—not the result one would expect.

It has been suggested that androstenol andandrostenone may serve territorial functions,rather than sexual attraction functions. Womenwaiting in a dentist’s office preferred chairssprayed with androstenone (Kirk-Smith &Booth, 1980), while men tended to avoid bath-room stalls sprayed with androstenol(Gustavson, Dawson, & Bonett, 1987). Again,these were instances in which exposure to anundetected airborne chemical influencedbehavior.

There are also reports that emotions may beinfluenced by these putative pheromones. Thefocus of research since the mid-1990s has beenon androstadienone (AND), another 16androstene steroid. Work with AND wassparked by the still controversial finding thatAND elicited reaction in the VNO of females,but not males (Grosser, Monti-Bloch, Jennings-White, & Berliner, 2000). The same researchreported a decrease in negative affect associ-ated with AND. The Grosser et al. study deliv-ered AND directly to the nasal cavity via a va-

por pulse. Using a different methodology, Ja-cob and McClintock (2000) suggested thatAND “modulates” negative mood in women.These researchers disguised the AND withclove oil and exposed the women to the stimuliby wiping a solution beneath the nose. Thewomen in the control condition reported in-creased negativity and decreased positive moodover several hours; the women in the ANDcondition remained stable over time. Althoughthese studies suggest that AND may indeedhave mood effects, any conclusion that AND isa human pheromone remains speculative. Nev-ertheless, while the question of whether peoplerespond to classical pheromones is debated, theemotional responses of humans to somesemiochemicals is less dubious.

Mood Odors

There is accumulating evidence that humanscommunicate emotionally with body odors.One of the earliest studies to reveal this wasconducted by Chen and Haviland-Jones(1999). As a part of this study, the effect of hu-man odors on mood reports and dream memo-ries was investigated. To this end, underarmpads (absorbing axillary odors) were collectedfrom nightshirts (provided and laundered withunscented products) worn for a number of daysby male and female odor donors who wereyoung children, young adults, or older adults.Participants who were exposed to these odorslater were asked to report their moods and towrite a dream memory. No reference to theodor pads was made; they were simply placedon the working desk in an open glass container(the jar held the pads of one set of donors—same age and gender). Despite the fact that theodors of the young men and older women werenot discriminated and were both rated as mostunpleasant and intense, they affected themoods and dream memories of the detectorsvery differently. Specifically, the odors from theyoung men elicited more negative moods on arating scale and more aggressive dream con-tent. The older women’s odors elicited morepositive moods and more affiliative dream con-tent. This initial study raised two key ques-tions. First, it appears to be a challenge to anygeneralization of the Proust association effect.The odors were not identified or discriminatedas old versus young or as male versus female,and yet there were different mood effects. Sec-ond, both odors were intense and unpleasant,


and yet one evoked fairly pleasant responses.This is also a challenge to the associative expla-nations.

Because the odors judged to be the same haddifferent mood effects, Chen and Haviland-Jones (1999, 2000) hypothesized that moodwas being communicated. The young men inthe study who were donors were all college stu-dents and at a time in their lives that was inher-ently stressful and required a goal-oriented atti-tude. Even their participation in our study wasmotivated by a course requirement rather thanpersonal interest. It seemed likely that theseyoung men might experience higher levels offrustration and aggression than other groupswhile wearing our axillary pads. On the otherhand, the older women, who were all veryhealthy and retired, were probably living theirlives at a more peaceful pace, focused on per-sonal fulfillment and possibly nurturing activi-ties. As such, the moods and emotions they ex-perienced while wearing the axillary padscould have been more positive than those of theyoung males.

This led to a more direct investigation of thepossibility that mood can be communicatedwith chemosensory products in axillary perspi-ration. To produce the mood odors, donors(nonsmokers, not wearing deodorant) cameinto the lab, placed the axillary pads undertheir arms, and watched a short video thatshowed repeated brief clips from popular mov-ies of actors showing an emotional reaction(e.g., fear, anger, or happiness). Detectors werelater asked to match the axillary pads to amood name (e.g., “fear,” “happy,” “control”).The detectors were able to identify the mood atlevels significantly above chance. A subsequentstudy by another lab (Ackerl, Atzmueller, &Grammer, 2002) has at least partially repli-cated the Chen and Haviland-Jones (1999)findings.

Additional studies have expanded our un-derstanding of the effects that body odors haveon those exposed to them. For example, Chen,Katdare, and Lucas (2006) asked how cogni-tive processing might be influenced by the pres-ence of mood odors. Participants were given aword association task in the presence ofaxillary pads from donors who watched a scaryor neutral video. The word task paired neutraland/or threat words in a counterbalanced de-sign. Although threat-related words were pro-cessed equally fast in all mood odor conditions,

participants were more accurate in the fearcondition. On the other hand, in ambiguouspairings (threat-neutral, for example), the fearodor slowed processing. According to Chen etal., exposure to the fear odor led the subjects toadopt a more efficient yet cautious cognitivestrategy for the task. In a similar vein, DeGroatand Haviland-Jones (2005) found a differencein reaction time between the fear and happyodor conditions in a visual image reaction task.Here the images were mixed, pleasant and un-pleasant. Again, the fear odor was associatedwith slower reaction times than the happyodor.

It seems counterintuitive that the fear odorssometimes would lead to slower responding.One explanation might be that sensing an odorwithout having a visual or auditory matchleads to a slow, cautious search for a matchrather than the rapid “fear–flight” response. Infact, it is possible that sensing a fear odor withno obvious source would lead to “freezing”rather than flight, because the direction offlight toward safety would not be known.

OF FLOWERS AND HUMANS

Turning to the influence of more pleasant odorson moods, we serendipitously had begun a re-search program in positive psychology that hadbroad environmental goals and implications.We (Haviland-Jones, Rosario, Wilson, &McGuire, 2005) investigated how flowers,some of which were fragrant, as comparedwith other objects in the environment, evokepositive emotional and social responses in hu-mans. The 2005 article describes three differentstudies conducted over several years. The firststudy in the series riveted our attention, be-cause 100% of the participants exposed to amixed flower bouquet responded with the Du-chenne (zygomatic/orbicularis orbis) smile.

In the first study (Haviland-Jones et al.,2005) two members of the research team (oneto code, one to present) brought a flower bou-quet, a fruit basket, or a pillar candle to eachparticipant to thank him or her for participa-tion. The stimulus was closed in a box to hide itfrom the coder (so the coding would be double-blind), as well as to prevent anticipatory re-sponses from the participant. All the partici-pants who received the floral stimulus re-sponded with the Duchenne smile within 5


seconds of the door on the box opening to re-veal the stimulus. The group receiving the flow-ers also had a pre- to poststimulus moodchange; only the group receiving the flowers re-ported more positive emotion over a 3-day pe-riod.

In the second study (Haviland-Jones et al.,2005), participants received “doses” of floralbouquets. Those who received more than oneover several weeks had the lowest reports ofdepression. Participants in this study, who wereall over age 60, scored higher on an episodicmemory task after receiving the bouquets. Thismay indicate some potential motivational ef-fects on cognitive tasks as well as emotional re-sponses.

In the third study, there were more positivesocial interactions when a single flower wasgiven to a person in an elevator than when apen was given (or nothing). Not only were par-ticipants in the elevator more likely to respondwith a Duchenne smile, but they were morelikely to move closer to the experimenter, initi-ate new conversations, and look directly at theexperimenter when a flower was given. Thiswas the case for men as well as women, eventhough most people in modern Western cultureassociate flowers with women and predomi-nantly give flowers to women.

A closer examination of earlier research(Dimberg & Thell, 1988) also suggested thatflower photographs presented subliminallymay elicit a positive response. Dimberg andThell exposed participants to fear stimuli (e.g.,snakes) and a “neutral” stimulus (flowers)tachistoscopically. They found that exposure toflowers elicited electromyographic (EMG) acti-vation of the zygomatic muscles (a smile), re-ferring to this as a positive response. They didnot, however, measure the EMG activation ofthe orbicularis orbis muscles, also required forthe Duchenne or “true” smile, so they couldnot conclude that flowers elicited happiness.The Duchenne smile reliably indicates happi-ness and positive affect, even in the absence ofa self-report of happiness (Dimberg, Thunberg,& Elmehed, 2000). Such a smile activates boththe zygomatic muscles of the cheeks and theorbicularis orbis muscles of the eyes.

There are several explanations for the flow-ers and positive emotion effect, only one ofwhich involves semiochemicals. There aresome theories that flowers mark the placeswhere fruit would be harvested, and thus are

an evolutionary food signal (Pinker, 1997) andnaturally associated with positive feelings. Butin that case, one would have expected the fruitgift itself to produce as much happiness as theflowers (if not more), and this was not the case.On its own, the flowers’ association with po-tential food cannot explain our results. Alter-natively, it may be that the symmetry offlowers—a cue for memory and recognizability,and an evolutionary signal for detecting foodand other significant events (Enquist & Arak,1994)—led to the smiles and happiness. Thisidea is also captured by Cupchik (2006), whopresents arguments for the evolutionary signifi-cance of aesthetics. However, it would be diffi-cult to argue that the floral bouquet had moresymmetry than the fruit basket or the pillarcandle. It is possible that the semiochemicalsand the fragrance of the flowers alone mighthave produced smiles and happy moods. Or allof the proposed effects may be synergistic,lending flowers a quality of “superstimulus”—appropriate odors, colors, symmetry, and asso-ciation with food rewards. The significance ofthe studies seems to be that a naturally occur-ring stimulus such as flowers, not necessarilyrewarding in any biological sense (not food,shelter, etc.), can affect mood rather dramati-cally.

We do not dismiss the possibility thatlearned associations accounted for some of theeffects. However, learned associations alonecannot explain why the effect of the flowerswas so much greater than that of the fruit andcandles. The reactions of participants who re-ceived flowers was astonishing; they invited theexperimenters into their homes and sent thank-you cards, sometimes with photographs of theflowers. In the elevator study, several partici-pants who received nothing or a pen were dis-appointed upon seeing other people emergingfrom the elevator with flowers and returned re-peatedly in an attempt to receive a flowerthemselves; no one returned for the attractivepen, seemingly more useful in the library.

It is likely that odor plays a very significantrole in the human preference for flowers. Notonly do people choose flowers because theylike the odor, but they also perfume themselvesand their personal belongings with floralodors. In one set of studies, both men andwomen who wore floral and other colognes fora month reported better moods than a similargroup of people who did not have this treat-


ment (Schiffman, Sattely-Miller, Suggs, & Gra-ham, 1995; Schiffman, Suggs, & Sattely-Miller,1995). The pleasant odors of flora may aloneaccount for the positive mood effects we havedescribed above. Cupchik has reported thatpleasant floral odors used when people werereading a passage from literature influenced theepisodes remembered from that passage(Cupchik & Phillips, 2005; Phillips &Cupchik, 2004). Those exposed to such odorsremembered the positive emotional materialbetter, as well as some types of social informa-tion. This once again suggests a strong associa-tive effect, because there is an association be-tween the pleasantness of the odor and thepositive nature of the episode.

In a report prepared for the Sense of SmellInstitute, we (Haviland-Jones & Wilson, 2005)found that the odors of some flowers (gardeniaand others), even when present at such a lowlevel as to be undetected by the participants,prevented negative emotional responses tostressful video material and also sent people to“search” their own memories for more positiveemotional associations. We have argued thatbecause the mood-enhancing effects of theseodors were found even when the participantscould not detect the presence of any odor, anonsemantic, possibly nonassociative processis necessary to explain the mood effects of flo-ral odors.

We suggest that flowering plants have founda positive “mood niche” to exploit: Humansprovide propagation and care for these specialflowering plants, which in turn improvemoods. Positive mood itself has survival bene-fits (see Lucas & Diener, Chapter 29;Fredrickson & Cohn, Chapter 48; andPanksepp, Chapter 4, this volume). There maybe some places in the world or some plants oranimals that evoke or provide more happinessfor people than other places, plants, or ani-mals. Some aspects of this process are likely toinclude semiochemicals and odors.

Interspecies communication through chemo-sensory signals is not uncommon in the insectphyla (Weller, Jacobson, & Conner, 2000) andis even noted in mammals (von Helversen et al.,2000). Some flowers give off an odor thatbrings large gatherings of bats to “party,” asvon Helversen et al. (2000) report. These flow-ers require multiple pollinators, and theswarms of bats comply in the requisite num-bers. They come in response to a chemo-

sensory, airborne signal that mimics bat odorsused for gathering. When the flowers no longerrequire pollination, they cease to put out thechemosensory signal. So it is known that plantsand mammals may have coevolved systemsthat require chemosensory information. Hu-mans and flowers may also be involved in sucha system, demonstrating another avenue for ol-factory and semiochemical functions—no mat-ter whether this effect is unknown to the recipi-ent of the chemical, amusingly suspected, oractually demonstrated as in our (Haviland-Jones et al., 2005) research.

OUR HYPOTHESESABOUT THE NOSE AND EMOTION

We humans are overwhelmed with semio-chemicals at all times in our lives. Some like tosay that we live in a “chemical soup”—not re-ferring to factory wastes, but to the everydaychemical information that plants, storms, ba-bies, books, foods, and so forth produce. Forthe most part, it seems that we do not need toattend to the specific information; it sums itselfas congruent or incongruent with other cuesand is generally part of the background ratherthan the foreground. Some of this “soup” ap-parently cues us about the moods in the air, butnot in such a way that we are usually aware ofit.

A number of studies have led us to a hypoth-esis we have called the “search” function ofsemiochemicals. The puzzle is that small, fairlyconsistent responses to semiochemicals orodorants are found in many behaviors, such asreaction times or semantic constructions. Thesebehaviors are not noticed by the people pro-ducing them, and in fact may be denied. For ex-ample, even though some say they can smellfear, for the most part they do not claim to bemade afraid by that. Almost no one claims tosmell happiness, and the potentiating effects ofa happy odor are summarily dismissed. Yetthese odors may in fact have unnoticed, small,consistent effects. The general phenomenonthat behavior may occur without actual under-standing of its stimulus is known in several sys-tems.

There is ample evidence that we can respondto stimuli that we do not recognize (e.g.,Öhman & Mineka, 2001; Zajonc, 1980), andthat we may attribute our responses—memo-


ries, associations—to the wrong stimulus(Tversky & Kahneman, 1992). Öhman (seeChapter 44, this volume) points our that thereare likely to be different brain structures in-volved in preattentive, potentiating stimulusprocessing versus focused attentive processes.In the case of long-term anxiety versus fear of aparticular stimulus, he argues that it is possibleto influence or potentiate the search foranxiety-causing stimuli, independently of at-tention to an actual fearful stimulus. Thepotentiation seems quite likely to interfere withprocessing nonfearful stimuli. In the case of ourmostly pleasant chemosensory stimuli, we sus-pect that even other emotions may be similarly“potentiated.” Pleasant stimuli with social sig-nificance may well lend themselves to thesearch for congruent information and may in-hibit the search for incongruent stimuli.

We propose that there must be an ongoing,not-quite-conscious match–mismatch systemthat is correlating bits of information acrossthe sensory systems. It is actively searching forexisting associations and making new associa-tions. If the semiochemical stimulus indicates aromantic possibility, the behavior of the personexposed to a romantic semiochemical may be-come more congruent with a search forromance—a bit of leaning forward, a littletouch of the hair, as in a tiny display. It is un-likely that anyone would be aware of this shiftin behavior. If there is no receptive person, orno time and place for romance, the behaviorselicited by the semiochemicals are short-circuited. There is not enough concurrence tosupport a romantic interlude. On the otherhand, if the search finds a response, a person,or an event, we conclude that the words of thepotential partner, or the mood music, or someother visible or audible event precipitated ro-mance, not the semiochemical and our ownsubtle “search” behavior. We seldom refer tothe romance in semiochemicals, even thoughwe say colloquially that “romance was in theair.” As another example, a man who searchesfor something pleasing to concur with a pleas-ant odor may think of an upcoming vacation. Ifasked why he is happy, he refers to the vaca-tion, not to what the nose has sensed. He cred-its the result of a search, not the impetus for thesearch.

Whether a search is initiated and abortedwhen no other sensory or ideational systemconfirms the search, or whether the search

leads to a target that is then presumed to be theevoker, the information from the nose is notcredited. In the theater, if an odor lingering af-ter a horror movie showing was different fromthe odor lingering after the comedy theater pa-trons departed, the nose might well know, butthe narrative mind behind the nose would takeall the credit for knowing and attributing theextra “frisson” to other things.

NEW DIRECTIONS

Based on the available research it is reasonableto believe that we humans have somechemosensory signals that contain mood infor-mation, and that we are generally able to haveour moods influenced by a variety of naturaland constructed odorants. We are a consider-able distance from knowing how influentialsuch communication might be, whether it islearned differently for men and women and indifferent cultures, or whether there are univer-sal elements. The genetic components of thechemical aspects are also still to be discovered.Could we vary behavioral tendencies or atti-tudes as well as mood by changing thechemosensory environment? To suggest thatthe environment, chemosensory or otherwise,is psychologically unimportant assumes thatwe are separate and apart from the environ-ment. But we are not separate; we interact withthe environment, and such interaction by defi-nition is bidirectional. That we can come to un-derstand the influences of such bidirectionalityon mood, behaviors, or human performance ingeneral enhances our well-being and is essentialto understanding the nature of who we are.

ACKNOWLEDGMENT

We would like to thank Linnea R. Dickson for herexpert commentary on and editing of this chapter.

REFERENCES

Ackerl, K., Atzmueller, M., & Grammer, K. (2002). Thescent of fear. Neuroendocrinological Letters, 23, 79–84.

Aggleton, J. P., & Waskett, L. (1999). The ability ofodours to serve as state-dependent cues for real-world memories: Can Viking smells aid the recall ofViking experiences? British Journal of Psychology,90, 1–7.


Baas, D., Aleman, A., & Kahn, R. S. (2004).Lateralization of amygdala activation: A systematicreview of functional neuroimaging studies. Brain Re-search: Brain Research Reviews, 45(2), 96–1003.

Bensafi, M., Brown, W. M., Tsutsui, T., Mainland, J. D.,Johnson, B. N., Bremner, E. A., et al. (2003). Sex-steroid derived compounds induce sex-specific effectson autonomic nervous system function in humans.Behavioral Neuroscience, 117(6), 1125–1134.

Benton, D., & Wastell, V. (1986). Effects of androstenolon human sexual arousal. Biological Psychology, 22,141–147.

Brand, G. (2006). Olfactory/trigeminal interactions innasal chemoreception. Neuroscience and Bio-behavioral Reviews, 30, 908–917.

Britton, J. C., Phan, K. L., Taylor, S. F., Welsh, R. C.,Berridge, K. C., & Liberzon, I. (2006). Neural corre-lates of social and nonsocial emotions: An fMRIstudy. NeuroImage, 31(1), 397–409.

Buck, L., & Axel, R. (1991). A novel multigene familymay encode odorant receptors: A molecular basis forodor recognition. Cell, 65, 175–187.

Cain, W. S. (1978). History of research on smell. In E. C.Carterette & M. P. Friedman (Eds.), Handbook ofperception (Vol. 6, pp. 197–229). New York: Aca-demic Press.

Carleton, A., Petreanu, L. T., Lansford, R., Alvarez-Buylla, A., & Lledo, P. (2003). Becoming a new neu-ron in the adult olfactory bulb. Nature Neuroscience,6(5), 507–518.

Chen, D., & Haviland-Jones, J. (1999). Rapid moodchange and human odors. Physiology and Behavior,68, 241–250.

Chen, D., & Haviland-Jones, J. (2000). Human olfac-tory communication of emotion. Perceptual and Mo-tor Skills, 91, 771–781.

Chen, D., Katdare, A., & Lucas, N. (2006) Chemo-signals of fear enhance cognitive performance in hu-mans. Chemical Senses, 31, 415–423.

Classen, C., Howes, D., & Synnott, A. (1994). Aroma:The cultural history of smell. London: Routledge.

Cowley, J. J., & Brooksbank, B. W. L. (1991). Humanexposure to putative pheromones and changes in as-pects of social behaviour. Journal of Steroid Bio-chemistry and Molecular Biology, 39(4), 647–659.

Cupchik, G. C. (2006). Emotion in aesthetics and theaesthetics of emotion. In P. Locher, C. Martindale, &L. Dorfman (Eds.), New directions in aesthetics, cre-ativity and the arts (pp. 209–224). Amityville, NY:Baywood.

Cupchik, G. C., & Phillips, K. (2005). The scent of liter-ature. Cognition and Emotion, 19(1), 101–119.

Degel, J., & Koster, E. P. (1999). Odors: Implicit mem-ory and performance effects. Chemical Senses, 24(3),317–325.

DeGroat, D. A., & Haviland-Jones, J. (2005, April). Ex-posure to human emotion semiochemicals (moododors) affects behavior. Poster presented at the an-nual meeting of the Association for ChemoreceptionSciences, Sarasota, FL.

Dimberg, U., & Thell, S. (1988). Facial electro-myography, fear relevance and the experience ofstimuli. Journal of Psychophysiology, 2, 213–219.

Dimberg, U., Thunberg, M., & Elmehed, K. (2000). Un-conscious facial reactions to emotional facial expres-sions. Psychological Science, 11(1), 86–89.

Dorries, K. M., Adkins-Regan, E., & Halpern, B. P.(1995). Olfactory sensitivity to the pheromone,androstenone, is sexually dimorphic in the pig. Physi-ology and Behavior, 57(2), 255–259.

Dorries, K. M., Adkins-Regan, E., & Halpern, B. P.(1997). Sensitivity and behavioral responses to thepheromone androstenone are not mediated by thevomeronasal organ in domestic pigs. Brain, Behaviorand Evolution, 49, 53–62.

Enquist, M., & Arak, A. (1994). Symmetry, beauty andevolution. Nature, 372, 169–172.

Filsinger, E. E., Braun, J. J., Monte, W. C., & Linder, D.E. (1984). Human (Homo sapiens) responses to pig(Sus scrofa) sex pheromone 5 alpha-androst-16-en-3-one. Journal of Comparative Psychology, 98(2), 219–222.

Fleming, A. S., Corter, C., Franks, P., Surbey, M.,Schneider, B., & Steiner, M. (1993). Postpartum fac-tors related to mother’s attraction to newborn infantodors. Developmental Psychobiology, 26(2), 115–132.

Gilbert, A. N., & Wysocki, C. J. (1987) The smell sur-vey. National Geographic, 172, 514–525.

Goodlin-Jones, B. L., Eiben, L. A., & Anders, T. F.(1997). Maternal well-being and sleep–wake behav-iors in infants: An intervention using maternal odor.Infant Mental Health Journal, 18(4), 378–393.

Graves, A. B., Bowend, J. D., Rajaram, L., McCormich,W. C., McCurry, S. M., Schellenberg, G. D., et al.(1999). Impaired olfaction as a marker for cognitivedecline: Interaction with apolipoprotein E ε4 status.Neurology, 53(7), 1480–1487.

Grosser, B., Monti-Bloch, L., Jennings-White, C., &Berliner, D. (2000). Behavioral and electro-physiological effects of androstadienone, a humanpheromone. Psychoneuroendocrinology, 25, 289–299.

Gustavson, A. R., Dawson, M. E., & Bonett, D. G.(1987). Androstenone, a putative humanpheromone, affects human (Homo sapiens) malechoice performance. Journal of Comparative Psy-chology, 101(2), 210–212.

Halpern, M. (1987). The organization and function ofthe vomeronasal system. Annual Review of Neuro-science, 10, 325–362.

Haviland-Jones, J., Rosario, H. H., Wilson, P. J., &McGuire, T. (2005). An environmental approach topositive emotion: Flowers. Evolutionary Psychology,3, 104–132.

Haviland-Jones, J., & Wilson, P. J. (2005). Fragrance:Emotion, sensuality and relationships. New York:Sense of Smell Institute.

Herz, R. S. (2000, July–August). Scents of time. The Sci-ences, pp. 34–39.


Herz, R. S., & Cupchik, G. C. (1995). The emotionaldistinctiveness of odor-evoked memories. ChemicalSenses, 20(5), 517–528.

Holland, R. W., Hendriks, M., & Aarts, H. (2005).Smells like clean spirit: Nonconscious effects of scenton cognition and behavior. Psychological Science,16(9), 689–693.

Hudson, J. A., Wilson, P. J., Freyberg, R., & Haviland-Jones, J. (2007). Scents and sensibility: Effects of im-plicit odor exposure on autobiographical memory.Manuscript submitted for publication.

Jacob, S., & McClintock, M. K. (2000). Psychologicalstate and mood effects of steroidal chemosignals inwomen and men. Hormones and Behavior, 37(1),57–78.

Karlson, P., & Luscher, M. (1959). Pheromones: A newterm for a class of biologically active substances. Na-ture, 183, 55–56.

Kirk-Smith, M. D., & Booth, D. A. (1980). Effects ofandrostenone on choice of location in others’ pres-ence. In H. van der Starre (Ed.), Olfaction and taste(Vol. 7, pp. 397–400). London: IRL Press.

LoSasso, G. L., Rapport, L. J., Axelrod, B. N., & Whit-man, R. D. (2002). Neurocognitive sequelae of expo-sure to organic solvents and (meth)acrylates amongnail-studio technicians. Neuropsychiatry, Neuropsy-chology, and Behavioral Neurology, 15(1), 44–55.

McClintock, M. (1971). Menstrual synchrony and sup-pression. Nature, 229, 244–245.

Meredith, M. (2001). Human vomeronasal organ func-tion: A critical review of best and worst cases. Chem-ical Senses, 26, 433–445.

Mesholam, R. I., Moberg, P. J., Mahr, R. N., & Doty, R.L. (1998). Olfaction in neurodegenerative disease: Ameta-analysis of olfactory functioning in Alzheimer’sand Parkinson’s diseases. Archive of Neurology, 55,84–90.

Ober, C., Weitkamp, L. R., Cox, N., Dytch, H., Kostyu,D., & Elias, S. (1997). HLA and mate choice in hu-mans. American Journal of Human Genetics, 61,497–504.

Öhman, A., & Mineka, S. (2001). Fears, phobias andpreparedness: Toward an evolved module of fearand fear learning. Psychological Review, 108, 438–522.

Phillips, K., & Cupchik, G. C. (2004). Scented memo-ries of literature. Memory, 12(3), 366–375.


Porter, R. H., Balogh, R. D., Cernoch, J. M., & Franchi,C. (1986). Recognition of kin through characteristicbody odors. Chemical Senses, 11, 389–395.

Porter, R. H., Cernoch, J. M., & Balogh, R. D. (1985).Odor signatures and kin recognition. Physiology andBehavior, 34, 445–448.

Preti, G., Wysocki, C. J., Barnhart, K. T., Sondheimer, S.J., & Leyden, J. J. (2003). Male axillary extracts con-tain pheromones that affect pulsatile secretion ofluteinizing hormone and mood in women recipients.Biology of Reproduction, 68(6), 2107–2113.

Proust, M. (1957). Swann’s way (C. K. S. Moncrieff,Trans.). New York: Penguin Books. (Original workpublished 1913)

Rolls, E. T. (1996). The orbitofrontal cortex. Philosoph-ical Transactions of the Royal Society of London: Se-ries B. Biological Sciences, 351(1346), 1433–1443.

Rosenkranz, H. C., & Cunningham, A. R. (2003). Envi-ronmental odors and health hazards. The Science ofthe Total Environment, 313, 15–24.

Rotton, J. (1983). Affective and cognitive consequencesof malodorous pollution. Basic and Applied SocialPsychology, 4, 171–191.

Rouhi, A. M. (2002). Exploring the chemical senses.Chemical and Engineering News, 80(1), 24–29.

Russell, M., Mendelson, T., & Peeke, H. (1983).Mother’s identification of their infant’s odors. Ethol-ogy and Sociobiology, 4, 29–31.

Savic, I. (2005). Brain imaging studies of the functionalorganization of human olfaction. Chemical Senses,30(Suppl. 1), i222–i223.

Schaal, B., & Porter, R. H. (1991). “Microsmatic hu-mans” revisited: The generation and perception ofchemical signals. In P. J. B. Slater & J. S. Rosenblatt(Eds.). Advances in the study of behavior (Vol. 20,pp. 135–199). San Diego, CA: Academic Press.

Schiffman, S. S., Sattely-Miller, E. A., Suggs, M. S., &Graham, B. G. (1995). The effect of pleasant odorsand hormone status on mood of women at mid-life.Brain Research Bulletin, 36, 19–29.

Schiffman, S. S., Suggs, M. S., & Sattely-Miller, E. A.(1995). Effects of pleasant odors on mood of males atmid-life: Comparison of African-American andEuropean-American men. Brain Research Bulletin,36(1), 31–37.

Schiffman, S. S., & Williams, C. M. (2005). Science ofodor as a potential health issue. Journal of Environ-mental Quality, 34(1), 129–138.

Steele, J. J. (1992). The anthropology of smell and scentin ancient Egypt and modern South American sha-manism. In S. Van Toller & G. H. Dodd (Eds.), Fra-grance: The psychology and biology of perfume(pp. 287–303). London: Elsevier.

Stern, K., & McClintock, M. K. (1998). Regulation ofovulation by human pheromones. Nature, 392, 177–179.

Stockhorst, U., & Pietrowsky, R. (2004). Olfactory per-ception, communication, and the nose-to-brain path-way. Physiology and Behavior, 83, 3–11.

Stoddart, D. M. (1990). The scented ape. Cambridge,UK: Cambridge University Press.

Swan, G. E., & Carmelli, D. (2000). Impaired olfactionpredicts cognitive decline in nondemented olderadults. Neuroepidemiology, 21(2), 58–67.

Tversky, A., & Kahneman, D. (1992). Advances inprospect theory: Cumulative representation of un-certainty. Journal of Risk and Uncertainty, 5, 297–323.

Van Toller, S., & Dodd, G. H. (Eds.). (1988). Perfumery:The psychology and biology of fragrance. London:Chapman & Hall.


Van Toller, S., & Dodd, G. H. (Eds.). (1992). Fragrance:The psychology and biology of perfume. London:Elsevier.

von Helversen, O., Winkler, L., & Bestmann, S. J.(2000). Sulfur containing “perfumes“ attract flowervisiting bats. Journal of Comparative Physiology A:Neuroethology, Sensory, Neural, and BehavioralPhysiology, 186(2), 143–153.

Wedekind, C., Seebeck, T., Bettens, F., & Paepke, A. J.(1995). MHC-dependent mate preferences in hu-mans. Proceedings of the Royal Society of London:Biological Sciences, 260, 245–249.

Weller, S. J., Jacobson, N. L., & Conner, W. E. (2000).The evolution of chemical defenses and mating sys-

tems in tiger moths. Biological Journal of theLinnaean Society, 68, 557–558.

Woodworth, R. S., & Schlosberg, H. (1955). Experi-mental psychology. London: Methuen.

Zajonc, R. B. (1980). Feeling and thinking: Preferences needno inferences. American Psychologist, 35, 151–175.

Zald, D. H. (2003). The human amygdale and the emo-tional evaluation of sensory stimuli. Brain ResearchReviews, 41, 88–123.

Zald, D. H., & Pardo, J. V. (1997). Emotion, olfaction,and the human amygdala: Amygdala activation dur-ing aversive olfactory stimulation. Proceedings of theNational Academy of Sciences of the USA, 94, 4119–4124.


C H A P T E R 1 5

The Neuroimaging of Emotion

TOR D. WAGER, LISA FELDMAN BARRETT,ELIZA BLISS-MOREAU, KRISTEN A. LINDQUIST, SETH DUNCAN,

HEDY KOBER, JOSH JOSEPH, MATTHEW DAVIDSON,and JENNIFER MIZE

Questions about the nature of emotion haveexisted since psychology emerged as a scientificdiscipline in the late 19th century (Darwin,1872/1998; Dewey, 1895; Irons, 1897; James,1884). For a century, scientists were unable tomeasure emotions at their source, and so theyrelied on measures of behavior, reported expe-rience, and activity of the peripheral nervoussystem to address fundamental questions aboutwhat emotions are and how they function inthe economy of the mind. Although much waslearned that is of both scientific interest andpractical value, questions about the nature ofemotion remained fundamentally unresolved.The relatively recent introduction of neuroim-aging techniques, particularly functional mag-netic resonance imaging (fMRI) and positronemission tomography (PET), provide a newperspective on the emotion in the intact humanbrain, and have the potential to identify whichbrain areas are consistently and specifically as-sociated with particular types of emotionalstates. Fifteen years’ worth of neuroimaging re-search has investigated the neural underpin-nings of emotions—including responses to ba-

sic affective stimuli such as pictures or odors;the experience and regulation of the discreteemotional events that we refer to as “disgust,”“anger,” and “desire”; and the perception ofemotion in others. These findings have the po-tential to shed new light on what emotions areand how they work.

We begin this chapter by sketching a hypoth-esized “neural reference space” for emotion(Barrett, Mesquita, Ochsner, & Gross, 2007;Edelman & Tononi, 2001), which refers to theset of brain structures thought to instantiateemotions and related affective states. We thenexamine how findings from neuroimagingstudies map onto this space, and what theycontribute to the understanding of the brainbases of affect and emotion. We take a meta-analytic approach, integrating the results ob-tained in 163 individual studies to locate the re-gions most consistently activated across arange of emotion-related tasks. We refer to thisset of regions as the “observed neural referencespace.” Next, we bring meta-analytic evidenceto bear on three unresolved issues in the emo-tion literature. First, we ask whether the expe-

249

rience and perception of emotion produce dif-ferent patterns of brain activation. Second, weask whether the experiences of pleasant andunpleasant affect are instantiated by distinctcircuitry in the human brain. Third, we addressthe methodological question of whether PETand fMRI are equally suitable for studyingemotion in the human brain, particularly inbrainstem and midbrain areas. In addressingthese three questions, we also touch on otherissues, such as the centrality of the amygdala inemotion, the representation and lateralizationof affect in the brain, and the validity of fMRIas a means of studying emotional experience.Other recent meta-analyses have tackled thestructure of emotion (Murphy, Nimmo-Smith,& Lawrence, 2003; Phan, Wager, Taylor, &Liberzon, 2002), so we do not address thisquestion here.

THE HYPOTHESIZED NEURALREFERENCE SPACE FOREMOTION-RELATED PHENOMENA

Animal and lesion studies guide and constraincurrent thinking about the neural systems thatgive rise to emotion. Animal models can pro-vide exquisite neurophysiological detail thatconstrains theories about mental processes,and neuropsychology provides unique evidenceon the brain components necessary for intactemotional processes in humans. Although neu-roimaging offers unique advantages, in that itoffers a probe of brain function in the intacthuman, interpretation of neuroimaging studiesrelies heavily on these complementary meth-ods. Below, we provide a few pointers to themassive body of evidence on the affective brainthat informs our interpretations of the body ofneuroimaging studies to date. We begin in theoldest parts of the brain and work our way upto the cortical centers that so markedly differ-entiate humans from other species.

Brainstem

Brainstem nuclei form the oldest centers relatedto affective processing and generate autonomicoutput to regulate the heart, vasculature, andother visceral organs. Nuclei within these re-gions have generally bidirectional connectionswith other emotion-related structures, such asthe medial prefrontal cortex (mPFC), insula(INS), and amygdala (Amy) (Amaral, Price,

Pitkanen, & Carmichael, 1992; Barbas, Saha,Rempel-Clower, & Ghashghaei, 2004; Mc-Donald, 1998; Ongur, An, & Price, 1998;Ongur & Price, 2000). (For a listing of all ab-breviations used for brain regions in this chap-ter, see Table 15.1.) A particularly importantstructure is the midbrain periaqueductal gray(PAG), which is thought to coordinate coherentphysiological and behavioral responses tothreat (Bandler & Shipley, 1994; Holstege &Georgiadis, 2004; Van der Horst & Holstege,1998). Stimulation of different longitudinalcolumns of PAG elicits distinct coordinated,organism-wide response “modes” that mirrorthe natural affective reactions elicited bythreat. For example, stimulation of lateral col-umns elicits “defensive” behaviors, such as fac-ing and backing away from a perceived at-tacker, hissing, and attack when approached(Gregg & Siegel, 2001); it also elicits auto-nomic responses that include tachycardia, in-creased blood pressure and blood flow to theface, pupillary dilation, and piloerection(Lovick, 1992), as well as analgesia. Together,these effects are consistent with a defensive–aggressive emotional response. As PAG re-ceives direct projections from numerous corti-cal regions, including the anterior cingulatecortex (ACC), mPFC, anterior INS (aINS), andmedial temporal lobe (MTL) (Shipley, Ennis,Rizvi, & Behbehani, 1991), the PAG might bethought of as an integrative emotional center,and it plays a central role in some conceptionsof emotion (Panksepp, 1998).

In spite of their prevalence in animal models,brainstem nuclei are rarely discussed in neuro-imaging studies, partly due to their small sizeand difficulty in localization. Many studies re-port activation in these subcortical areas, how-ever, and we analyze the specificity and reliabil-ity of these activations in this chapter. Wefind consistent activations around the humanPAG, particularly in studies of negative emo-tional experience. These results underscorehomologiesbetweenhumansandotheranimals.

Diencephalon

The hypothalamus (Hy) and thalamus consti-tute most of the diencephalon. Like the PAG,the Hy is a major player in animal models ofemotion. It governs the pituitary and therebythe body’s endocrine system; plays a major rolein the regulation of motivated behavior andhomeostatic processes (Sewards & Sewards,


2003; Valenstein, Cox, & Kakolewski, 1970);and interacts with the autonomic nervous sys-tem through large reciprocal connections withthe PAG and other brainstem nuclei (Saper,Loewy, Swanson, & Cowan, 1976). The lateralHy receives projections from diverse limbicstructures and projects to the midbrain (midb),middle hypothalamic zone, and medial Hy. Themiddle zone regulates autonomic functions andbottom-up forms of attention via connectionsto brainstem nuclei, including the PAG,reticular formation, parabrachial nucleus, ven-tral tegmental area (VTA), raphe nuclei, andspinal autonomic centers. And finally, the me-dial zone regulates endocrine function, such as

the release of cortisol during stress. As we showhere, activations in the human Hy and sur-rounding structures are reliable across studiesand show a preference in frquency of activationfor studies of positive emotional experience.

The thalamus is perhaps best known for itsrole in sensory processing, but it contains over30 distinct nuclei whose cortico-thalamic loopscover virtually the entire cortical mantle. Themediodorsal nucleus (MD) and the intralam-inar nuclei are most closely associated with af-fective processes. Human thalamic activationsare reliable in studies of emotion, and someparts show preference for emotional experi-ence.

15. The Neuroimaging of Emotion 251

TABLE 15.1. Abbreviations for Brain Regions, Organized by Anatomical Structure

Localization prefixes/suffixes

v ventrala anteriord dorsalr rostrals superiori inferiorfr frontallat. Lateralm or med. MedialBA Brodmann’s area

Brainstem

Midb midbrainPAG periaqueductal grayVTA ventral tegmental area

Diencephalon

Hy hypothalamus

ThalamusThal thalamusMD mediodorsal nucleusMGN medial geniculate nucleusCM centromedian nucleusSTN subthalamic nucleus

Subcortical telencephalon

Amy amygdalaHCMP hippocampusBF basal forebrain (cholinergic)BNST bed nucleus of the stria terminalisStr striatum (Cau/Put)Put putamenCau caudateGP globus pallidusNAC nucleus accumbens

Paralimbic

Ins or INS insulaAg agrancular region of insulamPFC medical prefrontal cortexvmPFC ventromedical prefrontal cortexOFC orbito-frontal cortexACC anterior cingulate cortexrdACC rostral dorsal anterior cingulatepgACC pregenual cingulatesgACC subgenual cingulateTP temporal polepHCMP parahippocampal cortexMTL medial temporal lobe

Other cortical regions

Lateral frontalIFG inferior frontal gyrusfrOP frontal operculumIFS inferior frontal sulcus

Medial walldmPFC dorsomedial prefrontal cortexPCC posterior cingulate cortexpre-SMA pre-supplementary motor area

TemporalMTL medial temporal lobeTC temporal cortexSTS sup. temporal culcusSTG sup. temporal gyrus

OccipitalOCC occipital cortexV1 primary visual cortex

Cerebellum

CB cerebellum

Subcortical Telencephalon

Overlying the brainstem and diencephalon area group of subcortical areas that are typicallyidentified as core limbic structures. These in-clude the amygdala, hippocampus, cholinergicbasal forebrain (BF) nuclei, and basal ganglia.The amygdala is well known for its role inemotion, particularly fear, though it also playsa prominent role in appetitive processes(Braesicke et al., 2005; Waraczynski, 2006).The basolateral complex plays a critical role infear conditioning—the learning of associationsbetween specific environmental cues and aver-sive outcomes (Anglada Figueroa & Quirk,2005; Davis, 1992; Goosens & Maren,2001; LeDoux, 2000; Nader, Majidishad,Amorapanth, & LeDoux, 2001). The centralnucleus, at the dorsal end of the amygdala, isimportant for the physiological and behavioralexpression of conditioned fear responses (Da-vis, 1992; Feldman, Conforti, & Saphier, 1990;Kalin, Shelton, & Davidson, 2004). However,the role of these structures in emotional experi-ence (e.g., the feeling of fear) is less certain, asfear-like responses to naturally threateningstimuli do not always require the amygdala(Davis & Lee, 1998; Walker & Davis, 1997;Wallace & Rosen, 2001). As we show below,imaging studies suggest that activations ob-served in the human amygdala are more likelyto relate to visual cues that signal affective sig-nificance than to those signaling negative expe-rience.

The amygdala, like many gross anatomicalstructures in the “affective brain,” plays rolesin both positive and negative affective pro-cesses. It is critical for the evaluation of sensorycues associated with reward (Cador, Robbins,& Everitt, 1989; Everitt, Cador, & Robbins,1989; Everitt et al., 1999) and the short-termupdating of the reward value of cues in context(Baxter & Murray, 2002; Schoenbaum, Chiba,& Gallagher, 1998; Schoenbaum, Setlow,Saddoris, & Gallagher, 2003). Recently Paton,Belova, Morrison, and Salzman (2006) showedthat separate populations of amygdala neuronsrespond to stimuli that predict positive andnegative future outcomes.

Taken together, all of this research suggeststhat the amygdala is important to the evalua-tion of sensory cues for relevance to the organ-ism, and directs an organism to learn moreabout a stimulus so as to better determine itspredictive value for well-being and survival

(Davis & Whalen, 2001; Kim, Somerville,Johnstone, Alexander, & Whalen, 2003;Whalen, 1998). This is consistent with the ideathat amygdala activations in human neuroim-aging studies are related to the salience or po-tential information value of visual stimuli(Amaral, 2003; Liberzon, Phan, Decker, &Taylor, 2003; Whalen et al., 2004).

Interspersed with cell groups in the extendedamygdala are a variety of cell groups spreadthroughout the BF. Some, such as the bed nu-cleus of the stria terminalis (BNST), are likelyto be important for fear and anxiety (Davis &Lee, 1998; Davis & Shi, 1999; Walker & Da-vis, 1997). Other nuclei serve as suppliers ofacetylcholine to the cortex and play a key rolein motivational modulation of attention(Sarter, Hasselmo, Bruno, & Givens, 2005) andsensory plasticity (Bear & Singer, 1986;Weinberger, 1995). These or other nuclei maybe important for human reward and pleasure:Early stimulation studies in humans suggestedthat stimulation of the septal region in the BFcan produce pleasurable responses (Heath,1972; Heath, Cox, & Lustick, 1974), thoughthese early observations may have actuallybeen related to seeking or appetitive behaviorrelated to stimulation of the visual striatumrather than to “pleasure” per se. We show be-low that the BF and septal regions are consis-tently activated in neuroimaging studies ofemotion, and that different parts of this regionare selective for positive and negative affect.Midline structures around the septal nuclei areactivated preferentially by studies of positiveemotional experience, whereas more lateral ar-eas around the BNST and extended amygdalaare activated preferentially by studies of nega-tive emotional experience.

The hippocampus, posterior to the amygdala,figures prominently in Gray’s (1978) affective/motivational theory of behavioral inhibition inanxiety. However, recent studies documentingthe role of the hippocampus in long-term mem-ory formation and consolidation (Squire &Zola-Morgan, 1991) have led researchers to sus-pect that its role in emotional behavior ismemory-related. Consistent with this view, thehippocampus is particularly important for con-textual fear conditioning in rodents (e.g.,Maren, Aharonov, & Fanselow, 1997). In hu-mans, hippocampal and medial temporal activa-tions are reliable in emotion-related tasks, butappear to be more related to perception of affec-tive stimuli than to emotional experience.


The basal ganglia are a set of subcorticalstructures that are critical for planning and ini-tiating motivationally relevant behaviors. Al-though they were once thought to be primarilyrelated to motor control, their functional role islikely to extend to the computation of affectivevalue in a more general sense. The striatum—consisting of the caudate, putamen, and nu-cleus accumbens (NAC)—and the globuspallidus (GP, not shown) constitute the majorpart of the basal ganglia. The ventral parts ofthe striatum—including NAC, ventral striatum(ventral caudate and putamen)—and ventralpallidum play important roles in motivation,reward, and learning. Along with VTA and lat-eral Hy, they form a network of regions rich indopamine and opioid receptors that might beconsidered the appetitive motivational “back-bone” of the brain (Berridge, 2004). Whereasthis system was originally thought to mediateprimary hedonic or “reward” responses, thereis now substantial evidence that dopamine sig-naling in this network—particularly in themesolimbic pathway from VTA to NAC—ismore closely related to the generation of moti-vated behavior than to “pleasure” per se(Berridge & Robinson, 1998; Salamone,Cousins, & Snyder, 1997). These areas aresome of the most frequently activated struc-tures in studies of human emotion, and differ-ent portions of these structures are preferen-tially activated by positive versus negativeemotional experience.

Paralimbic Cortex

The paralimbic “belt” is a set of phylogeneti-cally older cortex with large, direct projectionsto subcortical and brainstem nuclei. These in-clude the orbitofrontal cortex (OFC), rostralmPFC, aINS, and anterior temporal cortex(TC). Damage to OFC is associated with inap-propriate generation and regulation of affect,which may take the form of flattened affect, in-creased expression and reports of negativeemotion, or inappropriate emotion for the so-cial context, depending on the case (Beer,Heerey, Keltner, Scabini, & Knight, 2003;Berlin, Rolls, & Kischka, 2004; Hornak et al.,2003). Damage is also associated with reducedphysiological output (e.g., heart rate and skinconductance; Anderson, Damasio, Tranel, &Damasio, 2000; Angrilli, Palomba, Cantagallo,Maietti, & Stegagno, 1999; Roberts et al.,2004).

The ventral areas of the medial wall have di-rect projections to the Hy and lower brainstemautonomic effectors (Saper, 1995). Subgenualcingulate (sgACC) and ventromedial prefrontalcortex (vmPFC) are related to visceromotorcontrol in a number of animal studies (Vogt,Finch, & Olson, 1992), and subregions appearto play different and perhaps opposing roles inthe generation and regulation or extinction ofhypothalamic–pituitary–adrenal “stress” re-sponses (Sullivan & Gratton, 2002) and condi-tioned fear responses (Milad & Quirk, 2002).Rostral dorsal anterior cingulate (rdACC) andpregenual anterior cingulate (pgACC) havealso been associated with diverse affect-relatedfunctions, including maternal bonding, pain,and emotion; the rostral cingulate may be sub-divided into more rostral affect-related regionsand more posterior response-selection-relatedregions (Devinsky, Morrell, & Vogt, 1995;Vogt, Nimchinsky, Vogt, & Hof, 1995). In hu-man neuroimaging of emotion, distinct subre-gions of dorsal and ventral mPFC, rdACC,pgACC, and OFC are activated. DorsomedialPFC and multiple regions within OFC are se-lective for studies of emotional experience, butACC is not. Medial OFC and vmPFC are selec-tive for positive emotion.

aINS, shown in the top left portion of Plate15.1 (see color insert), is connected with di-verse subcortical “limbic” regions and projectsto brainstem autonomic centers. It has been as-sociated with interoception of affect-relatedbody states, including perception of pain anditch (Craig, 2002), and with visceromotor con-trol (Yasui, Breder, Saper, & Cechetto, 1991).The aINS can be divided based oncytoarchitecture and function into ventral anddorsal regions (Mesulam & Mufson, 1982).The evolutionarily older ventral portion,agranular insula (Ag), is a core paralimbic re-gion containing primary cortical regions forsensory–affective processing (taste and smell)and is particularly associated in human imag-ing studies with emotion (Wager & Barrett,2004). The dorsal region, by contrast, is acti-vated in a more diverse set of cognitive tasks; itcontains the operculo-insular junction, com-monly activated in tasks requiring the context-sensitive deployment of attention (Thompson-Schill, D’Esposito, Aguirre, & Farah, 1997;Wager, Reading, & Jonides, 2004). (In manycognitive studies, the area of operculo-insularactivation is referred to as the inferior frontalgyrus [IFG]). This distinction turns out to be


important for our analyses, as the human stud-ies we review show preference for positiveemotional experience in the ventral aINS (nearOFC and primary taste and smell cortices), andpreference for negative emotional experience inthe dorsal aINS regions (which are most oftenactivated by pain and tasks that elicit negativeemotions).

The medial and lateral anterior TC aredensely interconnected with OFC, and earlystudies of human TC stimulation producedparticularly strong and vivid emotional experi-ences (Sem-Jacobsen, 1968). These experiencesoften took the form of reliving a period of thepast (e.g., a scene from childhood) as thoughactually there, complete with emotive behaviorappropriate for the situation. We find reliableactivation of anterior TC in studies of emo-tional experience, with different portions pref-erentially activated by positive and negativeemotions.

Above, we have summarized some of thebroad roles thought to be played by the variousstructures that make up the affective brain.Next, we present the methods and results ofour meta-analysis in more detail. Though theresults are largely consistent with hypothesesbased on animal and lesion data, they providemaps of the human affective brain that in manycases indicate that different regions within thebroad structures discussed above may play dif-ferent and even opposing roles.

THE META-ANALYTIC APPROACH

In spite of the growing literature on the neuralbases of emotion, we still know very littleabout the precise functions of key brain regionsin the affective brain that instantiate the emo-tional lives of humans. A major difficulty in hu-man work is variability: Studies of what scien-tists assume are the same phenomenon (e.g.,the experience of fear) can produce widelyvarying results both across participants withina study and across studies. For the most part,this variability is treated as error, even thoughit may reflect real and important differences inbrain anatomy (at the individual participantlevel) or task requirements and therefore psy-chological process (at the level of the study).Single neuroimaging studies typically averagesignals across individuals to identify brain ar-eas that show consistent activity across a group

of participants. Similarly, with the use of meta-analytic techniques, we can average resultsacross studies to identify brain areas that showconsistent activity across a group of studies.Meta-analyses allow us not only to identifyconsistent brain–process correspondences (i.e.,the extent to which a brain area consistentlyshows increased activity with a psychologicalprocess), but also to identify the specificity ofsuch correspondences. Of course, many of thedetails and nuances of individual studies arelost in a meta-analytic approach, but itsstrength is that it allows us to view the affectivebrain painted in broad strokes, and provides uswith a means to address whether ask the majorpsychological categories that scientists typi-cally rely on—perception versus experience,positive versus negative affect, or varieties ofnegative emotion (e.g., sadness vs. anger vs.fear)—produce reliable and specific differencesin brain activity.

The Sample of Studies

We included findings from a total of 163 neu-roimaging studies (57 PET and 106 fMRI) onunmedicated, healthy adults published between1990 and 2005. These are summarized in Table15.2 and described in detail in Appendix 15.1.1

All studies were coded for whether they tar-geted “affect” (defined as a pleasant or un-pleasant state arising from presentation ofsurvival-related or social stimuli) or “emotion”(defined as instances of categories typically la-beled by the English words “anger,” “sadness,”“fear,” “disgust,” and “happiness”). We alsocoded other relevant study properties, such aswhether a study involved perception or experi-ence, whether positive or negative feelings wereevoked (for studies of experience), and whetherPET or fMRI methods were used.2 Papers werecoded first by a team of four trained raters.Each paper was coded by two different coders.We included only activations (omitting deacti-vations, because they were less consistently re-ported).

Multilevel Analytic Strategy:Peak Density Analysis

Meta-analyses of neuroimaging studies do notusually compute average effect sizes, as is donewith behavioral data, but instead summarizethe frequency with which studies report peakactivation coordinates in a particular brain lo-


cation (Fox, Parsons, & Lancaster, 1998; Lairdet al., 2005; Wager, Phan, Liberzon, & Taylor,2003; Wager, Reading, et al., 2004).3 Eachstudy reports one or more contrasts that mapthe difference in brain activity for two condi-tions (e.g., positive vs. neutral picture viewing).By convention, activated regions in each con-trast are summarized as coordinates in a stan-dardized, three-dimensional brain space di-vided in millimeters: x (left–right), y (back–front), and z (top–bottom) coordinates. In ameta-analysis, the brain is divided up into a setof three-dimensional cubic volumes (called“voxels”; each is 2 × 2 × 2 mm), and maps areconstructed of the density of reported activa-tion coordinates within a local volume (within10 mm) around each voxel.

Previous meta-analyses used activation coor-dinates as the unit of analysis. For example, if12 studies with one contrast map each reported18 peak coordinates within 10 mm of a voxelin frontal cortex, the density for that voxelwould be 18/(10 mm3). That value would becompared to the distribution of values ex-pected by chance in order to assess significance.Rather than treating individual peaks as theunit of analysis, our meta-analyses treat a con-trast as a random effect with activation coordi-nates nested within contrast.4 Thus the densitymeasure of interest is the number of contrasts(not the number of individual activation peaks)that produced activation near a voxel. In theexample above, there were 12 contrasts that re-ported activations near the frontal voxel of in-terest, which amounts to 12 nominally inde-pendent pieces of information, and our densitycount is 12/(10 mm3). Studies might reportmultiple nearby peaks for the same contrast be-cause of low spatial smoothness in the data, re-porting conventions, low thresholds, or evenvoxel sizes used. Our method is insensitive tothese types of variations across studies. As a re-sult, a single study can no longer dispropor-tionately contribute to the result by reportingmany nearby peaks in an area.

For the present dataset, we included 437contrasts, each associated with a set of re-ported coordinates. (Table 15.2 summarizesthe contrasts we analyzed.) Each set of coordi-nates was transformed into a map of “active”voxels that were within 10 mm of a reportedpeak for that contrast. We then computed asummary density map of the proportion ofcontrasts activating near each voxel by taking aweighted average of contrast activation maps.

The weight for each study was the square rootof the sample size, multiplied by an adjustmentweight for type of the analysis used for popula-tion inference.5 The values in this peak densitymap have a transparent interpretation: “Den-sity” refers to the number of contrasts or statis-tical parametric maps with a nearby (within 10mm) peak, weighted by the quality of informa-tion provided by the study.

Statistical Inference and Thresholding

The density map is then compared to MonteCarlo simulations to identify voxels with acti-vations that exceed the frequency expected bychance (i.e., a uniform distribution of activa-tion coordinates across the brain’s gray mat-ter). For the present dataset, using a MonteCarlo simulation in which the observed num-ber of activation coordinates from the 437 acti-vation maps were placed at random locationsthroughout the brain—and repeating this pro-cess 5,000 or more times—allowed us to deter-mine which locations in the brain showed agreater-than-chance number of nearby activa-tion coordinates, providing a stringent family-wise error rate correction for search across thelocations within the brain. Our approach wasfirst to locate regions that were consistently ac-tivated across a significant number of thesestudies (the “observed neural referencespace”), and then to analyze the likelihood ofactivating a region in relation to different cate-gories of emotional phenomena.

The maps in Plates 15.1, 15.2, and 15.3 (seecolor insert) and Figure 15.1 show voxels forwhich the density of reported peaks exceededthat expected by chance, corrected for searchacross the many voxels of the brain (i.e.,family-wise error rate correction). The yellowcolor in the plates shows regions in which thepeak density is high enough that the null-hypothesis chance of finding a single significantvoxel anywhere in the gray matter of the brainis p < .05. The other colors use an “extent-based” threshold, in which the number of con-tiguous voxels above a primary threshold (e.g.,p < .001) are counted and compared with thenumber of contiguous voxels expected bychance (Friston, Worsley, Frackowiak,Mazziotta, & Evans, 1994). Colored regionsshow areas in which a cluster of voxels thislarge is unlikely to occur by chance anywherein the brain. We have used the Monte Carlosimulation to establish extent thresholds at p <


.001 (orange in the plates) and p < .005 (pinkin the plates). We report the significance ofvoxels at the highest primary threshold forwhich the significance criteria are met.

Visualization and Localization

The three-dimensional illustrations of brainslices presented in this chapter were recon-structed from a canonical MRI image(colin27.img, the single-subject template inSPM2; www.fil.ion.ucl.ac.uk/spm/software/spm2).This brain was coregistered with the interna-tional standard Montreal Neurologic Institute(MNI) brain template (avg152T1.img), whichis itself based on the average of 152 brains reg-istered roughly to landmarks from the atlas ofTalairach and Tourneaux (1988). To localizehighly replicable regions, we overlaid signifi-cant voxels on the MNI average template anddetermined their locations by using the atlases

of Duvernoy (Duvernoy, 1995; Duvernoy &Bourgouin, 1999), Martin (1996), Haines(2000), and Ongur, Ferry, and Price (2003);this method provided more accurate resultsthan automated labeling systems. We use thesingle-subject brain only for visualization inour illustrations, because its anatomical detailmakes brain landmarks more identifiable toreaders. We do not report Brodmann’s areas(BAs) because their boundaries cannot be iden-tified with sufficient accuracy on this templatebrain (unless they are in regions shown inOngur et al., 2003, who provide labels regis-tered to the MNI brain). Although variationacross labs and software packages in nominallysimilar warping to “Talairach space” producesinconsistencies among reported coordinates(Brett, Johnsrude, & Owen, 2002), the MNItemplate brain is the most popular template forelectronic registration, so using it minimizes lo-calization errors in the meta-analysis results.


TABLE 15.2. Summary of Contrasts Analyzed

Experience versus perception

Experience Perception Mixed

PET 109 28 11fMRI 131 129 29

Affective/emotional valence

Negative Positive Mixed/nonspecific

PET 94 30 24fMRI 187 65 37

Specific emotion

Happiness Anger Disgust Fear Sadness Mixed/other

PET 17 10 4 11 23 9fMRI 19 16 40 57 22 34

Correction for multiple comparisons

Unknown Corrected Small-volume corrected Uncorrected

PET 7 21 0 121fMRI 8 73 28 186

Population inference

No (infer on sample only) Yes (infer on population)

PET 124 24fMRI 74 215

Note. A summary of the 437 contrast maps from 163 studies used in the meta-analysis. Numbers reflect the number of con-trasts in each category. “Population inference” refers to the number of contrasts that treated subject as a random effect, allow-ing valid population inference (“yes” in the table), as compared with contrasts that performed a “fixed-effects” analysis andwhose results cannot be generalized beyond the sample studied. Lower weights were given to “fixed-effects” and smaller-sam-ple studies in the meta-analysis.

THE OBSERVEDNEURAL REFERENCE SPACE FOREMOTION-RELATED PHENOMENA

Plate 15.1 summarizes the regions that wereconsistently activated in our database of neuro-imaging studies. (Stereotactic coordinates forthe most consistent activation foci are listed inAppendix 15.2.) The right lateral surface of thebrain (upper left panel), the left medial surface(upper right panel), the cerebellum (CB) (pur-ple shading, lower left panel), and the brain-stem and prominent subcortical regions (lowerright panel) are shown in three-dimensionalrendering.

There was remarkable consistency betweenthe hypothetical and observed neural referencespaces for emotion-related phenomena. As ex-pected, we observed consistent activity in orclosely associated with “limbic” areas. A strik-ing feature of the map is the inclusion ofdiencephalic and brainstem regions that havebeen identified in animal models of affectivebehavior, but are infrequently discussed in hu-man neuroimaging studies. In the brainstem,PAG and VTA were consistently activated,whereas lower brainstem centers in the pons andmedulla were not consistently activated (thoughsome pontine activations were consistent inother analyses presented below). Consistentlyactivated diencephalic regions included thesubthalamic nucleus (STN), Hy, and much of thedorsal thalamus, though the maximal consis-tency in the thalamus was in the central medialzone, around the “limbic” MD andcentromedian (CM) nuclei. In the telen-cephalon, large significant regions of activationwere observed in and around the amygdala ex-tending into the BF, NAC, hippocampus, andvStr and vGP. In the paralimbic belt, consistentfindings included vmPFC (10 m/r), multiple lat-eral OFC sites (42/12l and m; 13l), the aINS, andthe medial and lateral anterior TC (pHCMP andtemporal pole [TP]). Cingulate activations werelargely limited to the rostral half of the ACC,corresponding to the so-called “affective” zone(Bush, Luu, & Posner, 2000). Strikingly, how-ever, activations in the medial wall were clus-tered into at least three distinct groups, corre-sponding to pgACC, rACC, and sgACC.

The area of superior dmPFC above thecingulate sulcus (BA 9 extending back to BA32) was consistently activated and distinctfrom ACC activation. The functional contribu-tions of dmPFC have yet to be precisely deter-

mined, but recent research and theorizing sug-gest that these brain areas contribute jointly tomaking mental state attributions (for reviews,see Adolphs, 2001; Blakemore, Winston, &Frith, 2004; Lane & McRae, 2004)—such aswhen a person makes judgments about the psy-chological states of another person, or moni-tors, introspects, or makes inferences about hisor her own moment-to-moment feelings (seealso Mitchell, Banaji, & Macrae, 2005;Ochsner et al., 2004).

We also observed consistent activations inregions not traditionally considered part of theneural reference space for emotion, includingthose in the lateral frontal cortex, TC, occipitalcortex (OCC), and cerebellum. These findingssuggest that additional psychological processesinvolved in emotion-related phenomena maybe overlooked in existing neuroscience models.

Although individual studies have reportedlateral frontal activations in locations that spanthe expanse of cortex (see, e.g., Figure 15.1A),the only consistent activations across all studieslie in the bilateral IFG, extending from the parsopercularis (Broca’s area, BA 44) through parstriangularis (BA 45) and pars orbitalis on theinferior frontal convexity (BA 47/lateral 12).The activated region also extended into thefrontal operculum and was contiguous with ac-tivation in OFC and aINS. Neuroimaging stud-ies of response inhibition, response selection,task switching, and working memory havecommonly activated the area around BA 44–45and the underlying operculum (Badre,Poldrack, Pare-Blagoev, Insler, & Wagner,2005; Gabrieli, Poldrack, & Desmond, 1998;Martin & Chao, 2001; Poldrack et al., 1999;Wager, Jonides, Smith, & Nichols, 2005; Wag-ner, Maril, Bjork, & Schacter, 2001). Meta-analyses have suggested that the frontaloperculo-insular border contiguous with thedorsal aINS, rather than the lateral surface, isthe area most consistently activated acrossstudies (Wager, Reading, et al., 2004; Wager &Smith, 2003). A general role for BA 44–45 andthe operculum might be context-based selec-tion among competing stimulus–responsemappings or sets (Thompson-Schill et al.,1997). In emotion-related phenomena, this re-gion may be important for the information se-lection processes critical for conceptual pro-cessing associated with meaning analysis (suchas that associated with appraisal). In support ofthis notion, emotion- and pain-related activityin IFG and the operculum is modified by both


manipulations of the meaning context in whichaffective stimuli are presented (Benedetti,Mayberg, Wager, Stohler, & Zubieta, 2005;Kong et al., 2006; Wager, Rilling, et al., 2004)and voluntary regulation of emotional re-sponses (Ochsner, Bunge, Gross, & Gabrieli,2002).

Activation of superior temporal sulcus/supe-rior temporal gyrus (STS/STG) and inferiortemporal and occipital “association” corticescould be related to enhanced sensory integra-tion, but their precise role in affective process-ing is unclear. Although the function of the pos-terior cingulate cortex (PCC) remains unclear(Maddock, 1999), it may play a role inmemory-guided representation of context im-portant for conceptual processing in emotion(Maddock, 1999; Mantani, Okamoto, Shirao,Okada, & Yamawaki, 2005; Minoshima et al.,1997). Primary visual cortex (V1) also showedconsistent activation, suggesting that early vi-sual processing is enhanced when compared toneutral control conditions.

The consistent cerebellar activation we ob-served might be related to increased demandson motor planning during affective and emo-tional states, but there is accumulating evi-dence for a more direct cerebellar role inemotion-related phenomena. Electrical stimu-lation of deep cerebellar nuclei in humans, par-ticularly the fastigial nucleus, has been shownto induce activity in mesolimbic affect-relatedareas (Heath, Dempesy, Fontana, & Myers,1978), and in some cases it has elicited states ofprofound rage (Heath et al., 1974). Conversely,cerebellar damage often produces what mightbe considered a disorder of emotion regulation,characterized by fluctuations between flattenedaffect and inappropriate social behaviors (e.g.,“trying to kiss the experimenter”; Schmah-mann & Sherman, 1998) reminiscent of the so-cial and emotional deficits with OFC damage.The cerebellum is connected with specificprefrontal regions in topographically mappedreciprocal circuits (Middleton & Strick, 1994,2000) and with “limbic” regions, including theHy (Haines & Dietrichs, 1984), OFC (BA 12),dmPFC (BA 9 and BA 32), portions of IFG (BA46), and inferior frontal convexity (BA 46/12)(Middleton & Strick, 2001). Cerebellarefferents to these areas pass largely throughDM in the thalamus, which we also find is con-sistently activated in human emotion. One hy-pothesis is that the cerebellum might contributeto the processing of situational context

(Schmahmann & Sherman, 1998) as part of acomplex pattern recognition system (Albus,1971).

COMPARING EXPERIENCEAND PERCEPTION

Studies of human affect and emotion (and theirassociated contrast maps) can be categorized asinvestigations of “experience” (involving thegeneration of feelings in response to pictures,sounds, memories, imagery, or other stimuli) or“perception” (involving observation and judg-ment about the normative content in a stimu-lus, such as whether a picture of facial behavioris classified as an “expression” that depicts“fear,” “anger,” or “disgust”). The critical con-ceptual difference separating experience fromperception is whether a contrast (comparingactivity in an experimental vs. a control condi-tion) captures activity related to the generationof subjective feelings. Of course, this distinc-tion is often a matter of degree, because thepassive viewing of some stimuli that are used inperception-based studies (such as photos fromthe International Affective Picture System) willevoke an affective response regardless ofwhether participants are asked to report it, andstimuli used in experience-based studies (suchas memories of social situations or sounds) re-quire the perception of emotional content. Toperform this analysis, however, we distin-guished contrasts where affective experienceswere being generated from those wherechanges in the perceiver’s affect was unlikely.

Studies of experience in our sample (see Ap-pendix 15.1) typically involved the recall ofpersonal experiences, the viewing of stronglyevocative visual or auditory stimuli, or expo-sure to pleasant or unpleasant tastes or odors.Studies of perception involved judgmentsabout visual stimuli such as facial expressionsthat were unlikely to produce a strong changein experience (though such stimuli can serve asprimes that may influence subsequent behav-ior). Papers with ambiguous status were ex-cluded from this analysis.

Meta-analysis density maps showing signifi-cant differences between experience and percep-tion are shown in Plates 15.2A and 15.2B. Somesignificant regions are shown on the same axialbrain slices for direct comparison in Plate 15.2C,with experience versus perception in red andperception versus experience in blue. The den-


sity maps identified brain locations where therewas a relative difference in peak activity betweenthe experience and perception studies. Becausethe null hypothesis assumed that points fromeach condition were uniformly distributedthroughout the brain, this analysis shows brainlocations where differences in the relative distri-bution of peaks is large.6

The results show a striking dissociation.Contrasts for the experience of affect or emo-tion showed relatively greater activation for themedial subcortex (including Hy, VTA, PAG,dorsal pons, and surrounding reticular forma-tion), the diencephalon (including the BF/ex-tended amygdala, Hy, and thalamus), areas ofparalimbic cortex (OFC, aINS, vmPFC, andTP), MTL, dmPFC, ventral IFG, and the deepregions of the CB surrounding the deep cerebel-lar nuclei. These findings are consistent withother work suggesting prominent roles forOFC (Kringelbach, 2005), aINS (Craig, Chen,Bandy, & Reiman, 2000), and dmPFC (Barrettet al., 2007) in the experience of affect andemotion. Moreover, they are consistent withstimulation studies in humans. Numerous casereports by Sem-Jacobsen (1968) and othersshow effects on experience of both positive andnegative emotions with stimulation of the dor-sal and orbital medial wall, and vivid emo-tional memories with anterior and medial tem-poral stimulation. Experience-related increasesin subcortical and brainstem areas attest to thereliability of subcortical activation in neuroim-aging studies, although they are less frequentlydiscussed than cortical activations. Interest-ingly, rdACC did not differentiate between ex-perience and perception; it was activated byboth types of studies.

Perception-related contrasts more consis-tently activated amygdala, pHCMP, pgACC,dorsal IFG, inferior TC and OCC, and lateralcerebellum. These findings are consistent withthe idea that amygdala activations in humanneuroimaging studies are related to the salienceor potential information value of visual stimuli(Liberzon et al., 2003), rather than playingsome necessary role in emotional experience(Anderson & Phelps, 2002). Finding inferiortemporal and occipital specialization for per-ception does not come as a surprise, given thestrong relationship between these areas andvisuospatial processing. Dorsal IFG and lateralcerebellum specialization for perception mayrelate to pattern recognition and conceptualprocessing that is necessary for normal emo-

tion perception. These regions are connected intopographically mapped loops (Middleton &Strick, 2001), suggesting that they may be partof the same functional circuit. Of course, suchconclusions about the precise function of thesecircuits are speculative, and they remain to betested in focused individual studies. What isstriking, however, is that different regions ofthe IFG seem to be differentially involved in ex-perience and perception: the ventral andopercular parts for experience, and the moredorsal part for perception.

COMPARING PLEASANTAND UNPLEASANT EXPERIENCES

Using only the 240 contrasts (in 95 studies) in-volving experience, we next examined whetherpleasant and unpleasant experiences were im-plemented in separable distributed systems. Aprevious meta-analysis tackled the positive–negative distinction (Wager et al., 2003), butdid not separate experience from perception, aswe do here.

Reported activation peaks for unpleasant(blue) and pleasant (yellow) contrasts areshown in Plates 15.3A and 15.3B, and meta-analysis density maps showing relative differ-ences are shown in Plates 15.3C and 15.3D,with pleasant versus unpleasant in yellow andunpleasant versus pleasant in blue. Overall, theresults suggest dissociations based on valencethat are in general agreement with hypothe-sized specializations based on animal and le-sion studies. Pleasant experiences were associ-ated with relatively greater activation in medialdopamine-rich areas (VTA, NAC, and portionsof vStr), as well as in Hy, vmPFC, and rightOFC. Unpleasant experiences were associatedwith more consistent activation in amygdala,aINS, PAG, left OFC, and more posterior por-tions of vStr and vGP. The results provide apromising indication that different gross ana-tomical areas may be differentially sensitive topleasant and unpleasant stimuli, although theydo not imply that activation in any of these re-gions is uniquely associated with either cate-gory. In fact, chi-square tests revealed no regionwith greater absolute proportions of pleasantactivations, primarily because unpleasant expe-riences elicit more robust responses that engagemany parts of the brain. The relative specializa-tion we report here is important, however, be-cause studies of animal models have found that


neurons and nuclear groups that are specializedaccording to valence are often contained withinthe same gross anatomical structure—for ex-ample, intermixed positive and negativeexpected-value neurons within the amygdala(Paton et al., 2006), or rostrocaudal negative–positive gradients within the NAC (Reynolds& Berridge, 2002). This raises the question ofwhether neuroimaging truly provides the reso-lution to separate representations of positiveand negative affect. Our results are in line withthe idea that different brain regions have differ-ent relative concentrations of specialized neu-rons, and that large-scale structure (aggregatedacross millions of neurons) in the affect systemcan be detected via neuroimaging.

An additional point worth noting is thatwhile we did not compare left and right hemi-spheres using direct statistical contrasts, we didobserve a pattern of lateralization in the OFCand basal forebrain (and a lack of lateralizationin superior lateral cortex) that was differentfrom that predicted by previous theories.Studies based on electroencephalography(EEG; for a review, see, e.g., Davidson, 2000),and lesion studies linking left-hemisphere dam-age to depression (Borod, 1992), suggest thatthe left lateral prefrontal cortex supports pleas-ant moods and reactions to pleasant stimuli,whereas the right lateral prefrontal cortex sup-ports unpleasant moods processes. It is possiblethat lateralized EEG activity is predictive ofmood or affective style, but the pattern of spe-cific brain activation underlying those corticalpotentials is more complex. Alternatively,arousal differences across studies might masklateralized differences (Canli, Desmond, Zhao,Glover, & Gabrieli, 1998), or lateralizationmight be more closely related to approach–avoidance motivation than to affective valenceper se, as suggested by our previous meta-analyses and recent work (Pizzagalli,Sherwood, Henriques, & Davidson, 2005;Wager et al., 2003). These caveats notwith-standing, the neuroimaging correlates of ob-served lateralization of affect in EEG studies re-main to be elucidated.

THE SEARCH FOR SUBCORTICALCIRCUITS: COMPARING PETAND fMRI STUDIES

Most emotion-related research on rats and pri-mates pinpoints midbrain and brainstem areas

as important for emotional behavior, and thusfar our meta-analyses suggest that these areasare particularly active during the experience ofaffect and emotion. Compared to PET imaging,fMRI is less well suited to the study of the basaltelencephalon and brainstem, due to magneticsusceptibility artifacts.7 The hypothesis, then, isthat fMRI studies of emotion-related phenom-ena may underestimate midbrain and brain-stem contributions to affect and emotion.Alternatively, however, these issues may becounterbalanced by fMRI’s greater spatio-temporal precision and potential to collectlarger amounts of data.

Using chi-square analysis, we compared den-sity maps for contrasts from PET and fMRIstudies to determine whether one imagingmethod was more likely to activate particularbrain regions in absolute rather than relativeterms. The analysis tested the proportion ofcontrasts using each method that activatedwithin 10 mm of each voxel, and thus con-trolled for overall differences in the frequencyof use of PET and fMRI. The map of regionswith significant chi-square statistics, shown inFigure 15.1, reveals no significant differencesbetween PET and fMRI in the brainstem, BF,vStr, or OFC. Thus the benefits of fMRI in spa-tial resolution may compensate in part for in-creased artifacts, and thus fMRI can be a usefultool for examining the subcortex. Superior cor-tical regions, as well as amygdala and someother subcortical regions, appeared to be moreconsistent in fMRI studies (white in Figure15.1)—though we suspect that effects inamygdala in particular may be related to thewidespread use of a priori amygdala regions ofinterest, rather than to inherently more reliableactivation.

GENERAL DISCUSSION

The meta-analyses reported in this chapterbuild on previously published metaanalyses ofemotion (Murphy et al., 2003; Phan et al.,2002; Wager et al., 2003) in important ways.First, whereas those analyses combined percep-tion and experience of emotion and affect, weexplicitly compared the two types of studiesand found that they were distinguished by theirrelative concentration of peak activations:brainstem, hypothalamic, and paralimbic selec-tivity for experience, and amygdalar complexand posterior cortex selectivity for perception.


Second, we compared studies of pleasant andunpleasant experiences; we found selectivityfor pleasant experience in midline brainstem,hypothalamic, and ventromedial frontal re-gions, and selectivity for negative experience indistinct brainstem (PAG), insular, striatal, andorbital cortical regions. Third, we explicitly ex-amined the suitability of fMRI as a method forinvestigating brainstem and basal telenceph-alon; we found that fMRI may be comparableto PET in its effectiveness as a tool for studyingthe brainstem and subcortex.

Of course, there is still much to be done.First, neuroimaging studies must continue tomove away from referring to broad anatomicalregions that in actuality perform different andoften opposing functions. For example, theACC encompasses about 15,000 2 × 2 × 2voxels of brain tissue, but it is common for re-searchers to compare results from differentstudies at the gross structural level, trying to fita single common interpretation to results thatare in different areas with different anatomicalprojections and functional profiles. Using dataacross many studies to demarcate regions of

the brain precisely, as we have done here, is anessential step toward building a more system-atic method. Once a set of regions is identifiedfor study, it becomes possible to examine theirdynamics (e.g., via connectivity analyses andstructural models) in a much more meaningfulway.

Second, in interpreting findings from ourmeta-analyses (or from any single neuroimag-ing study, for that matter), it is important tokeep in mind that brain areas that span even afew millimeters are probably not consistentlydedicated to any one process. Even individualneurons may participate in a number of func-tional circuits, and fMRI measures activityintegrated over populations of neurons in-volved in different processes,8 including differ-ent types of affect (Paton et al., 2006). Theseminal work of Sem-Jacobsen (1968) in hu-mans, and of Valenstein, Cox, and Kakolewski(1968) in rats, has shown strikingly that stimu-lation of brain sites very close together (nomore than a few millimeters) can elicit vastlydifferent emotional responses. In the words ofSem-Jacobsen (1976), “An electrode 0.5–1 cm


FIGURE 15.1. (A) Reported activation coordinates for contrasts using PET (black circles) and fMRI(white triangles) on the right lateral surface. (B) Meta-analysis results showing significant differences inabsolute proportion of PET versus fMRI studies (black) and fMRI versus PET studies (white). The analy-sis used chi-square tests, controlling for the frequency of use of each method. (C) The comparisons in (C)shown on brain slices to reveal subcortical locations. Abbreviations are defined in the text and in Table15.1.

from a positive point may give the opposite[emotional] response with about the samestrength. There appears to be this dual arrange-ment in the ventromedial area of the frontallobe, the central part of the temporal lobe, aswell as other structures” (p. 516). In support ofthis notion, our summary of neuroimagingstudies shows nearby regions with differentfunctional specialization in the basal telen-cephalon, ventral basal ganglia, and inferiorfrontal cortex.

Furthermore, in early human electrical stim-ulation studies of emotion, although stimula-tion of the same site was often found to repro-duce an emotional experience in the samesession, elapsed time or variation of the behav-ioral contexts in which stimulation occurredmarkedly affected the emotional response.Documentation of this phenomenon in animalsled Valenstein et al. (1970) to question the ideaof fixed affective circuits. Thus, even with therelative precision of chronically implanted elec-trodes, the claim of Heraclitus that “you cannever step into the same river twice” (Plato,360 B.C.E.)may well apply to the emotionalbrain.

All these considerations lead us to believethat two kinds of paradigm shifts are essentialfor advancing the neuroimaging of emotion.One is that researchers need to move fromstudying brain areas in isolation to identifyinginterconnected, distributed circuits. Functionalconnectivity analysis may provide more preciseinformation about brain processing related tovarious affective states and events, and confirmor constrain inferences about regional activa-tions based on patterns of connectivity. In thisrespect, neuroimaging plays a unique and com-plementary role to lesion studies in animals, be-cause neuroimaging alone allows the simulta-neous measurement of the entire brain anddynamic patterns of functional connectivityacross diverse systems. Another is that re-searchers should move beyond mapping brainresponses to individual psychological phenom-ena to making inferences about psychologicalstates based on brain activity. Neuroimagingprovides unique and valuable informationabout the organization of the human brain;what is at stake here is the ability to learnabout the organization of the mind from braindata. Making psychological inferences frombrain activity is an extremely difficult task, andfunctional inferences in neuroimaging studieshave often been made in an ad hoc fashion.

Valid psychological inference requires compar-ing activations across a number ofpsychological states; meta-analyses of the typewe report here are one way to perform suchcomparisons. Formal inferences about psycho-logical states from neuroimaging data can bemade by using classifier systems, and this is apromising new direction.

Overall, the past 15 years have seen an ex-plosion in the application of brain imagingmethods to emotional phenomena. The emerg-ing field of affective neuroscience has been en-gaged in a search for answers to two interde-pendent questions: the locations of brainregions that represent affective information,and the psychological distinctions that definethe conditions for their activation. As informa-tion is accumulated, specific mappings can bemade among studies of nondisabled humans,studies of humans with psychological andbrain pathology, and animal models. Neuroim-aging paves the way for synergy across thesepreviously quite disparate fields by referring tothe common “language” of the brain. Whendata are aggregated across neuroimaging stud-ies, there is a remarkable and somewhatunderappreciated consistency between animaland human stimulation work. As such data areaccumulated, more elaborated and refinedmappings between brain activity and affectiveprocesses will yield yet greater synergy acrossthe neurosciences.

NOTES

1. In addition, there are certain methodological limita-tions to single neuroimaging studies that can be over-come with meta-analytic summaries. For example,current neuroimaging techniques are plagued by bothlow power and the presence of many false positives.Only 20% of the studies sampled corrected for multi-ple comparisons (i.e., tests of more than one brain re-gion). Of these, many used inappropriate methodsthat did not provide adequate correction, yieldingfalse-positive results.

2. Although emotions are the subset of affective re-sponses that are elaborated with various sources ofconceptual content (Barrett et al., 2007), we treatedaffect and emotion as separate categories in our anal-yses.

3. Effect size meta-analyses have been performed, butare problematic in general due to inconsistencies indata analysis across studies that can influence effectsizes. In addition, the peak density method avoids theneed to estimate effect sizes in regions for which ef-fects are not reported.


4. For simplicity, we assume that different contrasts re-ported by the same study are independent.

5. Larger studies have been weighted more heavily. Thesquare root transformation provides a measure closerto effect size. Studies that used “random-effects”models appropriate for population inference wereweighted 1.25 more heavily than those treating par-ticipants as a fixed effect. “Fixed-effects” modelswere commonly used in early neuroimaging studies,but are not appropriate for generalizing to a popula-tion. They generally produced much higher effectsizes (e.g., Z-scores) for the same data.

6. This analysis does not provide information aboutwhether more studies of experience (or perception)activate an area in absolute terms. Information aboutthe absolute frequencies of activation is provided by 2

(chi-square) analyses that test whether significantlymore activation maps from experience studies versusperception studies activate an area. These analysesare available from Wager and Barrett, and were notincluded here because of space limitations; the resultsagree in large part with the areas shown here. Also,these maps do not provide direct inferences about thelikelihood of experience (or perception), given activa-tion in a particular area (Poldrack, 2006). Density-based maps of relative distributions are very useful,however, because they control for the overall fre-quency of activation across conditions, and thereforeallow for more subtle differences to emerge. An ex-ample of this occurs in our analysis of pleasant versusunpleasant affective feelings.

7. Typical blood-oxygen-level-dependent (BOLD) fMRImeasures functional activity by being sensitive to lo-cal field inhomogeneities (Ogawa et al., 1992). A dif-ficulty is that transitions from air sinus space to tissuearound the base of the brain create local fieldinhomogeneities, resulting in both signal loss and dis-tortion, which limit both sensitivity and localization.

8. Counting neurons and synapses is difficult, but toprovide a general idea of neural connectivity, someestimates are about 13.7 × 109 neocortical neurons(Braendgaard, Evans, Howard, & Gundersen, 1990)and 164 × 1012 synapses (Tang, Nyengaard, DeGroot, & Gundersen, 2001), yielding an average ofnearly 12,000 synapses per neuron. Many cortical ar-eas have about 70,000–100,000 neurons per 1 mm2

of tissue, an area substantially smaller than a voxel inhuman neuroimaging studies.

REFERENCES

Adolphs, R. (2001). The neurobiology of social cogni-tion. Current Opinion in Neurobiology, 11(2), 231–239.

Albus, J. S. (1971). A theory of cerebellar function.Mathematical Biosciences, 10, 25–61.

Amaral, D. G. (2003). The amygdala, social behavior,and danger detection. Annals of the New York Acad-emy of Sciences, 1000, 337–347.

Amaral, D. G., Price, J. L., Pitkanen, A., & Carmichael,S. T. (1992). Anatomical organization of the primateamygdaloid complex. In J. P. Aggleton (Ed.), Theamygdala: Neurobiological aspects of emotion,memory, and mental dysfunction (pp. 1–66). NewYork: Wiley-Liss.

Anderson, A. K., & Phelps, E. A. (2002). Is the humanamygdala critical for the subjective experience ofemotion?: Evidence of intact dispositional affect inpatients with amygdala lesions. Journal of CognitiveNeuroscience, 14(5), 709–720.

Anderson, S. W., Damasio, H., Tranel, D., & Damasio,A. R. (2000). Long-term sequelae of prefrontal cor-tex damage acquired in early childhood. Develop-mental Neuropsychology, 18(3), 281–296.

Anglada Figueroa, D., & Quirk, G. J. (2005). Lesions ofthe basal amygdala block expression of conditionedfear but not extinction. Journal of Neuroscience,25(42), 9680–9685.

Angrilli, A., Palomba, D., Cantagallo, A., Maietti, A.,& Stegagno, L. (1999). Emotional impairment af-ter right orbitofrontal lesion in a patient with-out cognitive deficits. NeuroReport, 10(8), 1741–1746.

Badre, D., Poldrack, R. A., Pare-Blagoev, E. J., Insler, R.Z., & Wagner, A. D. (2005). Dissociable controlledretrievel and generalized selection mechanisms inventrolateral prefrontal cortex. Neuron, 47, 907–918.

Bandler, R., & Shipley, M. T. (1994). Columnar organi-zation in the midbrain periaqueductal gray: Modulesfor emotional expression? Trends in Neurosciences,17(9), 379–389.

Barrett, L. F., Mesquita, B., Ochsner, K. N., & Gross, J.J. (2007). The experience of emotion. Annual Reviewof Psychology, 58, 7.1–7.31.

Baxter, M. G., & Murray, E. A. (2002). The amygdalaand reward. Nature Reviews Neuroscience, 3(7),563–573.

Bear, M. F., & Singer, W. (1986). Modulation of visualcortical plasticity by acetylcholine and nor-adrenaline. Nature, 320(6058), 172–176.

Beer, J. S., Heerey, E. A., Keltner, D., Scabini, D., &Knight, R. T. (2003). The regulatory function of self-conscious emotion: insights from patients withorbitofrontal damage. Journal of Personality and So-cial Psychology, 85(4), 594–604.

Benedetti, F., Mayberg, H. S., Wager, T. D., Stohler, C.S., & Zubieta, J. K. (2005). Neurobiological mecha-nisms of the placebo effect. Journal of Neurosciences,25(45), 10390–10402.

Berlin, H. A., Rolls, E. T., & Kischka, U. (2004). Impul-sivity, time perception, emotion and reinforcementsensitivity in patients with orbitofrontal cortex le-sions. Brain, 127(Pt. 5), 1108–1126.

Berridge, K. C. (2004). Motivation concepts in behav-ioral neuroscience. Physiology and Behavior, 81(2),179–209.

Berridge, K. C., & Robinson, T. E. (1998). What is therole of dopamine in reward: Hedonic impact, reward


learning, or incentive salience? Brain Research Re-views, 28(3), 309–369.

Blakemore, S.-J., Winston, J., & Frith, U. (2004). Socialcognitive neuroscience: Where are we heading?Trends in Cognitive Sciences, 8, 216–222.

Borod, J. C. (1992). Interhemispheric and intra-hemispheric control of emotion: A focus on unilat-eral brain damage. Journal of Consulting and Clini-cal Psychology, 60(3), 339–348.

Braendgaard, H., Evans, S. M., Howard, C. V., &Gundersen, H. J. (1990). The total number of neu-rons in the human neocortex unbiasedly estimatedusing optical disectors. Journal of Microscopy,157(Pt. 3), 285–304.

Braesicke, K., Parkinson, J. A., Reekie, Y., Man, M.-S.,Hopewell, L., Pears, A., et al. (2005). Autonomicarousal in an appetitive context in primates: A behav-ioural and neural analysis. European Journal of Neu-roscience, 21(6), 1733–1740.

Brett, M., Johnsrude, I. S., & Owen, A. M. (2002). Theproblem of functional localization in the humanbrain. Nature Reviews Neuroscience, 3(3), 243–249.

Bush, G., Luu, P., & Posner, M. I. (2000). Cognitive andemotional influences in anterior cingulate cortex.Trends in Cognitive Sciences, 4(6), 215–222.

Cador, M., Robbins, T. W., & Everitt, B. J. (1989). In-volvement of the amygdala in stimulus–reward asso-ciations: Interaction with the ventral striatum. Neu-roscience, 30(1), 77–86.

Canli, T., Desmond, J. E., Zhao, Z., Glover, G., &Gabrieli, J. D. (1998). Hemispheric asymmetry foremotional stimuli detected with fMRI. NeuroReport,9(14), 3233–3239.

Craig, A. D. (2002). How do you feel? Interoception:The sense of the physiological condition of the body.Nature Reviews Neuroscience, 3(8), 655–666.

Craig, A. D., Chen, K., Bandy, D., & Reiman, E. M.(2000). Thermosensory activation of insular cortex.Nature Neuroscience, 3, 184–190.


Davidson, R. J. (2000). Anterior electrophysiologicalasymmetries, emotion, and depression: Conceptualand methodological conundrums. Psychophysiology,35(5), 607–614.

Davis, M. (1992). The role of the amygdala in fear andanxiety. Annual Review of Neuroscience, 15(1), 353–375.

Davis, M., & Lee, Y. L. (1998). Fear and anxiety: Possi-ble roles of the amygdala and bed nucleus of the striaterminalis. Cognition and Emotion, 12(3), 277–305.

Davis, M., & Shi, C. (1999). The extended amygdala:Are the central nucleus of the amygdala and the bednucleus of the stria terminalis differentially involvedin fear versus anxiety? Annals of the New York Acad-emy of Sciences, 877, 281–291.

Davis, M., & Whalen, P. J. (2001). The amygdala: Vigi-lance and emotion. Molecular Psychiatry, 6(1), 13–34.

Devinsky, O., Morrell, M. J., & Vogt, B. A. (1995).Contributions of anterior cingulate cortex to behav-iour. Brain, 118(Pt. 1), 279–306.

Dewey, J. (1895). The theory of emotion: 2. The signifi-cance of emotions. Psychological Review, 2, 13–32.

Duvernoy, H. M. (1995). The human brain stem andcerebellum: Surface, structure, vascularization, andthree-dimensional sectional anatomy with MRI.Vienna: Springer-Verlag.

Duvernoy, H. M., & Bourgouin, P. (1999). The humanbrain: Surface, three-dimensional sectional anatomywith MRI, and blood supply. Wien, Vienna: Springer-Verlag.

Edelman, G. M., & Tononi, G. (2001). A universe ofconsciousness: How matter becomes imagination.New York: Basic Books.

Everitt, B. J., Cador, M., & Robbins, T. W. (1989). In-teractions between the amygdala and ventralstriatum in stimulus–reward associations: studies us-ing a second-order schedule of sexual reinforcement.Neuroscience, 30(1), 63–75.

Everitt, B. J., Parkinson, J. A., Olmstead, M. C., Ar-royo, M., Robledo, P., & Robbins, T. W. (1999).Associative processes in addiction and reward. Therole of amygdala–ventral striatal subsystems. An-nals of the New York Academy of Sciences, 877,412–438.

Feldman, S., Conforti, N., & Saphier, D. (1990). Thepreoptic area and bed nucleus of the stria terminalisare involved in the effects of the amygdala on adreno-cortical secretion. Neuroscience, 37(3), 775–779.

Fox, P. T., Parsons, L. M., & Lancaster, J. L. (1998). Be-yond the single study: function/location metanalysisin cognitive neuroimaging. Current Opinion in Neu-robiology, 8(2), 178–187.

Friston, K. J., Worsley, K. J., Frackowiak, R. S. J.,Mazziotta, J. C., & Evans, A. C. (1994). Assessingthe significance of focal activations using their spatialextent. Human Brain Mapping, 1, 210–220.

Gabrieli, J. D., Poldrack, R. A., & Desmond, J. E.(1998). The role of left prefrontal cortex in languageand memory. Proceedings of the National Academyof Sciences of the USA, 95(3), 906–913.

Goosens, K. A., & Maren, S. (2001). Contextual andauditory fear conditioning are mediated by the lat-eral, basal, and central amygdaloid nuclei in rats.Learning and Memory, 8(3), 148–155.

Gray, J. A. (1978). The neuropsychology of anxiety.British Journal of Psychology, 69(4), 417–434.

Gregg, T. R., & Siegel, A. (2001). Brain structures andneurotransmitters regulating aggression in cats: Im-plications for human aggression. Progress in Neuro-psychopharmacology and Biological Psychiatry,25(1), 91–140.

Haines, D. E. (2000). Neuroanatomy: An atlas of struc-tures, sections, and systems. Philadelphia: LippincottWilliams & Wilkins.

Haines, D. E., & Dietrichs, E. (1984). An HRP study ofhypothalamo-cerebellar and cerebellohypothalamicconnections in squirrel monkey (Saimiri sciureus).


Journal of Comparative Neurology, 229(4), 559–575.

Heath, R. G. (1972). Pleasure and brain activity in man.Deep and surface electroencephalograms during or-gasm. Journal of Nervous and Mental Disease,154(1), 3–18.

Heath, R. G., Cox, A. W., & Lustick, L. S. (1974). Brainactivity during emotional states. American Journal ofPsychiatry, 131(8), 858–862.

Heath, R. G., Dempesy, C. W., Fontana, C. J., & Myers,W. A. (1978). Cerebellar stimulation: Effects onseptal region, hippocampus, and amygdala of catsand rats. Biological Psychiatry, 13(5), 501–529.

Holstege, G., & Georgiadis, J. R. (2004). The emotionalbrain: Neural correlates of cat sexual behavior andhuman male ejaculation. Progress in Brain Research,143, 39–45.

Hornak, J., Bramham, J., Rolls, E. T., Morris, R. G.,O’Doherty, J., Bullock, P. R., et al. (2003). Changesin emotion after circumscribed surgical lesions of theorbitofrontal and cingulate cortices. Brain, 126(Pt.7), 1691–1712.

Irons, D. (1897). The nature of emotion. PhilosophicalReview, 6(3), 242–256.

James, W. (1884). What is an emotion? Mind, 9(34),188–205.

Kalin, N. H., Shelton, S. E., & Davidson, R. J. (2004).The role of the central nucleus of the amygdala inmediating fear and anxiety in the primate. Journal ofNeuroscience, 24(24), 5506–5515.

Kim, H., Somerville, L. H., Johnstone, T., Alexander, A.L., & Whalen, P. J. (2003). Inverse amygdala and me-dial prefrontal cortex responses to surprised faces.NeuroReport, 14(18), 2317–2322.

Kong, J., Gollub, R. L., Rosman, I. S., Webb, J. M.,Vangel, M. G., Kirsch, I., et al. (2006). Brain activityassociated with expectancy-enhanced placebo anal-gesia as measured by functional magnetic resonanceimaging. Journal of Neuroscience, 26(2), 381–388.

Kringelbach, M. L. (2005). The human orbitofrontalcortex: Linking reward to hedonic experience. Na-ture Reviews Neuroscience, 6(9), 691–702.

Laird, A. R., Fox, P. M., Price, C. J., Glahn, D. C.,Uecker, A. M., Lancaster, J. L., et al. (2005). ALEmeta-analysis: Controlling the false discovery rateand performing statistical contrasts. Human BrainMapping, 25(1), 155–164.

Lane, R. D., & McRae, K. (2004). Neural substrates ofconscious emotional experience: A cognitive-neuroscientific perspective. In M. Beauregard (Ed.),Consciousness, emotional self-regulation and thebrain (pp. 87–122). Amsterdam: Benjamins.

LeDoux, J. E. (2000). Emotion circuits in the brain. An-nual Review of Neuroscience, 23(1), 155–184.

Liberzon, I., Phan, K. L., Decker, L. R., & Taylor, S. F.(2003). Extended amygdala and emotional salience:A PET activation study of positive and negative af-fect. Neuropsychopharmacology, 28(4), 726–733.

Lovick, T. A. (1992). Inhibitory modulation of the car-diovascular defence response by the ventrolateral

periaqueductal grey matter in rats. ExperimentalBrain Research, 89(1), 133–139.

Maddock, R. J. (1999). The retrosplenial cortex andemotion: New insights from functional neuroimagingof the human brain. Trends in Neurosciences, 22(7),310–316.

Mantani, T., Okamoto, Y., Shirao, N., Okada, G., &Yamawaki, S. (2005). Reduced activation of poste-rior cingulate cortex during imagery in subjects withhigh degrees of alexithymia: A functional magneticresonance imaging study. Biological Psychiatry, 57,982–990.

Maren, S., Aharonov, G., & Fanselow, M. S. (1997).Neurotoxic lesions of the dorsal hippocampus andPavlovian fear conditioning in rats. Behavioral BrainResearch, 88(2), 261–274.

Martin, A., & Chao, L. L. (2001). Semantic memoryand the brain: Structure and processes. CurrentOpinion in Neurobiology, 11, 194–201.

Martin, J. H. (1996). Neuroanatomy: Text and atlas(2nd ed.). Stamford, CT: Appleton & Lange.

McDonald, A. J. (1998). Cortical pathways to the mam-malian amygdala. Progress in Neurobiology, 55(3),257–332.

Mesulam, M. M., & Mufson, E. J. (1982). Insula of theOld World monkey: I. Architectonics in the insulo-orbito-temporal component of the paralimbic brain.Jounral of Comparative Neurology, 212(1), 1–22.

Middleton, F. A., & Strick, P. L. (1994). Anatomical evi-dence for cerebellar and basal ganglia involvement inhigher cognitive function. Science, 266(5184), 458.

Middleton, F. A., & Strick, P. L. (2000). Basal gangliaand cerebellar loops: Motor and cognitive circuits.Brain Research: Brain Research Reviews, 31(2–3),236–250.

Middleton, F. A., & Strick, P. L. (2001). Cerebellar pro-jections to the prefrontal cortex of the primate. Jour-nal of Neuroscience, 21(2), 700.

Milad, M. R., & Quirk, G. J. (2002). Neurons in medialprefrontal cortex signal memory for fear extinction.Nature, 420(6911), 70–74.

Minoshima, S., Giordani, B., Berent, S., Frey, K. A., Fos-ter, N. L., & Kuhl, D. E. (1997). Metabolic reductionin the posterior cingulate cortex in very early Alzhei-mer’s disease. Annals of Neurology, 42(1), 85–94.

Mitchell, J. P., Banaji, M. R., & Macrae, C. N. (2005).General and specific contributions of the medialprefrontal cortex to knowledge about mental states.NeuroImage, 28, 757–762.

Murphy, F. C., Nimmo-Smith, I., & Lawrence, A. D.(2003). Functional neuroanatomy of emotions: Ameta-analysis. Cognitive, Affective, and BehavioralNeuroscience, 3(3), 207–233.

Nader, K., Majidishad, P., Amorapanth, P., & LeDoux,J. E. (2001). Damage to the lateral and central, butnot other, amygdaloid nuclei prevents the acquisitionof auditory fear conditioning. Learning and Memory,8(3), 156–163.

Ochsner, K. N., Bunge, S. A., Gross, J. J., & Gabrieli, J.D. (2002). Rethinking feelings: An FMRI study of the


cognitive regulation of emotion. Journal of CognitiveNeuroscience, 14(8), 1215–1229.

Ochsner, K. N., Knierim, K., Ludlow, D. H., Hanelin, J.,Ramachandran, T., Glover, G., et al. (2004). Re-flecting upon feelings: An fMRI study of neural sys-tems supporting the attribution of emotion to selfand other. Journal of Cognitive Neuroscience,16(10), 1746–1772.

Ogawa, S., Tank, D. W., Menon, R., Ellermann, J. M.,Kim, S., Merkle, H., et al. (1992). Intrinsic signalchanges accompanying sensory stimulation: Func-tional brain mapping with magnetic resonance imag-ing. Proceedings of the National Academy of Sci-ences of the USA, 89(13), 5951–5955.

Ongur, D., An, X., & Price, J. L. (1998). Prefrontal cor-tical projections to the hypothalamus in macaquemonkeys. Jounrla of Comparative Neurology,401(4), 480–505.

Ongur, D., Ferry, A. T., & Price, J. L. (2003). Architec-tonic subdivision of the human orbital and medialprefrontal cortex. Journal of Comparative Neurol-ogy, 460(3), 425–449.

Ongur, D., & Price, J. L. (2000). The organization ofnetworks within the orbital and medial prefrontalcortex of rats, monkeys and humans. Cerebral Cor-tex, 10(3), 206–219.

Panksepp, J. (1998). Affective neuroscience: The foun-dations of human and animal emotions. New York:Oxford University Press.

Paton, J. J., Belova, M. A., Morrison, S. E., & Salzman,C. D. (2006). The primate amygdala represents thepositive and negative value of visual stimuli duringlearning. Nature, 439(7078), 865–870.

Phan, K. L., Wager, T., Taylor, S. F., & Liberzon, I.(2002). Functional neuroanatomy of emotion: Ameta-analysis of emotion activation studies in PETand fMRI. NeuroImage, 16(2), 331–348.

Pizzagalli, D. A., Sherwood, R. J., Henriques, J. B., &Davidson, R. J. (2005). Frontal brain asymmetry andreward responsiveness: A source-localization study.Psychological Science, 16(10), 805–813.

Poldrack, R. A. (2006). Can cognitive processes be in-ferred from neuroimaging data? Trends in CognitiveSciences, 10(2), 59–63.

Poldrack, R. A., Wagner, A. D., Prull, M. W., Desmond,J. E., Glover, G. H., & Gabrieli, J. D. (1999). Func-tional specialization for semantic and phonologicalprocessing in the left inferior prefrontal cortex.NeuroImage, 10(1), 15–35.

Reynolds, S. M., & Berridge, K. C. (2002). Positive andnegative motivation in nucleus accumbens shell: Bi-valent rostrocaudal gradients for GABA-elicited eat-ing, taste “liking”/”disliking” reactions, place prefer-ence/avoidance, and fear. Journal of Neuroscience,22(16), 7308–7320.

Roberts, N. A., Beer, J. S., Werner, K. H., Scabini, D.,Levens, S. M., Knight, R. T., et al. (2004). The impactof orbital prefrontal cortex damage on emotional ac-tivation to unanticipated and anticipated acoustic

startle stimuli. Cognitive, Affective, and BehavioralNeuroscience, 4(3), 307–316.

Salamone, J. D., Cousins, M. S., & Snyder, B. J. (1997).Behavioral functions of nucleus accumbens dopa-mine: Empirical and conceptual problems with theanhedonia hypothesis. Neuroscience and Bio-behavioral Reviews, 21(3), 341–359.

Saper, C. B. (1995). Central autonomic system. In G.Paxinos (Ed.), The rat nervous system (2nd ed., pp.107–135). San Diego: Academic Press.

Saper, C. B., Loewy, A. D., Swanson, L. W., & Cowan,W. M. (1976). Direct hypothalamoautonomic con-nections. Brain Research, 117(2), 305–312.

Sarter, M., Hasselmo, M. E., Bruno, J. P., & Givens, B.(2005). Unraveling the attentional functions of corti-cal cholinergic inputs: Interactions between signal-driven and cognitive modulation of signal detection.Brain Research: Brain Research Reviews, 48(1), 98–111.

Schmahmann, J. D., & Sherman, J. C. (1998). The cere-bellar cognitive affective syndrome. Brain, 121(4),561–579.

Schoenbaum, G., Chiba, A. A., & Gallagher, M. (1998).Orbitofrontal cortex and basolateral amygdala en-code expected outcomes during learning. NatureNeuroscience, 1(2), 155–159.

Schoenbaum, G., Setlow, B., Saddoris, M. P., &Gallagher, M. (2003). Encoding predicted outcomeand acquired value in orbitofrontal cortex during cuesampling depends upon input from basolateralamygdala. Neuron, 39(5), 855–867.

Sem-Jacobsen, C. W. (1968). Depth-electrographic stim-ulation of the human brain and behavior: From four-teen years of studies and treatment of Parkinson’sdisease and mental disorders with implanted elec-trodes. Springfield, IL: Thomas.

Sem-Jacobsen, C. W. (1976). Electrical stimulation andself-stimulation in man with chronic implanted elec-trodes. Interpretation and pitfalls of results. In A.Wauquier & E. T. Rolls (Eds.), Brain-stimulation re-ward (pp. 505–520). New York: Elsevier.

Sewards, T. V., & Sewards, M. A. (2003). Representa-tions of motivational drives in mesial cortex, medialthalamus, hypothalamus and midbrain. Brain Re-search Bulletin, 61(1), 25–49.

Shipley, M. T., Ennis, M., Rizvi, T. A., & Behbehani, M.M. (1991). Topographical specificity of forebrain in-puts to the midbrain periaqueductal gray: Evidencefor discrete longitudinally organized input columns.In A. Depaulis & R. Bandler (Eds.), The midbrainperiaqueductal gray matter: Functional, anatomical,and neurochemical organization (pp. 417–448). NewYork: Plenum Press.

Squire, L. R., & Zola-Morgan, S. (1991). The medialtemporal lobe memory system. Science, 253(5026),1380.

Sullivan, R. M., & Gratton, A. (2002). Prefrontal cor-tical regulation of hypothalamic–pituitary–adrenalfunction in the rat and implications for psychopath-


ology: Side matters. Psychoneuroendocrinology,27(1–2), 99–114.

Talairach, J., & Tournoux, P. (1988). Co-planarstereotaxic atlas of the human brain: 3Dimensionalproportional system. An approach to cerebral imag-ing. Stuttgart: Thieme.

Tang, Y., Nyengaard, J. R., De Groot, D. M. G., &Gundersen, H. J. G. (2001). Total regional and globalnumber of synapses in the human brain neocortex.Synapse, 41(3), 258–273.

Thompson-Schill, S. L., D’Esposito, M., Aguirre, G. K.,& Farah, M. J. (1997). Role of left inferior prefrontalcortex in retrieval of semantic knowledge: A reevalu-ation. Proceedings of the National Academy of Sci-ences of the USA, 94(26), 14792–14797.

Valenstein, E. S., Cox, V. C., & Kakolewski, J. W.(1968). Modification of motivated behavior elicitedby electrical stimulation of the hypothalamus. Sci-ence, 159(3819), 1119.

Valenstein, E. S., Cox, V. C., & Kakolewski, J. W. (1970).Reexamination of the role of the hypothalamus in mo-tivation. Psychological Review, 77(1), 16–31.

Van der Horst, V. G., & Holstege, G. (1998). Sensoryand motor components of reproductive behavior:Pathways and plasticity. Behavioural Brain Research,92(2), 157–167.

Vogt, B. A., Finch, D. M., & Olson, C. R. (1992). Func-tional heterogeneity in cingulate cortex: The anteriorexecutive and posterior evaluative regions. CerebralCortex, 2(6), 435.

Vogt, B. A., Nimchinsky, E. A., Vogt, L. J., & Hof, P. R.(1995). Human cingulate cortex: Surface features,flat maps, and cytoarchitecture. Journal of Compara-tive Neurology, 359(3), 490–506.

Wager, T. D., & Barrett, L. F. (2004). From affect tocontrol: Functional specialization of the insula in mo-tivation and regulation. Available online at www.co-lumbia.edu/cu/psychology.tor/

Wager, T. D., Jonides, J., Smith, E. E., & Nichols, T. E.(2005). Towards a taxonomy of attention-shifting:Individual differences in fMRI during multiple shifttypes. Cognitive, Affective, and Behavioral Neurosci-ence, 5(2), 127–143.

Wager, T. D., Phan, K. L., Liberzon, I., & Taylor, S. F.(2003). Valence, gender, and lateralization of func-tional brain anatomy in emotion: A meta-analysis offindings from neuroimaging. NeuroImage, 19(3),513–531.

Wager, T. D., Reading, S., & Jonides, J. (2004). Neuro-imaging studies of shifting attention: A meta-analysis. NeuroImage, 22(4), 1679–1693.

Wager, T. D., Rilling, J. K., Smith, E. E., Sokolik, A.,Casey, K. L., Davidson, R. J., et al. (2004).Placebo-induced changes in FMRI in the anticipa-tion and experience of pain. Science, 303(5661),1162–1167.

Wager, T. D., & Smith, E. E. (2003). Neuroimagingstudies of working memory: A metaanalysis. Cogni-tive, Affective, and Behavioral Neuroscience, 3(4),255–274.

Wagner, A. D., Maril, A., Bjork, R. A., & Schacter, D. L.(2001). prefrontal contributions to executive control:fMRI evidence for functional distinctions within lat-eral Prefrontal cortex. NeuroImage, 14(6), 1337–1347.

Walker, D. L., & Davis, M. (1997). Double dissociationbetween the involvement of the bed nucleus of thestria terminalis and the central nucleus of theamygdala in startle increases produced by condi-tioned versus unconditioned fear. Journal of Neuro-science, 17(23), 9375–9383.

Wallace, K. J., & Rosen, J. B. (2001). Neurotoxic le-sions of the lateral nucleus of the amygdala de-crease conditioned fear but not unconditioned fearof a predator odor: comparison with electrolyticlesions. Journal of Neuroscience, 21(10), 3619–3627.

Waraczynski, M. A. (2006). The central extendedamygdala network as a proposed circuit underlyingreward valuation. Neuroscience and BiobehavioralReviews, 30, 472–496.

Weinberger, N. M. (1995). Dynamic regulation of recep-tive fields and maps in the adult sensory cortex. An-nual Review of Neuroscience, 18(1), 129–158.

Whalen, P. J. (1998). Fear, vigilance, and ambiguity: Ini-tial neuroimaging studies of the human amygdala.Current Directions in Psychological Science, 7(6),177–188.

Whalen, P. J., Kagan, J., Cook, R. G., Davis, F. C., Kim,H., Polis, S., et al. (2004). Human amygdalaresponsivity to masked fearful eye whites. Science,306(5704), 2061.

Yasui, Y., Breder, C. D., Saper, C. B., & Cechetto, D. F.(1991). Autonomic responses and efferent pathwaysfrom the insular cortex in the rat. Journal of Com-parative Neurology, 303(3), 355–374.


268

APPENDIX 15.1. Studies Included in the Meta-Analysis

First author Year Imaging Sex N Valence

Emotion Induction method

Aff Ang Disg Fear Hap Sad Vis Aud T/Olf Rec Img

Studies of emotional experience

Schafer 2005 fMRI X 40 Neg ° ° °Grimm 2005 fMRI X 29 Neg ° °Hutcherson 2005 fMRI F 28 Neg ° °Cato 2004 fMRI X 26 Neg ° °Eugene 2003 fMRI F 20 Neg ° °Lang 1998 fMRI X 20 Neg ° °Levesque 2003 fMRI F 20 Neg ° °Stark 2003 fMRI X 19 Neg ° ° °Aron 2005 fMRI X 17 Pos ° °Simpson 2000 fMRI X 17 Neg ° °Anderson 2003 fMRI X 16 Neg ° °Dolcos 2004 fMRI F 16 Neg ° °Gottfried 2002 fMRI X 15 Pos ° °Stark 2005 fMRI X 15 Neg ° °Canli 1998 fMRI F 14 Pos ° °Goel 2001 fMRI X 14 Pos ° °Fulbright 1998 fMRI X 13 Neg ° °Goldin 2005 fMRI F 13 Pos ° ° °Heinzel 2005 fMRI X 13 X ° °Markowitch 2003 fMRI X 13 Pos ° °Moll 2005 fMRI X 13 Neg ° ° °Northoff 2004 fMRI X 13 X ° °Shirao 2005 fMRI F 13 Neg ° °Elliott 2000 fMRI X 12 Pos ° ° °Fitzgerald 2004 fMRI X 12 Neg ° °Maratos 2001 fMRI X 12 Pos ° ° °Schienle 2006 fMRI F 12 Neg ° ° °Schienle 2002 fMRI F 12 Neg ° °DeAraujo 2003 fMRI X 11 Pos ° °Hariri 2003 fMRI X 11 Neg °Beauregard 2001 fMRI M 10 Pos ° °Canli 2000 fMRI F 10 Neg ° °Klein 2003 fMRI F 10 Neg ° °Lee 2004 fMRI X 10 Pos ° °Maddock 1997 fMRI X 10 Neg ° °Ruby 2004 fMRI M 10 X ° °Wrase 2003 fMRI F 10 Neg ° ° °Yamasaki 2002 fMRI X 10 Neg ° °Kringelbach 2003 fMRI M 9 Pos ° °O’Doherty (b) 2001 fMRI X 9 Neg ° °Small 2003 fMRI X 9 Neg ° °Whalen 1998 fMRI M 8 Neg ° ° °Wright 2004 fMRI X 8 Neg °Beauregard 1998 fMRI X 7 Neg ° °Moll 2002 fMRI X 7 Neg ° °O’Doherty (a) 2001 fMRI ? 7 Neg ° °Phana 2004 fMRI X 7 Neg ° ° °Bystritsky 2001 fMRI X 6 Neg ° °Herpetz 2001 fMRI F 6 Neg ° ° °Nitschke 2004 fMRI F 6 Pos ° ° °Teasdale 1999 fMRI X 6 Pos °Francis 1999 fMRI ? 4 Pos ° °Lorberbaum 1999 fMRI F 4 Neg ° °Damasio 2000 PET X 25a Pos ° ° ° ° °George 1994 PET X 21 Neg ° °Kimbrell 1999 PET X 16 Neg ° °Paradiso 2003 PET X 17 Neg ° ° ° °Pietrini 2000 PET X 15 Neg ° ° °

(continued)

269

APPENDIX 15.1. (continued)




Taylor 2000 PET X 14 Neg ° °Lane (a) 1997 PET F 12 Pos ° °Lane 1998 PET F 12 X ° ° °Partiot 1995 PET ° 12 Neg ° °Reiman 1997 PET F 12 X ° ° °Zaid 1997 PET F 12 Neg ° °Aalto 2002 PET F 11 Neg ° ° °Aalto 2005 Pet F 11 Neg ° °Gemar 1996 PET M 11 Neg ° °George 1995 PET F 11 Neg ° °Lane (c) 1997 PET F 11 Pos ° ° °Baker 1997 PET M 10 Pos ° ° °Beauregard 1997 PET M 10 X ° °Blood 2001 PET X 10 Pos ° °Blood 1999 PET X 10 Pos ° °Dolan 2000 PET M 10 X ° °George 1996 PET M 10 Pos ° ° °Lane (b) 1997 PET M 10 X ° ° ° °Liberzon 2000 PET F 10 Neg ° °Liberzon 2000 PET F 10 Neg ° °Liberzon 2003 PET X 10 Neg ° °Taylor 2003 PET X 10 Neg ° °Redoute 2000 PET M 9 Pos ° °Zald 1998 PET F 9 Pos ° °Doughtery 1999 PET M 8 Neg ° °Liotti 2000 PET F 8 Neg ° °Mayberg 1999 PET F 8 Neg °Ottowitz 2004 PET F 8 Neg ° °Paradiso 1997 PET X 8 Pos ° °Rauch 1999 PET M 8 Pos ° °Shin 2000 PET M 8 Neg ° °Taylor 1998 PET F 8 Neg ° °Kosslyn 1996 PET M 7 Neg ° °Pardo 1993 PET X 7 Neg ° °Fischer 1996 PET X 6 Neg ° °Isenberg 1999 PET X 6 Neg ° ° ° ° °Lane 1999 PET M 6 X ° °

Studies of mixed or ambiguous perception/experience

Habel 2005 fMRI M 26 Pos ° °Kuchinke 2005 fMRI X 20 Pos ° °Crosson 1999 fMRI X 17 X ° °Wicker 2003 fMRI M 14 Neg ° °Hariri 2002 fMRI X 12 Neg ° °Zatorre 2000 fMRI X 12 X °Rolls 2003 fMRI X 11 Neg ° °Buchanan 2000 fMRI M 10 X ° ° ° °Tabert 2001 fMRI F 9 Neg ° °Paradiso 1999 PET X 17 Pos ° °Royet 2000 PET M 12 X ° °Royet 2001 PET M 12 X ° °Zatorre 2000 PET X 12 X °Frey 2000 PET F 11 Neg ° °

Studies of emotional perception

Das 2005 fMRI X 28 Neg ° °Tessitore 2005 fMRI X 27 X ° °

(continued)

270

APPENDIX 15.1. (continued)




Liddell 2005 fMRI X 25 Neg ° ° °Fischer 2004 fMRI X 24 Neg ° °Williams, L. 2004 fMRI X 22 Neg ° °Kesler-West 2001 fMRI X 21 Neg ° ° °Pessoa 2002 fMRI X 21 Neg ° ° °Fitzgerald 2005 fMRI X 20 Pos ° ° ° ° ° °Grobras 2005 fMRI X 20 Neg ° °Schroeder 2004 fMRI X 20 Neg °Hariri 2000 fMRI X 16 X ° ° °Somerville 2004 fMRI X 16 Pos ° ° ° °Fecteau 2005 fMRI X 15 X °Grandjean 2005 fMRI X 15 Neg ° °Reinders 2005 fMRI X 15 Neg ° °Gur 2002 fMRI X 14 X ° ° ° °Shin 2005 fMRI M 13 Neg ° °Williams, L. 2005 fMRI X 13 Neg ° ° ° °Williams, M. 2005 fMRI X 13 X ° °Dolan 2001 fMRI X 12 Neg ° ° °Iidaka 2001 fMRI X 12 Neg °Killgore 2004 fMRI F 12 X ° ° °Strange 2000 fMRI X 12 Neg ° °Vuilleumier 2001 fMRI X 12 Neg ° °Wang 2005 fMRI X 12 Neg ° °Adams 2003 fMRI X 11 Neg °Williams, L. 2001 fMRI M 11 Neg ° °Breiter 1996 fMRI M 10 Pos ° ° °Gorno-Tempini 2001 fMRI X 10 Pos ° ° °Hare 2005 fMRI X 10 Pos °Mccullough 2005 fMRI X 10 X ° °Sato 2004 fMRI X 10 Neg ° ° °Wildgruber 2005 fMRI X 10 X °Critchley 2000 fMRI X 9 X °Lange 2003 fMRI ? 9 Neg ° °Nomura 2004 fMRI X 9 Neg ° °Dolan 1996 fMRI M 8 Pos ° °Narumoto 2000 fMRI X 8 X ° °Phillips 2004 fMRI M 8 Neg ° ° °Phillips (b) 1998 fMRI X 8 Pos ° °Whalen 2001 fMRI X 8 Neg ° ° °Phillips 1997 fMRI X 7 Neg ° ° °Phillips (a) 1998 fMRI M 6 Neg ° ° °Sprengelmeyer 1998 fMRI X 6 Neg ° ° °Blair 1999 PET M 13 Neg ° ° °George 1996 PET X 13 X °Kilts 2003 PET X 13 Pos ° ° °George 1993 PET F 9 X °Pourtois 2005 PET M 8 Neg ° °Sergent 1994 PET M 8 X ° °Nakamura 1999 PET M 7 X °Imaizumi 1997 PET M 6 X ° °Morris 1999 PET M 6 X ° °Morris 1996 PET X 5 Pos ° ° ° °Morris 1998 PET X 5 X ° ° °

Totals 77 17 20 37 23 26 112 19 15 17 9

Note. Emotions: Aff, affect (see text); Ang, anger; Disg, disgust; Fear, fear; Hap, happiness; Sad, sadness. Induction methods: Vis, vi-sual; Aud, auditory; T/Olf, tactile/olfactory; Rec, recall; Img, mental imaging. X in Sex and Valence columns refers to “mixed” sex orvalence.aDamasio (2000) had differing numbers of subjects for each emotion, ranging from 16 for the lowest and 25 for the highest.

271

APPENDIX 15.2. Stereotactic Coordinates for Activation Foci

Name Lat X Y Z Vol %Act Thr Name Lat X Y Z Vol %Act Thr

Brainstem Paralimbic

dPAG/SC — 2 –30 –6 24 6 P aIns R 42 24 –8 1712 18 P

DiencephalonaIns R 44 16 –2 2032 17 PaIns/frOP L –40 24 –6 4248 25 P

STN R 12 –18 –6 280 14 P vaIns (ag), TP L –28 6 –22 2344 25 PThal (DM, CM) — 0 –16 4 1376 14 P mid-Ins R 40 6 0 48 7 PSTN R 16 –8 0 768 15 E OFC (47/12m, 12l) L –28 34 –18 8 4 PHy R 8 –10 –12 864 13 E pgACC — 2 40 2 168 8 PHy L –10 0 –14 512 13 E dACC — 0 24 32 64 8 P

Subcortical telencephalonrdACC — –4 34 18 8 4 PTP (lat.) R 50 8 –26 176 6 P

Amy/BF R 20 –4 –20 5744 25 P TP (med.) R 38 12 –24 280 11 PAmy/BF L –20 –6 –18 6800 30 P aIns L –34 12 –10 1160 19 EHCMP/vGP R 22 –12 –12 1944 21 P aIns L –32 12 0 1224 15 EvStr/BF L –30 2 –12 1832 26 P vaIns (Ag)/TP L –40 10 –20 2040 18 EvStr/vGP R 26 0 –10 1704 24 P vIns R 40 4 –14 1040 17 EvGP/vIns L –30 –8 –10 1984 24 P mid-Ins R 44 –4 0 1576 9 EvGP/Hy/STN L –10 –6 –6 1104 19 P mid-Ins L –40 0 –2 1752 15 EAmy (BL) R 30 –2 –28 576 15 E mid-Ins (dors.) R 42 4 10 736 12 EHCMP L –18 –14 –24 584 15 E vmPGC (24, 10m/r) — 2 32 –4 1672 11 EHCMP L –30 –12 –24 408 12 E dmPFC (BA 9) — –4 52 30 648 11 PvStr/HCMP R 32 –10 –8 624 14 E dmPFC (BA 9/32) — 0 52 20 24 5 PvGP R 14 6 –8 1208 17 E OFC (47/12m,l), vaIns R 36 26 –16 2160 14 evGP L –22 8 –12 432 13 E OFC (47/12L), IFG L –44 26 –16 1304 14 ePut R 26 2 0 1936 15 E Uncus R 12 –2 –26 480 12 EPut L –28 0 0 1008 14 E Uncus L –12 0 –26 440 13 ePara-HCMP L –28 –24 –18 200 5 E

Other cortical regions

IFG/frOP R 48 22 12 1744 19 PfrOP L –46 16 0 552 12 PIFS L –48 22 18 24 5 PIFS L –46 10 24 48 6 PPCC — 2 –52 24 32 6 PTC/OCC R 52 –58 8 280 9 PiTC L –36 –50 –20 8 5 PiTC/sCBLM R 40 –54 –22 2016 15 PiTC/sCBLM L –38 –64 –18 936 10 PSTS (post.) R 54 –46 10 720 10 POCC (lat.) R 48 –70 –2 128 9 POCC (lat.) L –48 –70 10 48 5 PiOCC L –32 –78 –18 16 6 PIFG L –50 24 6 400 10 EPCC R 6 –60 18 176 6 Epre-SMA — 4 12 48 1208 8 Epre-SMA — 0 10 60 1280 8 ETC/OCC R 44 –60 –14 792 12 ESTG L –48 –8 –2 1096 10 EV1 (BA 17) L –6 –88 0 1120 9 E

Note. Lat, lateralization; [X, Y, Z] are stereotactic coordinates; Vol, volume of region in mm3; %Act, percentage of contrast maps ac-tivating within 10 mm; Thr, threshold type; P, primary height threshold; E, extent threshold.

C H A P T E R 1 6

Interoception and EmotionA Neuroanatomical Perspective

A. D. (BUD) CRAIG

When someone asks you, “How do you feel?,”your answer includes your bodily feelings aswell as your emotional feelings. Influentialmodels of emotion view the embodiment ofemotions as fundamental; that is, many observ-ers believe that interoceptive awareness has acrucial role in emotional awareness (Barrett,Quigley, Bliss-Moreau, & Aronson, 2004;Damasio, 1994; James, 1890/2007; Philippot& Chalelle, 2002; Wiens, 2005). The neuro-anatomical findings described in this chapteraddress the organization of pathways in thebrain that engender feelings from the body, andthey support the view that the neural substratesresponsible for subjective awareness of youremotional state are based on the neural repre-sentation of your physiological state.

These data show that there is a phylogeneti-cally novel homeostatic sensory afferent path-way in primates, especially in humanoid pri-mates, that provides the basis for a sense of thephysiological condition of the body (“the mate-rial me”) in posterior insular cortex (Craig,2002). A posterior-to-anterior progression ofincreasingly complex re-representations in the

human insula provides a foundation for thesequential integration of your homeostatic con-dition with your sensory environment, withyour motivational condition, and with your so-cial condition. Based on considerable evidence,I propose that this integration generates a uni-fied penultimate metarepresentation of theglobal emotional moment at the junction of theanterior insula and the frontal operculum(Craig, 2004a, 2005). Convergent functionalimaging findings reveal that the anterior insulaand the anterior cingulate cortices are con-jointly activated during all human emotions(e.g., Murphy, Nimmo-Smith, & Lawrence,2003), which in my view indicates that thelimbic sensory representation of subjective“feelings” (in the anterior insula) and thelimbic motor representation of volitionalagency (in the anterior cingulate) together formthe fundamental neuroanatomical basis for allhuman emotions (Craig, 2002). This view isconsistent with the definition of an emotion inhumans as both a feeling and a motivation withconcomitant autonomic sequelae (Rolls, 1999).In this view, emotions are not simply occa-

272

sional events, but rather are ongoing and con-tinuous, and emotional behaviors can occurwithout subjective feelings (as they do con-stantly in animals other than humanoid pri-mates, or in unconscious human emotionalacts).

Thus I concur with the idea that emotionalbehaviors evolved as energy-efficient means ofproducing goal-directed actions that fulfillhomeostatic and social needs (Darwin, 1872/1965), and I believe that our capacity forawareness of our emotional behaviors (i.e., oursubjective feelings) evolved because it enor-mously enhanced the efficiency and complexityof emotional communication. I propose in thischapter that a neuroanatomical substrate forsubjective emotional awareness exists infronto-insular cortex as a finite set of repeatedpenultimate metarepresentations of globalemotional moments that extends across time,forming the basis not only for the continuity ofsubjective emotional awareness within a finitepresent, but also for the uniquely human fac-ulty of music. In fact, there are numerous con-vergent observations that lead to this proposal.Finally, these considerations also provide ahomeostatic model that can explain the asym-metrical roles of the left and right forebrain inhuman emotion (Craig, 2005). These ideasmay seem heuristic, philosophical, and specula-tive, but they arise directly from experimentalfindings in cats, monkeys and humans, as I willexplain. These insights afford an opportunityto reclassify functional psychiatric disorderson a neurobiological basis, and they identifyinteroceptive processing regions that can serveas treatment targets for emotional problems(e.g., Simmons, Strigo, Matthews, Paulus, &Stein, 2006).

As a functional neuroanatomist, I study theorganization of the brain by using anatomicalmethods (e.g., tract tracing) and physiologicalmethods (e.g., microelectrode recordings) incomparative animals. I also use psychophysicaland functional imaging methods (e.g., func-tional magnetic resonance imaging, or fMRI)in humans to test the predictions from compar-ative animal studies and to extrapolate thesefindings into more highly evolved regions ofthe human brain. My research is based on theknowledge that the brain is not a mysticalstructure, but rather is reproducibly andevolutionarily well organized for the purposeof maintaining and advancing both the individ-ual and the species. The brain is not color-

coded, its internal connections are not readilyvisible, its physiological operations are ephem-eral, and it is organized in series of processingareas and nested hierarchies that form net-works, so it is difficult to analyze. Studies ofthe effects of lesions and stimulation first iden-tified the sensory regions (for vision, audition,and touch) and the motor regions (for skeletalmovements and visceral activation) of the hu-man cerebral cortex. Modern functional imag-ing studies, which produce color-coded mapsof brain activation based on relative measure-ments of local cerebral blood flow during spe-cific tasks, have validated these insights and arenow being used in experimental psychology toreveal brain regions that are active during dif-ferent cognitive and emotional tasks. However,such studies are constrained by inherent limita-tions in temporal, spatial, and statistical resolu-tion, and most importantly by the fact that theunderlying neural organization is generally notrevealed by a top-down phenomenological ap-proach (sometimes referred to as “blob-ology”;see Sternberg, 2000). By contrast, the findingsdescribed in this chapter begin with a bottom-up view of ascending pathways associated withthe sensory representation of the physiologicalcondition of the body.

The data indicate that these pathways repre-sent the ongoing status of all tissues and organsof the body, including skin, muscle, and vis-cera. I use the term “interoception” specificallyto denote this generalized homeostatic sensorycapacity, redefining it from its original narrowusage to refer only to visceral sensation(Sherrington, 1948). The sensory afferents thatrepresent the condition of the body subserve“homeostasis,” which is the ongoing, hierar-chically organized neurobiological process thatmaintains an optimal balance in the physiologi-cal condition of the body. Homeostasis inmammals comprises many integrated functionsand includes autonomic, neuroendocrine, andbehavioral mechanisms. Thermoregulation is agood example of a homeostatic function. Thesalient purpose of thermoregulation, as of ho-meostasis, is optimal energy management insupport of life. All animals thermoregulate.The primordial means of thermoregulation invertebrates is motivated behavior, similar tohunger and thirst. In humans, such affectivemotivations are accompanied by distincthomeostatic (interoceptive) “feelings”; thesemodalities include not only temperature, pain,itch, hunger, and thirst, but all feelings from the

16. Interoception and Emotion 273

body, such as muscle ache, visceral urgency,and so-called “air hunger.” Consistent with theview that an emotion in humans consists of asensation and a motivation with direct auto-nomic sequelae (Rolls, 1999), I regard thesefeelings as “homeostatic emotions” that drivebehavior. In my opinion, they are virtuallyequivalent to the “background emotions” ofDamasio (1994), and they are incorporated inthe concept of “core affect” (Russell & Barrett,1999; Barrett et al., 2004). The neuroanatomyof the forebrain pathways described below pro-vides compelling support for this concept.

Temperature sensation is an instructive ex-ample of a homeostatic emotion, becausewe normally think of it as an exteroceptivediscriminative sensory capacity. However, theobligatory affect (pleasantness or unpleasant-ness) we feel with each temperature stimulus isthe perceptual correlate of behavioral thermo-regulatory motivation. This affect highlightsthe importance of temperature sensation forhomeostasis, because its valence depends di-rectly on your body’s thermoregulatory needs(Cabanac, 1971; Mower, 1976). Thus the coolglass of water that feels wonderful if you areoverheated feels gnawingly unpleasant if youare chilled. Conversely, if you are chilled, thena hot shower feels wonderful, even if it is sting-ing and prickly, but it would be called painful ifyou were too warm. Similarly, if you remain ina room that is too chilly (or too warm) forenergy-efficient thermoneutrality (or if youplace your hand on an object that is too cold ortoo hot), then you feel a growing discomfort(which as it increases may be called painful)until you respond in a behaviorally appropriatemanner. In the same way, eating salt or sugar ispleasant (and thus motivated) if the body needsit, but it becomes unpleasant and distasteful af-ter you’ve eaten enough. These affective feel-ings reflect behavioral motivations that aredriven by the homeostatic needs of the body,and the human perception of this combinationof a feeling and a motivation is a homeostaticemotion. Note that homeostatically motivatedbehaviors occur in all animals and require noawareness of accompanying feelings. In fact,the structural absence in subprimates of theforebrain homeostatic afferent pathways de-scribed below (specifically, the modality-selective interoceptive representation in thelimbic sensory insular cortex, and the directmotivational pathway to the limbic behavioralmotor cortex in the cingulate region) implies

that they cannot experience feelings from thebody in the same way that humans do.

In what follows, I describe first the func-tional neuroanatomical evidence for aninteroceptive (homeostatic afferent) pathwayin primates, beginning with the development ofspinal homeostatic afferent processing and theidentified spinal neurons that represent each ofthe specific feelings from the body. I detail theprojections of these neurons to spinal,brainstem, and thalamo-cortical levels in mon-keys, and then I describe functional imagingstudies in humans that validate these compara-tive findings. I emphasize a particular func-tional imaging experiment that demonstratesgraded activation of the dorsal posterior insulaby cooling stimuli, because this finding vali-dates the identification of the ascendinginteroceptive pathway and because it providescompelling neuroanatomical evidence for theconcept of homeostatic emotions. In addition,the data in this particular experiment reveal theneuroanatomical progression of activity to theanterior insula that underlies subjective aware-ness of feelings of temperature. Next, I summa-rize the convergent evidence for the role ofanterior insula in all subjective emotional feel-ings, which substantiates the interoceptive ba-sis for emotional awareness, and then I de-scribe how repeated metarepresentations ofglobal emotional moments can explain the ob-served role of this cortical region in awareness,time, and music. Finally, I summarize a neuro-anatomical model for left–right emotionalasymmetry based on the homeostatic afferentprocessing pathways.

THE FUNCTIONAL ANATOMY OFTHE ASCENDING INTEROCEPTIVEPATHWAY IN PRIMATES

Specialized peripheral and central neural sub-strates that represent all homeostatic afferentactivity generate the distinct feelings of pain,temperature, itch, muscle ache, sensual touch,and other bodily sensations by way of discretesensory channels. This modality-selective sen-sory activity is conveyed first of all to hierarchi-cally organized homeostatic integration andpreautonomic response regions in the spinalcord and brainstem. An encephalized forebrainsystem that evolved virtually uniquely in pri-mates surmounts the homeostatic response sys-tem; it provides a direct cortical sensory image


of the physiological condition of the body inthe limbic sensory insular cortex, and, in paral-lel, activates the limbic behavioral motor cor-tex in the medial frontal region (Figure 16.1;for complete descriptions, see Craig, 2002,2003).

Small-Diameter Afferents

Small-diameter (A-delta and C) primary affer-ent fibers (which include nociceptors, thermo-receptors, osmoreceptors, and metaborecep-tors) report the physiological status of thevarious tissues of the body. These fibers termi-nate monosynaptically on projection neuronsin the most superficial layer of the spinal (andtrigeminal) dorsal horn (which is called laminaI or the marginal zone). The way the spinalcord develops reveals the special role of thesefibers in homeostasis. During development,these small-diameter afferents originate from asecond wave of small neurons that emergewithin the dorsal root ganglia subsequent tothe large cells that generate the large-diameterfibers innervating mechanoreceptors and pro-prioceptors. The large-diameter fibers enter thespinal cord first and contact large dorsal hornneurons, and then the small-diameter fibers en-ter at the same time as the lamina I neurons ap-pear. The lamina I neurons originate not fromcells of the dorsal placode, but from progeni-tors of interneurons in the lateral horn (thesympathetic cell column), and they migrate tothe top of the dorsal horn during a ventro-medial rotation of the entire dorsal horn at pre-cisely the right time during development tomeet the incoming small-diameter afferents.This delicately coordinated ontogenetic se-quence indicates that, together, the small-diameter afferents and the lamina I neuronsconstitute a cohesive system for homeostaticafferent activity that parallels the efferent sym-pathetic nervous system. In contrast, the largeneurons that receive large-diameter fiber inputcome to lie at the base of the dorsal horn (be-cause of the rotation) and serve as interneuronsfor skeletal motoneurons. The small-diameterafferent fibers in cranial parasympatheticnerves (e.g., vagus and glossopharyngeal) thatinnervate visceral organs terminate similarly inthe nucleus of the solitary tract in the caudalmedulla. These pathways represent, respec-tively, the afferent inputs for the sympatheticand the parasympathetic halves of the auto-nomic nervous system.


FIGURE 16.1. A schematic representation of thehomeostatic (interoceptive) afferent pathway inprimates that represents the physiological condi-tion of the body. The main forebrain projectionsare to the dorsal posterior insula (interoceptivecortex) and to the medial frontal region (the ante-rior cingulate cortex, or ACC). The pathway thatascends from lamina I of the spinal cord via thelateral spinothalamic tract represents the sympa-thetic half of the homeostatic afferent system, anda parallel pathway via the nucleus of the solitarytract and the basal portion of the ventromedialnucleus in the thalamus represents its parasympa-thetic half (not depicted). From Craig (2003). Re-printed, with permission, from the Annual Re-view of Neuroscience, Volume 26. Copyright2003 by Annual Reviews (www.annualreviews.org).

Lamina I Neurons

Consistent with this developmental view, theprojections of lamina I neurons convey small-diameter homeostatic sensory afferent input di-rectly to the homeostatic integration and con-trol regions in the spinal cord and brainstem.Lamina I projections thus provide the substratefor the hierarchical, modality-selective somato-autonomic reflexes activated by spinal small-diameter afferents that are crucial for homeo-static function (Sato & Schmidt, 1973).

In the spinal cord, their major projection isto the autonomic cell column of the thoraco-lumbar spinal cord, where sympatheticpreganglionic output neurons are located. Inthe brainstem, lamina I neurons project exclu-sively to recognized homeostatic integrationsites, including the ventrolateral medulla, thecatecholamine cell groups A1–A2 and A5–A7,the parabrachial nucleus (PB), and theperiaqueductal gray (PAG). They converge atthese sites with afferent activity associated withthe parasympathetic system by way of the soli-tary nucleus. These sites are all heavily in-terconnected with the hypothalamus andamygdala. The crucial role of lamina I inhomeostatic reflexes is highlighted by the fact

that lamina I and the spinal autonomic cell col-umns are the only spinal sites that receivedescending modulation directly from brain-stem and hypothalamic preautonomic sources.

Lamina I neurons comprise severalmodality-selective classes that can be regardedas virtual “labeled lines” for distinct feelingsfrom the body, such as first (sharp) pain, sec-ond (burning) pain, cool, warm, itch, sensualtouch, muscle ache and cramp, and so on (al-though their activity must be integrated in theforebrain; see Craig, 2003). Each class consistsof a morphologically distinct type of neuronthat receives input from a specific subset ofsmall-diameter afferents. Quantitative analysesusing extracellular recordings from singlelamina I neurons indicate that the responsecharacteristics of these cells correspond verywell with the psychophysical characteristics ofthese distinct sensations (Craig, Krout, & An-drew, 2001; Craig, 2004b). For example, thecooling-sensitive thermoreceptive-specific neu-rons that are uniquely found in lamina I re-spond linearly to innocuous cooling (Figure16.2A) and saturate at noxious cold temper-atures, as human cooling sensitivity does.Intracellular recording and staining showedthat such cells are uniformly pyramid-shaped


FIGURE 16.2. (A) Original data showing that thermoreceptive-specific lamina I spinothalamic neuronslinearly encode skin temperature. The graphs on the left show the ensemble mean, and the graphs on theright show the responses of individual neurons to the characterizing staircase stimulus. From Craig,Krout, and Andrew (2001). Copyright 2001 by the American Physiological Society. Reprinted by permis-sion. (B) Examples of lamina I neurons that were individually characterized and stained intracellularly,showing the correspondence of the three main physiological classes of cells with the three main morpho-logical shapes. From Han, Zhang, and Craig (1998). Copyright 1998 by Nature Publishing Group. Re-printed by permission.

lamina I cells (Figure 16.2B). In contrast, thepolymodal nociceptive lamina I cells are selec-tively sensitive to noxious heat, noxious pinch,and noxious cold, and are monosynapticallyactivated by C-fibers; their responses to re-peated brief-contact heat and the thermal grill(an illusion of pain) correspond in all respectswith the psychophysical characteristics of thehuman sensation of second (“burning”) pain.Such cells are multipolar lamina I neurons (Fig-ure 16.2B). The selectivity of lamina Ispinothalamic tract neurons is convincinglydemonstrated by the subpopulation of cellsthat have characteristics uniquely correspond-ing with the human sensation of itch (Andrew& Craig, 2001).

Whereas lamina I is usually thought of as a“pain and temperature” site, its role in homeo-stasis is clearly revealed by the neurons that re-spond selectively to small-diameter afferentinput from muscle. Such afferents subserve on-going homeostatic adjustments to muscularwork, but when strongly activated they are alsodirectly responsible for the feelings of muscleburn, pain, and cramping. The commingling ofthese muscle-selective neurons in lamina I withother cells selectively responsive to temperatureor pain or itch emphasizes the fundamentalperspective that lamina I neurons convey affer-ent activity representing all aspects of the phys-iological condition of all tissues of the body. Itis also important to recognize that lamina Ineurons do not simply provide emergency sig-nals, but rather are engaged in homeostasis onan ongoing basis, just as cardiorespiratory ac-tivity is continuously affected by ongoing mus-cular activity and by ongoing temperaturechanges. Furthermore, many small-diameterafferents are sensitive to cytokines, opioids, ste-roids, hormones (e.g., somatostatin), and otherlocal and circulating immune modulators, con-sistent with the role of sympathetic and para-sympathetic autonomic efferents in the regula-tion of immune and neuroendocrine functions.

Lamina I Projectionsto the Primate Forebrain

High-resolution tract-tracing experiments haverevealed that in primates, lamina I neuronsproject to the contralateral thalamus in the lat-eral spinothalamic tract, which ascends at theprecise location within the spinal cord wherecordotomy lesions in humans interrupt feelingsfrom the body such as pain, temperature, itch,

sensual touch, and so on (Figure 16.1; Craig,2002, 2003). In subprimates, ascending laminaI activity is processed mainly in brainstem sites(e.g., A1, PB, and PAG), which then provide ahighly integrated signal to emotional behavior-al control regions in the forebrain. Encephal-ization in primates produced a somato-topographic, modality-selective lamina Ispinothalamic projection to a specific thalamo-cortical relay nucleus (VMpo), which in turnprojects to a discrete portion of dorsal poste-rior insular cortex (dpIns, or interoceptive cor-tex). A parallel pathway, also unique to pri-mates, conveys afferent input from the vagaland glossopharyngeal nerves by way of the nu-cleus of the solitary tract to a rostrally adjacentthalamo-cortical relay nucleus (VMb), whichin turn projects to a rostrally adjacent region ofdorsal insular cortex. (The insula is a cortical“island” buried within the lateral sulcus thathas intimate connections with amygdala, hypo-thalamus, and cingulate and orbito-frontal cor-tices.) Together, these pathways through VMpoand VMb provide a direct cortical image of allhomeostatic afferent activity that mirrors thesympathetic and the parasympathetic halves ofthe efferent autonomic nervous system. TheVMpo in the human thalamus is greatly en-larged relative to that of the macaque monkey.

The ascending projections of lamina I neu-rons in primates also provide a direct thalamo-cortical pathway that activates the anteriorcingulate cortex (ACC) by way of a medialthalamic relay (MDvc). By contrast, insubprimates the medial thalamic sources of in-put to the ACC receive integrated homeostaticinput from the brainstem (PB and PAG). (Incats and rats, there is instead a direct lamina Iprojection to a neighboring, developmentallyrelated nucleus in medial thalamus [nucleussubmedius], which projects to ventrolateralorbitofrontal cortex, associated with hedonicintegration and descending antinociception,rather than to the ACC.) I concur with the sug-gestion made previously by several neuro-anatomists (e.g., M. Mesulam, and G. VanHoesen) that the insula can be regarded aslimbic sensory cortex because of its associationwith homeostatic afferent activity and theamygdala, hypothalamus, and orbito-frontalcortex, and that the ACC can be regarded aslimbic motor cortex because of its associationwith homeostatic efferent (autonomic) activityand its descending projections to the PAG inthe brainstem. This fits well with the modern


view of the overall organization of frontal cor-tex into limbic sensory and limbic motor net-works (Ongur & Price, 2000).

FUNCTIONAL IMAGING OFINTEROCEPTION IN HUMANS

Activation of Dorsal Posterior Insula

Functional imaging, lesion, and stimulationstudies in humans confirm the crucial role ofthe dorsal insula in graded pain, graded tem-perature, graded itch, dynamic or painful mus-cle sensation, sensual touch, hunger, thirst, gus-tation, cardiorespiratory activity, “air hunger,”and so on (for references, see Craig, 2002,2003). It serves as primary sensory cortex foreach of these distinct interoceptive “feelings”from the body. Accumulating evidence indi-cates that it is activated in a graded mannerduring each of these sensations, that stimula-tion of this region (or VMpo/VMb) can pro-duce these sensations, and that lesions in-cluding dpIns (or VMpo/VMb) can selectivelydisrupt these sensations. Notably, dpIns isdelimited by labeling for the receptors ofcorticotropin-releasing factor, which many in-vestigators regard as a definitive marker for ho-meostasis (Sanchez, Young, Plotsky, & Insel,1999). The primordial role of the insular cor-tex is modulation of brainstem homeostatic in-tegration sites, which are the main targets of itsdescending projections. Thus the interoceptivecortical image of the physiological condition ofthe body in humans emerged evolutionarily asan extension of the hierarchical homeostaticsystem. In other words, feelings from the bodyin humans reflect its homeostatic condition.

The experimental tract-tracing findings inmonkeys indicate that the lamina I input toVMpo, and also the cortical projections ofVMpo and VMb to dpIns, are topographicallyorganized in the anterior-to-posterior direction(face to foot). The neuroanatomical findings inthe monkey predict that all homeostatic affer-ent modalities will produce graded activationin dpIns that is arranged somatotopo-graphically in the anterior-to-posterior direc-tion. Recent fMRI evidence has demonstratedthis somatotopic gradient for sensations of heatpain, innocuous cooling, and muscle pain inthe dpIns of humans (Brooks, Zambreanu,Godinez, Craig, & Tracey, 2005; Hua, Strigo,Baxter, Johnson, & Craig, 2005; Henderson,Gandevia, & Macefield, 2007). Significantly,

this gradient is orthogonal to the lateral-to-medial (face-to-foot) organization of the mainsomatosensory thalamus and the S1 and S2somatosensory cortices. This contrast substan-tiates the differentiation of dpIns; in particular,it highlights the distinction of dpIns from S2, towhich pain-related activation in the operculo-insular region has often been incorrectly as-cribed. This also means that the large-diameterafferent lemniscal pathway to the somato-sensory thalamus (VPM/VPL) and the small-diameter afferent pathway to VMpo/VMb de-velop along different neurotrophic gradients,which substantiates the concept that the VMband VMpo representation of homeostaticafferents (which control smooth muscle) is sep-arate from the main somatosensory representa-tion of mechanical and proprioceptive inputs(which control skeletal muscle). That is, thereare two fundamentally different thalamo-cortical representations of somatic afferent ac-tivity in the human brain, and the primaryinteroceptive cortex in the posterior insula is adistinct entity.

Activation of Anterior Cingulate

Many functional imaging studies have docu-mented the role of the human ACC in behav-ioral drive and volition (behavioral agency;which seems to be guided dynamically by errorlikelihood estimation, see Brown & Braver,2005), and the ACC is activated by homeo-static afferent activity. The ACC is activated innearly all functional imaging studies of pain,and it is similarly activated in studies of muscleand visceral sensation. Functional imagingstudies of the thermal grill illusion of pain(Craig, Reiman, Evans, & Bushnell, 1996) andof hypnotically modulated pain unpleasantness(Rainville, Duncan, Price, Carrier, & Bushnell,1997) associate ACC activity directly with theaffective motivation of pain (unpleasantness),and they indicate that this pathway engendersthe affect and motivation that render pain ahomeostatic emotional drive. The conjoint ac-tivation of dpIns and ACC by the distincthomeostatic modalities seems to correspondwell with the idea that the feelings from thebody constitute homeostatic emotions that si-multaneously generate both a sensation and amotivation. That is, temperature, pain, itch,and muscle ache are homeostatic emotions thatdrive behavior, just as hunger and thirst are,and the condition of our bodies directly affects


our feelings and our motivations on an ongoingbasis. The activation of the region of the ACCby innocuous thermal stimuli described in thenext section is particularly instructive.

FUNCTIONAL IMAGINGOF TEMPERATURE SENSATION:AN EXAMPLE OFA HOMEOSTATIC EMOTION

Activation of Posterior Insulaand Anterior Cingulate

Temperature, like pain and itch, is a feelingfrom our bodies that we normally externalize.We readily detect the temperature of an objector of the environment. Temperature sensationhas always been regarded as a discriminativeexteroceptive cutaneous sensory capacity alliedwith the sense of touch. However, the ther-moreceptors in our skin, muscles, and visceraactually report local tissue temperature, andthermoregulation is necessary for all animals touse energy most efficiently. The functionalanatomy of temperature sensation in humansreflects the primordial importance of tempera-ture sensibility for homeostasis. The location ofthe discriminative thermosensory cortex identi-fied by the PET imaging data in Plate 16.1 (seecolor insert) fits precisely with the location ofdpIns identified by the functional anatomicaltracing studies in the monkey (Craig, Chen,Bandy, & Reiman, 2000; see also Maihöfner,Kaltenhäuser, Neundörfer, & Lang, 2004; Huaet al., 2005). Plate 16.1 shows three differentanalyses of brain activation observed with PETimaging during the application of six differentcool temperatures on the palm of the hand. Thefirst row, a simple contrast between the highestand lowest temperatures, served as a mask forsubsequent analyses. The second row shows aregression analysis against thermode tempera-ture, which reveals that activation of one site inthe contralateral cortex was directly correlatedwith objective temperature (the red blob on theleft side at level 24). This site can therefore beregarded as primary thermosensory cortex. Itslocation in dpIns explains clinical reports ofstroke-induced lesions that produced therm-anesthesia (Schmahmann & Leifer, 1992;Greenspan, Lee, & Lenz, 1999). Strikingly, thisregion is not part of the somatosensory corticesthat represent touch sensation, which lie on theparietal surface (postcentral gyrus) and in theparietal operculum. Rather, the role of insular

cortex as limbic sensory cortex fits neatly withthe view that temperature sensation is first andforemost of importance for homeostasis. In ad-dition, more recent fMRI evidence reveals thatdynamic cooling produces graded activation ata site in the medial frontal cortex (in the limbicbehavioral motor cortex near the ACC) adja-cent to the site that is activated during pain(Hua et al., 2005). Together, these findings sup-port the view that temperature produces both afeeling and a behavioral motivation inhumans—that is, a homeostatic emotion.

Activation of Anterior Insula Relatedto Subjective Feelings of Temperature

So how do we “feel” temperature? Our subjec-tive sense of innocuous temperature is very lin-ear, but it is not perfect. The deviation is suffi-cient to provide a robust statistical differencebetween cortical activation associated withsubjective rather than objective temperature.The third row in Plate 16.1 shows a regressionanalysis of the same dataset against the partici-pants’ subjective ratings of coolness. A largearea of activation in the anterior insula andorbito-frontal cortex on the right side is con-spicuous in these data. Close comparison of theactivation blobs at level 24 in rows 2 and 3 re-veals that in the subjective regression analysis,a new activation site appears just anterior tothe main site activated in dpIns in the objectivetemperature regression analysis. Thus thesedata indicate that there is an immediate re-representation of thermosensory activation inthe middle insula on the homolateral side, justanterior to dpIns. This re-representationmust be an abstracted or integrated re-representation in some sense, because its activ-ity is significantly more closely related to sub-jective ratings than to objective temperatures.The data show a subsequent commissuraltransfer from the middle insula to a lateralizedseries of re-representations in the anteriorinsula and orbito-frontal cortex on the rightside, which are much more strongly correlatedwith subjective feeling. This is a neurobiologi-cally parsimonious pattern of activation, be-cause progressive re-representations that com-bine feature extraction and cross-modalityintegration are present in the serial processingstreams observed in the visual, auditory, andparietal somatosensory cortical regions, andare consistent with the evolutionary develop-ment of new processing regions in primate


cortex (Krubitzer & Kaas, 2005). Thelateralization is crucial evidence of hemisphericspecialization; I return to this issue later in thechapter.

FUNCTIONAL IMAGING OFSUBJECTIVE FEELINGS IN HUMANS

Thus the PET imaging data on temperaturesensation illuminate an anatomical model inwhich the subjective awareness of feelings fromthe body is generated directly from cortical re-representations of the interoceptive image ofthe body’s homeostatic condition. The resultsof many other imaging studies of subjectivefeelings in humans correspond with this ana-tomical model. These convergent data suggestthat the cortical interoceptive image of thebody provides the basis for awareness of the“material me” or the sentient self.

The data show an anatomical progression ofactivity from dpIns, first to the middle insula

on the homolateral side, and then, by wayof a commissural lateralization, to the rightanterior insula. The posterior-to-anterior pro-gression of processing in the insular cortex isconsistent with several neuroanatomical con-siderations, most particularly with the enor-mous phylogenetic expansion of anterior insulaacross humanoid primates (Allman, Watson,Tetrault, & Hakeem, 2005). (Notably, in mon-keys, interoceptive cortex seems to project di-rectly to orbito-frontal cortex and does notgenerate the successive re-representations inanterior insular cortex that seem to be uniqueto humanoid primates.) Other imaging findingssupport the convergence in the middle insula ofactivity associated with emotionally salientstimuli of all modalities, in part by way of in-terconnections with the amygdala. Figure 16.3summarizes my view of the anatomical organi-zation of progressive integration in insularcortex, which is based on numerous ana-tomical and imaging reports (e.g., Chikama,McFarland, Amaral, & Haber, 1997; Brooks,


FIGURE 16.3. A conceptual diagram of the progressive integration in posterior, middle, and anteriorinsula of interoceptive state with the sensory, autonomic, motivational, and contextual conditions of thebody and the brain to produce a finite set of global emotional representations. In this homeostatic model,the right anterior insula represents the energy-consuming (sympathetic) conditions, and the left anteriorinsula represents the energy-enriching (parasympathetic) conditions.

Nurmikko, Bimson, Singh, & Roberts, 2002;Adolphs, 2002; Schweinhardt et al., 2006;Henderson et al., 2007). This figure illustratesthe concept that progressive re-representationsof the physiological (interoceptive) conditionof the body provide the basis for the integra-tion of homeostatic afferent activity—first inthe middle insula with emotionally salient in-puts from all sensory modalities and fromsubcortical homeostatic control regions (hypo-thalamus and amygdala), and then in the ante-rior insula with emotionally salient activity inother limbic cortical regions (ACC, orbito-frontal cortex), as well as with the cortical re-gion involved in contextual planning(dorsolateral prefrontal cortex). Thus thisposterior-to-anterior progression of re-representations in the human insula provides afoundation for the sequential integration of thehomeostatic condition of the body with thesensory environment, with internal autonomicstate, with motivational conditions and finallywith social conditions. This progression culmi-nates in the most recently evolved regions ofthe anterior insula, situated at its junction withthe frontal operculum and orbito-frontal cor-tex.

This anatomical view of homeostatic emo-tional integration is directly supported by evi-dence that activation of right anterior insula iscorrelated not only with subjective ratings oftemperature, but also with subjective attentionto pain (of either cutaneous, muscular, or vis-ceral origin) (e.g., Brooks et al., 2002). The an-atomical posterior-to-anterior progression canexplain why activation associated with antici-pation of pain and feelings of chronic pain lo-calize in a more anterior portion of right an-terior insula than activation during acuteexperimental pain (Schweinhardt et al., 2006).Furthermore, right anterior insula activation iscorrelated with subjective ratings of disgust ortrustworthiness; with internally generatedemotions like anger, sadness, happiness, lust,and fear; with imitation or empathic referral ofthese emotions; with empathic feelings of painimputed to a loved one; and even with feelingsof social exclusion (for references, see Craig,2002). So the convergence of these findingssupports the view that the cortical re-representations in right anterior insula of theinteroceptive image of the body’s physiologicalcondition provides a basis for the subjectiveawareness of all emotional feelings. The culmi-nation of this integration may involve different

modules that together produce a unified penul-timate metarepresentation, which I refer to as a“global emotional moment.” This convergentintegration is supported by demonstrations ofsynergistic activation of anterior insula byinteroceptive feelings and emotions (e.g., Phil-lips et al., 2003).

These neuroanatomical data substantiate theview that the cortical image of subjective feel-ings is built upon the homeostatic (intero-ceptive) hierarchy. The association of right an-terior insula with subjective awareness of allemotions fits well with the idea that humanemotions are based in part on feelings from thebody, which is the essence of the James–Langetheory of emotion (James, 1890/2007). Themetarepresentations of the state of the body inright anterior insula seem to differentiate innerfrom outer conditions (self from nonself) andto provide a subjective mental image of the ma-terial self as a feeling (sentient) entity that isutilized during all emotional states. In otherwords, it seems to provide the anatomical basisfor emotional awareness. As discussed ingreater detail in Craig (2002), this fits with theanatomical features of the so-called “soma-tic marker” hypothesis of consciousness(Damasio, 1994). The latter concept predictsthat interoceptive state and emotional feelingsare directly related, which is supported by re-cent evidence of distinct patterns ofhomeostatic activity during different emotions(Deichert, Flack, & Craig, 2005; Rainville,Bechara, Nagvi, & Damasio, 2006). Theinteroceptive re-representations in right ante-rior insula provide a substantive neural basisfor these relationships, and also for the “as-if”loop that Damasio postulated would enablemental interoceptive predictions of the conse-quences of emotional behaviors based on priorexperiences.

It is important to note explicitly that this in-tegration in the anterior insula includes notonly emotionally salient sensory inputs, butalso the activity within the motivational re-gions of the brain itself (i.e., ACC, orbito-frontal cortex, ventral striatum). Notably, theintegration of the behavioral agent with the“feeling self” can provide an anatomical basisfor the representation in awareness of the illu-sory “I” postulated by Damasio. Interestingly,studies of placebo analgesia reveal coordinatedactivation of ACC and right anterior insula,suggesting that the integration in the latter ofthe activity in the former (the behavioral agent)


not only produces a representation of motiva-tion within subjective emotional awareness,but also enables motivation to modulate sub-jective feelings directly. This interpretation of-fers an explanation for the paradoxical findingin some studies that activity in right anteriorinsula increases with placebo analgesia, ratherthan being reduced (Petrovic, Kalso, Petersson,& Ingvar, 2002; Lieberman et al., 2004). Thisperspective also negates a main criticism of theJames–Lange theory (i.e., that it did not allowfor feelings of internally generated emotion),and by incorporating the likelihood that ante-rior insula stores representations of pastinteroceptive/emotional experiences, it enablesready explanations of the interactions betweenhomeostatic functions and emotional states—for example, in anxiety, somatization and psy-chosomatic illness.

AN ANATOMICAL MODELFOR SUBJECTIVE AWARENESSOF FEELINGS, AGENCY, AND TIME

Thus a refined and integrated image of the stateof the body seems to provide the basis forawareness of the “feeling self.” In this homeo-static model, the representation of the globalemotional moment includes feelings about mo-tivations, and it can be modified by expecta-tions and motivations, based on interconnec-tions between anterior insula and ACC (as wellas subcortical structures, such as amygdala andventral striatum, and the dorsolateral pre-frontal cortex). In this view, the coactivation ofthe anterior insula and the ACC during allemotions not only corresponds with the viewthat an emotion is both a feeling and a motiva-tion, but also provides the active behavioralagent that is missing from the “somaticmarker” hypothesis. This interpretation is ana-tomically consistent with the dual lamina Ispinothalamo-cortical projections (to insulaand cingulate) and with the overall anatomicalorganization of primate frontal cortex into me-dial (limbic behavioral motor) and insular/or-bital (limbic sensory) networks (Ongur &Price, 2000).

Anterior Insula and Awareness

The role of the right anterior insula in inter-oceptive awareness and subjective feelings issupported by the demonstration by Critchley,

Weins, Rothshtein, Öhman, and Dolon (2004)that its activation is uniquely correlated withsubjective perception of heartbeat timing, aclassic test of interoceptive awareness (see dis-cussion by Craig, 2004). This fits with psycho-logical evidence associating interoceptive(heartbeat) awareness with self-reports of emo-tional experience (Barrett et al., 2004). Yet theanalyses by Critchley et al. (2004) were basedon subjective perceptions of cross-modal tim-ing, and in fact, graded, selective activation ofright anterior insula (and ACC) is producedduring other cross-modal timing disparities—for instance, in the ventriloquism illusion(Bushara, Grafman, & Hallett, 2001). In addi-tion, research has shown strong evidence of adirect relationship between right anteriorinsula activation and subjective awareness oftiming: by using either long or short expectedintervals, by using injections of clonidine toproduce subjective temporal slowing, or by us-ing graded attention to timing in a multicuetask (summarized in Coull, 2004). Neverthe-less, selective activation of anterior insula andACC bilaterally was strongly inversely corre-lated with display time (and performance) in abackward-masking task that probed the tem-poral limits of visual observation during briefsubsecond presentations (Deary et al., 2004);that is, activation increased with shorter pre-sentation times. In my opinion, this very strik-ing result, considered together with these otherfindings, indicates that the fundamental role ofanterior insula may be described most simplyas subjective awareness. This inference is sup-ported by recent evidence indicating that theanterior insula and ACC are uniquely associ-ated with lapses in attention (Weissman, Rob-erts, Visscher, & Woldorff, 2006), with “mind-wandering” (Mason et al., 2007), and with anintrinsic default network for “emotional sa-lience” (Seeley et al., 2007). After highlightinga unique type of neuron in right anterior insula,I outline a structural proposal for this infer-ence.

Von Economo Neurons

Allman et al. (2005) recently reported that apopulation of large spindle neurons, whichthey call “Von Economo neurons” (after anearly neuroanatomist), is uniquely found in theanterior insula (or fronto-insular region) andACC of humanoid primates. Most notably,they reported a trenchant phylogenetic correla-


tion, in that spindle cells are most numerous inaged humans, but progressively less numerousin children, gorillas, bonobos, and chimpan-zees, and nonexistent in macaque monkeys.This progression parallels the results of themirror test for self-awareness (Macphail,1998). Allman et al. hypothesize that thefronto-insular cortex may subserve “intu-ition,” because it is activated in studies of un-certainty and anxiety, and they propose thatthe Von Economo spindle neurons provide thebasis for large-scale integration and fast com-munication between the two sides of cortex. Itseems likely to me that the numerous smallclusters of Von Economo spindle neurons inter-connect the most advanced portions of limbicsensory (anterior insula) and limbic motor(ACC) cortices, both ipsilaterally andcontralaterally. The limbic cortices, in sharpcontrast to the tightly interconnected and con-tiguous sensorimotor cortices, are situatedphysically far apart as a consequence of theirpattern of evolutionary development. Analo-gous to the need for fast interconnections be-tween somatosensory and motor cortices forplaying the piano, the spindle neurons couldenable fast, complex, and highly integratedemotional behaviors.

A Structural Model for Awarenessand Music in the Anterior Insula

Thus an alternative hypothesis to the conceptof “intuition” is the idea that the Von Economospindle neurons, by interconnecting advancedlimbic sensory and motor cortices, provide theanatomical basis for the generation of unifiedpenultimate metarepresentations of globalemotional moments, and indeed for a series ofsuch metarepresentations across time. This hy-pothesis is based not only on the foregoingfunctional anatomical considerations, but alsoon growing evidence that the anterior insula isassociated with music. Functional imagingstudies indicate that listening to music or co-vertly imagining music strongly activates ante-rior insula (Ackermann & Riecker, 2004), andclinical studies indicate that restricted lesionsof anterior insula can produce a conditionknown as “amusia,” or disruption of the abil-ity to appreciate the emotional content of mu-sic. Music is a uniquely human faculty that hasthe profound capacity to bind us together emo-tionally and to change both our emotions andour subjective perception of time. Music can be

described as the rhythmic temporal progressionof emotionally laden moments—which sug-gests a very straightforward structural modelfor awareness, because a rhythmic progressionof global emotional moments can easily be rep-resented by a sequence of quantal anatomicalunits (Craig, 2004). (It is interesting to notehere my observation that bonobos, who seemto me to be remarkably aware and who uni-formly pass the mirror test, clearly use rhythmto communicate—synchronous rhythms foragreement and contrapuntal rhythms for dis-agreement and negotiation—whereas commonchimpanzees do not; see also Macphail, 1998;De Waal, 2003; Williams, 1980.)

That is, if the progressive homeostatic emo-tional integration in the anterior insula does at-tain the concise representation of a global emo-tional moment (i.e., including all emotionallysalient aspects of the homeostatic, sensory, au-tonomic, motivational, and social conditions,as described above), then development of theability to represent each emotional momentacross time would obviously be an enormousevolutionary advantage. A quantal series ofsuch metarepresentations of feelings acrosstime (from the past into the future) would di-rectly enable the perception of autobiographi-cal emotional continuity, the backward recog-nition of emotionally salient behavioralpatterns, and the forward anticipation of emo-tional consequences (i.e., Damasio’s “as-if”loop). Such an anatomical structure could en-gender awareness, because it would provide acortical image of the self as a continuous se-quence of sentient emotional moments (thatcould not “see” itself, just like “conscious-ness”). This anatomical structure could also di-rectly underlie the rhythmic emotional progres-sion of music. And, of course, repetition of thestructural unit representing a global emotionalmoment (in this model, a cluster of VonEconomo spindle neurons) to produce a se-quence of such moments would be a naturaland easily realizable neuroanatomical develop-ment evolutionarily. Evidence for the associa-tion of anterior insula with emotional value es-timation in a temporal-difference model ofinteroceptive learning (Seymour et al., 2004)and with interoceptive prediction in anxiety(Paulus & Stein, 2006) provides support forthis model.

Because the number of quantal representa-tions of global emotional moments must be ana-tomically constrained and finite, its magnitude


would correlate with the duration of the subjec-tive emotional present (the “specious mo-ment”). Thus one prediction of this modelwould be that moments of heightened awarenesswould be characterized by an altered perceptionof time. Intriguingly, the phrase “Time stoodstill” is often used to describe moments ofheightened awareness. A finite absolute numberof quantal global moments would mean thatduring such moments of heightened awareness,the absolute extent of the subjective perceptualpresent must be shortened: As finite resourceswere allocated to maximize representation ofthe present in real time, each global emotionalmoment would represent a shorter temporal du-ration, and the total extent of time representedacross the total pool would be shorter. Thisseems intuitively obvious, and there are fewquantitative psychometric data addressing thisprediction (Tse, Intriligator, Rivest, &Cavanagh, 2004), but the only evidence ofwhich I am aware associating the temporal fo-cusing of awareness with neural activation is theaforementioned study by Deary et al. (2004).

Another prediction would be that the loss ofsuch an anatomical structure would be associ-ated not only with amusia, but also with a con-dition in which the self would not have emo-tional salience across time. In fact, whereaslesions of posterior insula can produce discreteloss of pain and temperature sensations, lesionsof anterior insula can reportedly produce con-ditions regarded clinically as “anosognosia”(the lack of emotional awareness of oneself;Karnath, Baier, & Nagele, 2005) or “anergia”(complete listlessness; Manes, Paradiso, &Robinson, 1999) or “pain asymbolia”(Berthier, Starkstein, & Leiguarda, 1988), or“loss of the feeling of cigarette cravings”(Naqvi, Rudrauf, Damasio, & Bechara, 2007).Congenital disruption of the insula bilaterallyis found in children with Smith–Magenis syn-drome, which is characterized by mental retar-dation and lack of emotional coordination(Boddaert et al., 2004). Most striking is the di-rect association by recent clinical evidence ofthe selective degeneration of Von Economoneurons in the ACC and the anterior insulawith the loss of emotional awareness that char-acterizes fronto-temporal dementia (Seeley etal., 2006). Of course, although I have focusedmy comments on the “feeling” side of thelimbic cortical system, it is crucial to rememberthat if the Von Economo neurons interconnectanterior insula and ACC, consistent with the

conjoint activation of both sites during all emo-tions and during the experiment of Deary et al.(2004), then lesions of the ACC could produceclosely related motivational dysfunctions. Forexample, the ACC has been associated withalexithymia in several studies (e.g., Kano et al.,2003). Further work in this direction is cer-tainly needed.

HOMEOSTATIC AND EMOTIONALASYMMETRY IN THE FOREBRAIN

The homeostatic model of human emotionalawareness presented in the preceding section islateralized to the anterior insula and ACC onthe right side. This fits with an establishedpsychophysiological model, in which emotionis more strongly associated with the rightforebrain. However, psychophysiological evi-dence has accumulated indicating that the leftand right halves of the human forebrain are dif-ferentially associated with particular emotionsand affective traits. That is, reviewers of litera-ture relating affect and emotion to electro-encephalographic activity, cortisol secretion,immune function, and functional imaging ofbrain activity have concluded that positive ver-sus negative valence, approach versus with-drawal behavior, and/or affiliative versus per-sonal relevance are associated with left- versusright-hemispheric forebrain activity, respec-tively, albeit with possible underlying circuitspredominantly used for particular emotions(Heilman, 2000; Davidson, 2004; Allen &Kline, 2004; Wager, Phan, Liberzon, & Taylor,2003; Murphy et al., 2003). I have recentlyproposed that forebrain emotional asymmetryin humans is anatomically based on an asym-metrical representation of homeostatic activitythat originates from asymmetries in the periph-eral autonomic nervous system (Craig, 2005).

Briefly, this proposal suggests that emotionsare organized according to the fundamentalprinciple of autonomic opponency for the man-agement of physical and mental energy. In thishomeostatic neuroanatomical model of emo-tional asymmetry, the left forebrain is associ-ated predominantly with parasympathetic ac-tivity, and thus with nourishment, safety,positive affect, approach (appetitive) behavior,and group-oriented (affiliative) emotions; theright forebrain is associated predominantlywith sympathetic activity, and thus witharousal, danger, negative affect, withdrawal


(aversive) behavior, and individual-oriented(survival) emotions. In this model, manage-ment of physical and mental (meaning neural)energy is the salient organizational motif (asfor homeostasis), such that energy enrichmentis associated with the left forebrain and energyexpenditure is associated with the rightforebrain, consistent with the respective rolesof the parasympathetic and sympathetic effer-ent systems. In this model, the evolution ofmatching homeostatic afferent and autonomicefferent asymmetries in the humanoid fore-brain provided another substantial improve-ment in the efficiency of emotional control andcommunication, thereby facilitating increas-ingly complex social interaction (and leadingultimately to deictic signaling, language, andcivilization). For example, it implies that thehomeostatically parsimonious role of the leftinsula in (parasympathetic) emotional affilia-tion may be the basis for the crucial associationof left anterior insula (which is now included inBroca’s area) in verbal emotional communica-tion and the predominant use of the contra-lateral (right) hand in deictic pointing amonghumanoid primates (De Waal, 2003).

Coordinated opponent interactions betweenthe two hemispheres, mirroring the autonomicprinciple of coordinated opponency, can pro-vide a fundamental management process. Theanalogy of driving a car is useful: Rather thanplacing your hands at opposite sides of thesteering wheel, it is more efficient if you placeyour hands at the 2:00 and 10:00 positions, be-cause then you can drive by exerting a steadydownward pressure with one hand and simplyvarying the downward pressure with the otherhand. This is exactly how the coordinated au-tonomic control of the heart functions; tonicsympathetic drive is modulated by rapid varia-tions in parasympathetic drive. I propose thatthe forebrain coordinates emotional control inthe same manner. Notably, the salient purposeof homeostasis is optimal energy management,and the efficient use of energy in the brain,which overall consumes about 25% of thebody’s entire energy budget, must be a potentevolutionary pressure. This model instantiatesneurobiologically the psychological proposalthat a hypothetical “calm and connection sys-tem” opposes the arousal–stress system(Uvnas-Moberg, Arn, & Magnusson, 2005). Itcan incorporate the bivalent concept of emo-tion, in which positive and negative affects aredifferent psychological dimensions, and also

the core affect concept, in which energization isa key dimension (Barrett et al., 2004; Zautra,2003; Philippot & Chalelle, 2002).

Thus the model provides a structural basisfor suggesting that coordinated opponent inter-actions between left and right insula andcingulate are fundamentally important. Posi-tive and negative affect interact in an opponentfashion; for example, it is well documentedthat social engagement (and oxytocin) can sup-press arousal, stress, depression, and cortisolrelease, whereas, conversely, the latter fac-tors can reduce mood, sociability, and im-mune function (Heinrichs, Baumgartner,Kirschbaum, & Ehlert, 2003; Zautra, 2003).This model proposes that the relative balanceof activity within symmetrical modules in theleft and right insula and cingulate may be ofcritical significance for neurophysiologicallycoordinated emotional complexity, or mentalhealth.

For further elucidation of this proposal, seeCraig (2005). Recent findings add to the evi-dence cited in that article by showing (1) thatasthmatic (vagal bronchiopulmonary) afferentactivity is associated with left insular activation(Rosenkrantz et al., 2005); (2) that emotionallypleasing music, if chosen individually by thesubjects, activates the left anterior insula morestrongly than the right (Koelsch et al., 2006);(3) that a patient with a selective lesion of theleft anterior insula experienced a selective lossof musical enjoyment (Griffiths, Warren, Dean,& Howard, 2005); and (4) that covert singingat different tempos produces a strikingly clearasymmetry in anterior insula activation, consis-tent with the temporal effects of sympatheticand parasympathetic activity on heart rate(Ackermann & Riecker, 2004). Finally, theseconsiderations provide a solid neurobiologicalfoundation for explaining how increases inparasympathetic afferent activity can synergis-tically enhance positive emotion and reducenegative emotion, for example by slow medita-tive breathing (e.g., Brown & Gerbarg, 2005)or by electrical stimulation of the vagus nerve(Nemeroff et al., 2006).

ACKNOWLEDGMENTS

I thank Lisa Feldman Barrett for her editorial sugges-tions, Martin Paulus and Alex Zantra for commentson the final draft, and the National Institutes ofHealth and the Barrow Neurological Foundation forsupport.


REFERENCES

Ackermann, H., & Riecker, A. (2004). The contributionof the insula to motor aspects of speech production:A review and a hypothesis. Brain and Language, 89,320–328.

Adolphs, R. (2002). Trust in the brain. Nature Neuro-science, 5, 192–193.

Allen, J. J., & Kline, J. P. (2004). Frontal EEG asymme-try, emotion, and psychopathology: The first, and thenext 25 years. Biological Psychology, 67, 1–5.

Allman, J. M., Watson, K. K., Tetreault, N. A., &Hakeem, A. Y. (2005). Intuition and autism: A possi-ble role for Von Economo neurons. Trends in Cogni-tive Sciences, 9, 367–373.

Andrew, D., & Craig, A. D. (2001). Spinothalamiclamina I neurons selectively sensitive to histamine: Acentral neural pathway for itch. Nature Neurosci-ence, 4, 72–77.

Barrett, L. F., Quigley, K. S., Bliss-Moreau, E., &Aronson, K. R. (2004). Interoceptive sensitivity andself-reports of emotional experience. Journal of Per-sonality and Social Psychology, 87, 684–697.

Berthier, M., Starkstein, S., & Leiguarda, R. (1988).Asymbolia for pain: A sensory-limbic disconnectionsyndrome. Annals of Neurology, 24, 41–49.

Boddaert, N., De Leersnyder, H., Bourgeois, M.,Munnich, A., Brunelle, F., & Zilbovicius, M. (2004).Anatomical and functional brain imaging evidence oflenticulo-insular anomalies in Smith–Magenis syn-drome. NeuroImage, 21, 1021–1025.

Brooks, J. C., Nurmikko, T. J., Bimson, W. E., Singh, K.D., & Roberts, N. (2002). fMRI of thermal pain: Ef-fects of stimulus laterality and attention. Neuro-Image, 15, 293–301.

Brooks, J. C., Zambreanu, L., Godinez, A., Craig, A. D.,& Tracey, I. (2005). Somatotopic organisation of thehuman insula to painful heat studied with high reso-lution functional imaging. NeuroImage, 27, 201–209.

Brown, J. W., & Braver, T. S. (2005). Learned predic-tions of error likelihood in the anterior cingulate cor-tex. Science, 307, 1118–1121.

Brown, R. P., & Gerbarg, P. L. (2005). Sudarshan Kriyayogic breathing in the treatment of stress, anxiety,and depression: Part I. Neurophysiologic model.Journal of Alternative and Complementary Medi-cine, 11, 189–201.

Bushara, K. O., Grafman, J., & Hallett, M. (2001).Neural correlates of auditory–visual stimulus onsetasynchrony detection. Journal of Neuroscience, 21,300–304.

Cabanac, M. (1971). Physiological role of pleasure. Sci-ence, 173, 1103–1107.

Chikama, M., McFarland, N. R., Amaral, D. G., &Haber, S. N. (1997). Insular cortical projections tofunctional regions of the striatum correlate with cor-tical cytoarchitectonic organization in the primate.Journal of Neuroscience, 17, 9686–9705.

Coull, J. T. (2004). fMRI studies of temporal attention:

Allocating attention within, or towards, time. BrainResearch: Cognitive Brain Research, 21, 216–226.

Craig, A. D. (2002). How do you feel? Interoception:The sense of the physiological condition of the body.Nature Reviews Neuroscience, 3, 655–666.

Craig, A. D. (2003). Pain mechanisms: Labeled linesversus convergence in central processing. Annual Re-view of Neuroscience, 26, 1–30.

Craig, A. D. (2004a). Human feelings: Why are somemore aware than others? Trends in Cognitive Sci-ences, 8, 239–241.

Craig, A. D. (2004b). Lamina I, but not lamina V,spinothalamic neurons exhibit responses that corre-spond with burning pain. Journal of Neurophysi-ology, 92, 2604–2609.

Craig, A. D. (2005). Forebrain emotional asymmetry: Aneuroanatomical basis? Trends in Cognitive Sciences,9, 566–571.

Craig, A. D., Chen, K., Bandy, D., & Reiman, E. M.(2000). Thermosensory activation of insular cortex.Nature Neuroscience, 3, 184–190.

Craig, A. D., Krout, K., & Andrew, D. (2001). Quanti-tative response characteristics of thermoreceptiveand nociceptive lamina I spinothalamic neurons inthe cat. Journal of Neurophysiology, 86, 1459–1480.

Craig, A. D., Reiman, E. M., Evans, A., & Bushnell, M.C. (1996). Functional imaging of an illusion of pain.Nature, 384, 258–260.

Critchley, H. D., Weins, S., Rotshtein, P., Öhman, A., &Dolan, R. J. (2004). Neural systems supportinginteroceptive awareness. Nature Neuroscience, 7,189–195.



Davidson, R. J. (2004). Well-being and affective style:neural substrates and biobehavioural correlates.Philosophical Transactions of the Royal Society ofLondon: Series B. Biological Sciences, 359, 1395–1411.

Deary, I. J., Simonotto, E., Meyer, M., Marshall, A.,Marshall, I., Goddard, N., et al. (2004). The func-tional anatomy of inspection time: An event-relatedfMRI study. NeuroImage, 22, 1466–1479.

Deichert, N. T., Flack, W. F., Jr., & Craig, F. W., Jr.(2005). Patterns of cardiovascular responses duringangry, sad, and happy emotional recall tasks. Cogni-tion and Emotion, 19, 941–951.

De Waal, F. B. M. (2003). Darwin’s legacy and the studyof primate visual communication. Annals of the NewYork Academy of Sciences, 1000, 7–31.

Greenspan, J. D., Lee, R. R., & Lenz, F. A. (1999). Painsensitivity alterations as a function of lesion locationin the parasylvian cortex. Pain, 81, 273–282.

Griffiths, T. D., Warren, J. D., Dean, J. L., & Howard,D. (2004). “When the feeling’s gone”: A selective lossof musical emotion. Journal of Neurology, Neurosur-gery and Psychiatry, 75, 344–345.


Han, Z.-S., Zhang, E.-T., & Craig, A. D. (1998). Noci-ceptive and thermoreceptive lamina I neurons are an-atomically distinct. Nature Neuroscience, 1, 218–225.

Heilman, K. M. (2000). Emotional experience: A neuro-logical model. In R. D. Lane & L. Nadel (Eds.), Cog-nitive neuroscience of emotion (pp. 328–344). NewYork: Oxford University Press.

Heinrichs, M., Baumgartner, T., Kirschbaum, C., &Ehlert, U. (2003). Social support and oxytocin inter-act to suppress cortisol and subjective responses topsychosocial stress. Biological Psychiatry, 54, 1389–1398.

Henderson, L. A., Gandevia, S. C., & Macefield, V. G.(2007). Somatotopic organization of the processingof muscle and cutaneous pain in the left and rightinsula cortex: A single-trial fMRI study. Pain, 128,20–30.

Hua, L. H., Strigo, I. A., Baxter, L. C., Johnson, S. C., &Craig, A. D. (2005). Anteroposterior somatotopy ofinnocuous cooling activation focus in human dorsalposterior insular cortex. American Journal of Physi-ology: Regulatory, Integrative, and ComparativePhysiology, 289, R319–R325.

James, W. (2007). The principles of psychology. Re-trieved from psychclassics.yorku.ca.James/Principles/index.htm (Original work published 1890)

Kano, M., Fukudo, S., Gyoba, J., Kamachi, M.,Tagawa, M., Mochizuki, H., et al. (2003). Specificbrain processing of facial expressions in people withalexithymia: An H2 15O-PET study. Brain, 126,1474–1484.

Karnath, H. O., Baier, B., & Nagele, T. (2005). Aware-ness of the functioning of one’s own limbs mediatedby the insular cortex? Journal of Neuroscience, 25,7134–7138.

Koelsch, S., Fritz, T., DY, V. C., Muller, K., & Friederici,A. D. (2006). Investigating emotion with music: AnfMRI study. Human Brain Mapping, 27, 239–250.

Krubitzer, L., & Kaas, J. (2005). The evolution of theneocortex in mammals: How is phenotypic diversitygenerated? Current Opinion in Neurobiology, 15,444–453.

Lieberman, M. D., Jarcho, J. M., Berman S., Naliboff,B. D., Suyenobu, B. Y., Mandelkern, M., et al.(2004). The neural correlates of placebo effects: Adisruption account. NeuroImage, 22, 447–455.

Macphail, E. M. (1998). The evolution of conscious-ness. Oxford: Oxford University Press.

Maihöfner, C., Kaltenhäuser, M., Neundörfer, B., &Lang, E. (2002). Temporo-spatial analysis of corticalactivation by phasic innocuous and noxious coldstimuli: A magnetoencephalographic study. Pain,100, 281–290.

Manes, F., Paradiso, S., & Robinson, R. G. (1999). Neu-ropsychiatric effects of insular stroke. Journal ofNervous and Mental Disease, 187, 707–712.

Mason, M. F., Norton, M. I., Van Horn, J. D.,Wegner, D. M., Grafton, S. T., & Macrae, C. N.(2007). Wandering minds: The default network and

stimulus-independent thought. Science, 315, 393–395.

Mower, G. (1976). Perceived intensity of peripheralthermal stimuli is independent of internal body tem-perature. Journal of Comparative and PhysiologicalPsychology, 90, 1152–1155.

Murphy, F. C., Nimmo-Smith, I., & Lawrence, A. D.(2003). Functional neuroanatomy of emotions: Ameta-analysis. Cognitive, Affective, and BehavioralNeuroscience, 3, 207–233.

Naqvi, N. H., Rudrauf, D., Damasio, H., & Bechara, A.(2007). Damage to the insula disrupts addiction tocigarette smoking. Science, 315, 531–534.

Nemeroff, C. B., Mayberg, H. S., Krahl, S. E.,McNamara, J., Frazer, A., Henry, T. R., et al. (2006).VNS therapy in treatment-resistant depression: Clini-cal evidence and putative neurobiological mecha-nisms. Neuropsychopharmacology, 31, 1345–1355.

Olausson, H., Lamarre, Y., Backlund, H., Morin, C.,Wallin, B. G., Starck, G., et al. (2002). Unmyelinatedtactile afferents signal touch and project to insularcortex. Nature Neuroscience, 5, 900–904.

Ongur, D., & Price, J. L. (2000). The organization ofnetworks within the orbital and medial prefrontalcortex of rats, monkeys and humans. Cerebral Cor-tex, 10, 206–219.

Paulus, M. P., & Stein, M. B. (2006). An insular view ofanxiety. Biological Psychiatry, 60, 383–387.

Petrovic, P., Kalso, E., Petersson, K. M., & Ingvar, M.(2002). Placebo and opioid analgesia: Imaging ashared neuronal network. Science, 295, 1737–1740.

Philippot, P., & Chalelle, G. (2002). Respiratory feed-back in the generation of emotion. Cognition andEmotion, 16, 605–627.

Phillips, M. L., Gregory, L. J., Cullen, S., Cohen, S., Ng,V., Andrew, C., et al. (2003). The effect of negativeemotional context on neural and behavioural re-sponses to oesophageal stimulation. Brain, 126, 669–684.

Rainville, P., Bechara, A., Naqvi, N., & Damasio, A. R.(2006). Basic emotions are associated with distinctpatterns of cardiorespiratory activity. InternationalJournal of Psychophysiology, 61, 5–18.

Rainville, P., Duncan, G. H., Price, D. D., Carrier, B., &Bushnell, M. C. (1997). Pain affect encoded in hu-man anterior cingulate but not somatosensory cor-tex. Science, 277, 968–971.

Rolls, E. T. (1999). The brain and emotion. Oxford:Oxford University Press.

Rosenkranz, M. A., Busse, W. W., Johnstone, T.,Swenson, C. A., Crisafi, G. M., Jackson, M. M., et al.(2005). Neural circuitry underlying the interactionbetween emotion and asthma symptom exacerba-tion. Proceedings of the National Academy of Sci-ences USA, 102, 13319–13324.


Sanchez, M. M., Young, L. J., Plotsky, P. M., & Insel, T.


R. (1999). Autoradiographic and in situ hybridiza-tion localization of corticotropin-releasing factor 1and 2 receptors in nonhuman primate brain. Journalof Comparative Neurology, 408, 365–377.

Sato, A., & Schmidt, R. F. (1973). Somatosympatheticreflexes: Afferent fibers, central pathways, dischargecharacteristics. Physiological Reviews, 53, 916–947.

Schmahmann, J. D., & Leifer, D. (1992). Parietalpseudothalamic pain syndrome: Clinical features andanatomic correlates. Archives of Neurology, 49,1032–1037.

Schweinhardt, P., Glynn, C., Brooks, J., McQuay, H.,Jack, T., Chessell, I., et al. (2006). An fMRI study ofcerebral processing of brush-evoked allodynia inneuropathic pain patients. NeuroImage, 32, 256–265.

Seeley, W. W., Carlin, D. A., Allman, J. M., Macedo, M.N., Bush, C., Miller, B. L., et al. (2006). Earlyfrontotemporal dementia targets neurons unique toapes and humans. Annals of Neurology, 60, 660–667.

Seeley, W. W., Menon, V., Schatzberg, A. F., Keller, J.,Glover, G. H., Kenna, H., et al. (2007). Dissociableintrinsic connectivity networks for salience process-ing and executive control. Journal of Neuroscience,27, 2349–2356.

Seymour, B., O’Doherty, J. P., Dayan, P., Koltzenburg,M., Jones, A. K., Dolan, R. J., et al. (2004). Temporaldifference models describe higher-order learning inhumans. Nature, 429, 664–667.

Sherrington, C. S. (1948). The integrative action of thenervous system. Cambridge, UK: Cambridge Univer-sity Press.

Simmons, A., Strigo, I., Matthews, S. C., Paulus, M. P.,& Stein, M. B. (2006). Anticipation of aversive visualstimuli is associated with increased insula activationin anxiety-prone subjects. Biological Psychiatry, 60,402–409.

Sternberg, R. J. (2000). Cognition: The holey grail ofgeneral intelligence [Comment]. Science, 289, 399–401.

Tse, P. U., Intriligator, J., Rivest, J., & Cavanagh, P.(2004). Attention and the subjective expansion oftime. Perception and Psychophysics, 66, 1171–1189.

Uvnas-Moberg, K., Arn, I., & Magnusson, D. (2005).The psychobiology of emotion: The role of theoxytocinergic system. International Journal of Be-havioural Medicine, 12, 59–65.

Wager, T. D., Phan, K. L., Liberzon, I., & Taylor, S. F.(2003). Valence, gender, and lateralization of func-tional brain anatomy in emotion: A meta-analysis offindings from neuroimaging. NeuroImage, 19, 513–531.

Weissman, D. H., Roberts, K. C., Visscher, K. M., &Woldorff, M. G. (2006). The neural bases of momen-tary lapses in attention. Nature Neuroscience, 9,971–978.

Wiens, S. (2005). Interoception in emotional ex-perience. Current Opinion in Neurology, 18, 442–447.

Williams, L. (1980). The dancing chimpanzee: A studyof the origins of primitive music. London: Allison &Busby.

Zautra, A. J. (2003). Emotions, stress, and health. NewYork: Oxford University Press.


PA R T I I I

DEVELOPMENTALCHANGES

C H A P T E R 1 7

The Developmentof Facial Expressions

Current Perspectives on Infant Emotions

LINDA A. CAMRAS and SERAH S. FATANI

Although the relevance of facial expression toemotion is widely acknowledged, the nature ofthe relationship between them has been subjectto considerable debate. Charles Darwin’s TheExpression of the Emotions in Man and Ani-mals (Darwin, 1872/1998) is most often citedas the seminal work on emotional expression,providing a detailed account of expressivebehavior across both animal species and hu-man cultures. Darwin argued that expressivebehavior in both animals and humans could beexplained in terms of three principles: service-able associated habits, antithetical actions, andnervous system excitation. According to theprinciple of serviceable associated habits, someexpressive behaviors are derived from instru-mental actions that can serve adaptive func-tions when carried out in their entirety. For ex-ample, one component of the anger expression(baring the teeth) originated in the action ofbiting. However, Darwin was notably ambigu-ous regarding the current adaptive value ofemotional expressions and their relationship to

human motivation (see Fridlund, 1994, fordiscussion). Nonetheless, starting with SilvanTomkins (1962), most recent and contempo-rary emotion theorists (e.g., Ekman, 1994,2003; Izard & Ackerman, 2000; Izard, 1991)implicitly share the view that emotions are pri-mary motivational forces in humans and thatmuch of human behavior is organized in theservice of emotion-related functions and goals.

In addition to advocating the central role ofemotion in human motivation, Tomkins alsorevived scientific interest in emotional facialexpression—a topic that had languished inthe 20th century. As cogently reviewed andcritiqued by Ekman (1982), several studiesconducted from the 1920s through the 1950shad purported to find no systematic relation-ship between facial expression and emotion ineither adults or infants. Nevertheless, Tomkinsencouraged two young investigators, PaulEkman and Carroll Izard, to pursue a series ofstudies on the recognition of emotional facialexpressions in a variety of Western and non-

291

Western cultures. These investigations includedEkman, Sorenson, and Friesen’s (1969) land-mark study of emotion recognition by the Forepeople, a preliterate New Guinea tribe. Sincethat time, considerable evidence (reviewed byMatsumoto, Keltner, Shiota, O’Sullivan, &Frank, Chapter 13, this volume) has been gar-nered in adult studies linking facial emotionalexpression to other components of emotion re-sponding (in particular, self-reported feelings).Although there are vocal opponents to both thethesis of universality and the notion that facialexpressions reflect one’s emotions rather thanmore specifically serve as messages of so-cial communication (Barrett, Lindquist, &Gendron, 2007; Fridlund, 1994; Russell,1994), many contemporary psychologists stillaccept the premise that the prototypic emo-tional expressions described for adults arerelated to human emotions as motivationalstates.

EMOTIONAL EXPRESSIONIN INFANCY

One plausible extension of the adult-based the-ories of facial expression is the proposal that asimilar relationship between expression andemotion exists throughout the lifespan. Yet theimplications of findings from the adult litera-ture for the development of expressive behav-ior are not straightforward. More than onedevelopmental pathway might lead to the out-comes reported for adults. In fact, as we de-scribe below, a number of investigators holdthat infant facial expressions do not corre-spond to adult-like discrete emotions and haveproposed alternative developmental models.

In this chapter, we review a number of thesetheoretical models, focusing on three questionsthat are central to understanding the develop-ment of emotional expression during infancyand beyond: (1) How do facial expressions be-come organized during the course of develop-ment? (2) What is the relation between facialexpressions and emotion-eliciting situations?(3) What is the relationship between facial ex-pressions and other emotion-related behavioralresponses (e.g., emotion “action tendencies”)?These three questions are not unrelated. In par-ticular, as described below, investigators oftenhave sought to address the first one by examin-ing evidence regarding the second and thirdquestions across development.

DIFFERENTIAL EMOTIONS THEORY

Although several related versions of discrete-emotion theories have been proposed in theadult literature, only Izard’s differential-emotions theory (DET; Ackerman, Abe, &Izard, 1998; Izard, 1991; Izard & Malatesta,1987) includes an explicit developmental com-ponent. According to DET, each discrete emo-tion consists of three constituents—neural, ex-pressive, and experiential—that operate as anintegral hard-wired system. Following Tomkins(1962), Izard has argued that emotions (ratherthan drives) are the primary motivators of hu-man behavior. Presumably an individual expe-riencing an emotion is motivated to pursuegoals associated with its adaptive function.Thus emotions motivate the individual’s selec-tion and organization of behaviors aroundthese adaptive goals. Selection and organiza-tion of such behaviors are also acknowledgedto depend upon situational factors, the individ-ual’s personality, and other aspects of his or herdevelopmental history.

As a core component of emotion, expressivebehavior is generally not discretionary, but in-stead is an automatic readout of the emotionsystem. However, several important caveats tothis principle are related to our three focalquestions regarding expressive development.First, DET proponents acknowledge that somefleeting facial expressions observed in the first2 months of life may not be expressions ofemotion (Izard & Abe, 2004). According toDET, at this very early age the neural circuitsinvolved in facial expressions at maturity arestill lacking in infants, and infant expressionssuch as smiles during sleep or during transi-tional states (e.g., waking up from a nap) mayreflect random central nervous system activity(Ackerman et al., 1998). Thus, in response toour first question concerning how facial ex-pressions become organized, DET proponentspropose that expressive reorganization occursearly in development as a result of neurologicalmaturation.

The second exception to the automatic read-out hypothesis suggests that expressive im-pulses sometimes may be too weak to produceobservable expressive behavior. Presumablythis might result in an absence of the predictedfacial expression—but not in the production ofa facial configuration corresponding to a dif-ferent emotion. Lastly, according to DET, ex-pression production may be regulated by older

292 III. DEVELOPMENTAL CHANGES

infants, children, and adults according to socialand personal display rules that are acquiredover the course of development. This is accom-plished by overriding the automatic output ofexpressive behavior via a voluntary control sys-tem that exists separately from emotion. As aresult, older infants, children, and adults maynot always produce facial expressions corre-sponding to their true emotions. However, in-fants past the neonatal stage but younger thanapproximately 1 year are presumed not to ex-ert such voluntary control over their facialbehavior (Izard et al., 1995). Thus, in responseto our second question concerning the relation-ship between expression and emotion-elicitingsituations, DET proposes that observable facialexpressions produced by infants during theirfirst year of life may not always occur inemotion-eliciting situations, but those that dooccur should invariably correspond to the elic-ited emotion. In contrast, older children andadults may voluntarily produce expressionscorresponding to nonexperienced emotions.

Perhaps surprisingly, DET proponents havenot specifically addressed our third focal ques-tion, concerning the relationship between in-fant facial expressions and emotion-related ac-tion patterns. Although DET acknowledgesthat emotional expression includes nonfacialbehaviors (e.g., vocalizations), instrumental ac-tions are not considered to be a core compo-nent of emotion. Thus relationships betweenfacial expressions and action patterns in in-fancy have not been specified or subjected toinvestigation. Instead, as described below, theprimary focus in DET-related research with in-fants has been on the patterning of emotionalexpression across development and the rela-tionship between expressive behavior andemotion-eliciting incentive events.

To facilitate the study of infants’ expressivedevelopment, Izard and his colleagues devel-oped the MAX and AFFEX coding systems(Izard, 1995; Izard, Dougherty, & Hembree,1983) for use in scoring infant expressivebehavior and identifying facial configurationsthat are the expressions of discrete emotions orblends of such emotions. Using these systems,Izard and his colleagues have measured expres-sive behavior in response to a number of elicit-ing situations. In several of their studies (Abe& Izard, 1999; Izard & Abe, 2004; Izard,Hembree, & Huebner, 1987), they have re-ported that infants produce facial expressionsrepresenting appropriate emotion responses to

the incentive events. Izard et al. (1995) havealso examined the relative frequency of full-face versus partial emotional expressions orblends in infants of different ages, as well as thestability of individual differences across infantsin the production of various facial configura-tions. According to Izard’s view, younger in-fants should produce more full-face expres-sions and fewer partial expressions or blends,because full-face expressions are presumed tobe less controlled and regulated. Furthermore,morphological stability of the MAX-specifiedfacial configurations across infancy has beenreported (Izard et al., 1995) and is consideredevidence for their sharing emotion meaningwith the prototypic expressions described foradults. Lastly, Izard and colleagues have con-ducted a number of judgment studies in whichthey report that observers perceive the MAX-specified facial configurations to representtheir presumed emotions (Huebner & Izard,1988; Izard, 1971; Izard et al., 1995; Izard,Huebner, Risser, McGinnes, & Dougherty,1980).

All of these forms of evidence are controver-sial. For example, Oster (2005) has pointed toimportant differences in the morphology ofemotional expressions described for adults andthose specified in the MAX and AFFEX codingsystems. In addition, Oster’s own judgmentstudies (Oster, Hegley, & Nagel, 1992) suggestthat several MAX- and AFFEX-specified ex-pressions are perceived as representing distressrather than discrete negative emotions. Otherinvestigators (Matias & Cohn, 1993) havefound that blends rather than full-face expres-sions are predominant during mothers’ interac-tions with young infants. Lastly, observation ofinfant expressive behavior in response to anumber of different eliciting situations has re-vealed patterns of expressive responses that donot appear to reflect the discrete negative emo-tions presumed to be experienced by the major-ity of infants. We return to these findings be-low.

DIFFERENTIATION THEORIESOF EMOTIONAL DEVELOPMENT

Prior to the advent of contemporary discrete-emotion theories, the most prominent view ofinfant emotional development portrayed on-togeny as a process of differentiation and inte-gration. In 1932 Katherine Bridges published a

17. The Development of Facial Expressions 293

highly influential monograph that dominatedthe literature on infant expressive developmentfor several decades. In this monograph, she de-scribed infant emotions as originating in a stateof diffuse excitement that differentiates first togenerate delight and distress, and then to pro-duce more distinct emotion states such as fear,anger, elation, and affection. More recentlyAlan Sroufe (1996) has produced a theory ofemotional development that retains the notionof specific emotions deriving from less differen-tiated earlier reactions. According to Sroufe,the emergence of emotions can be described interms of three developmental steps: pre-emotion reactions during the newborn period,precursor emotions during the first half year oflife, and more mature emotions during the sec-ond half year. Although he describes these stepsin detail only for joy, fear, and anger, Sroufecontends that the analysis can be extended toother emotions.

Differentiation theorists implicitly attributethe reorganization of emotional expressionduring early infancy to maturational processes.Thus their view of how expressions become or-ganized (and reorganized) over the course ofearly development (i.e., the mechanisms of de-velopmental change) is similar to that proposedby DET. However, in response to our secondtwo questions, they describe very different pat-terns of relationships between facial expres-sions and emotion-eliciting situations, as wellas between such expressions and emotion-related behavioral responses. Generally speak-ing, differentiation theorists have distinguishedbetween positive and negative expressions andbetween positive and negative emotional statesoccurring early in development. However, be-yond these broad-based valence distinctions,they do not propose a consistently tight corre-spondence between specific emotional expres-sions and the less differentiated emotionalstates that precede the development of maturediscrete emotions. For example, Sroufe distin-guishes between positive and negative facial ex-pressions (i.e., smiles vs. distress expressions),but does not distinguish between distress ex-pressions accompanying wariness (the precur-sor to fear) and frustration (the precursor ofanger). Instead, wariness and frustration aredistinguished in terms of both their eliciting cir-cumstances and an infant’s nonfacial responses(i.e., avoidance vs. diffuse attack). Thus facialexpressions are systematically related to both

emotion-eliciting situations and emotion-related behaviors, but not in the manner pro-posed by DET (i.e., as involving a one-to-onecorrespondence between specific expressionsand corresponding discrete emotions).

Because Sroufe has superseded Bridges as themost influential differentiation theorist, wepresent a more complete description of hismodel. According to Sroufe, neonatal pre-emotion reactions (i.e., smiling, distress) areautomatic reflexive responses to quantitative,rather than qualitative, aspects of stimulation(e.g., temporal and intensity features ofarousal). For example, smiling can be producedby gentle modulated arousal, while distress re-sults from more intense arousal buildup. Be-cause Sroufe believes that eliciting stimuli forthese reactions can be identified that are some-what analogous to the later elicitors of moremature emotions (e.g., physical restraint foranger), he considers these neonatal reactions tobe prototypes for later-developing, more dis-crete emotions. Nonetheless, because Sroufedefines emotions as subjective reactions requir-ing some degree of cognitive evaluation, theseearly precognitive reactions are not themselvesconsidered emotions.

Following the neonatal period, precursoremotions (e.g., pleasure, wariness, frustration)emerge. These are considered true emotions, be-cause the infant has begun to develop the cogni-tive ability to process stimulus content; however,they are still regarded as precursors, becausethey involve only the simple cognitive process ofrelating present experiences to past experiences.Thus pleasure arises from stimulus recognition,wariness from recognition failure, and frustra-tion from inability to execute a familiar (i.e., rec-ognizable) behavioral routine. Of particular rel-evance to the topic of this chapter, Sroufe statesthat some precursor emotions are not distin-guishable in terms of their facial or vocal expres-sive components. For example, wariness andfrustration reactions are similarly manifested incrying and distress. After the first 6 months, ba-sic emotions (e.g., joy, fear, and anger) begin toemerge. Sroufe views these as mature emotions,because they involve more complex cognitiveevaluation processes. For example, fear is elic-ited by the perception of a threat, as opposed to amore general failure to recognize a stimulus.Fear and anger are differentially manifested inmore specific behavioral responses (e.g., avoid-ance vs. diffuse attack).


Also of particular relevance for this chapter,Sroufe asserts that reliable differences in the fa-cial expressions for these emotions emergesome time after the emergence of the emotionsthemselves. Thus Sroufe’s theory is distinctfrom DET in its contention that there are notdistinct facial expressions corresponding to dis-tinct emotions at all ages. For example, accord-ing to Sroufe (1996), “few or no elements ofthe fear face are seen in the distress of wari-ness” (p. 65). Earlier and later forms of emo-tions and emotional expressions are relatedthrough their developmental history ratherthan through their morphology.

AN ONTOGENETIC VIEWOF EXPRESSIVE DEVELOPMENT

Focusing more specifically on expressive be-havior rather than emotional development ingeneral, Oster (2005) considers a broaderrange of facial expressions than either Sroufeor Izard does. Rather than confining her effortsto a search for the origins of adult emotionalfacial configurations, Oster seeks to investigatethe infant’s entire expressive repertoire. Ac-cording to her ontogenetic view, an infantshows a variety of distinctive facial expressionsthat have important signal value within thecontext of the infant’s world, going beyond oreven irrespective of their relationship to dis-crete emotions. Because the adaptational de-mands of the environment differ for infantsand adults, she argues that it is reasonable tobelieve that their emotional and expressive rep-ertoires may differ accordingly.

Although Oster has not articulated a specificposition on the question of how facial expres-sions get organized over the course of develop-ment (i.e., our first focal question), her viewimplicitly suggests that expressions are tied toemotion-related states from the beginning, butthat emotion states themselves may changewith development. In principle, this position issimilar to Sroufe’s, although Oster identifies adifferent set of emotion-related states preced-ing the development of discrete emotions (e.g.,attempts by the infant to regulate negative af-fect). In response to our second focal question(regarding relations between situations andemotional expressions), Oster’s discussion, likeSroufe’s, implies that facial expressions are sys-tematically related to situations that elicit these

emotional states. Unlike Sroufe, Oster does notdiscuss relationships between expressions andnonfacial emotion-related actions on the partof the infant (our third focal question). Instead,she relies on analyses of the temporal pattern-ing of facial expressions, their situational oc-currence, and observers’ interpretations of theexpressions to make inferences about the emo-tion states underlying infants’ expressivebehavior.

As an example of her perspective in practice,Oster has studied the situational occurrence of“pouting,” a mouth configuration that wouldbe identified as a component of sadness accord-ing to the MAX and AFFEX coding systems.However, based on her observations, Osterproposes that “pouting” reflects an effort bythe infant to regulate distress, and that the labelof “sadness” fails to capture its more specificmeaning and signal value. Examining infantexpressions by means of her fine-grained BabyFACS (an infant-oriented version of Ekman,Friesen, & Hager’s anatomically based FacialAction Coding System; Oster, 2006), Oster hasalso highlighted morphological differences be-tween the discrete negative emotional expres-sions proposed for infants according to MAXand adult facial configurations correspondingto the same emotions (e.g., adult anger and fearprototypes). These morphological differencessuggest that the meanings of adult and infantexpressions may not be identical. In addition,Oster notes the broad situational occurrence ofMAX-specified pain and anger faces that areoften produced in the context of infant crying.Like several other investigators (e.g., Camras,1991; Sroufe, 1996), she concludes that cur-rently there is no convincing evidence thatthese MAX-specified expressions correspondto discrete adult-like emotions. Instead, theymay reflect a more generalized distress reactionor a more specific level or type of distress thatdoes not correspond to a discrete-emotion label(e.g., intense distress or modulated distressrather than “anger” or “sadness”). To deter-mine the true meaning of infant emotional ex-pressions, Oster advocates an empirical ap-proach to expressive development that takesinfancy itself as its point of origin, seeks toidentify infants’ repertoire of expressive behav-iors, and attempts to determine their adaptivevalue for infants themselves as well as theirontogenetic relationship to expressive behaviorin older individuals. According to Oster,


whether or not this relationship involves a truedifferentiation process remains to be deter-mined.

THE FUNCTIONALIST FRAMEWORK

The functionalist approach to emotion wasproposed by Barrett and Campos (1987) to rec-tify discrete-emotion theories’ tendency to fo-cus narrowly on a set of core components situ-ated within the person. According to thefunctionalist approach, an emotion is a rela-tional process through which an individual at-tempts to establish, change, or maintain somesignificant aspect of his or her relationship tothe external or internal environment. Emotionscan acquire their significance via several pro-cesses. Perhaps the most important of these aresocial signaling (i.e., observation of others’ ex-pressive responses) and appraisal of the event’srelevance to one’s personal goals. The elicitedemotion reflects the nature of that relationshipand includes an open class of responses de-signed to attain (or maintain) a desired goalstate. For example, an event appraised as in-volving a significant loss might evoke responsesdesigned to recoup the loss or to acquire an ac-ceptable substitute. Of importance, the specificresponses themselves (e.g., crying, searchingfor the lost object, seeking comfort) are notpredetermined by an innate emotion program,but are drawn from the individual’s entire re-sponse repertoire in the service of achieving theperson’s particular goal for that emotion epi-sode. That is, emotion responses are functionalrather than fixed or preprogrammed. Nonethe-less, the functionalist approach acknowledgesthat there may be intrinsic links between spe-cific emotions (or emotion families) and partic-ular responses (e.g., action tendencies, facialexpressions, physiological patterning). How-ever, these links are not invariant and are sub-servient to the context-dependent selection ofresponses designed to achieve the individual’semotion goals.

With respect to how infant expressions be-come organized (our first focal question),functionalists propose that this occurs in con-junction with infants’ social development.Consistent with its overall conceptualization ofthe emotion process, functionalism views facialexpressions primarily as social signals servingto communicate emotions to others, ratherthan as direct readouts of emotions themselves.

Thus regarding relationships between facial ex-pressions and emotion situations (our secondfocal question), functionalists assert that facialexpressions may or may not be generated, de-pending upon whether they would be service-able in the particular emotion situation. Impor-tantly, functionalists do not believe that theabsence of facial signaling in an emotion epi-sode necessarily reflects the suppression of“naturally” produced expressive behavior. Al-though expressive behavior indeed may some-times be altered in accordance with social dis-play rules, this reflects a change in theevaluation of the expression’s functionality,rather than the suppression of an inherent au-tomatic readout of the emotion. According tothe functionalist perspective, socialization pro-cesses are important in establishing the condi-tions under which emotion signals (includingfacial expressions) are produced. Developmen-tal changes in expression–emotion relation-ships reflect an individual’s socialization expe-riences in conjunction with his or herevaluation of the effectiveness of expressivebehavior.

In response to our third focal question,functionalists would propose an indirectrather than a direct relationship between fa-cial expressions and other emotion-related be-haviors. Because functionalists view allemotion-related behaviors as discretionary,co-occurrences of specific facial expressionsand instrumental behaviors would not neces-sarily be expected. In fact, functionalists haveargued that facial expressions and instrumen-tal behaviors may be produced as alternativeresponses in emotion situations, dependingupon the individual’s assessment of their rela-tive functionality. Thus, in response to ourthird focal question, functionalists would pre-dict that relationships between facial expres-sions and other emotion-related behaviorsmight not be observed within an individual,but might be observed as substitute responsesto the same emotion situation.

Functionalism represents an important ad-vance in developmental theorizing about ex-pressive behavior. Although it does not elabo-rate on the specific mechanisms of expressionproduction (i.e., processes that determinewhether expressions are generated, and if so,which ones), the functionalist approach hasprovided an important alternative to theoriesthat focus on relatively rigid relationshipsamong a set of core emotion components. It en-


courages investigators to view emotion as amore flexible system of responses oriented to-ward achieving an individual’s goals.

A SOCIOCULTURALINTERNALIZATION MODEL

More recently, Holodynski and Friedlmeier(2006) have presented an integrative model ofexpressive development spanning infancy andchildhood. With respect to infancy, they pro-pose, as differentiation theorists do, that infantfacial expressions initially reflect precursoremotions (e.g., distress, pleasure, fearful ten-sion) and are not selectively associated withdiscrete-emotion-specific causes or coordinatedwith emotion-specific behaviors. However,caregivers interpret these diffuse infant expres-sions within their contexts of occurrence andthus respond to infants with appropriate ac-tions. At the same time, caregivers shape linksbetween specific facial expressions and discreteemotions through their own behavior. That is,caregivers mirror infants’ expressive behaviorduring their interactions with infants, but do soselectively and in exaggerated form. For exam-ple, when an infant cries in circumstances con-sidered sadness-appropriate by the caregiver,then that caregiver may respond with an exag-gerated sadness expression while offering com-fort to the infant. That is, the caregiver displaysboth exaggerated (i.e., prototypic) facial ex-pressions and emotion-appropriate motive-serving actions. Infants thus acquire theseexpression–emotion relationships, which arelater reflected in their own expressive and emo-tional behavior. Holodynski and Friedlmeier’smodel therefore provides a clear proposal foreach of the focal questions we address in thischapter. According to their model, socializationis the means through which diffuse expressionsof positive affect or distress are organized intothe specific facial configurations correspondingto discrete emotions that are selectively associ-ated with emotion-related situations andbehavior responses.

Holodynski and Friedlmeier’s (2006) modelincorporates a number of findings from the de-velopmental literature regarding expressivemirroring during mother–infant interaction. Atthe same time, their proposals regarding thelinking of discrete emotional expressions andemotion-specific functional actions requiresfurther substantiation. In particular, given that

early social interactions may differ widelyacross cultures, further research involving non-Western mothers and infants is required to fur-ther substantiate their proposal regardingmechanisms of expressive development. Never-theless, Holodynski and Friedlmeier present anintriguing model that has the potential tobridge the gap between infant and adult ex-pressive behavior.

THE DYNAMICAL SYSTEMSPERSPECTIVE

Like the functionalist perspective, thedynamical systems approach to emotional de-velopment emphasizes flexibility in the organi-zation of emotion responses and variabilityfrom individual to individual (Camras &Witherton, 2005; Haviland-Jones, Boulifard,& Magai, 2001; Magai & Haviland-Jones,2002). Originating outside the social sciences,the dynamical systems framework has beenproposed as a general model that can be usedto account for the organization of complex sys-tems of various sorts (see Kelso, 1995). Becauseit provides a unique and novel perspective onsystem organization and interrelationshipsamong system components, the dynamicalsystems approach presents a provocative alter-native to the theoretical models describedabove.

The dynamical systems perspective addressesthe question of system organization (our firstfocal question) by asserting a broader principleof self-organization, rather than narrower prin-ciples of either maturational control or shapingvia socialization (Fogel & Thelen, 1987; Fogelet al., 1992). With regard to relationships be-tween facial expressions and both eliciting cir-cumstances and emotion-related behavioral re-sponses (i.e., our second and third focalquestions), the principle of self-organizationimplies a flexible relationship among these sys-tems components, as determined by the partic-ulars of the situational context and the individ-ual’s developmental history. We expand uponthese ideas below.

According to the dynamical systems perspec-tive, emotions may be conceived of as “attrac-tor states”—that is, frequently observed orga-nizations of emotion system components(Fogel & Thelen, 1987). However, dynamicalsystems attractors may themselves involve con-siderable variability in their details. Thus in the


case of emotion, the dynamical systemsapproach—like the functionalist approach—asserts that any specific episode may or maynot include a particular emotion component(e.g., an emotional facial expression). Further-more, at a lower level of analysis, emotioncomponents themselves (e.g., facial expres-sions) may vary in their details. Such variabilityalso may reflect the influence of lower-ordercontextual factors that influence the formationof an attractor. For example, Fogel, Nelson-Goens, Hsu, and Shapiro (2000) found thatdifferent types of mother–infant interactionsare related to specific variants of infant smiling.Fogel, Messinger, and their colleagues (e.g.,Messinger, Fogel, & Dickson, 1999, 2001) pro-pose that such expressive variability engendersvariability in emotional experience, becauseemotion itself emerges from the interaction ofits constituent components.

Dynamical systems approaches are particu-larly concerned with processes leading to sys-tem change both across real time and across de-velopment (Fogel & Thelen, 1987; Thelen,1989). According to this perspective, qualita-tive shifts in the organization of a system(termed “phase shifts”) will occur when somekey system component (termed the “controlparameter”) reaches a critical threshold. Forexample, Wolff (1987) observed that increasingthe intensity of stimulus input sometimes itselfproduced a qualitative change in an infant’s ex-pressive behavior (e.g., from smiling to crying).Across development, a major reorganization ofemotion responding may also occur when somedevelopmental variable reaches its criticalthreshold. For example, a number of investiga-tors (Camras, Oster, Campos, Miyake, &Bradshaw, 1992; Emde, Izard, Huebner, Sorce,& Klinnert, 1985; Fogel & Thelen, 1987;Sroufe, 1996) have noted heterochronicity inthe development of infant emotional responses(e.g., the early appearance of smiles dissociatedfrom other components of the presumed corre-sponding emotion). According to a dynamicalsystems perspective, such expressive compo-nents may become part of an organized emo-tional response when some (as yet unidentified)component of the system achieves threshold.Importantly, control parameters may differacross episodes occurring both within a narrowtime frame and across development. For exam-ple, the intensification of either hunger or painabove a certain threshold may lead to infantdistress. Across development, a major reorga-

nization of emotion responding may occurwhen motor development, language develop-ment, or some other developmental variablereaches a critical threshold (Campos,Kermoian, & Zumbahlen, 1992). Two uniquefeatures of dynamical systems models ofchange are these: (1) The system demonstratesincreased instability as the control parameterreaches its critical threshold; and (2) as it ap-proaches the point of transition, the system be-comes more responsive to external perturba-tions. For example, as hunger increases, aninfant may be more likely to cry in response tobeing accidentally poked by the caregiver.

Various scholars have applied dynamicalsystems principles somewhat differently to pro-vide accounts of infant emotional develop-ment. For example, Fogel et al. (1992) viewemotions as emerging de novo from the self-organized interactions of components, includ-ing some that are not emotion-specific (e.g.,gaze, posture, instrumental actions). In con-trast, Marc Lewis retains the notion that dis-crete emotions are preexisting hard-wired as-semblies of expressive, physiological, andphenomenological components, but viewsthese assemblies as components withinlarger self-organized dynamical systems(i.e., emotion–appraisal amalgams; Lewis &Douglas, 1998). With respect to emotional fa-cial expressions, the former position impliesthat unique relationships between facial config-urations and emotional experience should notbe expected. The latter position implies that aninvariant concordance between certain facialexpressions and their corresponding affectiveexperiences should exist.

Because advocates of both these positionscan offer plausible (but different) sorts of evi-dence consistent with their view, we havesought to develop a third position that can ac-commodate the widest range of findings todate. This effort originated in Camras’s (1992)observational study of her daughter’s expres-sive behavior during the first 9 weeks of life.Initially adopting the perspective of discrete-emotion theories, Camras made a number ofunexpected observations that could not easilybe explained by DET’s account of infant ex-pressive behavior. Subsequent research (de-scribed below), as well as a review of the litera-ture (Camras, Malatesta, & Izard, 1991),similarly revealed phenomena indicating thatDET’s initial account of infant expressive de-velopment required modification.


In particular, three anomalous phenomenawere identified: (1) the systematic occurrenceof codeable “emotional” expressions in situa-tions unlikely to have elicited the correspond-ing discrete emotion (e.g., nonemotional neo-natal smiling [Emde et al., 1985], “surprise”expressions produced by infants who were un-likely to be experiencing surprise [Camras,Lambrecht, & Michel, 1996]); (2) the nonoc-currence of emotional expressions correspond-ing to the emotion presumably experienced byinfants (e.g., absence of fear expressions in in-fants judged to be afraid on the visual cliff[Hiatt, Campos, & Emde, 1979], absence ofsurprise expressions in infants judged to be sur-prised by an expectancy violation [Camras,2000; Camras et al., 2002]); and (3) rapidshifts between MAX-designated facial expres-sions for anger, sadness, and physical distress/pain during bouts of intense crying occurring inresponse to almost any form of negative elicitor(including MAX-specified pain expressions incircumstances during which pain was unlikelyto be experienced; Camras, 1992).

To explain these findings, Camras (2000)has proposed that infant emotions and infantfacial expressions constitute overlapping butpartly separate dynamical systems. More spe-cifically, to explain the first anomaly describedabove, Camras and her colleagues (Camras etal., 1996; Michel, Camras, & Sullivan, 1992)drew upon dynamical systems research in thearea of nonfacial motoric action, and in partic-ular upon the concept of the “coordinative mo-tor structure” (i.e., a grouping of muscle ac-tions that are synergistically linked to eachother; see also Fogel, 1985). Thus they pro-posed that infant facial expressions may some-times be produced in non-emotion-related cir-cumstances via self-organizing processes ofrecruitment among “lower-order” facial mus-cle movements. For example, Camras et al.(1996) showed that 5- and 7-month-old infantswill raise their brows as they open their mouthsto incorporate an object. That is, opening themouth recruits a synergistically-related raisedbrow movement producing an expression of“surprise” in nonsurprising situations.

To explain the second set of anomalous find-ings, Camras (2000; Camras et al., 2002) pro-posed that facial expressions are nonobligatorycomponents of emotion episodes and are pro-duced only when their corresponding controlparameters are present at the necessary thresh-old. In dynamical systems terms, discrete-

emotion attractors (i.e., larger ensembles ofemotion components) can be variable and thusmay or may not include a prototypic emotionalexpression in any one instance. This view issimilar to the functionalists’ notion that facialexpressions are produced only when theyserve a communicative function. However, dy-namical systems proponents would recognizethe potential for a broader set of control pa-rameters (both social and nonsocial) to deter-mine whether or not a facial expression is pro-duced. For example, Michel et al. (1992)showed that infants raised their brows whenlooking upward to view an attractive object,but not when looking downward. In this case,head position and gaze direction served as“control parameters” determining whether ornot the infant produced a facial expressioncodeable as “interest” in an interest-relevantsituation.

To explain the third set of anomalous find-ings, Camras (1992) proposed that the facialconfigurations depicted by DET (see above) asreflecting discrete anger, sadness, or physicalpain may represent negative affective statesthat do not correspond to traditional categoriesof discrete emotions. Like Oster (2005), shesuggested that some of these expressions mayreflect different intensities of a more generalstate of “distress”—in other words, negativeemotion that is relatively undifferentiated withrespect to distinguishable functional goals (e.g.,removing an obstacle vs. escaping danger).Some forms of negative emotional expressionalso may reflect infants’ efforts to modulatethat distress (Oster, 2005) and/or the influenceof non-emotion-related factors such as headposition and gaze direction. For example,Camras et al. (2007) have suggested that differ-ences in head position and/or gaze directionduring negative emotion episodes may deter-mine whether infants produces brow configu-rations associated with prototypic expressionsof fear or sadness or anger.

Recent studies have continued to explore therelationship between negative facial expres-sions and negative emotion in infants. In a col-laborative investigation of European Ameri-can, Japanese, and Chinese infants, Camras etal. (2007) found that more 11-month-olds pro-duced MAX-specified anger expressions thanfear expressions in both anger/frustration-eliciting situations and fear-eliciting situations.Other mismatches between infant facial ex-pressions and their situationally based pre-


dicted emotions have also been documented(Bennett, Bendersky, & Lewis, 2002, 2004). Atthe same time, some studies (Lewis, Alessandri,& Sullivan, 1990; Lewis & Ramsay, 2005;Sullivan & Lewis, 2003; Weinberg & Tronick,1994) have found differential relationships be-tween MAX-specified anger and sadness ex-pressions and other components of infant re-sponding. However, these important findingscan be accommodated within both a discrete-emotion model and a model that links specificinfant negative expressions to different levels ofdistress intensity and modulation. Because itcan incorporate the wider range of findings re-viewed above, we currently favor the latter in-terpretation.

Because at the present time there is no extantevidence suggesting that the “pain,” “anger,”and “sadness” configurations occur in infantswhen negative emotion is completely absent,we currently retain the notion of an invariantconcordance between these expressions andsome form of negative affect. This representsa compromise between the discrete-emotionsperspective (with its emphasis on hard-wiredrelationships between specific expressions andspecific negative emotions) and the purely dy-namical systems view espoused by Fogel andhis colleagues (a view that rejects the notion ofinvariant links controlled by preexisting emo-tion programs). However, we preserve the rightto revise our present position in light of futureresearch that may reveal examples of mis-matches between negative emotion and nega-tive infant facial expressions (e.g., analogous tocrying from happiness in adults).

To summarize our view of expressive devel-opment, we concur with other dynamical sys-tems thinkers in asserting that facial expres-sions and other components of discreteemotions develop heterochronically. That is,they are not initially associated with each otherin young infants. We further propose that dur-ing the course of development, these facial con-figurations do eventually become linked toother components of their corresponding dis-crete emotions. For example, anger expressionsbecome differentially linked to other compo-nents of anger rather than to more general dis-tress.

Although the mechanisms involved in such atransformation are currently unknown, theymay reasonably be proposed to involve infant–caregiver interactions such as those describedby Holodynski and Friedlmeier (2006). Cast in

dynamical systems terms, the control para-meters producing a phase shift in expressivebehavior would be embedded in such in-teractions. However, to justify employing dy-namical systems concepts to describe this de-velopmental change, researchers would need todemonstrate that such transformation involvesthe characteristics of a dynamical systemsphase shift (e.g., increased instability and re-sponsiveness to perturbations during the pe-riod of transformation).

Beyond these proposals, we also wish to sug-gest that a one-to-one invariant relationship be-tween the prototypic facial expressions andother components of discrete emotions maynever arise. In dynamical systems terms,discrete-emotion attractors (i.e., larger ensem-bles of emotion components) may be variableeven in adulthood and thus may or may not in-clude a prototypic emotional facial expression inany particular instance. In addition, as dynami-cal systems thinkers, we believe that prototypic“emotional” facial configurations may alsoserve as components of non-emotion-related“attractor states.” Thus they may be systemati-cally produced in non-emotion-related contexts(e.g., when women lift their brows and opentheir mouths to produce a “surprise” expressionwhile putting on mascara).

FUTURE DIRECTIONS

Like the other theoretical perspectives reviewedabove, the dynamical systems perspective hasmade a significant but limited contribution toour understanding of infant emotional and ex-pressive development. Although it serves as avaluable heuristic to those seeking an alterna-tive theory that can incorporate data not easilyaccommodated by other perspectives, currentlyproposed dynamical systems accounts arethemselves incomplete. Thus, although we our-selves favor the dynamical systems perspective,we believe that it is premature to adjudicate de-finitively among the several theoretical ap-proaches described above. Instead, we con-clude by identifying some directions for futureresearch that will enable scholars to developtheir thinking further within each perspectiveand to evaluate the relative merits and limita-tions of each approach.

First, we believe that it is critically impor-tant for future researchers to more exten-sively document the factors associated with


production of various emotional expressionsacross a wide variety of situations and acrossa broad developmental time period. This isan essential gap that must be filled in orderto proceed to create an account of expressivedevelopment within any theoretical frame-work. To provide a specific example from thedynamical systems perspective, once we deter-mine the age and circumstances under whichprototypic fear expressions are produced, wecan seek to identify developmental phaseshifts and determine the relevant “control pa-rameters” for this expression. Similarly, inves-tigators utilizing other approaches can deter-mine whether the eliciting factors for thisexpression are consistent with their own the-oretical perspective.

Second, we wish to reemphasize our beliefthat a satisfactory account of emotional ex-pression must also explain the variability de-scribed for the emotional expression proto-types themselves. For example, as highlightedby Oster (2005), several different facial config-urations are included as “sad” expressionswithin the MAX and AFFEX coding systems,and these may have very different meanings.With respect to smiling, Fogel and his col-leagues have conducted exemplary work inidentifying contextual factors systematicallyassociated with variants of this expression.Further studies should focus on explainingvariability within the prototypic expressionsdescribed for other discrete emotions. Oneunique contribution of the dynamical systemsapproach to the development of emotional ex-pression is to conceptualize both expressionsand emotions as variable attractor states ratherthan fixed entities, and thus to motivate effortsto understand their variability.

In conclusion, we also feel it is important toemphasize that none of the developmental per-spectives described above are necessarily incon-sistent with findings in the adult literature ofcorrespondences between emotional ex-pressions and self-reported feeling states(Matsumoto et al., Chapter 13, this volume).However, we also call for studies that furtherexplore the relationship between facial expres-sion and emotion in older children and adults.By generating more extensive data on children’sand adults’ production of facial expressions ina wide range of circumstances, we can best en-sure that our models accurately reflect the truerelationship between expression and emotionthroughout the lifespan.

REFERENCES

Abe, J. A., & Izard, C. E. (1999). A longitudinal studyof emotion expression and personality relations inearly development. Journal of Personality and SocialPsychology, 77(3), 566–577.

Ackerman, B. P., Abe, J. A., & Izard, C. E. (1998). Dif-ferential emotions theory and emotional develop-ment: Mindful of modularity. In M. F. Mascolo & S.Griffin (Eds.), What develops in emotional develop-ment? (pp. 85–106). New York: Plenum Press.

Barrett, K. C., & Campos, J. J. (1987). Perspectives onemotional development: II. A functionalist approachto emotions. In J. Osofsky (Ed.), Handbook of infantdevelopment (2nd ed., pp. 555–578). New York:Wiley.

Barrett, L. F., Lindquist, K., & Gendron, M. (2007).Language as a context for emotion perception.Trends in Cognitive Sciences, 11, 327–332.

Bennett, D., Bendersky, M., & Lewis, M. (2002). Facialexpressivity at 4 months: A context by expressionanalysis. Infancy, 3(1), 97–113.

Bennett, D., Bendersky, M., & Lewis, M. (2004). Onspecifying specificity: Facial expressions at 4 months.Infancy, 6(3), 425–429.

Bridges, K. M. B. (1932). Emotional development inearly infancy. Child Development, 3, 324–341.

Campos, J., Kermoian, R., & Zumbahlen, M. (1992).Socioemotional transformations in the family systemfollowing infant crawling onset. In N. Eisenberg &R. Fabes (Eds.), New directions for child develop-ment: Vol. 55. Emotion and its regulation in early de-velopment (pp. 25–40). San Francisco: Jossey-Bass.

Camras, L. A. (1991). Conceptualizing early infant af-fect: View II and reply. In K. Strongman (Ed.), Inter-national review of studies on emotion (pp. 16–28,33–36). New York: Wiley.

Camras, L. A. (1992). Expressive development and ba-sic emotion. Cognition and Emotion, 6(3/4), 269–283.

Camras, L. A. (2000). Surprise!: Facial expressions canbe coordinative motor structures. In M. Lewis & I.Granic (Eds.), Emotion, development and self-organization (pp. 100–124). New York: CambridgeUniversity Press.

Camras, L. A., Lambrecht, L., & Michel, G. (1996). In-fant “surprise” expressions as coordinative motorstructures. Journal of Nonverbal Behavior, 20, 183–195.

Camras, L. A., Malatesta, C., & Izard, C. (1991). Thedevelopment of facial expressions in infancy. In R.Feldman & B. Rime (Eds.), Fundamentals of nonver-bal behavior (pp. 73–105). Cambridge, UK: Cam-bridge University Press.

Camras, L. A., Meng, Z., Ujiie, T., Dharamsi, S.,Miyake, K., Oster, H., et al. (2002). Observing emo-tion in infants: Facial expression, body behavior andrater judgments of responses to an expectancy-violating event. Emotion, 2, 178–193.

Camras, L. A., Oster, H., Bakeman, R., Meng, Z., Ujiie,


T., & Campos, J. J. (2007). Do infants show distinctnegative facial expressions for different negativeemotions?: Emotional expression in European-American, Chinese, and Japanese infants. Infancy,11(2), 131–155.

Camras, L. A., Oster, H., Campos, J., Miyake, K., &Bradshaw, D. (1992). Japanese and American in-fants’ responses to arm restraint. Developmental Psy-chology, 28(4), 578–583.

Camras, L. A., & Witherington, D. C. (2005). Dynami-cal systems approaches to emotional development.Developmental Review, 25, 328–350.

Darwin, C. (1998). The expression of the emotions inman and animals. New York: Oxford UniversityPress. (Original work published 1872)

Ekman, P. (1982). Emotion in the human face (2nd ed.).Cambridge, UK: Cambridge University Press.

Ekman, P. (1994). All emotions are basic. In P. Ekman& R. Davidson (Eds.), The nature of emotion: Fun-damental questions (pp. 15–19). New York: OxfordUniversity Press.

Ekman, P. (2003). Emotions revealed: Recognizing facesand feelings to improve communication and emo-tional life. New York: Holt.

Ekman, P., Sorenson, E., & Friesen, W. (1969). Pan-cultural elements in facial display of emotion. Sci-ence, 164, 86–88.

Emde, R. N., Izard, C., Huebner, R., Sorce, J. F., &Klinnert, M. (1985). Adult judgments of emotions:Replication studies within and across laboratories.Infant Behavior and Development, 8, 79–88.

Fogel, A. (1985). Coordinative structures in the devel-opment of expressive behavior in early infancy. In G.Zivin (Ed.), The development of expressive behavior(pp. 249–267). New York: Academic Press.

Fogel, A., Nelson-Goens, G., Hsu, H., & Shapiro, A.(2000). Do different infants smiles reflect differentpositive emotions? Social Development, 9(4), 497–520.

Fogel, A., Nwokah, E., Dedo, J., Messinger, D., Dick-son, K., Matusov, E., et al. (1992). Social process the-ory of emotion: A dynamic systems approach. SocialDevelopment, 1(2), 122–142.

Fogel, A., & Thelen, E. (1987). The development ofearly expressive and communicative action. Develop-mental Psychology, 23, 747–761.

Fridlund, A. (1994). Human facial expression. SanDiego, CA: Academic Press.

Haviland-Jones, J. M., Boulifard, D., & Magai, C.(2001). Old–new answers and new–old questions ofpersonality and emotion: A matter of complexity. InH. A. Bosma & E. S. Kunnen (Eds.), Identity andemotion: Development through self-organization(pp. 151–171). Cambridge, UK: Cambridge Univer-sity Press.

Hiatt, S., Campos, J., & Emde, R. (1979). Facial pat-terning and infant emotional expression: Happiness,surprise, and fear. Child Development, 50, 1020–1035.

Holodynski, M., & Friedlmeier, W. (2006). Develop-ment of emotions and their regulation: A socio-culturally based internalization model. Boston:Kluwer Academic.

Huebner, R. R., & Izard, C. E. (1988). Mothers’ re-sponses to infants’ facial expressions of sadness, an-ger, and physical distress. Motivation and Emotion,12, 185–196.


Izard, C. E. (1991). The psychology of emotions. NewYork: Plenum Press.

Izard, C. E. (1995). The Maximally Discriminative Fa-cial Movement Coding System. Unpublished manu-script.

Izard, C. E., & Abe, J. A. (2004). Developmentalchanges in facial expression of emotions in thestrange situation during the second year of life. Emo-tion, 4(3), 251–265.

Izard, C. E., & Ackerman, B. P. (2000). Motivational,organizational, and regulatory functions of discreteemotions. In M. Lewis & J. M. Haviland-Jones(Eds.), Handbook of emotions (2nd ed., pp. 253–264). New York: Guilford Press.

Izard, C. E., Dougherty, L., & Hembree, E. (1983). Asystem for identifying affect expressions by holisticjudgments (AFFEX). Newark: Instructional Re-sources Center, University of Delaware.

Izard, C. E., Fantauzzo, C. A., Castle, J. M., Haynes, O.M., Rayias, M. F., & Putnam, P. H. (1995). The on-togeny and significance of infants’ facial expressionsin the first 9 months of life. Developmental Psychol-ogy, 31(6), 997–1013.

Izard, C. E., Hembree, E. A., & Huebner, R. R. (1987).Infants’ emotion expression to acute pain: Develop-mental changes and stability of individual differ-ences. Developmental Psychology, 23(1), 105–113.

Izard, C. E., Huebner, R. R., Risser, D., McGinnes, G.C., & Dougherty, L. M. (1980). The young infant’sability to produce discrete emotion expressions. De-velopmental Psychology, 16, 132–140.

Izard, C. E., & Malatesta, C. (1987). Perspectives onemotional development I: Differential emotions the-ory of early emotional development. In J. Osofsky(Ed.), Handbook of infant development (2nd ed.,pp. 494–554). New York: Wiley.

Kelso, J. (1995). Dynamic patterns. Cambridge, MA:MIT Press.

Lewis, M., Alessandri, S. M., & Sullivan, M. W. (1990).Violation of expectancy, loss of control, and anger inyoung infants. Developmental Psychology, 26, 745–751.

Lewis, M., & Ramsay, D. (2005). Infant emotional andcortisol responses to goal blockage. Child Develop-ment, 76, 518–530.

Lewis, M. D., & Douglas, L. (1998). A dynamic systemsapproach to cognition–emotion interactions in devel-opment. In M. F. Mascolo & S. Griffin (Eds.), What


develops in emotional development? (pp. 159–188).New York: Plenum Press.

Magai, C., & Haviland-Jones, J. M. (2002). The hiddengenius of emotion. Cambridge, UK: Cambridge Uni-versity Press.

Matias, R., & Cohn, J. (1993). Are MAX-specified in-fant facial expressions during face-to-face interactionconsistent with differential emotions theory? Devel-opmental Psychology, 29(3), 524–531.

Messinger, D., Fogel, A., & Dickson, K. (1999). What’sin a smile? Developmental Psychology, 35(3), 701–708.

Messinger, D., Fogel, A., & Dickson, K. (2001). Allsmiles are positive, but some smiles are more positivethan others. Developmental Psychology, 37(5), 642–653.

Michel, G., Camras, L., & Sullivan, J. (1992). Infantinterest expressions as coordinative motor struc-tures. Infant Behavior and Development, 15, 347–358.

Oster, H. (2005). The repertoire of infant facial expres-sions: An ontogenetic perspective. In J. Nadel & D.Muir (Eds.), Emotional development (pp. 261–292).Oxford: Oxford University Press.

Oster, H. (2006). Baby FACS: Facial Action Coding Sys-tem for infants and young children. Unpublishedmanuscript, New York University.

Oster, H., Hegley, D., & Nagel, L. (1992). Adult judg-ments and fine-grained analysis of infant facial ex-pressions: Testing the validity of a priori coding for-mulas. Developmental Psychology, 28, 1115–1131.

Russell, J. A. (1994). Is there universal recognition ofemotion from facial expression?: A review of thecross-cultural studies. Psychological Bulletin, 115,102–141.

Sroufe, L. A. (1996). Emotional development. NewYork: Cambridge University Press.

Sullivan, M., & Lewis, M. (2003). Contextual determi-nants of anger and other negative expressions in younginfants. Developmental Psychology, 39, 693–705.

Thelen, E. (1989). Self-organization in developmentalprocesses: Can systems approaches work? In M. R.Gunnar & E. Thelen (Eds.), Minnesota Symposiumon Child Psychology: Vol. 22. Systems and develop-ment (pp. 77–118). Hillsdale, NJ: Erlbaum.

Tomkins, S. (1962). Affect, imagery, consciousness: Vol.1: The positive affects. New York: Springer.

Weinberg, K., & Tronick, E. (1994). Beyond the face:An empirical study of infant affective configurationsof facial, vocal, gestural, and regulatory behaviors.Child Development, 65, 1503–1515.

Wolff, P. (1987). The development of behavioral statesand the expression of emotions in early infancy. Chi-cago: University of Chicago Press.


C H A P T E R 1 8

The Emergenceof Human Emotions

MICHAEL LEWIS

Observation of newborn infants reveals arather narrow range of emotional behavior.They cry and show distress when pained,lonely, or in need of food and attention. Theylook attentive and focused on objects and peo-ple in their world. They appear to listen tosounds, to look at objects, and to respond totickle sensations. Moreover, they seem to showpositive emotions, such as happiness and con-tentment. When fed, picked up, or changed,they show relaxed body posture, smile, and ap-pear content. Although they show a wide rangeof postural and even facial expressions, the setof discrete emotions that they exhibit is ratherlimited. Yet, in a matter of months and indeedby the end of the third year of life, these samechildren display a wide range of emotions. Infact, by this age, almost the full range of adultemotions can be said to exist (Lewis, 1992b).In 3 years, the display and range of humanemotions go from a few to the highly differenti-ated many, as well as from action patterns tofeelings. In order to understand this rapid de-velopment, it is necessary for us to consider theset of issues that will enable a careful articula-

tion of their development. The first issue to bediscussed, therefore, is the topic of the topol-ogy of emotional features. Embedded withinthis is a consideration of the development ofthese features. Finally, the developmental se-quence over the first 3 years of life is consid-ered.

THE TOPOLOGY OF EMOTION

In order to talk about developmental issues in-volved in the study of emotion, it is importantthat we first make clear what we mean by theterm “emotion.” “Emotion,” like the term“cognition,” refers to a class of elicitors, be-haviors, states, and experiences. If we do notdistinguish among these features of emotion,the study of them and their development be-comes difficult. For example, Zajonc (1980)—and, more recently, Barrett (see Lindquist& Barrett, Chapter 34, this volume) andNiedenthal (see Chapter 36, this volume)—argued that emotions can occur withoutcognitions, while Lazarus (1982) argued that

304

emotions require cognitions. As we shall see,Zajonc and Lazarus were describing differentfeatures of emotional life. Because of this, eachcould arrive at diametrically opposing posi-tions without endangering his own argument.The reasons for this are quite simple: As weshall see, one was arguing for emotions asstates, perhaps connected to action patternsand environmental triggers, whereas the otherargued for emotions as experiences.

Emotional Elicitors

In order for an emotion to occur, some stimulusevent—what I will call the “emotionalelicitor”—must trigger a change in the state ofthe organism. The state of the organism can bea change in an idea, or it can be a change in thephysiological state of the organism. The trig-gering event may either be an external orinternal stimulus. External elicitors may benonsocial (e.g., loud noise) or social (e.g., sepa-ration from a loved one). Internal elicitors mayrange from changes in specific physiologicalstates to complex cognitive activities. Since it isobviously much harder to identify and manipu-late an internal elicitor than an external one, itis not surprising that most research deals withexternal stimuli; that is, it attempts to deter-mine precisely which features of the elicitor ac-tivate the emotion.

A major problem in defining an emotionalelicitor is that not all stimuli can be character-ized as emotional elicitors. For example, a blastof cold air may cause a drop in body tempera-ture and elicit shivering, but one is reluctant toclassify this occurrence as an emotional event.In general, we use our “common sense” to de-fine an event as an emotional elicitor. Thus, forexample, the approach of a stranger or experi-ence on a visual cliff apparatus is usually theeliciting event for fear in the very young. Theapproach of a familiar parent is not used forfear, but is used as an elicitor to measure joy orhappiness. Events that we use to elicit particu-lar emotions grow out of our common experi-ences. Unfortunately, such experiences may notbe correct. As we can see in studies on fear, notall children show fear at a stranger’s approach(Lewis & Rosenblum, 1974).

The problem of the nature of elicitors be-comes even more serious when we try to mea-sure physiological reactions to emotionalevents. For example, in the presentation of ahorror film and the measurement of physiolog-

ical response to that horror film, one assumesthat the elicitor is a fearful one. What physio-logically appears is then taken as the responseto fear. When subjects are questioned as to thenature of the elicitor and what emotions it pro-duces, it is often the case that (1) they do notproduce the emotion believed associated withthe elicitor, or (2) they produce a diverse set ofemotional reactions. Schwartz and Weinberger(1980), for example, asked subjects what emo-tions they had to a set of different events andfound that they offered a variety of emotionsfor the same elicitor. Likewise, a colleague andI asked adults to mention the emotions pro-duced by such elicitors as going to the weddingof their child or the death of a parent, andfound that they gave a variety of emotions astheir responses (Lewis & Michalson, 1983).Such research as this suggests that we have lit-tle information, excluding our common experi-ence, in regard to the nature of emotionalelicitors.

While from a scientific viewpoint little infor-mation about which emotions are elicited bywhich stimulus events is available, it appears tobe the case that within a culture adult individu-als seem to possess common knowledge in re-gard to how they should react emotionally toparticular stimulus events. So, for example, atthe death of a friend’s parent, we know theemotion others either are likely to have or areexpected to show. The script learning of ap-propriate emotions by stimulus conditions,whether these are “true” emotions or simplyplay-acting, informs us that knowledge aboutemotional elicitors and appropriate emotionalresponses is something that is acquired. In re-views of this subject (see Harris, 1989; Lewis,1989), data on children’s acquisition of suchknowledge reveal that by the age of 10 yearschildren have a good sense of the appropriateemotions for appropriate stimulus elicitors.The learning of these emotional scripts appearsto take place quite early. For example, whenchildren were asked to choose which emotionalexpression was likely to go with a set of storiesthat included receiving gifts at a birthday party,being lost from your mother in a grocery store,and falling and hurting yourself, children asyoung as 3–5 years were already capable ofresponding appropriately—that is, the wayadults would respond to these same emotionalscripts (Lewis, 1989). Learning what is appro-priate vis-à-vis the culture is important for chil-dren, and they acquire such knowledge early.

18. The Emergence of Human Emotions 305

Serious child maltreatment results in failures inthis domain (Pollak, Cicchetti, Hornung, &Reed, 2000). The acquisition of such knowl-edge does not necessarily imply that the situa-tions do not produce the emotions that arecommonly believed to occur under such condi-tions. All that needs to be pointed out is thepossibility that specific stimulus events aremore likely to elicit some emotions rather thanothers. The emotion elicitors can be a functionof what a child has learned in terms of how tobehave, as well as a function of some automaticprocess whereby specific events elicit specificemotions.

Development of Elicitors

The problem of the development of elicitorsraises an important issue: Namely, what is therelation between emotions (states, experiences,and expression) and elicitors? Darwin (1872/1965), arguing for the bioevolutionary ap-proach, suggested that emotional behaviors oraction patterns are adaptive responses to spe-cific events (or elicitors). Thus one might sug-gest that anger is the result of a blocked goal,sadness the result of loss, and fear the result ofuncertainty or loss of control. Izard’s (1977)original differential-emotions theory held tothis view, although there may not be as close to aone-to-one relation as he proposed. To test thisview, my colleagues and I (Bennett, Bendersky,& Lewis, 2005; Lewis, Hitchcock, & Sullivan,2004; Lewis & Ramsay, 2005; Lewis, Ramsay,& Sullivan, 2006; Sullivan & Lewis, 2003) ex-amined 4-month-old infants’ response to frus-tration. Darwin argued that anger is the actionpattern that evolved to enable humans to over-come a barrier to a goal. After the infants in ourstudy were taught to pull a string to get to see apicture, and after they had learned the responseto the goal, the pull no longer resulted in the pic-ture’s coming on (frustration or goal blockage).Across many studies, we found that the actionpattern most frequently associated with thiselicitor was an anger-like face, increased pull-ing, increases in autonomic nervous system(ANS) response, and no increase in stress hor-mone. Children who showed not anger-likefaces but sadness-like faces showed no increasesin pulling nor in ANS response, but did showsignificant increases in stress hormone levels.Such findings support Darwin’s idea thatemotions—action patterns—evolved as adap-tive responses to particular events (elicitors).

Although these studies of anger in responseto a blocked goal appear clear in showing aperhaps innate connection between elicitor andaction pattern, more work with other elicitorsis needed. The difficulty is in the need to specifythe elicitor carefully—a problem that is notreadily addressed, and one that Gibson (1960)expressed concern about over 45 years ago inhis paper on the nature of the stimulus. Unlesswe can specify the elicitor carefully, much diffi-culty in understanding emotional behavior andits development will continue.

It would appear, therefore, that there is aclass of elicitors that has little developmentalhistory. A blocked goal causes angry actionpatterns in organisms throughout their lives.The looming of a visual event causes startle andattention, and perhaps fear as well. The sight offood always serves as a positive elicitor, if oneis hungry. It would therefore seem possible toimagine a class of events, biologically connect-ed, that would produce a particular emotionalstate. Even for this class of more automatic-seeming elicitors, the developmental events ofthe organisms may be such as to inhibit or re-strict the elicitors from operating in their natu-ral way.

In the class of elicitors with a developmentalcourse, the structure that supports the elicitor–response connection is likely to undergochange. Within this class are elicitors that arebiologically connected to a response, as well aselicitors that are connected to a responsethrough learned associations. For example, in-fants’ fear of strangers may be biologically pro-grammed; over time, stranger fear may declinebecause the biological structure supporting theelicitor–response connection has broken downor has been altered by experience. Learned as-sociations between elicitors and responses mayalso be subject to developmental change be-cause new structures are formed or old ones areextinguished; for instance, the formation ofnew structures can be predicated on cognitivechanges. The data from numerous sources sug-gest that important cognitive factors play a rolein mediating the effects of classes of events inthe elicitation of fear (see, e.g., Campos &Stenberg, 1981; Feinman & Lewis, 1984). Sev-eral of these cognitive processes are consideredhere, and more could probably be added to thelist. These capacities are regarded as criticaland serve as examples of the role that cognitivedevelopment may play in mediating the devel-opment of fear elicitors. First, memory must


play an important role in the elicitation of fear.Children must be able to recognize and associ-ate past events that were noxious. The whitecoats of doctors may be associated with painand thus acquire the capacity to elicit fear. Interms of cognitive expectancy, violation per sedoes not seem to be a fear elicitor. In fact, viola-tion of expectancy may be arousing, and theparticular emotion produced may depend onwhether the organism can assimilate and con-trol the event (Lewis & Goldberg, 1969). Someevents that are uncontrollable are likely to elicitfear as well (see Gunnar, 1980).

Tracing the developmental course of elicitorsis a difficult task to do. The development ofother cognitive processes—categorization, clas-sification, reasoning, and the like—is alsolikely to influence which elicitors produce whatemotional responses; for example, failure in atask produces sadness in children prior to 24months of age, while failure at a task after 24months is likely to produce shame or guilt, aswell as sadness. The same elicitor produces dif-ferent emotions, depending upon children’scognitive capacity. Before children can evaluatetheir actions against some standard, the successat achieving a goal results in happiness. Oncechildren are capable of this evaluation of self,the emotion as a consequence of the success islikely to be pride (Lewis, 1992b). Such findingsas these alert us to several problems concerningemotional elicitors. These include (1) that spe-cific elicitors may have an automatic biologi-cal adaptive connection to specific emotions,whereas others are connected through learnedassociations; (2) that individuals may differ inthe extent to which the same elicitor producesdifferent emotions; and (3) that the relation be-tween emotional elicitors and emotional out-comes changes as a function of the meaningsystem of a particular individual.

Emotional States

Emotional states are inferred constructs. Thesestates are defined as particular constellations ofchanges in somatic and/or neurophysiologicalactivity. Emotional states can occur without or-ganisms’ being able to perceive these states. In-dividuals can be angry as a consequence of aparticular elicitor and yet not perceive the an-gry state that they are in. An emotional statemay involve changes in neurophysiological andhormonal responses, as well as changes in fa-cial, bodily, and vocal behavior. As Darwin

(1872/1965) and more recently I (Lewis,2005a; Lewis et al., 2006) have discussed,states are considered action patterns thatinclude facial changes and physiological re-sponses.

Two views exist concerning emotional states.According to the first, states are associatedwith specific receptors; indeed, they constitutethe activation of these receptors (Izard, 1977;Tomkins, 1962, 1963). In the second, emo-tional states are not associated with specific re-ceptors and stimuli and do not exist as specificchanges; instead, they are general responsetendencies associated with specific cognitions(Mandler, 1975, 1980; Ortony, Clore, & Col-lins, 1988; Schachter & Singer, 1962).

In the first view, specific emotional states oraction patterns are postulated that have con-comitant physiological components and thatare expressed in specific facial and bodily be-haviors. There is a one-to-one correspondencebetween an emotion, such as anger, fear, sad-ness, or happiness, and some internal specificstate that matches this emotion. This view ofspecific emotional states has served, since Dar-win’s (1872/1965) initial formulation, as thebasis of what we believe to be the correspon-dence between the specific emotions we experi-ence and the functions of our bodies (seeNiedenthal, Chapter 36, this volume). Exceptfor bodily and facial expressions, no one-to-one correspondence has been found in adultsbetween such inferred physiological changesand emotions. Investigators exploring brainfunction (Davidson & Fox, 1982; Nelson &Bosquet, 2000; Nelson & Bloom, 1997) andthose looking at specific ANS changes (Ekman,1989) argue for some correspondence betweenspecific internal states and specific emotions.Even so, the evidence for specific states remainslacking.

The nonstate theories, cognitive in nature,argue less for a specific correspondence be-tween an internal state and emotions; rather,cognitive activity is seen as the determiner ofspecific emotions. Either general arousal mod-els, such as Schachter and Singer’s (1962)model, or cognitive theory models have as theirbasic tenet a denial of the existence of specificstates; rather, emotions occur as a consequenceof thinking (Elster, 1999; Lazarus, 2001;Ortony et al., 1988).

Specific states, having specific stimuli thatelicit them, have been suggested. The theory ofinnate releasing mechanisms (IRMs) suggests


that animals will show a fear response, given aparticular stimulus event. Thus one argumentis that there is a direct correspondence betweena specific elicitor and a specific state. Watson(1919) argued that there are specific elicitors ininfants; fear is produced by a falling sensationor by loud noises. Likewise, attachment theo-rists argued that children show joy or attach-ment to the object that takes care of them(Bowlby, 1969). Anger as a consequence of ablocked goal has been shown in animals andinfants.

On the other hand, it is quite clear that cer-tain specific emotions can be produced onlythrough cognitive processes. For example, cer-tain elicitors invoke cognitive processes, whichin turn may elicit or produce specific emotionalstates. In such cases, cognition is necessary forthe elicitation of a specific state, but may not bethe material of that state. Consider the emotionof shame. One must have certain cognitions forshame to occur. Shame occurs when personsevaluate their behavior against some standardand find that they have failed (Lewis, 1992band Chapter 46, this volume). Such cognitionscan lead to a specific emotional state, which islikely to have specific bodily activity. In thesecases cognitions are the elicitors of the emo-tion, the state being brought into existence bythoughts; thus there are many possibilities:

1. An emotional state or action pattern canbe elicited in some automatic fashion by certainstimulus events—for example, the case of fearwhen an animal sees a predator (IRM).

2. Emotional states can be elicited throughcognitive evaluative processes. They may beautomatic—that is, may be action patterns tocertain thoughts. Darwin believed that the self-conscious emotions are elicited by thoughtsabout others’ thinking about us. Plutchik(1980) and I (Lewis, 1992b) have argued fordistinguishing between different emotionalstates by using the difference between the levelsof cognitive activity involved in their elicita-tion. Fear of falling is elicited by little cogni-tion, shame by much cognition.

3. Alternatively, there are no specific emo-tional states or action patterns, but onlygeneral arousal, which is interpreted vis-à-visthe events surrounding the arousal (Clore &Ortony, 1991; Schachter & Singer, 1962). Inthis model, there is an emotional state, but notspecific action patterns associated with it.

What is clear is that even if there are specificemotional states, they may bear little corre-spondence to our emotional lives—either emo-tional expressions or our experience of emo-tions. So, for example, it may be quite possibleto have a specific emotional state but to be un-aware of it, ignore it, or even deny it. Likewise,we may have a specific emotional state butchoose not to express it. Thus, for example, Imay be angry at my dean for not giving me araise, but I am not likely to express that angerwhen I see her. Emotional states, then, are in-ferred, and whether they are specific, general,or nonexistent awaits further research.

If we hold to the existence of emotionalstates, then for the most part they must beviewed as transient action patterns that are al-terations in ongoing levels of neurophysio-logical and/or somatic activity. These transient,ongoing changes in our neurophysiological andsomatic activity levels imply that there is a con-stant stream of change. It becomes difficult toimagine, therefore, being awake and not beingin some emotional state or some level ofarousal. However, since there need not be anycorrespondence between an emotional stateand emotional experience and expression,there is no reason to assume that we are awareof the states that we are in. This does not meanthat these states are not affecting our ongoingbehavior—only that they are not apparent(Lewis, 1991).

Development of Emotional States

In a discussion of the developmental issues per-taining to emotional states, two issues need tobe addressed. The first concerns the nature ofthe different states and how they are derived;the second pertains to the developmentalcourse of states once they emerge. For example,if emotional states are viewed as specific, thequestion of how specific states develop needs tobe addressed. Two general models are possible.According to one, specific emotional states arederived from developmental processes. Suchprocesses may be purely maturational, or theymay be interactive, involving the organismwith its environment. The second model doesnot depict a role for development in the emer-gence of specific states; rather, discrete emo-tional states are assumed to be innate.

In the first model, the infant has two basicstates or one bipolar state at birth: a negative


or distress state and a positive or satiated state.Subsequent states emerge through the differen-tiation of this basic bipolar state. Differentia-tion theories focus on the modulation of boththe bipolar state and general arousal state.Hedonic tone and arousal may be the dimen-sions necessary to generate specific emotionalstates. This idea was proposed by Bridges(1932) and is considered a differentiation hy-pothesis. This theory has been adopted by oth-ers, including Spitz (1965), and more recentlySroufe (1996).

The way in which the interface of arousaland hedonic tone develops into specific emo-tional states remains speculative. It has been ar-gued that both mother–child interaction andmaturation underlie the process of differen-tiation (Als, 1975; Brazelton, Koslowski, &Main, 1974; Sander, 1977). The regulation ofthe child’s state may be the mechanism leadingto differentiation. Although some theoristsstress that emotional differentiation is deter-mined more by biological than by interactivefactors, the combination of the two forcesseems most likely. While such a theory is ap-pealing, the derivation of specific emotionalstates remains without empirical support.

A much simpler developmental model con-cerning differentiation can be considered froma purely biological perspective. Such a biologi-cal model can be imagined in which undifferen-tiated emotion becomes differentiated as afunction of maturation. According to such aview (see Lewis & Michalson, 1983), the rateof differentiation and the unfolding of differen-tiated emotion states are programmed accord-ing to some physiological timetable. The differ-entiation from general to specific structures is acommon process in morphology; there is noreason not to consider such a possibility inemotional development. The most likely expla-nation of emotional development is the differ-entiation of emotion states that occurs as afunction of maturation, socialization, and cog-nitive development. Whatever processes under-lie this differentiation, the model is develop-mental in nature.

An alternative model is that some discretestates are preprogrammed in some sense andneed not be further differentiated (Izard,1978). They exist at birth, even though theymay not emerge until a later point in develop-ment. The view is unlike the differentiationmodel in that discrete emotional states do not

develop from an original undifferentiated state,but are innate at birth in already differentiatedform. In this discrete-systems model, specificemotion states emerge either in some predeter-mined order or as needed in the life of the in-fant. They may co-occur with the emergence ofother structures, although they are independentof them. The emotional system essentially op-erates according to biological directives.

These different models address the concep-tual difference between experience and struc-ture found in the arguments of Hume andKant. In one case, experience produces a struc-ture (Hume, 1739/1888). In the other case, ex-perience is assimilated into innate structures(Kant, 1781/1958). In the study of emotionaldevelopment, the question is whether emo-tional states are preformed and depend only onthe development of cognitions, or whethercognitions themselves produce the emotionalstates or structures. Such a distinction is ratherfine, but has important theoretical implica-tions. Such a distinction can be seen in thestudy of fear: Is each fear state the same asother fear states, regardless of the circum-stances, or do fear states differ as a function ofthe elicitors? For example, is the fear state pro-duced by a loud noise the same as the fear stateproduced by the association of a doctor’s whitecoat with the pain of a needle? Are emotionalstates independent of or dependent on particu-lar cognitions? If emotional states are inde-pendent, they need not be created by thecognitions.

The first issue in the development of statesconcerns the origin of discrete emotional states.The second issue focuses on the developmentalchanges in emotional states once they haveemerged. For example, 8-month-old childrenmay show behaviors reflecting fear at the ap-pearance and approach of strangers, and 2-year-old children may exhibit fear behaviorswhen they have broken their parents’ favoritelamp. Do similar fear states underlie the fearexpressions in both cases? Although theelicitors of states and the children’s cognitivecapacities are different in these two cases, theunderlying emotional states may be similar.

Major developmental changes may occur in(1) the events that produce emotional states,(2) the behavioral responses used to referencestates, and (3) the cognitive structures of chil-dren. Whether the emotional state itselfchanges as a function of development is diffi-


cult to determine. However, there may well beimportant physiological and neural changesthat differentiate young and old organisms.Given that important physiological changes oc-cur over age, the physiological processes asso-ciated with emotional states may change overtime. If this were the case, then the consistencyof an emotion may be a function of our experi-ence of it more than the underlying state. Whatis clear and what will be shown below is thatthe appearance of particular emotions may de-pend on new cognitions, as well as the fact thatnew cognitions may allow for the developmentof new emotions. The former case can be seenagain in the example of fear. While 1-year-oldinfants may fear falling off a “visual cliff,” theydo not fear failing an exam or being caughtcheating on their income tax. Such fears in anadult are due to elaborate social and cognitivedevelopment. An example of the latter—that is,cognitions producing new emotions—has to dowith classes of emotions called “self-consciousevaluative emotions.” These emotions, such aspride and shame, cannot occur until elaboratecognitive processes have occurred (see Stipek,Recchia, & McClintic, 1992). Althoughcognitions are related to new emotions, the ma-terials of these emotional states or action pat-terns are not likely to be the materials of theemotional states. As such, it is hard to believethat experiences or even cognitions (thoughts)create the states. If that is the case, than statesare a feature of human life having an evolution-ary developmental history.

Emotional Expressions

Emotional expressions are those potentiallyobservable surface changes in face, voice, body,and activity level. Emotional expressions areseen by some as the manifestations of internalemotional states (Ekman & Friesen, 1974;Levenson, Ekman, & Friesen, 1990). In fact,no single measure of emotional states or actionpatterns is more differentiating than emotionalexpressions. The problem with emotional ex-pressions is that they are soon capable of beingmasked, dissembled, and in general controlledby an individual (Saarni, 1999). Moreover,emotional expressions are subject to wide cul-tural and socialization experiences. Thus therelationship between expressions and states re-mains somewhat vague (Lewis & Saarni,1993). The measurement of emotional expres-sion is reviewed in detail in other chapters, so I

spend little time on the definition of emotionalexpression except to make several points.

First, emotional expressions tend to be stud-ied in terms of facial expression, and whilebody postures have been studied (see, e.g.,Argyle, 1975), the study of children’s emo-tional expressiveness in terms of body posturesand activity has received little attention. Vocal-izations are one of the least understood aspectsof emotional expression, although they seem tobe important conveyors of emotional states. In-deed, vocal expressions are extremely powerfuland may have the ability to elicit similar emo-tional states in others. Vocalizations may bemuch more contagious than facial or bodily ex-pressions. For example, movies are much fun-nier when seen with others who laugh out loudthan when seen by oneself. Because of the con-tagious nature of vocalization, vocal expres-sion may be the target of early socialization ef-forts. Crying is a case in point. Crying behavioris quickly brought under control as parents so-cialize their children not to cry when distressedor in need. Locomotion may be another modeof expressing emotions. For example, movingaway from and moving toward an object arephysical responses associated with differentemotional states (Schneirla, 1959). Indeed, aninfant’s movement away from an unfamiliartoy or person, independent of facial expression,is often used to reference fear (Schaffer, Green-wood, & Parry, 1972). Interestingly, movementtoward can be part of two different action pat-terns: It can reflect either joy or anger.

Although there are some data on emotionalexpressions in each of these four modalities (fa-cial, postural, vocal, and locomotor), the rela-tions among them have received almost no atten-tion. It seems reasonable to assume thatsobering, crying, and running away form a co-herent response that reflects the emotional stateof fear. The particular modality used to expressan emotion may be a function of specific rules ofsocialization or of a response hierarchy in whichone modality has precedence over another. Sucha hierarchy may be determined either by a set ofbiological imperatives or by a set of socializationrules. The use of one or more channels to expressa particular emotion may be determined by acomplex set of interactions. One issue of particu-lar interest is the effect on some expressionswhen one modality is inhibited. Inhibition in aparticular modality can be experimentally pro-duced, for example, by preventing a child frommoving about. For instance, if children are pre-


vented from running away from an approachingstranger because they are restrained in a high-chair, they may express their internal state moreintensely through alternative means, such aschanges in facial musculature. Another exampleof the use of differential modalities in expressingemotions occurs in the work on stress. We(Lewis, Ramsay, & Kawakami, 1993), for ex-ample, found that infants who do not expressdistress when pained by inoculation are morelikely to show large adrenocortical responses.Suomi (1991) and Levine and Wiener (1989)have found that monkeys that do not show loudcries of distress upon being separated from theirmothers are much more likely to show higheradrenocortical responses. Thus the relationsamong modalities of expression may play an im-portant role in determining what emotional ex-pressions are presented and the intensity ofthem.

Development of Emotional Expressions

The question concerning the development ofemotional expressions takes many forms. Firstand foremost is the question of whether a par-ticular emotional expression—the synchroniza-tion of a set of particular facial musculaturemovements—appears de novo or has a devel-opmental course (see Camras & Fatani, Chap-ter 17, this volume). Izard (1977), for one, ar-gued for their appearing as part of the adaptiveevolutionary history of humans, while others(Camras, Lambrecht, & Michel, 1996; Oster,2005) have shown changes in the facialneuromusculature synchronization. This con-troversy is all the more interesting, given thatthere are no theories suggesting how the syn-chronization takes place except for proposalsof some environmental organizer. There are nodata on how environments organize these mus-cle movements, although there is a suggestionthat environments affect individual differencesin facial expressions—that is, in the likelihoodof an expression, given the same elicitor. Forexample, some infants show more joy than oth-ers at the approach of their mothers, and thisdifference may be a function of the infants’interaction with their environments. Even so,there is no evidence for environments’ organiz-ing the neuromusculature. Alternatively, thebelief in a connection between elicitors andspecific action patterns leads to the belief thatthe action pattern itself is a given, establishedby adaptive evolutionary processes.

Another issue related to the facial muscula-ture question is the nature of the associationamong facial, vocal, and physiological behav-iors that make up an action pattern. As de-scribed earlier, my colleagues and I have shownthat an anger-like face, increased action towarda goal, lower stress response, and greater ANSorganization appear to be such an action pat-tern evolved as an adaptive emotional responseto the blockage of a goal. How does the organi-zation of multiple modalities come about? Thefull developmental course of emotional expres-sions, then, is uncharted.

We are much more inclined to believe thatparticular emotional expressions reflect a spe-cific underlying state in infants and young chil-dren when we see particular faces in particularcontexts. Thus, for example, when childrenshow wary or fearful faces at the approach of astranger, we are more apt to credit those facesas meaning that the children are in a fear statethan if the children show those same faces to-ward their mothers, who are sitting next tothem. Faces, expressed in the context of partic-ular situations, lend validating meaning to theconnection between facial expression andinternal state. Nevertheless, the question ofwhether a facial expression truly reflects anemotional state cannot be readily answered,except through phenomenological report.

Emotional Experiences

Emotional experiences are the interpretationsand evaluations by individuals of their per-ceived situations, emotional states, and expres-sions. Emotional experiences require that indi-viduals attend to their emotional states (i.e.,changes in their neurophysiological behavior),as well as the situations in which the changesoccur, the behaviors of others, and their ownexpressions. Attending to these stimuli is nei-ther automatic nor necessarily conscious. Anemotional experience may not occur because ofcompeting stimuli to which the organism’s at-tention is drawn. For example, consider the fol-lowing scenario: The car a woman is drivingsuddenly has a blowout in the front tire; the carskids across the road, but the woman succeedsin bringing it under control and stopping thecar on the shoulder. Her physiological state aswell as her facial expression may indicate thatwhile she is bringing the car under control, herpredominant emotional state is fear. Becauseher attention is directed toward controlling the


car, however, she is not aware of her internalstate or of her expressions. She only experi-ences fear after she gets out of the car to exam-ine the tire. Emotional experiences thus requirepeople to attend to a select set of stimuli. With-out attention, emotional experiences may notoccur, even though an emotional state may ex-ist. Many other examples are possible. Fromthe clinical literature, a patient may be in a par-ticular emotional state (e.g., depression), butmay attend to select features of that state (e.g.,fatigue), and so may only experience tiredness.Or a patient may not experience pain at thedentist when distraction is provided throughthe use of earphones and loud music.

An emotional experience may not necessar-ily be conscious, either. If one is willing to dis-tinguish between explicit and implicit experi-ences, emotional experiences may occur atdifferent levels of consciousness. Such an anal-ysis forms the basis of much psychoanalyticthought. For example, individuals may be in anemotional state of anger. That is, with propermeasurement techniques, one would find a pat-tern of internal physiological responses indica-tive of anger. Moreover, these persons may acttoward those objects that or persons who havemade them angry in a way that suggests theyare intentionally behaving in response to an in-ternal state of anger. Nonetheless, the personsmay deny that they feel anger or are acting inan angry fashion. Within the therapeutic situa-tion, such people might be shown that (1) theyare angry, and (2) they are responding inten-tionally as a consequence of that anger. Thetherapeutic process may further reveal that un-conscious processes are operating in a fashionparallel to conscious ones. Defense mecha-nisms, for example, function to separate levelsof awareness. Although awareness may not beat a conscious level, unconscious awarenessmay still exert powerful effects. Slips of thetongue, accidents, and classes of unintentionalconscious behavior may all be manifesting in-tentional unconscious awareness (Freud, 1901/1960). Thus people may experience their inter-nal states and expressions and be aware of thisexperience, or they may experience them in anunconscious mode in which the conscious per-ception of the experience is unavailable.

Up until this point, we have assumed thatthere exists an internal state that is experi-enced. As some have argued, the experiencingof an emotion does not have to rely upon anyinternal state at all. In fact, no internal state

may exist. For those who do not believe in theconstruct of a unique set of variables marking aspecific state (Clore & Ortony, 1991; Ortony etal., 1988), the experience of the state is nothingmore than a cognitive construction, utilizingsuch perceptions as the nature of the experi-ence, past history, the responses of others, andso on. Under such a view, emotional experi-ences are the unique and specific states them-selves. From a cognitive-constructive per-spective, such a view of emotions is quitereasonable. In fact, data on patients with spinalinjuries suggest that emotional experiences canoccur without specific physiological states. Forexample, patients with such injuries who areincapable of receiving neural messages frombelow the waist report sexually orgasmic expe-riences, even though no state information isavailable to them. They construct the experi-ence from their past knowledge and not fromany change in their neurophysiological state.

Emotional experiences occur through the in-terpretation and evaluation of states, ex-pressions, situations, behaviors of others, andbeliefs about what ought to be happening.Emotional experiences therefore depend oncognitive processes. Cognitive processes in-volving interpretation and evaluation are enor-mously complex and involve various percep-tual, memory, and elaborating processes.Evaluation and interpretation not only involvecognitive processes that enable organisms toact on information, but are very much depen-dent on socialization to provide the content ofthe emotional experience. The particular so-cialization rules are little studied and not wellunderstood (Lewis & Saarni, 1985; Saarni,1999).

Not all theories of emotional experienceneed be tied to the context, nor do all suggestthat there is an underlying emotional state.However, all emotional experience does in-volve an evaluative interpretive process, includ-ing the interpretation of internal states, con-text, behavior of others, and meaning given bythe culture.

Development of Emotional Experiences

The development of emotional experiences isone of the least understood aspects of emotion.Emotional experiences require an organism topossess some fundamental cognitive abilities,including the ability to perceive and discrimi-nate, recall, associate, and compare. Emotional


experiences also require a particular cognitiveability—that is, the development of a conceptof self. Emotional experiences take the linguis-tic form “I am frightened” or “I am happy.” Inall cases, the subject and object are the same:that is, oneself. Until an organism is capable ofobjective self-awareness (Duval & Wicklund,1972), the ability to experience may be lacking.Emotional experience requires both generalcognitive capacities—something I touch uponbelow—and the specific cognitive capacity ofself-referential behavior, or what I have re-ferred to as “consciousness” (Lewis, 1992b,2003b).

General cognitive processes necessary for or-ganisms to perceive and discriminate elicitorsof particular behaviors (whether these be inter-nal or external to them), as well as overt emo-tional expressions of themselves and of others,have a developmental course. For example, in-fants younger than 6 months are generally un-able to discriminate between facial patternsand do so on the basis of discrete features(Caron, Caron, & Myers, 1982). Schaffer(1974) demonstrated that children cannotmake simultaneous comparisons prior to 7 or 8months of age. This would suggest that infantsare not capable of experiencing emotions priorto this point. Moreover, some emotional expe-riences may require a higher level of cognitiveprocessing than others, and some are likely todevelop earlier than others. For example, fearprobably emerges earlier than shame, since theformer requires less cognitive and evaluativeprocessing than the latter (Lewis, 1992b).

If emotional experience is the consequenceof an evaluation of one’s bodily changes, andalso of the context and the behaviors of others,then two processes are necessary for most emo-tional experiences: (1) the knowledge that thebodily changes are uniquely different fromother changes (i.e., they are internal rather thanexternal), and (2) the evaluation of thesechanges. The internal–external distinction foremotional development is important, because itaddresses the differences between experienceand expression. If we believe that facial expres-sion is equivalent to an emotional state or ex-perience, then it is possible to infer an internalevent by examining its external manifestation.If, however, we do not subscribe to the view ofa one-to-one correspondence between expres-sion and experience, then all we can say is thatthere is an external manifestation of someunperceived internal event. Emotional experi-

ences, by nature, are internal events. Moreover,the internal and external distinction can onlybe carried out by a self capable of making thedistinction between the self and the other. Suchevaluation may involve the process of self-awareness.

Self-awareness is an information-processingand decision-making event related to internalstimuli. It logically requires an organism topossess the notion of agency (Lewis, 2003a).The term “agency” refers to that aspect ofaction that makes reference to the cause ofthe action—not only who or what is causingthe stimulus to change, but who is evaluatingit. The stimulus change itself may have the ef-fect of alerting the organism and forcing it tomake some type of evaluation. Emotional ex-perience requires that the organism be capa-ble of attending to itself. Thus the statement“I am happy” implies two things: first, that Ihave an internal state called happiness, andsecond, that I perceive that internal state ofmyself. Until organisms are capable of thiscognitive capacity, they should not be capableof emotional experiences (Lewis, 2003b,2005b; Lewis & Brooks-Gunn, 1978, 1979).This does not mean that infants, prior to ac-quiring an objective self or consciousness ofthe self, do not have unique emotional states;they do. What seems reasonable to postulateis that an individual can be in a particularemotional state and yet not experience it. Justas we have seen in the example of the womanwhose car slides off the road, an emotionalstate can exist without experience, so we canimagine an infant’s having an emotional statewithout being able to experience it. Thisleads to the rather peculiar proposition that achild can be in a state of pain or can be in astate of fear, yet not experience that state, ifby “experiencing it” we mean being able tomake reference to the self as having thatstate. In a series of studies, my collaboratorsand I have demonstrated that the emergenceof this self-conscious process does not occurprior to 15 months of age, and that it seemsto emerge mostly as a function of maturationin the second half of the second year of life.It is only then that children can be both in aparticular emotional state and can be said toexperience that state. Moreover, the produc-tion of certain states requires self-awareness;therefore, certain emotions are unlikely to oc-cur until this cognitive process emerges(Lewis, Sullivan, Stanger, & Weiss, 1989).


Once the basic cognitive processes that allowfor objective self-awareness or consciousnessoccur, organisms are capable of experiencingemotions. As I have pointed out, they maybe capable of experiencing existing emotionalstates as well as capable of experiencing emo-tions that have no internal state, either becauseinternal emotional states do not exist or be-cause the organisms are experiencing a differ-ent emotional state from that which exists. Therules that govern how we experience our emo-tional states or how we create emotional expe-riences themselves are complex and varied.Clearly, socialization rules are involved, on acultural as well as on a familiar or individuallevel. For example, in cultures that do nottolerate interpersonal aggression—Japan, forexample—the experiencing of anger is cultur-ally inappropriate. It may be the case that Japa-nese children or adults may act in an angry wayand may even have an emotional state of anger.However, because having such a state is inap-propriate, they are not likely to have the emo-tional experience of anger. Exactly how the so-cialization process proceeds so as to influence,modify, alter, or accent emotional experiencesis little understood. Clearly, the topic of the so-cialization of emotion involves the socializa-tion of at least all four features of emotion dis-cussed here. It affects the meaning of stimuliand what we allow events to do in terms of act-ing as elicitors of particular emotions. It affectsthe emotional-expressiveness dimension ofemotion, and, finally, it affects the emotionalexperience.

From an interpersonal and intrapersonalpoint of view, the socialization rules that act onthe experiencing of emotion are somewhatbetter articulated. Freud’s theory of the uncon-scious and of defense mechanisms addressesthis point. Defense mechanisms have as theirchief function preventing individuals from ex-periencing emotions or, alternatively, from hav-ing emotions that they do not like to have. Forexample, denial and repression serve the func-tion of preventing people from having particu-lar emotional experiences that they deem unac-ceptable. They prevent it by not allowing themto become conscious or self-aware. Projection,on the other hand, allows for the experiencingof the emotion—not as the self’s experiencingit, however, but as the self’s experiencing it inanother. As we can see in each defense mecha-nism, the major function is to provide meansfor altering emotional experience.

REINTEGRATING EMOTIONAL LIFE

In the preceding discussion, I focused upon spe-cific features of emotional life in order to seehow the developmental process can affect eachof these components. Unfortunately, the focuson individual features does a disservice to thecomplexity of emotional life. Moreover, it doesnot allow us to look at developmental issuesthat may be related to the relationships be-tween various components, such as the relationbetween emotional expression and state (Lewis& Michalson, 1983). Very early in life, emo-tional expressions and states may have littlecorrespondence. At some point in develop-ment, there appears to be some coherencebetween emotional states and expressions—ayoung child smiling with joy at someone’s joke.With socialization and further development,the disassembling of expression from internalstate takes place. Children very quickly learn todetach expression from states, and thus to dis-semble: Children as young as 21

2 years of ageare quite capable of successfully lying aboutcommitting a transgression, through verbalresponse as well as through facial response(Lewis, 1993). Thus there is a developmentalcourse in the connection between expressionand state. A similar analysis can be made forthe coherence between internal state and expe-rience. Earlier in the developmental process,children may have internal states that they donot experience. There may then be a period inwhich internal states and experiences formsome coherence, only to change once again sothat experiences of emotion can take placewithout internal states. These developmentalsequences in the coherence between features ofemotional life need more careful study.

A MODEL OF EMOTIONALDEVELOPMENT

In what follows, I present a model of the emer-gence of different emotions over the first 3years of life. I choose this period because it rep-resents the major developmental leap of themajority of adult emotions in emotional devel-opment. This is not to say that past 3 years ofage other emotions do not emerge, or that theemotions that have emerged are not elaboratedmore fully. I suspect that both are the case.

One problem with articulating a model ofthe emergence of emotional life has to do with


the appropriate markers for the emotions. Arewe making reference solely to emotional ex-pressions, or are we talking about emotionalstates or experiences? The ability to do morethan observe the emitted behaviors of the childis needed, but behavioral observation is oftenall that is possible. In order to get at emotionalexperiences, we need language in the form of “Iam sad” or “I am ashamed.” Since during thisperiod the language of the child is quite limited,the study of emotional experience is difficult.Likewise, the study of emotional-state develop-ment is difficult because there has been littlesuccess to date in finding unique configura-tions of neurophysiological measures thatmark unique emotions in adults, let alone chil-dren and infants.

What we are left observing are emotional ex-pressions and behaviors in context. Observa-tion of a behavior in context allows us, at leastfrom the adult meaning system, to assume thatthe child’s expression reflects an emotion.Observation of fear over the approach of astranger, or joy when a mother appears, allowsus to accept that an internal state of fear or joyexists. With these limitations in mind, the fol-lowing discussion and mapping of emotionaldevelopment can take place.

Following Bridges (1932), as well as others,we assume that at birth the child shows a bipo-lar emotional life. On the one hand, there isgeneral distress marked by crying and irritabil-ity. On the other hand, there is pleasure markedby satiation, attention, and responsivity to theenvironment. Attention to/interest in the envi-ronment appears from the beginning of life,and we can place this in the positive pole; or, ifwe choose, we can separate this, thus suggest-ing a tripartite division with pleasure at oneend, distress at the other, and interest as a sepa-rate dimension (Lewis & Michalson, 1983; seeFigure 18.1).

By 3 months, joy emerges. Infants start tosmile and appear to show excitement/happi-ness when confronted with familiar events,such as faces of people they know or even unfa-miliar faces. Also by 3 months, sadnessemerges, especially around the withdrawal ofpositive stimulus events. Three-month-old chil-dren show sadness when their mothers stop in-teracting with them. Disgust also appears in itsprimitive form—a spitting out and getting ridof unpleasant-tasting objects placed in themouth. Thus by 3 months children are alreadyshowing interest, joy, sadness, and disgust, and

exhibiting these expressions in appropriatecontexts.

Anger has been reported to emerge between4 and 6 months (Stenberg, Campos, & Emde,1983). Anger is manifested when children arefrustrated—in particular, when their hands andarms are pinned down and they are preventedfrom moving. However, we (Lewis et al., 1990)have shown anger in 2-month-old infants whena learned instrumental act was removed. Thisstudy demonstrates the earliest known emer-gence of anger. Anger is a particularly interest-ing emotion, since, from Darwin (1872/1965)on, it has been associated with unique cognitivecapacity. Anger is thought to be both a facialand motor/body response designed to over-come an obstacle. Notice that in this definitionof anger the organism must have some knowl-edge regarding the instrumental activity to-ward a goal. For anger to be adaptive, it mustbe a response that attempts to overcome a bar-rier blocking a goal. In some sense, then,means–ends knowledge must be available, andthe demonstration of anger at this early pointin life reflects the child’s early knowledge ac-quisition relative to this ability (Lewis, 1991).

Fearfulness seems to emerge still later. Again,fearfulness reflects further cognitive develop-ment. Schaffer (1974) has shown that in orderfor children to show fear, they must be capableof comparing the event that frightens themwith some other event; for instance, in strangerfear, infants have to compare the face of astranger to their internal representation ormemory of faces. Fear occurs when the face isfound to be discrepant or unfamiliar relative toall other faces that the child remembers.Children’s ability to show fearfulness, there-fore, does not seem to emerge until this com-parison ability emerges. Children begin toshow this behavior at about 7–8 months, al-though it has been reported by some to occureven earlier, especially in children who seem tobe precocious. In the first 8–9 months of life,children’s emotional behavior reflects the emer-gence of the six early emotions, called by some“primary emotions” or “basic emotions” (see,e.g., Izard, 1978; Tomkins, 1962).

Surprise also appears in the first 6 months oflife. Children show surprise when there are vio-lations of expected events; for example, wheninfants see a midget (a small adult) walking to-ward them, they are reported to show interestand surprise rather than fear or joy (Brooks &Lewis, 1976). Surprise can be seen either when


there is violation of expectancy or as a responseto discovery, as in an “Aha!” experience. We(Lewis, Sullivan, & Michalson, 1984) showedthat when children were taught an instrumen-tal arm-pulling response, they showed surpriseat the point when they discovered that the armpull could turn on a slide. Surprise can reflecteither a violation or a confirmation of expec-tancy. Cognitive processes play an importantrole in the emergence of these early emotions,even though the cognitive processes are limited;this is not so for the next class of emotions.

Figure 18.1 indicates that a new cognitivecapacity emerges somewhere in the second halfof the second year of life. The emergence ofconsciousness or objective self-awareness (self-referential behavior) gives rise to a new classof emotions. These have been called “self-conscious emotions” and include embarrass-ment, empathy, and envy. Although little workexists in the development of these emotions,

several studies support the emergence of em-barrassment at this point in development. We(Lewis et al., 1989) have shown that the emer-gence of embarrassment only takes place afterconsciousness or self-recognition occurs.Empathy, too, emerges in relation to self-recognition (Bischof-Köhler, 1991).

Two points are to be noticed about this classof emotions. First, the observation of theseemotions requires measuring not only facial ex-pressions, but also bodily and vocal behaviors.Whereas the earlier emotions can be observedreadily in specific facial configurations, thesenew emotions require measurement of bodilybehaviors. Embarrassment, for example, is bestmeasured by nervous touching, smiling, gazeaversion, and return behaviors. The second im-portant point related to the emergence of theseemotions is that while they reflect self-consciousness, they do not require self-evaluation. The emergence of these self-


FIGURE 18.1. Development of emotions over the first 3 years of life.

conscious emotions is related uniquely to thecognitive milestone of paying attention to theself. This topic is taken up in more detail in an-other chapter (see Lewis, Chapter 46, this vol-ume).

Figure 18.1 also shows a second cognitivemilestone, which occurs sometime between 2and 3 years of age. This ability is characterizedby children’s capacity to evaluate their behav-ior against a standard; the standard can be ei-ther external, as in the case of parental orteacher sanction or praise, or internal, as in thecase of children’s developing their own stan-dards. This capacity to evaluate personalbehavior in relation to a standard develops inthe third year of life, and it gives rise to anotherset of emotions. We have called these “self-conscious evaluative emotions”; they includepride, shame, and guilt, among others. Theseemotions require that children have a sense ofself and be capable of comparing their ownbehavior against standards. If children fail vis-à-vis the standard, they are likely to feel shame,guilt, or regret. If they succeed, they are likelyto feel pride (Lewis, 1992a). It is important tonote that pride and shame are quite differentfrom happiness and sadness. For example, wecan win a lottery and feel quite happy aboutwinning the money; however, we would notfeel pride, because we would not view the win-ning of the lottery as having anything to dowith our behavior. The same is true for failure;we might feel sad if we were not able to dosomething, but if it was not our fault, then wewould not feel shame or guilt. These complexsocial-evaluative emotions make their appear-ance at about 3 years of age (see Lewis, 1992b;Stipek et al., 1992).

Thus, by 3 years of age, the emotional life ofa child has become highly differentiated. Fromthe original tripartite set of emotions, the childcomes within 3 years to possess an elaborateand complex emotional system. While the emo-tional life of the 3-year-old will continue to beelaborated and will expand, the basic struc-tures necessary for this expansion have alreadybeen formed. New experiences, additionalmeaning, and more elaborate cognitive capaci-ties will all serve to enhance and elaborate thechild’s emotional life. However, by 3 years ofage, the child already shows those emotionsthat Darwin (1872/1965) characterized asunique to our species—the emotions of self-consciousness. With these, the major develop-mental activity has been achieved.

REFERENCES

Als, H. (1975). The human newborn and his mother: Anethological study of their interaction. Unpublisheddoctoral dissertation, University of Pennsylvania.

Argyle, M. (1975). Bodily communication. New York:International Universities Press.

Bennett, D. S., Bendersky, M., & Lewis, M. (2005).Does the organization of emotional expressionchange over time?: Facial expressivity from 4 to 12months. Infancy, 8(2), 167–187.

Bischof-Köhler, D. (1991). The development of empathyin infants. In M. E. Lamb & H. Keller (Eds.), Devel-opment: Perspectives from German-speaking coun-tries (pp. 245–273). Hillsdale, NJ: Erlbaum.

Bowlby, J. (1969). Attachment and loss: Vol. 1. Attach-ment. New York: Basic Books.

Brazelton, T. B., Koslowski, B., & Main, M. (1974).The origins of reciprocity: The early mother–infantinteraction. In M. Lewis & L. A. Rosenblum (Eds.),The effect of the infant on its caregiver (pp. 49–76).New York: Wiley.

Bridges, K. M. B. (1932). Emotional development inearly infancy. Child Development, 3, 324–334.

Brooks, J., & Lewis, M. (1976). Infants’ responses tostrangers: Midget, adult and child. Child Develop-ment, 47, 323–332.

Campos, J., & Stenberg, C. (1981). Perception, apprais-al, and emotion: The onset of social referencing. InM. E. Lamb & L. R. Sherrod (Eds.), Infant socialcognition: Empirical and theoretical considerations(pp. 273–314). Hillsdale, NJ: Erlbaum.

Camras, L. A., Lambrecht, L., & Michel, G. F. (1996).Infant “surprise” expressions as coordinative motorstructures. Journal of Nonverbal Behavior, 20(3),183–195.

Caron, R. F., Caron, A. J., & Myers, R. S. (1982). Ab-straction of invariant face expressions in infancy.Child Development, 53, 1008–1015.

Clore, G. L., & Ortony, A. (1991). What more is thereto emotion concepts than prototypes? Journal of Per-sonality and Social Psychology, 60, 48–50.

Darwin, C. R. (1965). The expression of the emotions inman and animals. Chicago: University of ChicagoPress. (Original work published 1872)

Davidson, R. J., & Fox, N. A. (1982). A symmetricalbrain activity discriminate between positive versusnegative affective stimuli in human infants. Science,218, 1235–1237.

Duval, S., & Wicklund, R. A. (1972). A theory of objec-tive self-awareness. New York: Academic Press.

Ekman, P. (1989). The argument and evidence aboutuniversals in facial expressions of emotion. In J. Wag-ner & A. Manstead (Eds.), Handbook of socialpsychophysiology (pp. 143–164). New York: Wiley.

Ekman, P., & Friesen, W. V. (1974). Detecting deceptionfrom the body or face. Journal of Personality and So-cial Psychology, 29, 288–298.

Elster, J. (1999). Alchemies of the mind. Cambridge,UK: Cambridge University Press.


Feinman, S., & Lewis, M. (1984). Is there social life be-yond the dyad?: A social-psychological view of socialconnections in infancy. In M. Lewis (Ed.), Beyond thedyad (pp. 13–41). New York: Plenum Press.

Freud, S. (1960). The psychopathology of everyday life(A. Tyson, trans.). New York: Norton. (Originalwork published 1901)

Gibson, J. J. (1960). The concept of the stimulus in psy-chology. American Psychologist, 15, 694–703.

Gunnar, M. R. (1980). Control, warning signals anddistress in infancy. Developmental Psychology, 16,281–289.

Harris, P. (1989). Children and emotion: The devel-opment of psychological understanding. Oxford:Blackwell.

Hume, D. (1888). A treatise of human nature (L. A.Selby-Bigge, Ed.). Oxford: Clarendon Press. (Origi-nal work published 1739)

Izard, C. E. (1977). Human emotion. New York: Ple-num Press.

Izard, C. E. (1978). Emotions and emotion–cognitionrelationships. In M. Lewis & L. A. Rosenblum(Eds.), The genesis of behavior: Vol. 1. The devel-opment of affect (pp. 389–413). New York: Ple-num Press.

Kant, I. (1958). Critique of pure reason (N. KempSmith, Trans.). New York: Macmillan. (Originalwork published 1781)

Lazarus, R. S. (1982). Thoughts on the relations be-tween emotion and cognition. American Psycholo-gist, 37, 1019–1024.

Lazarus, R. S. (2001). Relational meaning and discreteemotions. In K. R. Scherer, A. Schorr, & T. Johnstone(Eds.), Appraisal processes in emotion: Theory, meth-ods, research (pp. 37–67). New York: Oxford Uni-versity Press.


Levine, S., & Wiener, S. G. (1989). Coping with uncer-tainty: A paradox. In D. S. Palermo (Ed.), Copingwith uncertainty: Behavioral and developmental per-spectives (pp. 1–16). Hillsdale, NJ: Erlbaum.

Lewis, M. (1989). Cultural differences in children’sknowledge of emotional scripts. In P. Harris & C.Saarni (Eds.), Children’s understanding of emotion(pp. 350–373). New York: Cambridge UniversityPress.

Lewis, M. (1991). Ways of knowing: Objective self-awareness or consciousness. Developmental Review,11, 231–243.

Lewis, M. (1992a). The self in self-conscious emotions.A commentary. In D. Stipek, S. Recchia, & S.McClintic (Eds.), Self-evaluation in young children.Monographs of the Society for Research in Child De-velopment, 57(1, Serial No. 226), 85–95.

Lewis, M. (1992b). Shame: The exposed self. NewYork: Free Press.

Lewis, M. (1993). The development of deception. In M.Lewis & C. Saarni (Eds.), Lying and deception in ev-eryday life (pp. 90–105). New York: Guilford Press.

Lewis, M. (2003a). The development of self-consciousness. In J. Roessler & N. Eilan (Eds.),Agency and self-awareness: Issues in philosophy andpsychology (pp. 275–295). New York: Oxford Uni-versity Press.

Lewis, M. (2003b). The emergence of consciousness andits role in human development. Annals of the NewYork Academy of Sciences, 1001, 1–29.

Lewis, M. (2005a, October 17). Early individual differ-ences in coping with stress: Emotional behavior, HPAand ANS reactivity. Paper presented at the New YorkAcademy of Sciences, Psychology Section, New York.

Lewis, M. (2005b). Origins of the self-conscious child.In W. R. Crozier & L. E. Alden (Eds.), The essentialhandbook of social anxiety for clinicians (pp. 81–98). Chichester, UK: Wiley.

Lewis, M., & Brooks-Gunn, J. (1978). Self-knowledgeand emotional development. In M. Lewis & L.Rosenblum (Eds.), The genesis of behavior: Vol. 1.The development of affect (pp. 205–226). New York:Plenum Press.

Lewis, M., & Brooks-Gunn, J. (1979). Social cognitionand the acquisition of self. New York: Plenum Press.

Lewis, M., & Goldberg, S. (1969). Perceptual–cognitivedevelopment in infancy: A generalized expectancymodel as a function of the mother–infant interaction.Merrill–Palmer Quarterly, 15, 81–100.

Lewis, M., Hitchcock, D. F. A., & Sullivan, M. W.(2004). Physiological and emotional reactivity tolearning and frustration. Infancy, 6(1), 121–143.

Lewis, M., & Michalson, L. (1983). Children’s emo-tions and moods: Developmental theory and mea-surement. New York: Plenum Press.

Lewis, M., & Ramsay, D. S. (2005). Infant emotionaland cortisol responses to goal blockage. Child Devel-opment, 76(2), 518–530.

Lewis, M., Ramsay, D. S., & Kawakami, K. (1993).Affectivity and cortisol response differences betweenJapanese and American infants. Child Development,64, 1722–1731.

Lewis, M., Ramsay, D. S., & Sullivan, M. W. (2006).The relation of ANS and HPA activation to infant an-ger and sadness response to goal blockage. Develop-mental Psychobiology, 48, 397–405.

Lewis, M., & Rosenblum, L. (Eds.). (1974). The originsof behavior: Vol. 2. The origins of fear. New York:Wiley.

Lewis, M., & Saarni, C. (Eds.). (1985). The socializa-tion of emotion. New York: Plenum Press.

Lewis, M., & Saarni, C. (Eds.). (1993). Lying and de-ception in everyday life. New York: Guilford Press.

Lewis, M., Sullivan, M. W., & Michalson, L. (1984).The cognitive–emotional fugue. In C. E. Izard, J.Kagan, & R. Zajonc (Eds.), Emotions, cognition,and behavior (pp. 264–288). New York: CambridgeUniversity Press.


Lewis, M., Sullivan, M. W., Stanger, C., & Weiss, M.(1989). Self-development and self-conscious emo-tions. Child Development, 60, 146–156.


Mandler, G. (1980). The generation of emotion: A psy-chological theory. In R. Plutchik & H. Kellerman(Eds.), Emotion: Theory, research, and experience(Vol. 1, pp. 219–244). New York: Academic Press.

Nelson, C. A., & Bloom, F. E. (1997). Child develop-ment and neuroscience. Child Development, 68,970–987.

Nelson, C. A., & Bosquet, H. (2000). Neurobiology offetal and infant development: Implications for infantmental health. In C. H. Zeanah (Ed.), Handbook ofinfant mental health (2nd ed., pp. 37–59). New York:Guilford Press.

Ortony, A., Clore, G. L., & Collins, A. (1988). The cog-nitive structure of emotions. New York: CambridgeUniversity Press.

Oster, H. (2005). The repertoire of infant facial expres-sions: An ontogenetic perspective. In J. Nadel & D.Muir (Eds.), Emotional development: Recent re-search advances (pp. 261–292). New York: OxfordUniversity Press.

Plutchik, R. (1980). Emotion: A psychoevolutionarysynthesis. New York: Harper & Row.

Pollak, S. D., Cicchetti, D., Hornung, K., & Reed, A.(2000). Recognizing emotion in faces: Developmen-tal effects of child abuse and neglect. DevelopmentalPsychology, 36(5), 679–688.

Saarni, C. (1999). The development of emotional com-petence. New York: Guilford Press.

Sander, L. W. (1977). Infant and caretaking environ-ment: Investigation and conceptualization of adap-tive behavior in a system of increasing complexity. InE. J. Anthony (Ed.), The child psychiatrist as investi-gator (pp. 170–183). New York: Plenum Press.


Schaffer, H. R. (1974). Cognitive components of theinfant’s response to strangeness. In M. Lewis & L.A. Rosenblum (Eds.), The origins of behavior: Vol.

2. The origins of fear (pp. 11–24). New York:Wiley.

Schaffer, H. R., Greenwood, A., & Parry, M. H. (1972).The onset of wariness. Child Development, 43, 165–175.

Schneirla, T. C. (1959). An evolutionary and develop-mental theory of biphasic processes underlying ap-proach and withdrawal. In M. R. Jones (Ed.), Ne-braska Symposium on Motivation (Vol. 7, pp. 1–42).Lincoln: University of Nebraska Press.

Schwartz, G. E., & Weinberger, D. A. (1980). Patternsof emotional responses to affective situations: Rela-tions among happiness, sadness, anger, fear, depres-sion, and anxiety. Motivation and Emotion, 4(2),175–191.

Spitz, R. A. (1965). The first year of life. New York: In-ternational Universities Press.

Sroufe, L. A. (1996). Emotional development: The orga-nization of emotional life in the early years. NewYork: Cambridge University Press.

Stenberg, C. R., Campos, J. J., & Emde, R. N. (1983).The facial expression of anger in seven-month-old in-fants. Child Development, 54, 178–184.

Stipek, D., Recchia, S., & McClintic, S. (Eds.). (1992).Self-evaluation in young children. Monographs of theSociety for Research in Child Development, 57(1, Se-rial No. 226).

Sullivan, M. W., & Lewis, M. (2003). Contextual deter-minants of anger and other negative expressions inyoung infants. Developmental Psychology, 39(4),693–705.

Suomi, S. (1991). Primate separation models of affectivedisorders. In J. Madden (Ed.), Neurobiology of learn-ing, emotion and affect (pp. 195–213). New York:Raven Press.

Tomkins, S. D. (1962). Affect, imagery, consciousness:Vol. 1. The positive affects. New York: Springer.

Tomkins, S. D. (1963). Affect, imagery, consciousness:Vol. 2. The negative affects. New York: Springer.

Watson, J. B. (1919). Psychology from the standpoint ofa behaviorist. Philadelphia: Lippincott.

Zajonc, R. B. (1980). Feeling and thinking: Preferencesneed no inferences. American Psychologist, 35, 151–175.


C H A P T E R 1 9

Children’s Understandingof Emotion

PAUL L. HARRIS

AWARENESS OF EMOTION

In this chapter, I discuss children’s developingawareness and understanding of emotions—both the emotions that they themselves feel andthe emotions expressed by other people. I firstdescribe children’s changing ability to put feel-ings into words. Next, I consider how children’sunderstanding of emotion changes with devel-opment. First, I emphasize that children cannotrely on a script-based conceptualization, butmust attend to the relation between appraisalprocesses and ensuing emotion. More specifi-cally, to understand how individuals may havedifferent emotional reactions to the same situa-tion, children need to understand the role ofappraisal in the elicitation of emotion. Second,by way of a detailed illustration of this claim, Iconsider children’s emerging understanding of arelatively complex but central emotion—namely, guilt. On the one hand, there are signsthat children express guilt even in the preschoolperiod. On the other hand, it is only in middlechildhood that children systematically recog-nize when a person might be prone to guilty feel-

ings. I discuss possible reasons for this lag be-tween the expression and the attribution ofguilt, eventually highlighting the importance ofunderstanding appraisal in the attribution ofguilt. Finally, I turn to a discussion of individualdifferences in children’s understanding of emo-tion. In that connection, I review the increas-ingly solid evidence that children who are givenopportunities to engage in family conversationabout emotion end up with a more accurate andcomprehensive understanding.

TALKING ABOUT EMOTION

Psychological theories of emotion, whether fo-cused on children or adults, have been influ-enced by Darwin’s (1872/1998) emphasis onthe continuities between human beings andnonhuman primates with respect to both thefunction and the communication of emotions.However, human beings, unlike other primates,can put their feelings into words. It could be ar-gued that this capacity only serves to amplify apreexisting mode of nonverbal communica-

320

tion. However, it is more likely that it producesa psychological revolution. After all, it allowshuman beings to communicate what they feelnot just about ongoing situations, but alsoabout past, future, recurrent, or hypotheticalsituations. These conversations—which beginin early childhood—provide our species with aunique opportunity to share, understand, andreconstitute emotional experience.

To document young children’s emergingability to talk about emotion, we (Wellman,Harris, Banerjee, & Sinclair, 1995) studied asmall group of children whose language pro-duction had been recorded on an intensive lon-gitudinal basis from 2 to 5 years of age. Weconcentrated on all those utterances in whichchildren referred either to an emotion or, forcomparison purposes, to another subjectivestate that is not an emotion (namely, pain). Thefindings revealed that even 2-year-olds talk sys-tematically about emotion. They refer to asmall set of emotional states—both positivestates (feeling happy or good; laughing; andfeeling love or loving) and negative states (feel-ing angry or mad; feeling frightened, scared, orafraid; and feeling sad or crying). Althoughchildren talk most often about their own feel-ings, they also talk about the feelings of otherpeople. Moreover, children’s attributions ofemotion are not triggered simply by the recog-nition of animate, expressive displays, becausethey readily attribute various emotions to dolls,stuffed animals, and made-up characters. Insum, almost as soon as they are able to talk,children begin to report on their own feelingsand on those of other people, and they projectsuch feelings onto nonhumans.

Despite this emerging communicative capac-ity, it could be argued that when children startto put their own feelings into words, they arenot engaged in any self-conscious reporting oftheir experience. Thus Wittgenstein (1953) sug-gested that early emotion utterances should beseen not as reports of emotion, but as vocal ex-pressions of emotion, on a par with exclama-tions such as “Ouch!” or “Ow!” A close exam-ination of 2-year-olds’ utterances shows thatWittgenstein’s proposal does not capture thefull complexity of young children’s talk aboutemotion. If their references to emotion weresimply supplements to, or substitutes for, theordinary facial and behavioral indices of emo-tion, we would expect those utterances to betriggered more or less exclusively by ongoingor current emotions. However, about half of 2-

year-olds’ references to emotion are concernedwith past, future, and recurrent feelings, andthe distribution of references is similar among3- and 4-year-olds.

This stable pattern shows that, from theirearliest emergence, we can think of children’sutterances about emotion as referential reportsand not as lexical substitutes for scowls andsmiles. Indeed, Wittgenstein’s analysis is noteven appropriate for children’s pain utterances.Here too, children talk not only about currentpains; they also refer to pains that they mightexperience in the future or have experienced inthe past. More generally, analysis of children’sreferences to emotion shows that these refer-ences can be mainly categorized as descriptivestatements, even if they are sometimes used inan instrumental fashion—to obtain sympathy,or to influence the emotional state of anotherperson (Dunn, Brown, & Beardsall, 1991;Wellman et al., 1995). Indeed, this bias towardcommentary is evident even below 2 yearsof age. Dunn, Bretherton, and Munn (1987)found that children between 18 and 24 monthsused conversation about feelings primarily tocomment on their own feelings or those of an-other person, even though their mothers—towhom most of these comments were directed—used such conversations in a more didactic orpragmatic fashion.

In a follow-up analysis of children’s everydayconversations about emotion, Lagattuta andWellman (2002) looked more closely at theway that children talk about negative as com-pared to positive emotions. As in the Wellmanet al. (1995) study, they looked at the utter-ances produced by a small group of childrenwhose language production had been studiedon an intensive longitudinal basis from 2 to 5years of age. Overall, children and their parentsdiscussed positive and negative emotions atabout the same rate. Nevertheless, when pastemotions were discussed, there was a tendencyto focus on negative rather than positive emo-tions. This bias toward the negative was truefor children and adults alike. Talk about nega-tive emotions also included about three timesas many causal elaborations as talk about posi-tive emotions did, and again this bias emergedamong both children and adults. Moreover,when children and adults posed open-ended (asopposed to a closed-ended) questions aboutemotions, such questions were about threetimes more frequent for negative than for posi-tive emotions. Overall, then, Lagattuta and

19. Children’s Understanding of Emotion 321

Wellman (2002) found that conversationsabout past emotional experiences are especiallyfrequent, elaborate, and unconstrained in thecase of negative emotions.

It seems plausible that conversations aboutpast emotions may help children understandhow an emotion can be reactivated long afterthe precipitating situation is over. This facilita-tion may occur in two ways. First, childrenmay be drawn into conversation about pastevents and thereby experience a reactivation ofthe emotion that they felt earlier. To the extentthat elaborate conversations about the past fo-cus on negative events, children may be espe-cially alert to the reactivation of negative emo-tion as compared to positive emotion. Second,when ruminating about a past event, childrenmay display the emotion associated with thatevent and thereby puzzle their parents with ademeanor that is not consonant with presentcircumstances. Parental questioning mayprompt children to think about their emotionalstate and its cause. Such questions (“What’s thematter?”, “What’s bothering you?”) are espe-cially likely to occur when a child appears up-set in otherwise positive circumstances. By con-trast, a child who remains cheerful in the faceof negative circumstances is less likely to pro-voke parental questioning. Thus if either ofthese speculations is correct, children may beespecially aware of the way in which rumina-tion about the past can evoke emotion in thecase of negative as opposed to positive feelings.

Lagattuta and Wellman (2001) examinedthis issue with children ranging from 3 to 5years of age. Children listened to stories inwhich the protagonist experienced an emotionthat was either positive or negative and eitherdid or did not match the current situation.Consider, for example, the following storyabout a negative emotion that did not matchthe current situation: “Suzie feels sad when theneighbor’s black spotted dog scares away herrabbit. Many days later, the neighbor’s dogslowly walks over, sits down, and wags his tail‘real friendly.’ Suzie starts to feel sad. . . . Whydoes Suzie start to feel sad right now?”Children were scored for the frequency withwhich they produced so-called “cognitivecueing” explanations, involving references to acue in the present situation that made the pro-tagonist think about a past event—for exam-ple, “The dog makes her think about the lostrabbit.” Lagattuta and Wellman (2001) foundthat children were especially likely to produce

such cognitive cueing explanations for storiesinvolving a negative mismatch, as exemplifiedby the story about Suzie’s dog. For stories in-volving a positive emotion, such cognitivecueing explanations were less frequent, al-though among older children they becamemore frequent. By implication, conversationsabout negative emotions offer children an ini-tial foothold toward the insight that one’s cur-rent emotion is not tied to present circum-stances, but is markedly influenced by thoughtsand reminders of noncurrent situations.

Overall, the findings on children’s conver-sations about emotion highlight the extentto which children’s emotional experience andemotional reflection are not tied to the currentmoment. Language does, of course, allow chil-dren to put their ongoing feelings into words.In addition, however, it allows them to talkabout future emotions and to revisit past emo-tions. It is probably by virtue of such conversa-tions that children become aware of the factthat past emotion can be reactivated by re-minders.

BEYOND SCRIPTS: DESIRES, BELIEFS,AND EMOTION

In the preceding section, I have talked in globalterms about children’s ability to report on andto understand emotion. In this section I con-sider in more detail the nature of this under-standing and the way that it changes in thecourse of development. One simple and attrac-tive proposal is that children develop an in-creasingly elaborate set of scripts for variousemotions. Thus they identify and remember thetype of situations that elicit particularemotions—fear, sadness, happiness, guilt, andso forth (Barden, Zelco, Duncan, & Masters,1980; Harris, Olthof, Meerum Terwogt, &Hardman, 1987)—and they also identify andremember the typical actions and expressionsthat accompany a particular emotional state(Trabasso, Stein, & Johnson, 1981).

This notion of script-based knowledge hasseveral advantages. It assimilates children’s un-derstanding of emotion to a wider body of re-search on children’s recall and understandingof sequentially organized events (cf. Nelson &Gruendel, 1979). It highlights the fact that anunderstanding of emotion calls for a causal un-derstanding of the connections among its se-quential components. It is sufficiently flexible


to be of service if we look outside the Westernworld to children’s understanding of emotionin cultures where different emotional themesare prominent; for example, Lutz (1987) hasused this approach in her analysis of the emo-tion concepts of children on the island of Ifalukin the Western Pacific. Finally, the notion of anemotion script fits comfortably with a possibil-ity raised in the preceding section—that chil-dren’s understanding may be elaborated notjust in the context of emotionally charged en-counters, but in the context of family discus-sions in which past episodes are likely to be re-hearsed and organized into a coherentnarrative sequence.

However, closer scrutiny of the script con-cept reveals a conceptual difficulty. The samesituation can elicit different emotions in differ-ent individuals, depending on the appraisalthat a particular individual makes of the situa-tion. This means that if a child attempts tostore a list of scripts for emotion, it will be nec-essary to store different scripts for differentpeople. An alternative, and more economical,solution is to define the eliciting situation inmore abstract terms. For example, it is possibleto define situations that provoke happiness as“situations that are judged by an actor to bringabout the fulfillment of his or her goals.” Amove in this direction, however, tacitly ac-knowledges that emotions are very specialkinds of scripts. They do not begin with thekind of objective event that we normally asso-ciate with scripts (e.g., the action of sittingdown at a table might be seen as the first movein the dinner script). Rather, they begin with anevent that is inherently psychological (namely,a person appraising a situation). A more fruit-ful approach to children’s understanding ofemotion, therefore, is to acknowledge that chil-dren may indeed construct scripts for givenemotions—but that key elements of thosescripts will include a diagnosis not of the objec-tive situation that faces the actor, but rather ananalysis of how the actor appraises that situa-tion. To make the same point differently, it isnot just psychologists who have to recognizethe role of appraisal processes in emotion.Young children must do the same.

The limitations of the script-based approachcan be highlighted in another way. Childrenwith autism are often good at remembering re-current sequences of events. Indeed, part of theclinical picture of autism is a disposition to be-come upset at an unexpected departure from a

routine sequence. Their script-sensitive mem-ory appears to serve children with autism quitewell with respect to emotion. Thus they readilyjudge that certain situations (getting nice thingsto eat, birthday parties) make people happy,whereas other situations (having to go to bedearly, falling over) make people unhappy(Baron-Cohen, 1991; Tan & Harris, 1991).Using a different technique, Ozonoff, Penning-ton, and Rogers (1990) showed that autisticchildren could select the appropriate facialexpressions to go with various emotionallycharged pictures. For example, they chose a sadface for a picture of a child looking at a brokentoy and an angry face for a picture of two chil-dren fighting. Despite this apparent familiaritywith routine emotion scripts, children withautism perform poorly, compared with non-autistic controls, when a correct attribution ofemotion requires them to go beyond the objec-tive situation and to consider how a protago-nist’s beliefs influence his or her appraisal ofthat objective situation (Baron-Cohen, 1991;Harris, 1991). The clear implication is thatnonautistic children do go beyond a script-based analysis and take into account the pro-tagonist’s appraisal of the situation.

If we accept this argument, we can ask inmore detail how children make sense of theprocess of appraisal. First, 2- and 3-year-oldsappreciate the role that desires or goals play indetermining a protagonist’s appraisal and ensu-ing emotion. For example, they understandthat a toy elephant may feel happy to be givenmilk if she wants milk, whereas another animalmay feel upset if he prefers juice instead (Har-ris, Johnson, Hutton, Andrews, & Cooke,1989; Yuill, 1984). By 4 and 5 years of age, thissimple desire-based concept of emotion is elab-orated to include beliefs and expectations.Children realize that it is not the match be-tween desire and actual outcome that triggersemotion, but the match between desire and ex-pected outcome. Suppose, for example, that theelephant wants some milk and is about to getit, so that if the match between desire and ac-tual outcome is the only factor taken into con-sideration, she should feel happy. Suppose fur-ther, however, that the elephant wronglyexpects to get something other than milk. Inthat case, 4- and 5-year-olds realize that the ele-phant will feel upset rather than happy (Harriset al., 1989). They appreciate that her appraisalof the situation, and her ensuing emotion, arebased on the mismatch between her desire and


the expected outcome, even when the expecta-tion is ill founded.

The shift from a desire to a belief–desire con-ception of mind and emotion is now well estab-lished. Emotion judgment tasks such as the onejust described are useful sources of evidence,because whether children are asked to takeonly desires into account or beliefs and desires,they can still be asked to make the same simplebinary judgment—namely, whether the animalis happy rather than sad. Another importantsource of evidence is children’s spontaneoustalk about psychological states. In the preced-ing section, I have described the way that chil-dren report on emotional states (Wellman etal., 1995). Using a similar database, Bartschand Wellman (1995) have examined children’sreferences to other mental states. Their analysisreveals that children talk systematically aboutdesires and goals throughout most of the thirdyear, chiefly using the term “want.” Then,starting at about the third birthday, childrenalso begin to make reference to beliefs, mainlyusing the terms “know” or “think.” Even-tually, by about the fifth birthday, talk aboutbeliefs becomes as frequent as talk about de-sires.

This developmental pattern is probably uni-versal. Tardif and Wellman (2000) report thatchildren learning to speak Cantonese and Man-darin display a similar progression: Talk aboutgoals and desires emerges early; talk about be-liefs and expectations shows a later onset.These data help to rule out various possible in-terpretations of the lag between talk about de-sires and talk about knowledge and belief. Forexample, it might be argued that it arises be-cause in English the predicate complementstructure is simpler for the verb “want” thanfor the verbs “think” and “know.” However, inMandarin and Cantonese, the predicate com-plement structure is relatively simple across ref-erences to both desires and thoughts. Indeed, itis worth noting that in both Mandarin andCantonese, some polysemous mental verbs canbe used to indicate either desire or thought.Yet, despite the availability of the same lexicalitem for both meanings, the lag between refer-ences to desires and thoughts still emerges. Insum, whether we focus on children’s emotionjudgments by using experimental tasks involv-ing a simple binary judgment, or on children’sspontaneous references to mental processes inthe course of their everyday conversation (be itin English, in Mandarin, or in Cantonese), the

evidence is robust that children focus initiallyon an agent’s goals, but increasingly take intoaccount his or her thoughts and beliefs.

It is worth noting that in the context of emo-tion judgments, the development of children’sunderstanding of the role of thoughts and be-liefs is quite protracted. Children ages 4–5years are generally accurate in recognizingthe impact of thoughts and beliefs on action(Wellman, Cross, & Watson, 2001). However,it is only at about 5–6 years of age that childrenrecognize the impact of beliefs on emotion.Several studies support this conclusion. For ex-ample, Hadwin and Perner (1991) found thatvirtually all 5-year-olds could appreciate astory character’s mistaken belief, but only at 6years of age did a significant majority makecorrect attributions of surprise that were con-sistent with the story character’s mistaken be-lief. Bradmetz and Schneider (1999) replicatedthe same lag between belief and emotion attri-butions across a series of five experiments. Forexample, when given a version of the story ofLittle Red Riding Hood, children frequently re-alized that Little Red Riding Hood mistakenlythinks it is her grandmother in the bed, butthen went on to say that she was afraid—andinvoked the wolf to explain her fear (e.g., “Be-cause it is a wolf!” or “Because the wolf wantsto eat her”). On the other hand, no child madea correct emotion attribution but failed thefalse-belief test. Finally, we (de Rosnay & Har-ris, 2002) compared children’s performance ontwo versions of a nasty-surprise task (involvinga protagonist’s mistakenly expecting a posi-tive outcome that turns out to be negative).Children frequently erred by incorrectly identi-fying the protagonist’s emotion as negative,despite correctly identifying the protagonist’spositive belief. In summary, whereas youngpreschoolers (ages 3–4 years) grasp that an in-dividual will appraise a situation in terms of hisor her desires—and will react with differentemotions, depending on whether those desiresare frustrated or fulfilled—children’s appraisalof the impact of beliefs on emotion emergesmore slowly. Only at about 5–6 years of age dothey realize that the appraisal of frustration orfulfillment, rather than the objective situation,is what dictates emotional reactions.

Children’s developing appreciation of ap-praisal processes can be further highlighted bytaking a closer look at children’s attributions ofan important and complex social emotion—namely, guilt. In a pioneering study, Nunner-


Winkler and Sodian (1988) found a surprisingage change. Children ages 4–5 years consis-tently claimed that a story protagonist who hadcommitted a serious transgression (e.g., delib-erately lied, pushed another child, or stolensomething) would feel happy. The children jus-tified this by noting that the outcome of theprotagonist’s transgression had produced posi-tive results: He or she had successfully stolensomething or had managed to displace anotherchild on the swing. By about the age of 8 years,children were more likely to claim that the pro-tagonist would feel bad and to refer to the storycharacter’s transgression in explaining that at-tribution.

One plausible explanation of this age changeis that older children increasingly expect that atransgression will provoke a bad conscience;they have an understanding of what it means tofeel guilty. Still, before focusing in more detailon that interpretation, I believe it is worth con-sidering various alternatives. First, it is conceiv-able that younger children regard the transgres-sions as trivial. However, a long tradition ofresearch on moral development shows that pre-schoolers actually think of lying, hitting, andstealing as serious transgressions (Smetana,1981). Not surprisingly, therefore, Keller,Lourenço, Malti, and Saalbach (2003) couldfind no age change in children’s castigation ofsuch basic transgressions.

A second possibility is that older children ex-pect the protagonist to feel bad because theyare more alert to the risk of punishment. In-deed, because they are older and arguably ex-pected to “know better,” older children mayexpect more severe punishment for such trans-gressions than younger children. However,children’s justifications lend little support tothis explanation. They rarely refer to punish-ment or fear of punishment when explainingwhy the perpetrator feels bad.

A third possibility is that older children aremore likely to interpret the question in terms ofthe emotion that the perpetrator should feel,whereas the younger children focus on whatthe perpetrator does feel. However, whenKeller et al. (2003) asked children howthey themselves would feel after such atransgression—a question format that shouldpresumably have prompted the children to fo-cus on socially desirable feelings—the familiarage change reemerged.

A fourth possibility is that older childrenhave greater empathy: They may more readily

acknowledge the suffering experienced by thevictims of the transgressions, and they may at-tribute bad feelings to the person who hascaused that distress. However, when Arsenioand Kramer (1992) explicitly asked children ofvarious ages about the feelings of the victim, allage groups acknowledged his or her distress. Afurther problem for the empathy hypothesis isthat a similar age change in children’s emo-tion attributions emerges when they are askedabout transgressions that do not involve anysuffering by a victim. Lagattuta (2005) pre-sented children from 4 to 7 years of age withstories involving a conflict between the protag-onist’s desire and various nonmoral rules con-cerning, for example, safety (“Don’t run intothe street”) or nutrition (“Don’t eat cookies be-fore dinner”). Older children were more likelyto acknowledge that the protagonist could feelbad after breaking the rule. Equally important,they were also more likely to acknowledge thatthe protagonist could feel good about resistingthe temptation to break the rule. Clearly, in nei-ther of these cases is it feasible to explain theage change in terms of increased empathy for avictim. In these episodes, there was no thirdparty in the role of victim.

The most plausible explanation of the find-ings is that older children increasingly conceiveof agents as engaging in a particular kind ofappraisal process. Whereas younger childrenare inclined to think that wrongdoers focusmainly on their goals and feel happy or sad de-pending on whether those goals are fulfilled ornot, older children increasingly acknowledgean additional appraisal process in which thewrongdoer appraises his or her actions in termsof whether or not they conform to various rulesand obligations (Harris, 1989). In probing chil-dren’s explanations for the emotion felt by thestory characters, Lagattuta (2005) obtainedfirm support for this interpretation. Older chil-dren were more likely than younger children tofocus on rules and obligations. They said, forexample, “She feels a little bad because sheshouldn’t have done that,” or “ . . . because hismom said he had to stay out of the street.”)

If the analysis above is correct, it implies thatchildren come to attribute guilt only in middlechildhood. Yet recent evidence suggests thateven preschool children feel and express guilt.For example, Kochanska, Gross, Lin, andNichols (2002) report that preschoolers show afairly stable tendency toward displays of dis-comfort following a mishap; these displays cor-


relate moderately with maternal ratings of achild’s tendency to feel guilt; and children whodisplay more discomfort are more likely to con-form to adult-imposed rules. Taken together,these results suggest that even preschoolers dofeel guilt. If that is the case, why do childrenonly start to attribute guilt several years later?The most plausible explanation is that this lagis simply one more example of a very generaldevelopmental pattern. Children enter into avariety of mental states quite early in life: Theyentertain false beliefs, they experience surprise,and they feel badly about a misdemeanor. Theability to make sense of and attribute thosemental states—whether to the self or to an-other person—is far from being an automaticaccompaniment. It is only after constructing atheory or model of those mental states, includ-ing the appraisal processes that give rise tothem, that children can make appropriate attri-butions of belief, surprise, or guilt.

INDIVIDUAL DIFFERENCESIN UNDERSTANDING EMOTION

So far, I have focused on important age changesin children’s understanding of emotion. Thisemphasis reflects the research program that hasdominated research for the past 20 years ormore: Investigators have aimed to identify asuccession of conceptual insights that childrencome to master in the course of development.However, more recently, attention has increas-ingly turned to individual differences in chil-dren’s mastery of those insights. On the onehand, investigators have developed tools tomeasure variation in children’s understandingof emotion; on the other hand, they have begunto analyze the reasons why some children areadvanced in their understanding, whereas oth-ers are much slower.

We (Pons, Harris, & de Rosnay, 2004) havereported on a Test of Emotion Comprehension(TEC) composed of nine different components:(1) recognizing facial expressions of emotion;(2) understanding situational causes; (3) under-standing the effect of external reminders onemotion; (4) understanding the link betweendesire and emotion; (5) understanding the linkbetween belief and emotion; (6) understandingthe potential discrepancy between felt and ex-pressed emotion; (7) understanding guilt; (8)understanding the regulation of emotion; and

(9) understanding mixed or ambivalent emo-tions. We tested children between 3 and 11years of age for their mastery of each compo-nent. The main findings were that children dis-played a clear improvement with age on eachcomponent, and that the components them-selves can be plausibly grouped into three de-velopmental phases. The first period is charac-terized by the understanding of key publicaspects of emotion—their mode of expression,their situational causes, and the effect of exter-nal reminders. The second period is character-ized by mastery of the mentalistic nature ofemotion—the role of desires and beliefs, andthe distinction between felt and expressed emo-tion. The third period is characterized by anunderstanding of how the same individual canreflect on a situation from different points ofview or in terms of different criteria andthereby evoke different feelings—either at thesame time or successively. We (Pons et al.,2004) observed a hierarchical relationshipamong these three phases. By implication, un-derstanding key external aspects of emotionis a prerequisite for understanding the morementalistic aspects, which in turn is a prerequi-site for understanding the impact of reflectionand rumination on emotion.

In a follow-up study, we (Pons & Harris,2005) looked at longitudinal change and stabil-ity in children’s performance on the TEC.Children ages 7, 9, and 11 years at Time 1 wereretested 13 months later at Time 2. More thanhalf of the 7- and 9-year-olds showed gains atretest (although the majority of 11-year-oldsperformed about the same—probably reflect-ing the absence of challenging components forthis older group). Individual differences wereconsiderable at both Times 1 and 2 for all threeage group. Thus children in the youngest groupvaried by as much as six components, and chil-dren in the two older groups varied by as muchas four components. Moreover, when adjacentage groups were compared, the highest-scoringchildren in the younger group scored higherthan the lowest-scoring children in the oldergroup (by two to three components, dependingon which two age groups were compared). In-deed, some of the 7-year-olds had an overalllevel of emotion understanding that was higherthan some of the 11-year-olds’. These individ-ual differences remained quite stable over the13-month period. Thus, despite the gains overtime made by many children in their under-


standing of emotion, their level of understand-ing at Time 1 was a good predictor of their un-derstanding at Time 2.

Do these marked and stable individual dif-ferences in children’s understanding of emotionhave any implications for their behavior? Inparticular, we may ask whether children’s un-derstanding of emotion has an impact on theirsocial relationships when they move outsidethe family and start to form relationships withpeers. Several studies have explored this possi-bility. Denham, McKinley, Couchoud, andHolt (1990) tested preschoolers (mean age = 44months) for their emotion knowledge:Children had to identify a puppet’s emotion (ofhappiness, sadness, anger, or fear), both whenit exhibited a prototypical reaction (e.g., fearduring a nightmare) and when it showed anatypical reaction (e.g., sadness at going to pre-school). In addition, a sociometric measure wasused to assess children for their acceptance asplaymates among their peers. Children withhigher scores on the emotion test proved to bemore popular among their peers, even whenthe contributions of age and gender were re-moved. Cassidy, Parke, Butkovsky, andBraungart (1992) obtained very similar resultswith first-grade children. Children’s overallscore in an interview about the causes, conse-quences, and associated expression of emotionwas correlated with popularity. Finally, in alongitudinal study of 4- and 5-year-olds, Ed-wards, Manstead, and MacDonald (1984)found that children who were accurate at iden-tifying facial expressions of emotion proved tobe more popular 1–2 years later (even whentheir initial popularity was taken into account).The consistency among these three studies isstriking.

Nevertheless, caution is needed in interpret-ing the findings (Manstead, 1994). First, we donot yet understand the causal link between theunderstanding of emotion and peer relation-ships. Acceptance by peers may increase chil-dren’s opportunities for learning about emo-tion. Thus children’s understanding of emotionmay not promote their friendships; rather,causal arrow may move in the reverse direc-tion. Alternatively, a third variable (such as in-telligence or verbal ability) may underpin bothemotion understanding and popularity.

Second, it would be premature to concludethat children with more advanced insight intoemotion inevitably end up having healthy and

positive relationships with their peers. Con-sider the thought-provoking study of bullyingcarried out by Sutton, Smith, and Swettenham(1999). Bullies are sometimes characterized asawkward children who resort to aggression be-cause of their limited social skills. Yet it is alsoconceivable that bullying calls for an astuteanalysis of whom to victimize and how by-standers will react. With this in mind, Sutton etal. (1999) administered a set of stories designedto assess the understanding of emotions andcognitions among 7- to 10-year-olds. Childrenwho were “ringleader” bullies scored higherthan several other groups: “follower” bullies(i.e., those who helped or supported the ring-leaders), victims, and defenders of the victims.

The implication of all these studies is that in-sight into other children’s emotional and men-tal states is associated with social adroitness.Thus it is linked to popularity and acceptance,but also to leadership in the context of bully-ing. Although it is tempting to assume that amore advanced or precocious understanding ofemotion invariably yields positive social out-comes, such an assumption is probably too op-timistic.

To turn now to the origins of such individualdifferences, recent evidence increasingly pointsto a key role for family conversation. Considera child with a parent who frequently discussesemotion—by drawing out the child’s own feel-ings, by calling attention to the way that his orher actions may have emotional implicationsfor other members of the family, or by elabo-rating on the feelings of story characters. Con-sider, on the other hand, a parent who is moreconstrained in talking about emotion, whetherwith respect to the child or to other people.These two different conversation partnersmight be expected to have a differential impacton the extent to which the child understandshow an emotion comes about, or is prepared totalk about emotion, or both.

Certainly there is marked variation amongfamilies in the frequency with which emotionsare discussed. Dunn, Brown, and Beardsall(1991) found that some children never madeany mention of emotion during an hour-longhome visit, whereas others made more than 25such references; variation among the motherswas equally great. Accumulating evidence indi-cates that the frequency with which preschoolchildren engage in family discussion aboutemotions and their causes is correlated with


their later ability to identify how someonefeels. The link has been found over a relativelyshort period straddling the third birthday—that is, from 33 to 40 months (Dunn, Brown,Slomkowski, Tesla & Youngblade, 1991)—aswell as over a more extended period from 3 to6 years (Dunn et al., 1991; Brown & Dunn,1996).

Such correlational data are, of course, opento various interpretations. One possibility isthat the correlation reflects some stable attrib-ute of the child that manifests itself both in psy-chological talk and in sensitivity to emotion.For example, some children may be naturallyempathic: They may seek out and elicit moreconversations about emotion, and also displaya keen ability to assess how other people feel,as measured by standard tests of emotion un-derstanding. However, it is also plausible tosuppose that the correlation reflects the didac-tic role that conversation can play for children.Frequent family discussion, particularly whenparents are involved, may prompt children totalk about emotion and to increase their under-standing and perspective taking. One piece ofevidence that fits this second proposal has beenreported by Garner, Jones, Gaddy, and Reddie(1997): They found that children’s perspectivetaking is correlated with family discussion ofemotion that focuses not simply on what a per-son feels, but rather on why someone feels agiven emotion. We (de Rosnay, Pons, Harris, &Morrell, 2004) obtained a second piece of evi-dence consistent with the didactic role of fam-ily members: Mothers’ use of mentalistic termswhen describing their children (i.e., referencesto their children’s psychological characteristics,rather than their physical characteristics ortheir behavior) were positively correlated withchildren’s correct emotion attributions to storyprotagonists. Thus, even when a mother wasnot engaged in conversation with her child,characteristics of her discourse style neverthe-less predicted the child’s understanding of emo-tion.

Such an emphasis on the didactic role of pa-rental conversation (especially maternal con-versation) is consistent with a larger body ofresearch that has investigated children’s devel-oping understanding of various mental states,including beliefs as well as emotions. Three im-portant conclusions have emerged from thesestudies. First, when longitudinal data are col-lected, they confirm that the mothers’ discourseis what predicts children’s later understanding

of mental states; there is no indication that chil-dren’s understanding of mental states predictslater patterns of discourse by their mothers.Second, it is the mothers’ focus on mentalstates that appears to be critical, rather thanany generalized disposition to engage childrenin conversation. Third, this influence appearsto have a sustained impact; it is evident among3- and 6-year-olds alike (Harris, de Rosnay, &Pons, 2005; Ruffman, Slade, & Crowe, 2002).

In the coming years, we may expect to seemore research on the questions of how childrencome to vary in their understanding of emo-tion, and what part family talk may play inpromoting that variation. In that regard, wecan anticipate an increasing confluence of find-ings from research on early attachment, onchildren’s conversations about emotionallycharged events (especially past negativeevents), and on individual differences in chil-dren’s understanding of mental states (includ-ing emotion). For the most part, these topicshave been studied independently of one an-other, but that is likely to change. Such a con-fluence is likely to yield practical as well as the-oretical benefits. If we know more about howchildren’s understanding of emotion can be fa-cilitated in the context of the family, especiallyfamily conversation, we may be able to repro-duce some of those beneficial effects throughdeliberate and systematic therapeutic interven-tion.

CONCLUSIONS

In this chapter, I have examined several interre-lated aspects of children’s understanding ofemotion: their ability to report emotion inwords and to understand the way that pastemotions can be reactivated; their sensitivity tokey components of the appraisal processes thatmodulate a person’s emotional response to agiven situation (namely, the person’s desires,the person’s beliefs, and the person’s evaluationof his or her own standing in relation to vari-ous rules and obligations); and individual dif-ferences in emotion understanding, their linkswith peer relationships, and the key role offamily conversation in promoting children’sunderstanding of emotion.

At certain points, I have touched on a largertheme that deserves more attention in future re-search. Arguably, children’s developing un-derstanding of emotion is simply an epi-


phenomenon of the underlying process. Thisunderstanding may operate at a “meta” level,sealed off from the underlying emotional pro-cess that is its subject matter. To take a concreteexample, it is possible to assert that a childfunctions at two separate levels: On one level,there is the child’s experience and display ofguilt; at a separate level, there is the child’s ca-pacity for reporting on, attributing, and rumi-nating about the experience of guilt. Increasedsophistication at the “meta” level may havefew or no effects on processing at the primary,lower level.

Such a clear-cut separation between levels ofprocessing may simplify our analysis of devel-opment, but it probably distorts or ignoressome important features of human emotion. Iteffectively predicts that a disruption or delay inthe development of an understanding of emo-tion need have no repercussions for the basicemotional processes themselves. However,there are several reasons for thinking that suchrepercussions do exist. First, there is a thera-peutic tradition suggesting that intense emo-tional experiences that are reworked in thecontext of communication and ruminationhave different sequelae from those that are not.Such reworking need not be in the context ofdiscussion with a trained therapist; it can alsooccur in the context of a privately written nar-rative (Pennebaker, 1996). One plausible ex-trapolation of these findings is that the emo-tional lives of children who grow up in homeswhere there is open discussion of emotionallycharged encounters will be different from thosewhere such discussion does not occur. They arelikely to be prompted to engage in the type ofinsightful thinking about the causes of theiremotions that has been shown to be benefi-cial for adults’ physical and mental health(Pennebaker, Barger, & Tiebout, 1989).

Second, this capacity for communicationand rumination dramatically alters the con-texts in which children can seek support andreassurance. Attachment theorists have empha-sized the ways that a caretaker may or may notprovide reassurance at moments of distress.Typically, they have focused on those momentswhen the precipitating factor is fairly easy forthe caretaker to discern: The child is unnervedby a stranger, or distressed by the caregiver’s re-cent absence, or fretful about the caregiver’simminent departure. However, the emotionalhorizon of an older child is much larger; he orshe can be distressed or fearful about events

that might happen in the future or happened inthe more distant past. In such contexts, chil-dren who can articulate their anxieties and dis-cuss their causes are clearly better placed to re-ceive reassurance.

Finally, children’s ability to understand andpredict their own emotions likely effects theirdecision making about what course of action totake. In its turn, this chosen course of actionwill lead to—or prevent—certain emotionalconsequences. For example, the ability to antic-ipate guilt can serve as a brake or warning sig-nal when a guilt-inducing transgression is con-templated (Lake, Lane, & Harris, 1995). Thiswarning signal is sufficient to help children toinhibit the transgression, and to avoid any sub-sequent guilt. Stated in more general terms,children’s insight into their emotional lives doesnot simply enable them to foresee the inevita-ble; it allows them to look into the future andto make choices about what their emotionallife should be like. In that respect, children’sunderstanding of emotion enables them to altertheir experience of emotion.

REFERENCES

Arsenio, W. F., & Kramer, R. (1992). Victimizers andtheir victims: Children’s conceptions of the mixedemotional consequences of moral transgressions.Child Development, 63, 915–927.

Barden, R. C., Zelko, F. A., Duncan, S. W., & Masters,J. C. (1980). Children’s consensual knowledge aboutthe experiential determinants of emotion. Journal ofPersonality and Social Psychology, 39, 968–976.

Baron-Cohen, S. (1991). Do people with autism under-stand what causes emotion? Child Development, 62,385–395.

Bartsch, K., & Wellman, H. M. (1995). Children talkabout the mind. New York: Oxford University Press.

Bradmetz, J., & Schneider, R. (1999). Is Little RedRiding Hood afraid of her grandmother?: Cognitivevs. emotional response to a false belief. British Jour-nal of Developmental Psychology, 7, 501–514.

Brown, J. R., & Dunn, J. (1996). Continuities in emo-tion understanding from three to six years. Child De-velopment, 67, 789–802.

Cassidy, J., Parke, R. D., Butkovsky, L., & Braungart, J.M. (1992). Family–peer connections: The roles ofemotional expressiveness within the family and chil-dren’s understanding of emotions. Child Develop-ment, 63, 603–618.

Darwin, C. (1998). The expression of the emotions inman and animals (3rd ed.). London: HarperCollins.(Original work published 1872)

Denham, S. A., McKinley, M., Couchoud, E. A., &Holt, R. (1990). Emotional and behavioral predic-


tors of preschool peer ratings. Child Development,61, 1145–1152.

de Rosnay, M., & Harris, P. L. (2002). Individual differ-ences in children’s understanding of emotion: Theroles of attachment and language. Attachment andHuman Development, 4, 39–54.

de Rosnay, M., Pons, F., Harris, P. L., & Morrell, J. M.B. (2004). A lag between understanding false beliefand emotion attribution in young children: Relation-ships with linguistic ability and mothers’ mental statelanguage. British Journal of Developmental Psychol-ogy, 22, 197–218.

Dunn, J., Bretherton, I., & Munn, P. (1987). Conversa-tions about feeling states between mothers and theiryoung children. Developmental Psychology, 23, 132–139.

Dunn, J., Brown, J., & Beardsall, L. (1991). Family talkabout feeling states and children’s later understand-ing of others’ emotions. Developmental Psychology,27, 448–455.

Dunn, J., Brown, J., Slomkowski, C., Tesla, C., &Youngblade, L. (1991). Young children’s understand-ing of other people’s feelings and beliefs: Individualdifferences and their antecedents. Child Develop-ment, 62, 1352–1366.

Edwards, R., Manstead, A. S., & MacDonald, C.(1984). The relationship between children’s socio-metric status and ability to recognize facial expres-sions. European Journal of Social Psychology, 14,235–238.

Garner, P. W., Jones, D. C., Gaddy, G., & Rennie, K. M.(1997). Low-income mothers’ conversations aboutemotions and their children’s emotional competence.Social Development, 6, 37–52.

Hadwin, J., & Perner, J. (1991). Pleased and surprised:Children’s cognitive theory of emotion. British Jour-nal of Developmental Psychology, 9, 215–234.

Harris, P. L. (1989). Children and emotion: The devel-opment of psychological understanding. Oxford:Blackwell.

Harris, P. L. (1991). The work of the imagination. In A.Whiten (Ed.), Natural theories of mind (pp. 283–304). Oxford: Blackwell.

Harris, P. L., de Rosnay, M., & Pons, F. (2005). Lan-guage and children’s understanding of mental states.Current Directions in Psychological Science, 14, 69–73.

Harris, P. L., Johnson, C. N., Hutton, D., Andrews, G.,& Cooke, T. (1989). Young children’s theory of mindand emotion. Cognition and Emotion, 3, 379–400.

Harris, P. L., Olthof, T., Meerum Terwogt, M., &Hardman, C. E. (1987). Children’s knowledge of thesituations that provoke emotion. International Jour-nal of Behavioral Development, 10, 319–343.

Keller, M., Lourenço, O., Malti, T., & Saalbach, H.(2003). The multifaceted phenomenon of ‘happy vic-timizers’: A cross-cultural comparison of moral emo-tions. British Journal of Developmental Psychology,21, 1–18.

Kochanska, G., Gross, J. N., Lin, M.-H., & Nichols, K.E. (2002). Guilt in young children: Development, de-terminants, and relations with a broader system ofstandards. Child Development, 73, 461–482.

Lagattuta, K. H. (2005). When you shouldn’t do whatyou want to do: Young children’s understanding ofdesires, rules and emotions. Child Development, 76,713–733.

Lagattuta, K. H., & Wellman, H. M. (2001). Thinkingabout the past: Early knowledge about links betweenprior experience, thinking, and emotion. Child De-velopment, 72, 82–102.

Lagattuta, K. H., & Wellman, H. M. (2002). Differ-ences in early parent–child conversations about nega-tive versus positive emotions: Implications for thedevelopment of psychological understanding. Devel-opmental Psychology, 38, 564–580

Lake, N., Lane, S., & Harris, P. L. (1995). The expecta-tion of guilt and resistance to temptation. Early De-velopment and Parenting, 4, 63–73.

Lutz, C. (1987). Goals, events, and understanding inIfaluk emotion theory. In D. Holland & N. Quinn(Eds.), Cultural models in language and thought(pp. 290–312). Cambridge, UK: Cambridge Univer-sity Press.

Manstead, A. S. R. (1994). Children’s understanding ofemotion. In J. J. Russell, J.-M. Fernández-Dols, A. S.R. Manstead, & J. C. Wellenkamp (Eds.), Everydayconceptions of emotion (pp. 315–331). Dordrecht,The Netherlands: Kluwer.

Nelson, K., & Gruendel, J. (1979). At morning it’slunchtime: A scriptal view of children’s dialogues.Discourse Processes, 2, 73–94.

Nunner-Winkler, G., & Sodian, B. (1988). Children’sunderstanding of moral emotions. Child Develop-ment, 59, 1323–1338.

Ozonoff, S., Pennington, B. F., & Rogers, S. J. (1990).Are there emotion perception deficits in young autis-tic children? Journal of Child Psychology and Psychi-atry, 31, 343–361.

Pennebaker, J. W. (1996). Cognitive, emotional, andlanguage processes in disclosure. Cognition andEmotion, 10, 601–626.

Pennebaker, J. W., Barger, S. D., & Tiebout, J. (1989).Disclosures of traumas and health among Holocaustsurvivors. Psychosomatic Medicine, 51, 577–589.

Pons, F., & Harris, P. L. (2005). Longitudinal changeand longitudinal stability of individual differences inchildren’s emotion understanding. Cognition andEmotion, 19, 1158–1174.

Pons, F., Harris, P. L., & de Rosnay, M. (2004). Emotioncomprehension between 3 and 11 years: Develop-mental periods and hierarchical organization. Euro-pean Journal of Developmental Psychology, 1, 127–152.

Ruffman, T., Slade, L., & Crowe, E. (2002). The rela-tion between children’s and mother’s mental statelanguage and theory-of-mind understanding. ChildDevelopment, 74, 734–751.


Smetana, J. G. (1981). Preschool children’s conceptionof moral and social rules. Child Development, 52,1333–1336.

Sutton, J., Smith, P. K., & Swettenham, J. (1999). Socialcognition and bullying: Social inadequacy or skilledmanipulation? British Journal of Developmental Psy-chology, 17, 435–450.

Tan, J., & Harris, P. L. (1991). Autistic children under-stand seeing and wanting. Development and Psycho-pathology, 3, 163–174.

Tardif, T., & Wellman, H. M. (2000). Acquisition ofmental state language in Mandarin- and Cantonese-speaking children. Developmental Psychology, 36,25–43.

Trabasso, T., Stein, N. L., & Johnson, L. R. (1981).Children’s knowledge of events: A causal analysis of

story structure. In G. Bower (Ed.), Learning andmotivation (Vol. 15, pp. 237–282). New York: Aca-demic Press.

Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory-of-mind development: The truthabout false belief. Child Development, 72, 655–684.

Wellman, H. M., Harris, P. L., Banerjee, M., & Sinclair,A. (1995). Early understanding of emotion: Evidencefrom natural language. Cognition and Emotion, 9,117–149.

Wittgenstein, L. (1953). Philosophical investigations.Oxford: Blackwell.

Yuill, N. (1984). Young children’s coordination of mo-tive and outcome in judgments of satisfaction andmorality. British Journal of Developmental Psychol-ogy, 2, 73–81.


C H A P T E R 2 0

The Interfaceof Emotional Development

with Social Context

CAROLYN SAARNI

When the second edition of this handbook waspublished in 2000, I wrote a chapter with asimilar title, and at that time I thought I knewsomething about the reciprocal influence ofcontext and emotion (i.e., contexts provide the“stage” as well as the “audience” for emotions,and emotions influence both the “audience”and the selection of “stages” to play themselvesout upon). Although I did not use that particu-lar dramaturgical metaphor in the earlier chap-ter, it sums up what I thought captured the reci-procity between context and emotion. Now mythoughts about the relationship between con-text and emotion, especially from a develop-mental perspective (i.e., how this relationshipchanges over time), have taken a definite turn:Action—and thus motivation—play a largerrole in my construal of how emotions dependupon context and vice versa. I now assign alarger role to intentionality, to self-organizingprinciples, and to systems of interaction. Inshort, emotional development needs to be con-sidered from within a “bioecological” frame-work; this conceptual platform regards livingorganisms as dynamic systems reciprocally em-

bedded in a community. Clements and Shelford(1939), early plant and animal bioecologists,provided the theoretical underpinnings for theconcept of bioecology, and their emphasis wasvery clearly on the synthesis of habitat and or-ganism. Habitat for plant and animal commu-nities is analogous to context for human com-munities (ranging from families to societies),and the adjustments and adaptation made byplants and animals in tandem with their habi-tats correspond to the developmental relation-ships between genotypes and likely or possi-ble phenotypes in human development thatare manifested in particular social contexts(Bronfenbrenner & Ceci, 1994).

In this chapter, I review some of the tradi-tional and contemporary viewpoints on therelationship between context and emotionaldevelopment. I include some particularly inter-esting studies that illustrate the bioecologicalframework for understanding emotional devel-opment in its social context. I also discuss sev-eral features of the cultural context that consti-tute important aspects of the bioecologicalframework.

332

WHY IS CONTEXT SIGNIFICANTFOR EMOTIONAL DEVELOPMENT?

Some studies explicitly invoke the role of con-text as an illustration of the plasticity of emo-tional development. For example, the classiclongitudinal research by Murphy and Moriarty(1976) provided many rich descriptions of chil-dren who maintained a distinctive style of emo-tional regulation that at its core containedsome degree of continuity; yet across varyingcontexts and developmental stages, the chil-dren’s actual expression of their regulatorystyle was quite variable. I quote one such de-scription below:

Moreover, forms of expression of reactionchanged while the essential reaction continued.Sheila protested loudly, vociferously, and deci-sively as an infant; at three she was equally deci-sive in her unequivocal “no!”; by age ten shecould cooperate yet still convey her clear negativereaction by “making a face.” In other words, feel-ings and attitude patterns could persist in very dif-ferent forms. (Murphy & Moriarty, 1976, p. 197)

In this example (and many others like it pro-vided by the authors), I see plasticity of emo-tional development as reflecting a style ofemotional responding that has multiple mani-festations, which vary with development andeliciting situation. Murphy and Moriarty alsodescribed a different sort of plasticity of emo-tional development across children, wherebythe same stimulus elicited widely differingemotional-expressive behavior. Some childrenreacted to a stressful stimulus by becomingrigid, as though freezing; other children re-vealed more disorganized cognitive function-ing, manifested in slow, fuzzy speech or in-creased stammering. In this case, emotionalplasticity is reflected in the same elicitor’s beingresponded to by different emotional copingstyles. Relative to a bioecological approachtoward emotional development, I foundMurphy’s and Moriarty’s work also firmlyanchored in an appreciation of the infant’s“organicity”—that is, its temperamental dispo-sitions for how to respond to variable contexts.They too saw how the subsequent developmentof any given infant depended on the daily inter-actions, both social and object-oriented, thatthe infant experienced. Opportunities to accesssocial and inanimate environments were pos-sible insofar as these contexts provided“affordances” for interaction (Gibson, 1982),

and affordant characteristics of an environ-ment were often made feasible by sensitivecaregivers who could determine when a babywas overstimulated or understimulated.

Still another kind of emotional plasticity isdescribed in early work by Fraiberg (1971). Shefound that infants with very low vision wereminimally facially expressive in interactionswith their caregivers, but they were very activewith their fingers; indeed, they relied on theirfingers’ sensory capacities to explore theirworlds. Thus, for a more typical expressivepart of the body (the face), these infants substi-tuted other parts of their body that remainedmaneuverable by them—namely, their fingers.These visually impaired infants were experi-enced by their parents as somehow emotionallyunresponsive, which was a judgment or “rat-ing” of their facial expressiveness in contextsdeemed by “raters” (their parents) as somehowatypical. This idea that a judgment of emotionis made in the eyes of the beholder has been atthe center of much debate in the field of emo-tion, pitting universalists (e.g., Ekman, 1989)against cultural relativists (e.g., Russell, 1991).The former minimize the role of context as wellas the plasticity of emotion; the latter empha-size the significance of context, including cul-ture, in understanding emotion and the plas-ticity of emotion expression, appraisal, andmeaning (Fernández-Dols, 1999; White, 2000;Wierzbicka, 1999).

A relevant study by Camras et al. (2002) ex-amined directly the role of context and cultureon raters’ judgments of the emotion surprise ininfants close to 1 year of age. The data werecollected in Japan, in China, and in the UnitedStates with European American infants. Thenaive raters (in the United States) were to judgewhen the infants reacted with surprise when atoy they had seen on four previous trials afterlifting off a cloth covering it was covertly re-placed with a new toy. This expectancy-violating event was thought to elicit surprise,but the infants showed classic surprise expres-sions about as often in the four baseline ses-sions as they did in the “surprising-toy-switch”condition (about 30%). More often, the Japa-nese and European American infants showed abody stilling (a cessation of movement), ac-companied by a sober facial expression. TheChinese infants did not show this bodily stillingor sobering of facial expression, because forthese two behaviors to be observed, there mustbe more body activity and expressivity preced-

20. Interface of Emotional Development with Social Context 333

ing the expectancy-violating event for stillingor sobering to be judged as occurring. The Chi-nese infants appeared more placid or “self-contained,” or perhaps their behavioral cuesfor reacting to a discrepancy were too subtle tobe observed by Western judges. Were the Japa-nese and European American babies actuallysurprised by the toy switch, or were they justfocusing their attention on this discrepancy asthey stopped wiggling and displayed a somberexpression? They already possessed object per-manence, but perhaps surprise is an emotionthat develops later and may well depend oncontextual features such as proximity to care-givers (because otherwise an “impossibleevent” could be experienced as a source of dis-tress), temperamental proclivities toward reac-tivity, and the extent to which the individualhas some control over the elicitor of the sur-prising event (e.g., the classic Jack-in-the-Box,which pops out of the lid when the crank han-dle is turned and after the music plays to a cer-tain point). Maybe we need to watch 15- to 18-month-olds play with the Jack-in-the-Box and,with caregivers nearby, watch how the toddlersrespond to Jack reliably popping out fromunder the open lid as they themselves crank thehandle the requisite number of times. I predictthat by the second or third pop-up, we mightsee more classic surprise faces, but then withsubsequent repetition (again under the tod-dlers’ control), there would be a transforma-tion of the expressive response toward positiveanticipation and eventually habituation. Surelythere is a thesis somewhere in such a follow-upproject; for a related study with much youngerinfants with a Jack-in-the-Box manipulated byand coached by their mothers to associate sur-prise with the popping up of Jack’s head, see re-search undertaken by Reissland and Shepherd(2002).

To sum up, plasticity of emotional develop-ment relative to context can be seen in fourways: (1) in a particular style of emotion regu-lation manifesting itself in different forms inthe same individual at different developmentalperiods, whereby the eliciting situations sharesome common relational theme for the individ-ual; (2) in the same or very similar emotion-evoking context eliciting different expressivereactions, suggestive of the same hedonic tonebut with varying manifestations across differ-ent individuals (e.g., a “freezing” reaction to astressful stimulus vs. an agitated reaction); (3)in the “equipotentiality” of emotion-expressivechannels (e.g., facial expressions vs. vocal

channel or kinesthetic channel); and (4) injudges’ ratings of emotion in others that are in-formed by the context in which the others ap-pear to be responding to relational goals. Thesefour different ways of looking at plasticity inemotional development vis-à-vis shifting con-texts are compatible with functionalist (Cam-pos, Mumme, Kermoian, & Campos, 1994)and dynamic-systems (Camras, 1991, 1992)theoretical perspectives. Each of the four waysis linked to how the developing child respondsto the environment in an intentional, goal-directed way, thus rendering that environmentemotion-evocative, which feeds back into thechild’s repertoire of what he or she brings to thecontextual interaction. This notion of a feed-back system is not new; indeed, it is part of theself-organizing approach to emotional develop-ment (e.g., Fogel et al., 1992), and is also re-flected in Larsen and Prizmic-Larsen’s (2006)discussion of the necessity for multimethodmeasurement in emotion research. These au-thors all address emotion as a multiresponsesystem in which the various components ofemotion interact with one another (e.g., facialexpressions may amplify subjective experienceof emotion). Larsen and Prizmic-Larsen (2006)also raise further measurement issues ofwhether emotion is being investigated cate-gorically (e.g., discrete or basic emotions) ordimensionally (e.g., hedonic tone, degree ofarousal), and whether emotion is viewed froma state or trait perspective. Context furthercomplicates the measurement of emotion andundoubtedly contributes to the fluidity or loosecoupling of the varied response systems sub-sumed under the construct “emotion.”

CONTEXT AND ITS ATTRIBUTES

Many studies emphasize how emotional devel-opment is simultaneously embedded in bothverbal and nonverbal interpersonal exchanges,and in contexts that children respond to inrelationship to their goals (for reviews, seeDenham, 1998; Saarni, Campos, Camras, &Witherington, 2006). We can look at emotionaldevelopment as a series of accumulatedchanges that reflect past opportunities forlearning emotion-laden meaningful connec-tions, but that only become manifest or ex-pressed in the context of the moment. Such aviewpoint has been adopted by theorists suchas Lerner and Kauffman (1985), who have de-scribed contextualism relative to development


as a view that emphasizes successive (as op-posed to progressive or endpoint-oriented)probabilistic change in the transactions be-tween individual and environment over time.This view allows for more plasticity in develop-ment than do more constrained developmentalmodels, which focus primarily on structuralchange (e.g., a cognitive-developmental ap-proach assumes that the endpoint is the acqui-sition of formal operations). Contextualismmay also be thought of as pragmatic and in-strumental, with philosophical roots in thework of William James (1907/1975) and espe-cially John Dewey (1925, 1934). To under-stand the pragmatics of a given instance ofbehavior, we need to consider the interpen-etration of subject and context; that is, the per-son acts on his or her world even as the worldreciprocally acts on the person. This again ech-oes the relational–functionalist position in de-velopmental theory, insofar as behavior is re-sponsive to an affordant environment andbehavior functions so as to engage instrumen-tally with that affordant environment (Camposet al., 1994).

Time and Context

Lewis (1997) has argued that a contextual viewof development means that earlier events areunlikely to have much relation to later ones.This is especially so “if the earlier events thatare studied are not related to the needs andplans of the individual as they exist now or inthe future” (p. 68). This is a potent idea foremotional development, for it means thatemotion-related behavior at Time 1 need notinfluence or be related to emotion-relatedbehavior at Time 2, unless Time 1 and Time 2are both occasions that are defined by the sameneeds and plans of the individual expressingthe emotion-related behavior. But this raises adouble-barreled question: How might needsand plans be continuous across time and thusrecurrently elicit similar emotional responses,and exactly how much time are we talkingabout? One possible solution is to think interms of how the future can be embedded in thepresent: Our emotional functioning is revealedin how we strive to reach our goals, or arefaced with having to revise them, or may beblocked from attaining them. This process mayyield consequences that prove to be relativelydesirable or undesirable in the here-and-now,but we do not live only in the present. The veryfact that we have goals and intentions means

that the future affects our present actions. Ourwants and desires are the sources of our mo-tives, and thus they orient us toward thefuture—providing us, so to speak, with somenavigational strategies for making it throughthe contextual landscape presently facing us(see also the discussion by Josephs, 1998).Thus, although we may debate whether theonly known endpoint toward which our devel-opment “progresses” is our own death, we arenot devoid of a future in the shorter run. It isfirmly entrenched in our goals, and from thisstandpoint, our present adaptive efforts arewedded to our future. How far into thefuture—10 seconds or 10 years—is not known,for the dynamic flux inherent in context (espe-cially in those circumstances not under ourcontrol) can intervene and lay waste to the bestof our long-range plans. This is not to implythat contextual shifts are somehow undesir-able; indeed, they can also provide unexpectedopportunities for change or even release froman otherwise dreary and emotionally numbinglife situation.

To complicate our thinking about emotionaldevelopment still further, Lewis (1997) has alsoargued that a given behavior can instrumen-tally serve as the means to multiple goals, justas a variety of different behaviors may be use-ful for reaching the same goal. An examplecommon in Western societies that illustrates thefirst point is crying: Tears may be shed in re-sponse to a loss or in response to being deeplymoved (as in awe-inspiring events). Similarly, asocial smile may function as a metacommuni-cative comment about one’s own minor socialgaffe or as a signal to another that his or her so-cial gaffe was noticed but excused. As a furtherillustration, different emotional-expressive be-haviors may be recruited to reach the same gen-eral goal; for example, when children want tobe accepted by their friends (the goal), theyvariously adopt “cool emotional fronts”(Gottman, Katz, & Hooven, 1997; Saarni &von Salisch, 1993) but can also smile engag-ingly and genuinely (von Salisch, 1996). Bothstrategies are useful for fostering relationshipswith peers.

Communication of Emotion and Context

A number of developmental theorists have as-serted that emotion communication cannot beseparated from its context (Barrett, 1993;Fabes, 2002; Saarni, 1989, 1990; Trevarthen,1984, 1993). Certainly the meaning of a partic-


ular facial expression is qualified by the con-text in which it occurs, as was described byCamras (1991): Her young daughter’s expres-sion of disgust was revealed both when her facewas washed and when she was merely pulledinto an upright position. It is the onlooker orrecipient of the communicative message whomust infer what the sender is communicatingabout his or her emotional experience. Did lit-tle Justine Camras experience both face wash-ing and being pulled upright as distressinglyaversive, or were these events simply interrupt-ing whatever she was doing at the time? Or wasone aversive and the other effortful? It isthe parents, upon witnessing their baby’semotional-expressive behavior in conjunctionwith some situation, who attribute distress, ir-ritability, or effort to their infant. Thus emo-tion communication becomes more compli-cated, for now we must add to the contextsurrounding the sender’s emotional-expressivemessage the context surrounding the recipientas he or she attributes meaningfulness to themessage.

Illustrative of the kind of complexity thatone encounters if one seriously wants to exam-ine how context affects emotion communica-tion is the investigation of family conflictundertaken by Noller, Feeney, Sheehan, and Pe-terson (2000) with two-parent households andtheir adolescent twins. Although neither con-flict between the spouses nor a mother’s reportsof conflict with her twins correlated with stylesof conflict between the twins, a father’sreport of negative emotional communication(demand–withdraw patterns that were experi-enced as hostile or invalidating) with his ado-lescent twins was associated with increased ex-pression of negativity and hostility between thetwins. Noller et al. argued that this styleof emotion communication was largely trans-mitted in nonverbal emotionally expressivebehaviors—unpleasant vocalics (e.g., a sarcas-tic tone of voice), negative facial expressions,and other nonsupportive behaviors (e.g.,closed, rigid body posture). In short, partici-pants in emotional communication are both“senders” and “receivers” of emotional-expressive messages, and these messages areimbued with the individuals’ own emotionalexpectancies, projections, and associations.The emotional-expressive exchanges can them-selves be elicitors of emotion in the participantsthat influence subsequent coping efforts (e.g.,withdrawal after being the recipient of a sar-castic comment accompanied by a sneering fa-

cial expression), and obviously relationshipsacquire meaning and nuance as various stylesof emotion communication are exchanged (seealso Gottman et al., 1997; Katz & Woodin,2002; Ratner & Stettner, 1991; Saarni &Weber, 1999; Steinberg & Laird, 1989).

Reciprocal Emotion Communicationand Context

Cross-cultural research on emotion socializa-tion suggests that “emotions can be seen asboth the medium and the message of socializa-tion. Their uniqueness, [as well as] their crucialimportance for understanding development,lies in this dual and encompassing role” (Lutz,1983, p. 60). Thus, even as we may observeemotional development in a child, those whointeract with the child are communicating theirown emotions to the child, often elicited bytheir evaluation of the child’s emotional be-havior (Fabes, Poulin, Eisenberg, & Madden-Derdich, 2002; McDowell, Kim, O’Neil, &Parke, 2002; Spinrad & Stifter, 2002). Paren-thetically, infancy researchers have long notedthis sort of complementary, reciprocal, and in-cremental “dovetailing” of infants’ and moth-ers’ responses to one another (e.g., Cappella,1981; Wasserman & Lewis, 1982). In additionto parents and other family members who areengaged in this reciprocal emotion-socializingprocess with children, the larger world ofpeers, the mass media, and other adult figures(e.g., teachers) are also part of the emotionallycommunicative social context. Thus childrenacquire both emotion-laden beliefs andemotional-expressive behaviors that reflectthese different influences. At the same time,their cultural beliefs about feelings and howthey have learned to express their emotionsconverge toward (sub)cultural norms (Gallois,1994; Kirouac & Hess, 1999; Wierzbicka,1994). Children become culturally predictable—a view elaborated by McNaughton (1996) onparenting practices and cultural identity, andby Tomasello, Kruger, and Ratner (1993) onthe significance of intersubjectivity in culturallearning and human development.

The Taxonomy of Context

In an early paper, Lewis (1978) noted how dif-ficult was the task of creating a meaningful tax-onomy of situations. One could do so by usingphysical properties, such as the location of asituation; the functional activities associated


with situations, such as eating, playing, orworking; or the social/relational aspect of situ-ations, such as being with family members ver-sus with peers. But these dimensions interactwith one another as well: For instance, playingwith peers may occur away from home moreoften for older children, but less often foryounger children. The more complex interac-tion of situational features in older children’splay means that there will be less adult supervi-sion, whereas for younger children there ismore supervision; what then are the implica-tions for the nature of their emotional ex-perience in these different play settings? Thisapproach—taking salient features of a contextand combining them to yield researchquestions—is less often pursued than the re-verse, which is to compare older and youngerchildren’s play with peers and “discover” thatthe context affects what they do!

Examples of how investigators have exam-ined emotional development relative to differ-ent definitions of context include emphasizingthe verbal/sociolinguistic environment that thechild is exposed to (e.g., Denham & Auerbach,1995; Dunn, Brown, & Beardsall, 1991; Lewis& Freedle, 1973; Miller & Sperry, 1987), thepeer group setting as a mutually influencingcontext for emotional experience and/or emo-tion understanding (e.g., Asher & Rose, 1997;Saarni, 1988; Underwood, Hurley, Johanson,& Mosley, 1999), and the influence ofemotional-expressive signals on emotionalfunctioning (e.g., Cassidy, Parke, Butkovsky, &Braungart, 1992; Halberstadt, 1991; Hubbard,2001; Lewis & Michalson, 1985; Saarni,1992).

Researchers often explicitly invoke contextas a substantial part of their investigation intosome emotional process, and in recent years anumber of studies have been carried out in dif-ferent social contexts (e.g., within the family,between peers), and others have described indetail the kinds of processes that characterizethe emotion-laden transactional flow back andforth between child and context (for a review,see Saarni et al., 2006). Process-oriented re-search seems especially well suited to an exami-nation of emotional development in context,but this sort of research approach is not com-mon in investigations of emotional develop-ment (however, see Bainum, Lounsbury, &Pollio, 1984, for a naturalistic study of smilingand laughing in young children; Gentzler,Contreras-Grau, Kerns, & Weimer, 2005, forstudy of parent–child dyadic emotional com-

munication; and Saarni, 1992, for a study ofschool-age children’s attempts to influence theemotional state of a depressed adult). In thenext section, I review several relevant studiesthat seem particularly illustrative of how con-text has been incorporated into research onemotional development.

SOME EXEMPLARY STUDIES

I examine in this section several important cat-egories of how investigators have integratedthe critical role played by social context intotheir research on emotional development. Thecategories into which I have rather arbitrarilyplaced the sorts of context involved include (1)relationship processes and attributes (such asfriendship) that overlap with relationship di-mensions (such as degree of closeness or degreeof conflict); (2) expressive behavior in the“sender” as context for the child recipient; (3)environmental contexts such as culture; and (4)dispositions in the child that interact with so-cial contexts. Due to space constraints, I amnot be able to address the rich literature onyoung children’s acquisition of discourse strate-gies and narrative practices as they relate toa contextual view of emotional development(e.g., Fivush, 1991; Nelson, 1996; Oppenheim,Nir, Warren, & Emde, 1997). Nor can I reviewhere the many fascinating investigations ofhow young children employ language to medi-ate their understanding of emotions in context(e.g., Bretherton, Fritz, Zahn-Waxler, &Ridgeway, 1986; Cervantes & Callanan, 1998;Denham & Auerbach, 1995; Dunn, Brether-ton, & Munn, 1987).

Relationship Processes and Attributesas Mediating or Moderating Contexts

One of the major ways that relationships ascontexts for emotional development have beenstudied in children has been to look at linkagesbetween children’s relationships (most often,those with their parents) and the sorts of emo-tional functioning the children subsequentlylearn or express. To illustrate, Cassidy (1994)has argued that attachment style is related tochildren’s subsequent emotion regulation, andothers suggest that parent–child relationshipsare linked to children’s subsequent emotionalcompetence (e.g., Denham, Mitchell-Copeland,Strandberg, Auerbach, & Blair, 1997), empa-thy (Strayer & Roberts, 2004), and vicarious


emotional responses (Eisenberg, Fabes,Schaller, Carlo, & Miller, 1991). These rela-tional linkages have been construed either as“moderators” (they enhance, amplify, de-crease, or inhibit the expression of the relation-ship) or as “mediators” (they are more directcausal links in the relationship). A considerableliterature exists on young children’s attachmentstatus (secure, insecure/resistant, insecure/avoidant) and their social-emotional compe-tence (reviewed in Contreras & Kerns, 2000).However, it is not always clear how attachmentexerts this influence on children’s subsequentsocial competence; that is, is it mediatedthrough some process of caregiving or throughits influence on children’s emotion regulation,or is it moderated by the child’s temperamentalproclivities? Thompson (2006) argues convinc-ingly that attachment security predicts chil-dren’s social competence (defined as enhancedunderstanding of others’ feelings as well associometric status), and that this relationship ismediated by sensitive caregiving that empha-sizes open discourse about emotions (Cassidy,Kirsh, Scolton, & Parke, 1996; Denham, Blair,Schmidt, & DeMulder, 2002; Raikes &Thompson, 2005). However, it should bepointed out that this relationship appears an-chored in sensitive caregiving that is relativelystable and continuous in children’s lives.

Further support for the important role thatemotion communication plays in relationshipsfor children and their subsequent emotionalfunctioning comes from a recent study byGentzler et al. (2005) and a related study (usingthe same sample of children and parents)by Contreras, Kerns, Weimer, Gentzler, andTomich (2000). The sample consisted of 75fifth graders and their parents, and in theGentzler et al. (2005) study, the children andparents were videotaped dyadically while dis-cussing a distressing episode. A couple of mea-sures were given to the parents to report ontheir children’s tendency to experience negativeemotions and their children’s likely use of vari-ous coping strategies. The children also com-pleted a measure that indexed the degree towhich they were comfortable in affectivelysharing with their parents. The investigatorsglobally scored the observational task, with theintent of establishing the degree to which theparent and child in each dyad could communi-cate emotions openly to one another. Their re-sults indicated that parents’ reports of theirchildren’s emotions and their openness with

their children in the observation task bothcorrelated with their children’s constructivecoping: Those children who were the “bestconstructive copers” were those who werecomfortable communicating their emotions totheir parents, and their parents in turn weresupportive and responsive in emotionally dis-tressing contexts.

In the earlier Contreras et al. (2000) study,the investigators wanted to determine whethermother–child attachment would influence chil-dren’s constructive coping strategies, andwhether these in turn would influence the chil-dren’s peer competence (judged by their teach-ers). The children’s attachment security (as-sessed with a scale developed by Kerns, Klepac,& Cole, 1996) and attachment-related state ofmind (assessed with a semistructured projectiveinterview developed by Resnick, 1993) were in-deed found to be related to their use of con-structive coping strategies (reported by theirmothers), which was found to partially mediatethe children’s peer competence. What occurredwas that the children’s degree of negative emo-tionality (here viewed as a temperamental dis-position) moderated the association betweenconstructive coping and peer competence. Ifchildren who were prone to negative emotion-ality could draw upon constructive copingstrategies, they were as likely to be judged ashaving competent social behavior as childrenlow in negative emotionality were. Their at-tachment relationship was related to their con-structive coping repertoire, but not to theirproneness to negative emotionality. Thus theattachment relationship buffered children whowere disposed toward frequent experiences ofnegative emotion, such that they acquired con-structive coping strategies that they could thenaccess in their social relationships with peers.

From a very different perspective, the vari-ous strategies used by children and youths toregulate anger in the context of close friendshipwas examined by von Salisch (2005). Shedeveloped a self-report scale for use with chil-dren, and a slightly modified version for ado-lescents, that examined anger regulation stra-tegies within friendships—specifically, thestrategies of ignoring, confrontation, redirec-tion of attention, self-blaming, explanation,reconciliation, and humor. Her scale differsfrom other scales assessing anger reactions be-cause of its emphasis on relational context(friendship); also, it does not view anger reac-tions as a trait of the person, unlike measures


such as the State–Trait Anger Expression In-ventory (Forgays, Forgays, & Spielberger,1997) or the Anger Response Inventory forchildren and adolescents, which is similarlydecontextualized (Tangney et al., 1996).

Von Salisch followed 85 children longitudi-nally for 5 years, from late childhood to mid-adolescence, and found that the adolescentshad changed their self-reported regulationstrategies: They were significantly more likelyto use explanation/reconciliation and humor,and less likely to use confrontation, ignoring,redirection of attention, and self-blaming.What this suggests about the developmentalchange in anger regulation is that when chil-dren and youths are invested in a relationship,they are more likely to use emotion regulationstrategies that maintain the relationship. This isnot to say that the more “childish” strategies ofanger regulation that they reported using whenthey were younger disappear from their reper-toire, but they do not think they work well formaintaining or enhancing a relationship thatthey care about. Thus relationship context can-not be separated from the emotion regulationstrategies or coping strategies that an individ-ual uses, even though the temperamental dispo-sition for emotional reactivity may predisposethe individual to more frequent emotionality.Perhaps when the relationship context is one ofanonymity or of an “outsider group,” then themore childish anger regulation strategies be-come more likely (e.g., as in road rage inci-dents).

Emotional-Expressive Behavior as Context

Here I describe three studies examining theinfluence of adult interactants’ emotional-expressive behavior on children’s emotional de-velopment, which in turn influenced the chil-dren’s social behavior. The first study was un-dertaken by McDowell and Parke (2005) with76 fourth-grade children and their parents.They used various global ratings of parents’ in-teractive behaviors with their children whendiscussing a difficult and emotion-evocativetopic, and developed from these ratings twogeneral composite scores. The first was paren-tal use of positive affect, consisting of positiveemotional expressions, negative emotional ex-pressions (reverse-coded), clarity of expression,intensity of expression, and awareness of achild’s feelings. The second was degree of con-trol, consisting of parental regulation of a

child’s emotions and adopting a controlling/di-rective style with a child. The children alsoparticipated in a “disappointing-gift” experi-ence (Saarni, 1984), yielding four ratings: posi-tive expressive responses, negative expressiveresponses, degree of tension, and degree of so-cial monitoring. Their social competence wasalso evaluated via both teacher ratings and peersociometric ratings 1 year later.

The results indicated that parental positiveaffect in the parent–child interaction taskstrongly predicted their children’s use offewer negative expressive behaviors in thedisappointing-gift situation; that is, the chil-dren were able to mask their disappointment atreceiving an inappropriate toy, and thus toavoid violating the common etiquette rule of“Smile and be gracious when given a gift, evenif you don’t like it.” The influence of the paren-tal control variable was also significant, butvaried by parental gender: Higher father con-trol was associated with fewer child positiveexpressive behaviors, and higher mother con-trol was associated with a greater degree of ten-sion manifested by a child.

For the prediction of the children’s socialcompetence a year later, children’s display ruleuse predicted both teacher and peer ratings:Children’s positive expressive behaviors in thedisappointing-gift situation were strongly asso-ciated with positive peer ratings, and their lowuse of negative expressive behaviors in thesame situation was more strongly associatedwith their teachers’ positive social competenceratings of them. Parental positive affect wasalso significantly associated with higherpositive teacher ratings of social competence,whereas for the peer ratings only mothers’positive affect had some influence. Themediational analyses revealed that children’snegative expressive behaviors mediated the ef-fects that their parents’ positive affect and con-trol (especially for fathers) had on their socialcompetence with peers. What this complex pat-tern of results suggests is that as children ma-ture, their parents are less likely to be presentwhen their children are engaged in peer interac-tion, and thus we would presume that they areless likely to directly influence their children’ssocial competence with peers. However, the ef-fects of parents’ emotional-expressive style arestill manifest in terms of how their childrenmanage their own emotional-expressive style,and it is this style that in turn affects children’ssocial skills and acceptance by their peers. Par-


ents’ influence on their children’s socially com-petent behavior with their peers is thus indi-rect.

Another intriguing study on the effects ofcaregivers’ emotional-expressive behavior onchildren’s social behavior was undertaken byKarrass and Walden (2005). They exposed 4-and 5-year-old children to an experimenterwho adopted either a warm, friendly, and re-sponsive (nurturant) expressive style or a cold,unfriendly, and unresponsive (non-nurturant)style. The children’s facial expressions werecoded for 30 seconds during part of this inter-action. Then a second experimenter came inand interacted with the child over a mock task,and the child’s social initiatives toward the sec-ond experimenter were observed. Their resultsindicated that when children interacted with anurturant adult, they then initiated signifi-cantly more overtures toward a second adultthan did children exposed to a non-nurturantadult. The non-nurturant adult also elicited sig-nificantly less happy facial expressions. Lastly,the children’s happiness (as indicated by facialexpressions) with the first experimenter signifi-cantly predicted their social initiatives with thesecond experimenter, and, indeed, their ex-pressed happiness partially mediated the rela-tionship between nurturing caregiving styleand number of social initiatives with the sec-ond adult. Interestingly, expressed sadness didnot partially mediate the relationship betweennurturing style and social initiation. From acontextual point of view, what was apparentlyhappening for these children was that exposureto a short social interaction yielding a happyexperience acted as a prime for subsequent so-ciability, whereas a distressing social interac-tion elicited a more cautious (perhaps tense orsad) response in subsequent social interactionwith another, similar female adult.

Karrass and Walden’s (2005) results havesome parallels with data I obtained with ele-mentary school children (Saarni, 1992) in anearly observational study where the childrenwere faced with having to cheer up a despon-dent “market researcher” who previously hadbeen very friendly. Whereas Karrass andWalden had randomly assigned their childrento either the nurturing or the non-nurturingstyle of interaction, I had children act as theirown controls by having them interact twicewith the same experimenter, who varied her ex-pressive behavior from happy and friendly inthe first meeting to sad and withdrawn in the

second. The social task to be accomplishedseemed straightforward, but the managementof one’s own emotions and behavior presentedcomplex challenges in this case. Being con-fronted with the social context of a depressedwoman could make oneself sad, angered thatone had glibly agreed to try to cheer her up, orchallenged to try to make her happier. Eightychildren across three age groups (7–8, 9–10,and 11–12 years) participated, and all sessionswere held at their school. An assistant accom-panied the children to and from their class-rooms, and prior to the second (“sad”) meetingasked the children to help cheer up her col-league, who was feeling quite “down.” Allagreed to do so.

The children were videotaped throughouttheir interaction with the market researcherwhile doing a task for her. This permitted theestablishment of a base rate of emotional-expressive behaviors when the children werewith the market researcher in her happy state.This base rate could then be compared to whatthey attempted to do when trying to cheer herup in her sad state. As a check on the manipula-tion, a mirror was placed behind the children ina slightly offset position, so that the videocamera could also simultaneously capture the“market researcher’s” expressive behavior to-ward the child.

Results indicated that the oldest children(11–12 years) were the most positive in theiremotional-expressive behavior in both thehappy and sad sessions, and that the middleage group (9–10 years) revealed a curiouslyflat emotional profile in response to the twoemotional-state variations. Among the youn-gest children (7–8 years) were those who didappear to become “engulfed” by the sad re-searcher’s demeanor and looked as though theywould very nearly cry, crawl under the table, ortry to leave. Interestingly, the preadolescents,while showing the most positive behavior to-ward the sad researcher, also revealed the mosttension-filled expressive behaviors—for exam-ple, biting their lips, touching themselves, rub-bing their fingers together, and so forth. Froman impressionistic standpoint, these oldest chil-dren seemed more self-contained and less in-fluenced by the sad researcher’s emotional-expressive behavior, despite their tension-ladennonverbal behaviors. The content of their con-versation with her was also more often “up-beat,” and they talked more than the twoyounger groups, who tended to clam up when


faced with the sad woman. Few gender differ-ences were found in this study, but older girlsdid smile more at the sad market researcher asa strategy to try to cheer her up. In contrast,some of the older boys appeared annoyed atthe prospect of having to help cheer someoneup, despite having agreed to do so (e.g., theydrummed their fingers on the table, bumpedaround in their seats).

This observational study raised many ques-tions about the mutual interaction of two indi-viduals’ emotional-expressive cues, as well asthe added effect on children when they areaware of being videotaped. The contextmanipulation—interacting with either a happyor a sad person—overlapped with other con-textual features: This was a relatively unfamil-iar individual, who was an adult (more power-ful status relative to the child), and the largersetting was that the children were at theirschool. Dealing with a depressed adult wellknown to a child while the child and adult areat home is a very different context, and onethat has been productively explored by Zahn-Waxler and her colleagues with younger chil-dren (Zahn-Waxler, Cole, & Barrett, 1991;Zahn-Waxler, Kochanska, Krupnick, &McKnew, 1990). But in both settings, the effectof having to interact with a depressed adult isnoticeably more negative for younger children.Perhaps one of the emotional tasks that youngadolescents in North American society begin tolearn is how, when, and with whom to “discon-nect” their emotions from those of another, soas to maintain their own emotional boundarieswhen it is adaptive for them to do so (e.g., ado-lescents are relatively less involved in parentalconflict; Cummings, Ballard, & El-Sheikh,1991).

Cultural Practices as Context

Miller and Goodnow (1995) have argued thatthe very “normativeness” of cultural practicesmeans that they are embedded in expectationsabout appropriate behavior, and that such ex-pected behavior is in turn embedded in valuesand assumed conventions that help to give thecultural group its identity. They have arguedfurther that children’s development cannot beseparated from cultural context and its associ-ated practices, for these constitute the route bywhich the children become members of a com-munity, sharing its beliefs and values. At thesame time, developing children have the poten-

tial to change or modulate cultural practices(perhaps especially in adolescence); in otherwords, cultural values, beliefs, and practicescan and do change.

Richly illustrative of the importance of cul-tural values and practices on emotional devel-opment is a recent study by Cole, Tamang, andShrestha (2006) with 3- to 5-year-old childrenfrom two subcultures within Nepal: theTamang, who practice Tibetan Buddhism, andthe Brahman, who are high-caste Hindus andhave higher status in Nepal than the Tamang. Iconcur with Cole and her colleagues when theystate that “cultural values, which organize acommunity’s beliefs and practices for maintain-ing standards of conduct, penetrate caregiverbehavior. In the process of childrearing [whichitself reflects many diverse cultural practices]children’s emotions are socialized” (p. 1239).Cole et al. found that the Tamang householdswere characterized by high adult nurturanceand child trust toward adults, whereas theBrahman households were more likely to pro-mote autonomy and differentiation betweenadults and children. When the researchers ex-amined their exhaustive observations of emo-tion episodes within the households, Tamangadults were more likely to respond to children’sshame with teaching and nurturing, and totheir anger with rebuking and teasing. Consis-tent with some of Cole and colleagues’ earlierwork with this population, Tamang children’sscripts for anger tended to dismiss or minimizeanger (Cole, Brushi, & Tamang, 2002), and an-ger was viewed as undesirable and not sociallycompetent by village elders. The Brahmanadults were more likely to respond to their chil-dren’ anger with positive attention, as in a“teachable moment,” and to children’s shamewith ignoring. Not surprisingly, Tamang chil-dren expressed shame more frequently than an-ger, and Brahman children expressed angermore often than shame. Cole et al. (2006) haveexplained these subcultural differences as re-flecting important values differences and statusdifferences between the two communities. TheTamang value self-effacement and compassion-ate tolerance, and anger is viewed as destruc-tive to social harmony. For the privileged Brah-mans, properly regulated anger may helpto establish dominance and competence; theadults attribute greater value to academic com-petence in their children, and regulated angermay play a part in establishing or claimingone’s societal and educational success. Cole et


al. have also noted that Brahmans’ expressionof anger to lower-status people would not bereacted to negatively, since they are the ones inpower and thus in control. These political–societal differences and religious value differ-ences play themselves out in culturally imbuedcaregiving practices, such that within only afew years the children of the Tamang and theBrahmans reflect these values in their ownscripted emotional responses.

Child Disposition in Interactionwith Context

Children’s temperament and biological disposi-tion toward “emotionality” or behavioral inhi-bition have been extensively investigated, andthe reader is referred to two recent reviews fora far more complete discussion than will beprovided here (Kagan & Fox, 2006; Rothbart& Bates, 2006). Recently Ellis, Boyce, and theircolleagues have proposed that children can alsodiffer in genetically linked reactivity to stressfulcontexts (Ellis, Essex, & Boyce, 2005; Ellis,Jackson, & Boyce, 2006); they suggest that“biological sensitivity to context confers fitnessbenefits not only in highly stressful environ-ments (by augmenting vigilance to threats anddangers) but also in highly protective environ-ments (by increasing permeability to socialresources and support)” (Ellis et al., 2006,p. 205). Their data on preschool children andtheir families revealed a curvilinear relation-ship with reactivity: Disproportionately morechildren were highly reactive either in highlysupportive families or in highly stressed andadverse family environments. In “average”family environments, children were relativelylow in reactivity. Ellis et al.’s theoretical posi-tion is that in environments that offer manypossibilities (“multiniche”), as found in familyrearing situations, many phenotypic expres-sions of stress reactivity will be favored, andthat at the extremes (highly supportive vs.highly adverse), gene-linked regulatory effectswill be switched on (as in reaching a threshold).

Bugental and Beaulieu (2004) adopted thiscontext sensitivity position to explain their re-sults in following medically at-risk children(born prematurely or with poor Apgar scores).They found that as toddlers these children’shealth prospered if their mothers had learnedto provide sensitive and supportive care (and toavoid depression) by having participated in ahome visitation program. In a control group,

mothers did not receive such support, and as aresult their toddlers fared far worse in healthoutcomes. Bugental and Beaulieu interpretedtheir results as showing support for contextsensitivity could be “switched on” by exposureto early stress. The toddlers’ reactivity wasmoderated by supportive and sensitive parent-ing, such that very positive outcomes for thechildren were obtained as they thrived in thissort of protective family environment—in con-trast to biologically vulnerable children rearedin adverse family environments (see alsoBugental, 2003). In a similar vein, early re-search by Riksen-Walraven (1978) hadworking-class mothers trained to provide par-ticularly responsive and contingent caregivingto their infants, with the result that their in-fants demonstrated more exploratory behaviorand learned more quickly on a contingencytask than a similar sample of infants whoselow-income mothers were not trained andwhose caregiving was not optimal. The pointto be made here is that these very immediateand contingent social-interactive responsesprovided an affordant environment for theinfants to respond to, and thus to fulfilltheir potential. A low-affordance, reduced-contingency/nonstimulating social environ-ment is simply not designed to maximize chil-dren’s potential.

Two other studies deserve mention here interms of how children’s emotional dispositionsinteract with social contexts. The first study, byArsenio, Cooperman, and Lover (2000), ob-served 4- to 5-year-olds’ emotional dispositionsin free-play situations at their preschool andcategorized them into anger, happiness, sad-ness, and so forth. They also observed chil-dren’s emotional displays during aggressive epi-sodes (gleeful vs. angry); obtained a measure ofchildren’s understanding of emotion; and hadtheir teachers rate them on social competence,as well as having their peers rate themsociometrically. Their results showed that therelationship among peer acceptance, emotionunderstanding, and emotional disposition wasmediated by the children’s aggression, and alsothat children’s displays of happiness during ag-gression directly affected (negatively) their peeracceptance. These latter children appeared toenjoy being “mean” and may have been dis-playing instrumental, hostile aggression ten-dencies rather than the more ordinary reactiveor provoked aggression. Not surprisingly, theywere not well liked. However, this happiness


displayed during aggressive interactions was it-self not related to the children’s dispositiontoward happiness as displayed during the base-line observational period. Instead, what pre-dicted aggression were the children’s baselinedisposition toward anger, displays of happinessduring the aggressive episode, and deficientemotion understanding. This combination inpreschool children of angry emotional propen-sities, gleeful victimization or hostile powerassertion, and deficient emotion knowledgeappeared to push the children toward increas-ingly negative social outcomes, with the resultthat their social context for emotional develop-ment becomes channeled into one of peer rejec-tion (Dodge et al., 2003; Katz, 2000).

The second study was an examination ofemotion-regulating strategy use relative to dif-ferent emotion-evocative situations. Zimmer-mann and Stansbury (2003) exposed 53 youngpreschoolers to three different sorts of chal-lenging or stress-inducing situations: twoanxiety-inducing situations (approaching amale stranger, interrupting a busy female ex-perimenter who did not want to be interrupted)and a frustration-inducing situation (delay ofgratification with mother present). Their re-sults indicated that both context and tempera-mental reactivity (jointly measured by parentreport and a lab-based observation) influencedwhat sort of emotion-regulating strategy wasused. Self-comforting emotion regulation wasmost often elicited in the approaching-male-stranger situation, and distraction and instru-mental regulatory strategies were more oftenused in the busy-experimenter and delay-of-gratification situations, which were alsoviewed as less distressing to the young children.(More specifically, the busy experimenter wasfemale, she was not trying to engage the child,and a box of toys was nearby; in the frustrationtask, the mother was in the room, and the childhad been asked to wait before eating somecandy.) Temperament influenced this generalpattern, such that shyer children more oftenused the instrumental tactic of withdrawal.Bolder children more often used distractiontactics. Attentional focusing was also related touse of a self-comforting strategy (e.g., suckingone’s thumb), and it appeared as though thesechildren had less flexible deployment of atten-tion; that is, they focused on the stress-inducingsituation instead of accessing available optionsin the situation for distraction purposes (e.g.,exploring the box of toys). The degree of chal-

lenge or stress in the situation tended to makethe effect of temperament more salient, whichmakes sense insofar as low-challenge situations(e.g., the delay-of-gratification situation withmother present) are not likely to elicit dis-positional biases for emotion regulation.

What is potentially of interest here is thattemperament- or disposition-related emotionalstyle may contribute to children’s avoidingsome situations and choosing other situationsthat are conducive to their experiencing lessnegative emotional arousal or distress. Thistendency for emotional disposition and contextto begin showing coherence is borne out inFabes and his colleagues’ research on pre-schoolers’ negative emotionality and their in-creasing experience of social isolation over a 3-month period observation of their socialinteraction—or lack thereof (Fabes, Hanish,Martin, & Eisenberg, 2002). It was not clear inthis research whether these children who wereprone to negative affect were choosing to en-gage more in solitary activities, or whether theywere increasingly being rejected by their peersand thus left alone, but the outcome was thesame: A disposition to display strong and nega-tive emotion was associated with aloneness. In-deed, Asher and colleagues’ work has long in-dicated that rejected children do feel lonely(e.g., Asher, Parkhurst, Hymel, & Williams,1990). Lastly, a longitudinal study of adoles-cents (Murphy, Shepard, Eisenberg, & Fabes,2004) indicated considerable continuity amongreduced social functioning, a bias toward nega-tive emotionality, and reduced regulatory skills(according to parent and teacher reports),again implying an increasing coherence overtime of disposition and social contexts in whichreduced social skills are displayed.

CONCLUSION

The bioecological platform that I endorse forunderstanding the interface (indeed, theinterpenetration) of emotional developmentand context rests upon the ideas put forwardby Bronfenbrenner and Ceci (1994), who pro-posed that proximal processes mediate thelinks between genotype and phenotype. Theseproximal processes are the kinds of social-contextual variables that have been addressedin this chapter—namely, caregiver behaviors,communication within relationships, interac-tion with peers and siblings, and so forth.


These immediate and also motivating interac-tive processes create dynamic environmentsthat become affordant for infants and childrento respond to; these responses in turn changethose dynamic environments, and the ensuingchain of affordances stimulates the inter-actants’ responses to one another further overtime (see also Larsen & Almeida, 1999). Theseaffordant social contexts are often contingenton the dovetailing of interpersonal communi-cative behaviors, and thus the emotional dis-plays of interactants play a very salient role inthe proximal processes mediating between ourgenotypes and what we in fact reveal as ourbehavioral phenotypes. The research describedpreviously by Ellis et al. (2005, 2006) on con-text sensitivity is relevant here as well: Emo-tional reactivity was elevated both in highlyfavorable environments and in adverse envi-ronments, but the former produced highlyadaptive outcomes and the latter poor out-comes (see also Bugental & Beaulieu, 2004). Itis my hope that researchers studying emotiondevelopment will remember that emotions donot occur only in regions of the brain. The de-veloping brain, affected as it is by its geneticheritage, still needs an environment in which togrow and reveal its complexity and plasticity indynamic interpersonal exchanges with othersin an interpersonal context.

REFERENCES

Arsenio, W., Cooperman, S., & Lover, A. (2000). Affec-tive predictors of preschoolers’ aggression and peeracceptance: Direct and indirect effects. Developmen-tal Psychology, 36, 438–448.

Asher, S., Parkhurst, J., Hymel, S., & Williams, G.(1990). Peer rejection and loneliness in childhood. InS. Asher & J. Coie (Eds.), Peer rejection in childhood(pp. 253–273). New York: Cambridge UniversityPress.

Asher, S., & Rose, A. (1997). Promoting children’ssocial-emotional adjustment with peers. In P. Salovey& D. Sluyter (Eds.), Emotional development andemotional intelligence (pp. 196–224). New York: Ba-sic Books.

Bainum, C. K., Lounsbury, K., & Pollio, H. (1984). Thedevelopment of laughter and smiling in nurseryschool children. Child Development, 55, 1946–1957.

Barrett, K. (1993). The development of nonverbal com-munication of emotion: A functionalist perspective.Journal of Nonverbal Behavior, 17, 145–169.

Bretherton, I., Fritz, J., Zahn-Waxler, C., & Ridgeway, D.(1986). Learning to talk about emotions: A function-alist perspective. Child Development, 57, 529–548.

Bronfenbrenner, U., & Ceci, S. J. (1994). Nature–nurture reconceptualized in developmental perspec-tive: A bioecological model. Psychological Review,101, 568–586.

Bugental, D. B. (2003). Thriving in the face of child-hood adversity. New York: Psychology Press.

Bugental, D. B., & Beaulieu, D. (2004). Maltreatmentrisk among disabled children: A bio-social-cognitiveapproach. In R. Kail (Ed.), Advances in child devel-opment and behavior (Vol. 31, pp. 129–164). SanDiego: Academic Press.

Campos, J., Mumme, D., Kermoian, R., & Campos, R.(1994). A functionalist perspective on the nature ofemotion. In N. Fox (Ed.), The development of emo-tion regulation. Monographs of the Society for Re-search in Child Development, 59(2–3, Serial No.240), 284–303.

Camras, L. (1991). Conceptualizing early infant affect:Emotions as fact, fiction or artifact? In K. T. Strong-man (Ed.), International review of studies on emo-tion (Vol. 1, pp. 16–28). New York: Wiley.

Camras, L. (1992). Expressive development and basicemotions. Cognition and Emotion, 6, 269–283.

Camras, L., Meng, Z., Ujiie, T., Dharamsi, S., Miyake,K., Oster, H., et al. (2002). Observing emotion in in-fants: Facial expression, body behavior, and raterjudgments of responses to an expectancy-violatingevent. Emotion, 2, 179–193.

Cappella, J. (1981). Mutual influence in expressivebehavior: Adult–adult and infant–adult dyadic inter-action. Psychological Bulletin, 89, 101–132.

Cassidy, J. (1994). Emotion regulation: Influences of at-tachment relationships. In N. Fox (Ed.), The develop-ment of emotion regulation. Monographs of the Soci-ety for Research in Child Development, 59(2–3,Serial No. 240), 228–249.

Cassidy, J., Kirsh, S., Scolton, K., & Parke, R. D.(1996). Attachment and representations of peer rela-tionships. Developmental Psychology, 32, 892–904.

Cassidy, J., Parke, R., Butkovsky, L., & Braungart, J.(1992). Family–peer connections: The roles of emo-tional expressiveness within the family and children’sunderstanding of emotions. Child Development, 63,603–618.

Cervantes, C., & Callanan, M. (1998). Labels and ex-planations in mother–child emotion talk: Age andgender differentiation. Developmental Psychology,34, 88–98.

Clements, F. E., & Shelford, V. E. (1939). Bio-ecology.New York: Wiley.

Cole, P. M., Brushi, C., & Tamang, B. (2002). Culturaldifferences in children’s emotional reactions to diffi-cult situations. Child Development, 73, 983–996.

Cole, P. M., Tamang, B., & Shrestha, S. (2006). Culturalvariations in the socialization of young children’s an-ger and shame. Child Development, 77, 1237–1251.

Contreras, J., & Kerns, K. (2000). Emotion regulationprocesses: Explaining links between parent–child at-tachment and peer relationships. In K. Kerns, J.Contreras, & A. Neal-Barnett (Eds.), Family and


peers: Linking two social worlds (pp. 1–25).Westport, CT: Praeger.

Contreras, J., Kerns, K., Weimer, B., Gentzler, A., &Tomich, P. (2000). Emotion regulation as a mediatorof associations between mother–child attachmentand peer relationships in middle childhood. Journalof Family Psychology, 14, 111–124.

Cummings, E. M., Ballard, M., & El-Sheikh, M. (1991).Responses of children and adolescents to interadultanger as a function of gender, age, and mode of ex-pression. Merrill–Palmer Quarterly, 37, 543–560.

Denham, S. (1998). Emotional development in youngchildren. New York: Guilford Press.

Denham, S., & Auerbach, S. (1995). Mother–child dia-logue about emotions. Genetic, Social, and GeneralPsychology Monographs, 121, 311–338.

Denham, S., Blair, K., Schmidt, M., & DeMulder, E.(2002). Compromised emotional competence: Seedsof violence sown early? American Journal of Ortho-psychiatry, 72, 70–82.

Denham, S., Mitchell-Copeland, J., Strandberg, K., Au-erbach, S., & Blair, K. (1997). Parental contributionsto preschoolers’ emotion competence: Direct and in-direct effects. Motivation and Emotion, 21, 65–86.

Dewey, J. (1925). Experience and nature. La Salle, IL:Open Court Press.

Dewey, J. (1934). Art as experience. New York: Berke-ley.

Dodge, K. A., Lansford, J., Burks, V., Bates, J. E., Pettit,G., Fontaine, R., et al. (2003). Peer rejection and so-cial information-processing factors in the develop-ment of aggressive behavior problems in children.Child Development, 74, 374–393.


Dunn, J., Brown, J., & Beardsall, L. (1991). Family talkabout feeling states and children’s later understand-ing of others’ emotions. Developmental Psychology,27, 448–455.

Eisenberg, N., Fabes, R., Schaller, M., Carlo, G., &Miller, P. A. (1991). The relations of parental charac-teristics and practices to children’s vicarious emo-tional responding. Child Development, 62, 1393–1408.

Ekman, P. (1989). The argument and evidence aboutuniversals in facial expressions of emotion. In H.Wagner & A. Manstead (Eds.), Handbook of socialpsychophysiology (pp. 143–163). New York: Wiley.

Ellis, B. J., Essex, M. J., & Boyce, W. T. (2005). Biologi-cal sensitivity to context: II. Empirical explorationsof an evolutionary-developmental theory. Develop-ment and Psychopathology, 17, 303–328.

Ellis, B. J., Jackson, J. J., & Boyce, W. T. (2006). Thestress response systems: Universality and adaptive in-dividual differences. Developmental Review, 26,175–212.

Fabes, R. (Ed.). (2002). Emotions and the family. NewYork: Haworth Press.

Fabes, R., Hanish, L., Martin, C. L., & Eisenberg, N.(2002). Young children’s negative emotionality andsocial isolation: A latent growth curve analysis.Merrill–Palmer Quarterly, 48(3), 284–307.

Fabes, R., Poulin, R., Eisenberg, N., & Madden-Derdich, D. (2002). The Coping with Children’sNegative Emotions Scale (CCNES): Psychometricproperties and relations with children’s emotionalcompetence. In R. Fabes (Ed.), Emotions and thefamily (pp. 285–310). New York: Haworth Press.

Fernández-Dols, J. M. (1999). Facial expression andemotion: A situationist view. In P. Philippot, R. S.Feldman, & E. J. Coats (Eds.), The social context ofnonverbal behavior (pp. 242–261). Cambridge, UK:Cambridge University Press.

Fivush, R. (1991). The social construction of per-sonal narratives. Merrill–Palmer Quarterly, 37, 59–82.

Fogel, A., Nwokah, E., Young-Dedo, J., Messinger, D.,Dickson, K. L., Matusov, E., et al. (1992). Social pro-cess theory of emotion: A dynamic systems approach.Social Development, 1, 122–150.

Forgays, D., Forgays, D. K., & Spielberger, C. (1997).Factor structure of the State–Trait Anger ExpressionInventory. Journal of Personality Assessment, 69,497–507.

Fraiberg, S. (1971). Insights from the blind. New York:Basic Books.

Gallois, C. (1994). Group membership, social rules, andpower: A social psychological perspective on emo-tional communication. Journal of Pragmatics, 22,301–324.

Gentzler, A., Contreras-Grau, J., Kerns, K., & Weimer,B. (2005). Parent–child emotional communicationand children’s coping in middle childhood. Social De-velopment, 14, 591–612.

Gibson, E. J. (1982). The concept of affordance in de-velopment: The renascence of functionalism. In A. W.Collins (Ed.), Minnesota Symposium on Child Psy-chology: Vol. 13. Theoretical perspectives on devel-opment (pp. 55–81). Hillsdale, NJ: Erlbaum.

Gottman, J., Katz, L. F., & Hooven, C. (1997). Meta-emotion. Hillsdale, NJ: Erlbaum.

Halberstadt, A. G. (1991). Toward an ecology of ex-pressiveness: Family socialization in particular and amodel in general. In R. S. Feldman & B. Rime (Eds.),Fundamentals of nonverbal behavior (pp. 106–160).New York: Cambridge University Press.

Hubbard, J. (2001). Emotion expression processes inchildren’s peer interaction: The role of peer rejection,aggression, and gender. Child Development, 72,1426–1438.

James, W. (1975). Pragmatism. Cambridge, MA: Har-vard University Press. (Original work published1907)

Josephs, I. (1998). Constructing one’s self in the city ofthe silent: Dialogue, symbols, and the role of “as-if”in self-development. Human Development, 41, 180–195.

Kagan, J., & Fox, N. A. (2006). Biology, culture, and


temperamental biases. In W. Damon & R. Lerner (Se-ries Eds.) & N. Eisenberg (Vol. Ed.), Handbook ofchild psychology: Vol. 3. Social, emotional, and per-sonality development (6th ed., pp. 167–225). NewYork: Wiley.

Karrass, J., & Walden, T. (2005). Effects of nurturingand non-nurturing caregiving on child social initia-tive: An experimental investigation of emotion as amediator of social behavior. Social Development, 14,685–700.

Katz, L. F. (2000). Living in a hostile world: Toward anintegrated model of family, peer, and physiologicalprocesses in aggressive preschoolers. In K. Kerns, J.Contreras, & A. Neal-Barnett (Eds.), Family andpeers: Linking two social worlds (pp. 115–136).Westport, CT: Praeger.

Katz, L. F., & Woodin, E. M. (2002). Hostility, hostiledetachment, and conflict engagement in marriages:Effects on child and family functioning. Child Devel-opment, 73, 636–652.

Kerns, K., Klepac, L., & Cole, A. (1996). Peer relation-ships and preadolescents’ perceptions of security inthe child–mother relationship. Developmental Psy-chology, 32, 457–466.

Kirouac, G., & Hess, U. (1999). Group membershipand the decoding of nonverbal behavior. In P.Philippot, R. S. Feldman, & E. J. Coats (Eds.), Thesocial context of nonverbal behavior (pp. 182–210).Cambridge, UK: Cambridge University Press.

Larsen, R. W., & Almeida, D. (1999). Emotional trans-mission in the daily lives of families: A new paradigmfor studying family process. Journal of Marriage andthe Family, 61, 5–20.

Larsen, R. W., & Prizmic-Larsen, Z. (2006). Measuringemotions: Implications of a multimethod perspective.In M. Eid & E. Diener (Eds.), Handbook ofmultimethod measurement in psychology (pp. 337–351). Washington, DC: American Psychological As-sociation.

Lerner, R. M., & Kauffman, M. (1985). The concept ofdevelopment in contextualism. Developmental Re-view, 5, 309–333.

Lewis, M. (1978). Situational analysis and the study ofbehavioral development. In L. Pervin & M. Lewis(Eds.), Perspectives in interactional psychology (pp.49–66). New York: Plenum Press.

Lewis, M. (1997). Altering fate: Why the past does notpredict the future. New York: Guilford Press.

Lewis, M., & Freedle, R. (1973). Mother–infant dyad:The cradle of meaning. In P. Pliner, L. Krames, & T.Alloway (Eds.), Communication and affect: Lan-guage and thought (pp. 127–155). New York: Aca-demic Press.

Lewis, M., & Michalson, L. (1985). Faces as signs andsymbols. In G. Zivin (Ed.), The development of ex-pressive behavior: Biology–environment interactions(pp. 153–182). New York: Academic Press.

McDowell, D. J., Kim, M., O’Neil, R., & Parke, R. D.(2002). Children’s emotional regulation and socialcompetence in middle childhood: The role of mater-

nal and paternal interactive style. In R. Fabes (Ed.),Emotions and the family (pp. 345–364). New York:Haworth Press.

McDowell, D. J., & Parke, R. D. (2005). Parental con-trol and affect as predictors of children’s display ruleuse and social competence with peers. Social Devel-opment, 14, 440–457.

McNaughton, S. (1996). Ways of parenting and culturalidentity. Culture and Psychology, 2, 173–202.

Miller, P., & Sperry, L. L. (1987). The socialization ofanger and aggression. Merrill–Palmer Quarterly, 33,1–31.

Miller, P. J., & Goodnow, J. (1995). Cultural practices:Toward an integration of culture and development.New Directions for Child Development, 67, 5–16.

Murphy, B., Shepard, S., Eisenberg, N., & Fabes, R.(2004). Concurrent and across time prediction ofyoung adolescents’ social functioning: The role ofemotionality and regulation. Social Development,13, 56–86.

Murphy, L., & Moriarty, A. (1976). Vulnerability, cop-ing, and growth. New Haven, CT: Yale UniversityPress.

Nelson, K. (1996). Language in cognitive development:Emergence of the mediated mind. New York: Cam-bridge University Press.

Noller, P., Feeney, J. A., Sheehan, G., & Peterson, C.(2000). Marital conflict patterns: Links with familyconflict and family members’ perceptions of one an-other. Personal Relationships, 7, 79–94.

Oppenheim, D., Nir, A., Warren, S., & Emde, R. (1997).Emotion regulation in mother–child narrative co-construction: Associations with children’s narrativesand adaptation. Developmental Psychology, 33,284–294.

Raikes, H. A., & Thompson, R. A. (2005). Links be-tween risk and attachment security: Models of influ-ence. Applied Developmental Psychology, 26, 440–455.

Ratner, H., & Stettner, L. (1991). Thinking and feeling:Putting Humpty Dumpty together again. Merrill–Palmer Quarterly, 37, 1–26.

Reissland, N., & Shepherd, J. (2002). Gaze directionand maternal pitch in surprise-eliciting situations. In-fant Behavior and Development, 24, 408–417.

Resnick, G. (1993). Manual for the administration, cod-ing and interpretation of the Separation Anxiety Testfor 11–14 Year Olds. Rockville, MD: Westat.

Riksen-Walraven, J. M. (1978). Effects of caregiverbehavior on habituation rate and self-efficacy in in-fants. International Journal of Behavioral Develop-ment, 1, 105–130.

Rothbart, M. K., & Bates, J. E. (2006). Temperament.In W. Damon & R. Lerner (Series Eds.) & N.Eisenberg (Vol. Ed.), Handbook of child psychology:Vol. 3. Social, emotional, and personality develop-ment (6th ed., pp. 99–166). New York: Wiley.

Russell, J. A. (1991). Culture and the categorization ofemotion. Psychological Bulletin, 110, 426–450.

Saarni, C. (1984). An observational study of children’s


attempts to monitor their expressive behavior. ChildDevelopment, 55, 1504–1513.

Saarni, C. (1988). Children’s understanding of the inter-personal consequences of dissemblance of nonverbalemotional-expressive behavior. Journal of NonverbalBehavior, 12(4, Pt. 2), 275–294.

Saarni, C. (1989). Children’s understanding of strategiccontrol of emotional expression in social transac-tions. In C. Saarni & P. Harris (Eds.), Children’s un-derstanding of emotion (pp. 181–208). New York:Cambridge University Press.

Saarni, C. (1990). Emotional competence: How emo-tions and relationships become integrated. In R. A.Thompson (Ed.), Nebraska Symposium on Motiva-tion: Vol. 36. Socioemotional development (pp. 115–182). Lincoln: University of Nebraska Press.

Saarni, C. (1992). Children’s emotional-expressive be-haviors as regulators of others’ happy and sad states.New Directions for Child Development, 55, 91–106.

Saarni, C., Campos, J., Camras, L., & Witherington, D.(2006). Emotional development: Action, communi-cation, and understanding. In W. Damon & R.Lerner (Series Eds.) & N. Eisenberg (Vol. Ed.), Hand-book of child psychology: Vol. 3. Social, emotional,and personality development (6th ed., pp. 226–299).New York: Wiley.

Saarni, C., & von Salisch, M. (1993). The socializationof emotional dissemblance. In M. Lewis & C. Saarni(Eds.), Lying and deception in everyday life (pp. 106–125). New York: Guilford Press.

Saarni, C., & Weber, H. (1999). Emotional displays anddissemblance in childhood: Implications for self-presentation. In P. Philippot, R. S. Feldman, & E.Coats (Eds.), The social context of nonverbal behav-ior (pp. 71–105). Cambridge, UK: Cambridge Uni-versity Press.

Spinrad, T., & Stifter, C. (2002). Maternal sensitivityand infant emotional reactivity: Concurrent and lon-gitudinal relations. In R. Fabes (Ed.), Emotions andthe family (pp. 243–264). New York: Haworth Press.

Steinberg, S., & Laird, J. (1989). Parent attributions ofemotion to their children and the cues children use inperceiving their own emotions. Motivation and Emo-tion, 13, 179–191.

Strayer, J., & Roberts, W. L. (2004). Children’s anger,emotional expressiveness, and empathy: Relationswith parents’ empathy, emotional expressiveness,and parenting practices. Social Development, 13,229–254.

Tangney, J. P., Hill-Barlow, D., Wagner, P., Marschall,D., Borenstein, J. K., Sanftner, J., et al. (1996). As-sessing individual differences in constructive versusdestructive responses to anger across the lifespan.Journal of Personality and Social Psychology, 70,780–796.

Thompson, R. (2006). The development of the person:Social understanding, relationships, conscience, self.In W. Damon & R. Lerner (Series Eds.) & N.Eisenberg (Vol. Ed.), Handbook of child psychology:

Vol. 3. Social, emotional, and personality develop-ment (6th ed., pp. 24–98). New York: Wiley.

Tomasello, M., Kruger, A. C., & Ratner, H. H. (1993).Cultural learning. Behavioral and Brain Sciences, 16,495–552.

Trevarthen, C. (1984). Emotions in infancy: Regulatorsof contact and relationships with persons. In K.Scherer & P. Ekman (Eds.), Approaches to emotion(pp. 129–157). Hillsdale, NJ: Erlbaum.

Trevarthen, C. (1993). The function of emotions in earlyinfant communication and development. In J. Nadel& L. Cumaioni (Eds.), New perspectives in earlycommunicative development (pp. 48–81). London:Routledge.

Underwood, M., Hurley, J., Johanson, C., & Mosley, J.(1999). An experimental, observational investigationof children’s responses to peer provocation: Develop-mental and gender differences in middle childhood.Child Development, 70, 1428–1446.

von Salisch, M. (1996). Relationships between children:Symmetry and asymmetry among peers, friends, andsiblings. In A. E. Auhagen & M. von Salisch (Eds.),The diversity of human relationships (pp. 59–77).New York: Cambridge University Press.

von Salisch, M. (2005). Anger regulation amongfriends: Assessment and development childhood toadolescence. Journal of Social and Personal Relation-ships, 22, 837–855.

Wasserman, G., & Lewis, M. (1982). The effects of situ-ations and situation transitions on maternal and in-fant behavior. Journal of Genetic Psychology, 140,19–31.

White, G. M. (2000). Representing emotional meaning:Category, metaphor, schema, discourse. In M. Lewis& J. M. Haviland-Jones (Eds.), Handbook of emo-tions (2nd ed., pp. 30–44). New York: GuilfordPress.

Wierzbicka, A. (1994). Emotion, language, and culturalscripts. In S. Kitayama & H. Markus (Eds.), Emotionand culture (pp. 133–196). Washington, DC: Ameri-can Psychological Association.

Wierzbicka, A. (1999). Emotions across languages andcultures. Cambridge, UK: Cambridge UniversityPress.

Zahn-Waxler, C., Cole, P. M., & Barrett, K. C. (1991).Guilt and empathy: Sex differences and implicationsfor the development of depression. In J. Garber & K.Dodge (Eds.), The development of emotion regula-tion and dysregulation (pp. 243–272). New York:Cambridge University Press.

Zahn-Waxler, C., Kochanska, G., Krupnick, J., &McKnew, D. (1990). Patterns of guilt in children ofdepressed and well mothers. Developmental Psychol-ogy, 26, 51–59.

Zimmermann, L. K., & Stansbury, K. (2003). The influ-ence of temperamental reactivity and situational con-text on the emotion-regulatory abilities of 3-year-oldchildren. Journal of Genetic Psychology, 164, 389–409.


C H A P T E R 2 1

Young Children’s Understandingof Others’ Emotions

SHERRI C. WIDEN and JAMES A. RUSSELL

As adults, we not only have emotions but wetry to understand emotions—in a variety ofways. We have concepts for mood and temper-ament and for specific categories of emotions(happiness, sadness, anger, hope, envy, etc.).We judge emotions along broad dimensions ofvalence (feeling good vs. feeling bad) andarousal (feeling low vs. high in energy). Foreach emotion, we know a script, including itseliciting event, conscious feeling, facial expres-sion, vocalization, action, physiological mani-festation, and so on, aligned in a causal andtemporal order. We understand that one’s emo-tional reaction to an event depends on howthat event is appraised, and that the reactioncan be regulated or faked. We use these variousconcepts to understand and predict emotionalreactions and to guide our behavior accord-ingly.

Adult understanding of others’ emotion ispreceded by a long developmental path. Ourchapter discusses that path from infancythrough the preschool years, with a focus on is-sues of taxonomy. Do children begin with aninnate, or at least prepared, set of mental cate-gories for basic emotions? Or do these catego-ries themselves develop out of an earlier, more

primitive understanding? More generally,which aspects of this understanding developearlier and thus are easier to acquire? Andwhat then propels change down this path?

Our perspective in this chapter is unusual, inthat we believe that understanding emotion interms of adult-like discrete categories is a rela-tively late development—an endpoint of a pro-cess of differentiation rather than a startingpoint. We focus on 2-year-olds, who, we claim,see the emotional world largely in terms ofbroad dimensions of valence and arousal. Webegin, however, by examining relevant researchon infants and toddlers, and we provide an in-terpretation different from that usually offered.We end with a brief survey of development inchildren 3 years and older, including the forma-tion of discrete emotion categories.

OUR CONCEPTUAL APPROACH

As adults, we place emotions in a hierarchy, asimplified version of which is shown in Figure21.1. The broadest categories are at a super-ordinate level. These broad categories are sub-divided into more specific ones (at a basic

348

level), which are further subdivided into evenmore specific ones (at a subordinate level). Thequestion is where in this hierarchy children be-gin. Can they begin anywhere? Or do they be-gin in the middle (basic level) and only later ac-quire higher and lower levels? Or do they beginat the top and then differentiate (subdivide) thesuperordinate categories into ever more spe-cific categories?

The most commonly assumed possibility is adiscrete-category account. In this view, childreninitially understand emotions by means of dis-crete basic-level categories that can legitimatelybe labeled “angry,” “scared,” “sad,” and thelike. Children later come to understand thatthose emotions can be grouped together ac-cording to valence and arousal, thereby form-ing the superordinate level. They also subdividethe basic level to form a subordinate level (e.g.,fear is subdivided into anxiety, panic, etc.).This basic-level account is consistent with thegeneral idea that young children often start ac-quiring labels at the basic level of any concep-tual hierarchy (e.g., Markman, 1989). Fordiscrete-category accounts of emotion under-standing, see Denham (1998), Izard (1994),Pons, Harris, and de Rosnay (2004), Saarni(1999), and Walker-Andrews (1997).

Another possibility, a differentiation ac-count, is consistent with the general notion thatcognitive development proceeds through differ-entiation (e.g., Werner, 1948) and is the basicassumption of our approach. Our specific dif-ferentiation account (Widen & Russell, 2003)is supplemented by a circumplex structuralmodel (Bullock & Russell, 1984), according to

which the superordinate level of Figure 21.1 isactually more complex: It consists of the twobroad dimensions of valence and arousal. Theresult can be thought of as four broad catego-ries (Figure 21.2), although without sharpboundaries between them: On the pleasant sideare pleasure + high arousal (which we mightcall “excitement” broadly construed) and plea-sure + low arousal (“serenity”). On the un-pleasant side are displeasure + high arousal(“distress”) and displeasure + low arousal(“depression”).

We believe that this simple scheme is a child’sstarting point and captures a 2-year-old’s men-tal taxonomy for emotions. Children then dif-ferentiate within these broad categories, even-tually arriving at discrete concepts such asanger and fear. Thus the mental categories of

21. Young Children’s Understanding of Others’ Emotions 349

FIGURE 21.1. A hierarchical model of emotion.

FIGURE 21.2. The circumplex model of emotion.Adapted from Bullock and Russell (1984). Copy-right 1984 by the International Society for theStudy of Behavioral Development. Adapted bypermission of Sage Publications, Ltd.

anger, fear, jealousy, and so on are not pre-formed, but must be constructed though a pro-cess we describe as building a script.

Our topic in this chapter is the nature of chil-dren’s understanding of emotion, rather thanthe nature of emotion per se or children’s pro-duction of emotion. Still, we need to refer toemotions, and so, throughout this chapter, wewrite of emotions as if they could be unequivo-cally divided into discrete categories. We writeof facial expressions as if a certain one con-veyed exactly one discrete emotion. And we re-fer to happiness, surprise, fear, anger, disgust,and sadness as basic-level emotions. We writein this way not because we accept these as-sumptions, but for convenience, and to showthat our developmental account of under-standing is largely independent of assumptionsabout the emotions themselves.

INFANTS AND TODDLERS

Many researchers have claimed to find categor-ical understanding of emotion in infants andtoddlers. In this section, we offer a different,more cautious interpretation of these studies.We rely on distinctions among “detection” (asensory system is affected by information),“discrimination” (the ability to tell the differ-ence between two stimuli), and “recognition”(the ability to attribute emotional meaning toa stimulus detected and discriminated); seeWalker-Andrews (1997). It is recognition thatcorresponds to understanding.

Young Infants (<10 Months)

A common method used with the youngest in-fants is habituation: Infants are shown repeatedtrials of one kind of facial expression (e.g., hap-piness) until they habituate (i.e., until the timethey spend looking at each face drops below acriterion). Then a different kind of facial expres-sion (e.g., anger) is shown. Infants, including ne-onates, look longer at the new expression, indi-cating that they detect the change. Although it istempting to suppose that the infants recognizehappiness and anger from those expressions, themore justified interpretation is that they dis-criminate between features or patterns of fea-tures (e.g., see Caron, Caron, & Meyers, 1985).Perhaps infants begin by discriminating facialexpressions on the basis of single features, such

as curvature of the mouth, openness of eyes, orshape of brows, and later discriminate on thebasis of combinations of such features. The in-fants may thus have formed a category of“smile”—but without any corresponding rec-ognition of its meaning or any understanding ofhappiness or feeling good.

Infants have also been said to behaviorallymatch the emotional display that is presentedto them in a way that reveals their recognitionof that display. Thus infants (as young as 10weeks) smile more and show more interestwhen viewing a positive emotional display andhearing pleasant vocalizations; infants aremore agitated and distressed when witnessingan adult frowning or crying (e.g., D’Entremont& Muir, 1999; Kahana-Kalman & Walker-Andrews, 2001). Most such studies have beenrestricted to a positive–negative coding of in-fants’ responses, and thus, if the infants recog-nized the adult display, it might well have beenin terms of valence. To our knowledge, onlyone such study coded infants’ responses interms of specific discrete emotions (Haviland& Lelwica, 1987). This study used Izard’s(1979) Maximally Discriminative Facial Move-ment Coding System (MAX), and thus their re-sults are based on the assumption that certainconfigurations of facial movements can be usedto infer one specific, discrete emotion (for achallenge to this assumption, see Camras,1992). Haviland and Lelwica found that 5-month-olds showed increased happiness andinterest when their mothers displayed happi-ness; that they showed increased anger but de-creased interest when their mothers displayedanger; but that they did not show sadness whentheir mothers displayed sadness. Indeed, thebabies rarely showed sadness at all. Thus, al-though the results of this study give some sup-port to a discrete category interpretation, thesupport is weak. Furthermore, infants’ reac-tions may not require any understanding of theemotion displayed. That is, even if it could beshown that infants react with anger to adultanger displays, the adult displays may simplyelicit emotional reactions from the infants—reactions not mediated or accompanied by un-derstanding. In addition, in these studies, in-fants’ reactions may also have included a com-ponent of imitation.

In intermodal matching, infants are simulta-neously shown two videos of facial expressions(e.g., happiness and sadness), accompanied by


a single voice that matches the emotion of oneof the faces (e.g., Kahana-Kalman & Walker-Andrews, 2001; Soken & Pick, 1999). Themeasure of interest is whether infants look lon-ger at the face that matches the emotion of thesound track than at the nonmatching face. Twoconclusions emerge from this research: First,the youngest infants (from 3.5 to 5 months ofage) do not show intermodal matching; in-stead, they show a preference for a smiling faceor for the face on the right, regardless of vocal-ization. Second, infants about 5–7 months ofage can match the voice to the target facial ex-pression when given two emotions of oppositevalence—evidence consistent with our dimen-sional view (e.g., Kahana-Kalman & Walker-Andrews, 2001; Walker-Andrews, 1986).

The key question is how infants respond toemotions of the same valence. To our knowl-edge, only one study provides such data. Sokenand Pick (1999) investigated 7-month-olds’intermodal matching of all possible pairs ofhappiness, interest (a second positive emotion),anger, and sadness. Overall, there was a match-ing effect plus a preference for specific facesover others: The order of preference was inter-est, happiness, anger, and (least preferred) sad-ness. (In our account, this order translates topositive over negative and high- over low-arousal faces.) For our purposes, the interestingresults came from the same-valence trials (hap-piness vs. interest, sadness vs. anger): Again, in-fants tended to match (i.e., they looked longerat the face whose emotion matched that of thevoice). This finding can be interpreted as recog-nizing discrete emotion categories, but it couldalso be interpreted as matching by arousal (thesecond dimension in our circumplex). An evenless generous interpretation is that, within va-lence, infants matched the level of animation inthe face to the level of animation in the vocal-ization, without any reference to dimensions orcategories of emotion.

To summarize, in our interpretation, adultemotional displays influence young infants(< 10 months) (Witherington, Campos, &Hertenstein, 2001). They respond behaviorallyand emotionally to emotional actions of others.Vocal stimuli can capture their attention. Vo-cal, tactile, and visual displays alter their affec-tive state (Owren, Rendall, & Bachorowski,2003) and regulate their behavior (Campos,Thein, & Owen, 2003). Infants also evidencean early-emerging perceptual ability to detect

and to discriminate between classes of displays.If they recognize any emotional meaning inothers’ displays, that meaning is in terms of va-lence. But neither the infants’ perceptions northeir emotional responses need be mediated byany understanding of the emotional meaning ofthe display. We question whether before about10 months infants are in any way recognizingan emotional message conveyed by a facial orvocal communiqué.

Older Infants and Toddlers(10–24 Months)

By about 10 months, infants build on theirperceptual abilities to begin to find emotionalmeaning in faces and voices. Infants use thisinformation to guide their own behavior andto predict the behavior of another. At thisage, infants begin triadic interactions: An in-fant and caregiver can jointly focus on a thirdstimulus. Infants use these interactions to aidthem in learning which events are rewardingand which are punishing, which to approachand which to avoid. Infants begin to under-stand the referential nature of a caregiver’ssignals (Moses, Baldwin, Rosicky, & Tidball,2001). Older infants and toddlers (10–24months) adjust their behavior according toadults’ emotional displays and thus allow amore convincing demonstration that they findemotional meaning in the events they witness.This adjustment can be seen in the social ref-erencing paradigm.

In the typical social referencing study, an in-fant is presented with an ambiguous stimulus(e.g., a novel toy, a visual cliff). The infanttends to look to the caregiver for clarification.The caregiver, in turn, has been instructed todisplay a particular emotion (e.g., happiness orfear). The measure of interest is whether the in-fant then approaches or avoids the ambiguousstimulus, based on the caregiver’s display (e.g.,Feinman & Lewis, 1983; Klinnert, Emde,Butterfield, & Campos, 1986). By 10–12months of age, infants can indeed use another’semotional display to guide their own behaviorin this situation (e.g., Sorce, Emde, Campos, &Klinnert, 1985). As age increases, toddlers en-gage in social referencing more reliably andmore quickly (e.g., Walden & Kim, 2005).

In most studies, the caregiver displays eithera positive or negative (usually fear or disgust)signal. Thus the reliable finding that infants use


this information is consistent with the idea thatinfants of this age interpret faces in terms of va-lence. For present purposes, the key questionagain is the infants’ response to displays of thesame valence. To our knowledge, only three so-cial referencing studies have compared infants’responses to displays of more than one negativeemotion (Bradshaw, 1986, as cited in Camposet al., 2003; Sorce et al., 1985; Svejda & Cam-pos, 1982, as cited in Campos et al., 2003).Again, in these three studies, infants did re-spond differently to positive and negative dis-plays. Indeed, from their review of how a care-giver’s vocally expressed emotions regulated aninfant’s reaction to an ambiguous toy, Camposet al. (2003) concluded that “behavior regula-tion was a function only of the hedonic tone ofthe signal” (p. 117). None of the three studiesfound a clear behavioral difference in infants’responses to displays of the same valence, butthe study by Sorce et al. (1985) comes closest.Sorce et al. included three negative emotions(fear, anger, sadness) and two positive ones(happiness, interest). The ambiguous stimuluswas a visual cliff. The percentages of infantswho crossed the visual cliff were 0% after afear expression, 11% after an anger expres-sion, 33% after a sadness expression, 74% af-ter a happiness expression, and 73% after aninterest expression. The large difference be-tween negative and positive displays is againconsistent with our dimensional perspective.The much smaller differences within the nega-tive conditions provide limited support for adiscrete-categories interpretation. Our alterna-tive explanation for these within-valence differ-ences is that sadness is a lower-arousal emotionthan either anger or fear. Thus the sadness dis-plays might have simply produced less of a re-sponse in infants. Recall that infants fail to re-spond to sadness faces in intermodal matchingstudies.

On one interpretation, social referencingstudies suggest that toddlers find some mean-ing in adult displays. Because it is limited to abehavioral approach–avoidance measure, so-cial referencing research cannot provide un-equivocal evidence on just what the child un-derstands. An alternative interpretation is thatthe adult facial display elicits a particular statein the infants (e.g., comfort or upset), which inturn influences their willingness to cross the vi-sual cliff. If so, no understanding of emotionwould be necessary in the social referencing sit-uation.

The social referencing paradigm has been ex-panded to show that infants grasp the link be-tween adult emotional displays and the objectof that display (e.g., Moses et al., 2001;Mumme & Fernald, 2003). In these studies, in-fants witnessed an adult emotional display di-rected toward one of two objects but not theother. Both 12- and 18-month-olds were morelikely to play with an object toward which theadult showed happiness and less likely to playwith an object toward which the adult showeda negative emotion. None of these studies com-pared infants’ reactions to different negativeemotions. Thus these studies are again consis-tent with a valence interpretation and leaveopen the question of discrete emotion catego-ries.

With another twist, the social referencingparadigm was used to examine children’s un-derstanding of the connection between an emo-tional display and the displayer’s intentions ordesires. In these referential understanding stud-ies, infants (9–18 months of age) looked longerwhen a person’s emotional display (e.g., sad-ness, happiness) did not predict their behavior(Barna & Legerstee, 2005; Phillips, Wellman,& Spelke, 2002). By 18 months, toddlers un-derstood that someone could want more of afood that the children found undesirable (rawbroccoli), based on emotional displays of plea-sure and disgust (Repacholi & Gopnik, 1997).Thus in the second year of life, an infant isforging connections between others’ emotionaldisplays and their desires. But each of thesestudies again compared only happiness and onenegative emotion—providing, in our interpre-tation, further support for the valence interpre-tation, but remaining silent on infants’ and tod-dlers’ understanding of discrete, basic-levelemotions.

To summarize, it is plausible to suppose thatolder infants and toddlers (10–24 months) findmeaning in emotional displays, but the ques-tion is the precise nature of that meaning. Ofcourse, alternative interpretations are possible,and the studies reviewed here may not be capa-ble of revealing an infant’s full understandingof emotion. Still, it is interesting that in a litera-ture with a fair number of studies conducted byresearchers coming from a categorical perspec-tive, there are no reports of infant behavior forwhich the only explanation would be couchedin terms of their understanding discrete catego-ries of emotions. The same data are consistentwith a dimensional account.


TWO-YEAR-OLDS

Characterizing the emotion knowledge of 2-year-olds is key to capturing the developmentalsequence of emotion understanding. Two-year-olds are beginning to talk, and we begin there.

Children typically begin using emotion la-bels at 18–20 months of age, but their useof these labels is infrequent (e.g., Bretherton,Fritz, Zahn-Waxler, & Ridgeway, 1986; Dunn,Bretherton, & Munn, 1987). Moreover, ac-cording to parental report, most childrenbefore their second birthday have only oneemotion-related word in their vocabulary:“good” (Ridgeway, Waters, & Kuczaj, 1985).Between 24 and 36 months, children add“happy,” “sad,” “angry,” and “scared”(Ridgeway et al., 1985). This, more generally,is a time when children describe others as thesubjects of subjective experiences, such as feel-ing “sleepy,” “tired,” “hot,” and “cold”(Huttenlocher & Smiley, 1990). Children alsopossess a lexical class of “feeling”: Whenasked, “How does this person feel?” andshown facial expressions of emotions, 2-year-olds provided a feeling word on 70.2% of 336trials; more telling, 92.8% of 153 “errors”(those trials on which they provided a non-target verbal response, other than “I don’tknow”) were feeling words, including “tired,”“bored,” and other nonemotion words (Widen& Russell, 2007).

Longitudinal data on five children betweentheir second and fifth birthdays provided bythe Child Language Data Exchange System(CHILDES) allowed Wellman, Harris,Banerjee, and Sinclair (1995) to trace the devel-opment of emotion vocabularies in spontane-ous speech and to infer some of the assump-tions the children were making when they usedemotion terms. Two-year-olds used not just“feel good” (“okay,” “better”) and “happy,”but also “love” for positive feelings. They used“fear,” “anger,” and “sadness” for negativefeelings. They understood that emotions aredistinct from the causes eliciting them and fromthe behaviors and expressions resulting fromthem. These children did not simply use theiremotion words to label their own reactions,but attributed emotions to other people, as wellas to dolls and imaginary friends. They spokeof past and future emotions, not just presentones. The ways they spoke of emotions presup-posed that emotions have “intentional objects”(things the emotions are about) and distin-

guished the objects of the emotions from theircauses. In short, the CHILDES data showedthat 2-year-olds evidenced a mentalistic con-ception of emotion.

What of their implicit emotion taxonomy?On the basis of evidence from spontaneousspeech, Dunn et al. (1987) concluded that 2-year-olds could “distinguish and discuss”(p. 139) a variety of emotions. This conclusion iswarranted to some degree, but the difficulty withstudies of spontaneous speech is that we rarelyknow precisely what children mean by the emo-tion labels they use. When a child uses, say, “an-gry,” the child has not been shown to mean thediscrete emotion category of anger. We need ex-perimental tasks that test this assumption.

Unfortunately, 2-year-olds are rarely in-cluded in experimental studies, and even whenthey are included, they produce so many “er-rors” that their results have been thought diffi-cult to interpret. Traditionally, their responsesare scored “correct” or “incorrect.” Whentheir “correct” responses are greater than ex-pected by chance, this result too has been as-similated into the prevailing presumption thatthey understand the emotional world in termsof discrete emotions, albeit with many errors.When this assumption is examined in 2-year-olds, however, a very different conclusionemerges. When these children do not respond“correctly,” they are not always silent. A closerlook at 2-year-olds’ responses, especially their“incorrect” ones, supports three complemen-tary conclusions.

First, children add emotion words to theirvocabulary in a systematic fashion. In a studyin which children were asked to label the emo-tion conveyed by prototypical facial expres-sions of each of six basic-level emotions, chil-dren varied in the number of emotion labelsthey used (Widen & Russell, 2003). Some usednone. Of those who used one label (regardlessof age), that label was most likely to be“happy.” Of those who used two, some used“happy” and “angry”; the others used“happy” and “sad.” Children then added theother, either “angry” or “sad,” as the third,but, on average, 10 months elapsed before thethird label was added. The same pattern of re-sults was found when children were told storiesof prototypical emotional events and asked tolabel the emotion of the protagonist (Nelson,Widen, & Russell, 2006).

Second, when these children use the labels“sad” and “angry,” they do not mean what


adults mean. For example, in the free-labelingtask just described, children use their few labelsfor all or most of the emotional stimuli pre-sented, not just for the ones adults would labelas happiness, sadness, or anger. The same con-clusion is illustrated by a categorization task(Russell & Widen, 2002a). The category waspresented as a box into which only people whofelt a target emotion could go. The childrenwere then shown, one at a time, photographsof various persons, each with a prototypicalfacial expression of an emotion, and askedwhether each person should go into the box orbe left out. The verbal demands were low, andchildren understood the task. But they did notshow a discrete adult-like category of anger.Rather, for a 2-year-old, “angry” was muchbroader; they were as likely to include sad,fearful, and disgusted faces as angry ones (Fig-ure 21.3). Similarly, shown an array of proto-typical facial expressions and asked to find allthe angry persons, 2-year-olds rarely selectedpositive faces, but did select the full range ofnegative ones, with about equal probability(Bullock & Russell, 1984; Denham &Couchoud, 1990).

The third conclusion from these studies isthat 2-year-olds understand emotional stimulimainly in terms of the broad dimensions ofpleasure (valence) and arousal. In the studiesdescribed above, the breadth of the children’scategories must be understood in terms of va-

lence. In Figure 21.3, for example, the “an-ger” category is extended to include all andonly faces of negative valence. Other studieslead to the same conclusion. In a forced-choice study, each face was paired with eachof eight other faces on different trials (Bull-ock & Russell, 1985). Two-year-olds’ perfor-mance was above chance levels for “happy,”“excited,” “surprised,” “scared, “mad,” “dis-gusted,” “sad,” and “calm” (but not for“sleepy”). At first glance, this result seems tosupport a basic-level-categories approach: 2-year-olds could select the facial expressionthat matched the label for eight of nine emo-tion categories with above-chance accuracy.However, analyses of all their responses (both“correct” and “incorrect”) suggests that thebroad dimensions of valence and arousal mayprovide the better interpretation. Two-year-olds’ performance was higher when the simi-larity of the pair of faces according to thecircumplex model decreased. Similarity in thecircumplex, in turn, can be interpreted interms of similarity along the dimensions ofvalence and arousal. When the 2-year-oldswere labeling facial expressions, their “incor-rect” responses were more likely to be labelsof the same valence than ones of the oppositevalence in both free-labeling studies andforced-choice studies (Denham & Couchoud,1990; Widen & Russell, 2003, Study 2; Nel-son et al., 2006; Bullock & Russell, 1985).


FIGURE 21.3. Faces that 2-year-olds included in the angry box.

The studies reviewed so far have relied onchildren’s production or comprehension ofemotion words. The fourth conclusion is thatthis evidence captures their underlying concep-tual structure and not simply something oftheir word use. Recall the categorization task.An interesting feature of that study was thatthe anger box had been defined in two ways(Russell & Widen, 2002a): For half of the chil-dren the experimenter used the word “angry,”and for the other half the experimenter usedtwo photographs of prototypical anger facialexpressions (“feel like this” [pointing]). Bothways of defining the angry box produced simi-lar results—2-year-olds included the majorityof all negative expressions in this box—eventhough, in the latter condition, the problemcould have been solved for a discrete anger cat-egory simply by perceptual matching.

Bullock and Russell (1984; Russell & Bull-ock, 1986b) used yet another method that didnot rely on emotion words: multidimensionalscaling of judged similarity between facial ex-pressions. Although the stimuli were facial ex-pressions thought to convey basic discrete emo-tions, multidimensional scaling showed that 2-year-olds judge similarity on the basis of twobroad dimensions—valence and arousal.

Two-year-olds are also learning about thecauses and consequences of emotions. Wellmanand Woolley (1990), in their research on under-standing of desires, convincingly demonstratedthat 2-year-olds can associate happiness withdesired outcomes and sadness with undesiredoutcomes. In a study of slightly older children(as young as 3 years), Stein and Levine (1989)found that children could select a positive facefor a positive outcome (e.g., receiving a toy)and a negative face for a negative outcome(e.g., losing a puppy); these same children didnot, however, distinguish within negative out-comes. That is, they were as likely to select asad as an angry face for outcomes that adultsjudged to elicit sadness, and the same for out-comes adults judged to elicit anger. Similarly,Trabasso, Stein, and Johnson (1981) foundthat 3- and 4-year-olds labeled events that frus-trated their goals as both “sad” and “angry,”and only older children distinguished the two.Other studies (e.g., Borke, 1971; Denham &Couchoud, 1990) have obtained similar re-sults, again with children older than 2. We be-lieve that 2-year-olds would show similar re-sults, but the two studies we know of thatactually tested 2-year-olds yielded mixed re-

sults (Denham & Couchoud, 1990; Widen &Russell, 2006)—presumably because tasks thatare manageable for 3-year-olds might not be sofor 2-year-olds.

Summary of the Research

Two-year-olds are surprisingly sophisticated intheir mentalistic conception of emotion as astate separate from its causes and behavioralconsequences. On the other hand, they are sur-prisingly limited taxonomically to the broad di-mensions of valence and arousal. Children latermove toward more specific discrete categoriesof emotion, but slowly and with difficulty (formore on this, see below). We found little evi-dence that children younger than 3 years un-derstand anger, fear, and other discrete catego-ries of emotion as such. When they use thewords “scared,” “angry,” and so on, they seemto have in mind not discrete emotions but verybroad categories of emotion, perhaps initiallysimply happy and unhappy ones. The evidenceon 2-year-olds reinforces our cautious interpre-tation of evidence on those younger than 2years.

This perspective is also consistent with thefinding from the theory-of-mind perspectivethat 2-year-olds, lacking a full notion of belief,are limited to desire and perception (Wellman& Woolley, 1990). They attribute desires toothers, and understand that fulfillment of de-sire leads to positive emotion, whereas unful-filled desire leads to negative emotion. Further-more, 2-year-olds understand desires as mentalstates, attributing different desires to differentindividuals. They therefore judge that “Billwho wants a bunny and finds one will behappy whereas Mary who wants a kitty andfinds a bunny—the exact same bunny that Billfound—will be sad” (Wellman, 1995, p. 302).Two-year-olds also understand that others canperceive or fail to perceive an event (thebunny), but these children fail to understandthat others can have different beliefs. To the ex-tent that specific discrete emotions presupposecertain beliefs (Ortony, Clore, & Collins,1987), the theory-of-mind research thus rein-forces our doubts about how much 2-year-oldscan understand about specific discrete emo-tions. That is, 2-year-olds routinely fail false-belief tasks, suggesting that they would fail toappreciate that a dog they know to be harmlesscould elicit fear in someone who believes itdangerous.


PRESCHOOLERS

In this section, we consider children 3 to 5years of age, and build on the premise that 2-year-olds understand others’ emotions in termsof the broad dimensions of pleasure andarousal—an understanding that complementstheir perception-and-desire theory of the mindsof others. During their fourth year, children be-gin to evidence fuller understanding of beliefs(Wellman, 1995); this advance complementsand perhaps underlies their fuller understand-ing of emotion concepts they already have in el-ementary form and their division of emotioninto ever finer discrete categories. We presentfive hypotheses, which together constitutewhat we call the “differentiation model” ofemotion understanding (Widen & Russell,2003).

First Hypothesis

Valence and arousal dimensions continue to beimportant. When preschoolers make “errors,”or when they judge similarity among emotions,they continue to show the influence of pleasureand arousal (Bullock & Russell, 1984, 1985;Widen & Russell, 2003). In all, seven studies todate have analyzed preschoolers’ “incorrect”emotion responses on labeling faces (Bullock &Russell, 1984, 1985; Denham & Couchoud,1990; Widen & Russell, 2002; 2003, Study 2,Study 3; 2004). At every age, valence continuedto dominate children’s “errors.” Similar resultsoccurred in two studies that specified the targetemotion with stories rather than faces (Den-ham & Couchoud, 1990; Widen & Russell,2004). Before children know other features ofshame, gratitude, pride, and jealousy, theyknow their valence (Russell & Paris, 1994). Va-lence can be seen in even older children andadults (Bullock & Russell, 1984; Coren & Rus-sell, 1992; Russell & Bullock, 1986a, 1986b;Russell & Fehr, 1994) and in a variety of cul-tures with a variety of languages (e.g., Russell,Lewicka, & Niit, 1989). Evidence for valence isubiquitous.

Second Hypothesis

Children use different emotion labels with dif-ferent frequencies. In spontaneous speech, chil-dren use some emotion labels more frequentlythan others. But the children may simply be ex-periencing or witnessing some emotions more

frequently than others. However, in a study inwhich an equal number of emotions (repre-sented as facial expressions) were presented(Widen & Russell, 2003, Study 3), the samedifferential frequency was found. Labels wereused in the following rank order, starting withthe most frequent: “happy,” “sad,” “angry,”“scared,” “surprised,” and “disgusted.” Dif-ferential frequency of label use for faces hadbeen reported before (e.g., Gosselin & Simard,1999; Izard, 1994), and the pattern was inter-preted in one of two ways: (1) Some facial ex-pressions (i.e., happiness, sadness, anger) areeasier to recognize than others, or (2) somechildren lack some words in their vocabulary.We showed that the same order of use occurredfor both “correct” and “incorrect” uses; differ-ential use of labels was therefore not a result ofthe faces per se, but rather of children’s inter-pretation of those faces. Furthermore, all chil-dren in the study had been shown in a priortask to have all six labels in their vocabulary.We thus interpret differential use of emotion la-bels as reflecting differences in accessibility ofthe emotion concepts, which in turn are corre-lated with the order in which the categories areacquired, as detailed in the next hypothesis.

Third Hypothesis

Emotion categories enter a child’s taxonomy ina systematic order. As we have mentioned ear-lier, when children (2–5 years of age) weresorted, irrespective of age, by the number ofdifferent emotion category labels they used forfacial expressions (or, in a separate study, emo-tion stories), labels were found to emerge in asystematic order (Figure 21.4) (Nelson et al.,2006; Widen & Russell, 2003). Earlier, wehave described the steps up to three labels—“happy,” “angry,” and “sad.” The next step al-lows two paths: Some children added “sur-prised,” and some “scared.” For five labels, thetwo paths merged. In the last step, “disgusted”was added. Age increased with the number oflabels used, from a mean age of 30 months forthose children producing no labels to 62months for those producing five labels. Over81% of the children (vs. the 23% expected bychance) fit the pattern of Figure 21.4. We havesince replicated this result three times, and eachtime the proportion of children who fit the pre-dicted pattern has been high: 78% (Widen &Russell, 2007), 86% (Widen & Russell, inpress), and 91% (Nelson et al., 2006). Further-


more, children’s position in the scheme of Fig-ure 21.4 predicts their categorization of emo-tional stimuli in separate tasks (Widen &Russell, 2007, in press), even when age is con-trolled for.

In many studies, children have been asked toassociate emotions with emotion labels. Theubiquitous conclusion, starting with Gates(1923), is that the proportion of correct associ-ations increases with age. A second conclusionhas been that the proportion correct varieswith emotion. This latter finding provides anindirect test of our scheme in Figure 21.4.Earlier-emerging categories should be morepracticed and accessible; if so, then we can ex-pect that children will be more often correctwith the earlier-emerging categories. In a re-view of 19 studies (Widen, 2005), preschoolers’performance was highest on happiness, sad-ness, and anger, followed by fear, disgust, andsurprise, although the order within these twosets varied with task. This overall pattern heldfor three response formats, whether emotionswere represented by facial expressions or sto-ries, and whether emotion labels were the inde-pendent or dependent variable. Thus our devel-opmental sequence shown in Figure 21.4 isprobably not limited to the experimental con-text in which it was discovered (free labeling offaces and stories).

Of course, there are far more emotions thanthe six shown in Figure 21.4. The evidence sug-gests that additional emotion concepts enterthe developmental picture after Labeling Level5. For example, given stories about emotions,preschoolers were able to label our first fiveemotions (happiness, sadness, anger, fear, andsurprise) earlier than shame, contempt, or love(e.g., Wintre & Vallance, 1994; cf. Brody &Harrison, 1987). Similarly, they are able to de-scribe the causes for our first five emotions ear-lier than those for shame, gratitude, pride, orjealousy (Harris, Olthof, Meerum Terwogt, &Hardman, 1987; Russell & Paris, 1994).

Fourth Hypothesis

Categories begin broad, but then narrow. As isevident from Figure 21.4, emotion categoriesbegin broad. During one period in the life ofmany 2-year-olds, the category labeled “angry”includes all negative emotions. With time, andas new categories emerge, this broad “anger”category gradually narrows. Children continueto use “angry” for the anger face, but are lesslikely to use it for sadness and fear faces on thelabeling task (Widen & Russell, 2003) and onthe box (categorization) task (Russell &Widen, 2002a). Narrowing has also beenfound when children were presented with an


FIGURE 21.4. Systematic emergence of emotion labels. LL, Labeling Level. Adapted from Widen andRussell (2003). Copyright 2003 by the American Psychological Association. Adapted by permission.

array of facial expressions and asked to find allwho displayed a particular emotion (Bullock &Russell, 1984, 1985; Bormann-Kischkel,Hildebrand-Pascher, & Stegbauer, 1990).

In our interpretation, narrowing beginswhen preschoolers begin to use the arousal di-mension to distinguish between negative emo-tions. This process can be seen, for example,when preschoolers begin to exclude nontargetfacial expressions from the anger box (Russell& Widen, 2002a). Thus, for the anger box, theorder in which faces are excluded is predictableon the basis of arousal: Children first excludesad faces, which are the most dissimilar to an-ger faces on arousal. They next exclude fearfulfaces and finally disgusted faces, although most5-year-olds still include the disgusted face (nar-rowing is gradual and incomplete even at theend of the preschool years).

Later-emerging emotion categories also nar-row with age. We had thought, based on chil-dren’s free-labeling responses, that fear, sur-prise, and disgust began narrower (Widen &Russell, 2003), but a later categorization studyshowed that they begin just as broad as do hap-piness, anger, and sadness (Widen & Russell,2007). Pride, jealousy, and so on similarly be-gin as broad concepts defined by nothing ex-cept valence, but then narrow as they take onmore adult meaning (Russell & Paris, 1994). Inthese cases, narrowing is unlikely to be basedon arousal but on other factors, as detailednext.

Fifth Hypothesis

Children form a script for each emotion cate-gory. As adults, we know that each emotion is asequence of subevents. In fear, prototypically, adanger occurs; the person orients to it, freezesor flees, and feels unhappy; physiologicalarousal increases; face and voice change. Theconcept of fear is thus a script in whichsubevents unfold in a temporal and causal or-der. The script for sadness may include a loss,resulting in feeling bad, pouting or crying,whining, withdrawal, tears, and slow or sup-pressed movement. Children must acquirethese scripts and their labels. In the studies de-scribed so far, it is not clear just how much ofeach script a child knows. For example, associ-ating the word “scared” with a type of facialexpression need not imply that the child knowsother subevents of the fear script. Our fourthhypothesis suggests that scripts begin with few

components (which explains their initialbreadth), but then acquire new components(and hence narrow).

How is a script built? Which parts enter thescript earlier and which later? Answers to suchquestions can hint at the process of building ascript. For example, some theorists have as-sumed that facial expressions are the bases forconstructing scripts (e.g., Harris, 1989; Izard,1994). If so, then a reasonable hypothesiswould be that witnessing different facial ex-pressions would compel the child to differenti-ate a currently broad script. According to thishypothesis, a child at Figure 21.4’s LabelingLevel 2a (who labels all negative emotions as“angry”) would soon notice that some negativefacial expressions involve downcast eyes,downturned mouth, and tears (or perhaps,more generally, facial signs of low arousal),whereas other faces involve knitted brows,staring eyes, bared teeth, and clenched jaw (or,more generally, facial signs of high arousal). Asa consequence, the child would then divide theinitially broad category into two separate cate-gories and then become receptive to differentlabels (“sad,” “angry”) for this second cate-gory of negative emotions (although at thislevel both of these categories remain broaderthan the adult version).

Studies of children’s scripts for emotionsmust present a child with one part of a script(mode of presentation) and ask the child foranother part (mode of response). This face-early hypothesis can then be tested by includingfacial expressions in one mode or the other. Forexample, Camras and Allison (1985) told chil-dren (preschoolers to second graders) very briefstories about a girl (e.g., “Her mother hasdied”). Children were asked to identify thegirl’s emotion, using one of two responsemodes: an array of labels (“happy,” “angry,”“sad”) or an array of corresponding faces(smiling, frowning, or crying). Much to every-one’s surprise, children did better overall giventhe label response format than given the faceresponse format. Ten studies have now com-pared facial expressions to at least one othermode: (1) emotion labels (Camras & Allison,1985; Russell, 1990; Russell & Widen, 2002a,2002b); (2) emotion stories describing thecauses and/or consequences (Markham & Ad-ams, 1992; Nelson et al., 2006; Widen & Rus-sell, 2002, 2006); or (3) tone of voice (Stifter &Fox, 1987). All but one study (Stifter & Fox,1987) found a difference between modes, but


none found that facial expressions were thestronger cues to emotion. This face inferiorityeffect was particularly strong for fear. The infe-riority of faces was robust whether mode wasthe independent or dependent variable, andwhether children were asked to categorize,choose from an array, label, or describe thecause or consequence of the emotion.

Instead of faces as the bases of emotionscripts, there is emerging evidence that no onecue type is the strongest at all ages or for allemotions. We (Widen & Russell, 2004) askedchildren (3 or 4 years old) to describe thecauses of six different emotions, specified byfacial expressions, labels, or brief stories de-scribing behavioral consequences. Again, faceswere not the strongest cues for any emotion orfor either age group. Behavioral consequenceswere the strongest cues for 3-year-olds, espe-cially for anger. Emotion labels were the stron-gest cues for 4-year-olds, especially for fear anddisgust. It remains possible that facial expres-sions are stronger cues for even younger chil-dren and for the earliest-emerging broad emo-tion categories (e.g., feeling good, feeling bad),however a child may label them.

Further evidence that faces are relativelyweak cues to emotion is that, even given proto-typical facial expressions of emotion, childrentake into account the apparent sex of the per-son whose emotion they are categorizing. We(Widen & Russell, 2002) asked children to la-bel someone’s emotion on the basis of either aprototypical facial expression or a brief stereo-typical story. The “boy’s” and the “girl’s” facesand expressions were actually identical, but thefaces were made to appear “male” and “fe-male” by adding gender-appropriate hairstyles.The boy’s and girl’s stories for each emotionwere also identical except for names and pro-nouns. Even given these clear cues to emotion,the perceived sex of the protagonist influencedchildren’s emotion attributions. (This evidencealso implies that scripts contain gender stereo-types and are thus not equivalent to scientificaccounts of each emotion.)

Insufficient evidence is available to provide aclear account of how the scripts are acquired.Perhaps the scripts for happiness and sadnessbegin by linking facial expressions (smiling vs.crying) to feeling good versus bad, which thensoon become linked to their causes (meeting vs.not meeting desires). For anger, the script maybegin instead by linking hostile behavior tofeeling bad. For fear and disgust, the script may

begin with a child’s hearing the labels “scared”and “disgusted,” which prompts him or her tosearch for causes that differentiate these emo-tions from sadness and anger. In each case,once the first link is formed, the child can thenadd more components. Because different cuesinitiate the process for different emotions, dif-ferent cues are more powerful in eliciting theconcepts for different emotions. Because con-cepts for different emotions are acquired at dif-ferent ages, different cues are more powerful atdifferent ages.

Harris (2000) rightly objected to the notionof scripts as we have so far characterized themon the standard grounds offered for appraisaltheories of emotion: The emotion that occursdepends not so much on the reality of the pre-cipitating event as on how the precipitatingevent is appraised. An approaching dog elicitsjoy in the person who appraises the dog asplayful, but elicits fear in the person who ap-praises it as dangerous. Understanding howemotions depend on appraisals is one of thetasks for the child developing an understandingof emotion. Although they are not physicalevents, appraisals can be considered assubevents of the emotion process and hence aselements in the script.

Appreciating the role of appraisals presup-poses an understanding that different peoplecan have different appraisals. A person’sappraisal of an event is a belief about thatevent. Several sources of evidence hint at thedependence of emotion understanding on beliefunderstanding. Of course, the two are corre-lated, but the dependency may go in one direc-tion. First, belief understanding does not seemto depend on emotion understanding, as shownby evidence from children with developmentaldisorders (e.g., conduct disorder, autism) whocan understand belief but not emotion (e.g.,Blair, 2002). Second, there is clear evidencethat children understand false beliefs beforethey can make the corresponding belief-basedemotion attributions (e.g., Harris, Johnson,Hutton, Andrews, & Cooke, 1989; de Rosnay,Pons, Harris, & Morrell, 2004). For example,Bradmetz and Schneider (1999) demonstratedthat children understood that when she first ar-rived, Little Red Riding Hood believed that thewolf was her grandmother. Nonetheless, thesesame children said that Little Red Riding Hoodfelt afraid rather than happy to see her grand-mother. Thus they had mastered the false-belieftask, but could not then attribute happiness to


Little Red Riding Hood at seeing her grand-mother when they knew that it was in fact thewolf. Third, Rieffe, Meerum Terwogt, andCowan (2005) asked children as young as 4years to explain various emotional episodes.For happiness, anger, and sadness, childrenspontaneously referred to desires but not be-liefs. In contrast, in explaining fear, they re-ferred to beliefs. This evidence suggests that thethree earliest concepts remain tied for a time tothe desire psychology of their origin. In addi-tion, many later, so-called “social” or “self-conscious” emotions are largely defined by be-liefs about the eliciting situation. Regret, for in-stance, has no uniquely identifying facial, vo-cal, physiological, or behavioral action patternassociated with it, but centrally involves feelingbad based on the belief that a prior decision orevent did not turn out well.

Of course, children’s understanding of emo-tion continues to develop beyond what we havedescribed so far. Pons et al. (2004) identifiednine different components of children’s under-standing of emotion. In addition to the compo-nents we have already discussed here, childrencome to understand (1) how reminders of anevent can reactivate an emotion, (2) how emo-tions can be controlled, (3) that there can be adiscrepancy between outer appearance of emo-tion and inner experience, (4) that mixed emo-tions can exist, and (5) that emotions dependon the morality of the precipitating event.

CONCLUSION

One particularly interesting perspective onconceptual development in general is called“theory theory” (e.g., Gopnik & Wellman,1994). From this perspective, a child’s under-standing of emotion is a theory; changes in un-derstanding are changes in the theory. A child’sinitial theories are powerful, are often biologi-cally given, and aid learning, but they arerelatively simple and cannot explain all theevidence the child encounters. As children en-counter unexplained events and even counter-evidence, auxiliary hypotheses may be added tothe original theory, without seriously alteringit. But eventually, as in science, the old theoryproves to be inadequate and a new theory is de-veloped.

For emotion understanding, we propose thatchildren’s earliest theory includes the conceptsof valence and arousal. As shown in Figure

21.2, these two dimensions can also be thoughtof as four broad categories: pleasure/higharousal, pleasure/low arousal, displeasure/higharousal, and displeasure/low arousal. Althoughsimple, this theory allows a child to place theemotions of others in these broad categoriesand thus to anticipate the affective quality oftheir subsequent behavior, to acquire knowl-edge of the positive and negative quality of cur-rent events (such as a visual cliff), and to gainknowledge of the others’ desires. Evidence indi-cates that this dimensional theory dominatesthe child’s thinking about emotion for the sec-ond and most of the third years of life.

Emotion researchers have long debatedwhether emotions are understood in terms ofdimensions or categories. When that debate isrephrased as a developmental question, the twosides of the debate can be reconciled. How doesa child’s initial broad dimensional understand-ing turn into the adult division of emotion intomore discrete categories? Or, phrased in termsof the four broad categories, how are the broadcategories differentiated into more specificones? How are new categories acquired? Andwhat propels these changes?

To answer the question of change, the theorytheory points to the evidence facing the child.The four-category system does not allow suffi-ciently precise accounts of this evidence, and sothe theory must be changed. The kind of evi-dence faced is what is eventually incorporatedinto emotion scripts: different kinds ofemotion-eliciting situations, different overtbehavior, different expressive behavior, and dif-ferent labels used by the language community.This new theory still has valence and arousal asits bases, but can better accommodate the newevidence.

Another force on the child’s developing the-ory of emotion is the development of other per-ceptual and cognitive abilities. Infants must ac-quire perceptual categories of facial actionsbefore being able to attach emotional signifi-cance to those categories. Older children mustdevelop a theory of mind. The concept of desireis needed for a child to understand why thesame event can bring positive feelings to onebut negative feelings to another. The concept ofbelief is needed to understand that differentpeople can appraise the same situation differ-ently and thus emotionally respond differently.

Gopnik and Wellman (1994) contrasted thetheory theory with two other broad perspec-tives on conceptual development, both of


which could incorporate emotion understand-ing. According to the notion of innate modules,the prototype of which is a Chomskian accountof language, more of emotion understandingwould be universal and innate. If the categoriesfor the basic emotions and their links to spe-cific expressions are so viewed, then this per-spective is similar to the usually encounteredstandard account of emotion understanding. Incontrast, according to the simulation account(Harris, 1992), emotion understanding beginsnot with witnessing the emotions of others, butwith experiencing them oneself. Presumably,the discrete categories of emotion experienceare biologically given. In trying to anticipatethe emotions of others, one does not rely on aconceptual theory or an innate module, but cansimulate the experience in oneself. The contrastbetween the theory theory and the simulationtheory points to the importance of studyingpersonal experiences of emotion and their rolein understanding others.

ACKNOWLEDGMENTS

We thank Lisa Feldman Barrett, Paul L. Harris,Mathew J. Hertenstein, Nicole L. Nelson, and Caro-lyn Saarni for their help and comments.

REFERENCES

Barna, J., & Legerstee, M. (2005). Nine- and twelve-month-old infants relate emotions to people’s ac-tions. Cognition and Emotion, 19, 53–67.

Blair, R. J. R. (2002). Theory of mind, autism, and emo-tional intelligence. In L. F. Barrett & P. Salovey(Eds.), The wisdom in feeling: Psychological pro-cesses in emotional intelligence (pp. 406–434). NewYork: Guilford Press.

Borke, H. (1971). Interpersonal perception of youngchildren: Egocentrism or empathy? DevelopmentalPsychology, 5, 263–269.

Bormann-Kischkel, C., Hildebrand-Pascher, S., &Stegbauer, G. (1990). The development of emotionalconcepts: A replication with a German sample. Inter-national Journal of Behavioral Development, 13,355–372.

Bradmetz, J., & Schneider, R. (1999). Is Little RedRiding Hood afraid of her grandmother?: Cognitivevs. emotional response to a false belief. British Jour-nal of Developmental Psychology, 17, 501–514.

Bretherton, I., Fritz, J., Zahn-Waxler, C., & Ridgeway,D. (1986). Learning to talk about emotions: A func-tionalist perspective. Child Development, 57, 529–548.

Brody, L. R., & Harrison, R. H. (1987). Developmental

changes in children’s abilities to match and labelemotionally laden situations. Motivation and Emo-tion, 11, 347–365.

Bullock, M., & Russell, J. A. (1984). Preschool chil-dren’s interpretation of facial expressions of emotion.International Journal of Behavioral Development, 7,193–214.

Bullock, M., & Russell, J. A. (1985). Further evidenceon preschoolers’ interpretations of facial expressionsof emotion. International Journal of Behavioral De-velopment, 8, 15–38.

Campos, J. J., Thein, S., & Owen, D. (2003). A Darwin-ian legacy to understanding human infancy: Emo-tional expressions as behavior regulators. Annals ofthe New York Academy of Sciences, 1000, 110–134.

Camras, L. A. (1992). A dynamic systems perspective onexpressive development. In K. Strongman (Ed.), In-ternational review of studies on emotion (pp. 16–28).New York: Wiley.

Camras, L. A., & Allison, K. (1985). Children’s under-standing of emotional facial expressions and verballabels. Journal of Nonverbal Behavior, 9, 84–94.

Caron, R. F., Caron, A. J., & Myers, R. S. (1985). Do in-fants see emotional expressions in static faces? ChildDevelopment, 56, 1552–1560.

Coren, S., & Russell, J. A. (1992). The relative domi-nance of different facial expressions of emotionunder conditions of perceptual ambiguity. Cognitionand Emotion, 6, 539–556.

Denham, S. A. (1998). Emotional development inyoung children. New York: Guilford Press.

Denham, S. A., & Couchoud, E. A. (1990). Young pre-schoolers’ ability to identify emotions in equivocalsituations. Child Study Journal, 20, 153–169.

D’Entremont, B., & Muir, D. (1999). Infant responsesto adult happy and sad vocal and facial expressionsduring face-to-face interactions. Infant Behavior andDevelopment, 22, 527–539.

de Rosnay, M., Pons, F., Harris, P. L., & Morrell, J. M.B. (2004). A lag between understanding false beliefand emotion attribution in young children: Relation-ships with linguistic ability and mothers’ mental-statelanguage. British Journal of Developmental Psychol-ogy, 22, 197–218.


Feinman, S., & Lewis, M. (1983). Social referencing atten months: A second-order effect on infants’ re-sponses to strangers. Child Development, 54, 878–887.

Gates, G. S. (1923). An experimental study of thegrowth of social perception. Journal of EducationalPsychology, 14, 449–461.

Gopnik, A., & Wellman, H. M. (1994). The theory the-ory. In L. A. Hirschfeld & S. A. Gelman (Eds.), Map-ping the mind: Domain specificity in cognition andculture (pp. 257–293). New York: Cambridge Uni-versity Press.


Gosselin, P., & Simard, J. (1999). Children’s knowledgeof facial expressions of emotions: Distinguishing fearand surprise. Journal of Genetic Psychology, 160,181–193.

Harris, P. L. (1989). Children and emotion: The devel-opment of psychological understanding. Oxford:Blackwell.

Harris, P. L. (1992). From simulation to folk psychol-ogy: The case for development. Mind and Language,7, 120–144.

Harris, P. L. (2000). The work of the imagination: Un-derstanding children’s worlds. Malden, MA:Blackwell.

Harris, P. L., Johnson, C. N., Hutton, D., Andrews, G.,& Cooke, T. (1989). Young children’s theory of mindand emotion. Cognition and Emotion, 3, 379–400.

Harris, P. L., Olthof, T., Meerum Terwogt, M., &Hardman, C. E. (1987). Children’s knowledge of thesituations that provoke emotion. International Jour-nal of Behavioral Development, 10, 319–343.

Haviland, J. M., & Lelwica, M. (1987). The induced af-fect response: 10-week-old infants’ responses to threeemotion expressions. Developmental Psychology, 23,97–104.

Huttenlocher, J., & Smiley, P. (1990). Emerging notionsof persons. In N. L. Stein, B. Leventhal, & T.Trabasso (Eds.), Psychological and biological ap-proaches to emotion (pp. 283–295). Hillsdale, NJ:Erlbaum.

Izard, C. E. (1979). The Maximally Discriminative Fa-cial Movement Coding System (MAX). Newark: Uni-versity of Delaware, Instructional Resources Center.

Izard, C. E. (1994). Innate and universal facial expres-sions: Evidence from developmental and cross-cultural research. Psychological Bulletin, 2, 288–299.

Kahana-Kalman, R., & Walker-Andrews, A. S. (2001).The role of person familiarity in young infants’ per-ception of emotional expressions. Child Develop-ment, 72, 352–369.

Klinnert, M. D., Emde, R. N., Butterfield, P., & Cam-pos, J. J. (1986). Social referencing: The infant’s useof emotional signals from a friendly adult withmother present. Developmental Psychology, 22,427–432.

Markham, R., & Adams, K. (1992). The effect of typeof task on children’s identification of facial expres-sions. Journal of Nonverbal Behavior, 16, 21–39.

Markman, E. M. (1989). Categorization and naming inchildren: Problems of induction. Cambridge, MA:MIT Press.

Moses, L. J., Baldwin, D. A., Rosicky, J. G., & Tidball,G. (2001). Evidence for referential understanding inthe emotions domain at twelve and eighteen months.Child Development, 72, 718–735.

Mumme, D. L., & Fernald, A. (2003). The infant as on-looker: Learning from emotional reactions observedin a television scenario. Child Development, 74,221–237.

Nelson, N. L., Widen, S. C., & Russell, J. A. (2006).Children’s understanding of emotions’ causes andconsequences: Labeling facial expression and stories.Poster presented at the 18th Annual Convention ofthe American Psychological Society, New York.


Owren, M. J., Rendall, D., & Bachorowski, J.-A.(2003). Nonlinguistic vocal communication. In D.Maestripieri (Ed.), Primate psychology (pp. 359–394). Cambridge, MA: Harvard University Press.

Phillips, A. T., Wellman, H. M., & Spelke, E. S. (2002).Infants’ ability to connect gaze and emotional expres-sion to intentional action. Cognition, 85, 53–78.

Pons, F., Harris, P. L., & de Rosnay, M. (2004). Emotioncomprehension between 3 and 11 years: Develop-mental periods and hierarchical organization. Euro-pean Journal of Developmental Psychology, 1, 127–152.

Repacholi, B. M., & Gopnik, A. (1997). Early reasoningabout desires: Evidence from 14- and 18-month-olds.Developmental Psychology, 33, 12–21.

Ridgeway, D., Waters, E., & Kuczaj, S. A., II. (1985).Acquisition of emotion-descriptive language: Recep-tive and productive vocabulary norms for ages 18months to 6 years. Developmental Psychology, 21,901–908.

Rieffe, C., Meerum Terwogt, M., & Cowan, R. (2005).Children’s understanding of mental states as causesof emotions. Infant and Child Development, 14,259–272.

Russell, J. A. (1990). The preschooler’s understandingof the causes and consequences of emotion. ChildDevelopment, 61, 1872–1881.

Russell, J. A., & Bullock, M. (1986a). Fuzzy conceptsand the perception of emotion in facial expressions.Social Cognition, 4, 309–34l.

Russell, J. A., & Bullock, M. (1986b). On the dimen-sions preschoolers use to interpret facial expressionsof emotion. Developmental Psychology, 22, 97–l02.

Russell, J. A., & Fehr, B. (1994). Fuzzy concepts in afuzzy hierarchy: Varieties of anger. Journal of Person-ality and Social Psychology, 67, 186–205.

Russell, J. A., Lewicka, M., & Niit, T. (1989). A cross-cultural study of a circumplex model of affect. Jour-nal of Personality and Social Psychology, 57, 848–856.

Russell, J. A., & Paris, F. A. (1994). Do children acquireconcepts for complex emotions abruptly? Interna-tional Journal of Behavioral Development, 17, 349–365.

Russell, J. A., & Widen, S. C. (2002a). A label superior-ity effect in children’s categorization of facial expres-sions. Social Development, 11, 30–52.

Russell, J. A., & Widen, S. C. (2002b). Words versusfaces in evoking children’s knowledge of the causes ofemotions. International Journal of Behavioral Devel-opment, 26, 97–103.


Saarni, C. (1999). The development of emotional com-petence. New York: Guilford Press.

Soken, N. H., & Pick, A. D. (1999). Infants’ perceptionof dynamic affective expressions: Do infants distin-guish specific expressions? Child Development, 70,1275–1282.

Sorce, J. F., Emde, R. N., Campos, J. J., & Klinnert, M.D. (1985). Maternal emotional signaling: Its effect onthe visual cliff behavior of 1-year-olds. Developmen-tal Psychology, 21, 195–200.

Stein, N. L., & Levine, L. J. (1989). The causal organi-sation of emotional knowledge: A developmentalstudy. Cognition and Emotion, 3, 343–378.

Stifter, C. A., & Fox, N. A. (1987). Preschool children’sability to identify and label emotions. Journal ofNonverbal Behavior, 11, 43–54.

Trabasso, T., Stein, N. L., & Johnson, L. R. (1981).Children’s knowledge of events: A causal analysis ofstory structure. In G. H. Bower (Ed.), The psychol-ogy of learning and motivation (Vol. 15, pp. 237–282). New York: Academic Press.

Walden, T. A., & Kim, G. (2005). Infants’ social lookingtoward mothers and strangers. International Journalof Behavioral Development, 29, 356–360.

Walker-Andrews, A. S. (1986). Intermodal perception ofexpressive behaviors: Relation of eye and voice? De-velopmental Psychology, 22, 373–377.

Walker-Andrews, A. S. (1997). Infants’ perception ofexpressive behaviors: Differentiation of multimodalinformation. Psychological Bulletin, 121, 437–456.

Wellman, H. M. (1995). Young children’s conception ofmind and emotion: Evidence from English speakers.In J. A. Russell, J. M. Fernández-Dols, A. S. R.Manstead, & J. C. Wellenkamp (Eds.), NATO ASISeries D: Behavioural and social sciences. Vol. 81.Everyday conceptions of emotion: An introductionto the psychology, anthropology and linguistics ofemotion (pp. 289–313). New York: Kluwer Aca-demic.

Wellman, H. M., Harris, P. L., Banerjee, M., & Sinclair,A. (1995). Early understanding of emotion: Evidence

from natural language. Cognition and Emotion, 9,117–149.

Wellman, H. M., & Woolley, J. D. (1990). From simpledesires to ordinary beliefs: The early development ofeveryday psychology. Cognition, 35, 245–275.

Werner, H. (1948). Comparative psychology of mentaldevelopment. Oxford: Follett.

Widen, S. C. (2005). Preschoolers’ use of emotion labelsand understanding of emotion categories. Disserta-tion Abstracts International, 65, 6692.

Widen, S. C., & Russell, J. A. (2002). Gender and pre-schoolers’ perception of emotion. Merrill–PalmerQuarterly, 48, 248–262.

Widen, S. C., & Russell, J. A. (2003). A closer look atpreschoolers’ freely produced labels for facial expres-sions. Developmental Psychology, 39, 114–128.

Widen, S. C., & Russell, J. A. (2004). The relativepower of an emotion’s facial expression, label, andbehavioral consequences to evoke preschoolers’knowledge of its cause. Cognitive Development, 19,111–125.

Widen, S. C., & Russell, J. A. (2006). Free labeling ofemotions by 2- and 3-year-olds: A face superiority ef-fect. Manuscript in preparation.

Widen, S. C., & Russell, J. A. (2007). Children’s andadults’ understanding of the “disgust face.” Cogni-tion and Emotion.

Widen, S. C., & Russell, J. A. (in press). Evidence on thenature of children’s understanding of emotion: Thecase of anger and disgust. Manuscript submitted forpublication.

Wintre, M. G., & Vallance, D. D. (1994). A develop-mental sequence in the comprehension of emotions:Intensity, multiple emotions, and valence. Develop-mental Psychology, 30, 509–514.

Witherington, D. C., Campos, J. J., & Hertenstein, M. J.(2001). Principles of emotion and its development ininfancy. In G. Bremner & A. Fogel (Eds.), Hand-books of developmental psychology: Blackwell hand-book of infant development (pp. 427–464). Malden,MA: Blackwell.


C H A P T E R 2 2

Intermodal Emotional Processesin Infancy

ARLENE S. WALKER-ANDREWS

Very early in life, human infants interact withtheir caregivers in what is called “proto-conversation” (Rochat, Quierdo, & Striano,1999; Trevarthen, 1979). In these social inter-actions, an infant and adult gaze, touch, smile,and vocalize toward one another in turn-takingsequences (Stern, Jaffe, Beebe, & Bennett,1975). Some authors have likened thiscoregulated activity to a dance with a sense ofshared rhythm and movement (e.g., Schore,1994; Stern, 1985). Contingency is importantto these interactions (Bigelow & Rochat,2006), but perhaps more critical is that themembers of the dyad are sharing affect.1 Adultand infant do not merely mimic one another,nor do they respond randomly. Often they ex-press the same affect but in different modali-ties; for example, the adult smiles, and the in-fant chortles in response. The interaction ismultimodal. These temporally coordinated in-teractions between infants and adult caregivershave been described extensively (Trevarthen,1993; Tronick, 1989). Furthermore, somestudies have identified patterns indicating thatthe movements and vocalizations of infants are

contingent with the temporal characteristics ofthe adults’ actions (Jaffe, Beebe, Feldstein,Crown, & Jasnow, 2001; Stern, Hofer, Haft, &Dore, 1985). This process of mutual attune-ment “reflects the role of emotion in communi-cation. It can be used to share feelings with an-other, to empathize, to mock, to respondcontingently, to change the other’s arousal levelor emotion, to change the other’s goal, to teach,or to play” (Fogel, 2000, p. 132). To summa-rize, infants are exposed to rich, multimodal af-fective displays in the first few months of life,and they respond reciprocally to that informa-tion. The experimental results to be summa-rized in this chapter support the premise thatvery young infants recognize the meaning of af-fective expressions, provided that the infantsobserve multimodal expressions in familiarcontexts.

In an experimental study of affect as mani-fested in such social interactions, Haviland andLelwica (1987) found that by 2½ months, in-fants are tiny experts in affective turn taking andsharing. Haviland and Lelwica videotapedmother–infant interactions in which mothers

364

acted out three facial and vocal expressions, asassigned, for 20 seconds each. Infants respondeddifferently to happy, angry, and sad facial–vocalexpressions, sometimes matching or imitating.For example, to a maternal portrayal of sadness,an infant was likely to look down, show a soberface, and produce more mouthing. These re-sponses were coordinated temporally with themothers’ affective displays, as indicated by eventanalyses. These authors concluded:

First, by 10 weeks of age, infants respond differ-ently to three maternal affect expressions whenthe presentation is simultaneously facial and vo-cal—joy, anger, and sadness. Second, these infantscan match or mirror joy and anger expressions.Third, the infants’ matching responses to the ma-ternal affects are only part of complex but predict-able behavioral patterns that seem to indicatemeaningful affect states and possibly self-regulation in the infants. (Haviland & Lelwica,1987, p. 102)

More recently, in another study of infants’responsiveness to live affective expressions, we(Montague & Walker-Andrews, 2001) con-ducted a study in which 4-month-olds werepresented with affective expressions embeddedin the familiar game of peek-a-boo. Peek-a-booprovides a rich natural context characterizedby substantial multimodal information andspatiotemporal structuring, and in which theexcitement builds on the expectation of thedisappearance–reappearance of a face and theaccompanying vocalizations. We asked a fe-male experimenter unfamiliar to an infant toact out the typical happy–surprised facial–vocal expression used in the game as she hidher face and then reappeared with a vocalized“peek-a-boo.” On designated reappearance tri-als, however, she portrayed an expression thatwould be unanticipated in the standard game(a sad, angry, or fearful expression). Infants re-acted differentially to the alteration of the typi-cal peek-a-boo interaction. Infants in the angryand fearful conditions increased their lookingtime to these expressions, while those infantswho viewed either typical happy–surprised orsad expressions on the designated trials de-creased their looking time over the sequence. Inaddition, infants assigned to the angry condi-tion increased their looking time on all trialssubsequent to the first angry expression, as ifthe appearance of anger increased their overalllevel of arousal. Analysis of the infants’ facialexpressiveness demonstrated that infants in the

sad condition were more labile; they lookedaway more and they responded with more fa-cial movements, as well as more expressionsidentifiable as interest (Izard, 1995). Insummary, 2½-month-old infants who werepresented with multimodal expressions by afamiliar person (Haviland & Lelwica, 1987)and 4-month-olds presented with multimodalexpressions by a stranger in a familiar setting(Montague & Walker-Andrews, 2001) discrim-inated the expressions. In addition, they re-sponded differentially to those expressions inboth contexts. These results show that younginfants can discriminate and respond appropri-ately even during posed interactions, as long asaffective information is multimodal and is pre-sented in familiar contexts.

INFANTS’ PERCEPTION OF EMOTION

That infants are successful participants inprotoconversations suggests that the percep-tion of affect occurs early in life and involvesall of the senses. Despite this possibility, thestudy of infants’ perception of affect in con-trolled experiments in the laboratory startedwith a focus on whether infants simply couldtell the difference between posed facial expres-sions. As with much early infant research, theassumption was that one must simplify thestimuli for infants so as not to overwhelm theirsenses. Infants demonstrated an ability todiscriminate such stimuli. For example,LaBarbera, Izard, Vietze, and Parisi (1976) pre-sented infants with black-and-white slides ofthree facial expressions one at a time. Theyfound that 4- and 6-month-olds looked longerat the joy expression than at anger and neutralexpressions, but that there was no difference inlooking time to the anger and neutral expres-sions. Similarly, Young-Browne, Rosenfeld,and Horowitz (1977) obtained some evidenceof discrimination of facial expressions by 3months. They reported that infants increasedlooking time to a slide of a surprised face afterviewing a happy face and to a happy face afterviewing a surprised face. Infants also increasedtheir looking time when a surprised face fol-lowed a sad face, but not vice versa. No incre-ments in looking time were found when happyand sad were contrasted in either order. Never-theless, both studies showed that infants coulddiscriminate photographed, static facial ex-pressions.

22. Intermodal Emotional Processes in Infancy 365

Nelson, Morse, and Leavitt (1979) posed amore critical question: Can infants move be-yond discrimination and actually recognize af-fective expressions? They argued that if infantscould detect common information for affectacross several photographed facial expressionexemplars and generalize that information to anew exemplar, evidence of recognition of affectwould be obtained. Seven-month-old infantsviewed photographed facial expressions, but tointroduce variability, several different personsposed for each photograph used during the fa-miliarization interval. Infants increased theirlooking time to fearful expressions after view-ing a set of happy expressions, although theydid not look longer to a change in expression ifthese sets were presented in the reverse order.Nelson and Dolgin (1985) and others (e.g.,Serrano, Iglesias, & Loeches, 1992) reportedsimilar patterns with even younger infants (3months), and concluded that young infants aresophisticated perceivers of facial affect. Reex-amination of these results, however, suggestedthat infants might have been responding to fea-ture differences between facial expressionsrather than to affect. To test this alternative in-terpretation, Caron, Caron, and Myers (1985)varied the “toothiness” of various exemplarsof angry and happy facial expressions. Theyfound that infants grouped the expressions bythe visibility of molars rather than the emotiondisplayed. In general, the early research on in-fants’ recognition of affect indicated that in-fants relied primarily on feature information,unless affect was made much more salient (cf.Ludemann & Nelson, 1988). These findingsappear to contradict data from mother–infantinteraction studies suggesting that infants arequite accomplished at affect perception.

MULTIMODAL, DYNAMICPRESENTATIONS

A major difference between the two types ofstudies summarized above—laboratory exami-nations of the visual perception of photo-graphed facial expressions, and naturalistic ob-servations of mother–infant interactions—canbe found in the emotion exemplars presented.Infants appear more sophisticated when dy-namic, live stimulus displays are used in sup-portive contexts. Although the cited laboratorystudies provided stimulus control, they eradi-cated important information for affect. In the

natural world, a young infant usually observesunified expressive behaviors. That is, the infanttypically views a person’s face and gestures,hears that person’s voice, and is touched bythat person during an interaction. Rarely is aninfant presented with only a voice or only aface and gestures when another person claimshis or her attention. Infants do hear voices em-anating from adjacent rooms on occasion, butthese vocalizations are usually directed atsomeone else. Indeed, it is quite possible that avery young infant requires all of the informa-tion available multimodally for accurate dis-crimination and comprehension of expressionsto develop. As Charlesworth and Kreutzer(1973) suggested, infants may need the “totalexpressive behavior (facial, vocal, verbal, andpostural) of those around them before they candiscriminate such behavior from other behav-iors (e.g., instrumental behaviors such as chew-ing, drinking, reaching) and recognize theirmeaning” (p. 125). Facial expressions, whichcarry information about intentions, are a sub-set of facial actions that also include eye move-ments and instrumental behaviors such as eat-ing, speaking, and sneezing. The infant mustabstract the affectively relevant informationfrom such facial actions.

The specific information available in a faceor a voice and gestures is not precisely equiva-lent, but rather invariant, across behaviors. Forexample, the location in space of the voice andface are identical, and location is specified byboth visual and auditory information. Further-more, acoustic characteristics of the voice (e.g.,volume, variability in pitch, rhythm, tempo)may be related to aspects of facial movementsand gestures (e.g., extent, fluidity, tempo). Themeaning of a behavior that is expressed orallyis also expressed in behaviors that can be seen.With respect to affect, to perceive the unity ofan expressive behavior, an infant must detectinformation that is invariant over more thanone modality. The infant must perceive oneevent, specified multimodally, if that event is tohave meaning. Evidence from studies con-ducted over the past 30 years suggests that thedevelopment of affect perception moves fromdetection of intermodal correspondences dur-ing multimodal interactions to the ability torecognize affect in almost any expression, in-cluding the information from gestures, briefvocal signals, or information preserved in astatic drawing. For example, an adult observercan detect anger at a distance, as when witness-


ing someone slamming a fist on a surface. Thegesture alone conveys the meaning to an adult.

Evidence for the early emergence of inter-modal perception and sensitivity to amodal in-formation for a property such as texture(which can be seen or felt) or affect (which canbe seen or heard or felt) has grown dramati-cally over the past few decades. “There is com-pelling neural, electrophysiological, and behav-ior evidence of strong intermodal linkages innewborns and young of a variety of avianand mammalian species, including humans”(Lickliter & Bahrick, 2000, p. 644). Young in-fants are sensitive to correspondences betweenauditory, visual, and tactual stimulation (Gib-son & Walker, 1984; Lewkowicz, 1992, 1996;Meltzoff & Borton, 1979; Pickens, 1994;Walker-Andrews & Lennon, 1985); they alsoshow the effects of intersensory facilitation, inwhich stimulation in one modality enhances re-sponsiveness to stimulation in other modalities(Bahrick, Flom, & Lickliter, 2002; Lewkowicz& Turkewitz, 1981; Lickliter & Stoumbos,1991), across a range of situations.

Why multimodal presentations are more ef-fective (especially during early development)remains an open question, however. Drawingfrom the theories of E. J. Gibson (1969) and J.J. Gibson (1979), I have proposed that detec-tion of invariant relations underlies perceptualdevelopment and learning, including in the per-ception of emotion (Walker-Andrews, 1994).That is, the world itself is multimodal, with re-dundant information specifying objects andevents. In the world of affect, shared meaningis carried by face, voice, gestures, posture, andtouch. In some instances the correspondencesare physical (such as synchrony between visiblelip movements and onset of sounds), and inothers they are more abstract. Infants appear toexperience a world of perceptual unity basedon the detection of amodal information and/or intermodal correspondences that specifyproperties of objects and events and their“affordances”:

The infant does not begin by first discriminatingthe qualities of objects and then learning the com-binations of qualities that specify them. Phenome-nal objects are not built up of qualities; it is theother way around. The affordance of an object iswhat the infant begins by noticing. The meaning isobserved before the substance and surface, thecolor and form, are seen as such. An affordance isan invariant combination of variables, and onemight guess that it is easiest to perceive such an in-

variant unit than it is to perceive all the variablesseparately. (J. J. Gibson, 1979, pp. 134–135)

With experience, infants become sensitive tonew invariant relationships and come to detectadditional properties of objects and events, in-cluding social-emotional interactions. Learningitself results from changes in perceptual selec-tion, leading to progressive differentiation ofthe environment. In essence, an infant perceivesthe meaning of an event prior to the separatephysical properties of that event. The inter-modal correspondences marking the visual andauditory information for a single affective ex-pression allow or compel the infant to attend tothat expression and explore it fully, even in thepresence of other, competing events. When aninfant observes someone laughing and speak-ing in an animated fashion, the shared proper-ties are highlighted, while other information(such as the person’s hair color or height)becomes background (Bahrick, Walker, &Neisser, 1981).

In summary, during multimodal stimulation,processing and learning of amodal propertiesis facilitated. Therefore, affective information,which is presented across modalities, should beeasiest to detect when it is highlighted in multi-modal presentations. Differentiation of therhythm, animation, timing, and intensity of anaffective expression is supported when the in-fant observes a speaking, gesturing, and frown-ing or smiling person during an interactivebout. Parten and Marler (1999) introduced aclassification scheme for categorizing and com-paring how the combination of unimodal sig-nals detected in a multimodal communicativeinteraction is perceived across a wide range ofspecies at all ages. They suggest that redun-dancy in meaning across the components of amultimodal signal may have one of several ef-fects. The perceiver may respond to the redun-dant multimodal signals as if to any single com-ponent. Alternatively, the response may beenhanced. When the signals are nonredundant,there may be specific interactions: Each com-ponent may act independently, one may domi-nate or modulate another, or an entirely newresponse may emerge. An analogous examplewithin vision only would be the presence of asmile in conjuction with fully crinkled eyes,compared to the smile unaccompanied by eyecontractions.

Neurophysiological evidence suggests thatthe brain is set up for multimodal processing of


the types of relationships Parten and Marlerdescribe. For example, there are sites of multi-sensory convergence at the cortical level in cats(Wallace, Meredith, & Stein, 1992), monkeys(Mistlin & Perrett, 1990), and humans(Calvert, 2001), and many brain areas respondto multisensory inputs. For human infants, theperception of facial and vocal affective infor-mation may lead to greater responsiveness thanthe observation of either facial expressions orvocal expressions separately—a multiplicativeor enhanced effect such as that found duringneural processing of simultaneous visual andauditory stimuli in the superior colliculus ofcats (Stein, Wallace, & Meredith, 1995).Finally, recent results suggest that humans andothers experience an emotion upon viewing an-other’s emotional response (Wicker et al.,2003).

INFANTS’ PERCEPTIONOF FACIAL–VOCAL EXPRESSIONS

Studies of infants’ perception of affective ex-pressions have borne out the proposal that in-fants are more sensitive to affective informa-tion when it is presented multimodally. Forexample, Caron, Caron, and MacLean (1988)showed infants videotapes of six women whoacted out sad, happy, or angry facial–vocal ex-pressions continuously until the infants lost in-terest in the specific expression (i.e., habitu-ated). At that point, Caron et al. showed theinfants a new set of videotaped expressions.The infants responded differently to thischange, depending on their age and which ex-pression they observed initially. Infants asyoung as 4 months of age could tell dynamichappy and sad expressions apart, and coulddistinguish angry and happy expressions by 7months, as long as these expressions were pre-sented in specific orders. That is, 4-month-oldsdiscriminated happy and sad expressions whenhappy ones were presented first, and at 5months when either order was used. Similarly,7-month-olds discriminated angry and happyfacial–vocal expressions when happy ones oc-curred first, but not without the presence of avoice. In each case, the presence of both faceand voice was critical, as was the dynamicquality of the expressions. Infants failed to dis-criminate happy and angry facial expressionsdepicted in photographs at 9 months in an ear-lier study (Caron et al., 1985). More re-cently, my colleagues and I (Walker-Andrews,

Mayhew, Coffield, & Krogh-Jespersen, 2004)found that infants as young as 3½ monthscould recognize happy and sad expressions pre-sented by their own parents. Infants were ha-bituated to the face and voice of each parentacting out “happy” or “sad” in alternation. Attesting, infants responded to a change in emo-tion, person, or emotion and person in eitherorder. A second group of infants viewed thesame videotaped expressions, but failed toshow recognition of expressions presented by(to them) unfamiliar males and females. Theydishabituated (i.e., regained visual interest)only to a change in both expression and per-son.

To determine how important it is that faceand voice be presented together, we (Walker-Andrews & Lennon, 1991) studied the processby which 5-month-olds might come to discrim-inate happy and sad vocal expressions. Weshowed 5-month-olds a happy or angry vocalexpression accompanied by an affectivelymatching, but static, facial expression; a facialexpression depicting a different emotion; ora checkerboard until the infants’ attentionwaned. Infants in this study increased theirlooking time to any change in vocal expression,except when a checkerboard had accompaniedthe vocal expression during the habituationphase. We speculated that the results reflectedhow infants might discover the differentaffordances of emotional expressions in theirnatural environment. This discovery would bemade in the context of a communicative inter-action where an infant is able to look at anadult’s face while listening to the adult’s voice.In this view, the face acts as a setting for attend-ing to the affective quality of the voice.

Flom and Bahrick (2007) recently focusedon the developmental course of infants’ detec-tion of affective information in bimodal(auditory–visual) and unimodal (auditory- orvisual-only) displays. They presented infants of3, 4, 5, and 7 months with dynamic color videofilms of an adult female portraying happy, an-gry, or sad affective expressions for habitua-tion. Across experiments, infants were habitu-ated to bimodal or unimodal presentations. Attesting, they were presented a change in thehabituated emotion expression. Flom andBahrick found that infants discriminated affectby 4 months if bimodal displays were pre-sented. For unimodal displays, affect discrimi-nation did not appear until 5 months forauditory-only displays, and at 7 months forvisual-only displays. Temporal synchrony was


also necessary for the youngest infants’ dis-crimination; 3-month-olds did not respond to achange in affect when the face and voice weretemporally disaligned, although 4-month-oldswere successful in this task.

Infants also show intermodal matching ofaffect expressions. In a series of studies, I(Walker, 1982; Walker-Andrews, 1986, 1988)showed infants ranging in age from 2 to 7months pairs of facial expressions accom-panied by a single vocal expression thataffectively matched one of the facial expres-sions. For example, one infant saw a happy andan angry facial expression side by side while lis-tening to an angry vocal expression for 2 min-utes, followed by a second presentation inwhich the happy vocal expression was played.In such cases, 2-month-olds looked almost ex-clusively at the happy expression, regardlessof which vocal expression was placed. Four-month-olds increased looking time to thehappy facial expression when it was sound-matched, and 5- and 7-month-olds increasedfixation to any of the facial expressions thatwere sound-matched. I (Walker, 1982) alsoconducted two experiments that focused on therole of temporal synchrony in intermodalmatching of affective expressions. In one ex-periment, happy and angry facial expressionswere presented upside down with a temporallysynchronized vocal expression that affectivelymatched one of the faces. Infants failed to showa looking preference in this situation, althoughsynchrony was present. In another experiment,temporal synchrony was disrupted, by delayingthe voice by 2 seconds. In this case, infantsfailed to show visual preferences on the firsttrial, but on the second trial, the preference forthe sound-matched facial expression emergedand increased steadily over the trial. In a finalstudy (Walker-Andrews, 1986), 5- and 7-month-olds observed facial expressions inwhich the lower part of the face was blockedoff, so that synchrony between lip movementsand vocalizations was not visible. Older infantslooked preferentially to the sound-matched fa-cial expression, although 5-month-olds failedto do so.

These results, in combination with those ofFlom and Bahrick (2007) and Caron et al.(1988), suggest that during the first year in-fants are developing the ability to detect com-mon affect across bimodal, dynamic presenta-tions of an emotional expression. Infantsappear to detect affective information first inbimodal (face and voice) expressions, and they

seem to recognize happy expressions earliest. Itis as if the redundancy between the sight andsound guides infants’ perceptual selectivity andpermits them to attend to the affective intent(see Bahrick et al., 1981; Bahrick, Lickliter, &Flom, 2004). The presence of temporal syn-chrony between face and voice contributes toperception in younger infants, but is neitherimperative nor sufficient for matching to occur.For example, Soken and Pick (1992) demon-strated that 7-month-olds can match happyand angry facial expressions to their vocalcounterparts based solely on motion informa-tion. They presented infants with point-light vi-sual displays, created by placing luminescentdots on a blackened face, thereby eliminatingfeature information while preserving the mo-tion information for affect. These displays wereaccompanied by affectively matching vocal ex-pressions. Soken and Pick found that infantsdetected correspondences between facial andvocal expressions based on affective meaning,even when each component was produced by adifferent person.

As indicated earlier, when infants observeexpressions presented by familiar persons suchas their mothers, they show greater sensitivity.We (Kahana-Kalman & Walker-Andrews,2001) found that infants could match happyand sad facial and vocal expressions that werepresented by their own mothers. When 3-month-olds observed maternal affective ex-pressions, they increased their looking time tothe sound-specified film. Even when the faceand voice were presented out of synchrony, 3-month-olds showed the matching effect. In-fants failed to show intermodal matching whenthe expressions were portrayed by an unfamil-iar woman, even when the vocal and facial dis-plays were synchronized temporally. Infants’affective responsiveness also differed acrossemotions and according to the familiarity ofthe actress. When happiness was presented fa-cially and vocally, infants were more expres-sive: They showed greater variability of affec-tive expression and increased the number ofexpressions they showed. Judges rated infantsas experiencing more positive affect and asmore interested and engaged, especially wheninfants were viewing their mothers. Affect cod-ing demonstrated that infants who viewed theirmothers spent more time smiling at the happysound-specified films than did infants viewingan unfamiliar woman. There were no differ-ences in smiling duration when sadness was thesound-specified emotion. Infants produced


more full and bright smiles when happinesswas sound-specified, particularly for maternalhappy expressions. Finally, the mean length ofdistress was protracted for infants who ob-served unfamiliar women, compared to thosewho observed their mothers. It is possible thatthe familiar context (mother) supports infants’perceptual skills, acting as a scaffold fromwhich to explore other facets of the world.That infants respond affectively to their moth-ers’ expressions underscores the conclusionthat they detect affective meaning in those ex-pressions.

We (Montague & Walker-Andrews, 2002)investigated infants’ intermodal matching, us-ing happy and angry as well as happy and sadexpressions posed by both mothers and fathers.The participating 3-month-olds demonstratedintermodal matching for their mothers’ expres-sions, but they did not make intermodalmatches for expressions presented by unfamil-iar males and females. For fathers, the patternwas more complex. Overall, the infants did notshow intermodal matching for paternal expres-sions, but infants whose fathers engaged in40% or more of routine caregiving activitiesshowed the same intermodal matching formothers and fathers. An ongoing longitudinalstudy (Montague, Kahana-Kalman, Goldman,Long, & Walker-Andrews, 2005) with infantsat 3, 4, 5, and 7 months is aimed at document-ing the developmental trajectory of suchintermodal matching and expressiveness. At 4months, infants smile more to maternal happyexpressions (facial–vocal expressions with thevoice presented out of synchrony) and showmore lability to maternal sad expressions. Nodifferences have been found in infants’ re-sponses to unfamiliar adults’ expressions atthis age.

Even older infants are more sensitive to themeanings expressed in multimodal portrayals.Vocal expressions alone, and vocal expressionscombined with facial expressions, are more in-fluential in regulating infants’ behavior insocial-referencing studies (Barrett, Campos, &Emde, 1996; Mumme, Fernald, & Herrera,1996). Vaish and Striano (2004) examined therole of facial versus vocal information on in-fants’ responses in a potentially threatening sit-uation. Mothers acted out positive facial-only,vocal-only, or facial–vocal expressions towardtheir 12-month-olds on a visual cliff apparatus.Infants crossed the cliff more quickly in thefacial–vocal and vocal-only conditions, and

looked more often to their mothers in thefacial–vocal condition.

INFANTS’ RESPONSESTO TACTUAL INFORMATION

Most studies that use multimodal expressionsof emotion have used auditory–visual presenta-tions. However, a few investigators have exam-ined the effects of tactual information. For ex-ample, it has been established that contact(tactual stimulation) acts to soothe infants dur-ing still-face experiments, in which an infantbecomes distressed when a caregiver becomesunresponsive facially and vocally (Gusella,Muir, & Tronick, 1988; Stack & Muir, 1990,1992). Infants show more smiling, vocalizing,and gazing when touch is used in a still-face ep-isode than when no touch is provided. Stack(2001) reported that mothers can modify theirinfants’ behavior (e.g., elicit smiles) when themothers are asked to do so by stroking or oth-erwise touching the infants. Recently, Moreno,Posada, and Goldyn (2006) explicitly exam-ined the role of touch in the interactions of 3½-month-olds and their mothers. Mothers wereasked to play with their infants in a face-to-faceinteraction for 3 minutes, once with touch andonce without touch. Moreno et al. found, as inmost studies investigating the perception of af-fect, that context is extremely important to af-fect recognition and regulation. Symmetry be-tween mother and infant was influenceddifferentially, depending on the type of touch(playful vs. affectionate) and an infant’s as-signed condition. The authors concluded thatdyads “make constant ‘online’ adjustments totheir contribution of the interaction based ontheir own internal state as well as feedbackthey receive” (p. 16).

In summary, experimental studies suggestthat infants can recognize emotional expres-sions early in development when affective in-formation is presented multimodally. A num-ber of factors influence the facility with whichinfants respond to that affective information,including the child’s stage of development, thespecific emotion, the familiarity of the personenacting an affective expression, and the con-text in which the expression is encountered.Pinning down the various factors and explicat-ing mechanisms by which they act on infants’perception of affective information continue tobe of interest to researchers.


CHILDREN’S PERCEPTIONOF EMOTION

Although the focus of this chapter is on the de-velopment of affect understanding during in-fancy, it is worth noting that the understandingof emotional expressions also seems influencedby modality in young children. Childrenreadily perceive, recognize, and label emotionsacross a wide range of conditions, and theyshow development in these abilities throughoutchildhood (Denham, 1998). Children are morecapable of recognizing happy than sad expres-sions, and better at labeling both of these thanangry and fearful expressions (Denham, Zoller,& Couchoud, 1994). Shackman and Pollak(2005), who examined learning factors thatmight influence the perception of multimodalaffective expressions among school-age chil-dren conclude that contextual cues such as fa-miliarity enhance children’s ability to extractmeaningful information from affective dis-plays, as is the case with infants, and that the“importance of auditory versus visual perceptsis influenced by the meaning attached to partic-ular emotions contained in each expression andby the familiarity of the individual expressingemotion” (p. 1124).

SINGLE-MODALITY PRESENTATIONS

As alluded to in the review above, infants alsocome to recognize affective information insingle-modality displays. The literature on theperception of facial expressions suggests thatinfants come to detect affective information inthe face sometime in the second half of the firstyear, and in voices sometime before that. As de-scribed previously, Flom and Bahrick (2007)examined this development within a singlestudy, finding that infants appear to recognizeemotion in bimodal presentations first, fol-lowed by unimodal vocal expressions, and fi-nally unimodal facial expressions. This is thesequence that Buhler (1930) and I (Walker-Andrews, 1997) have proposed, but there havebeen other proposals as well. Lewkowicz(1988a, 1988b) advanced a theory of early au-ditory dominance in infants’ perception of so-cial and nonsocial events, which was extendedto the perception of affective expressions byCaron (1988). More recent results, however(e.g., Lewkowicz, 2000; Lewkowicz &Edmondson, 1993), have indicated that for

speech and affective information, the sequenceoriginally proposed by Buhler (1930) is morelikely to be the case. Although infants detectunimodal information that potentially specifiesthe meaning of an affective expression, there isno evidence that very young infants perceivethis unimodal information as informativeabout affect. That is, although a neonate candiscriminate the fundamental frequencies oftwo tones or the pitch of two different voices,or distinguish the upturned curve of a smilefrom the bared teeth of a snarl, the infant doesnot seem to perceive these differences as affec-tive. Rather, the infant seems to rely on the re-dundant information in multimodal displays.Others (e.g., Klinnert, Campos, Sorce, Emde,& Svejda, 1983; Nelson, 1987) have alluded tothis developmental pattern as well. As such, thedevelopmental pattern may be no differentfrom all sorts of perceptual learning in whichinformation may be detected first in intermodaldisplays (Bahrick & Lickliter, 2002; Walker-Andrews, 1994).

EXPERIENCE OF EMOTION

The first part of this chapter has focused on theperception of emotions portrayed by others, aswell as the responsiveness of infants and youngchildren to that affective information. How-ever, the ease with which human infants learnabout emotion may stem from their own multi-modal experience of emotion as well. At a min-imum, the experience of anger at one’s desiresbeing thwarted can result in a plethora of phys-ical and sensory signals: One’s own angry cry isheard, the turmoil in the stomach is felt, thenarrowing of one’s eyes and hard stare are alsoperceived, and the consequences of one’s ac-tions are seen. Faced with an angry person,similar effects may occur. One hears the other’sangry voice, sees the angry face, and feels one’sown shrinking or firmly held stance. Duringprotoconversations, the infant observes thetypical happy facial expression, exaggeratedpitch, and volubility of the happy voice, andmay be tickled or playfully poked by a caregiv-er. The infant responds with a happy smile, en-ergetic movements of his or her limbs, andlaughter.

Stern (1985) has gone beyond this mere de-scription of how one might experience an emo-tion multimodally, to the suggestion that theexperience of emotion, because it is multimod-


al, provides a sense of self. He first talks of themultimodal experience of all kinds: “For in-stance, the actual experience of looking for thefirst time at something that, on the basis ofhow it felt to the touch, should look a certainway, and having it, indeed, look that way issomething like a déjà vu experience” (p. 52).Moreover, what Stern calls “vitality affects”(feelings such as those experienced with thecoming and going of emotions and thoughts)are experienced by the infant from within, aswell as in the behavior of other persons. Recentevidence from Fogel and his colleagues demon-strates that infants have a rich emotional life,as demonstrated by the complexity of theirsmiles and other expressive behaviors in spe-cific interactive contexts. In one study, Fogel etal. (2000) found that infants showed differenttypes of smiles, depending on the context. Thatis, infants as young as 6 months emittedsmiles—simple, Duchenne, play, and duplay—that varied with the game being played (peek-a-boo or tickle), portion of the game, trial, anddirection of an infant’s gaze. The authors con-cluded that when smiling, infants were experi-encing “qualitatively different kinds of enjoy-ment during these two games” (p. 497)—enjoyment of readiness to engage in play, of re-lief, of participation and agency, of escape, andof buildup. The results indicate that the “samefacial action, smiling, can reflect different posi-tive emotions depending upon cooccurring fa-cial actions and the dynamics of the social pro-cess, and that the positive emotional experienceof infants as young as six months is more com-plex than previously reported” (p. 497). Ofparticular importance are (1) the demonstra-tion that smiles are systematically related tocontext and have specific meanings during so-cial communication, and (2) the suggestionthat the smiles themselves are associated withfeelings such as the enjoyment of relief or en-joyment of participation. The physical actionof the tickle game is accompanied by the smilesof relief following the climax. The smile in-dexes a particular emotion. Again, as proposedby Stern (1985), infants have a sense of them-selves as participants in an interaction. Infants’responses during interactions are coordinated,multimodal actions involving vocalizations,gaze, facial expressions, body movements, andgestures, which relay information to the otheras well as to the infants themselves (Malatesta,1981; Weinberg & Tronick, 1994) about inter-nal states, and that regulate the behavior ofboth partners. Affective arousal may mediate

associations among infants’ vocalizations, fa-cial expressions, gaze, postural changes, andother responses (Fogel et al., 1997).

Smith (2005) speaks of the powerful compu-tational system that is at work: “Developing ina social world does not just mean that develop-ment is guided by a mature partner. It also pro-vides an additional level of higher-order multi-modal correlations critical to concepts of self,other, intention, and what is commonly knownin the literature as theory of mind” (p. 292).The interaction of baby and other providesmultimodal, time-locked correlations that pro-vide information about the appearance, ac-tions, and internal states of the self and others.Perhaps one of the most remarkable demon-strations of the integration of information fromthe self and from others comes from the studyof neonatal imitation. Meltzoff and Moore(1977) demonstrated that newborns can imi-tate simple movements of the face, head, andhand modeled by an adult. Meltzoff andMoore labeled this a “supramodal” process,because infants could imitate a seen protrusionof the tongue by an adult with their own (felt)tongue protrusion. Infants also mimic the pho-netic and prosodic structure in their mothers’speech (Bloom, 1988; Kuhl & Meltzoff, 1996;Papousek & Papousek, 1989). Meltzoff and hiscolleagues suggest that information about fa-cial acts is somehow fed into a supramodal rep-resentation that allows for imitation of what ismodeled by another, even though one’s ownperformance cannot be seen. Finally, recentfindings regarding “mirror neurons” (e.g.,Wicker et al., 2003) indicate that humans mayrespond with emotion to the observed emo-tional expressions of others. Whatever themechanism, it is clear that intermodal pro-cesses play a critical role in imitation.

CONCLUSION

In conclusion, intermodal processes are criticalto the development of emotion, including theperception of others’ affective expressions andthe growth of one’s own emotional system. In-fants are quick to detect amodal relations andintermodal correspondences in the fully multi-modal packages presented during social inter-actions. The infant is gifted further with the ex-perience of his or her own affective behavior,which includes facial, vocal, visceral, and kin-esthetic responses. Just as an infant approach-ing the age of 1 is more influenced by vocal and


facial expressions together than by either alone(Vaish & Striano, 2004), so is a young infantreliant on multimodal information for disam-biguating or recognizing others’ affective ex-pressions. The term “resonance” as used by J.J. Gibson (1979) is apt for describing theperception–action coupling that is manifestedin the perception and expression of affect. Theemphasis on intermodal processes captures thefact that emotion is inherently multimodal. In-fant and maternal facial expressions, gestures,and vocalizations are tightly coupled in a socialinteraction (Gogate, Walker-Andrews, &Bahrick, 2001). Moreover, a mother does notsimply react to an infant’s behavior; she alsobuilds on it. Caregivers “create a conversation-like exchange by weaving their own behavioraround the child’s natural activity patterns. Butbabies’ behaviors are both entrained by themother’s pattern and educated by the multi-modal correspondences those interactions cre-ate” (Smith, 2005, p. 294). During the process,infants and caregivers react to, imitate, andscaffold their responses on one another, inways that enrich the interaction and allow theinfants to learn about emotion.

NOTE

1. In this example, the term “affect” is being used to de-scribe more general states that include emotion andmood, and “affective information” the sensory stim-uli specifying affect; “emotion” is being reserved fordiscrete experiences such as joy, anger, and fear.

REFERENCES

Bahrick, L. E., Flom, R., & Lickliter, R. (2002).Intersensory redundancy facilitates discrimination oftempo in 3-month-olds. Developmental Psychology,41, 352–363.

Bahrick, L. E., & Lickliter, R. (2002). Intersensory re-dundancy guides early perceptual and cognitive de-velopment. Advances in Child Development andBehavior, 30, 153–187.

Bahrick, L. E., Lickliter, R., & Flom, R. (2004).Intersensory redundancy guides the development ofselective attention, perception, and cognition in in-fancy. Current Directions in Psychological Science,13(3), 99–102.

Bahrick, L. E., Walker, A. S., & Neisser, U. (1981). Se-lective looking by infants. Cognitive Psychology, 13,377–390.

Barrett, K. C., Campos, J. J., & Emde, R. N. (1996). In-fants’ use of conflicting emotion signals. Cognitionand Emotion, 10, 113–135.

Bigelow, A. E., & Rochat, P. (2006). Two-month-old in-

fants’ sensitivity to social contingency in mother–infant and stranger–infant interaction. Infancy, 9,313–325.

Bloom, K. (1988). Quality of adult vocalizations affectsthe quality of infant vocalizations. Journal of ChildLanguage, 14, 211–227.

Buhler, C. (1930). The first year of life. New York: JohnDay.

Calvert, G. A. (2001). Crossmodal processing in the hu-man brain: Insights from functional neuroimagingstudies. Cerebral Cortex, 11, 1110–1123.

Caron, A. J. (1988). The role of face and voice in infantdiscrimination of naturalistic emotional expressions.Paper presented at the International Conference onInfant Studies, Washington, DC.

Caron, A. J., Caron, R. F., & MacLean, D. J. (1988). In-fant discrimination of naturalistic emotional expres-sions: The role of face and voice. Child Development,59, 604–616.

Caron, R. F., Caron, A. J., & Myers, R. S. (1985). Do in-fants see emotional expressions in static faces? ChildDevelopment, 56, 1552–1560.

Charlesworth, W. R., & Kreutzer, M. A. (1973). Facialexpressions of infants and children. In P. Ekman(Ed.), Darwin and facial expression (pp. 91–167).New York: Academic Press.

Denham, S. A. (1998). Emotion development in youngchildren. New York: Guilford Press.

Denham, S. A., Zoller, D., & Couchoud, E. A. (1994).Socialization of preschooler’s emotion understand-ing. Developmental Psychology, 30, 928–936.

Flom, R., & Bahrick, L. E. (2007). The development ofinfant discrimination of affect in multimodal andunimodal stimulation: The role of intersensory re-dundancy. Developmental Psychology, 43, 238–252.

Fogel, A., Dickson, K. L., Hsu, H., Messinger, D. S.,Nelson-Goens, G. C., & Nwokah, E. E. (1997).Communication of smiling and laughter in mother–infant play: Research on emotion from a dynamicsystems perspective. In K. Barrett (Ed.), New direc-tions for child development: No. 77. The communi-cation of emotions: Current research from diverseperspectives (pp. 5–24). San Francisco: Jossey-Bass.

Gibson, E. J. (1969). Principles of perceptual learningand development. Englewood Cliffs, NJ: PrenticeHall.

Gibson, E. J., & Walker, A. S. (1984). Development ofknowledge of visual–tactual affordances of sub-stance. Child Development, 55, 453–460.

Gibson, J. J. (1979). The ecological approach to visualperception. Hillsdale, NJ: Erlbaum.

Gogate, L. J., Walker-Andrews, A. S., & Bahrick, L. E.(2001). The intersensory origins of word comprehen-sion: An ecological–dynamic systems view. Develop-mental Science, 4(1), 1–18.

Gusella, J. L., Muir, D. W., & Tronick, E. Z. (1988). Theeffect of manipulating maternal behavior during aninteraction on 3- and 6-month-olds’ affect and atten-tion. Child Development, 59, 1111–1124.

Haviland, J. M., & Lelwica, M. (1987). The induced af-fect response: 10-week-old infants’ responses to three


emotion expressions. Developmental Psychology, 23,97–104.

Izard, C. E. (1995). The Maximally Discriminative Fa-cial Movement Coding System (MAX) (rev. ed.).Newark: University of Delaware, Information Tech-nologies and University Media Services.

Jaffe, J., Beebe, B., Feldstein, S., Crown, C., & Jasnow,M. D. (2001). Rhythms of dialogue in infancy.Monographs of the Society for Research in Child De-velopment, 66(2, Serial No. 265).


Klinnert, M., Campos, J. J., Sorce, J., Emde, R. N., &Svejda, M. (1983). Emotions as behavior regulators:Social referencing in infancy. In R. Plutchik & H.Kellerman (Eds.), The emotions: Vol. 2. Emotions inearly development (pp. 57–86). New York: AcademicPress.

Kuhl, P. K., & Meltzoff, A. N. (1996). Infant vocaliza-tions in response to speech: Vocal imitation and de-velopmental change. Journal of the Acoustical Soci-ety of America, 100, 2425–2438.

LaBarbera, J. D., Izard, C. E., Vietze, P., & Parisi, S. A.(1976). Four- and six-month-old infants visual re-sponses to joy, anger, and neutral expressions. ChildDevelopment, 47, 535–538.

Lewkowicz, D. J. (1988a). Sensory dominance in in-fants: 1. Six-month-old infants’ response toauditory–visual compounds. Developmental Psy-chology, 24, 155–171.

Lewkowicz, D. J. (1988b). Sensory dominance in in-fants: 2. Ten-month-old infants’ response toauditory–visual compounds. Developmental Psy-chology, 24, 172–182.

Lewkowicz, D. J. (1992). Infants’ response to tempo-rally based intersensory equivalence: The effect ofsynchronous sounds on visual preferences for movingstimuli. Infant Behavior and Development, 15, 297–324.

Lewkowicz, D. J. (1996). Infants’ response to the audi-ble and visible properties of the human face: I. Roleof lexical/syntactic content, temporal synchrony, gen-der, and manner of speech. Developmental Psychol-ogy, 32, 347–366.

Lewkowicz, D. J. (2000). Infants’ perception of the au-dible, visible, and bimodal attributes of multimodalsyllables. Child Development, 71, 1241–1257.

Lewkowicz, D. J., & Edmondson, D. (1993). Infants’responsiveness to the multimodal properties of thehuman face. Paper presented at the biennial meetingof the Society for Research in Child Development,New Orleans.

Lewkowicz, D. J., & Turkewitz, G. (1981). Intersensoryinteraction in newborns: Modification of visual pref-erences following exposure to sound. Child Develop-ment, 52, 827–832.

Lickliter, R., & Bahrick, L. E. (2000). Perceptual devel-opment and the origins of multisensory responsive-

ness. In G. A. Calvert, C. Spence, & B. E. Stein (Eds.),The handbook of multisensory processes (pp. 643–654). Cambridge, MA: MIT Press.

Lickliter, R., & Stoumbos, J. (1991). Enhanced prenatalauditory experience facilitates species-specific visualresponsiveness in bobwhite quail chicks (Colinusvirginianus). Journal of Comparative Psychology,105, 89–94.

Ludemann, P., & Nelson, C. (1988). Categorical repre-sentation of facial expressions by 7-month-old in-fants. Developmental Psychology, 24, 492–501.

Malatesta, C. A. (1981). Infant emotion and the vocalaffect lexicon. Motivation and Emotion, 5, 1–23.

Meltzoff, A. N., & Borton, R. W. (1979). Intermodalmatching by human neonates. Nature, 282, 403–404.

Meltzoff, A. N., & Moore, M. K. (1977). Imitation offacial and manual gestures by human neonates. Sci-ence, 198, 75–78.

Mistlin, A. J., & Perrett, D. I. (1990). Visual andsomatosensory processing in the macaque temporalcortex: The role of ‘expectation.’ Experimental BrainResearch, 83, 437–450.

Montague, D. P. F., Kahana-Kalman, R., Goldman, S.,Long, L., & Walker-Andrews, A. S. (2005). Infants’recognition of emotion expressions and the role ofparental involvement and emotionality. Poster pre-sented at the biennial meeting of the Society for Re-search in Child Development, Atlanta, GA.

Montague, D. P. F., & Walker-Andrews, A. S. (2001).Peekaboo: A new look at infants’ perception of emo-tion expression. Developmental Psychology, 37,826–838.

Montague, D. P. F., & Walker-Andrews, A. S. (2002).Mothers, fathers, and infants: The role of familiarityand parental involvement in infants’ perception ofemotion expressions. Child Development, 73, 1339–1352.

Moreno, A. J., Posada, G. E., & Goldyn, D. T. (2006).Presence and quality of touch influence coregulationin mother–infant dyads. Infancy, 9, 1–20.

Mumme, D. L., Fernald, A., & Herrera, C. (1996). In-fants’ responses to facial and vocal emotional signalsin a social referencing paradigm. Child Development,67, 3219–3237.

Nelson, C. A. (1987). The recognition of facial expres-sions in the first two years of life: Mechanisms of de-velopment. Child Development, 58, 889–909.

Nelson, C. A., & Dolgin, K. (1985). The generalizeddiscrimination of facial expressions by seven-month-old infants. Child Development, 56, 58–61.

Nelson, C. A., Morse, P. A., & Leavitt, L. A. (1979).Recognition of facial expressions by seven-month-old infants. Child Development, 50, 1239–1242.

Papousek, M., & Papousek, H. (1989). Forms and func-tions of vocal matching in interactions betweenmothers and their precanonical infants. First Lan-guage, 9, 137–158.

Parten, S., & Marler, P. (1999). Communication goesmultimodal. Science, 2823, 1272–1273.


Pickens, J. (1994). Perception of auditory–visual dis-tance relations by 5-month-old infants. Developmen-tal Psychology, 30, 537–544.

Rochat, P., Quierdo, J. G., & Striano, T. (1999).Emerging sensitivity of the timing and structure ofprotoconversation in early infancy. DevelopmentalPsychology, 35, 950–957.

Schore, A. (1994). Affect regulation and the origins ofthe self: The neurobiology of emotional develop-ment. Hillsdale, NJ: Erlbaum.

Serrano, J. M., Iglesias, J., & Loeches, A. (1992). Visualdiscrimination and recognition of facial expressionsof anger, fear and surprise in four- to six-month-oldinfants. Developmental Psychobiology, 25, 411–425.

Shackman, J. E., & Pollak, S. D. (2005). Experiential in-fluences on multimodal perception of emotion. ChildDevelopment, 76, 1116–1126.

Smith, L. B. (2005). Cognition as a dynamic system:Principles from embodiment. Developmental Review,25, 278–298.

Soken, N. H., & Pick, A. D. (1992). Intermodal percep-tion of happy and angry expressive behaviors byseven-month-old infants. Child Development, 63,787–793.

Stack, D. M. (2001). The salience of touch and physicalcontact during infancy: Unraveling some of the mys-teries of the somesthetic sense. In G. Bremner & A.Fogel (Eds.), Handbooks of developmental psychol-ogy: Blackwell handbook of infant development(pp. 351–378). Malden, MA: Blackwell.

Stack, D. M., & Muir, D. W. (1990). Tactile stimulationas a component of social interchange: New interpre-tations for the still-face effect. British Journal of De-velopmental Psychology, 8, 131–145.

Stack, D. M., & Muir, D. W. (1992). Adult tactile stimu-lation during face-to-face interactions modulatesfive-month-olds’ affect and attention. Child Develop-ment, 63, 1509–1525.

Stein, B. E., Wallace, M. T., & Meredith, M. A. (1995).Neural mechanisms mediating attention and orienta-tion to multisensory cues. In M. Gazzaniga (Ed.), Thecognitive neurosciences (pp. 683–702). Cambridge,MA: MIT Press.

Stern, D. N. (1985). The interpersonal world of the in-fant. New York: Basic Books.

Stern, D. N., Hofer, L., Haft, W., & Dore, J. (1985). Af-fect attunement: The sharing by means of intermodalfluency. In T. M. Field & N. A. Fox (Eds.), Social per-ception in infants (pp. 249–286). Norwood, NJ:Ablex.

Stern, D. N., Jaffe, J., Beebe, B., & Bennett, S. L. (1975).Vocalizing in unison and alternation: Two modes ofcommunication within the mother–infant dyad. An-nals of the New York Academy of Sciences, 263, 89–100.

Trevarthen, C. (1979). Communication and coopera-tion in early infancy: A description of primaryintersubjectivity. In M. Bullowa (Ed.), Before speech:

The beginning of interpersonal communication.Cambridge, UK: Cambridge University Press.

Trevarthen, C. (1993). The self born in intersubjectivity:The psychology of an infant communicating. In U.Neisser (Ed.) Ecological and interpersonal knowl-edge of self (pp. 121–173). New York: CambridgeUniversity Press.

Tronick, E. Z. (1989). Emotions and emotional commu-nication in infants. American Psychologist, 44(2),112–119.

Vaish, A., & Striano, T. (2004). Is visual reference neces-sary? Developmental Science, 7, 261–269.

Walker, A. S. (1982). Intermodal perception of expres-sive behaviors by human infants. Journal of Experi-mental Child Psychology, 33, 514–535.

Walker-Andrews, A. S. (1986). Intermodal perception ofexpressive behaviors: Relation of eye and voice? De-velopmental Psychology, 22, 373–377.

Walker-Andrews, A. S. (1988). Infants’ perception ofthe affordances of expressive behaviors. In C. K.Rovee-Collier (Ed.), Advances in infancy research(Vol. 5, pp. 173–221). Norwood, NJ: Ablex.

Walker-Andrews, A. S. (1994). Taxonomy forintermodal relations. In D. J. Lewkowicz & R.Lickliter (Eds.), The development of intersensory per-ception (pp. 39–56). Hillsdale, NJ: Erlbaum.

Walker-Andrews, A. S. (1997). Infants’ perception ofexpressive behaviors: Differentiation of multimodalinformation. Psychological Bulletin, 121, 437–456.

Walker-Andrews, A. S., & Grolnick, W. (1983). Dis-crimination of vocal expression by young infants. In-fant Behavior and Development, 6, 491–498.

Walker-Andrews, A. S., & Lennon, E. M. (1985).Auditory–visual perception of changing distance byhuman infants. Child Development, 56, 544–548.

Walker-Andrews, A. S., & Lennon, E. (1991). Infants’discrimination of vocal expressions: Contributions ofauditory and visual information. Infant Behavior andDevelopment, 14, 131–142.

Walker-Andrews, A. S., Mayhew, E., Coffield, C., &Krogh-Jespersen, S. (2004). Familiarity allows con-tempt: Infants’ understanding of the emotion of fa-miliar people. Poster presented at the meeting of theInternational Society for Infant Studies, Chicago.

Wallace, M. T., Meredith, M. A., & Stein, B. E. (1992).Integration of multiple sensory modalities in cat cor-tex. Experimental Brain Research, 91, 484–488.

Weinberg, M. K., & Tronick, E. Z. (1994). Beyond theface: An empirical study of infant affective configura-tion of facial, vocal, gestural, and regulatory behav-iors. Child Development, 65, 1503–1515.

Wicker, B., Keysers, C., Plailly, J., Royet, J.-P., Gallese,V., & Rizzoulatti, G. (2003). Both of us disgusted inmy insula: The common neural basis of seeing andfeeling disgust. Neuron, 40, 655–664.

Young-Browne, G., Rosenfeld, H. M., & Horowitz, F.D. (1977). Infant discrimination of facial expres-sions. Child Development, 48, 555–562.


C H A P T E R 2 3

Long-Lived EmotionsA Life Course Perspectiveon Emotional Development

CAROL MAGAI

INTERVIEWER: Do you believe there is such a thing as a limited life spanfor humans?

MICHAEL R. ROSE, Professor of Evolutionary Biology: No. Life span istotally tunable. In my lab, we tune it up and down all the time.

—DREIFUS (2006, reporting on life extension in fruit flies)

If we could physically retard the process of ag-ing in humans, which appears increasinglylikely, would the landscape of our emotionallives change? In this chapter, I argue that re-tarding physical aging would indeed alter theemotion system, though perhaps in some unan-ticipated ways. In order to assess my predic-tion, we need to start with what is alreadyknown about emotion and the aging process.Accordingly, I review the literature on theoriesof emotion and aging, as well as what is knownabout change and continuity in emotion acrossthe adult lifespan based on the empirical litera-ture.

ADULT DEVELOPMENTAL MODELSOF AFFECT AND AGING

Over the past several decades, three bodies oftheory concerning emotion and aging have

risen to prominence: (1) differential-emotionstheory (DET) and related functionalist ac-counts, (2) optimization and selectivity theo-ries, and (3) cognitive–affective developmentaltheory. DET is given prominence in this reviewbecause, of the three theories, it is the only onethat actually has its empirical roots in infancyand early development, articulates with adultdevelopment and development in later life, andthus encompasses the lifespan. Because ofspace limitations, however, I attend mainly toits applicability to adulthood.

Differential-Emotions Theory

Although a number of theories regarding hu-man emotions were propounded during the1960s and 1970s, Carroll Izard’s DET (Izard,1971, 1977, 1991) was unique in that it had adistinctly developmental emphasis. My col-leagues and I have made other contributions in

376

this vein, especially in the area of adult devel-opment and aging (Magai & Haviland-Jones,2002; Magai & Nusbaum, 1996; Malatesta-Magai, Jonas, Shepard, & Culver, 1992). DET,a variant of discrete-emotions theory, derivesits name from the theory’s emphasis on thequalitatively different nature of the primary orbasic emotions: joy, interest, surprise, anger,contempt, disgust, fear, sadness, shame, andguilt, each with a distinctively different motiva-tional, expressive, and phenomenological as-pect. Although the preponderance of researchand theory emanating from Izard’s laboratoryhas dealt with infancy and early childhood,more recent formulations by Izard andcolleagues—and related discrete-emotionsaccounts—have explicitly addressed the issueof adult development and aging (Consedine &Magai, 2006; Izard & Ackerman, 1998; Izard,Ackerman, Schoff, & Fine, 1999; Magai &Haviland-Jones, 2002; Magai & McFadden,1995; Malatesta, 1982, 1990; Malatesta,Izard, Culver, & Nicolich, 1987; Malatesta &Wilson, 1988; Malatesta-Magai et al., 1992).

One aspect of the emotion system that isconsidered stable over developmental time in-volves the feeling states that accompany theprimary or fundamental emotions. Each pri-mary emotion is associated with a distinctqualitative feeling state, and each is paired witha particular motivational vector. These pairingsare held to be hard-wired in the nervous systemand to remain stable from infancy to old age.The constancy of emotional phenomenologyacross developmental epochs is attributed tothe fact that feeling–motivation linkages areevolutionarily grounded and adaptive in na-ture. For example, the feeling of fear is funda-mentally about the experience of threat and in-vokes the motivation to defend the self—at anystage of development. Similarly, the feeling ofsadness involves sensations of loss, whetherone is a child, a middle-aged adult, or an el-derly person; these sensations of loss are ac-companied by psychomotor slowing and, un-less they are suppressed, facial expressionsindicative of loss (horseshoe mouth andoblique brows). Expressive behavior—especially facial expressions of emotion—arealso seen as hard-wired; that is, the basic neu-rological substrate of emotion gives rise to pat-terned facial expressions that are recognizeduniversally (Izard, 1971), as in the sadness ex-pression just described. According to the the-ory, facial expressive behavior acts as both a

signal to the self (motivating acquisitive,affiliative, or self-protective behavior) and as asignal to the social environment (communicat-ing needs, desires, fears, and intentions). Thecritical motivational and communicative as-pects of facial expressive behavior, as evolu-tionarily grounded, are basic to survival andadaptation (Darwin, 1872); what DET makesexplicit is that these expressions are critical toadaptation and survival not only in infancy, butthroughout the lifespan (Izard & Malatesta,1987).

Going beyond these fixed elements of theemotion system, the theory specifies domainsin which developmental changes are to be ex-pected. Also, these changes are held to be in-herent not at the local level of the hard-wiredemotions themselves, but in their linkages withthe cognitive and behavioral subsystems, whichlead to elaboration and complexity. For exam-ple, individuals acquire an increasingly sophis-ticated ability to anticipate the emotional re-sponses of others over time, which accruesfrom accumulated interpersonal experienceand the increasingly dense and elaborated con-nections that are forged among the emotional,cognitive, and behavioral subsystems of per-sonality. Thus, in theory, we may anticipatethat greater age will be accompanied by an in-creasingly sophisticated and nuanced ability torecognize, appreciate, and relate to the emo-tions of others. As well, such experience-basedand experience-enriched interconnections shouldenable the internal perception of a more richlytextured emotional phenomenology—orgreater emotional complexity—over develop-mental time.

Another aspect of the emotion system thatundergoes change over developmental time oc-curs in the area of facial expressive behavior.Although the neuromuscular patterning of thebasic emotions is hard-wired, expressivebehavior becomes modulated in accord withfamilial and cultural norms and with advanc-ing skill at understanding interpersonal pro-cess; while this appears to occur most strikinglyin infancy and childhood, it is a lifespan pro-cess, according to research on younger, middle-aged, and older adults (Malatesta & Izard,1984). In turn, these developmental acquisi-tions create feedback and feedforward loopsthat have an impact on interpersonal relations,potentially facilitating the development, main-tenance, and protection of social relations andsocial support systems. These changes in the

23. Long-Lived Emotions 377

formal aspects of expressive behavior includethe ability to minimize or amplify expressivebehavior and the capacity to reflect mixed,nuanced, or more complex emotions.

Differential emotions theory also touchesupon how emotions relate to the acquisition ofpersonality traits, following on the work ofSilvan Tomkins (1962, 1963, 1991, 1992) andhis discussion of “ideo-affective structures.”Izard’s “affective–cognitive structures” (Izard,1971, 1977), like Tomkins’s ideo-affectivestructures, are construed as relatively stable as-pects of personality in which early and repeti-tive emotional experiences result in structuralchanges that become consolidated in personal-ity. These “structures,” or internal models, giverise to enduring expectations that color the in-dividual’s affective world and bias informationprocessing and behavior in predictable ways. Itis this aspect of DET that has received particu-lar elaboration in our work (e.g., Magai &Haviland-Jones, 2002; Magai & Nusbaum,1996; Malatesta, 1982, 1990), where refer-ences to “structures” are exchanged for themore plastic concepts of “affective traits” or“biases.” This work has detailed a set of pat-terned affect-specific perceptual biases andaffect-specific productive biases or emotion-based personality traits (Malatesta, 1990;Magai & McFadden, 1995; Malatesta & Wil-son, 1988), which are manifested in the behav-iors, cognitions, and interpersonal processes ofordinary individuals in everyday life. Anotherderivation from DET leads us to expect thatdifferent emotion-based aspects of personalityshould lead to qualitatively different kinds ofsocial relations (Magai, Hunziker, Mesias, &Culver, 2000; Izard, 1991). Because emotionsare fundamentally social in nature, and becausedifferent emotions communicate qualitativelydifferent kinds of social messages, they serve asimportant informational signals to social part-ners. If particular emotion biases are trait-likein nature and are emitted repetitively, they arelikely to establish different patterns of social re-lations.

Our more recent work (e.g., Magai &Haviland-Jones, 2002; Magai & Nusbaum,1996) has incorporated elements of dynamic-systems theory, such as attractors and repellors,phase shifts, and nonlinear effects, to model notonly personality stability but also personalitychange. In brief, the theory suggests that whenintense, unanticipated, or surprising emotional

experiences of a positive or negative nature areencountered and are cognitively elaborated inthe context of an interpersonally supportive en-vironment, internal models of affect and inter-personal process can be substantively reworkedand result in personality change.

In summary, DET postulates that feelingstates and patterned expressive behavior are in-nate aspects of the emotion system; however,the latter retains some degree of plasticity,which results in cultural variants of expressivebehavior, as well as idiosyncratic familial andindividual variants. Because of the intrinsicinterconnectivity of the nervous system and thelinkages among the emotional, cognitive, andbehavioral subsystems of personality (Izard,1971, 1977), emotional experiences at the indi-vidual level over developmental time give riseto the capacity for increasingly complex emo-tional phenomenology and increasing sophisti-cation in noticing, recognizing, and respondingto the emotions of others. The capacity forchange as a function of accumulative experi-ence suggests that, all other things being equal,as people age they should show an enhancedcapacity to experience and express more com-plex emotions, demonstrate greater emotion-regulatory capacities, experience and reflectgreater empathy to the emotional distress ofothers, and sustain and enrich interpersonal re-lations with age. However, the theory alsonotes that the magnitude of these developmen-tally enabled capacities is, to a certain extent,constrained at the individual level by the for-mation of emotional biases in the personality.These biases, in turn, condition the extent towhich, and manner in which, these capacitieswill be manifested. Finally, the biases them-selves are also subject to modification overtime, again depending on emotional events atthe level of individual lives.

Optimization and Selectivity Theoriesof Emotion

Another set of theoretical formulations aboutaffect and aging, which I refer to as “optimiza-tion and selectivity theories,” has centered onemotional regulatory capacities of individualsacross the adult years and their social environ-ments. The works of Powell Lawton and LauraCarstensen are most noteworthy in this respect.Both authors propose that emotions are centralpsychological processes, and that adults be-


come increasingly skilled emotion regulatorsover the lifespan, particularly in the later years.

Lawton (e.g., Lawton, 2001; Lawton,Kleban, & Dean, 1993; Lawton, Kleban,Rajagopal, & Dean, 1992; Lawton, Moss,Winter, & Hoffman, 2002; Lawton, Ruck-deschel, Winter, & Kleban, 1999) has proposedthat there are three dimensions of affective ex-perience that undergo change during adult-hood, especially in later life: (1) responsivenessor reactivity to emotional stimuli, (2) dynamicfeatures of affects (the frequency, rise time, du-ration, and intensity of emotions), and (3) self-regulatory processes. However, it is the thirdaspect of affective experience that has seen hisgreatest theoretical elaboration. According toLawton, as adults age they acquire greater mas-tery over their emotions, and this occurs in thecontext of adaptations to changes in social con-texts and life events. In this view, older adultsare not only more skilled at emotion regula-tion; they deploy their regulatory skills deliber-atively and proactively toward the general goalof “affective optimization.” That is, they seekto create and maintain the kinds of social envi-ronments in which there is a balance betweenemotionally stimulating and insulating fea-tures, and in which they can avoid the occasionof conflict and negative affect. AlthoughLawton has viewed this as the normative devel-opmental trajectory, he has also made allow-ances for individual differences. Indeed,Lawton et al. (1999) hypothesized that olderpeople may vary in their mix of openness to theexperience of emotion and in the degree towhich they control emotional expression, andthat these idiosyncratic patterns can conceiv-ably affect what they actually experience.

Carstensen’s socioemotional selectivity the-ory (Carstensen, 1988, 1992, 1993, 1995;Carstensen, Fung, & Charles, 2003), whichtakes its inspiration from a proposed model ofsuccessful aging (Baltes & Baltes, 1990)—selective optimization with compensation—is alifespan theory of motivation focused primarilyon how adults create conditions in which theymay optimize their social and emotional lives.Starting with the well-established observationthat the social networks of adults tend to narrowin the later part of the lifespan, Carstensen hasproposed that social network attrition is an elec-tively driven process that is related to changes inthe nature of goals in later life. That is, accordingto Carstensen, there are three primary human

motives: (1) emotion regulation, (2) develop-ment and maintenance of the individual’s self-concept, and (3) information seeking. The prior-ity of these motives changes in middle and laterlife, in the context of an age-related diminutionin physiological and psychological resources. Assuch, information seeking recedes in impor-tance, and emotional goals take on primacy;thus changes in emotions are consequences ofmotivational changes. Emotional goals includemaintaining optimal emotion regulation di-rected at a positively valenced emotional life,and achieving emotional satisfaction in the con-text of close and emotionally gratifying interper-sonal relationships.

These shifts in prioritization from knowl-edge acquisition to maintenance of social rela-tions have three sources. One relates to the ma-turing individual’s lifetime accrual ofexperience and knowledge, such that fewer andfewer people can provide information that isnew and personally meaningful. A secondsource is the reality of physical aging, which re-sults in energy depletion, heightened physiolog-ical arousability, and increased physical debili-tation. A third source is the growing realizationthat the individual’s remaining lifetime is finite.Although a sense of imminent mortality can beactivated in adults of any age, it is held to beparticularly salient in older adults and accentu-ates the need to prioritize emotional goals.Thus the theory provides a theoretical rationalefor the narrowing of social networks in laterlife, as an alternative to disengagement theory(Cumming & Henry, 1961). Instead of a turn-ing away from social engagement, as disen-gagement theory would have it, Carstensen’stheory proposes that older adults selectively re-duce their social networks at the service of in-vesting in and deepening those that are mostemotionally gratifying. That is, social contactsbecome limited to relationships that offer theopportunity to maximize positive outcomes—social support, companionship, and assistance—and to avoid or reduce resource-depleting neg-ative affect and conflict (Carstensen, 1992;Carstensen, Graff, Levenson, & Gottman,1996; Carstensen, Gottman, & Levenson,1995). Controlling the quality of social engage-ments constitutes an adaptive strategy to regu-late emotion, which is important to the mainte-nance of well-being in later life.

In summary, the theoretical formulations ofLawton and Carstensen have focused princi-


pally on emotion-regulatory characteristics,processes, and functions in the context of socialrelations over the adult years, especially in laterlife. Both view aging as involving shifts in goalpriorities and the concurrent acquisition ofmore accomplished and efficient emotion regu-latory skills. Together, these changes both moti-vate and assist aging individuals to reduceperipheral social interactions and avoid the ex-perience of negative affect. Lawton’s formula-tions acknowledge the existence of interper-sonal differences in the qualities of affectregulation that distinguish individual personal-ity profiles, and he proposes that this shouldhave an impact on what different individualsactually experience. In contrast, Carstensen’stheory assumes that adaptive mechanisms suchas socioemotional selectivity operate in similarways across different personality characteris-tics (Lang, Staudinger, & Carstensen, 1998;Mariske, Lang, Baltes, & Baltes, 1995). Thetheory assumes that all older individuals valueclose relations and are motivated to selectivelycultivate those relationships that are the mostemotionally gratifying. Both theories offerclear predictions regarding emotional experi-ence and social behaviors in older adults. Olderversus younger adults (or those confrontedwith the imminence of their own mortality)should orient more toward emotional stimuli,should place a priority on social versus infor-mational goals, and should engage in activitiesthat are likely to reduce negative affect and in-crease positive affect.

Cognitive–Affective Developmental Theory

Gisela Labouvie-Vief offers a theory of emo-tion and aging that situates emotional develop-ment within the context of evolving cognitiveand ego processes (Labouvie-Vief, 2003;Labouvie-Vief, Chiodo, Goguen, Diehl, &Orwoll, 1995; Labouvie-Vief, DeVoe, &Bulka, 1989; Labouvie-Vief & Diehl, 2000;Labouvie-Vief, Hakin-Larson, DeVoe, &Schoeberlein, 1989; Labouvie-Vief & Medler,2002). Based on neo-Piagetian and postformalcognitive models, her theory suggests that emo-tional experiences are qualitatively restruc-tured as the maturing individual acquires morecomplex forms of cognition with which to eval-uate the world and develops a more differenti-ated and integrated self-concept. The increas-ing cognitive sophistication that comes with

maturation provides ever more complex scaf-folding for more differentiated experiences ofemotion and for the development of more so-phisticated emotion-regulatory capacities.

Although there are no fixed stages inLabouvie-Vief’s theory, she proposes a trajec-tory of adult emotional development alignedwith levels of cognitive and personality de-velopment. In infancy and early childhood,emotion regulation is said to operate at apresystemic cognitive level, being governed bythe actions of external social agents. With theadvent of formal operations during adoles-cence and early adulthood, an “intrasystemic”period emerges, allowing the individual to reg-ulate emotions in the context of abstract idealsand cultural standards; here emotional processis characterized by internal regulation that re-mains highly dependent on social conventionswhich define standards of comportment withrespect to emotional speech and behavior. Assuch, the cognitive–affective system remainsbounded by conventional language, symbols,and norms; behavior is conformity-drivenrather than oriented toward change and trans-formation. Labouvie-Vief asserts that for indi-viduals to achieve true emotional maturity,they must advance beyond conventional gover-nance of behavior to one in which the individu-als integrate self-reflective internal subjectivityinto the cognitive–emotional system.

From early adulthood through middle age,there is an emerging intersystemic level of inte-gration, in which there is a gradual shift fromthe conventional to the contextualistic orienta-tion. That is, rules and standards for behaviorare no longer viewed as absolute, but rather asrelativistic and moderated by context. Subjec-tivity, autonomy, and self-exploration increaseas cognitions and behaviors become more indi-vidualized and less driven by convention. Suchadvances in cognitive and ego complexity areheld to produce important reorganizations ofthe self that ultimately have an impact on emo-tional development. This description of a hier-archical integration represents what Labouvie-Vief calls the “de-repression of emotion,”which comes to be more prominent dur-ing middle adulthood. In this normative-developmental aspect of the model, greatercognitive and emotional complexity during themiddle years gives rise to enhanced flexibilityof self-regulation and the capacity for moremodulated emotion expression.


In more recent formulations, Labouvie-Vief’swork on affective organization and change incognitive–affective development differentiatesbetween two aspects of affect regulation re-lated to the self, having different aims and pos-sibly different trajectories. One aspect seeks tomaximize positive affect and minimize or at-tenuate negative affect; here the goal is tomaintain a positively valenced self-concept thatentails regulating behaviors and values so as toconform to social norms and to avoid interper-sonal conflict. The other aspect of emotionaldevelopment, affective complexity, involves in-trapsychic differentiation, greater mixing ofpositive and negative affect, greater toleranceof ambiguity, and more flexible affect regula-tion. Its aim is to maintain an open and undis-torted view of reality. The greater awareness ofdifferentiated inner states and de-repression ofaffect during the middle years, in which oppos-ing feelings may conflict with one another, isnot without cost, as this situation makes for acertain degree of tension. Some individualsmay not be capable of tolerating or resolvingthe tension, which may provoke a retreat intomore defensive modes of coping. Other indi-viduals may find that the internal tension inten-sifies their efforts at reconciliation and integra-tion, producing advances in ego level and gainsin emotional maturity. Moreover, as a develop-mental trajectory, the enhanced growth incognitive–affective complexity may or may notcontinue to accrue beyond middle age, whenfluid intelligence (i.e., abstract cognitive abili-ties) starts to decline, since emotional maturityin this theoretical context is defined as inte-grally interknit with cognitive capacities.Hence this kind of emotional complexity maypeak in the middle years.

In summary, Labouvie-Vief’s theory enter-tains the thesis that there may be both general,developmental trends reflecting improvementwith age—although experience, ego level, andother cognitive phenomena, rather than age perse, are the agents of change—as well as individ-ual differences in the extent to which an indi-vidual has a preference for positivity or com-plexity. The theory maintains that the course ofemotional development over the adult years in-volves qualitative changes in the subjective as-pects of emotional life that entail more comp-lexly textured emotional experience andgreater capacity for affect regulation. The latterchanges accrue from more general develop-

mental changes in cognitive capacities and egodevelopment.

In concluding this section on theoretical con-tributions to the field of emotions and aging,we can see that there are a number of testabletheoretical formulations about change andcontinuity in different components of the emo-tion system over the adult years. In broadstrokes, the theories converge on aspects ofphenomenology, expressive behavior, sensitiv-ity to behavioral emotional cues, emotion regu-lation, and the role of emotions in personalityconsolidation and change. This then becomesthe organizing framework for a review of theempirical literature.

EMPIRICAL WORKON EMOTIONAL DEVELOPMENTOVER THE LIFE COURSE

Continuity and Change in SubjectiveEmotional Experience

Researchers have typically assessed age-relateddifferences in the phenomenological experienceof emotion through self-reports, using ratingscales of emotional frequency or intensity. An-other, less often used methodology is ratingemotional experience by examining narrativesabout emotional experiences. At times the fo-cus is on dimensional aspects of emotion; atother times the focus is at the discrete-emotionlevel. In the service of being comprehensive, inthe present chapter I distinguish between agedifferences in terms of broad hedonic tone(positive vs. negative affect) and at the level ofdiscrete or differential emotions. I also considerthe case of blended or mixed emotions, to as-sess whether theoretical formulations aboutemotional complexity at the individual levelcan be supported.

Positive and Negative Affect

There is now a considerable literature onchanges in positive and negative affect over theadult lifespan, though most of it is cross-sectional in nature. In brief, reviews of this lit-erature (Consedine & Magai, 2006; Magai,2001) indicate that positive affect increasesover the adult years, and that negative affectdecreases or remains level. Although the litera-ture is more limited, longitudinal studies sug-gest a somewhat different picture. There are six


such studies that have tracked changes overtime, ranging from 4 to 22 years.

The Berlin Aging Study, which drew its par-ticipants from a stratified probability sample ofolder adults (Baltes & Mayer, 1999), yieldedboth cross-sectional and longitudinal data onage-related differences in levels of positive andnegative affect as measured by the Positive andNegative Affect Schedules. Although cross-sectional data from this study indicated thatpositive affect scores declined across severalgroups of older adults, including individualsages 70–75 and 90–100 years, and that thesedeclines were substantial (Smith, Fleeson,Geiselmann, Settersten, & Kunzmann, 1999),longitudinal analyses over a 4-year period wereinterpreted as suggesting that positive andnegative affect remain relatively stable(Kunzmann, Little, & Smith, 2000). Studiesthat have tracked changes over longer periodsof time, however, tend to find a tapering off ofpositive and negative affect. The Los AngelesLongitudinal Study of Generations (Charles,Reynolds, & Gatz, 2001), which surveyedgroups of individuals from young adulthood tothe mid-80s, found that positive affect re-mained stable for young, middle-aged, andolder adults, but decreased slightly among theoldest individuals (mid-60s to mid-80s) whilenegative affect declined in all groups, includingthe oldest sample. The Boston Department ofVeterans Affairs-based Normative Aging Study(Mroczek & Spiro, 2005) assessed lifesatisfaction—a measure of global positiveaffect—over a 22-year period of time; the par-ticipants ranged in age from 40 to 85. The au-thors found that life satisfaction peaked at age65 and then declined. The Baltimore Long-itudinal Study of Aging (Davey, Halverson,Zonderman, & Costa, 2004; Terracciano, Mc-Crae, Brant, & Costa, 2005) collected data onneuroticism and extraversion—which load onnegative and positive affect, respectively—from1958 to the most recent wave of investigation;the authors found declines in both measures (atleast until age 80) in this highly educated co-hort of adults.

In summary, although the data are not alwaysconsistent, cross-sectional data indicate thatthere are increases in positive affect and de-creases in negative affect over the adult years,though some of these effects may be due to sam-ples with a restricted age range or to the exis-tence of cohort effects (Charles et al., 2001;Costa, McCrae, & Zonderman, 1987). The lon-

gitudinal data, especially data focused on thelater years, suggests that there are declines inpositive and negative affect toward the end ofthe lifespan, although there are exceptions tothis general trend. Declines in the frequency orlevel of affect in the later years may be due to thedepletion of physiological resources or to thenarrowing of social networks in later life that oc-curs as a normative aspect of aging (Carstensen,1992; Carstensen, Isaacowitz, & Charles,1999); reduced social networks conceivablylimit theoccasion for the elicitationof emotion.

Changes in Discrete Emotions

In some ways it is harder to evaluate changes inemotion at the discrete level, because the meth-ods of assessment are more varied, spanningretrospective ratings (e.g., Mroczek, 2001),experimental manipulations (e.g., Tsai,Carstensen, & Levenson, 2000), and onlineratings (e.g., Carstensen, Mayr, Pasupathi, &Nesselroade, 2000). In general, when the re-search is cross-sectional and the measures areself-reports of emotions, the most reliable find-ing is a reduction in the frequency or levelof anger with increasing age (Birditt &Fingerman, 2003; Gross et al., 1997; Lawton etal., 1993; Schieman, 1999; Stoner & Spencer,1987). Other cross-sectional data suggest re-ductions in shame/shyness (Consedine &Magai, 2003; Lawton et al., 1993), and possi-bly in sadness and fear, but not in disgust(Gross et al., 1997). The literature on longitu-dinal studies of discrete emotions is far morelimited. In fact, at this writing, there are onlythree relevant studies. One assessed changes inanxiety, sadness, interest, anger-in, anger-out,and total anger over an 8-year period of time ina sample of individuals whose mean ages were56 years when first interviewed and 63 years atthe second interview (Magai, 1999). Only an-ger-out showed a significant decline over the 8-year period, with no signs of change in theother emotions. Another study, the BaltimoreLongitudinal Study of Aging Terracciano et al.,2005), tracked changes in the facets constitut-ing the five dimensions of the NEO PersonalityInventory over time. Self-consciousness, a facetof neuroticism, which is related to the discrete-emotion construct of shame, showed a lineardecline with age, as did hostility; however, thiseffect appeared to taper off at the upperreaches of the age range (80–90 years), with thepossible exception of anxiety.


This picture of decreases in a number of dis-crete negative emotions as based on retrospec-tive self-reports is not sustained when one ex-amines studies in which emotions are elicitedunder emotion induction conditions. Instead,this research indicates either no age differencesin emotional experience or an enhanced effectamong older adults once emotion is aroused.For example, one study found that older andyounger adults reported equivalent levels of an-ger, sadness, fear, and interest in response to arelived-emotions task in which autobiographi-cal memories involving anger, sadness, fear,and interest were elicited (Malatesta-Magai etal., 1992); in another study using the relived-emotions task for memories involving fear, sad-ness, anger, and happiness, there were no age-related differences for fear, sadness, and happi-ness ratings, though there was a reduction ofanger ratings (Labouvie-Vief, Lumley, Jain, &Heinze, 2003). Several studies have used filmstimuli to elicit emotions under laboratory con-ditions. One study found that older versusyounger adults reported greater sadness in re-sponse to age-relevant sadness themes such asthe loss of loved ones (Kunzmann & Grühn,2005), whereas another found that youngerand older adults had equivalent levels of sub-jective responses to sad and amusing films (Tsaiet al., 2000). Finally, another study found thatolder adults had higher ratings on contemptthan younger adults in response to film clips ofinjustice (Charles, 2005). Other studies haveexamined age differences in response to politi-cal or historical events. For example, olderadults reported feeling just as sad, angry, andhopeful as middle-aged and younger adults inresponse to a political event (Levine & Bluck,1997). Thus it appears that when discreteemotions are aroused, older adults experiencethe same or greater intensities or levels ofsuch emotions as younger and middle-agedadults.

Emotional Experience as Sampledin Emotional Narratives

Another approach to assessing people’s phe-nomenological experience of emotion is toevaluate the quality of narrative productions.Here an individual is not constrained by afixed-format choice of items to rate, but gener-ates spontaneous material that reflects directlyon the nature of his or her inner affectiveworld. I first summarize the findings with re-

spect to positive and negative affect, and thenturn to those studies that have examined dis-crete emotions.

In one of the most extensive analyses of narra-tive material, researchers culled through over3,000 narrative transcripts of individuals rang-ing in age from 8 to 85 years for their use of affectwords; this was complemented by a similar anal-ysis of the collected works of 10 eminent authorsacross their professional lives (Pennebaker &Stone, 2003). The analysis of the self-disclosurematerial indicated that the use of positive affectincreased with age, with the magnitude of the in-crease from ages 55–69 to age 70+ being quitepronounced; the use of negative affect words de-clined with age. In the analysis of the output ofeminent authors such as Jane Austen, the re-searchers found a pattern similar to the findingsbased on the self-disclosure material. In anotherstudy (Schredl & Doll, 1998), the dream contentof people ranging in age from adolescence to the80s was content-analyzed; this research yieldeddata consistent with those from the foregoingstudy, in that age was once again negatively cor-related with negative affect. Of particular inter-est is a study that compared the self-report rat-ings of emotion with those derived fromnarratives in a sample of 87 younger and olderadults (mean ages of 20 and 62 years, respec-tively). In this study of reactions to the verdict ofthe O. J. Simpson trial as it was disclosed on TV,participants were interviewed and tape-recorded for later content analysis, and alsorated the intensity and frequency of their anger,sadness, happiness, and surprise (Alea, Bluck, &Semegon, 2004). The researchers found thatboth scalar and narrative measures indicated agreater emotional response among the olderadults and more frequent and more intense expe-rience of sadness in the older adults; a similartrend occurred for anger, but it did not reach sig-nificance. Finally, my colleagues and I tabulatedthe number of instances of 10 discrete-emotionwords in the narratives of young, middle-aged,and older adults in a relived-emotionstask (Magai, Consedine, Krivoshekova,McPherson, & Kudajre-Gyamfi, 2006). Wefound that 9 of the 10 emotion words were usedat equivalent frequencies; the only exceptionwas that younger adults used more expressionsof contempt than the other two age groups.

In summary, the bulk of evidence from nar-rative analyses points to no real declines in thecapacity for emotion across the adult years; ifanything, older adults sometimes express more


acute emotional experience than youngeradults. Thus, if retrospective ratings of the fre-quency or level of emotion indicate greater pla-cidity in older adults (lower positive and nega-tive affect), this could be a function of fadingmemory due to the passage of time. In contrast,if the eliciting circumstances are immediate, areautobiographical, or involve personally mean-ingful life events, the emotional capacities ofolder adults appear to be preserved.

Emotional Complexity

In a recently developing line of research, inves-tigators have sought to assess whether or notthere are age-related changes in the complexityof people’s emotional experiences. Research inthis vein is variously referred to as research on“affective complexity” (Labouvie-Vief,Chiodo, et al. (1995) and Labouvie-Vief, Diehl,et al. (1995), “emotional heterogeneity”(Charles, 2005), and “emotional differentia-tion” (Carstensen et al., 2000; Ong &Bergeman, 2004). Labouvie-Vief views theability to tolerate and sustain ambivalent ormixed emotions as a hallmark of cognitive–affective complexity (Labouvie-Vief, Chiodo,et al., 1995; Labouvie-Vief, DeVoe, et al.,1989; Labouvie-Vief, Diehl, Chiodo, & Coyle,1995), and others have suggested that the ca-pacity for more complex or variegated emo-tional experience may be functionally relatedto greater resiliency (Ong & Bergeman, 2004),increased life satisfaction (Carstensen et al.,2000, 2003), or the development of greater di-alectical, conflict-engaging thinking over time(Magai & Haviland-Jones, 2002).

The relation between complexity and age isnot altogether clear at this point, which in partmay be related to the fact that emotional com-plexity has been measured in several differentways. In Labouvie-Vief’s work, complexityscores are derived from the unique patterningof cognitive and affective processes as evincedin coded narratives. In her empirical work,complexity tends to increase over the adultyears and peaks in middle adulthood, withsome decline thereafter. The apparent declinein later life may be in part a function of the wayin which the construct is measured. Becausethis index of affective complexity has a strongcognitive component, and because fluid intelli-gence declines in later life (as noted earlier), it isperhaps not surprising that this measure of af-

fective complexity follows the trajectory ofcognitive decline.

In other research on emotional complexity,there appear to be linear increases with age, al-though the relevant longitudinal work has yetto be done. In these studies, “complexity” isdefined as involving an increased range of emo-tions, or as consisting of blends or overlapsamong negative emotions or between positiveand negative affects (Carstensen et al., 2000;Charles, 2005; Ong & Bergeman, 2004). Twostudies have assessed this kind of complexity inthe context of experience-sampling research(Carstensen et al., 2000; Ong & Bergeman,2004). In both studies, the indices of blendedor overlapping emotions were based on the av-erage intraindividual correlation between posi-tive and negative affect. Both found a linear in-crease in complexity/poignancy with age; thatis, older respondents tended to relate moremixed and bittersweet emotions and more poi-gnant experiences within the same sampledmoment than younger respondents did. A re-lated construct—“emotion heterogeneity,” de-fined as an overlap between a dominant nega-tive emotion (i.e., the most intense emotion)and other negative emotions—was found tocharacterize the experiences of older adults to agreater extent than those of younger adults(Charles, 2005; Magai et al., 2006). This kindof emotional complexity is apparently not lim-ited to the subjective aspect of emotion, but isalso found when experience is sampled fromnarrative reports (Magai et al., 2006), and alsomay be seen in expressive behavior (as detailedbelow).

Continuity and Changein Emotional Behavior

Although emotional behaviors encompass fa-cial, vocal, bodily, and gestural expressions, thebulk of the literature on age-related differencesin emotional behavior during the adult yearsinvolves research on facial expressions. Re-searchers have typically assessed age differ-ences either in response to film stimuli, duringthe course of a relived-emotions task (autobio-graphical memories), or during in vivo inter-personal interactions. One research teamstudied the expressions of older and youngerEuropean and Chinese American adults in re-sponse to a film-based emotion induction (Tsaiet al., 2000). The data indicated equivalent lev-


els of expressed happiness, laughing, smiling,and crying in the two age groups. In addition,four studies have used autobiographical memo-ries to elicit emotion. In one of the earlier stud-ies, spontaneous facial activity was monitoredduring a relived-emotions task (Levenson,Carstensen, Friesen, & Ekman, 1991); no dif-ferences between younger and older adultswere found. Another group of researchers(Moreno, Borod, Welkowitz, & Alpert, 1990)examined age differences in facial symmetry ofexpressive behavior in young, middle-aged,and older women during the induction of fouremotions, surprise, joy, sadness, and disgust;though they found no age differences inlaterality, they found that older faces wererated as showing more intense disgust. In athird study (Malatesta-Magai et al., 1992),younger adults (< 50 years old) and olderadults (> 50 years old) were asked to recountfour kinds of autobiographical emotionalevents involving interest, sadness, fear, and an-ger; older adults were found to express greaterinterest, sadness, fear, and anger, respectively,than younger adults. A more recent study inwhich young, middle-aged and older adults re-lated sad and angry events (Magai et al., 2006)found similar durations of anger, sadness, dis-gust, fear, and interest in all three age groups.However, there were two departures from thisgeneral trend: a greater duration of shame,contempt, and joy expressions in the youngersample, and a greater duration of knit browin the oldest adults. The shame–contempt–joy triad was interpreted as indexing a cynical/self-conscious, perhaps mocking, facialpresentation—a pattern that is common in ad-olescents and young adults (Kahlbaugh &Haviland, 1994; Malatesta-Magai & Dorval,1992). The greater duration of knit brow wasinterpreted as indexing greater concentration,thought, or conscientiousness on the part ofolder adults, or alternatively as indicating thatthe task was more demanding of the olderadults’ more limited cognitive resources.

Finally, in terms of in vivo interactions, astudy of married couples showed that oldercouples relative to middle-aged couples dis-played less anger and disgust during discussionof a marital problem (Carstensen et al., 1995),though there were no differences in joy, con-tempt, and sadness, and older adults displayedmore expressions of affection. Another study,this time involving the coding of spontaneous

facial expressions of emotion in patients withmid- to late-stage dementia during a familyvisit (Magai, Cohen, Gomberg, Malatesta, &Culver, 1996), found that negative emotionswere preserved across levels of cognitive de-cline; only joy expressions were lower in fre-quency among the most deteriorated patients.In that study, facial expressions of emotionswere also found to be functionally related toemotional contexts; that is, sadness expres-sions, which were rare during the beginningand middle stages of the visit, increased as fam-ily members prepared to leave. In general, then,the data from the above-described kinds ofstudies, although not always consistent at thelevel of discrete emotions—which may be at-tributable to differential eliciting conditionsor to the number and types of expressionscoded—suggest that there are no real decre-ments in the level of facial emotional ex-pressivity across the adult years or in the capac-ity to express emotion in emotional contexts.Indeed, in some cases, older adults show moreintense or more frequent emotional expressionsthan younger adults.

A smaller body of literature, however, pointsto a somewhat different aspect of facial expres-sions that may change with age. An early study(Malatesta & Izard, 1984), which involved film-ing young, middle-aged, and older women asthey described emotional events in their lives,undertook a component versus a global analysisof facial expressions. This study found that al-though there were fewer component musclemovements in the facial expressions of themiddle-aged and older women, their expressionswere more complicated, with more instances ofdifferent emotions combined within the samecoding unit. As well, in a second study, naïvejudges found it more difficult to accurately judgethe type of emotional events these women weretalking about from their facial expressions(Malatesta, Fiore, & Messina, 1987).

Continuity and Change in Sensitivityto Emotion Signals

The literature on sensitivity to emotional infor-mation has ranged from studies of “emotionalsalience”—that is, the degree to which peopleorient to emotional content—to the perceptionof facial, vocal, and bodily expressions of emo-tion. Here I describe only studies of emotionalsalience and studies of sensitivity to facial ex-


pressions of emotion; studies of vocal andbodily expressions have tended to use nonstan-dard stimuli.

A limited body of research has indicated thatemotional content may be more salient for olderversus younger adults. In an incidental-memoryparadigm, adults ranging from 20 to 83 yearswere asked to recall as much content as theycould (Carstensen & Charles, 1994); the studyfound a linear increase with age in the propor-tion of emotional to neutral phrases that were re-called, which was interpreted as indicating thatemotional content is more salient to older adults.A more recent study (May, Rahhat, Berry, &Leighton, 2005) assessed younger and olderadults’ ability to remember three types of con-textual information about an event—perceptual, conceptual nonemotional, and emo-tional. Results indicated that younger adultswere better than older adults at rememberingperceptual and conceptual source information,but that the age difference in source memory waseliminated when the informational content wasemotional, suggesting that emotional informa-tion differentially engages older adults. Otherresearch that has examined younger and olderadults’ processing of emotionally positive versusnegative material indicates that older adults tendto show a positivity bias. That is, they tend toorient to and remember positively valenced ma-terial more than negatively valenced material,while such a positivity effect does not seem to ap-ply in younger adults (see Mather & Carstensen,2005, for a review). Moreover, some have sug-gested that the older adults’ tendency to orient toemotional material may help compensate forsome of the losses in cognitive function that ac-crue with age, such as declines in working mem-ory, attention, and sensory functioning(Blanchard-Fields, 2005; Carstensen & Mikels,2005). However, there is also evidence that whenthe cognitive processing demands are too strong,and/or emotion is too arousing, it exerts a pen-alty on the executive functions of older adults. Ina study of attentional processes involvingemotion-laden words, known as the “emotionalStroop task” (Wurm, Labouvie-Vief, Aycock,Rebucal, & Koch, 2004), older adults showedpoorer performance than younger adults. Oneform of the task required participants to makelexical decisions to emotion words spoken inseveral tones of voice; in the other form of thetask, participants had to name the font color ofemotionally arousing words displayed on a com-puter screen. Latencies were longer for test

words spoken in an incongruent tone of voiceand for test words high on arousal, but only inthe case of older adults.

Studies have also examined adults’ capacityto read the emotional content of faces accu-rately. Two studies have found that older adultsare just as accurate as younger adults at recog-nizing a range of discrete emotions, but thatthey are poorer at recognizing sadness(MacPherson, Phillips, & Della Sala, 2002;Moreno, Borod, Welkowitz, & Alpert, 1993).Moreover, two studies of emotion recognitionamong patients in the early stages of Alzhei-mer’s disease showed that these patients per-formed at a level comparable to that of olderadult controls (Lavenu, Pasquier, Lebert, Petit,& Van der Linden, 1999). In contrast, anotherteam of investigators (McDowell, Harrison, &Demaree, 1994) found that older adults wereless accurate than younger adults in the identi-fication of sad, angry, fearful, and neutral ex-pressions, and one study of patients with de-mentia found that these patients recognizedfacial emotion at a lower rate than non-demented older adults and younger adults (Al-len & Brosgole, 1993).

In summary, emotion appears to be particu-larly salient to older adults, and this appears tohelp offset declines in cognitive function. Aswell, although there is some inconsistency inthe literature, there do not appear to be greatage differences in the ability to recognize facialexpressions of emotion, with the possible ex-ception of sadness.

Continuity and Changein Emotion Regulation

A body of literature has been accumulating onthe emotion regulation capacities of olderadults. The term “emotion regulation” encom-passes a range of behavioral responses, includ-ing deliberate suppression of ongoing experi-ence and/or expression, as well as anticipatoryregulation and the active avoidance of emo-tionally provocative situations. Self-reportstudies suggest that older adults control theiremotions more than younger adults do (Grosset al., 1997; Labouvie-Vief, DeVoe, et al.,1989; Lawton et al., 1992), that they may bebetter at anticipating and avoiding interper-sonal conflict (Birditt & Fingerman, 2005;Birditt, Fingerman, & Almeida, 2005; Birditt,Fingerman, & Almeida, 2005; Carstensen etal., 1995), and that they may be better at delay-


ing emotion expression (Diehl, Coyle, &Labouvie-Vief, 1996). Self-report studies are,however, vulnerable to presentational biases.The few experimental studies in which partici-pants have been required to control behavioralexpressions of emotion have primarily beenconducted with younger adults in response tofilm clips (Gross, 1998; Gross & Levenson,1993, 1997). Only two studies have directlytested the thesis that the capacity to regulateemotion improves with age. In one study(Kunzmann, Kupperbusch, & Levenson,2005), younger and older adults viewed threefilms designed to elicit disgust; no age differ-ences were found in the ability to suppress oramplify emotional expression, nor in subjectiveemotional experience. In the second study(Magai et al., 2006), young, middle-aged, andolder adults were asked to relate personal sto-ries involving anger and sadness, but were toldnot to let their feelings show behaviorally. Theimpact of the experimental manipulation wasonly evident in the older adults; the inhibitioninstructions had no effect on young andmiddle-aged adults. Taken together, these stud-ies suggest then that the ability to regulate emo-tion voluntarily does not decline with age andperhaps may even improve.

Continuity and Change in Emotion Traits

There is now fairly consistent evidence thatbroad dimensions of positive and negative af-fect are stable over both short and longer peri-ods of time. Studies tracking change over a pe-riod of weeks in the frequency or intensityof emotion states have found test–retestreliabilities in the .70s and .80s (Diener &Larsen, 1984; Epstein, 1979). Longer-termstudies have returned similar results. For exam-ple, a longitudinal study of subjective well-be-ing in a national sample (Costa et al., 1987)showed substantial stability in positive andnegative affect over a period of 7–12 years, andmeasures of extraversion and neuroticism(which tap positive and negative affect, respec-tively, as noted earlier) have also been found tobe stable over a 10-year period (Costa & Mc-Crae, 1980). In terms of discrete emotions, onestudy (Berenbaum, Fujita, & Pfennig, 1995)found stability coefficients in the .70s and .80sfor sadness, anger, and fear over a period ofweeks. A study of postpartum women (Izard,Libero, Putnam, & Haynes, 1993) assessed thestability of 12 emotions over 3 years; the stabil-

ities ranged from a low of .33 for fear to a highof .71 for contempt, with an average of .56.Another study (Magai, 1999) obtained 8-yearstability coefficients for anxiety, depression, in-terest, anger-in, anger-out, and total anger,ranging from .47 to .75. Finally, the BaltimoreLongitudinal Study of Aging, which assesseddepressive symptoms at five occasions over an8-year period of time, found that depressivesymptoms were trait-like in nature (Davey etal., 2004).

Balancing this picture of general stability area few studies that have examined inter-individual differences in intraindividual stabil-ity and change. Such studies have indicatedthat changes do occur at the level of the indi-vidual over the adult years in such broadcategories as positive–negative affect andneuroticism–extraversion (Charles et al., 2001;Kunzmann et al., 2000). Such changes havebeen attributed to changes in health, as well asto emotional events that involve the initiationor termination of intimate relationships orother intense affiliative experiences (Magai,1999; Magai & Haviland-Jones, 2002), al-though additional longitudinal research is nec-essary to flesh out this very preliminary work.

SUMMARY OF SUPPORTFOR THEORETICAL MODELS

Three leading bodies of theory, although theyhave different emphases, collectively suggest(1) that the capacity for emotion regulationshould improve over time, with a correspond-ing improvement in the balance of positive tonegative affect; (2) that persons should orientmore to the emotional aspects of life and hu-man relations as they age; (3) that basic emo-tional processes, such as the ability to recognizeemotion and the capacity to experience emo-tional arousal in response to emotional stimuli,should remain relatively intact; (4) that emo-tion experience and expression should becomemore complex, nuanced, and elaborated; and(5) that individual-level changes should consistof emergent stability in patterns of both emo-tion experience and expression, while still pre-serving the capacity for change.

The empirical literature, in the main, sug-gests broad support for these theoretical prop-ositions, though more longitudinal and cross-sequential studies are necessary to confirm thepatterns disclosed through cross-sectional


work and the limited longitudinal studies thatcurrently exist. With respect to changes inhedonic tone and discrete emotions, the mostconsistent findings indicate that there are in-creases in the level or frequency of positiveemotion and declines in negative emotionbroadly across the adult years, and especially adecrease in anger; these effects appear relatedto gains in the ability to regulate emotion overthe adult years. However, among the oldest-old, there appear to be declines in both positiveand negative emotions, which may be linked tothe shrinkage of social networks or to the de-pletion of physiological resources in late life. Interms of the salience of emotion, age appears tobring with it an increasing appreciation for theaffective aspects of life and social relations. Theliterature also seems to suggest greater emo-tional complexity with age, and these effectsare found at the phenomenological and expres-sive levels; however, depending on how com-plexity is defined and measured, this greater af-fective complexity may continue across the lifecourse or may taper off after middle age.Finally, it appears that emotional biases in ex-perience and expression become consolidatedin personality over developmental time, thoughchange can occur at the level of individual livesunder certain theoretically specified conditions.

CONCLUDING THOUGHTS:EMOTION, FRUIT FLIES,AND LIVES IN PROGRESS

The epigraph to this chapter alludes to the factthat recent advances in longevity-enhancing re-search have proven effective in retarding theaging process in fruit flies, not to mentionin mammals (e.g., Rikke & Johnson, 2004;Weindruch, 1996), and hold promise for in-creasing the human lifespan in the not-too-distant future. In the event of an extended life-span, would the landscape of our emotionallives change? In all likelihood, the answer isyes, on several counts. First, if time left to live isof the essence, as the formulations of socio-emotional selectivity theory suggest, we mightanticipate that the postponement of imminentmortality would raise the age of divestment ininformational goals (pursuit of novelty, sensa-tion seeking) in favor of socioemotional goals.In turn, the prolonged pursuit of novelty andsensation seeking might take a toll on pair-

bonded relationships and also lead to furtherchanges in career paths, if trends over the last50 years are any guide. But for the most part,the form of the emotional change would also“depend” on whether life extension would oc-cur in the context of stretching out the yearsof physical decline—an extended-senescencemodel—or whether, like Oscar Wilde’s charac-ter Dorian Gray, people would retain the healthand physical vitality of their youth. In the past,life extension accruing from decreased mortal-ity during infancy and life-saving medical ad-vances prolonged the length of the adult yearsoverall, but ultimately also led to lengtheningthe time spent in physical decline. MichaelRose’s (Dreifus, 2006) research on fruit flies, inwhich he extended the age over which repro-duction can occur through selective breedingover multiple generations, has shown that lifecan be significantly extended without incurringthe ravages of senescence. Research may ulti-mately reveal the genetic mechanisms that un-derlie this Dorian Gray effect, and thus theprospect of a materially enhanced human life-span is not inconceivable. In any case, it is clearthat the extended-senescence effect would re-sult in a different pattern of affective agingthan the Dorian Gray effect would. The lattereffect would preserve and extend all the bene-fits derived from the accumulated experienceand emotional wisdom into grand old ageswithout the penalty of physical decline,whereas the former would not. In the senes-cence model, an extended proportion of thelifespan would be spent in disability, depletionof physiological resources, and narrowing ofsocial networks, which would result in declinesin levels of emotion and perhaps reductions inemotional complexity. Finally, one possibly un-expected effect of the Dorian Gray model—under conditions of an extended sexual and re-productive life, and the attendant vicissitudesof serial partnerships and blended families—would be greater bouts of human drama. Butthen, as some would say, what’s wrong withthat?

ACKNOWLEDGMENTS

Work on this chapter was supported by grants fromthe National Institute of General Medical Scienceand the National Institute on Aging (Nos. SO6GM54650, 1KO7 AG00921, and RO1 AG021017).


REFERENCES

Alea, N., Bluck, S., & Semegon, A. B. (2004). Youngand older adults’ expression of emotional experience:Do autobiographic narratives tell a different story?Journal of Adult Development, 11, 235–250.

Allen, R., & Brosgole, L. (1993). Facial and auditory af-fect recognition in senile geriatrics, the normal el-derly and young adults. International Journal ofNeuroscience, 68, 33–42.

Baltes, P. B., & Baltes, M. M. (1990). Psychological per-spectives on successful aging: The model of selectiveoptimization with compensation. In P. B. Baltes &M. M. Baltes (Eds.), Successful aging: Perspectivesfrom the behavioral sciences (pp. 1–34). New York:Cambridge University Press.

Baltes, P. B., & Mayer, K. U. (Eds.). (1999). The BerlinAging Study: Aging from 70 to 100. New York:Cambridge University Press.

Berenbaum, H., Fujita, F., & Pfennig, J. (1995). Consis-tency, specificity, and correlates of negative emotions.Journal of Personality and Social Psychology, 68,342–352.

Birditt, K., & Fingerman, K. L. (2003). Age and genderdifferences in adults’ descriptions of emotional reac-tions to interpersonal problems. Journal of Gerontol-ogy: Series B. Psychological Sciences, 58B, 237–245.

Birditt, K., & Fingerman, K. L. (2005). Do we get betterat picking our battles?: Age group differences in de-scriptions of behavioral reactions to interpersonaltensions. Journals of Gerontology: Series B. Psycho-logical Sciences and Social Sciences, 60B, P121–P128.

Birditt, K., Fingerman, K. L., & Almeida, D. M. (2005).Age differences in exposure and reactions to interper-sonal tensions: A daily diary study. Psychology andAging, 20, 330–340.

Blanchard-Fields, F. (2005). Introduction to the specialsection on emotion–cognition interactions andthe aging mind. Psychology and Aging, 20, 539–541.

Carstensen, L. L. (1988). A life-span approach to socialmotivation. In J. Heckhausen & C. S. Dweck (Eds.),Motivation and self-regulation across the life span(Vol. 9, pp. 341–364). New York: Cambridge Uni-versity Press.

Carstensen, L. L. (1992). Social and emotional patternsin adulthood: Support for socioemotional selectivitytheory. Psychology and Aging, 7, 331–338.

Carstensen, L. L. (1993). Motivation for social contactacross the lifespan: A theory of socioemotional selec-tivity. In J. E. Jacobs (Ed.), Nebraska Symposium onMotivation: Vol. 40. Developmental perspectives onmotivation (pp. 209–254). Lincoln: University of Ne-braska Press.

Carstensen, L. L. (1995). Evidence for a life-span theoryof socioemotional selectivity. Current Directions inPsychological Science, 4, 151–156.

Carstensen, L. L., & Charles, S. T. (1994). The salience

of emotion across the adult life span. Psychology andAging, 9, 259–264.

Carstensen, L. L., Fung, H. H., & Charles, S. T. (2003).Socioemotional selectivity theory and the regulationof emotion in the second half of life. Motivation andEmotion, 27, 103–123.

Carstensen, L. L., Gottman, J. M., & Levenson, R. W.(1995). Emotional behavior in long-term marriage.Psychology and Aging, 10, 140–149.

Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T.(1999). Taking time seriously: A theory ofsocioemotional selectivity. American Psychologist,54, 165–181.

Carstensen, L. L., Mayr, U., Pasupathi, M., &Nesselroade, J. R. (2000). Emotional experience ineveryday life across the adult life span. Journal ofPersonality and Social Psychology, 79, 644–655.

Carstensen, L. L., & Mikels, J. A. (2005). At the inter-section of emotion and cognition. Aging and thepositivity effect. Current Directions in PsychologicalScience, 14, 117–121.

Carstensen, L. O., Graff, J., Levenson, R. W., &Gottman, J. M. (1996). Affect in intimate relation-ships: The developmental course of marriage. In C.Magai & S. McFadden (Eds.), Handbook of emo-tion, adult development, and aging (pp. 227–247).San Diego: Academic Press.

Charles, S. T. (2005). Viewing injustice: Greater emo-tion heterogeneity with age. Psychology and Aging,20, 159–164.

Charles, S. T., Reynolds, C. A., & Gatz, M. (2001). Age-related differences and change in positive and nega-tive affect over 23 years. Journal of Personality andSocial Psychology, 80, 136–151.

Consedine, N. S., & Magai, C. (2003). Attachment andemotion experience in later life: The view from emo-tions theory. Attachment and Human Development,5, 165–187.

Consedine, N. S., & Magai, C. (2006). Emotion de-velopment in adulthood: A developmental function-alist review and critique. In C. Hoare (Ed.), TheOxford handbook of adult development and learn-ing (pp. 123–148). New York: Oxford UniversityPress.

Costa, P. T., Jr., & McCrae, R. R. (1980). Influence ofextraversion and neuroticism on subjective well-be-ing: Happy and unhappy people. Journal of Person-ality and Social Psychology, 38, 668–678.

Costa, P. T., Jr., McCrae, R. R., & Zonderman, A. B.(1987). Environmental and dispositional influenceson well-being: Longitudinal follow-up of an Ameri-can national sample. British Journal of Psychology,78, 299–306.

Cumming, E., & Henry, W. (1961). Growing old. NewYork: Basic Books.

Darwin, C. E. (1872). The expression of the emotions inman and animals. London: Murray.

Davey, A., Halverson, C. F., Jr., Zonderman, A. B., &Costa, P. T., Jr. (2004). Change in depressive symp-


toms in the Baltimore Longitudinal Study of Aging.Journals of Gerontology: Series B. Psychological Sci-ences and Social Sciences, 59B, P270–P277.

Diehl, M., Coyle, N., & Labouvie-Vief, G. (1996). Ageand sex differences in strategies of coping and de-fense across the life span. Psychology and Aging, 11,127–139.

Diener, E., & Larsen, R. J. (1984). Temporal stabilityand cross-situational consistency of affective, behav-ioral, and cognitive responses. Journal of Personalityand Social Psychology, 47, 871–883.

Dreifus, C. (2006, December 6). Live longer with evolu-tion?: Evidence may lie in fruit flies. The New YorkTimes, p. 2.

Epstein, S. (1979). The stability of behavior: I. On pre-dicting most of the people much of the time. Journalof Personality and Social Psychology, 37, 1097–1126.

Gross, J. J. (1998). Antecedent- and response-focusedemotion regulation: Divergent consequences for ex-perience, expression, and physiology. Journal of Per-sonality and Social Psychology, 74, 224–237.

Gross, J. J., Carstensen, L. L., Pasupathi, M., Tsai, J.,Goetestam-Skorpen, C. G., & Hsu, A. Y. C. (1997).Emotion and aging: Experience, expression, and con-trol. Psychology and Aging, 12, 590–599.

Gross, J. J., & Levenson, R. W. (1993). Emotional sup-pression: Physiology, self-report, and expressivebehavior. Journal of Personality and Social Psychol-ogy, 64, 970–986.

Gross, J. J., & Levenson, R. W. (1997). Hiding feelings:The acute effects of inhibiting negative and positiveemotion. Journal of Abnormal Psychology, 106, 95–103.

Izard, C. E. (1971). The face of emotion. New York:Appleton-Century Crofts.



Izard, C. E., & Ackerman, B. P. (1998). Emotions andself-concept across the life-span. Annual Review ofGerontology and Geriatrics, 17, 1–26.

Izard, C. E., Ackerman, B. P., Schoff, K. M., & Fine, S.E. (1999). Self-organization of discrete emotions,emotion patterns, and emotion–cognition relations.In M. D. Lewis & I. Granic (Eds.), Emotion, develop-ment, and self-organization: Dynamic systems ap-proaches to emotional development (pp. 12–36).New York: Cambridge University Press.

Izard, C. E., Libero, D. Z., Putnam, P., & Haynes, O. M.(1993). Stability of emotion experiences and their re-lations to traits of personality. Journal of Personalityand Social Psychology, 64, 847–860.

Izard, C. E., & Malatesta, C. (1987). Emotional de-velopment in infancy. In J. Osofsky (Ed.), Hand-book of infant development (pp. 494–554). NewYork: Wiley.

Kahlbaugh, P. E., & Haviland, J. M. (1994). Nonverbalcommunication between parents and adolescents: A

study of approach and avoidance behaviors. Journalof Nonverbal Behavior, 18, 91–113.

Kunzmann, U., & Grühn, D. (2005). Age differences inemotional reactivity: The sample case of sadness.Psychology and Aging, 20, 47–59.

Kunzmann, U., Kupperbusch, C., & Levenson, R. W.(2005). Behavioral inhibition and amplification dur-ing emotional arousal: A comparison of two agegroups. Psychology and Aging, 20, 144–158.

Kunzmann, U., Little, T. D., & Smith, J. (2000). Is age-related stability of subjective well-being a paradox?:Cross-sectional and longitudinal evidence from theBerlin Aging Study. Psychology and Aging, 15, 511–526.

Labouvie-Vief, G. (2003). Dynamic integration: Affect,cognition, and the self in adulthood. Current Direc-tions in Psychological Science, 12, 201–206.

Labouvie-Vief, G., Chiodo, L. M., Goguen, L. A., Diehl,M., & Orwoll, L. (1995). Representations of selfacross the life span. Psychology and Aging, 10, 404–415.

Labouvie-Vief, G., DeVoe, M., & Bulka, D. (1989).Speaking about feelings: Conceptions of emotionacross the lifespan. Psychology and Aging, 4, 425–437.

Labouvie-Vief, G., & Diehl, M. (2000). Cognitive com-plexity and cognitive–affective integration: Relatedor separate domains of adult development? Psychol-ogy and Aging, 15, 490–504.

Labouvie-Vief, G., Diehl, M., Chiodo, L. M., & Coyle,N. (1995). Representations of self and parents acrossthe life span. Journal of Adult Development, 2, 207–222.

Labouvie-Vief, G., Hakin-Larson, J., DeVoe, M., &Schoeberlein, S. (1989). Emotions and self-regulation: A lifespan view. Human Development,32, 279–299.

Labouvie-Vief, G., Lumley, M. A., Jain, E., & Heinze,H. (2003). Age and gender differences in cardiac re-activity and subjective emotion responses to emo-tional autobiographical memories. Emotion, 3, 115–126.

Labouvie-Vief, G., & Medler, M. (2002). Affect optimi-zation and affect complexity: Modes and styles ofregulation in adulthood. Psychology and Aging, 17,571–588.

Lang, F. R., Staudinger, U. M., & Carstensen, L. L.(1998). Perspectives on socioemotional selectivity inlate life: How personality and social context do (anddo not) make a difference. Journal of Gerontology:Series B. Psychological Sciences, 53B, P21–P30.

Lavenu, I., Pasquier, F., Lebert, F., Petit, H., & Van derLinden, M. (1999). Perception of emotion in fronto-temporal dementia and in Alzheimer’s disease. Alz-heimer’s Disease and Associated Disorders, 13, 96–101.

Lawton, M. P. (2001). Emotion later in life. Current Di-rections in Psychological Science, 10, 120–123.

Lawton, M. P., Kleban, M. H., & Dean, J. (1993). Af-fect and age: Cross-sectional comparisons of struc-


ture and prevalence. Psychology and Aging, 8, 165–175.

Lawton, M. P., Kleban, M. H., Rajagopal, D., & Dean,J. (1992). The dimensions of affective experience inthree age groups. Psychology and Aging, 7, 171–184.

Lawton, M. P., Moss, M. S., Winter, L., & Hoffman, C.(2002). Motivation in later life: Personal projects andwell-being. Psychology and Aging, 17, 539–547.

Lawton, M. P., Ruckdeschel, K., Winter, L., & Kleban,M. H. (1999). Affect-experiential personality types inmiddle and late adulthood. Journal of Mental Healthand Aging, 5, 223–239.

Levenson, R. W., Carstensen, L. L., Friesen, W. V., &Ekman, P. (1991). Emotion, physiology, and expres-sion in old age. Psychology and Aging, 6, 28–35.

Levine, L. J., & Bluck, S. (1997). Experienced and re-membered emotional intensity in older adults. Psy-chology and Aging, 12, 514–523.

MacPherson, S. E., Phillips, L. H., & Della Sala, S.(2002). Age, executive function, and social decisionmaking: A dorsolateral prefrontal theory of cognitiveaging. Psychology and Aging, 17, 598–609.

Magai, C. (1999). Personality change in adulthood:Loci of change and the role of interpersonal process.International Journal of Aging and Human Develop-ment, 49, 339–352.

Magai, C. (2001). Emotions over the lifespan. In J. E.Birren & K. W. Schaie (Eds.), Handbook of the psy-chology of aging (5th ed., pp. 310–344). San Diego,CA: Academic Press.

Magai, C., Cohen, C., Gomberg, D., Malatesta, C., &Culver, C. (1996). Emotional expression during mid-to late-stage dementia. International Psychogeri-atrics, 8, 383–395.

Magai, C., Consedine, N. S., Krivoshekova, Y. S.,McPherson, R., & Kudadjie-Gyamfi, E. (2006).Emotion experience and expression across the adultlifespan: Insights from a multimodal assessmentstudy. Psychology and Aging, 21, 303–317.

Magai, C., & Haviland-Jones, J. (2002). The hidden ge-nius of emotion: Lifespan transformations of person-ality. New York: Cambridge University Press.

Magai, C., Hunziker, Mesias, W., & Culver, L. C.(2000). Adult attachment styles and emotional bi-ases. International Journal of Behavioral Develop-ment, 24, 301–309.

Magai, C., & McFadden, S. H. (1995). The role of emo-tions in social and personality development: History,theory and research. New York: Plenum Press.

Magai, C., & Nusbaum, B. (1996). Personality changein adulthood: Dynamic systems, emotions and thetransformed self. In S. H. McFadden (Ed.), Hand-book of emotion, adult development, and aging (pp.403–420). San Diego: Academic Press.

Malatesta, C. (1982). The expression and regulation ofemotion: A lifespan perspective. In T. Field & A.Fogel (Eds.), Emotion and early interaction (pp. 1–24). Hillsdale, NJ: Erlbaum.

Malatesta, C. (1990). The role of emotion in the devel-opment and organization of personality. In R.

Thompson (Ed.), Nebraska Symposium on Motiva-tion: Vol. 36. Socioemotional development (pp. 1–56). Lincoln: University of Nebraska Press.

Malatesta, C., Fiore, M. J., & Messina, J. (1987). Af-fect, personality, and facial expressive characteristicsof older individuals. Psychology and Aging, 1, 64–69.

Malatesta, C., & Izard, C. E. (1984). Facial expression ofemotion in young, middle-aged, and older adults. InC. Malatesta & C. E. Izard (Eds.), Emotion and adultdevelopment (pp. 253–274). Beverly Hills, CA: Sage.

Malatesta, C., Izard, C. E., Culver, L. C., & Nicolich,M. (1987). Emotion communication skills in young,middle-aged, and older women. Psychology andAging, 2, 193–203.

Malatesta, C., & Wilson, A. (1988). Emotion–cognitioninteractions in personality development: A discreteemotions, functionalist analysis. British Journal ofSocial Psychology, 27, 92–112.

Malatesta-Magai, C., & Dorval, B. (1992). Language,affect, and social order. In M. R. Gunnar & M.Maratsos (Eds.), Modularity and constraints in lan-guage and cognition (pp. 139–177). Hillsdale, NJ:Erlbaum.

Malatesta-Magai, C., Jonas, R., Shepard, B., & Culver,C. (1992). Type A personality and emotionalexpressivity in younger and older adults. Psychologyand Aging, 7, 551–561.

Mariske, M., Lang, F. R., Baltes, P. B., & Baltes, M. M.(1995). Selective optimization with compensation:Life-span perspectives on successful human develop-ment. In L. Backman (Ed.), Psychological compensa-tion: Managing losses and promoting gains (pp. 25–49). Hillsdale, NJ: Erlbaum.

Mather, M., & Carstensen, L. L. (2005). Aging and mo-tivated cognition: The positivity effect in attentionand memory. Trends in Cognitive Sciences, 9, 496–502.

May, C. P., Rahhat, T., Berry, E. M., & Leighton, E. A.(2005). Aging, source memory, and emotion. Psy-chology and Aging, 20, 571–578.

McDowell, C. L., Harrison, D. W., & Demaree, H. A.(1994). Is right hemisphere decline in the perceptionof emotion a function of aging? International Journalof Neuroscience, 70, 1–11.

Moreno, C., Borod, J. C., Welkowitz, J., & Alpert, M.(1990). Lateralization for the expression and percep-tion of facial emotion as a function of age.Neuropsychologia, 28, 199–209.

Moreno, C., Borod, J. C., Welkowitz, J., & Alpert, M.(1993). The perception of facial emotion across theadult life span. Developmental Neuropsychology, 9,305–314.

Mrozek, D. K. (2001). Age and emotion in adulthood.Current Directions in Psychological Science, 10(3),87–90.

Mroczek, D. K., & Spiro, A. (2005). Change in life sat-isfaction during adulthood: Findings from the Vet-erans Affairs Normative Aging Study. Journal of Per-sonality and Social Psychology, 88, 189–202.


Ong, A. D., & Bergeman, C. S. (2004). The complexityof emotions in later life. Journals of Gerontology: Se-ries B. Psychological Sciences and Social Sciences,59B, P117–P122.

Pennebaker, J. W., & Stone, L. D. (2003). Words of wis-dom: Language use over the lifespan. Journal of Per-sonality and Social Psychology, 85, 291–301.

Rikke, B. A., & Johnson, T. E. (2004). Lower body tem-perature as a potential mechanism of life extension inhomeotherms. Experimental Gerontology, 39, 927–930.

Schieman, S. (1999). Age and anger. Journal of Healthand Social Behavior, 40, 273–289.

Schredl, M., & Doll, E. (1998). Emotions in diarydreams. Consciousness and Cognition, 7, 634–646.

Smith, J., Fleeson, W., Geiselmann, B., Settersten, R. A.,& Kunzmann, U. (1999). Sources of well-being invery old age. In P. B. Baltes & K. U. Mayer (Eds.),The Berlin Aging Study: Aging from 70 to 100(pp. 450–471). New York: Cambridge UniversityPress.

Stoner, S. B., & Spencer, W. B. (1987). Age and genderdifferences with the Anger Expression Scale.Educational and Psychological Measurement, 47,487–492.

Terracciano, A., McCrae, R. R., Brant, L. J., & Costa,

P. T., Jr. (2005). Hierarchical linear modelinganalyses of the NEO-PI-R scales in the BaltimoreLongitudinal Aging Study. Psychology and Aging,20, 493–506.

Tomkins, S. S. (1962). Affect, imagery, consciousness:Vol. 1. The positive affects. New York: Springer.

Tomkins, S. S. (1963). Affect, imagery, consciousness:Vol. 2. The negative affects. New York: Springer.

Tomkins, S. S. (1991). Affect, imagery, consciousness:Vol. 3. The negative affects: Fear and anger. NewYork: Springer.

Tomkins, S. S. (1992). Affect, imagery, consciousness:Vol. 4. Cognition: Duplication and transformation ofinformation. New York: Springer.

Tsai, J., Carstensen, L. L., & Levenson, R. W. (2000).Automatic, subjective, and expressive responses toemotional films in older and younger Chinese Ameri-cans and European Americans. Psychology andAging, 15, 684–693.

Weindruch, R. (1996). Caloric restriction and aging.Scientific American, 274, 46–52.

Wurm, L. H., Labouvie-Vief, G., Aycock, J., Rebucal, K.A., & Koch, H. E. (2004). Performance in auditoryand visual emotional Stroop tasks: A comparison ofolder and younger adults. Psychology and Aging, 19,523–535.


PA R T I V

SOCIAL PERSPECTIVES

C H A P T E R 2 4

Gender and Emotion in Context

LESLIE R. BRODY and JUDITH A. HALL

Gender differences in emotional functioningare widely documented, but are often inconsis-tent across personality, social, cultural, and sit-uational variables, as well as types of emo-tional processes, quality of emotions, and taskcharacteristics. This is not surprising, consider-ing the adaptive communicative and motiva-tional functions that emotions serve. Sincemales and females are often socialized to havedifferent motives and goals—depending ontheir ages, cultural backgrounds, and socializa-tion histories—gender differences should occurin emotional processes, but should also fail togeneralize broadly, instead varying as a func-tion of these same factors. Social/interpersonalgoals that may differ for males and femalesinclude culturally prescribed gender roles(e.g., the role of child caretaker vs. economicprovider); social motives, such as needs forintimacy versus control; and adapting to thepower and status imbalances between the twosexes, in which men typically have higherpower and status than do women (see Brody,1999). Intrapersonal processes may also differfor males and females, including the ways in

which emotions and conflict are regulated andthe types of self-schemas (e.g., independencevs. interdependence) that are maintained(Cross & Madson, 1997). Both interpersonaland intrapersonal processes may be influencedby a complex interaction or feedback loop be-tween gender differences in underlying biologi-cal processes (such as temperament) and socialand cultural responses to those differences (es-pecially on the part of caretakers), which are inturn influenced by cultural values surroundinggender, gender roles, and differential emphaseson collectivism versus individualism.

Recent research has focused on the complex-ities of when and how gender differences vary,with the theoretical frame being that genderdifferences in emotional functioning are bothmediated and moderated by sociocultural, cog-nitive, biological, and behavioral variables. Inturn, the variables that are found either tomoderate or to mediate gender differences inemotional functioning give clues as to their eti-ology. In our chapter, we highlight contextualvariations in gender differences in several dif-ferent emotional processes, focusing on normal

395

adult populations. We also present an updatedtheoretical model for the etiology of gender dif-ferences.

STEREOTYPES AND DISPLAY RULES

The stereotype that females are more emo-tional than men is pervasive across several dif-ferent cultures (Timmers, Fischer, & Manstead,2003). Among North American samples,women are believed to be more emotionally in-tense (Robinson & Johnson, 1997), as well asmore emotionally expressive (e.g., smiling,laughing, crying) and more skilled in the use ofnonverbal cues related to emotion (Briton &Hall, 1995). However, stereotypes are alsoemotion-specific: Happiness, embarrassment,surprise, sadness, fear, shame, and guilt are be-lieved to occur more in women, and anger, con-tempt, and pride more in men (e.g., Hess et al.,2000; Plant, Hyde, Keltner, & Devine, 2000).Although the distinction between expressionand experience is not always made, when it is,the results consistently show stereotypes to bestronger for expression than for experience(Plant et al., 2000).

How well are the stereotypes supported byself-reports and behavioral data? In general,rather well. Stereotypes about gender differ-ences in expression tend to correspond withself-reports of expression (see below); stereo-types about gender differences in nonverbalbehavior correspond well with measured gen-der differences (Briton & Hall, 1995); and thebelief that gender differences in expression arestronger than gender differences in experienceis corroborated when researchers measure bothsimultaneously. For example, Kring andGordon (1998) documented gender differencesin facial expressions, but not in self-reports ofexperience, in response to evocative films.

To find that stereotypes are somewhat con-firmed in self-reports or in actual behaviorshould be relatively unsurprising, because ourmost automatically encoded and retrievablememories are based on frequently occurringbehaviors, which may form the basis for manystereotypes (Hasher & Zacks, 1984). On theother hand, as pointed out by Brody (1997),gender and emotion stereotypes are imprecise,are overly general, and ignore the importanceof the modality in which an emotion is ex-pressed, as well as the situational and culturalcontext within which emotional expression oc-

curs. Because stereotypes ignore both the socialcontext and individual differences, they oftenlead to the erroneous assumption that genderdifferences are exclusively biological in origin.

Despite these cautions, stereotypes warrant acloser analysis, because they powerfully shapethe reality of gender differences in at least twoways. First, in any given interaction, genderstereotypes can generate expectancies aboutour same- and opposite-sex partners that influ-ence and elicit particular behaviors and emo-tional expressions, becoming self-fulfillingprophecies (Hall & Briton, 1993). Second, ste-reotypes have a strong implicit prescriptive as-pect, taking the form of “display rules,” whichare cultural norms regulating how, when, andwhere emotions can be expressed by males andfemales in any particular culture. For example,across 48 countries, adults reported that happi-ness was more desirable for girls and that fear-lessness and anger suppression were moredesirable for boys (Diener & Lucas, 2004). Vi-olating stereotypic display rules can lead tonegative social consequences, such as social re-jection and discrimination.

SELF-REPORT MEASURES

Self-report measures, though serving as the ba-sis of much of the available evidence aboutemotion and expression, are problematic forthree reasons. First, gender stereotypes maycolor participants’ self-concepts and thereforetheir self-descriptions (Robinson & Clore,2002). Some research supports this idea. Theextent to which students endorsed gender ste-reotypes related to the extent to which theythemselves reported experiencing differentemotions from those of the opposite sex(Grossman & Wood, 1993). Second, stereotyp-ical responding may be exacerbated by socialdesirability motives, which in turn may be in-fluenced by display rules. And third, the impor-tant conceptual distinction between emotionalexperience and emotional expression is fre-quently blurred. The items to which partici-pants are asked to respond may not make thedistinction; the commonly used term “emo-tional,” for example, implies both experienceand expression. Or participants may have diffi-culty making this distinction even if they areasked to do so. With these cautions in mind, wereview studies that have used self-report mea-sures.

396 IV. SOCIAL PERSPECTIVES

Self-Reports of General EmotionalExperience and Expression

Many studies find that women rate themselvesas more emotionally expressive than men re-port themselves to be (e.g., Simon & Nath,2004). Gross and John (1998) factor-analyzedsix frequently used self-report measures ofemotional expression and identified five fac-tors: positive expressivity, negative expressivity,the intensity of emotional expression, expres-sive confidence (such as enjoying acting), andmasking or emotional regulation (such as sup-pressing anger). Women rated themselves sig-nificantly higher on the first three of these fac-tors. Greater intensity is found for women bothin their descriptions of specific emotional expe-riences (Hess et al., 2000; Tobin, Graziano,Vanman, & Tassinary, 2000) and on globalself-report measures such as the Affect Inten-sity Measure (AIM; Diener, Sandvik, & Larsen,1985). Intensity must be distinguished fromfrequency, however. In the 1996 General SocialSurvey, involving more than 1,300 respon-dents, there was no overall gender difference inreports of overall frequency of emotional expe-rience (Simon & Nath, 2004). Estimates ofoverall frequency may mask differences forspecific emotions, however, as we demonstratein the next section.

Finally, there is emotional contagion:Women report a higher likelihood of “catch-ing” the emotions of others than men report(Doherty, 1997). This self-reported differenceis corroborated, for certain emotions, when fa-cial muscle activity is recorded through electro-myography (Dimberg & Lundquist, 1990).Emotional contagion combined with facialfeedback processes (Strack, Martin, & Stepper,1988) could contribute to women’s greater re-ported intensity of experience.

Self-Reports of Specific Emotions

The specific positive emotions reported moreintensely or more frequently by women includejoy, love, affection, warmth, and feelings ofwell-being (see Brody, 1993; Fischer &Manstead, 2000). Higher female reporting ofpositive emotions emerges most clearly in situ-ations involving intimate interpersonal rela-tionships. When interpersonal situations arenot the focus—for example, when participantsare asked about total frequency in the pastweek—women report significantly less positive

affect than men report (Simon & Nath, 2004).Females also generally report more empathyand sympathy than do males; these emotionsare hard to classify as either positive or nega-tive (see Lennon & Eisenberg, 1987). However,gender differences in empathy become smallerfrom ages 22 to 92 (Schieman & Van Gundy,2000), reflecting either developmental or gen-erational effects. Many negative emotions—including disgust; sadness; feelings of vulnera-bility, such as fear, anxiety, and hurt; and feel-ings of dysphoric self-consciousness, such asshame and embarrassment—are generally re-ported more by women than by men (see, e.g.,Brody, 1999; Hess et al., 2000; Simon & Nath,2004; Fischer, Rodriguez Mosquera, vanVianen, & Manstead, 2004), although genderdifferences in shame are inconsistent (Simon &Nath, 2004) and can vary depending on thegender-role-related relevance of the target situ-ation (Ferguson, Eyre, & Ashbasker, 2000).Sadness, depression, and dysphoria are also re-ported to be more intense and of longer dura-tion by women than by men (Scherer, Wallbott,& Summerfield, 1986).

Although men may express more angerthrough vocal, facial, and behavioral modali-ties than women, the data on gender differ-ences in anger from research using self-reportquestionnaires are inconsistent. When askedgeneral questions about how many days perweek they get angry, men and women report nodifferences (Simon & Nath, 2004). However,when asked about the intensity of their anger,sometimes women report getting more in-tensely angry than men (e.g., Simon & Nath,2004) and sometimes they show no gender dif-ferences, as was shown across 37 differentcountries (Fischer et al., 2004). Reports of an-ger are sometimes, but not always, heightenedin women and girls when a situational contextis specified, especially one involving inter-personal relationships (e.g., Buntaine &Costenbader, 1997; Chaplin, Cole, & Zahn-Waxler, 2005; Kring, 2000). Compared to men,women also report more enduring experiencesof anger (Simon & Nath, 2004), are morelikely to report hurt or disappointment in re-sponse to anger-inducing situations (Brody,1993), and are more likely to report feelingashamed after expressing anger (Kring, 2000),but are also more likely to view their anger asappropriate (Simon & Nath, 2004).

Emotions that males sometimes report ex-pressing or are reported by others to express

24. Gender and Emotion in Context 397

more frequently or intensely than females doare contempt, loneliness, pride, confidence,guilt, and excitement (Brody, 1993, 1999; Si-mon & Nath, 2004). However, gender differ-ences in contempt, guilt, and loneliness havebeen inconsistent across studies, dependingon situational circumstances, the characteris-tics of the particular samples assessed, andmethodological variables (including differ-ences between scenario-based methods andtrait measures of these emotions) (Benetti-McQuoid & Bursik, 2005; Ferguson et al.,2000).

What Do Self-Reports Measure?

Women’s reports of higher affective intensityon global self-report measures such as the AIMmay not accurately reflect sex differences inemotion at the time feelings are initially ex-pressed or evoked. When daily logs are used toreport momentary emotions, either no genderdifferences have been found (Barrett, Robin,Pietromonaco, & Eyssell, 1998), or men havereported positive events in their lives to bemore intense than women have reported theirsto be (Seidlitz & Diener, 1998). Moreover, sub-sequent global self-reports of emotion do notsignificantly relate to the intensity of emotionalreactions reported at the time events occurred(Seidlitz & Diener, 1998). Robinson, Johnson,and Shields (1998) found that men and womenretrospectively remembered their emotionsas more gender-stereotypic than they actuallywere. In a theoretical review, Robinson andClore (2002) argue that global and retrospec-tive self-reports of emotion partially reflectmemories for the contextual details of events.Women may have more sophisticated emotionconcepts that can serve as retrieval cues, orthey may encode emotional experiences inmore detail than men do (Seidlitz & Diener,1998). This may subsequently contribute totheir reports of more intense emotions relativeto men on global measures, even in the absenceof gender differences in emotional intensity atthe time feelings are actually expressed. Thesehypotheses are consistent with data that girlsand women in the United States and Australiaare faster in accessing, and are able to recallmore, childhood memories of emotional expe-riences than their male counterparts (Davis,1999), and also with data from Poland thatwomen use more positive and negative emotion

words when recalling vivid memories than mendo (Niedzwienska, 2003). It is also possiblethat there are gender differences in the mentalimagery surrounding emotional events, and en-hanced mental imagery has been found to berelated to heightened affective responding(Miller et al., 1987).

Robinson and Clore (2002) also argue thatas the delay lengthens between the occurrenceof an emotional event and later recall, detailedmemories fade, and self-reports of emotionincreasingly rely on belief- and identity-consistent generalizations. Thus gender-stereotypic beliefs and identity may contributeto gender differences in global self-report mea-sures. Alternative explanations for the discrep-ancy between global and specific measures arealso possible, including the idea that in the timeelapsed since an event, women may cumula-tively experience more emotion than men—perhaps, for example, ruminating over theevent, which retriggers emotional experiences.And, as pointed out earlier, global self-reportmeasures often blur the distinction betweenemotional experience and emotional expres-siveness.

However, the gender differences that appearon global self-report measures—with femalesreporting a wide range of both more frequentand more intense emotions than men—are un-likely to be solely determined by stereotypes,self-presentation biases, memory encoding, orother cognitive differences between males andfemales. Gender differences also appear onother measures of expressiveness, including ob-served interactions, the verbalization of emo-tion, facial expressiveness, and nonverbal mea-sures.

VERBALIZATION OF EMOTION

Consistent with self-report data, women havebeen found to refer to both positive and nega-tive emotions more often in conversations andin writing samples. For example, in writing aresponse story to a scenario in which they dealtwith an obstructive travel agent, females mademore emotional references than did males(Girdler, Turner, Sherwood, & Light, 1990; seealso Brody, 1999). Female physicians have beenfound to engage patients in more talk aboutemotions than male physicians do (Roter, Hall,& Aoki, 2002).


In both self-descriptions and observations ofmarital interactions, wives are more willing totell their husbands when they are feeling tense;they are more apt to disclose their feelings; andthey are more apt to try to explain their feelingsthan are husbands (Burke, Weir, & Harrison,1976). Observations of marriages corroboratethat women express more emotions in words—especially more negative emotions, includingmore distress and anger—than men do. Menhave been found to withdraw from criticismand marital conflict by “stonewalling,” whichinvolves inhibiting facial action and minimiz-ing listening and eye contact, more than theirwives do (Levenson, Carstensen, & Gottman,1994). A recent study of dating couples indi-cates that gender differences in interactions aremoderated by levels of stress/discomfort withthe discussion: In high-stress situations, genderdifferences are maximized, with men express-ing less emotion, more restricted affect, andmore withdrawal (Vogel, Wester, Heesacker, &Madon, 2003).

FACIAL EXPRESSIONS AND OTHERNONVERBAL BEHAVIORS

Women are more accurate facial expressers ofmost emotions, both when posing deliberatelyand when being observed unobtrusively (Hall,1984). A measure integrating facial, vocal, andpostural expressions shows that girls expressmore sadness and anxiety than boys at ages 4and 6, and over this 2-year period, boysdecrease their expression of these emotions(Chaplin et al., 2005). Quantitative reviewshave also concluded that women are more gen-erally expressive with their faces and hands,and that they smile, laugh, and nod more thanmen do (Hall, 1984; LaFrance, Hecht, & LevyPaluck, 2003). It is important to note, however,that these behaviors do not have to reflectemotional states (Chovil, 1991–1992; Krauss,Chen, & Chawla, 1996). Smiling is notablyambiguous as to its “real” emotional meaning,with some authors suggesting the possibilitythat smiling in women reflects false positivity(women’s smiles were less congruent with thecontent of verbal statements than was the casefor men; Bugental, Love, & Gianetto, 1971).However, women’s facial expressions were lessdiscrepant from their words than were men’s inthe research of Halberstadt, Hayes, and Pike

(1988), and women and men displayed Du-chenne (enjoyment) and non-Duchenne (social)smiles in approximately equal proportions inHecht and LaFrance’s (1998) study, meaningthat women did not show an excessive propor-tion of emotionally artificial smiles.

Men may convey anger more clearly in theirfacial expressions than women do. For exam-ple, when participants were videotaped as theydiscussed angry, sad, and happy emotionalmemories, a panel of judges was subsequentlyable to identify men’s facial displays of anger(independent of verbal content) more accu-rately than women’s (Coats & Feldman, 1996).Men are also more facially reactive in responseto angry stimuli than are women, as measuredby facial electromyographic activity (Dimberg& Lundquist, 1990).

NONVERBAL DECODING SKILL

Across many studies, females score higher thanmales in identifying the meanings of nonverbalcues of face, body, and voice (Hall, 1978, 1984;McClure, 2000). Interestingly, most of thestudies have tested sensitivity to expressions ofaffect, which is a female-stereotypic knowledgedomain. When the knowledge domain is male-stereotypic (such as judging status or domi-nance), the differences appear to be muchsmaller or nonexistent (Schmid Mast & Hall,2004). Evidence is also accumulating that per-formance on nonverbal decoding tasks is sus-ceptible to motivational influences, which mayhave implications for the size of the obtainedgender differences (Klein & Hodges, 2001).

EMOTIONAL COMPETENCE

Components of Emotional Competence

Theories of emotional intelligence define sev-eral emotion-related traits and skills to be im-portant for adaptive functioning (Matthews,Zeidner, & Roberts, 2002), including perceiv-ing emotions accurately, using emotion to facil-itate thought, and understanding and manag-ing emotion. Women score higher on all ofthese components on the Mayer, Salovey, andCaruso Emotional Intelligence Test (Day &Carroll, 2004; Mayer, Caruso, & Salovey,2000). Consistent with this result is the largeliterature showing female advantage in perceiv-


ing nonverbally communicated emotions (seeabove).

Several other constructs can be seen as fall-ing generally under the heading of emotionalcompetence. Gohm and Clore (2000) foundthat women reported a greater tendency to payattention to their emotions. When participantswere clustered according to their pattern ofscores on attention, clarity, and intensity,women predominated among those who werehigh on all three and those who were high onintensity but low on clarity (called “over-whelmed”), while men predominated amongthose who were low on all three and those whowere low on intensity and high on clarity(called “cerebral”).

Women and girls display more complexemotion knowledge than men and boys dowhen asked to describe emotional reactionsof self and others in hypothetical scenarios(Ciarrochi, Hynes, & Crittenden, 2005).Ciarrochi et al. (2005) showed that a motiva-tional manipulation brought men’s perfor-mance up to the level of women’s, but only af-ter men spent a significantly longer amount oftime on the task. As mentioned above with re-gard to nonverbal decoding accuracy, motiva-tional factors may play an important role inemotion-related tasks.

Emotion Regulation

Emotion regulation or management consists ofbehavioral, cognitive, attentional, physiologi-cal, or emotional strategies to eliminate, main-tain, or change emotional experience and/orexpression (Ochsner & Gross, 2005), withclosely related constructs being coping strate-gies and defense mechanisms. These constructsinclude the ideas that people attempt to controlemotional processes in accordance with cul-tural pressures (display rules, stereotypes, andpower/status imbalances), and/or in accor-dance with personality-related factors (self-construals, motives, conflicts and goals, with aprimary goal being the avoidance of painfulaffect; Cramer, 2002; Matsumoto, Takeuchi,Andayani, Kouznetsova, & Krupp, 1998). Be-cause cultural pressures and personality-relatedfactors differ by gender, it should not be sur-prising to find gender differences in emotionregulation strategies. Men report or are ob-served to use more problem-solving, behavior-al, suppression, and externalizing emotion reg-ulation strategies than women do, including

blaming others, taking active steps, and engag-ing in distracting or avoidance activities suchas exercise (Brody, Muderrisoglu, & Nakash-Eisikovits, 2002; Cramer, 2002; Gross & John,2003). Women report or are observed to usemore social support strategies; internalizingstrategies, such as blaming themselves; andemotion-focused strategies, such as ruminat-ing, consisting of passively focusing attentionon negative affect rather than taking activesteps (Cramer, 2002; Nolen-Hoeksema &Jackson, 2001; Thayer, Newman, & McClain,1994). The gender difference in rumination hasbeen found to be mediated by trauma andchronic strain histories (Nolen-Hoeksema,Larson, & Grayson, 1999), as well as by sev-eral attributions, including the uncontrollabili-ty of negative emotions and feelings of respon-sibility for social relationships (Nolen-Hoeksema & Jackson, 2001).

Gender differences in emotion regulationstrategies need to be qualified by type of emo-tion and situation. For example, women reportthat they exert more control over anger, con-tempt, and disgust than do men, and men re-port that they exert more control over fear andsurprise than do women across four differentcultures (Matsumoto et al., 1998). Moreover,women choose rumination strategies when inneutral or depressed moods, but choose dis-traction when in angry moods (Rusting &Nolen-Hoeksema, 1998).

Personality factors, including motives, mayalso moderate gender differences in emotionregulation. Females report regulating angerand sadness to protect others’ feelings, whereasmales attempt to maintain control and toavoid nonsupportive interpersonal reactions(Timmers, Fischer, & Manstead, 1998; Zeman& Shipman, 1998). Agreeableness, one of theBig Five personality factors consisting of suchcharacteristics as helpfulness and sympathy, isa stronger predictor of self-reported efforts tocontrol emotions for women than for men(Tobin et al., 2000).

Although both avoidance and emotion-focused coping strategies tend to be non-adaptive for both sexes (Thayer et al., 1994),evidence indicates that some regulation and de-fense strategies may be differentially adaptivefor each gender, varying in complex ways as afunction of type of situation, how gender-stereotypic the defense is, and the quality ofthe emotion being regulated (Brody et al.,2002).


SITUATIONAL ANDRELATIONSHIP SPECIFICITY

In What Situation Is EmotionBeing Expressed?

Gender differences in each modality of emo-tional expression shift depending on the partic-ular situation. For example, when participantsrecorded their emotions in response to randombeeps by pagers for a 1-week period, womenreported more positive affect states (e.g., happyand friendly, as opposed to unhappy and an-gry) while at work than they did while at home.The opposite was true of men: They reportedmore positive affect states while at home(Larson, Richards, & Perry-Jenkins, 1994).That women might experience more negativestates in marriage than men do is consistentwith Stets’s (1997) observational study ofmarried couples in videotaped discussions.Women’s verbal and nonverbal behavior wasmuch more negative than that of their hus-bands.

Studies also indicate that the meaning of asituation for the two sexes affects patternsof emotional expressiveness. For example,women reported relatively more hurt and sad-ness when partners rejected them, in accor-dance with interdependent self-construals,while men reported relatively more hurt andsadness when partners demanded more inti-macy, in accordance with independent self-construals (Brody et al., 2002).

To Whom Is Emotion Being Expressed?

Some critical aspects of context affecting emo-tional expressiveness are the characteristics ofthe participants in the interaction and the na-ture of their relationship, including their levelof familiarity and intimacy, their power andstatus with respect to each other, and their gen-ders. For example, both men and women ex-press more emotions and more intense emo-tions to people they know intimately and feelcloser to (Barrett et al., 1998). Barrett et al.(1998) speculate that women’s tendencies torate their interpersonal interactions as more in-timate than men’s may partially mediate genderdifferences in emotional intensity. Women froma wide variety of cultures also express emotionsto a greater number of people than men, whotend to limit themselves to expressing emotionsonly to intimate partners (Rimé, Mesquita,Philippot, & Boca, 1991). Women also report

controlling their emotions less with familymembers than males do (Matsumoto et al.,1998).

Both sexes are also more comfortable dis-closing feelings (with the possible exception ofanger) to women than to men (Timmers et al.,1998). In a meta-analysis of sex differences inself-disclosure (which includes but is not lim-ited to the disclosure of feelings), women self-disclosed more to female partners, but notmore to male partners, than males did (Dindia& Allen, 1992). Anger may be the only feelingthat is verbally disclosed or directed more to-ward men than toward women, especially insituations in which no provocation is involved(Bettencourt & Miller, 1996).

CULTURAL SPECIFICITY

Across 37 countries, women report more in-tense emotions that last longer and are ex-pressed more overtly than do men (Fischer &Manstead, 2000). In other cross-cultural stud-ies, females express more nonverbal emotionalreactions—including facial reactions, vocal re-actions, body movements, laughing, andsmiling—when expressing joy, sadness, fear,and anger than males do (Scherer et al., 1986).Moreover, in a six-nation study using U.S. andJapanese college students as posers of facial ex-pressions, the emotions portrayed by femaleswere more accurately judged by every culturalgroup, even though the photographs were in-tended to be standardized exemplars (Biehl etal., 1997). However, interactions among gen-der of judge, gender of poser, and culture havealso been found to exist for at least some emo-tions (Matsumoto, 1992).

Gender differences in emotional expressionacross cultures are likely to vary as a functionof cultural values, especially individualistic ver-sus collectivistic values (giving priority to per-sonal goals vs. loyalty to collective/groupgoals). Collectivism needs to be distinguishedfrom relational values, which prioritize main-taining intimate relationships and which aremore characteristic of women than men acrosscultures (Kashima et al., 1995). In Fischer andManstead’s (2000) data, the extent of genderdifferences in the intensity and duration of joy,shame, disgust, and guilt, and in the nonverbalbehaviors associated with those emotions, weregreater in individualistic than in collectivisticcountries. Similarly, gender differences in the


reported intensity of emotion in response toscenarios depicting joy, fear, and anger werenot significant in a sample of American blacks,who are hypothesized to have more collec-tivistic values than other American ethnicgroups (Vrana & Rollock, 2002). Fischer andManstead (2000) theorize that males in indi-vidualistic cultures are especially likely to mini-mize emotional expressions, because express-ing emotions might threaten the control that iscritical to their status.

PHYSIOLOGICAL AROUSAL

Research suggests that gender differences inphysiological arousal, including changes inheart rate, blood pressure, skin conductance,and levels of catecholamines (epinephrine andnorepinephrine), are specific to particularphysiological measures and emotions, as wellas to particular tasks and circumstances (seeBrody, 1999). In the same situations, somemeasures of arousal (such as neuroendocrinefunctioning or blood pressure) show men to bemore aroused than women, while others (suchas cardiovascular reactivity) show inconsistentor contradictory gender difference patterns(Polefrone & Manuck, 1987; Neumann &Waldstein, 2001). Type of emotion, age, andethnicity may moderate gender differences. Forexample, men show higher levels of skin con-ductance to fearful films than women, but notto films evoking other emotions (Kring &Gordon, 1998). Although 15- to 50-year-oldwomen show higher cardiac reactivity thansame-age men when recalling angry or scarymemories, no gender differences are evidentwhen recalling sad or happy memories. In menand women over 50, no gender differences incardiac reactivity are found across emotions(Labouvie-Vief, Lumley, Jain, & Heinze,2003). The moderating effects of ethnicity areevident in a study showing that American blackmen exhibit greater cardiovascular reactivity toimagined emotional situations than othergroups do (Vrana & Rollick, 2002).

Internalizers, Externalizers, Generalizers,and Low Responders

Gender differences in the patterns of corre-spondence between physiological arousal andother modes of emotional expression (e.g., self-reports) are consistent with gender differences

in emotion regulation, and in fact often provideclues as to emotion control strategies. Earlierwork on gender differences in patterns of rela-tionships (Buck, 1977; Manstead, 1991) sug-gested that men were more often “internal-izers” (showing physiological arousal with noovert emotional expressions), whereas womenwere more often “externalizers” (showingovert emotional expressions with no corre-sponding physiological arousal). More recentstudies confirm these patterns, but also indicatethat women are relatively more likely than mento be “generalizers” (Brody, 1999)—that is, toshow concordance in their expression of emo-tion, even at young ages (Quas, Hong, Alkon,& Boyce, 2000), and especially at high levels ofphysiological reactivity (Avero & Calvo,1999). In contrast, men are more likely thanwomen to be “low responders” (Kring &Gordon, 1998), showing no or low levels of ex-pression across modalities. Discordance amongmales may be related to maintaining control orsuppressing the behaviors and self-reports thatcorrespond to arousal (Avero & Calvo, 1999),whereas concordance among females may berelated to heightened emotional awareness ofself and others and to female-stereotypic gen-der roles that encourage emotional expression.

Exceptions to these patterns have been notedin the literature on marital interaction, inwhich husbands’ arousal has sometimes beenfound to be more likely to correspond to nega-tive affect than wives’ (Levenson et al., 1994,but see Kiecolt-Glaser et al., 1996). Moreover,men’s cardiovascular reactivity is more oftenrelated to their expression and suppression ofanger than is women’s (Burns & Katkin, 1993).

NEURAL SUBSTRATESOF EMOTIONAL FUNCTIONING

With the advent of new technology, particu-larly functional magnetic resonance imaging(fMRI) and positron emission tomography(PET), researchers are studying potential gen-der differences in the brain regions involved inemotional expression, perception, and experi-ence, especially in limbic system activation andbrain lateralization (Schienle, Schafer, Stark,Watler, & Vaitl, 2005). Although recent fMRIand PET data are intriguing, they are alsoplagued with small sample sizes, the lack of acoherent theoretical model incorporating theidea that biological development is both influ-


enced by and influences the social context, andinconsistencies in interpretation. Researchersoften fail to consider that gender differences inthe activation of a specific brain region in re-sponse to emotional stimuli can result from amultitude of processes, including differences inattention, the quality of emotional experience,the imagery associated with the experience, orthe expression of emotion in different modali-ties (Wager & Ochsner, 2005). Activation itselfhas been confusingly interpreted to indicatethat an area of the brain is both strong or weakfor a particular function (Brody, 1999). Wehope that these limitations will be addressed asthe field progresses.

ETIOLOGY OFGENDER DIFFERENCES

Gender-role-related differences in motives,goals, and social status are consistent withmany of the data we have reviewed about gen-der differences in emotional processes. For ex-ample, the emotions that women display morethan men (e.g., warmth, happiness, shame,fear, and nervousness), their relatively strongerabilities in emotional decoding, and theirhigher facial and nonverbal expressiveness maybe related to motives for affiliation and inti-macy; to a self-schema based on interdepen-dence; to perceived vulnerability in the face oflower power; and to their traditional genderroles (including child caretaking and socialbonding, which necessitate reading others’emotion signals). Greater male pride, loneli-ness, and contempt are consistent with themale roles of differentiating from and compet-ing with others; with maintaining a relativelyhigh-status position; and with a self-schemabased on individualism or independence.Moreover, the differing types of emotion regu-lation strategies used by women and men, andtheir differing rationales for using them (avoid-ing interpersonal conflict vs. maintaining con-trol), are also consistent with gender-role-related motives and undoubtedly contribute tosome of the gender differences in patternsof expressiveness, including concordance/externalizing and discordance/internalizing.That gender differences in various aspects ofemotional functioning become minimized asadults age is also consistent with the idea thatemotions are adaptive for gender roles, whichbecome less rigid in later life (Gutmann, 1987).

However, social and gender-role-relatedvariables cannot always account for gender dif-ferences in emotion. For example, evidence in-dicates that status differences do not accountwell for gender differences in nonverbal behav-ior (Hall, Coats, & Smith LeBeau, 2005). Inour view, multiple interrelated factors contrib-ute to the etiology of gender differences thatspan cultural, biological, societal, interper-sonal, and intrapersonal levels of analysis. Fur-thermore, these multiple causes coexist withmultiple moderating factors. We propose twoetiological models that encompass proximaland distal factors; interpersonal and intra-personal feedback processes; and the complexintertwining of situational, sociocultural, bio-logical, personality, and cognitive factors bothover time and in specific situations.

A Developmental Perspective

The first model includes distal factors, such asgender differences in temperament, family so-cialization history, gender-segregated play pat-terns, and cultural values, all of which contrib-ute to the nature of gender differences. Anintegration of these factors involves a feedbackloop in which differing temperamental charac-teristics of male and female infants elicit differ-ential responses from caretakers and peers,who are also conforming to cultural pressuresand display rules for gender socialization. Dif-fering temperamental characteristics includehigher activity and arousal levels in males andfaster maturation rates for effortful controlprocesses in females (see Brody, 1999; Else-Quest, Hyde, Goldsmith, & Van Hulle, 2006).Although in the past this model also includedhigher levels of sociability/empathy in femalesas a gender difference possibly rooted in tem-perament (Brody, 1999; Brody & Hall, 2000),a recent meta-analysis (Else-Quest et al., 2006)suggests that there are gender differences in in-fant arousal (favoring boys) and in infanteffortful control processes, including inhibitorycontrol and attention focusing (favoring girls),but not in infant sociability. Infant gender dif-ferences in self-control may partly contributeto the higher levels of agreeableness and socia-bility characterizing females later in develop-ment (Goodwin & Gotlib, 2004) becausehigher levels of self-control (along with earlylanguage development) would make it morelikely that girls would attend to socioemotionalrelationships and rules. In turn, agreeableness


is a significant predictor of other emotionalprocesses that are heightened in women, in-cluding emotional intensity and efforts to regu-late emotions (Tobin et al., 2000).

The socialization of emotion is especially in-fluenced by characteristics of the family sys-tem, including the parents’ own temperaments,their gender role attitudes and behaviors, thequality of their marital relationships, their cul-tural and socioeconomic backgrounds, and thegender constellation of the children in theirfamilies (Brody, 1999). The quality of parent–child narrative discourse and interaction hasbeen found to vary as a function of the gendercomposition of the parent–child dyad and thetype of emotion displayed or discussed. For ex-ample, fathers attend more to their preschooldaughters’ emotions of sadness and anxietythan to their sons’, and to their sons’ expres-sions of anger than to their daughters’; parentalattention also predicts the later expression ofsadness and anxiety 2 years later (Chaplin etal., 2005). This and other research suggeststhat boys learn to be less expressive of, andmore controlling of, emotions communicatingvulnerability as they develop. Mothers usemore emotion labels in conversations with pre-school daughters than with sons, and mothers’use of emotion labels significantly predicts in-dividual differences in children’s use of emotionlabels (Cervantes & Callanan, 1998). And,consistent with feminist object relations analy-ses of development, when fathers are more in-volved in child care, sons and daughters ex-press less gender-role-stereotypic emotions(Brody, 1997). Finally, gender-segregated peergroups and differentiated patterns of play bothelicit and reinforce gender-role-specific emo-tional styles (Rose & Rudolph, 2006).

Putting Distal and Proximal CausesTogether: The Example of Smiling

The second model integrates distal factors(such as gender differences in social roles andcultural values, social knowledge, and develop-mental history) with proximal factors (includ-ing characteristics of the situation, especiallyquality of affect and others’ expectations andtreatment by others) to account for gender dif-ferences in expressiveness, using smiling as anillustration (Hall, Carter, & Horgan, 2000). Akey feature is a set of feedback processes thatintensify women’s positive affect during socialinteraction and thereby increase their smiling.

Smiling itself enhances positive affect throughboth physiological mechanisms and attribu-tional processes (Strack et al., 1988).

If we take women’s greater smiling as a start-ing point, regardless of its immediate cause, fa-cial feedback would produce more positive af-fect in women than in men. Positivepsychological feedback can also follow fromsmiling due to gender-related motives andtraits. For example, if women smile partly tofulfill their internalized conception of “femi-ninity,” this would reinforce their feelings offemininity and generate positive affect, whichwould produce more smiling. Other gender-related motives are interpersonal trust, likingfor others, and capacity for intimacy. Some ofthese motives intrinsically imply more positiveaffect and smiling (such as liking others). But,in addition, the knowledge that one has actedon these motives (showing that one is trusting,that one likes others, etc.) produces positive af-fect, because one is acting in concordance witha gender-relevant value (Wood, Christensen,Hebl, & Rothgerber, 1997). In turn, others re-spond favorably, contributing to the cycle.Women may also use smiling in the service ofsocial skills to put others at ease, facilitateinteraction, and defuse conflict. Again therewould be positive feedback, because it is rein-forcing to feel socially competent and to knowthat one has promoted comfort and communi-cation, which in turn promotes more smiling.And others’ favorable reactions produce posi-tive affect and more smiling. Finally, smiling it-self is highly reciprocal: The more one is smiledat, the more one will smile back (Hinsz &Tomhave, 1991). Reciprocity, combined withemotional contagion, should increase the inten-sity of women’s positive affect and smiling.

Thus women experience numerous positivefeedback cycles involving their own behavior,their cognitions, their physiological processes,and others’ behaviors. These sum to create en-hanced positive affect in their immediate socialinteractions compared to men’s, which ulti-mately influences how much they smile relativeto men.

Hankin and Abramson’s (2001) model forwomen’s heightened depression is also basedon feedback cycles involving frequent exposureto negative events, heightened negative affect,and cognitive vulnerability factors. The simi-larity between the two models highlights theneed for research on how the quality of imme-diate social interactions relates to or predicts


long-term emotional styles, as well as how non-verbal expressions (such as smiling) relate togeneralized affective states (such as depres-sion).

CONCLUDING THOUGHTS

As in our chapters in the first two editions ofthis volume (Brody & Hall, 1993, 2000), wehave continued to find consistent gender differ-ences in several different emotional processesacross several types of data. For example, gen-der stereotypes have frequently been borne outby data on actual patterns of behavior. We havealso emphasized that gender differences in anyparticular modality of emotional expressionare culturally and situationally specific. Wehave argued that the differential expression ofemotions for the two sexes is adaptive for thesuccessful fulfillment of gender roles, and wehave alluded to a developmental etiologicalmodel that integrates a multiplicity of vari-ables, including temperament and socializationfactors. Finally, using smiling as an example,we have shown that proximal affective experi-ences, including motivational, personality, andcognitive determinants and feedback cycles, arelikely to be important determinants of differen-tial affective experiences.

The numerous variables involved in under-standing both how and when gender differ-ences occur, as well as their origins, make theresearch process in this area particularly com-plex. The most productive research strategy inthis area is one that investigates how the corre-lates and patterns of emotional functioning dif-fer for each gender, incorporating a diverse setof biological, personality, social, cognitive, andcultural variables as both mediators and mod-erators.

REFERENCES

Avero, P., & Calvo, M. G. (1999). Emotional reactivityto social-evaluative stress: Gender differences in re-sponse systems concordance. Personality and Indi-vidual Differences, 27, 155–170.

Barrett, L. F., Robin, L., Pietromonaco, P. R., & Eyssell,K. M. (1998). Are women the more emotional sex?:Evidence from emotional experiences in social con-text. Cognition and Emotion, 12, 555–578.

Benetti-McQuoid, J., & Bursik, K. (2005). Individualdifferences in experiences of and responses to guiltand shame: Examining the lenses of gender and gen-der role. Sex Roles, 53, 133–142.

Bettencourt, B. A., & Miller, N. (1996). Gender differ-ences in aggression as a function of provocation: Ameta-analysis. Psychological Bulletin, 119, 422–447.

Biehl, M., Matsumoto, D., Ekman, P., Hearn, V.,Heider, K., Kudoh, T., et al. (1997). Matsumoto andEkman’s Japanese and Caucasian Facial Expressionsof Emotion (JACFEE): Reliability data and cross-national differences. Journal of Nonverbal Behavior,21, 3–21.

Briton, N. J., & Hall, J. A. (1995). Beliefs about femaleand male nonverbal communication. Sex Roles, 32,79–90.

Brody, L. R. (1993). On understanding gender differ-ences in the expression of emotion: Gender roles, so-cialization and language. In S. Ablon, D. Brown, E.Khantzian, & J. Mack (Eds.), Human feelings: Ex-plorations in affect development and meaning(pp. 89–121). Hillsdale, NJ: Analytic Press.

Brody, L. R. (1997). Beyond stereotypes: Gender andemotion. Journal of Social Issues, 53, 369–393.

Brody, L. R. (1999). Gender, emotion and the family.Cambridge, MA: Harvard University Press.

Brody, L. R., & Hall, J. A. (1993). Gender and emotion.In M. Lewis & J. M. Haviland (Eds.), Handbook ofemotions (pp. 447–460). New York: Guilford Press.

Brody, L. R., & Hall, J. A. (2000). Gender, emotion, andexpression. In M. Lewis & J. M. Haviland-Jones(Eds.), Handbook of emotions (2nd ed., pp. 338–349). New York: Guilford Press.

Brody, L. R., Muderrisoglu, S., & Nakash-Eisikovits, O.(2002). Emotions, defenses, and gender. In R. F.Bornstein & J. M. Masling (Eds.), The psychodynam-ics of gender and gender role (pp. 203–249). Washing-ton, DC: American Psychological Association.

Buck, R. (1977). Nonverbal communication accuracy inpreschool children: Relationships with personalityand skin conductance. Journal of Personality and So-cial Psychology, 33, 225–236.

Bugental, D. E., Love, L. R., & Gianetto, R. M. (1971).Perfidious feminine faces. Journal of Personality andSocial Psychology, 17, 314–318.

Buntaine, R. L., & Costenbader, V. K. (1997). Self-reported differences in the experience and expressionof anger between girls and boys. Sex Roles, 36, 625–637.

Burke, R. J., Weir, T., & Harrison, D. (1976). Disclosureof problems and tensions experienced by maritalpartners. Psychological Reports, 38, 531–542.

Burns, J. W., & Katkin, E. S. (1993). Psychological, situ-ational, and gender predictors of cardiovascular reac-tivity to stress: A multivariate approach. Journal ofBehavioral Medicine, 16, 445–465.

Cervantes, C. A., & Callanan, M. A. (1998). Labels andexplanations in mother–child emotion talk: Age andgender differentiation. Developmental Psychology,34, 88–98.

Chaplin, T. M., Cole, P. M., & Zahn-Waxler, C. (2005).Parental socialization of emotion expression: Genderdifferences and relations to child adjustment. Emo-tion, 5, 80–88.


Chovil, N. (1991–1992). Discourse-oriented facial dis-plays in conversation. Research on Language and So-cial Interaction, 25, 163–194.

Ciarrochi, J., Hynes, K., & Crittenden, N. (2005). Canmen do better if they try harder?: Sex and motiva-tional effects on emotional awareness. Cognition andEmotion, 19, 133–141.

Coats, E. J., & Feldman, R. S. (1996). Gender differ-ences in nonverbal correlates of social status. Person-ality and Social Psychology Bulletin, 22, 1014–1022.

Cramer, P. (2002). The study of defense mechanisms:Gender implications. In R. F. Bornstein & J. M.Masling (Eds.), The psychodynamics of gender andgender role (pp. 81–128). Washington, DC: Ameri-can Psychological Association.

Cross, S. E., & Madson, L. (1997). Models of the self:Self-construals and gender. Psychological Bulletin,122, 5–37.

Davis, P. J. (1999). Gender differences in autobiographi-cal memory for childhood emotional experiences.Journal of Personality and Social Psychology, 76,498–510.

Day, A. L., & Carroll, S. A. (2004). Using an ability-based measure of emotional intelligence to predict in-dividual performance, group performance, andgroup citizenship behaviors. Personality and Individ-ual Differences, 36, 1443–1458.

Diener, E., Sandvik, E., & Larsen, R. J. (1985). Age andsex effects for emotional intensity. DevelopmentalPsychology, 21, 542–546.

Diener, M. L., & Lucas, R. E. (2004). Adults’ desires forchildren’s emotions across 48 countries. Journal ofCross-Cultural Psychology, 35, 525–547.

Dimberg, U., & Lundquist, L. (1990). Gender differ-ences in facial reactions to facial expressions. Biolog-ical Psychology, 30, 151–159.

Dindia, K., & Allen, M. (1992). Sex differences in self-disclosure: A meta-analysis. Psychological Bulletin,112, 106–124.

Doherty, R. W. (1997). The Emotional Contagion Scale:A measure of individual differences. Journal of Non-verbal Behavior, 21, 131–154.

Else-Quest, N. M., Hyde, J. S., Goldsmith, H. H., &Van Hulle, C. A. (2006). Gender differences in tem-perament: A meta-analysis. Psychological Bulletin,132, 33–72.

Ferguson, T. J., Eyre, H. L., & Ashbasker, M. (2000).Unwanted identities: A key variable in shame–angerlinks and gender differences in shame. Sex Roles, 42,133–157.

Fischer, A. H., & Manstead, A. S. R. (2000). The rela-tion between gender and emotion in different cul-tures. In A. H. Fischer (Ed.), Gender and emotion:Social psychological perspectives (pp. 71–98). NewYork: Cambridge University Press.

Fischer, A. H., Rodriguez Mosquera, P. M., van Vianen,A. E. M., & Manstead, A. S. R. (2004). Gender andculture differences in emotion. Emotion, 4, 87–94.

Girdler, S. S., Turner, J. R., Sherwood, A., & Light, K. C.(1990). Gender differences in blood pressure control

during a variety of behavioral stressors. Psychoso-matic Medicine, 52, 571–591.

Gohm, C. L., & Clore, G. L. (2000). Individual differ-ences in emotional experience: Mapping availablescales to processes. Personality and Social Psychol-ogy Bulletin, 26, 679–697.

Goodwin, R. D., & Gotlib, I. H. (2004). Gender differ-ences in depression: The role of personality factors.Psychiatry Research, 126, 135–142.

Gross, J. J., & John, O. P. (1998). Mapping the domainof expressivity: Multimethod evidence for a hierar-chical model. Journal of Personality and Social Psy-chology, 74, 170–191.

Gross, J. J., & John, O. P. (2003). Individual differencesin two emotion regulation processes: Implications foraffect, relationships, and well-being. Journal of Per-sonality and Social Psychology, 85, 348–362.

Grossman, M., & Wood, W. (1993). Sex differences inintensity of emotional experience: A social role inter-pretation. Journal of Personality and Social Psychol-ogy, 65, 1010–1022.

Gutmann, D. (1987). Reclaimed powers: The new psy-chology of men and women in later life. New York:Basic Books.

Halberstadt, A. G., Hayes, C. W., & Pike, K. M. (1988).Gender and gender role differences in smiling andcommunication consistency. Sex Roles, 19, 589–604.

Hall, J. A. (1978). Gender effects in decoding nonverbalcues. Psychological Bulletin, 85, 845–857.

Hall, J. A. (1984). Nonverbal sex differences: Commu-nication accuracy and expressive style. Baltimore:Johns Hopkins University Press.

Hall, J. A., & Briton, N. J. (1993). Gender, nonverbalbehavior, and expectations. In P. D. Blanck (Ed.), In-terpersonal expectations: Theory, research, and ap-plications (pp. 276–295). Cambridge, UK: Cam-bridge University Press.

Hall, J. A., Carter, J. D., & Horgan, T. G. (2000). Gen-der differences in the nonverbal communication ofemotion. In A. H. Fischer (Ed.), Gender and emotion:Social psychological perspectives (pp. 97–117). NewYork: Cambridge University Press.

Hall, J. A., Coats, E. J., & Smith LeBeau, L. (2005).Nonverbal behavior and the vertical dimension of so-cial relations: A meta-analysis. Psychological Bulle-tin, 131, 898–924.

Hankin, B. L., & Abramson, L. Y. (2001). Developmentof gender differences in depression: An elaboratedcognitive vulnerability–transactional stress theory.Psychological Bulletin, 127, 773–796.

Hasher, L., & Zacks, R. T. (1984). Automatic process-ing of fundamental information: The case of fre-quency of occurrence. American Psychologist, 39,1372–1388.

Hecht, M. A., & LaFrance, M. (1998). License or obli-gation to smile: The effect of power and sex onamount and type of smiling. Personality and SocialPsychology Bulletin, 24, 1332–1342.

Hess, U., Senecal, S., Kirouac, G., Herrera, P., Philippot,P., & Kleck, R. E. (2000). Emotional expressivity in


men and women: Stereotypes and self-perceptions.Cognition and Emotion, 14, 609–642.

Hinsz, V. B., & Tomhave, J. A. (1991). Smile and (half)the world smiles with you, frown and you frownalone. Personality and Social Psychology Bulletin,17, 586–592.

Kashima, Y., Yamaguchi, S., Kim, U., Choi, S., Gelfand,M. J., & Yuki, M. (1995). Culture, gender and self: Aperspective from individualism–collectivism re-search. Journal of Personality and Social Psychology,69, 925–937.

Kiecolt-Glaser, J. K., Newton, T., Cacioppo, J. T.,MacCallum, R. C., Glaser, R., & Malarkey, W. B.(1996). Marital conflict and endocrine function: Aremen really more physiologically affected thanwomen? Journal of Consulting and Clinical Psychol-ogy 64, 324–332.

Klein, K. J. K., & Hodges, S. D. (2001). Gender differ-ences, motivation, and empathic accuracy: When itpays to understand. Personality and Social Psychol-ogy Bulletin 27, 720–730.

Krauss, R. M., Chen, Y., & Chawla, P. (1996). Nonver-bal behavior and nonverbal communication: Whatdo conversational hand gestures tell us? In M. P.Zanna (Ed.), Advances in experimental social psy-chology (Vol. 28, pp. 389–450). San Diego, CA: Aca-demic Press.

Kring, A. M. (2000). Gender and anger. In A. H. Fischer(Ed.), Gender and emotion: Social psychological per-spectives (pp. 211–231). New York: Cambridge Uni-versity Press.

Kring, A. M., & Gordon, A. H. (1998). Sex differencesin emotion: Expression, experience, and physiology.Journal of Personality and Social Psychology 74,686–703.

Labouvie-Vief, G., Lumley, M. A., Jain, E., & Heinze,H. (2003). Age and gender differences in cardiac re-activity and subjective emotion responses to emo-tional autobiographical memories. Emotion 3, 115–126.

LaFrance, M., Hecht, M. A., & Levy Paluck, E. (2003).The contingent smile: A meta-analysis of sex differ-ences in smiling. Psychological Bulletin, 129, 305–334.

Larson, R. W., Richards, M. H., & Perry-Jenkins, M.(1994). Divergent worlds: The daily emotional expe-rience of mothers and fathers in the domestic andpublic spheres. Journal of Personality and Social Psy-chology, 67, 1034–1046.

Lennon, R., & Eisenberg, N. (1987). Gender and agedifferences in empathy and sympathy. In N.Eisenberg & J. Strayer (Eds.), Empathy and its devel-opment (pp. 195–217). Cambridge, UK: CambridgeUniversity Press.

Levenson, R. W., Carstensen, L. L., & Gottman, J. M.(1994). The influence of age and gender on affect,physiology, and their interrelations: A study of long-term marriages. Journal of Personality and SocialPsychology, 67, 56–68.

Manstead, A. S. R. (1991). Expressiveness as an individ-

ual difference. In R. S. Feldman & B. S. Rimé (Eds.),Fundamentals of nonverbal behavior (pp. 285–328).Cambridge, UK: Cambridge University Press.

Matsumoto, D. (1992). American–Japanese cultural dif-ferences in the recognition of universal facial expres-sions. Journal of Cross-Cultural Psychology, 23, 72–84.

Matsumoto, D., Takeuchi, S., Andayani, S.,Kouznetsova, N., & Krupp, D. (1998). The contribu-tion of individualism vs. collectivism to cross-national differences in display rules. Asian Journal ofSocial Psychology, 1, 147–165.

Matthews, G., Zeidner, M., & Roberts, R. D. (2002).Emotional intelligence: Science and myth. Cam-bridge, MA: MIT Press.

Mayer, J. D., Caruso, D. R., & Salovey, P. (2000). Emo-tional intelligence meets traditional standards for anintelligence. Intelligence, 27, 267–298.

McClure, E. B. (2000). A meta-analytic review of sexdifferences in facial expression processing and theirdevelopment in infants, children, and adolescents.Psychological Bulletin, 126, 424–453.

Miller, G. A., Levin, D. N., Kozak, M. J., Cook, E. W.,McLean, A., & Lang, P. J. (1987). Individual differ-ences in imagery and the psychophysiology of emo-tion. Cognition and Emotion, 1, 367–390.

Neumann, S. A., & Waldstein, S. R. (2001). Similar pat-terns of cardiovascular response during emotionalactivation as a function of affective valence andarousal and gender. Journal of Psychosomatic Re-search, 50, 245–253.

Niedzwienska, A. (2003). Gender differences in vividmemories. Sex Roles, 49, 321–331.

Nolen-Hoeksema, S., & Jackson, B. (2001). Mediatorsof the gender difference in rumination. Psychology ofWomen Quarterly, 25, 37–47.

Nolen-Hoeksema, S., Larson, J., & Grayson, C. (1999).Explaining the gender difference in depressive symp-toms. Journal of Personality and Social Psychology,77, 1061–1072.

Ochsner, K. N., & Gross, J. J. (2005). The cognitivecontrol of emotion. Trends in Cognitive Sciences, 9,242–249.

Plant, E. A., Hyde, J. S., Keltner, D., & Devine, P. G.(2000). The gender stereotyping of emotions. Psy-chology of Women Quarterly, 24, 81–92.

Polefrone, J. M., & Manuck, S. B. (1987). Gender dif-ferences in cardiovascular and neuroendocrine re-sponse to stressors. In R. Barnett, L. Biener, & G.Baruch (Eds.), Gender and stress (pp. 13–38). NewYork: Free Press.

Quas, J., Hong, M., Alkon, A., & Boyce, W. T. (2000).Dissociations between psychobiologic reactivity andemotional expression in children. DevelopmentalPsychobiology, 37, 153–175.

Rimé, B., Mesquita, B., Philippot, P., & Boca, S. (1991).Beyond the emotional event: Six studies on the socialsharing of emotion. Cognition and Emotion, 5, 435–465.

Robinson, M. D., & Clore, G. L. (2002). Belief and feel-


ing: Evidence for an accessibility model of emotionalself-report. Psychological Bulletin, 128, 934–960.

Robinson, M. D., & Johnson, J. T. (1997). Is it emotionor is it stress?: Gender stereotypes and the perceptionof subjective experience. Sex Roles, 36, 235–258.

Robinson, M. D., Johnson, J. T., & Shields, S. A.(1998). The gender heuristic and the database: Fac-tors affecting the perception of gender-related differ-ences in the experience and display of emotions. Ba-sic and Applied Social Psychology, 20, 206–219.

Rose, A. J., & Rudolph, K. D. (2006). A review of sexdifferences in peer relationship processes: Potentialtrade-offs for the emotional and behavioral develop-ment of girls and boys. Psychological Bulletin, 132,98–131.

Roter, D. L., Hall, J. A., & Aoki, Y. (2002). Physiciangender effects in medical communication: A meta-analytic review. Journal of the American Medical As-sociation, 288, 756–764.

Rusting, C. L., & Nolen-Hoeksema, S. (1998). Regu-lating responses to anger: Effects of rumination anddistraction on angry mood. Journal of Personalityand Social Psychology, 74, 790–803.

Scherer, K. R., Wallbott, H. G., & Summerfield, A. B.(1986). Experiencing emotion: A cross-culturalstudy. New York: Cambridge University Press.

Schieman, S., & Van Gundy, K. (2000). The personaland social links between age and self-reported empa-thy. Social Psychology Quarterly, 63, 152–174.

Schienle, A., Schafer, A., Stark, R., Walter, B., & Vaitl,D. (2005). Gender differences in the processing ofdisgust and fear-inducing pictures: An fMRI study.NeuroReport, 16, 277–280.

Schmid Mast, M., & Hall, J. A. (2004). Who is the bossand who is not?: Accuracy of judging status. Journalof Nonverbal Behavior, 28, 145–165.

Seidlitz, L., & Diener, E. (1998). Sex differences in therecall of affective experiences. Journal of Personalityand Social Psychology, 74, 262–271.

Simon, R. W., & Nath, L. E. (2004). Gender and emo-tion in the United States: Do men and women differin self-reports of feelings and expressive behavior?American Journal of Sociology, 109, 1137–1176.

Stets, J. E. (1997). Status and identity in marital in-teraction. Social Psychology Quarterly, 60, 185–217.


Thayer, R. E., Newman, J. R., & McClain, T. M.(1994). Self-regulation of mood: Strategies for chang-ing a bad mood, raising energy, and reducing tension.Journal of Personality and Social Psychology, 67,910–925.

Timmers, M., Fischer, A. H., & Manstead, A. S. R.(1998). Gender differences in motives for regulatingcloseness. Personality and Social Psychology Bulle-tin, 24, 974–985.

Timmers, M., Fischer, A. H., & Manstead, A. S. R.(2003). Ability versus vulnerability: Beliefs aboutmen’s and women’s emotional behavior. Cognitionand Emotion, 17, 41–63.

Tobin, R. M., Graziano, W. G., Vanman, E. J., &Tassinary, L. G. (2000). Personality, emotional expe-rience, and efforts to control emotions. Journal ofPersonality and Social Psychology, 79, 656–669.

Vogel, D. L., Wester, S. R., Heesacker, M., & Madon, S.(2003). Confirming gender stereotypes: A social roleperspective. Sex Roles, 48, 519–528.

Vrana, S. R., & Rollock, D. (2002). The role of ethnic-ity, gender, emotional content, and contextual differ-ences in physiological, expressive, and self-reportedemotional responses to imagery. Cognition and Emo-tion, 16, 165–192.

Wager, T. D., & Ochsner, K. N. (2005). Sex differencesin the emotional brain. NeuroReport, 16, 85–87.

Wood, W., Christensen, P. N., Hebl, M. R., &Rothgerber, H. (1997). Conformity to sex-typednorms, affect, and the self-concept. Journal of Per-sonality and Social Psychology, 73, 523–535.

Zeman, J., & Shipman, K. (1998). Influence of socialcontext on children’s affect regulation: A functional-ist perspective. Journal of Nonverbal Behavior, 22,141–165.


C H A P T E R 2 5

The Cultural Psychologyof the Emotions

Ancient and Renewed

RICHARD A. SHWEDER, JONATHAN HAIDT, RANDALL HORTON,and CRAIG JOSEPH

Great, deep, wide and unbounded, the ocean is nevertheless drunk byunderwater fires; in the same way, Sorrow is drunk by Anger.

—Translation of an unidentified Sanskrit stanza from Indiain the early Middle Ages (GNOLI, 1956, p. 35)

This chapter elaborates, revises, and partiallyrecapitulates an evolving description of the cul-tural psychology of the emotions, versions ofwhich appeared in the first two editions of thishandbook. We define and illustrate a cultural/symbolic/meaning-centered approach to thestudy of the emotions, using some sources thatare quite ancient (e.g., the 3rd-century A.D.Sanskrit text, the “Ras�dhy�ya” of theN��ya��stra) and others that are quite new. Thechapter updates a componential approach tothe cultural study of emotions, with special at-tention to comparative analyses of two emo-tion categories: those often translated and la-beled in English as “anger” and “shame.” Thechapter also examines the moral context ofemotional functioning in different cultural andreligious traditions (e.g., Hinduism, Buddhism,and Islam), while suggesting that the character

and meaning of particular emotions are sys-tematically related to the ethics (e.g., the ethicsof autonomy, community, or divinity) prevalentin a cultural community (Haidt, 2001, 2003;Haidt, Koller, & Dias, 1993; Haidt & Joseph,2004; Jensen, 1995, 1998, 2005; Shweder,1990b, 1994a, 2002; Shweder & Haidt, 1993;Shweder, Much, Mahapatra, & Park, 1997).

The major goals of “cultural psychology”are to spell out the implicit meanings that giveshape to psychological processes, to examinethe distribution of those meanings across eth-nic groups and temporal–spatial regions of theworld, and to identify the manner of their so-cial acquisition. Related goals are to reassessthe principle of psychic unity or uniformity,and to develop a credible theory of psychologi-cal diversity or pluralism. The emphasis in cul-tural psychology is upon the way the human

409

mind can be transformed and made functionalin a number of different ways, which are notequally distributed across ethnic and culturalcommunities around the world (see Markus &Kitayama, 1991; Markus, Kitayama, &Heiman, 1998; Shweder, 1991, 1996; Shweder& LeVine, 1984; Shweder et al., 1998).

One hallmark of cultural psychology is aconception of “culture” that is symbolic andbehavioral at the same time. Culture, so con-ceived, can be defined as the range of ideasabout what is true, good, beautiful, and effi-cient that are made manifest in the speech,laws, customary practices, and other purposiveactions of the members of any norm-sensitiveand self-policing group (Goodnow, Miller, &Kessel, 1995; Shweder et al., 1998; Shweder,1999a, 1999b). In research on cultural psy-chology, “culture” thus consists of meanings,conceptions, and interpretive schemes that areactivated, constructed, or brought “online”through participation in normative social insti-tutions and routine practices (including lin-guistic practices) (see, e.g., D’Andrade, 1984;Geertz, 1973; LeVine, 1984; Miller, Potts,Fung, Hoogstra, & Mintz, 1990; Shweder,1991, 1999a, 1999b). According to this view, aculture is the subset of humanly possible oravailable meanings that, by virtue of encultur-ation (informal or formal, implicit or explicit,unintended or intended), has become valuedand active in giving shape to the psychologicalprocesses of the individuals in a particularnorm-sensitive group.

A second hallmark of cultural psychology isthe idea that interpretation, conceptualization,and other “acts of meaning” can take placerapidly, automatically, and un-self-consciously.Indeed, it is assumed that “acts of meaning”(e.g., the judgment that the human body maybecome polluted or desanctified because it is atemple for the soul; or that illness is a means ofempowerment because it unburdens a personof accumulated spiritual debts; or that shyness,shame, modesty, and embarrassment are goodand strong emotions because they are displaysof civility signaling that people are playingtheir part in upholding and controlling the so-cial order) can take place so rapidly, automati-cally, and un-self-consciously that from thepoint of view of an individual person they areindistinguishable from “raw” experience or“naked” consciousness itself (see, e.g., Geertz,1984, on “experience-near” concepts, andKirsh, 1991, on “thought in action”; see also

Fish, 1980; Nisbett & Wilson, 1977). Accord-ing to this view, many rapid, automatic, andun-self-conscious psychological processes arebest understood not as “pure,” “basic,” “fun-damental,” or “intrinsic” processes, but ratheras content-laden processes, which are contin-gent on the implicit meanings, conceptualschemes, and ideas that give them life (Haidt,2001; Markus et al., 1998; Mesquita, 2003;Nisbett & Cohen, 1995; Shweder, 1990a;Stigler, 1984; Stigler, Chalip, & Miller, 1986;Stigler, Nusbaum, & Chalip, 1988).

As an initial illustration of these points, webegin our discussion in the 3rd century A.D. inIndia with a brief examination of a Sanskrittext (the “Ras�dhy�ya” of the N��ya��stra)that was written relatively early in the his-torical record of systematic human self-consciousness about the emotions. It is throughan analysis of this venerable text—an ancientexample of a cultural psychology—that weaddress contemporary concerns. The “Ras-�dhy�ya” is a useful intellectual pole star onwhich to concentrate a discussion of the cul-tural psychology of the emotions, for three rea-sons: (1) The text, although ancient, comparesfavorably with any contemporary treatise onthe symbolic character of emotional experi-ence; (2) the text, although famous amongSanskritists and scholars of South Asian civili-zation, is hardly known at all by emotion re-searchers in anthropology and psychology; and(3) the text provides the opportunity for an ob-ject lesson about the universally appealing yet,in some sense, culture-specific character of allaccounts about what is “basic” to the emo-tional nature of human beings.

THE CULTURALLY CONSTRUCTED“BASIC EMOTIONS”OF THE “RAS�DHY�YA”

Between the 3rd and 11th centuries A.D.,Hindu philosophers of poetics and drama, in-terested in human emotions as objects of aes-thetic pleasure, posited the existence of eight ornine basic emotions (sth�yi-bh�va) and devel-oped a relatively detailed account of the sym-bolic structures that give them shape andmeaning. There is no standard English transla-tion of the Sanskrit terms for the postulated ba-sic emotions. Indeed, there is no agreementabout whether they should be translated as“emotions,” “mental states,” or “feelings,” or


about whether they should be translated as“basic,” “dominant,” “permanent,” “univer-sal,” “natural,” or “principal” emotions (ormental states or feelings). The eight basic (ordominant) emotions (or mental states or feel-ings) are variously translated as follows: (1)sexual passion, love, or delight (rati); (2)amusement, laughter, humor, or mirth (h�sa);(3) sorrow (�oka); (4) anger (krodha); (5) fearor terror (bhaya); (6) perseverance, energy, dy-namic energy, or heroism (uts�ha); (7) disgustor disillusion (jugups�); and (8) amusement,wonder, astonishment, or amazement(vismaya). Some early medieval commentatorsmention an additional basic (or dominant)emotion (or mental state or feeling), (9) seren-ity or calm (sama). To simplify our exegesis, werefer to the eight (or nine) as “basic emotions,”and we label them “sexual passion,” “amuse-ment,” “sorrow,” “anger,” “fear,” “persever-ance,” “disgust,” “wonder,” and “serenity.”

The canonical Sanskrit text on the “emo-tions,” attributed to Bharata, is the sixth chap-ter, the “Ras�dhy�ya,” of the N��ya��stra,which is a book about drama. In Sanskritdrama, the primary aim of the aesthetic experi-ence was psychological; indeed, it was the sym-bolic representation of emotional states per sethat set the stage for aesthetic and revelatoryexperience (see Dimock et al., 1974). The fa-mous sixth chapter of the N��ya��stra is aboutthe narrative structure (the causes, conse-quences, and concomitants) of eight basic emo-tional states and the most effective means (viafacial expression, voice, posture, setting, char-acter, action, and physiological response) oftheir representation in the theatre. TheN��ya��stra was probably written some timebetween the 3rd and 5th centuries A.D. Themost famous of several commentaries on thetext is by the 10th- and 11th-century KashmiriBrahman philosopher Abhivanagupta (partialtranslations and contemporary commentariescan be found in Masson & Patwardhan, 1970,and Gnoli, 1956; see also Dimock et al., 1974,and Keith, 1924).

THE WONDER OF THE SANSKRITEMOTIONS: A CULTURAL ACCOUNT

Contemporary non-Hindu researchers in theUnited States and Europe are likely to find theaccount of the “basic emotions” in the“Ras�dhy�ya” both familiar and strange. In-

deed, one of the hazards of doing research onthe emotions is the temptation to presump-tively universalize a content-laden and culture-specific mental process and theorize that it is abasic or intrinsic mental process. Here we findit instructive (and a useful corrective to un-bounded generalizations) to compare two suchposited theories about “basic emotions” acrosshistorical time and cultural space. If we com-pare the Sanskrit list of nine (eight plus one)basic emotions (sexual passion, amusement,sorrow, anger, fear, perseverance, disgust, won-der, and sometimes serenity) with Paul Ekman’swell-known contemporary list of nine (six plusthree) basic emotions (anger, fear, sadness, hap-piness, surprise, and disgust, plus interest,shame, and contempt), which Ekman (1980,1984) has derived from the analysis of every-day facial expressions, the two lists do notseem to us to be closely coordinated, althoughthey are not totally disjoint either.

In his volume Performance Theory, RichardSchechner (1988) presents a series of photo-graphs of facial expressions that he claims areiconic representations of the nine basic emo-tions of the N��ya��stra. This, of course, is arisky thing to do. The N��ya��stra never ab-stracts out facial expressions as the key mark-ers of the basic emotions, but rather treatsthem as one element in an array of constitu-ents; and there is every reason to believe that inHindu drama facial expressions unfold dynam-ically in a sequence of movements, which arenot easily frozen into a single frame (Hejmadi,Davidson, & Rozin, 2000). Nevertheless,Schechner posits direct analogies between sixof his facial expressions for the Sanskrit basicemotions and the six facial expressions fromEkman’s primary scheme—equating, for exam-ple, Ekman’s representation of the face of sur-prise with the Sanskrit face of wonder, andEkman’s representation of the face of happinesswith the Sanskrit face of sexual passion.Schechner thinks he sees a universal pattern re-flected in the two schemes: He states, “Human-kind has countless gods, but I would be verysurprised if there were not some agreementconcerning the basic emotions” (1988, p. 266).But how much agreement?

In our view, several of Schechner’s equationsare dubious. For example, in Ekman’s face ofsurprise, the mouth is wide open; it is not simi-lar to the mouth of the Sanskrit emotion ofwonder, which is closed and faintly suggestiveof a smile. (The mouth is closed in all of the fa-

25. The Cultural Psychology of the Emotions 411

cial expressions of the medieval Hindu emo-tions, which, we speculate, may be related to acultural evaluation concerning the vulgarity ofan open mouth.) And in Ekman’s photo of theface of happiness, the eyes are directly frontal;they are not similar to the eyes of the Sanskritemotion of sexual passion, where the gaze isconspicuously averted to one side, perhaps sug-gestive of coyness, secrecy, or conspiracy. Moreimportantly, because Schechner’s equation ofAmerican “happiness” with Sanskrit “sexualpassion” seems peculiar from the start, itshould also be noted that Ekman’s photo of theface of happiness bears no resemblance what-soever to the face of amusement, which is theSanskrit emotion one might have intuitively ex-pected to be connected to the Western concep-tion of “happiness.”

We doubt that most Americans could spon-taneously generate accurate descriptions forthe majority of the nine facial icons of the San-skrit “basic emotions” displayed in Schechner’sbook. (Curiously, one of the faces that Ameri-can graduate students seem to identify withoutmuch difficulty is the Sanskrit face of serenity,which as far as we know is not a “basic emo-tion” on any Western list. In informal experi-ments conducted in classes at the University ofChicago, they also converge in their responsesto faces of fear, disgust, and sorrow, but not tothe other five.) Indeed, we believe one can plau-sibly argue that happiness, surprise, and mostof the other basic emotions on Ekman’s list donot have close analogues among the basic emo-tions of the “Ras�dhy�ya,” and any sense ofeasy familiarity with the Sanskrit list is moreapparent than real.

As we read the “Ras�dhy�ya” and commen-taries, three of the nine basic emotions (anger,fear, and sorrow) are genuinely familiar, in thesense of possessing an equivalent shape andmeaning for medieval Hindus and contempo-rary Americans. Of course, to acknowledgethose three points of dense similarity is not tosuggest that those three emotional meaningsmust be cross-cultural universals. Wierzbicka(1992) has brought to a halt facile claims abouttranslation equivalence by showing that “sad-ness” as understood in European and Ameri-can conceptions of the emotions is not an em-pirical universal and is neither lexicalized,important, nor salient in most of the languagesof the world. She claims that from the point ofview of the study of the linguistic semantics ofemotion terms around the world, there may be

no basic or universal emotions, although sheallows that feelings that are more or less“shame-like” are quite widespread.

Nevertheless, anger, fear, and sorrow areeasy to recognize in the “Ras�dhy�ya.” Sorrow,for example, is said to arise from misfortune,calamity, and destruction, and from “separa-tion from those who are dear, [their] downfall,loss of wealth, death and imprisonment.” “Itshould be acted out by tears, laments, dryingup of the mouth, change of color, languor inthe limbs, sighs, loss of memory, etc.” (Masson& Patwardhan, 1970, p. 52). Sorrow is said tobe accompanied by other mental states, includ-ing world-weariness, physical weariness, life-lessness, tears, confusion, dejection and worry.Anger and fear are also easy to recognize in thetext (see Masson & Patwardhan, 1970, pp. 52–53).

For three of the nine basic emotions de-scribed in the “Ras�dhy�ya,” it is easy to recog-nize the underlying script, to readily see the selfin the other, and to arrive at a cross-culturaland transhistorical agreement about what isbasic in emotional functioning (at least forthem and us). Yet as one moves beyond sorrow,anger, and fear to disgust, amusement, wonder,perseverance, sexual passion, and serenity, theway in which consciousness is partitioned orhierarchically structured into basic andnonbasic states in the “Ras�dhy�ya” seems lessand less familiar, despite any initial appear-ances to the contrary. This decline in familiarityis similar to the “gradient of recognition” thatHaidt and Keltner (1999) found when studyingfacial expressions in India and the UnitedStates: Some expressions are very well recog-nized across cultures, some are less well recog-nized, and there is no neatly bounded set of“universal” facial expressions.

Thus it becomes clear upon examination ofthe relevant Sanskrit texts and commentariesthat medieval Hindu “disgust” overlaps withbut also differs from modern American “dis-gust.” Medieval Hindu disgust is partitionedinto two subtypes. The first includes aspects ofhorror and disillusionment, as well as world-weariness associated with the quest for detach-ment, transcendence, and salvation; the secondincludes horror at the sight of blood. MedievalHindu disgust is, as the anthropologist McKimMarriott has suggested to us, more like a do-main of the loathsome, and it gathers togetherwithin its territory a broad range of human re-sponses to the ugly, the nasty, and the odious.


It also becomes clear upon close examina-tion that nuances make a difference, and thatmedieval Hindu “wonder” is not contempo-rary American “surprise,” but rather a state ofmind closer to admiration than to startle orshock. For Hindu wonder has less to do with asudden violation of expectations and more todo with one’s reactions to the opportunity towitness divine, heavenly, or exalted feats,events, or beings (including, e.g., the amazingfeats of a juggler). It is even possible to do suchwitnessing with the mouth closed, as long asthe eyes are wide open!

Upon closer examination, it becomes appar-ent as well that medieval Hindu “amusement”(which includes contemptuous, indignant, orderisive laughter at the faults and inferior sta-tus of others) is not contemporary American“happiness,” which has celebratory implica-tions. Indeed, happiness, shame, indignation,arrogance, and some contempt-like emotionsare explicitly mentioned in the “Ras�dhy�ya”for inclusion among 33 nonbasic (“accompa-nying”) mental states. Thus it seems reasonableto assert that the basic emotion designated bymedieval Hindu philosophers as “amusement”is not adequately translated as “happiness” oras “contempt.” (It should be noted that whilethe text provides little basis for determiningequivalence of meaning for the terms used totranslate the 33 nonbasic mental states, there isgood reason to doubt that “shame” or “happi-ness” have the same implications and associa-tions, or play the same psychological role, inIndia as they do in the contemporary UnitedStates. See Menon & Shweder, 1994, andShweder, 1996, 2003, on the positive qualitiesof “shame” in India, where it is a virtue associ-ated with civility, modesty and an ability to reinin one’s destructive powers in support of the so-cial order rather than with the diminishment ofthe ego; see also Parish, 1991, and below.)

Similarly, it becomes clear upon examinationof the text that medieval Hindu “perseverance”is not contemporary American “interest,” butis rather deeply connected to heroic determina-tion and a willfulness to engage in acts requir-ing endurance and self-sacrifice. In the contextof the early medieval Hindu scriptures, whenthe Hindu goddess Durga (or Kali) endures tri-als and tribulations yet persists in a seeminglyhopeless battle against uncountable demons inan effort to save the world, her efforts are saidto display the heroic rasa of perseverance.Mere interest has very little to do with it. She

would probably rather be doing something else(see below).

In summary, the two lists of nine basic hu-man emotions closely and truly overlap at onlythree points. All the other apparent points ofsimilarity (amusement as happiness, their dis-gust as our disgust, wonder as surprise, perse-verance as interest) turn out to be merely ap-parent; and for several of the emotions (sexualpassion, serenity, shame, contempt), there isnot even an illusion of transcultural equiva-lence. In the end, most of the items cannot beeasily mapped across the two lists without agood deal being lost in translation.

There are other ways in which the “Ras-�dhy�ya” presents us with a somewhat unfa-miliar portrait of the way consciousness is or-ganized. According to the text and commentar-ies, the four primary basic emotions are sexualpassion, anger, perseverance, and disgust. Thefour secondary basic emotions are amusement,sorrow, wonder, and fear. The ninth basic emo-tion, serenity, is sometimes viewed as a primarybasic emotion and either substituted for disgustor associated with disgust (through a causal se-quence that begins with horror and revulsionover attachments in the world, and ends withthe serenity of ego alienation, detachment, andsalvation).

In commenting on this scheme, it is worthnoting that Sigmund Freud might find much ofvalue in a conception of human personalitythat treats sexual passion and anger (and perse-verance and disgust) as the deepest aspects ofhuman experience. One wonders whetherFreud would have interpreted perseverance anddisgust as analogues to the life and death in-stincts. More notable, however, is the fact thatthe primary basic emotions are thus named be-cause they are the “emotions” associated inclassical and folk Hindu thought with the fourworthy ends or goals of life. One of those goalsof life—pleasure (k�ma)—is linked to sexualpassion. A second goal—control, autonomy,and power (artha)—is linked to anger. A thirdgoal—social duty and moral virtue (dharma)—is linked to perseverance. The forth and per-haps highest goal—purity, sanctity, salvation,or the attainment of divinity (moksha)—islinked to disgust and/or serenity. In otherwords, presupposed by this famous formula-tion about the organization of human emotionsis a special and local theory of morality and hu-man motivation and a specific way of life. Thusit is hardly surprising that this particular medi-


eval South Asian conception of the hierarchicalstructuring of consciousness into basics versusnonbasics and primary basics versus secondarybasics should seem somewhat strange or aliento emotion researchers in North America,and vice versa. In other words, in the“Ras�dhy�ya” one finds a relatively elaborateaccount of the symbolic structures that giveshape and meaning to a selected subset of men-tal experiences, which, because they have beenprivileged for local symbolic elaboration, havebecome transformed into mental experiencesthat people regard as “basic” in their particularculturally constituted world.

COMPARING AMERICAN “ANGER”AND TIBETAN

The strategy adopted in the “Ras�dhy�ya” is todefine a basic emotion by the implicit symbolicstructure that gives shape and meaning to thatemotion, and then to define that symbolicstructure by resolving it into its determinants,consequences, and accompanying side effects.This strategy is directly parallel to various con-temporary approaches to the cultural psychol-ogy of the emotions.

One aspect of this symbolic (or “cognitive,”“interpretive,” or “meaning-centered”) ap-proach is the view that kinds of emotions arenot kinds of things like plants or animals. In-stead, they are embodied interpretive schemesof a particular script-like or narrative form thatgive shape and meaning to the human experi-ence of those conditions of the world that havea bearing on self-esteem (see Shweder, 1994b).The components that are proposed as slots inthese emotion schemes may vary slightly fromscholar to scholar, although most of the com-ponents or slots in use today can be found inthe “Ras�dhy�ya.”

Mesquita and Frijda (1992; see alsoEllsworth, 1991; Frijda, 1986; Lazarus, 1991;Lewis, 1989; Lewis, Sullivan, & Michalson,1984; Lutz, 1985; Russell, 1991; Stein & Levine,1987), for example, parse each emotion scriptinto a series of components including “anteced-ent events,” “event coding” (type of condition ofthe world), “appraisal” (judged implications forself-esteem and well-being), “physiological re-action patterns,” “action readiness,” “emo-tional behavior,” and “regulation.” Shweder(1994b) suggests a parsing of emotion scriptsinto components such as “self-involving condi-

tions of the world” (e.g., loss and gain, protec-tion and threat), “somatic feelings” (e.g., muscletension, pain, dizziness, nausea, fatigue, breath-lessness), “affective feelings” (e.g., agitation,emptiness, expansiveness), “expressive modes”(e.g., face, posture, voice), and “plans for self-management” (e.g., to flee, to retaliate, to cele-brate, to invest). (See also Shweder, 1991, wherea slot is provided in the emotion narrative forvariations in “social regulation” or the norma-tive appropriateness of certain emotions beingexperienced or expressed.)

The primary assumption of the symbolic ap-proach is that the “emotion” (e.g., sadness,fear, or love) is not something independent ofor separable from the conditions that justify it,from the somatic and affective events that areways of being touched by it, from the actions itdemands, or the like. The “emotion” is thewhole story: a kind of somatic event (fatigue,chest pain, goose flesh) and/or affective event(panic, emptiness, expansiveness) experiencedas a perception of some antecedent conditions(death of a friend, acceptance of a book manu-script for publication, a proposition to go outto dinner) and their implications for the self(e.g., as loss, gain, threat, possibility), and ex-perienced as well as a social judgment (e.g., ofvice or virtue, sickness or health) and as a kindof plan for action to preserve one’s self-esteem(attack, withdraw, confess, hide, explore). The“emotion” is the entire script. It is the unfold-ing experience of all the components, or, per-haps more accurately, the cohesive experienceof the whole package deal.

A second aspect of the symbolic approach isthe view that any “emotion” is decomposableinto its components, and that these compo-nents are what must be compared when we askwhether two emotion experiences are the sameor different across cultures. Based on earlierwork by Shweder (1994b) and recent work byHorton (2006) attempting to integrate perspec-tives on emotion from across the fields of cog-nitive psychology and psychological anthropol-ogy, we find it useful to posit eight relevantcomponents, which we describe below. Wethen illustrate the utility of this approach by us-ing it to compare the emotion of “anger” (as itis experienced and understood by a sample ofurban American adults) with the Tibetan emo-tion lung lang (as it is experienced and under-stood by Tibetan refugees [laity and religiousvirtuosos] settled in India). The eight compo-nents are as follows:


• Component 1: Somatic experience. Arepeople alike or different in introspectively andobjectively observable physical changes (e.g.,muscle tension, headaches, blood pressureshifts, activation of specific neural pathways)when they experience the emotion?

• Component 2: Affective phenomenology.Are people alike or different in their affectiveexperiences (e.g., feelings of emptiness, calm,pleasantness, derealization, soul loss) whenthey experience the emotion?

• Component 3: Environmental determi-nants. Are people alike or different in the ante-cedent conditions associated with the emotion(e.g., winning the lottery, a remark from a sub-ordinate, birth of a child, physical contact witha member of an outcaste group)?

• Component 4: Appraisals of significance.Are people alike or different in the appraisalsof the antecedent conditions that elicit the emo-tion, and in ongoing construals that may in-flect, extend, transform, or truncate the experi-ence (e.g., others’ actions were intentional,unwanted, goal-enhancing, expected, disre-spectful, or status-degrading; the outcome canor cannot be changed)?

• Component 5: Normative social apprais-als. Are people alike or different in the extentto which showing, displaying, or merely expe-riencing the emotion has been socially desig-nated as a vice or virtue or as a sign of sicknessor health?

• Component 6: Self-management. Are peo-ple alike or different in the impulses to actionand plans for self-management that get acti-vated in association with the emotion (e.g., tocelebrate, to attack, to disengage and avoid theother person, to engage in problem solving)?

• Component 7: Communication and sym-bolization. Are people alike or different in theiconic and symbolic vehicles used for giving ex-pression to the emotion (e.g., facial expres-sions, voice, posture, and action)?

• Component 8: Social management. Arepeople alike or different in the ways they re-spond to and manage the communication andsymbolization of the emotion by others (e.g.,empathically mirroring the emotion, cowering,withdrawing, discussing an individual’s behav-ior with others, collectively shunning the indi-vidual)?

Depending on the interests and methodolog-ical commitments of investigators, any of theseeight domains can be elaborated further. In ex-

tending the seventh component domain (com-munication and symbolization), for instance,one could ask: Are the cultures alike or differ-ent in the symbolic resources they accord theirmembers for naming, evoking, and manipulat-ing discrete facets of the emotional experiencefor the achievement of important social and in-dividual goals (e.g., through meditations oncompassion, death metal concerts, workshopson assertiveness training or anger management,mass political demonstrations, initiation rites,or vulnerability to dissociative states in whichthe emotion is prominent)?

We recognize that this componential modelincludes facets of emotion-related experiencethat many psychological researchers might re-sist including in a conceptual or analytic defini-tion of emotion. From the perspective of thehybrid symbolic/interpretive/meaning-centeredview of emotion that we are advancing, emo-tional experience is not analytically dissolublefrom either the conditions that justify it orthe social meaning systems that sustain it.This model offers a context-rich, maxi-mally inclusive characterization of emotionalexperience—one in which elements ofsociocultural and linguistic context provide thenecessary background against which one canperceive local variations and transformationsof the figural center of emotive processes.

The model can provide a useful frameworkfor comparing emotional experiences not justacross cultures, but within cultures as well.Horton (2006) has recently used the model tocompare “anger” across three groups of indi-viduals: a sample of American adults living in amixed urban ethnic community, a sample of layTibetan refugees living in long-term settlementsin India, and a sample of Tibetan Buddhistmonks and nuns living in the same Indian com-munities.

The Tibetan emotion term lung lang (Wylie,1959, gives it as rlung langs) is used by allclasses of Tibetans living in the exile settle-ments of South Asia. It has a denotativebreadth similar to that of the English “anger,”and almost all modern bilingual dictionariesrender it in English as simply “to get angry.”The Tibetan expression, however, is actually aconceptual composite derived from twolexemes. The term rlung, which denotes thewind humor in the Tibetan ethnomedical sys-tem, is joined with the intransitive verb langspa, which means “to rise.” The combined ex-pression lung lang thus invokes an underlying


psychophysiological model of the emotion as arising movement of the wind that animatesconsciousness, upward from the chest. This un-derlying model articulates with cultural under-standings of the sources of vulnerability tochronic anger, the expected long-term effects,and the phenomenology of the experience ofthe emotion.

In the following discussion, we compare an-ger and lung lang, using the eight componentslisted above to reveal a complex pattern of sim-ilarities and differences. We begin with the do-main of normative social appraisals. We do sobecause we believe that examining the moraland ethical construals of anger in Americanculture versus lung lang in Tibetan culture pro-vides a crucial background for understandingobservations in the other component domains.

Component 5: Normative Social Appraisals

Tibetan and American respondents were asked,“In general, if you think about anger/lung lang,do you think of it as a good or a bad thing?”They were then asked, “For what reasons is itgood or bad?” Citing views grounded inMahayana Buddhist ethical and metaphysicalthought, Tibetan respondents, both lay andclerical, unequivocally viewed lung lang asmorally bad. They assimilated it to the senti-ment she dangs (anger/hatred), one of the“three moral poisons” (dug gsum) that arecommonly accepted by Buddhists to be the rootsources of all suffering for sentient beings.Americans, by contrast, viewed anger as a mor-ally ambivalent, neutral, or natural process. Al-though Americans recognized the potentialharmful effects of anger for others, they wereless likely than Tibetans to insist on its harmfuleffect for the person who experiences it. In-deed, Americans emphasized several positiveaspects of anger: It gives one an energy that canbe used in a positive way; it can lead to prob-lems’ being addressed that might otherwisepersist; it can be beneficial to society. Tibetans,by contrast, viewed lung lang as a fundamen-tally destructive sentiment, equally harmful toself and others. They viewed it as arising froman intrinsically flawed motivational state (a de-sire to harm another sentient being) and gener-ative of ultimately bad results. Reflecting uponmetaphysical understandings of karma, theyinsisted upon the symmetry of lung lang‘s ill ef-fects for all parties involved.

In response to the question “If a person getsangry a lot, over and over again, what kinds ofthings might happen to that person?”, Tibetansand Americans were alike in predicting adversesocial and health effects. In fact, the most com-mon metaphorical expressions for the antici-pated adverse social effects of chronic anger/lung lang in the two cultures were identical. Ti-betans and Americans agreed that people will“become more distant” (Tibetan: thags ring pochags) from the chronically angry individual.For Tibetans, these predicted social effectstended to involve community-level judgmentsand processes.

Although norms and expectancies were rela-tively easy to compare across cultural groups,comparing the two emotions across many ofthe remaining component domains required asomewhat different procedure. The researcherasked American and Tibetan respondents todiscuss in detail a recent situation in which theyhad felt anger/lung lang. The interviewerprobed for background information on the cir-cumstances in which an incident occurred, whowas involved, and particular judgments andappraisals that might have guided respondentsas they felt the emotion. The interview assessedtheir subjective physical and affective feelingsin the situation, as well as their fantasies, ac-tions and impulses, the reactions of bystandersand other individuals in the situation, and theeventual resolution of the situation. Coded andscored, these data provided the basis for sys-tematic tests of differences across samplegroups.

Component 1: Somatic Experience

Tibetan lung lang and American anger dis-played considerable overlap in the domain ofsomatic experience. Feelings of tension, shak-ing/nervousness, and heat were reported ascommon somatic feelings experienced with an-ger/lung lang in all three groups. Americans,however, produced a broader, more detailedrange of descriptions of physical feelings asso-ciated with anger.

In terms of the long-term anticipated so-matic effects of anger/lung lang, individualsfrom both cultures predicted bad health effectsfor chronically angry individuals. Some pre-dicted effects were common across all threegroups (e.g., heart disease, blood pressureproblems) while some differed. Many Tibetans


predicted that such individuals would sufferfrom srog lungs na tsha (literally, a life-wind ill-ness), a serious condition defined in Tibetanethnomedical tradition. Several Tibetan re-spondents also asserted that the chronically an-gry individuals would be likely to die prema-turely. Given the list of illnesses that Americansassociate with chronic anger, it likewise wouldhave been logically consistent for Americans toconnect chronic anger directly with prematuremorbidity. Yet no American respondents madethis connection.

Component 2: Affective Phenomenology

In contrast with other groups, several individu-als from the Tibetan clerical sample describedthe experience of anger in dissociative terms(e.g., “It felt as though I were drunk or crazy atthe time”). When asked, “When that situationhad just ended, how did you feel?”, Americanswere far more likely than were Tibetans to re-port feelings of lingering anger. Tibetans (par-ticularly the Buddhist clergy) were more likelythan Americans to report feeling a host of otherdysphoric emotional states at the end of the an-ger incidents. These included emotions similarto the sentiments lexicalized in English as “re-gret,” “shame,” and “unhappiness.”

Lingering differences in the encoding andretrieval of memories of experiences of anger/lung lang were suggested by the fact thatwhen respondents were asked, “If you thinkabout that situation now, do you still feel alittle angry?”, Americans were much morelikely than Tibetans were to say that they stillfelt angry when recalling the original situa-tions. Americans did so whether the originalfeelings of anger were intense or mild. For Ti-betans, the likelihood of feeling lung langupon recalling the situation appeared drivenby the strength of the feelings of lung lang inthe original situation. Only in situationswhere original feeling of lung lang had beenstrong were Tibetan respondents likely to feelanger on recall.

Component 3: Environmental Determinants

The failure to meet obligations and disrespect-ful treatment by others were among the mostcommon provocations to anger/lung lang inboth cultures. For Tibetans, the experience ofpublic criticisms and teasing (kyag kyag) that

had gotten out of hand played a disproportion-ate role as provocations. For Americans, awaste of the respondents’ time served as a morecommon provocation. Tibetans reported par-ticular difficulties with outgroup incidents ofanger, reflecting tensions between the them-selves and members of the local ethnically In-dian communities where they how live. So-cially, lung lang incidents displayed anasymmetric, hierarchical character not appar-ent in American anger incidents. Lay Tibetanrespondents reported no incidents of feeling theemotion toward Buddhist monks or nuns, yetTibetan monks and nuns readily reported suchfeelings toward lay individuals.

Component 4: Appraisals of Significance

Respondents from both cultures showed a re-luctance to attribute a deliberate intent to harmthe other party in the incidents they described.Across all groups, however, respondents ratedtheir feelings of anger/lung lang as strongerwhen they said they had made such an attribu-tion. Cultural differences in at least one varietyof secondary appraisals were evident as well.Tibetans, both lay and clergy, were much lesslikely than Americans to judge the other per-son’s provocative actions as typical or usual forthat person. American respondents, by con-trast, tended—chronically and spontaneouslyin open narratives—to connect the other per-son’s provocative behavior in the current situa-tion with the person’s past behavior, and to as-sert that a dispositional pattern existed for theindividual to act in that way. This attributionbias is consistent with prior cross-cultural re-search on the fundamental attribution error (cf.Shweder & Bourne, 1984; Miller, 1984). Col-lectively, this line of emerging research suggeststhat Americans make character-based, endur-ing dispositional attributions far more fre-quently than members of certain other culturalgroups do.

Component 6: Self-Management

Tibetans were much more likely than Ameri-cans to believe that anger/lung lang can be pre-vented and even permanently transcended. Fur-thermore, many were able to point toindividuals whom they believed had achievedsuch a state. Americans, by contrast, doubtedwhether anger-free living was either possible or


desirable. When the hypothetical question“What kind of person would it be who neverbecomes angry?” was posed, Americans of-fered responses like these: “People who don’tshow it and then one day they explode,” or“Maybe someone who was severely abused asa child.” Some rejected the question outright,saying, for example, “We shouldn’t be talkingabout this like it’s a good thing.” American re-spondents thus actively pathologized thehypoexpression of anger; in contrast, the no-tion that the absence of feelings of anger couldbe pathological was rejected quite thoroughlyby the Tibetans.

Although Americans and Tibetans endorseddifferent ideal strategies for managing the emo-tion, in practice they appeared similar in manyof the action tendencies and behaviors they re-ported engaging when the emotional experi-ence had been triggered. Actions ranged fromtaking time out, practicing patience (a set ofspecific Tibetan Buddhist techniques), andseeking mediation, to issuing open criticisms orthreats and (in some cases) exchanging blowswith the other party. Tibetans reported a signif-icantly shorter duration for feeling anger/lunglang than Americans in the incidents they de-scribed. Tibetan clergy reported significantlyless intensity of anger/lung lang feelings thanAmericans or lay Tibetans.

Component 7: Communicationand Symbolization

The two cultures accord their members radi-cally different resources for naming, evoking,and manipulating discrete facets of the experi-ence of anger/lung lang. If one considers the di-verse American social practices in which angerplays a central role (e.g., in spectacle entertain-ments like The Jerry Springer Show, deathmetal concerts, or professional wrestling; intherapeutic contexts like psychotherapy groupsfor children of alcoholics or anger managementclasses; or in diagnostic categories like inter-mittent explosive disorder), the ambivalentquality of the American view of anger is appar-ent. These widely differing American culturalpractices offer individuals varied opportunitiesto engage in expressing, channeling, harness-ing, directing, and controlling anger.

Tibetans, by contrast, possess a conceptuallyrich and elaborate tradition of Buddhist ideasand ethical practices—such as the mind-training (blo sbyong) tradition—and a set of

cultural institutions and rituals dedicated to thegoal of eliminating or transforming lung lang.Heavy metal and gangster rap have not caughton with Tibetan settlement youths, amongwhom performers like the Backstreet Boys andBryan Adams represent the transgressive edgeof global youth culture impingements. Concep-tual resources and social practices in Tibetanexile society reflect an unequivocal moral con-demnation of anger/lung lang and related senti-ments. Particularly through the institutionalstructures and practices of Buddhist monasti-cism, resources in Tibetan exile society are ded-icated to transforming, calming, preventing,and extinguishing lung lang, rather than chan-neling, cultivating, harnessing, expressing anddirecting it effectively.

Component 8: Social Management

Consistent with the normative ethical rejectionof anger in Tibetan culture, during incidents inwhich anger/lung lang was openly expressedand witnesses were present, Tibetan bystanderswere more likely to show disapproval of opendisplays of anger than were American bystand-ers.

These comparisons have been extended withethnographic and ethnolinguistic data (seeHorton, 2006). We believe that they illustratethe value of adopting a componential, sym-bolic/interpretive model of emotions when oneis seeking to compare emotional experiencesacross cultural groups.

BITE YOUR TONGUE:THE CASE OF HINDU

When emotions are analyzed in terms of theirconstituent components, the issue of transla-tion equivalence for mental states becomes amatter of pattern matching. One tries to deter-mine whether the variables in each of thosecomponent slots are linked in similar waysacross cultures. One benefit of this approach isthat it makes it possible to elucidate the waythe abstract conceptual or definitional core ofany particular emotion takes on a culture-specific character in different historical tradi-tions and is associated with a somewhat differ-ent set of mental states across cultural groups.Consider, for example, the contemporaryHindu conception of lajja (or lajya), which has


been explicated for two communities in SouthAsia: the Newars of Bhaktapur in Nepal (Par-ish, 1991), and the Oriyas of Bhubaneswar inOrissa, India (Menon & Shweder, 1994, 1998;Shweder, 2004). Lajja is often translated by bi-lingual informants and dictionaries as though itwere equivalent in meaning to the Englishword “shame”; it is also sometimes translatedas though it meant the same thing as the Ameri-can English words “embarrassment,” “shy-ness,” “modesty,” or “coyness.” Yet, as shouldbecome obvious from the following bit of cul-tural exegesis, the translation of the meaning ofmental states across languages and cultures is afar more subtle and hazardous process thanmany suppose.

For starters, somewhat unlike the meaningof “shame” current in contemporary Anglo-American circles, lajja is something one delib-erately shows or puts on display the way wemight show our “gratitude,” “loyalty,” or “re-spect.” It is a state of consciousness that hasbeen elevated in South Asia as a supreme vir-tue, especially for women, and it is routinelyexhibited in everyday life—for example, everytime a woman covers her face or ducks out of aroom to avoid direct affiliation with thosemembers of her family she is supposed toavoid. Parish (1991, p. 324) describes it as anemotion and a moral state. It is by means oftheir lajja that those who are civilized upholdthe social order by showing perseverance in thepursuit of their own social role obligations; bydisplaying respect for the hierarchical arrange-ment of social privileges and responsibilities;by acting shy, modest, or deferential and notencroaching on the prerogatives of others; orby covering one’s face, remaining silent, or low-ering one’s eyes in the presence of superiors.Like gratitude, loyalty, or respect, lajja (whichis a way of showing one’s civility and commit-ment to the maintenance of social harmonythrough displays of respectful restraint) isjudged in South Asia to be a very good thing.

While lajja may be experienced by both menand women, it is an emotion and a virtue asso-ciated with a certain feminine ideal. It is talkedabout as a lovely ornament worn by women.Lajja is the linguistic stem for the name of a lo-cal creeper plant (a “touch-me-not”), which isso demure that upon the slightest contact itcloses its petals and withdraws into itself. Tosay of a woman that she is full of lajja is a verypositive recommendation. Here is one reasonwhy.

Perhaps the most important collective repre-sentation of lajja in various regions of easternIndia is the Tantric icon portraying the mothergoddess Kali, brandishing weapons and a de-capitated head in her 10 arms, eyes bulging andtongue out, with her foot stepping on the chestof her husband, the god Siva, who is lying onthe ground beneath her. Based on interviewswith 92 informants in Orissa, India, Menonand Shweder (1994, 1998, 2003) have exam-ined the meaning of this icon and its signifi-cance for our understanding of lajja.

The gist of the story, as it is narrated by localexperts, is that once upon a time the male godsgave a boon to a minor demon, Mahisasura, tothe effect that he could only be killed at thehands of a naked female. They thereby turnedMahisasura into a major demon capable ter-rorizing all the male gods. In order to destroythe demon, the male gods pooled all their en-ergy and powers to create the goddess Durgaand arm her with their own weapons. On theirbehalf, they sent Durga into battle againstMahisasura, but they neglected to tell herabout the boon. She fought bravely, but couldnot kill the demon; he was too strong andclever. In desperation, Durga appealed forguidance from an auspicious goddess, who lether in on the secret. As one informant narratedthe story:

So Durga did as she was advised to [she stripped],and within seconds after Mahisasura saw her [na-ked], his strength waned and he died under hersword. After killing him a terrible rage enteredDurga’s mind, and she asked herself, “What kindsof gods are these that give to demons such boons,and apart from that what kind of gods are thesethat they do not have the honesty to tell me thetruth before sending me into battle?”

Durga felt humiliated by her nakedness andby the deceit. She decided that such a worldwith such gods did not deserve to survive; shetook on the form of Kali and went on a madrampage, devouring every living creature thatcame in her way. The gods then called on Siva,Kali’s husband, to do something to save theworld from destruction at the hands of themother goddess. Siva lay in her path as shecame tramping along, enraged. Absorbed inher wild dance of destruction, Kali accidentallystepped on Siva and placed her foot on her hus-band’s chest, an unspeakable act of disrespect.When she looked down and saw what she had


done she came back to her senses—in particu-lar to her sense of lajja, which she expressed bybiting her tongue between her teeth. She reinedin her anger and became calm and still. To thisday in Orissa, India, “Bite your tongue” is anidiomatic expression for lajja, and the biting ofthe tongue is the facial expression used bywomen as an iconic apology when they realizeor are confronted with the fact that they havefailed to uphold social norms.

One moral of the story is that men are inca-pable of running the world by themselves, eventhough they are socially dominant. They relyon women to make the world go ’round. Yet ina patriarchal society, men humiliate women bythe way they exploit female power, strength,and perseverance. This leads to anger or rage inwomen, which is highly destructive of every-thing of value and must be brought under con-trol for the sake of the social order. Lajja is asalient ideal in South Asia, because it preservessocial harmony by helping women to swallowtheir rage.

If we decompose lajja into its component do-mains, it becomes apparent just how hazardousit can be to assume that one can render theemotional meanings of others with terms fromour received English lexicon for mental states.One is reminded here of Geertz’s discussion(1984, p. 130) of the difficulties of translatingthe Balinese term lek. Balinese lek seems muchlike Hindu lajja. Geertz notes that lek has beenvariably translated and mistranslated, and that“shame” is the most common attempt. He triesto render it as “stage fright.” Hindu lajja doesnot map well onto words like “shame,” “em-barrassment,” “shyness,” “modesty,” or even“stage fright.” An analysis of the constituentsof lajja helps us see why.

The normative social appraisal of lajja, forexample, is quite positive. To be full of lajja isto possess the virtue of behaving in a civilizedmanner and in such a way that the social orderand its norms are upheld. It is not a neurosis,and it does not connote a reduction in thestrength of the ego. Indeed, lajja promotes self-esteem. To experience lajja is to experience thatsense of graceful submission and virtuous,courteous, well-mannered self-control that ledKali to rein in her rage. One might try to renderits meaning as “respectful restraint.”

The environmental determinants of lajja as asense of one’s own virtue and civility are as var-ied as the set of actions that are dutiful and re-sponsible, given one’s station in life in a world

in which all people are highly self-consciousabout their social designation (see Geertz,1984, for a brilliant attempt to capture the dra-matic qualities of such a world). They includeevents that we would find familiar (not beingseen naked by the wrong person in the wrongcontext), as well as many events that mightseem alien or strange (never talking directly toone’s husband’s elder brother or to one’s father-in-law; never being in the same room with bothone’s husband and another male to whom hemust defer).

In terms of self-management strategies,South Asian lajja may appear at first glance tobe similar to American shame or embarrass-ment. It activates a habit or routine that some-times results in hiding, covering up, and with-drawing from the scene. Yet what is reallybeing activated by lajja is a general habit of re-spect for social hierarchy and a consciousnessof one’s social and public responsibilities,which in the context of South Asian norms maycall for avoidance, silence, withdrawal, orother deferential, protective, or nonaggressivegestures and actions.

Finally, consider the semantic structure ofthe American English word “shame” and theIndian Oriya word lajja in the minds of par-ticular informants. When middle-class Anglo-American college students are presented withthe triad of terms “shame–happiness–anger”and asked, “Which is most different from theother two?”, they are most likely to respondthat either “happiness” or “shame” is mostdifferent from the other two, perhaps on thehedonic grounds that “shame” and “anger”go together (in contrast to “happiness”) be-cause they are both unpleasant feelings, orthat “happiness” and “anger” go together (incontrast to “shame”) because they are bothego-expanding emotions. Neither response istypical of responses in the South Asian com-munity where Menon and Shweder (1994)have worked, where lajja (shame?) and suka(happiness?) are thought to go together in thetriad test, and raga (anger?), perceived Ti-betan-like as destructive of society, is viewedas the odd emotion out. Here somethingseems to be amiss in the translation process.Something may well have been amiss in mostpast attempts to equate emotions across lan-guages and across local cultural worlds (seeLutz, 1985, 1988; Lutz and White, 1986;Rosaldo, 1984; Wierzbicka, 1990, 1992,1997).


THE SOCIAL AND MORAL CONTEXTOF EMOTIONAL EXPERIENCE

The cases of lung lang and lajja illustrate thedependence of emotional experience on its so-cial and moral context. To understand lajja, forexample, one must understand the moral goodsthat Oriyas strive to achieve. This strategy ofviewing emotions against the background oftheir associated moral goods can be extendedto other emotions by using a framework thathas proved useful in cultural psychologicalwork. Shweder et al. (1997; see also Shweder,1990b; Haidt, 2001; Haidt & Joseph, 2004;Haidt et al., 1993; Jensen, 1995, 1998, 2005)suggest that moral goods do not vary randomlyfrom culture to culture, but rather tend to clus-ter into three sets of related goods or three eth-ics, known as the ethics of “autonomy,” theethics of “community,” and the ethics of “di-vinity.” Cultures rely upon the three ethics tovarying degrees. The relative weights of thethree ethics within a culture appear to affectthe experience and expression of emotion, aswell as the way emotions are conceptualized byboth local folk and local experts.

In cultures that emphasize an ethic of auton-omy, the central object of value is the individ-ual conceptualized as a preference structure.Within that type of cultural world, the most sa-lient moral goods are those that promote theautonomy, freedom, and well-being of the indi-vidual, with the result that nothing can be con-demned that does not demonstrably harm oth-ers, restrict their freedom, or impinge on theirrights. Haidt et al. (1993), for example, foundthat American college students (a populationsteeped in the ethics of autonomy) respondedto stories about violations of food and sexualtaboos (e.g., eating one’s already dead pet dog)with disgust. Nevertheless, these students feltcompelled by the logic of their ethical stance toseparate their feelings of disgust from theirmoral judgments. As a result, they held firmlyto the view that their personal emotional reac-tions did not imply that the actions werewrong. They spoke exclusively in the languageof the ethics of autonomy, pointing out that no-body was hurt, and that the people involvedhad a right to do as they pleased in a privatesetting. Disgust plays an ambiguous role insuch an autonomy-based cultural world (seeRozin, Haidt, & McCauley, Chapter 47, thisvolume). In such a world, the moral domain islimited to issues of harm, rights, and justice

(Turiel, 1983), and the emotions that areexperienced as moral emotions (e.g., anger,sympathy, and guilt) are those that respond to arather narrow class of ethical goods (e.g., jus-tice, freedom, and the avoidance of harm). Insuch a world, the focus of ordinary folk and so-cial scientists alike is upon individuals’ strivingto maximize their personal utility (e.g., Laza-rus, 1991; Plutchik, 1980; Stein, Hernandez, &Trabasso, Chapter 35, this volume). Happi-ness, sadness, pride, and shame are viewed asresponses to individual gains and losses, suc-cesses and failures. Other moral goods (such asloyalty, duty, and respect for status) that mightbe linked to the emotions are either lost orundertheorized.

Nevertheless, in many parts of the world themoral domain has been constructed in such away that it is broader than, or at least differentfrom, an ethic of autonomy. In cultures thatemphasize an ethic of community, ontologicalpriority is given to collective entities (the fam-ily, guild, clan, community, corporation, or na-tion), and the central moral goods are thosethat protect these entities against challengesfrom without and decay from within (e.g.,goods such as loyalty, duty, honor, respectful-ness, chastity, modesty, and self-control). Insuch a world, individual choices (what to wear,whom to marry, how to address others) take ona moral significance (Shweder, Mahapatra, &Miller, 1987), and the successful pursuit of in-dividual goals may even be a cause for embar-rassment or shame. Haidt et al. (1993), for ex-ample, found that outside of college samples,people of lower socioeconomic status generallythought it was morally wrong to eat one’s al-ready dead pet dog or to clean one’s toilet withthe national flag. Even when these actions werejudged to be harmless, they were still seen asobjectively disgusting or disrespectful andhence as morally wrong. In a cultural worldbased on an ethics of community, emotionsmay exist that are not fully felt by those whosemorality is based on an ethics of autonomy.Lajja is a clear example, since it is not the typeof emotion that will be experienced in a worldthat sees hierarchy and the exclusive preroga-tives of others as unjust or as a form of oppres-sion, rather than as a powerful and legitimateobject of admiration and/or respect (Menon &Shweder, 1998).

Similarly, emotions related to honor and her-oism may require a strong attachment and ded-ication to a collectivity or group, for whom


the hero lays down his or her life. TheN��ya��stra’s otherwise puzzling inclusion ofperseverance or heroism as a basic emotion,equal to anger and fear, seems more intelligibleagainst the backdrop of the ethics of commu-nity. A James Bond-type hero may display per-severance as he battles to save the “freeworld,” yet we do not think he inspires thesame emotional experience in an American au-dience that an Indian audience savors when aHindi film hero battles to avenge the death ofhis father. Many older classic American filmsraised themes of family honor, but such themeshave become less common in recent decades, asthe ethics of autonomy has pushed back theethics of community. Unlike Hindi films, mod-ern American films rarely embed the hero inthe thick traditions and obligations of familyhistory. It is a rare movie indeed when we meetthe hero’s parents.

The third ethic, the ethic of divinity, mayhave a similar differential activation and en-abling effect on emotional life. In the ethic ofdivinity, people (and sometimes animals) areseen as containing a bit of God (or a god) with-in them, and the central moral goods are thosethat protect and dignify the person’s inherentdivinity. The body is experienced as a temple,so matters that seem to be personal choiceswithin the ethics of autonomy (e.g., food andsexual choices, personal hygiene) becomemoral and spiritual issues associated with suchgoods as sanctity, purity, and pollution. Givensuch an ethic, cleanliness is indeed next to god-liness.

Within the terms of a cultural world focusedon an ethic of divinity, even love and hate maylose their simple positive versus negativehedonic valences. A modern spiritual guide forHindus (Yatiswarananda, 1979, p. 187) saysthat hatred and attachment are both fettersthat “degrade the human being, preventing himfrom rising to his true stature. Both must be re-nounced.” Hindu scriptures contain many sto-ries about people who come to feel disgust attheir own greedy and carnal attachments. Thisdisgust helps them to renounce their attach-ments, and the renunciation floods them with apositive feeling. Of course, the very idea of anemotion connected with renunciation seemsparadoxical, since spiritual progress in manyEastern religions is measured by the degree towhich one moves beyond the experience ofemotions. Only once this paradox is graspeddoes the mysterious ninth emotion of the

N��ya��stra—serenity—begin to make sense,along with its otherwise puzzling textual asso-ciation with disgust. Serenity or calmness is animportant part of Hindu emotional life andemotional discourse precisely because of thecentrality of an ethics of divinity in everydayHindu life. Not surprisingly, it is on no Westernlists of basic emotions.

THE MORAL CONTEXTOF EMOTIONAL EXPERIENCEIN ISLAMIC THOUGHT

Emotions can be moralized in various ways.One can formulate rules for the experiencing orexpression of an emotion; one can make themthe subject of obligations; one can frame theirethical significance in terms of their conse-quences; and so on. But perhaps the most pow-erful and pervasive way of moralizing emotionsis by linking them with virtues—morally goodtraits or states of character.

The classic Western statement of the rela-tionship between emotions and virtues is Aris-totle’s formulation:

. . . fear and confidence and appetite and angerand pity and in general pleasure and pain may befelt both too much and too little, and in both casesnot well; but to feel them at the right times, withreference to the right objects, towards the rightpeople, with the right motive, and in the rightway, is what is both intermediate and best, andthis is characteristic of virtue. (NichomacheanEthics, 1106b)

For Aristotle, emotions are not morally signifi-cant per se, but rather with respect to theiralignment with ethical imperatives indexed bysuch phrases as “the right times,” “the rightobjects,” and so on. Virtue consists in largepart of being “properly affected,” which is aproduct of moral training and habituation.1

The great moral and religious traditions havebeen especially concerned with the moral signifi-cance of emotions. Here we briefly examine Is-lam’s long-standing “hypercognition” (Levy,1984) of the virtues and highly elaborated con-cern with the development of character, which isstill very much alive and kept in the mainstreamof everyday Islamic discourse through the foun-dational texts of the religion (the Qur’an and thehadith, or sayings of the Prophet), as well asthrough the many contemporary popular worksof devotional literature that deal with ethics.2


An understanding of Islamic morality mightusefully begin with the concept of akhlaq. Forthe vast majority of Muslims, morality or eth-ics is simply akhlaq, and akhlaq is almost uni-versally the translation offered by Muslims forboth “morality” and “ethics.” This is signifi-cant for a study of virtue concepts in Islam be-cause of the etymology of akhlaq and its re-lated words. Akhlaq comes from the Arabicroot kh-l-q, which means “to create, shape,make, form, or mold.” Akhlaq is the pluralform of khuluq, which denotes “an innate pe-culiarity, natural disposition, character, or na-ture.” The Arabic translation of “ethics” in thesense of a philosophical discipline is ‘ilm al-akhlaq, or “science of akhlaq.” Thus, whilethere are several Arabic words that translate“morality,” the predominant one reflects anidentification of morality and ethics with hu-man nature as authored or divinely createdwith a specific telos, function, end, or purposein mind that is definitive of the full realizationof one’s nature.

Since the time of Muhammad, Muslims haveseen Islam as being, in some significant mea-sure, a means for the transformation of emo-tional life from one of ignorance and back-wardness to one appropriate to a divinelycreated nature. This theme has been elaboratedthroughout Islamic history by some of its great-est philosophers, some of whom even adoptedAristotelian ways of thinking about emotionand virtue. For example, the great Muslim phi-losopher al-Ghazali offered a typology of thevirtues organized around three cardinal virtues:wisdom, courage, and temperance (Sherif,1975). Table 25.1 shows al-Ghazali’s typology.The translations of some of al-Ghazali’s termsmay seem somewhat stilted, being literal trans-lations without the nuance that comes withcontext. But one thing that is immediatelystriking is how many of the virtues listed hereare directly concerned with emotional statesand their management (we have put them inbold type). This is especially the case with thecategory connected with “temperance,” whichis perhaps not surprising, since temperance ingeneral connotes moderation.

al-Ghazali was also a Sufi—a Muslimmystic—and thus was especially concernedwith the fusion of the mind and the soul, andwith the practices through which perfect Gnos-tic communion with God could be achieved.Naturally, one of the aspects of the self thatmust be worked on in pursuit of this goal is

emotional experience, and elsewhere in hiswork al-Ghazali elaborates on the need for“disciplining the heart,” “breaking desires,”and “cultivating the emotions.”

There are some surprising, or at leastcounterintuitive, aspects of al-Ghazali’s system.For example, it is probably not obvious (atleast to contemporary Western readers) why“gentleness” would be classified under “cour-age,” or why “righteous indignation” would belisted as a form of “temperance.” For contem-porary American English speakers, “courage”seems to be about the management of fear inthe service of (reasonable) action, and indigna-tion hardly seems to be a temperate emotion—though perhaps “righteous” is meant to indi-cate the right and just kind of moderation re-


TABLE 25.1. al-Ghazali’s Classificationof Virtues

Cardinal virtue Related virtues

Wisdom DiscretionExcellence of discernmentPenetration of thoughtCorrectness of opinion

Courage MagnificenceIntrepidityGreatness of soulEnduranceGentlenessFortitudeSuppression of angerCorrect evaluation of selfAmiabilityNobilityManliness

Temperance ModestyShameRemissionPatienceLiberalityGood calculationContentmentAbstinenceCheerfulnessJoyTenderness of characterSelf-disciplineGood appearanceTranquilityHonest dealingRighteous indignationWit

Note. Virtues directly related to emotional states and theirmanagement are given in bold.

quired. But this initial opacity is itself an illus-tration of the entanglement of moral conceptswith emotion concepts. Part of the explanationof the classification of gentleness (as courage)and righteous indignation (as temperance) is Is-lam’s conception of itself as a divinely givenmeans for reforming what the early followersof Muhammed perceived to be the immoral,undisciplined ways of the pre-Islamic Bedouin.It becomes easier to promote the virtue of gen-tleness if it can be persuasively classified as aform of courage.

al-Ghazali’s table of the virtues is only oneinstance of a widespread preoccupation withvirtue ethics among Muslim philosophers.Miskawayh, for example, in his treatiseTahdhib wal-Akhlaq, offered a somewhat dif-ferent classification, while other authors con-tributed in-depth analyses of specific virtues,such as patience (sabr) and gratitude (shukr)(Miskawayh, 1968). One might well wonderwhether these medieval theories have any rele-vance for the moral thinking and emotion con-cepts of contemporary Muslims. Joseph (2001)studied the semantic organization of the virtueconcepts of contemporary Arab Muslims livingin the Bridgeview suburbs of Chicago. Usingfree-listing and similarity-sorting tasks and hi-erarchical cluster analysis, he found an intu-itively available typology of the virtues inwhich, again, the emotions play a prominentrole. The study discovered seven clusters of vir-tue concepts, which are listed below with theirconstituents. Concepts that seem clearly relatedto emotion are underlined.

Intellectual: wisdom and reason

Patience: optimism and patience

Self-control: strength, self-control, modesty,cleanliness, courage

Forbearance: humility, respect, gentleness, tol-erance, forbearance, forgivingness

Attitudes toward others: good neighbor, love ofothers, good relations with kin, compassion

Altruistic: generosity, helping others, doinggood works, sacrifice, gratitude

“Religious” virtues: sincerity, fidelity, truthful-ness, trustworthiness, justice, taqwa3

As with al-Ghazali’s philosophical classi-fication of the virtues, the latent or implicittypology revealed here reinforces the observa-tion that many emotion concepts, and the com-

ponents of them described above, are deeply te-leological: They reveal the goals or ends to-ward which human beings are meant to striveby virtue of their supposed divinely authorednature, and at the same time identify defectsand excesses that must be corrected if theirteloi are to be achieved. As our discussion ofemotion concepts like lajja and lung lang hassuggested, components such as environmentaldeterminants, appraisals of significance, self-management, social management, and commu-nication are significantly constituted by norma-tive ideals and concrete strategies for aligningemotional experience with ethical constraints.

Virtues, of course, are not themselves emo-tions. Rather, they constitute one powerful nor-mative language for articulating and thinkingabout the ethical and moral dimensions ofemotional experience.

CONCLUSION:THE CULTURAL PSYCHOLOGYOF THE EMOTIONS RENEWED

As we enter a new era of collaborative researchamong anthropologists, psychologists, lin-guists, philosophers, and biologists concernedwith similarities and differences in emotionalfunctioning on a worldwide scale, a major goalfor the cultural psychology of the emotionswill be to decompose the emotions (and thelanguages of the emotions) into multiple com-ponents for the sake of comparative under-standing. We hope that by means of thisdecomposition of the emotions’ symbolic struc-ture (and the recognition of the connection be-tween local moral worlds and the social con-struction of emotional experience), it will bepossible to render the meaning of other peo-ple’s mental states without assimilating them inmisleading ways to an a priori set of lexicalitems available in the language of the research-ers. In connection with this chapter, we hopethe reader will have become somewhat morealert to all that is potentially lost or mislead-ingly accrued in the process of translation—as,for example, when Hindu lajja is rendered asAmerican “shame,” or Tibetan lung lang asAmerican “anger.”

It is one of the great marvels of life thatacross languages, cultures, and history, it ispossible, with sufficient knowledge, effort, andinsight, to truly understand the meanings ofother people’s emotions and mental states. Yet


one must also marvel at one of the great ironiesof life—namely, that the process of understand-ing the consciousness of others can deceptivelyappear to be far easier than it really is, therebymaking it even more difficult to achieve a genu-ine understanding of “difference.” Thus, in theend, this chapter is really a plea for a decompo-sition of emotional states into their multiplecomponents of cultural meaning (from norma-tive social appraisals to somatic and affectiveexperience and more) and the application ofthose components of meaning in comparativeresearch. Unless we take that step, we will con-tinue to be prone to the bias that the emotionallife of human beings is “basically” the samearound the world. The truth may well be thatwhen it comes to “basic emotions,” we (medi-eval Hindus and contemporary secular Anglo-Americans, Tibetan Buddhists living in Indiaand Arab Muslims living in Chicago, aborigi-nal Pintupis and indigenous Russians, migra-tory Inuit and settled Balinese) are not onlybasically alike in many ways, but are also basi-cally different from each other in many signifi-cant ways as well.

NOTES

1. For general discussions of Aristotle’s views of emo-tions and virtues, see, for example, Kosman (1980)and Roberts (1989).

2. For some classical and modern treatments of the Is-lamic concept of akhlaq and of the role of the virtuesin morality, see, for example, Busool (1998), Lemu(1997), and Sherif (1975).

3. Taqwa is a difficult-to-translate term that roughlymeans “God-consciousness” and serves for many asthe ultimate virtue—the one to which all the othersare means.

REFERENCES

Aristotle. (1961). Nichomachean Ethics (W. D. Ross,Trans.). Oxford: Oxford University Press.

Busool, A. N. (1998). Prophet Muhammad’s high man-ners: Theory and practice. Chicago: Qur’an Society.

D’Andrade, R. G. (1984). Cultural meaning systems. InR. A. Shweder & R. A. LeVine (Eds.), Culture theory:Essays on mind, self, and emotion (pp. 88–119).Cambridge, UK: Cambridge University Press.

Dimock, E. C., Gerow, E., Naim, C. M., Ramanujan, A.K., Roadarmel, G., & van Builenen, J. A. B. (1974).Literatures of India: An introduction. Chicago: Uni-versity of Chicago Press.

Ekman, P. (1980). Biological and cultural contributionsto body and facial movement in the expression of

emotions. In A. Rorty (Ed.), Explaining emotions(pp. 73–101). Berkeley: University of CaliforniaPress.

Ekman, P. (1984). Expression and the nature of emo-tion. In K. Scherer & P. Ekman (Eds.), Approaches toemotion (pp. 319–343). Hillsdale, NJ: Erlbaum.

Ellsworth, P. (1991). Some implications of cognitiveappraisal theories of emotion. In K. Strongman (Ed.),International review of studies on emotion (Vol. 1,pp. 143–161). Chichester, UK: Wiley.

Fish, S. (1980). Is there a textbook in this class?: On theauthority of interpretive communities. Cambridge,MA: Harvard University Press.

Fridja, N. (1986). The emotions. Cambridge, UK: Cam-bridge University Press.

Geertz, C. (1973). The interpretation of culture. NewYork: Basic Books.

Geertz, C. (1984). From the nature point of view. In R.A. Shweder & R. A. Levine (Eds.), Culture theory:Essays on mind, self, and emotion (pp. 123–136).Cambridge, UK: Cambridge University Press.

Gnoli, R. (1956). The aesthetic experience according toAbhinavagupta. Rome: Istituto Italiano per Il Medioed Estremo Oriente.

Goodnow, J., Miller, P., & Kessel, F. (Eds.). (1995). Newdirections for child development: Vol. 67. Culturalpractices as contexts for development. San Francisco:Jossey-Bass.

Haidt, J. (2001). The emotional dog and its rational tail:A social intuitionist approach to moral judgment.Psychological Review, 108, 814–834.

Haidt, J. (2003). The moral emotions. In R. J.Davidson, K. R. Scherer, & H. H. Goldsmith (Eds.),Handbook of affective sciences (pp. 852–870). Ox-ford: Oxford University Press.

Haidt, J., & Joseph, C. (2004, Fall). Intuitive ethics:How innately prepared intuitions generate culturallyvariable virtues. Daedalus, pp. 55–66.

Haidt, J., & Keltner, D. (1999). Culture and facial ex-pression: Open-ended methods find more faces and agradient of recognition. Cognition and Emotion, 13,225–266.

Haidt, J., Koller, S., & Dias, M. (1993). Affect, culture,and morality, or is it wrong to eat your dog? Journalof Personality and Social Psychology, 65, 613–628.

Hejmadi, A., Davidson, R. J., & Rozin, P. (2000). Ex-ploring Hindu Indian emotion expressions: Evidencefor accurate recognition by Americans and Indians.Psychological Science, 11, 183–187.

Horton, R. (2006). Refining theory and practice in thecultural psychology of emotion: Tibetan “anger” andthe roots of the modern Tibetan commitment to non-violence. Unpublished doctoral dissertation, Univer-sity of Chicago.

Jensen, L. A. (1995). Habits of the heart revisited: Au-tonomy, community, divinity in adults’ moral lan-guage. Qualitative Sociology, 18, 71–86.

Jensen, L. A. (1998). Different habits, different hearts:The moral languages of the culture war. AmericanSociologist, 29, 83–101.


Jensen, L. A. (2005). Through two lenses: A cultural-developmental approach to moral psychology.(Available from Lene Jensen, Department of Psychol-ogy, Clark University, Worcester, MA)

Joseph, C. (2001). The virtues as a cultural domain: Astudy of Arabic speaking Muslims. Unpublished doc-toral dissertation, University of Chicago.

Keith, A. B. (1924). The Sanskrit drama. London: Ox-ford University Press.

Kirsh, D. (1991). Today the earwig, tomorrow man? Ar-tificial Intelligence, 47, 161–184.

Kosman, L. A. (1980). Being properly affected: Virtuesand feelings in Aristotle’s ethics. In A. Rorty (Ed.),Essays on Aristotle’s ethics (pp. 103–116). Berkeley:University of California Press.

Lazarus, R. S. (1991). Progress on a cognitive–motivational–relational theory of emotion. AmericanPsychologist, 46, 819–834.

Lemu, B. A. (1997). Islamic tahdhib and akhlaq: Theoryand practice. Chicago: IQRA’ International Educa-tional Foundation.

LeVine, R. A. (1984). Properties of culture: Anethnographic view. In R. A. Shweder & R. A. LeVine(Eds.), Culture theory: Essays on mind, self, andemotion (pp. 67–87). Cambridge, UK: CambridgeUniversity Press.

Levy, R. I. (1984). Emotion, knowing, and culture. In R.A. Shweder & R. A. LeVine (Eds.), Culture theory:Essays on mind, self, and emotion (pp. 214–237).Cambridge, UK: Cambridge University Press.

Lewis, M. (1989). Cultural differences in children’sknowledge of emotional scripts. In P. Harris & C.Saarni (Eds.), Children’s understanding of emotion(pp. 350–374). New York: Cambridge UniversityPress.

Lewis, M., Sullivan, M., & Michalson, L. (1982). Thecognitive–emotional fugue. In C. E. Izard, J. Kagen,& R. Zajonc (Eds.), Emotions, cognition, and behav-ior (pp. 264–288). New York: Cambridge UniversityPress.

Lutz, C. (1985). Depression and the translation of emo-tional worlds. In A. Kleinman & B. Good (Eds.), Cul-ture and depression: Studies in the anthropology andcross-cultural psychiatry of affect and disorder(pp. 63–100). Berkeley: University of CaliforniaPress.

Lutz, C. (1988). Unnatural emotions: Everyday senti-ments on a Micronesian atoll and their challenge toWestern theory. Chicago: University of ChicagoPress.

Lutz, C., & White, G. (1986). The anthropology ofemotions. Annual Review of Anthropology, 15, 405–436.

Markus, H. R., & Kitayama, S. (1991). Culture and theself: Implications for cognition, emotion and motiva-tion. Psychological Review, 98, 224–253.

Markus, H. R., Kitayama, S., & Heiman, R. (1998).Culture and “basic” psychological principles. In E. T.Higgins & A. W. Kruglanski (Eds.), Social psychol-

ogy: Handbook of basic principles (pp. 857–913).New York: Guilford Press.

Masson, J. L., & Patwardhan, M. V. (1970). Aestheticrapture: The Ras�dhy�ya of the N��ya��stra. Poona,India: Deccan College.

Menon, U., & Shweder, R. A. (1994). Kali’s tongue:Cultural psychology and the power of “shame” inOrissa, India.” In S. Kitayama & H. Markus (Eds.),Emotions and culture (pp. 241–284). Washington,DC: American Psychological Association.

Menon, U., & Shweder, R. A. (1998). The return of the“white man’s burden”: The encounter between themoral discourse of anthropology and the domesticlife of Hindu women. In R. A. Shweder (Ed.), Wel-come to middle age! (and other cultural fictions)(pp. 139–188). Chicago: University of Chicago Press.

Menon, U., & Shweder, R. A. (2003). Dominating Kali:Hindu family values and Tantric power. In R. F.McDermott & J. J. Kripal (Eds.), Encountering Kali:In the margins, at the center, in the west (pp. 80–99).Berkeley: University of California Press.

Mesquita, B. (2003). Emotions as dynamic cultural phe-nomena. In R. J. Davidson, K. R. Scherer, & H. H.Goldsmith (Eds.), Handbook of affective sciences(pp. 871–890). Oxford: Oxford University Press.

Mesquita, B., & Frijda, N. (1992). Cultural variationsin emotions: A review. Psychological Bulletin, 112,179–204.

Miller, J. G. (1984). Culture and the development of ev-eryday social explanation. Journal of Social and Per-sonality Psychology, 46, 961–978.

Miller, P., Potts, R., Fung, H., Hoogstra, L., & Mintz, J.(1990). Narrative practices and the social construc-tion of self in childhood. American Ethnologist, 17,292–311.

Miskawayh, A. M. (1968). The refinement of character(C. Zurayk, Trans.). Beirut: The American Universityof Beirut.

Nisbett, R. E., & Cohen, D. (1995). The culture ofhonor: The psychology of violence in the South.Boulder, CO: Westview Press.

Nisbett, R. E., & Wilson, T. D. (1977). Telling morethan we can know: Verbal reports on mental pro-cesses. Psychological Review, 84(3), 231–259.

Parish, S. (1991). The sacred mind: Newar cultural rep-resentations of mental life and the production ofmoral consciousness. Ethos, 19(3), 313–351.


Roberts, R. C. (1989). Aristotle on virtues and emo-tions. Philosophical Studies, 56(3), 293—306.

Rosaldo, M. Z. (1984). Toward an anthropology of selfand feeling. In R. A. Shweder & R. A. LeVine (Eds.),Culture theory: Essays on mind, self, and emotion(pp. 137–157). Cambridge: Cambridge UniversityPress.

Russell, J. A. (1991). Culture and the categorization ofemotions. Psychological Bulletin, 110(3), 426–450.


Schechner, R. (1988). Performance theory. London:Routledge.

Sherif, M. A. (1975). Ghazali’s theory of virtue. Albany:State University of New York Press.

Shweder, R. A. (1990a). Cultural psychology: What isit? In J. Stigler, R. A. Shweder, & G. Herdt (Eds.),Cultural psychology: Essays on comparative humandevelopment (pp. 1–43). New York: Cambridge Uni-versity Press.

Shweder, R. A. (1990b). In defense of moral realism.Child Development, 61, 2060–2068.

Shweder, R. A. (1991). Thinking through cultures: Ex-peditions in cultural psychology. Cambridge, MA:Harvard University Press.

Shweder, R. A. (1994a). Are moral intuitions self-evident truths? Criminal Justice Ethics, 13(2), 24–31.

Shweder, R. A. (1994b). You’re not sick, you’re just inlove”: Emotion as an interpretive system. In P.Ekman & R. Davidson (Eds.), The nature of emo-tions: Fundamental questions (pp. 32–44). NewYork: Oxford University Press.

Shweder, R. A. (1996). True ethnography: The lore, thelure and the law. In R. Jessor, A. Colby, & R. A.Shweder (Eds.), Ethnography and human develop-ment: Context and meaning in social inquiry (pp. 15–52). Chicago: University of Chicago Press.

Shweder, R. A. (1999a). Cultural psychology. In R. A.Wilson & F. Keil (Eds.), MIT encyclopedia of cogni-tive sciences (pp. 211–213). Cambridge, MA: MITPress.

Shweder, R. A. (1999b). Why cultural psychology?Ethos, 27, 62–73.

Shweder, R. A. (2002). The nature of morality: The cate-gory of bad acts. Lahey Clinic Medical Ethics News-letter, 9, 6–7.

Shweder, R. A. (2003). Toward a deep cultural psychol-ogy of shame. Social Research, 70, 1109–1130.

Shweder, R. A. (2004). Deconstructing the emotions forthe sake of comparative research. In A. S. R.Manstead, N. Frijda, & A. Fischer (Eds.), Feelingsand emotions: The Amsterdam Symposium (pp. 81–97). New York: Cambridge University Press.

Shweder, R. A., & Bourne, E. (1984). Does the conceptof the person vary cross-culturally? In R. A. Shweder& R. A. LeVine (Eds.), Culture theory: Essays onmind, self, and emotion (pp. 158–199). Cambridge,UK: Cambridge University Press.

Shweder, R. A., Goodnow, J., Hatano, G., LeVine, R.A., Markus, H. R., & Miller, P. (1998). The culturalpsychology of development: One mind, many men-talities. In W. Damon (Series Ed.) & R. Lerner (Vol.

Ed.), Handbook of child psychology: Vol. 1. Theoret-ical models of human development (5th ed., pp. 865–937). New York: Wiley.

Shweder, R. A., & Haidt, J. (1993). The future of moralpsychology: Truth, intuition and the pluralist way.Psychological Science, 4, 360–365.

Shweder, R. A., & LeVine, R. A. (Eds.). (1984). Culturetheory: Essays on mind, self and emotion. Cam-bridge, UK: Cambridge University Press.

Shweder, R. A., Mahapatra, M., & Miller, J. G. (1987).Culture and moral development. In J. Kagan & S.Lamb (Eds.), The emergence of morality in youngchildren (pp. 1–83). Chicago: University of ChicagoPress.

Shweder, R. A., Much, N. C., Mahapatra, M., & Park, L.(1997). The “big three” of morality (autonomy, com-munity, divinity) and the “big three” explanations ofsuffering. In A. M. Brandt & P. Rozin (Eds.), Moralityand health (pp. 119–169). New York: Routledge.

Stein, N., & Levine, L. J. (1987). Thinking about feel-ings: The development and organization of emo-tional knowledge. In R. E. Snow & M. J. Farr (Eds.),Aptitude, learning and instruction. Hillsdale, NJ:Erlbaum.

Stigler, J. (1984). Mental abacus: The effect of abacustraining on Chinese children’s mental calculation.Cognitive Psychology, 16, 145–176.

Stigler, J., Chalip, L., & Miller, K. (1986). Culture andmathematics learning. Review of Research in Educa-tion, 15, 253–306.

Stigler, J., Nusbaum, H., & Chalip, L. (1988). Develop-mental changes in speed of processing: Central limit-ing mechanisms on shell transfer. Child Develop-ment, 59, 1144–1153.

Turiel, E. (1983). The development of social knowledge:Morality and convention. Cambridge, UK: Cam-bridge University Press.

Wierzbicka, A. (Ed.). (1990). The semantics of the emo-tions [Special issue]. Australian Journal of Linguis-tics, 10(2).

Wierzbicka, A. (1992). Talk about emotions: Semantics,culture and cognition. Cognition and Emotion, 6(3–4), 285–319.

Wierzbicka, A. (1997). Understanding cultures throughtheir key words: English, Russian, Polish, Germanand Japanese. New York: Oxford University Press.

Wylie, T. (1959). A standard system of Tibetan tran-scription. Harvard Journal of Asiatic Studies, 22,261–267.

Yatiswarananda, S. (1979). Meditation and spirituallife. Bangalore, India: Sri Ramakrishna Ashrama.


C H A P T E R 2 6

Intergroup Emotions

ELIOT R. SMITH and DIANE M. MACKIE

Current emotion research and theory generallyassume that emotions are adaptive and func-tional, with a strong focus on their functional-ity at the biological or individual level. This fo-cus has often involved the use of nonsocial“prototype” emotion situations (such as en-countering a bear while walking in the woods)as conceptual touchstones. However, social as-pects of emotions are increasingly being inves-tigated, following from the recognition that (inhumans) emotions involve socially constructedmeanings and most often occur in a social con-text. In this chapter, we describe one specificclass of social emotions, which we have called“intergroup emotions”—emotions that arisewhen people identify with a social group andrespond emotionally to events or objects thatimpinge on the group. After introducing theconcept of intergroup emotions, we describe itsrelationship to theoretical models of emotionsin general. We review evidence (from our ownwork and that of others) regarding the keyproperties of intergroup emotions, and con-clude by summarizing the implications of thisconceptualization, both for emotion theory ingeneral and for intergroup relations. As we will

show, this model shares the assumption thatemotions are generally functional, but at thelevel of social groups and not solely at the levelof the individual.

WHAT AREINTERGROUP EMOTIONS?Group Identification and Depersonalization

The fundamental insight underlying the con-cept of intergroup emotions derives from a ma-jor line of social-psychological theory and re-search inspired by social identity theory (Tajfel,1978) and self-categorization theory (Turner,Hogg, Oakes, Reicher, & Wetherell, 1987).This is the idea that important group member-ships (those with which people psychologicallyidentify) become a part of a person’s identity,along with the person’s unique, personal, indi-viduating attributes. The “group” in this con-ceptualization can be a relatively small numberof people who interact face to face (such as acommittee or a sports team), or a larger socialcategory of people who do not interact face toface (such as a national, ethnic, gender, or reli-

428

gious identity). Either type of group can serveas a vehicle for social identification and canconstitute an important and meaningful aspectof people’s identity.

Under circumstances that make a particulargroup membership or social identity salient,people do not think of themselves as unique in-dividuals, but rather as relatively interchange-able members of the group (a process known as“depersonalization”). This occurs primarily inwhat is termed an “intergroup situation,” onein which social comparisons, competition, orconflict between groups are salient—hence thename of our theoretical model, “intergroupemotions theory” (IET). The consequences ofdepersonalization are many (for a review, seeOakes, Haslam, & Turner, 1994). People con-form to the norms of the activated group intheir beliefs, attitudes, and behaviors. They be-come motivationally aligned with the group, sothey see actions that advance the group’s inter-ests as desirable and beneficial. They perceiveother members of the same group (the“ingroup”) as similar to themselves and as lik-able, and tend to treat them with justice andfairness, while withholding these benefits frommembers of outgroups.

Tajfel’s (1978) original formulation of socialidentity theory included the idea that whengroup identification turns a group into an im-portant social identity for an individual, thegroup takes on emotional significance. How-ever, compared to the cognitive, motivational,and behavioral consequences of group identifi-cation, emotional consequences received al-most no theoretical or empirical attention untilrecently in the large literature inspired by socialidentity theory. This seems odd in view of theobvious fact that intergroup conflicts (whetherlabor–management disputes, ethnic or religiousconflicts, clashes between street gangs and po-lice, or international wars) are regularly char-acterized by extreme levels of emotion. Both di-rect participants and bystanders who identifywith one of the groups involved but are not di-rectly affected tend to experience strong nega-tive feelings, including anger, fear, resentment,contempt, and so on.

Group Identification as a Basisfor Intergroup Emotions

The core concepts underlying IET, introducedby Smith (1993), flow from this theoreticalbackground. The fundamental idea is that

when someone in an intergroup situation is re-sponding in terms of a social identity, objects orevents that affect the ingroup will elicit emo-tional responses, because the group becomes ina real sense an aspect of the person’s psycholog-ical self (Smith & Henry, 1996). Simply put,under these conditions depersonalizationcauses the person to react to the world as agroup member rather than a unique individual,and thus objects and events have emotionalconsequences based on the way they relate tothe group, not the individual. These group-based emotions can be understood and ana-lyzed in the same way as any others—by usingtheories of emotions in general, such asappraisal theories. At the heart of IET, there-fore, is the distinction between intergroup emo-tions and the more commonly studiedindividual-level emotions.

Let us clarify two important points. First, inour conceptualization, intergroup emotions areexperienced by individuals (when they identifyas members of a group), not by some kind ofgroup mind. Whether intergroup emotions areactually shared across many members of agroup is an empirical question, and we presentrelevant evidence later. Second, we assume thatintergroup emotions are generally similar toindividual-level emotions in the ways they areexperienced, the effects they have on cognitive,perceptual, and motor processes, and so forth.They differ in the ways they are elicited (i.e., bygroup-relevant rather than personally relevantevents) and in their functions (i.e., in regulatinggroup-related or collective behavior ratherthan purely individual behavior).

As background for IET, the broad idea thatintergroup relations and especially intergroupconflict might trigger emotional responses hasbeen investigated since the middle 1980s, withseminal work by Dijker (1987), Gaertner andDovidio (1986), and Stephan and Stephan(1985). Their research examined the usuallynegative emotions that people can experiencewhen encountering individual members ofoutgroups, triggered by cultural differences(e.g., the outgroup member may violateingroup norms) or by an anxiety-toned desireto make the interaction go smoothly and not togive offense. Research has amply demonstratedthat these emotional reactions can occur andcan motivate avoidance of outgroupmembers—demonstrating that avoidance neednot necessarily be attributed to sheer antipathyfor the outgroup.

26. Intergroup Emotions 429

From the standpoint of IET, an importantquestion is whether these negative emotions ex-perienced during intergroup contact areindividual-level or intergroup emotions. Theymay often remain at the individual level, if theemotions are generated by perceptions of anoutgroup member’s individual annoying or of-fensive characteristics, or by one’s perceivedpersonal lack of knowledge about the outgroupand uncertainty about how to interact appro-priately. Still, the definition of a situation as in-terpersonal or intergroup depends on the waythe perceiver thinks about it. So a white per-son’s concern about personally appearing prej-udiced in a cross-group interaction would bean individual emotion, but if he or she feels likea representative of the entire white group and isconcerned about confirming the stereotype thatwhites are generally racist, this would be agroup-level emotion. A woman’s anger at beingmade to look stupid by a man’s cutting com-ment might be individual, but if she feels thatshe was made to look stupid because she was awoman, the emotion might be group-based.Whether the emotion is individual or inter-group, if repeated encounters with members ofa particular group consistently give rise to feel-ings of annoyance, anxiety, and the like, theemotion may become associated with the men-tal representation of the group in general (justas for some individuals feelings of anxiety maybecome associated with dentists in general). Asa result, the perceiver might start to experienceanxiety, irritation, and so forth, on any thoughtabout or encounter with that particularoutgroup.

Intergroup Emotions and the Regulationof Behavior

Just as any emotion motivates people to takespecific sorts of action (Frijda, Kuipers, & terSchure, 1989), intergroup emotions should dothe same. Generally, intergroup emotionsshould lead to actions specifically related to theintergroup situation. Examples include desiresto confront or attack an outgroup, to avoid anoutgroup, to affiliate with ingroup members,or to support or oppose government policiesthat have an impact on entire social groups. Inthis way, intergroup emotions involve apprais-als, feelings, and action tendencies that are allat the group level: The individual who identi-fies with an ingroup may feel that they are

threatening us; we feel angry at them; we sup-port policies designed to keep them fromimmigrating to our country, cut off their gov-ernment benefits, and so on. The distinctionbetween individual and intergroup emotions isparallel to that between individual and frater-nal (i.e., group-based) relative deprivation(Runciman, 1966). The former is the percep-tion that one is not doing as well as other indi-viduals, and the latter is the perception thatone’s group is not doing as well as othergroups; fraternal deprivation has more potenteffects on political and social attitudes.

Two Foci of Intergroup Emotions

The fundamental idea that group identificationcan lead to group-level emotional responseshas both a narrower and a broader version.First, emotions such as anger, fear, disgust, orenvy targeted specifically at an outgroup mayrelate to perceptions, prejudiced attitudes, ordiscriminatory behaviors directed at theoutgroup. This line of thinking leads to the im-portant hypothesis that all negatively evaluatedoutgroups are not treated the same way. Forexample, outgroups that are targets of angermay be actively attacked or confronted, whileoutgroups that are targets of disgust may beavoided (see Mackie & Smith, 2002, for rele-vant evidence). Outgroup-directed emotionsneed not always be negative, for in some cir-cumstances people feel sympathy or pity to-ward an outgroup, leading to desires to offerhelp.

Second, people may experience more generalemotional feelings when they are thinking ofthemselves in terms of a particular social iden-tity. These may include not only emotions di-rected at outgroups, but also positive or nega-tive ingroup-directed emotions (such as group-based pride or collective guilt) and general af-fective feelings (such as happiness, anxiety, orirritation) that are based on group member-ship. Thus each group may have a typical “pro-file” of emotional tendencies, such that whenthinking of themselves in terms of that groupmembership, people often feel proud, irritated,anxious, and so on, in a pattern that systemati-cally differs from what they would report whenthinking of themselves as individuals (Smith,Seger, & Mackie, 2007).

The same distinction between specificallytargeted emotions and general feeling states


also characterizes individual-level emotions, ofcourse. Individuals at times experience emo-tions targeted at specific other individuals orobjects (e.g., anger at a threatening rival, sym-pathy for a person in distress), and also possessrelatively stable profiles of emotional feelings(e.g., general tendencies to feel angry, anxious,or happy) that relate to their personality char-acteristics (Watson & Clark, 1992).

RELATION OF IETTO EMOTION THEORY IN GENERAL

Appraisal Theories of Emotion

Our original outline of IET (Smith, 1993) andsubsequent work (Mackie, Devos, & Smith,2000; Mackie, Silver, & Smith, 2004; Smith,1999) adopted the assumptions of appraisaltheories of emotions, because they seemed to usto fit well with our general theoretical struc-ture. Our sole modification to appraisal theorywas the assumption that in an intergroup situa-tion, events, objects, and groups are appraisedin terms of their implications for the ingroup(not just the individual self). Intergroup emo-tions are generated by this appraisal process,just as individual-level emotions are generatedby appraisals of objects or events that impingeon the individual self (Scherer, 1984; Roseman,1984). For example, a person who identifieswith a group and sees the ingroup as threat-ened by a powerful outgroup’s actions may ex-perience intergroup anxiety or fear. One conse-quence may be a desire or impulse to avoid orescape from the outgroup. As we have dis-cussed earlier, intergroup emotions may be di-rected at the ingroup as well. For example, inan intergroup situation that makes social com-parisons between groups salient, people mayfeel collective pride if they believe that theirgroup has succeeded in an important task. Orfeelings of collective guilt may result if peopleappraise their group as having violated impor-tant moral principles (Doosje, Branscombe,Spears, & Manstead, 1998).

Although many appraisal theories (e.g.,Roseman, 1984) suggest that appraisals causeemotional states, the causal direction is notone-way. Tiedens and Linton (2001) have dem-onstrated that emotional states influence peo-ple’s judgments on appraisal-related dimen-sions; for example, because anxiety is related tothe appraisal of uncertainty, people in an anx-

ious state judge events as highly uncertain. Al-though research has not yet tested this idea,there is every reason to think that this processshould operate in the same way with regard tointergroup emotions as it does with individual-level emotions.

Finally, appraisal theories suggest that emo-tions lead to specific behavioral action tenden-cies (Frijda, 1986). Intergroup anger may leadto a desire to attack or confront the outgroup,disgust to a desire to avoid, or guilt to a desireto make reparations for the ingroup’s actions.By adopting the assumptions of appraisal theo-ries in this way, IET makes predictions aboutlinkages of appraisals of the intergroup situa-tion, specific emotional experiences that peoplereport when they are thinking of themselves asgroup members, and motives or tendencies toengage in collective (group-level) actions (thesepredictions have been tested; see, e.g., Mackieet al., 2000).

Core Affect Model of Emotion

However, IET does not necessarily rest solelyon appraisal theories. Its fundamental insight isthe idea that group identification makes theingroup part of the psychological self andhence makes group-relevant events or objectsable to trigger emotions (just as self-relevantevents or objects always do). The process of“triggering” could be described by using theconcepts of appraisal theories, as in our earlierpresentations of the IET model, or can be de-scribed in alternative ways. As an illustration,we consider the “core affect” model (Russell &Barrett, 1999; Russell, 2003; Barrett, 2006). Inthis model, states of core affect, which are de-scribed in a two-dimensional space whose axesare pleasantness and arousal, are fundamentalcomponents of all experienced emotions (aswell as moods). Core affect can change formany reasons: in response to external stimulisuch as pleasant or unpleasant environmentalstates or positive or negative events, or in re-sponse to internal physiological processes suchas diurnal rhythms. Core affect is assumed tobe subjectively perceptible (as the sense of feel-ing good, bad, energized, tired, etc.).

An experienced episode of emotion beginswith a change of core affect that is consciouslynoted and attributed to some cause. This attri-bution makes the difference between just feel-ing negatively aroused and feeling negatively


aroused because of a specific event. Attributingaffective states to external objects or events isoften unproblematic, because of the frequentclose co-occurrence in time of external eventsand the affective reactions that they cause. Andthese attributions are adaptive, for they allowthe perceiver to direct attention and behaviorappropriately with regard to the object that isresponsible for the feeling. However, the truecauses of affective states are not always obvi-ous, so people may make misattributions. Afeeling of negative arousal that is actually dueto irrelevant reasons (such as unpleasantly hottemperatures) may be misattributed to anotherperson’s annoying behavior, leading to the ex-perience of anger, and potentially to aggression(Berkowitz, 1998).

How is an emotional experience labeled asan instance of a specific emotion such as an-ger, fear, sadness, or guilt? The core affectmodel holds that various factors—includingthe core affect and its perceived cause, as wellas the situational context, one’s overt behav-iors, and bodily experiences (such as physio-logical changes)—become input to a percep-tual categorization process in which theexperience is categorized as a discrete emo-tion. Russell (2003) holds that the episode iscategorized on the basis of the resemblancebetween these factors and the person’s mentalrepresentation of a given emotion’s prototype.Barrett’s (2006) model differs in postulatingnot a fixed prototype representation, but asituated reconstruction—an emotion represen-tation that is created online and flexiblytuned to the constraints and goals of the situ-ation. In either case, these mental representa-tions (prototype or situated reconstruction)can and will vary between people because ofindividual experiences and cultural differ-ences. As a result of the categorization, a per-son could say that he or she feels afraid ofthe specific object. This labeling process pro-duces the conscious awareness of having anemotional episode.

The core affect model incorporates manysubstantive predictions of appraisal theories,by postulating that appraisals (the person’s in-terpretations of various aspects of the situa-tion) relate to the process of categorization andself-perception of a specific emotion, ratherthan seeing appraisals as initiating the emo-tional state with all its concomitants (subjectivefeelings, autonomic changes, instrumental ac-tions, etc.). See Russell (2003) for further dis-

cussion of relations between the core affectmodel and appraisal theories.

IET is broadly consistent with the core affectmodel as well as with appraisal theories. Howmight the process differ if a group (rather thanan individual) identity is at stake? First, let usassume that an event occurs with implicationsfor a person’s important group membership,such as a reminder of a nation’s colonial pastfor a person who identifies strongly with thenational group. The event may lead to changesin core affect—in this example, negativearousal. Next, the person may search for acause of that change, perhaps making amisattribution or perhaps correctly attributingit to the event in question. The individual willthen draw on his or her affective feelings andattributions, as well as other aspects of the situ-ation, to label the emotion as an instance ofguilt—in this case, collective guilt. Desires totake collective actions, such as to make repara-tions or to offer aid to the victims of theingroup’s actions, may ultimately result. Theoverall process is not greatly different fromthat postulated by the core affect model in gen-eral; the key differences are that the self-relevance that triggers affective changes in thefirst place is relevance to a collective (ratherthan an individual) self, and that the interpre-tive and categorization processes that followidentify the emotion as an intergroup ratherthan individual-level emotion. In our example,the perceiver will be readily able to categorizethe experience as one of collective guilt, be-cause he or she will have no reason to feel indi-vidually guilty about ingroup actions that oc-curred perhaps hundreds of years ago.

Another example illustrates how the processof making an attribution can switch a perceiverbetween an individual or intergroup emotion.Someone who learns that he or she was notchosen for a much-wanted promotion (andhence feels strong negative arousal) may decidethat the reason was his or her boss’s ignoranceand failure to recognize talent—and may thusexperience anger or disappointment at the indi-vidual level. Alternatively, the same event maybe attributed to minority groups’ pressing foraffirmative action through the promotion ofunderqualified minority candidates—in whichcase the individual’s anger may be at the inter-group level (directed at the minority outgroupin general). The action tendencies will be quitedifferent in this case than in the case of individ-ual anger.


One test of whether an emotion is individualor intergroup is to ask this question: Would theemotional response be similar if the same eventhappened to some other ingroup member? Foran intergroup emotion, the answer is probablyyes; someone might be angered by perceivingthat another ingroup member was “victim-ized” by an unfair affirmative action program,even if the perceiver was personally unaffected.In fact, U.S. Senator Jesse Helms ran a TV ad inhis 1990 campaign, showing a pair of whitehands crumpling a letter (presumably a rejec-tion letter) while a voice-over stated, “Youneeded that job, but they had to give it to a mi-nority” (the ad can be viewed at www.pbs.org/30secondcandidate/timeline/years/1990.html).Evidently the intention was to anger manywhite viewers at the thought that other mem-bers of their ingroup were being treated un-fairly.

As this discussion indicates, each of the threemain elements of the core affect model can beinfluenced by a person’s group identifications.Core affect changes can arise from events thathave an impact on the ingroup and not just theperson as an individual (a common example isthe way we feel great when our team wins). At-tributions concerning the cause of a change incore affect may implicate social groups—forexample, when we decide that an outgroupmember is treating us unfairly because he orshe is prejudiced against our group, not just foridiosyncratic personal reasons. And categoriza-tion processes may tell us that we are feeling anemotion as an individual (e.g., individual guiltfor some personal action) or as a group mem-ber (e.g., collective guilt for an ingroup’s histor-ical acts).

EVIDENCE REGARDINGINTERGROUP EMOTIONS

In this section, we review evidence regardingthe properties of intergroup emotions, espe-cially those that distinguish them from individ-ual emotions. We describe recent studies thatoffer evidence on four fundamental questions:(1) Are intergroup emotions distinct from thesame person’s individual emotions? (2) Arethey related to group identification? (3) Arethey shared among members of the samegroup? (4) Do they functionally regulate intra-group and intergroup attitudes and behaviors?Our discussion also emphasizes studies that

demonstrate roles for intergroup emotions inseveral specific phenomena relevant to preju-dice and intergroup relations.

Differences from Individual-Level Emotions

One important distinction between intergroupand individual emotions is demonstrated byfindings that people can experience emotionson behalf of a group or fellow group memberseven when the perceivers are not personally af-fected (e.g., Mackie et al., 2004; Yzerbyt,Dumont, Wigboldus, & Gordijn, 2003). Forexample, people report feeling unhappy andangry when they learn of an event that harmsother members of a situationally salientingroup, even though the event has no conceiv-able implications for the perceivers as individu-als (Yzerbyt et al., 2003).

In a more direct examination of relations be-tween group and individual-level emotions, we(Smith et al., 2007) had students report the ex-tent to which they generally felt 12 specificemotions, as individuals and as members ofseveral different ingroups (university, national,and political party groups). As might be ex-pected, reports of emotions at the individualand group level were correlated at about the .3level (e.g., people who reported feeling moreangry as individuals also tended to report feel-ing more angry as group members). Despitethis correlation, analyses showed that profilesof group emotions and individual emotionswere meaningfully distinct—qualitatively dif-ferent, not merely differing in the overall levelor intensity of individual versus group emo-tions.

Relationship to Group Identification

For all groups, we (Smith et al., 2007) found astrong relation between positive group emo-tions and ingroup identification. This makessense both because group identification is acondition that enables the experience of inter-group emotions, and because positive inter-group emotions such as pride and satisfactionare likely to reward and encourage strong iden-tification with a particular group. The relationof anger at the outgroup to group identificationwas also positive, consistent with some otherresearch (Kessler & Hollbach, 2005). In con-trast, negative group emotions other thanoutgroup anger (such as guilt, anxiety, and irri-tation) were more weakly and generally nega-


tively related to identification. The data pat-terns suggest that the negative correlation islargely due to a type of motivated cognition:Strong group identification leads people to re-interpret and reappraise group-related eventsto avoid negative feelings (Doosje et al., 1998).An additional process that may contribute tothe negative correlation is that negative emo-tions may motivate a decrease in identificationwith the particular group.

Social Sharing within an Ingroup

Our results (Smith et al., 2007) also demon-strate that people’s group-level emotions aresocially shared, and shared more strongly bypeople who identify more with the group. Thatis, each social group (Americans, Democrats,university students, etc.) has a specific profileof group-level emotions (e.g., high happiness,low anger, moderate guilt). When reportingtheir emotions for that particular group, mem-bers tend to converge toward the group profile,and those who identify more strongly with thegroup converge to a greater extent.

Three distinct processes may contribute tothis convergence. First, the convergence may bedue to “emotional contagion” (Neumann &Strack, 2000), meaning that people tend totake on the emotions displayed by fellowingroup members with whom they interact.However, the contagion mechanism may bemore relevant to face-to-face interactinggroups than to the social category groups usedin this study. Still, members of larger groupsmay be affected by emotional contagion whenleaders or prototypical members are portrayedin the media and other group members modeltheir emotions. Second, people could be re-garded as conforming to ingroup norms withregard to their group-level emotions. It haslong been known that when people stronglyidentify with a group, and their membership ismade salient in a specific situation, they tend toconform to group norms or move closer to thegroup prototype in their behaviors and their at-titudes (Hogg & Turner, 1987; Simon & Ham-ilton, 1994). Although existing research hasnot considered emotions as a domain in whichpeople move toward a group prototype, thesame process should operate with emotions.Finally, a third possibility is that group emotionconvergence occurs because thinking about agroup membership makes the same key group-relevant events and appraisals salient to differ-

ent perceivers, so that emotional responses tosuch events are generally shared among groupmembers. For example, when many individualAmericans think of themselves as Americans,the same limited number of events and situa-tions may be salient, such as attacks by militantanti-Americans. So these individuals may all re-port feeling angry as Americans because theyare all responding to more or less the same sa-lient events in more or less the same way. Ingeneral, all three of these processes (emotionalcontagion, conformity to emotion profiles thatfunction as group norms, and shared reactionsto salient group-relevant objects or events) maybe important causes of group emotion conver-gence.

Regulation of Intergroup and IntragroupAttitudes/Behaviors

We (Smith et al., 2007) found that group emo-tions predicted both ingroup-directed andoutgroup-directed action tendencies, aboveand beyond the relatively weak predictivepower of individual-level emotions. Thesestudies investigated a wide range of specific ac-tion tendencies involving ingroup support andsolidarity, outgroup confrontation, andoutgroup avoidance. The results suggested thatanger at the outgroup (and, to a lesser extent,anger at the ingroup) was the most powerfulpredictor across all categories of action tenden-cies. The fact that anger at the outgroup pre-dicts the desire to engage in confrontationalbehavior would be expected from standard the-ories relating emotion to action tendencies(Frijda et al., 1989; Mackie et al., 2000). Therelation of anger at the outgroup to tendenciesto support and affiliate with the ingroup is alsoconsistent with previous research (Kessler &Hollbach, 2005) showing that this group emo-tion tends to increase ingroup identification. Itis less clear why outgroup anger leads to ten-dencies to avoid the outgroup. We (Smith et al.,2007) suggested that this effect could also beregarded as part of a behavioral regulationstrategy: Someone who thinks that beingaround outgroup members might cause angermay choose to avoid such situations, to mini-mize those negative feelings and perhaps thedanger of acting inappropriately.

The regulation of intergroup behavior by in-tergroup emotions has also been demonstratedin other research (Maitner, Mackie, & Smith,2007). Three studies investigated the role of in-


tergroup satisfaction in intergroup conflict. Af-ter reading about real acts of aggression com-mitted by an ingroup, participants reportedhow those actions made them feel and howmuch they would support similar aggression inthe future. In all three studies, experiencing in-tergroup satisfaction increased support for sim-ilar aggression, whereas experiencing inter-group guilt decreased support for similaraggression. Study 2 showed that ingroup iden-tification increased appraisals of the aggressionas justified, which increased satisfaction anddecreased guilt, and thus increased support forfuture aggression. Study 3 provided an experi-mental test of the model: When justificationappraisals were manipulated, emotion and sup-port for further aggression changed accord-ingly. These findings demonstrate conditionsunder which intergroup satisfaction can influ-ence people’s support for their group’s aggres-sion, and therefore can facilitate and sustain in-tergroup conflict.

Yet another aspect of behavioral regulationby intergroup emotions is that if intergroupemotions are functional, successfully imple-menting an emotion-linked behavioral ten-dency should allow the emotional feelings todissipate, whereas impeding the behavioral ten-dency should intensify the emotion. We(Maitner, Mackie, & Smith, 2006) investigatedthe emotional consequences of satisfying orthwarting behavioral intentions related to in-tergroup emotions. Study 1 showed that if anattack on the ingroup produced anger, retalia-tion increased satisfaction, but if an attack pro-duced fear, retaliation increased fear and guilt.Study 2 showed that outgroup-directed angerinstigated via group insult dissipated when theingroup successfully responded, but was exac-erbated by an unsuccessful response. Re-sponding in an emotionally appropriate waywas satisfying, but ingroup failure to respondelicited anger directed at the ingroup. Study 3showed that intergroup guilt following aggres-sion was diminished when the ingroup madereparations, but was exacerbated when theingroup aggressed again. These findings dem-onstrate that satisfying behavioral intentionsassociated with intergroup emotions fulfills aregulatory function.

Besides these examples from our own re-search, many other studies show that inter-group emotions are related to desires to takeactions relevant to group memberships. For ex-ample, feelings of fear versus anger as Ameri-

cans experienced in response to the attacks ofSeptember 11, 2001, predicted respondents’support for restrictions on civil liberties mea-sured several months later (Skitka, Bauman, &Mullen, 2004). A number of other researchershave demonstrated potent effects of group-level guilt and ingroup-directed anger on actiontendencies related to apologizing and makingreparations for past group-based offenses(Branscombe & Doosje, 2004; Leach & Iyer,2006).

Relations of Intergroup Emotionsto Prejudice

The theme that intergroup emotions regulateand direct people’s attitudes and behaviors to-ward an outgroup has specific implications forunderstanding prejudice—a topic that has beenthe focus of much of our research. We (Miller,Smith, & Mackie, 2004) examined the role ofintergroup emotions directed at a specificoutgroup (African Americans) in predictingEuropean Americans’ prejudice against thatgroup. The studies replicated large bodies of re-search that show effects of past intergroup con-tact on reducing prejudice, and of “social dom-inance orientation” (SDO), a personality-likeindividual difference reflecting desires to main-tain hierarchies of group inequality, on increas-ing prejudice. The studies also demonstratedthat intergroup emotions (self-reports of differ-ent emotions experienced by European Ameri-cans when encountering or thinking aboutAfrican Americans) played a major role in me-diating both of those effects. Specifically, the ef-fect of past intergroup contact was mediated bypositive intergroup emotions, and there wasmore tentative evidence suggesting mediationby negative emotions. In other words, inter-group contact increases positive emotions anddecreases negative emotions, and both of theseeffects tend to reduce prejudice against theoutgroup. The effect of SDO was mediated byboth positive and negative intergroup emo-tions. People high in SDO tended to perceivecertain outgroups as threatening, leading tonegative emotions such as fear, anger, or resent-ment, which in turn increased prejudiced atti-tudes. Those high in SDO also appeared to ex-perience fewer positive emotions such assympathy or pride with regard to outgroups.Thus intergroup emotions play a key role inmediating the known powerful effects of bothintergroup contact and SDO on prejudice.


The Miller et al. (2004) studies also exam-ined stereotypes of the outgroup as an alterna-tive potential mediator, permitting a compari-son of their role with that of intergroupemotions. Stereotypes had no significant role inmediating the effects of past intergroup con-tact, and mediated part of the effect of SDO inone of the two studies. These results fit with theexisting empirical and conceptual evidence forthe priority of affective (emotional) over cogni-tive (stereotype beliefs) mediators of effects onprejudice. Pettigrew (1998), for example, notedthat effects of intergroup contact on affectivedependent variables are generally stronger thaneffects on measures of stereotypes, underliningthe importance of understanding group-levelemotions as part of the entire picture of preju-dice and intergroup relations.

SUMMARY AND IMPLICATIONS

The perspective described in this chapter restson the simple yet powerful idea that a socialidentity based on group identification—notjust a biological/individual self—can have emo-tional implications. Many concrete researchhypotheses can be generated by combining thisfundamental idea with assumptions based onemotion theory in general, yielding the insightthat appraisals of social situations, experiencedemotions, and emotion-driven desires for ac-tion can all occur at the level of the socialgroup as well as the individual.

Implications for the Studyof Intergroup Relations

This set of ideas has several implications forour understanding of prejudice and intergrouprelations. First, let us consider the relations ofgroup-based emotions directed at a particularoutgroup and the more traditional concept ofstereotypes of the outgroup. IET holds that ste-reotypes are far from irrelevant, for they mayfeed into appraisals. For example, an outgroupmay be seen as hard-working and achievement-oriented, which could lead to its being ap-praised as a potent threat to the ingroup’s eco-nomic status in society. Note that, as thisexample implies, there is no necessary direct re-lationship between the valence of stereotypesand the valence of people’s emotional responsesto groups. Positive traits ascribed to the

outgroup can lead to negative responses, be-cause emotional responses rest on relationalappraisals (what does the outgroup meanfor us?) rather than on the valence of theoutgroup’s characteristics in isolation.

A related point is that, as mentioned previ-ously, IET predicts that people may have quali-tatively different types of prejudices against dif-ferent groups, which may typically lead todifferent types of actions. We may wish, for ex-ample, to discriminate angry/resentful preju-dice from anxious/fearful prejudice fromdisgust-based prejudice. The appraisals that re-late to these, the individual-difference charac-teristics that may moderate them, and the ac-tion tendencies that will result may all lookstrikingly different from one another (seeNeuberg & Cottrell, 2002). In this way, IETtakes us beyond the simpler idea that prejudiceis simply a negative attitude or antipathy to-ward a group, and toward a more specific, dif-ferentiated understanding of prejudice and itscauses and effects (Mackie & Smith, 2002).

Finally, thinking about emotions as an im-portant component of people’s reactions tooutgroups also leads to new conceptions re-garding time. In traditional views, stereotypesand prejudiced attitudes toward an outgroupare typically regarded as highly stable (in fact,their resistance to change has been a core issuemotivating much research). In contrast, emo-tions are labile, varying over seconds and min-utes. This raises new research questions, suchas whether someone’s thoughts, evaluations,and behaviors toward an outgroup may differspecifically when the individual is in an emo-tional state from when he or she is not. At leastone study suggests that the answer to this ques-tion may be yes (DeSteno, Dasgupta, Bartlett,& Cajdric, 2004).

Implications for Emotion Theoryand Research

Just as IET encourages a rethinking of some is-sues related to prejudice and intergroup behav-ior, it also has implications for theory and re-search regarding emotion more generally.Perhaps the most fundamental implication isthe idea that emotions pertain to an identityand not to a biological individual. An individ-ual typically has multiple identities (including apersonal self, many group memberships thatmay become salient in different circumstances,


and possibly several important relational iden-tities as well; Sedikides & Brewer, 2001). Eachof these identities is an aspect of the self, so itremains true that, as emotion researchers as-sume, emotions always implicate the self. Inaddition, there will generally be a degree ofcontinuity between emotions experienced indifferent identities (illustrated by the .3 correla-tions between group-level and individual-levelemotions; Smith et al., 2007), perhaps partlydue to biologically based differences in emo-tional reactivity. Still, each identity may have adistinctive emotion profile, both short-termand chronically. This idea has many as-yet-unexplored implications—for example, in re-gard to emotional vulnerability.

The idea that emotions are rooted in identi-ties also has implications for emotion regula-tion: Because people can shift rapidly betweenidentities (e.g., between one group membershipand another, or between a group-level and anindividual-level identity), such shifts can bepart of an emotion regulation process (Smith &Mackie, 2006). People may be expected toadopt identities associated with positive groupemotions or disidentify from groups associatedwith negative group emotions, and there is evi-dence consistent with this idea (Kessler &Hollbach, 2005). Research might productivelyexamine more deta ils of this identity-basedemotion regulation strategy, and compare itseffectiveness against other strategies on whichthe literature has focused (Gross, 1998).

IET encourages theorists and researchers tolook into social influences (especially influ-ences of the intergroup context) on all aspectsof emotional responding. Whereas a number oftheorists (e.g., Smith & Lazarus, 1991;Mesquita, 2003) have argued that emotions areinherently relational or interpersonal in nature,our argument is slightly different—that emo-tions arise from group memberships or collec-tive (not only relational) aspects of the self(Sedikides & Brewer, 2001). Group member-ships may influence all aspects of emotional re-sponding: changes in core affect, the causal at-tributions people make for such changes, theway they categorize their emotional experi-ences, and the types of actions that they mayseek to perform when in emotional states. In-vestigations of these issues may clarify the waysemotions are functional at the level of groupsand not solely individuals. Indeed, such group-level functionality is to be expected, for group

living is evolutionarily ancient in humans aswell as related primate species, and it would besurprising indeed if emotions did not play arole in shaping our feelings about and reactionsto other ingroup members as well as out-group members (Cottrell & Neuberg, 2005;Caporael, 1997).

We close with one obvious point. For emo-tion researchers, intergroup relations offers apowerful venue in which to investigate emo-tions and their effects. This is a domain of so-cial life in which emotions are often intense,compelling—and highly consequential (with ef-fects including persecutions, pogroms, andgenocides). It is also a domain in which, as weargue, emotions are often driven by people’sself-identification with important socialgroups, rather than by their individual selves.Increases in our understanding of emotionsthemselves, as well as understanding of the so-cial phenomena to which they contribute, mayresult from an increased focus on social groupmemberships as bases for emotions.

ACKNOWLEDGMENTS

Preparation of this chapter was supported by GrantNo. R01 MH-63762 from the National Institute ofMental Health. We are grateful to Lisa FeldmanBarrett, Angie Maitner, and Charlie Seger for helpfulcomments and assistance.

REFERENCES

Barrett, L. F. (2006). Solving the emotion paradox: Cat-egorization and the experience of emotion. Personal-ity and Social Psychology Review, 10(1), 20–46.

Berkowitz, L. (1998). Affective aggression: The role ofstress, pain, and negative affect. In R. G. Geen & E.Donnerstein (Eds.), Human aggression: Theories, re-search, and implications for social policy (pp. 49–72). San Diego: Academic Press.

Branscombe, N. R., & Doosje, B. (Eds.). (2004). Collec-tive guilt: International perspectives. Cambridge,UK: Cambridge University Press.

Caporael, L. R. (1997). The evolution of truly socialcognition: The core configurations model. Personal-ity and Social Psychology Review, 1, 276–298.

Cottrell, C. A., & Neuberg, S. L. (2005). Different emo-tional reactions to different groups: A sociofunc-tional threat-based approach to “prejudice.” Jour-nal of Personality and Social Psychology, 88, 770–789.

DeSteno, D., Dasgupta, N., Bartlett, M. Y., & Cajdric,A. (2004). Prejudice from thin air: The effect of emo-


tion on automatic intergroup attitudes. PsychologicalScience, 15, 319–324.

Dijker, A. J. (1987). Emotional reactions to ethnic mi-norities. European Journal of Social Psychology, 17,305–325.

Doosje, B., Branscombe, N. R., Spears, R., &Manstead, A. S. R. (1998). Guilty by association:When one’s group has a negative history. Journal ofPersonality and Social Psychology, 75, 872–886.


Frijda, N. H., Kuipers, P., & ter Schure, E. (1989). Rela-tions among emotion, appraisal, and emotional ac-tion readiness. Journal of Personality and Social Psy-chology, 57, 212–228.

Gaertner, S. L., & Dovidio, J. F. (1986). The aversiveform of racism. In J. F. Dovidio & S. L. Gaertner(Eds.), Prejudice, discrimination, and racism (pp. 61–90). Orlando, FL: Academic Press.

Gross, J. J. (1998). Antecedent- and response-focusedemotion regulation: Divergent consequences for ex-perience, expression, and physiology. Journal of Per-sonality and Social Psychology, 74, 224–237.

Hogg, M. A., & Turner, J. C. (1987). Intergroup behav-ior, self-stereotyping and the salience of social catego-ries. British Journal of Social Psychology, 26, 325–340.

Kessler, T., & Hollbach, S. (2005). Group-based emo-tions as determinants of ingroup identification. Jour-nal of Experimental Social Psychology, 41, 677–685.

Leach, C. W., & Iyer, A. (2006). Anger and guilt aboutin-group advantage explain the willingness for politi-cal action. Personality and Social Psychology Bulle-tin, 32, 1232–1245.

Mackie, D. M., Devos, T., & Smith, E. R. (2000). Inter-group emotions: Explaining offensive action tenden-cies in an intergroup context. Journal of Personalityand Social Psychology, 79, 602–616.

Mackie, D. M., Silver, L. A., & Smith, E. R. (2004). In-tergroup emotions: Emotion as an intergroup phe-nomenon. In L. Z. Tiedens & C. W. Leach (Eds.), Thesocial life of emotions (pp. 227–245). Cambridge,UK: Cambridge University Press.

Mackie, D. M., & Smith, E. R. (Eds.). (2002). Fromprejudice to intergroup emotions: Differentiated re-actions to social groups. New York: PsychologyPress.

Maitner, A. T., Mackie, D. M., & Smith, E. R. (2006).Evidence for the regulatory function of intergroupemotion: Emotional consequences of implemented orimpeded intergroup action tendencies. Journal of Ex-perimental Social Psychology, 42, 720–726.

Maitner, A. T., Mackie, D. M., & Smith, E. R. (2007).Antecedents and consequences of satisfaction andguilt following ingroup aggression. Group Processesand Intergroup Relations, 10, 223–237.

Mesquita, B. (2003). Emotions as dynamic cultural phe-nomena. In R. J. Davidson, K. R. Scherer, & H. H.Goldsmith (Eds.), Handbook of affective sciences(pp. 871–890). New York: Oxford University Press.

Miller, D. A., Smith, E. R., & Mackie, D. M. (2004). Ef-fects of intergroup contact and political predisposi-tions on prejudice: Role of intergroup emotions.Group Processes and Intergroup Relations, 7, 221–237.

Neuberg, S. L., & Cottrell, C. A. (2002). Intergroupemotions: A biocultural approach. In D. M. Mackie& E. R. Smith (Eds.), From prejudice to intergroupemotions: Differentiated reactions to social groups(pp. 265–283). New York: Psychology Press.

Neumann, R., & Strack, F. (2000). “Mood contagion”:The automatic transfer of mood between persons.Journal of Personality and Social Psychology, 79,211–223.

Oakes, P., Haslam, S. A., & Turner, J. C. (1994). Stereo-typing and social reality. Oxford: Blackwell.

Pettigrew, T. F. (1998). Intergroup contact theory. An-nual Review of Psychology, 49, 65–85.

Roseman, I. J. (1984). Cognitive determinants of emo-tion: A structural theory. Review of Personality andSocial Psychology, 5, 11–36.

Runciman, W. G. (1966). Relative deprivation and so-cial justice. Berkeley: University of California Press.



Scherer, K. R. (1984). Emotion as a multicomponentprocess: A model and some cross-cultural data. Re-view of Personality and Social Psychology, 5, 37–63.

Sedikides, C., & Brewer, M. B. (Eds.). (2001). Individ-ual self, relational self, collective self. Philadelphia:Psychology Press.

Simon, B., & Hamilton, D. L. (1994). Self-stereotypingand social context: The effects of relative in-groupsize and in-group status. Journal of Personality andSocial Psychology, 66, 699–711.

Skitka, L. J., Bauman, C. W., & Mullen, E. (2004). Po-litical tolerance and coming to psychological closurefollowing the September 11, 2001 terrorist attacks:An integrative approach. Personality and Social Psy-chology Bulletin, 30, 743–756.

Smith, C. A., & Lazarus, R. S. (1991). Appraisal com-ponents, core relational themes, and the emotions.Cognition and Emotion, 7, 233–269.

Smith, E. R. (1993). Social identity and social emotions:Toward new conceptualizations of prejudice. In D.M. Mackie & D. L. Hamilton (Eds.), Affect, cogni-tion, and stereotyping: Interactive processes in groupperception (pp. 297–315). San Diego: AcademicPress.

Smith, E. R. (1999). Affective and cognitive implicationsof a group becoming part of the self: New models ofprejudice and of the self-concept. In D. Abrams & M.A. Hogg (Eds.), Social identity and social cognition(pp. 183–196). Oxford: Blackwell.

Smith, E. R., & Henry, S. (1996). An in-group becomes


part of the self: Response time evidence. Personalityand Social Psychology Bulletin, 22, 635–642.

Smith, E. R., & Mackie, D. M. (2006). It’s about time:Intergroup emotions as time-dependent phenomena.In R. Brown & D. Capozza (Eds.), Social identities:Motivational, emotional, cultural influences(pp. 173–187). New York: Psychology Press.

Smith, E. R., Seger, C., & Mackie, D. M. (2007). Canemotions be truly group-level? Evidence regardingfour conceptual criteria. Journal of Personality andSocial Psychology, 93, 431–446.

Stephan, W. G., & Stephan, C. W. (1985). Intergroupanxiety. Journal of Social Issues, 41, 157–175.

Tajfel, H. (1978). Differentiation between social groups:Studies in the social psychology of intergroup rela-tions. London: Academic Press.

Tiedens, L. Z., & Linton, S. (2001). Judgment underemotional certainty and uncertainty: The effects ofspecific emotions on information processing. Journalof Personality and Social Psychology, 81, 973–988.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D.,& Wetherell, M. S. (1987). Rediscovering the socialgroup: A self-categorization theory. Oxford:Blackwell.

Watson, D., & Clark, L. A. (1992). On traits and tem-perament: General and specific factors of emotionalexperience and their relation to the five-factor model.Journal of Personality, 60, 441–476.

Yzerbyt, V. Y., Dumont, M., Wigboldus, D., & Gordijn,E. (2003). I feel for us: The impact of categorizationand identification on emotions and action tendencies.British Journal of Social Psychology, 42, 533–549.


C H A P T E R 2 7

Empathy and Prosocial Behavior

MARTIN L. HOFFMAN

The interest in empathy is an offshoot of re-search on helping others in distress that tookoff with the brutal murder of a young womanin Queens, New York in the mid-1960s, in fullview of 30 people watching from their apart-ments. The early research on why bystandersdon’t help focused on the context and foundthat the presence of others can interfere withone’s helping by activating assumptions of “dif-fusion of responsibility” (“I’m sure someonecalled the police”) or “pluralistic ignorance”(“No one is doing anything, so it can’t be anemergency”).

Subsequent research has focused onprosocial motivation, mainly affective empa-thy. This chapter is concerned with empathy’sarousal, development, and contribution tohelping. It highlights the role of cognition inarousing empathy, developing it, and shaping itinto related emotions (sympathy, empathic an-ger and feeling of injustice, guilt). It brieflysummarizes gender differences and the role ofsocialization, and points up empathy’s limita-tions. Finally, new directions for research aresuggested.

EMPATHY DEFINED

“Empathy” can be defined as an emotionalstate triggered by another’s emotional state orsituation, in which one feels what the otherfeels or would normally be expected to feel inhis situation. Since empathic moral issues in-volve people in distress (pain, danger, poverty),our primary concern is empathic distress.Finally, mature empathy is metacognitive: Oneis aware of empathizing—that is, one feels dis-tressed but knows this is a response to an-other’s misfortune, not one’s own. One also hasa sense of how one as well as others might feelin the victim’s situation, and awareness that avictim’s outward behavior and facial expres-sion may not reflect how he or she feels.

EMPATHY AS A MORAL MOTIVE

One might expect some people to feel pleasureat another’s misfortune, and some do undercertain conditions, such as anger, dislike, andcompetition (Hareli & Weiner, 2002). The

440

overwhelming evidence, however, is that mostpeople feel empathically distressed and moti-vated to help (Hoffman, 1978, 2000; Eisenberg& Miller, 1987): The greater and the more in-tense their empathic distress is, and the moreintense the victim’s actual distress is, thequicker they are to help. Moreover, theirempathic distress decreases more quickly andthey feel better when they help than when theydon’t. Empathic distress, in short, has all the at-tributes of a prosocial moral motive.

There are also evolutionary grounds for em-pathy’s being necessary for human survival andtherefore part of human nature (Hoffman,1981). Experiments have suggested empathy’spresence in primates (Brothers, 1989), andthere is recent evidence that primates and hu-mans share “mirror neurons,” which may bethe neural substrate for mimicry (Iacoboni &Lenzi, 2002; Iacoboni et al., 2005). Finally,brain scan experiments and clinical studiesshow empathy’s neural basis (Gallese, 2003);identical-twin research supports a hereditarycomponent for empathy (Zahn-Waxler, Robin-son, Emde, & Plomin, 1992); and evidence hasbeen found for an empathy precursor at birth(Martin & Clark, 1982; Simner, 1971; Sagi &Hoffman, 1976). It seems reasonable to con-clude that human nature is not just egoistic, aslong assumed in Western psychology, but in-cludes an empathic moral dimension—a built-in “empathic morality.”

MODES OF EMPATHIC AROUSAL

Five empathy-arousing modes have been iden-tified (Hoffman, 1978, 2000). The first three—mimicry, conditioning, and direct association—are automatic and preverbal.

Mimicry

Mimicry was intuitively understood by AdamSmith (1759/1976, p. 4): “The mob, when theyare gazing at a dancer on the slack rope, natu-rally writhe and twist and balance their ownbodies as they see him do.” It was defined over100 years later by Lipps (1906) as an innate,involuntary, isomorphic response to another’sexpression of emotion that occurs in two stepsoperating in close sequence: One automaticallychanges one’s facial expression, voice, and pos-ture in synchrony with slightest changes in themodel’s facial, vocal, postural expressions of

feeling; the resulting muscle movements triggerafferent feedback to the brain, producing feel-ings in observers that match the model’s. Thediscovery of mirror neurons (noted above) sug-gests a simpler neural substrate: The same neu-ral pattern is involved in feeling an emotionand observing someone else feeling it; there-fore, observing another’s emotional expressionis all it takes to feel it. This may also explain in-fants’ imitating their mothers’ facial and handgestures and the synchronicity of changes in fa-cial expressions between infants and mothers,each responding to affect exhibited by theother (Jaffe, Beebe, Feldstein, Crown, &Jasnow, 2001; Stern, Hofer, Haft, & Dore,1985).

Conditioning

One can acquire empathic distress as a condi-tioned response when one’s actual distress ispaired with another’s expression of distress.This is inevitable in mother–infant interactions,as when a mother’s distress stiffens her bodyand is transferred to the infant in the course ofphysical handling. The mother’s facial and ver-bal expressions of distress then become condi-tioned stimuli that can subsequently evoke dis-tress in the child even in the absence of physicalcontact. They can also be generalized, so thatfacial and verbal signs of distress from anyonecan arouse distress in the infant.

Direct Association

Direct association does not require prior pair-ing with another’s distress—just having distressfeelings that can subsequently be evoked insimilar situations by another’s expression ofdistress. For example, having the experience ofbeing separated from one’s parents may helpone empathize with a friend in the midst of aseparation experience.

Empathy aroused by these three modes ispassive, involuntary, and based on surface cues;it requires little cognitive processing or aware-ness that the source of one’s distress is someoneelse’s pain. Still, these modes are important, be-cause they enable a primitive form of empathyin young infants. And since they continue be-yond infancy, they give empathy an involuntarydimension through life. They are limited, how-ever, because the victim must be present, andthey only allow empathy with very simple emo-

27. Empathy and Prosocial Behavior 441

tions. These limitations are overcome by lan-guage and cognitive development, which sup-port the two remaining modes of empathyarousal.

Verbally Mediated Association

In verbally mediated association, the victim’sdistress is communicated through language.When it is communicated only through lan-guage (a letter from the victim or someoneelse’s description of the victim’s plight), seman-tic processing is necessary to mediate the con-nection between the victim’s feeling and theobserver’s empathic response. This semanticprocessing may put distance between observerand victim; however, this distancing is reducedwhen the decoded message enables the ob-server to construct visual or auditory images ofthe victim (sad face, blood, cries, moans) andrespond empathically to these images throughthe preverbal modes. When the victim is pres-ent, the distancing effect is mitigated bypreverbal arousal modes activated directly bythe facial and other expressions of distress thataccompany the victim’s verbal communication.Or one’s empathy may first be aroused by therelatively quick-acting preverbal modes andthen fine-tuned by semantic processing.

Perspective Taking

The mode of perspective taking is not new. Da-vid Hume (1751/1957) thought that becausepeople are constituted similarly and have simi-lar life experiences, imagining oneself in an-other’s place converts the other’s situation intomental images that evoke the same feeling inoneself. Adam Smith (1759/1976, p. 261) wentfurther: “By the imagination we place ourselvesin the other’s situation, we conceive ourselvesenduring all the same torments, we enter, as itwere, into his body, and become in some mea-sure the same person with him, and thenceform some idea of his sensations, and even feelsomething which, though weaker in degree, isnot altogether unlike them.” The modern re-search, begun in the 1950s, reveals three typesof perspective taking. The first is self-focused:Imagining that the stimuli impinging on thevictim are impinging on oneself evokes anempathic response, which can be enhanced byassociation with similar events in one’s ownpast. The second is other-focused; it consists ofattending to the victim’s feelings, current life

condition, and behavior in similar situations.This may be more cognitive than affective em-pathy, except when the victim is present and af-fect is recruited from preverbal modes acti-vated by the victim’s face, voice, and posture.Research by Batson, Early, and Salvarani(1997) suggests that self-focused perspective-taking arouses more intense empathic distress(including its physiological manifestations),perhaps because self-focused perspective takingis more likely to evoke associations with pain-ful events in one’s own past. The third type ofperspective taking focuses on both self andother; it consists of co-occurring, parallel pro-cesses that benefit from the emotional intensityof self-focused and the sustained attention tothe victim of other-focused perspective taking.All three types of perspective taking are undervoluntary control and may be drawn out overtime, although they can also be triggered imme-diately on witnessing another’s distress or bypreverbal modes that draw attention to the vic-tim’s distress.

To summarize, empathic distress is amultidetermined and hence reliable prosocialmotive (Hoffman, 2000). The arousal modescan operate alone or in any combination. Thepreverbal modes allow empathic arousal in in-fants, but continue operating in childhood andadd an important involuntary dimension toempathy in adults. The cognitive modes en-large empathy’s scope to include subtle types ofdistress (e.g., disappointment) and allow empa-thy with victims who are absent. Multiplemodes thus enable one to respond empathicallyto whatever distress cues are available: Facial,vocal, and postural cues are picked up throughmimicry; situational cues through conditioningand association; distress expressed orally, inwriting, or by someone else can arouse empa-thy through the cognitive modes. Multiplemodes not only enable but often compel one torespond empathically—instantly, automati-cally, with or without conscious awareness.Even the cognitive modes, often drawn out andvoluntarily controlled, can kick in immediatelyif one attends closely to the victim.

EMPATHIC OVERAROUSAL

Though empathic distress increases with the in-tensity of victims’ distress it can become soaversive—a condition I have termed “empathic


overarousal” (EOA)—that bystanders shift at-tention to their own distress, leave the victim,or think of other things to turn off the image ofthe victim (Hoffman, 1978). Strayer (1993)showed 5- to 13-year-olds film clips of dis-tressed children (one child was unjustly pun-ished; another was forcibly separated fromfamily; a disabled child was climbing stairs).Subjects’ empathic distress and attention to vic-tims increased with the intensity of victims’ dis-tress until the subjects’ empathic distressreached the level of victims’ distress, afterwhich subjects’ focus shifted to themselves.Bandura and Rosenthal (1966) gave adultswatching someone being given electric shocks adrug that intensified their empathic distress,which the subjects reduced by engaging in dis-tracting thoughts and attending to lab details.

People are more vulnerable to EOA whenthey feel unable to reduce victims’ distress or toconfine empathic distress within a tolerablelevel of arousal. In one study, nursing traineesnew to hospital wards were so empathicallyoveraroused by terminally ill patients that theytried to avoid them; they changed when theyfound they could improve patients’ quality oflife (Williams, 1989). Children who exert emo-tional control and are taught coping strategiesfor handling anxiety by their parents are lessvulnerable to EOA, can keep empathic distresswithin a tolerable range, and focus more on thevictim’s than on their own distress (Eisenberg,Fabes, Schaller, Carlo, & Miller, 1991; Fabes,Eisenberg, Karbon, Troyer, & Switzer, 1994;Valiente et al., 2004).

VICARIOUS TRAUMATIZATIONOF CLINICIANS

I originally advanced the concept of EOA toexplain bystanders’ turning away from victims.But this does not hold for certain highly com-mitted people (“witnesses”; see below) or peo-ple whose role requires staying and helping (cli-nicians, nurses, rescue workers). There is agrowing literature on trauma clinicians’ “com-passion fatigue” and “vicarious trauma”(Figley, 1995; Pearlman & Saakvitne, 1995),which I have suggested may be due to EOA(Hoffman, 2000, 2002). I asked 125 clinicianshow they felt and coped in their last therapysession with a trauma patient.1 They reported alot of EOA, with cognitive disruptions and hor-rible images, nightmares, and physical symp-

toms, which were often hard to shake off after-ward (“I felt the sadness resurge and envelopme . . . almost impossible to concentrate andattend properly,” “My neck felt strained, tired,stomach ached, dizzy”). Proximal causes ofthis EOA were patients’ facial, vocal, and pos-tural expressions of pain (“Her tears, descrip-tion of childhood events, crying, and saying,‘Why don’t they understand how what they doaffects me?’ ”); vivid trauma narratives thatevoked painful images, especially when associ-ated with clinicians’ own past traumas (“I stillsee the picture I saw as she spoke of the manwho hurt her. He looks so malevolent. I imag-ined her small body size with a grown adult,her grimace of pain”); and some patients’ calmdemeanor that masked intense suffering (“Oh,my God, he’s speaking as though he were de-scribing the weather [drunken father threat-ened patient and mother with gun]. He’s fullydissociated from the feelings, been wounded sobadly”). The last example shows that a clini-cian’s empathic distress can be more intensethan the patient’s actual distress that evoked it.

The coping strategies clinicians reported us-ing to keep EOA under control included vari-ous information-processing techniques: gainingdistance by imagining that a patient’s traumanarrative was just a movie; splitting one’s focusso that one was partly an “objective observer”;taking time out by pushing trauma imagesaside, thinking about other things, and then re-grouping; and reminding oneself of past suc-cesses. These clinicians also used breathing andother relaxation techniques; consulted withcolleagues and supervisors; talked things overwith their therapists, spouses, and/or friends(“If I can’t get the terrible images out of mymind, I seek coworkers for debriefing”); joinedor started self-help groups; did volunteer com-munity service; or started a strenuous exerciseroutine. One clinician (a witness?—see below)took to political action on behalf of peoplewith her patient’s problem (“I pictured that ter-rified child who had been physically and sexu-ally abused, neglected, being threatened withremoval from her foster/adoptive family. I hadthoughts of lashing out at the system, and as aresult of many such cases, I did work to changeit”).

Clinicians’ EOA may best be viewed as partof an interactive process between intenseempathic distress and attempts to control it tomaintain their professional focus. That mostclinicians stay with patients despite the some-


times intense pain of EOA suggests that traumatherapy is a prosocial moral encounter andshould perhaps be added to the five typesof such encounters previously identified(Hoffman, 2000). It also suggests that the an-swer to the question “Does empathy lead toanything more than superficial helping?”(Neuberg et al., 1997), is yes.

EMPATHY DEVELOPMENT

Mature empathy, as noted earlier, has a meta-cognitive dimension. This requires having acognitive sense of oneself and others as sepa-rate beings with independent inner states (feel-ings, thoughts, perceptions) that are only partlyreflected in outward behavior, and with sepa-rate identities and life conditions. Before 4 or 5years of age, one can empathize, but with littleor no metacognitive awareness. This suggestedto me that empathy develops along with the de-velopment of cognitive self–other concepts, insix stages (Hoffman, 1978, 2000).

Global Empathic Distress:Newborn Reactive Cry

The well-known cry in response to the sound ofanother’s cry by alert, content newborns is notsimply imitation of the cry sound or a painfulreaction to a noxious stimulus. It is vigorous,intense, and identical to spontaneous cries ofinfants in actual distress (Sagi & Hoffman,1976; Simner, 1971). Martin and Clark (1982)found the same thing and also that infants cryless in response to the sound of their own cry(tape-recorded as in the previous cry research)or to a chimpanzee’s (which adults find moreaversive than human infant cries). This maymean that the cry is an innate, isomorphic re-sponse to a conspecific’s cry, which presumablysurvived natural selection. The underlyingmechanism could be mimicry, which wouldsuggest that mirror neurons are in place atbirth; or it could be conditioning, which new-borns are also capable of (Blass, Ganchrow, &Steiner, 1984).

Whatever the cause, newborns respond toanother’s distress by feeling distressed. Thissuggests that the newborn cry may be the firstinstance of empathy without awareness. Wemight expect it to be undermined by around 6months, due to the dawning awareness of oth-ers as physically separate from oneself. Indeed,

Hay, Nash, and Pedersen (1981) found that 6-month-olds looked at a crying child withoutimmediately crying; in response to prolongedcries, they looked sad and puckered up theirlips before starting to cry.

Egocentric Empathic Distress

By 11–12 months, infants do the same as Hayet al.’s (1981) 6-month-olds, but they alsowhimper and silently watch the victim (Radke-Yarrow & Zahn-Waxler, 1984). Some act, buttheir actions seem designed to reduce their owndistress. A 1-year-old daughter of a student ofmine saw a friend fall and cry, stared at thefriend, began to cry, then put her thumb in hermouth and buried her head in her mother’slap—as she did when she hurt herself. A parsi-monious explanation would be that like mostinfants her age, she still hadn’t fully “gradu-ated” from global empathic distress and re-mained unclear about the difference betweensomething happening to another and to herself.Distress cues from others still elicited globalempathic distress—a fusion of unpleasant feel-ings and stimuli from the dimly perceivedother, her own body, and the situation. She feltupset and sought to comfort herself. I call thistype of reaction “egocentric empathic dis-tress,” because it is both egocentric (there is amotive to reduce one’s own distress) andempathic (it is contingent on another’s dis-tress). The contingency is what justifies callingit a precursor of empathic morality.

Quasi-Egocentric Empathic Distress

About 2 months later, still early in the secondyear, children’s empathic crying, whimpering,and staring become less frequent. They beginmaking helpful advances toward a victim (ten-tative physical contact such as a pat or touch),which soon give way to more differentiatedpositive interventions: kissing, hugging, gettingsomeone else to help, physically assisting, ad-vising, reassuring. A 14-month-old boy re-sponded to a crying friend with a sad look, andthen gently took the friend’s hand and broughthim to his own mother, although the friend’smother was present (Hoffman, 1978). A 15-month-old girl “watches a visiting baby who iscrying: she watched him carefully, followedhim around, kept handing him toys and otheritems she’s fond of (her bottle, a string ofbeads)” (Radke-Yarrow & Zahn-Waxler,


1984). These actions showed that the childrennow realized that others were physical entitiesindependent of themselves, though they did notyet grasp that others have their own indepen-dent inner states. The actions were clearly de-signed to help another in distress and thusshowed empathic distress operating as aprosocial motive.

Beginning wiht sympathetic distress, whathappens when children realize that it is anotherwho in distressed? They may feel relieved andturn away, but the mental image of the victimmay still feed their empathic distress. In anycase, the evidence is that children generallycontinue to feel distressed and offer help. Thecombination of feeling distressed, knowing thisfeeling is due to another’s actual distress, andtrying to help suggested to me (Hoffman,1975a) that empathic distress is changing froma feeling more or less like the victim’s feeling toa more reciprocal feeling of concern for the vic-tim. This fits how most people report they feelwhen seeing someone in distress: They feel dis-tressed themselves, but also feel “sympatheticdistress” and a desire to help the victim. In anycase, from then on through life, empathic dis-tress may include a sympathetic component;people want to help because they feel sorry forthe victim, not just to relieve their own em-pathic distress. The sympathetic component ofempathic distress may be the child’s first trulyprosocial motive.

Qualitative-change hypotheses like this aredifficult to test. At first there was just anec-dotal evidence (Hoffman, 1975a), but threeindependent investigations have since pro-vided more convincing support: Self–otherdifferentiation (measured indirectly by mirrorself–other recognition) predates sympatheticdistress and helping (Bischof-Köhler, 1991;Johnson, 1992; Zahn-Waxler, Radke-Yarrow,& King, 1979). Explaining the underlyingprocess is even more difficult (see Hoffman,1976, 2007, for my attempt to do this). Inany case, from now on in this chapter, theterm “empathic distress” stands for empathic/sympathetic distress.

Veridical Empathy

Major advances in the self–other concept occurnear the end of the second year and continueinto the third. Children recognize themselves ina mirror (Lewis & Brooks-Gunn, 1979), re-vealing a sense of the body as a physical entity

that exists outside one’s subjective self and canbe seen by others. Children are also becomingaware that others have inner states (thoughts,feelings, desires) independent of their own.This allows more accurate empathy and effec-tive helping behavior. Sarah, age 2 years 3months, was riding in a car when her cousinbecame upset at losing his teddy bear. Someonesaid that it was in the trunk and could be re-trieved when they get home. About 10–15 min-utes later, when the car approached the house,Sarah said, “Now you can get your bear”(Blum, 1987).

The transition from quasi-egocentric toveridical empathic distress is illustrated by aslightly younger child, 2-year-old David, whobrought his own teddy bear to comfort a cryingfriend. When it didn’t work, he paused, ran tothe next room, and returned with the friend’steddy bear; the friend hugged it and stoppedcrying. I suggest that David was cognitively ad-vanced enough to wonder why his teddy didn’tstop his friend’s crying, to realize that he wouldwant his own teddy and so might his friend,and to remember seeing his friend playing hap-pily with his own teddy in the next room. Hisbehavior may show how cognitively ready chil-dren learn from “quasi-egocentric” mistakesand corrective feedback (in David’s case, thefriend’s continued crying) that another’s needsmay at times differ from theirs, and thus ad-vance to veridical empathy.

Children can now empathize with aware-ness, take others’ perspectives, and help moreappropriately. Veridical empathy has the basicfeatures of mature empathy, but becomes morecomplex with age. The growing understandingof causes, consequences, and correlates of emo-tions allows one to empathize not only withsimple but with subtle distress feelings: Pre-schoolers can empathize with missing one’sparents, older children with mixed feelings ordisappointment in one’s performance. Adoles-cents can empathize with another’s need for in-dependence, and even with fear of losing faceby accepting help. Moreover, from 14 to 20years of age they learn about many complex,highly specific distresses, such as failing testsimportant for one’s future, parents’ divorcing,or having less sexual experience than one’speers (Pasupathi, Staudinger, & Baltes, 2001).Apart from content, children in time realizethat their inner self has a reflective part—an“I” that thinks, feels, plans, and remembers—and that other people do too. Their empathy is


affected by knowing that people can displayemotions not felt and feel emotions not dis-played.2

Empathic Distress beyond the Situation

At some point, due to the emerging conceptionof self and others as continuous persons withseparate histories and identities, children be-come aware that others feel joy, anger, sadness,fear, and low esteem not only in a particular sit-uation, but also in their lives. Consequently,they not only respond empathically to an-other’s immediate distress, but also to whatthey imagine is the other’s chronically sad orunpleasant life.

When can they do this? The research on gen-der and ethnic identity suggests that children’ssense of self as coherent, continuous, and stableis hazy until about 6–9 years (Ruble & Martin,1998). We might expect them to know that oth-ers have identities soon afterward. This fitsGnepp and Gould’s (1985) research: They de-scribed a child’s prior experience (bitten by a ger-bil or rewarded for an excellent dive), then askedsubjects to predict the child’s emotional reactionin a related subsequent event (e.g., the previouslybitten child’s turn to feed the class gerbil). Halfthe second graders and most of the fifth graderscorrectly used the prior information (e.g., real-ized that the previously bitten child would beafraid to feed the gerbil). This suggests that theages of 7–10 years may be when knowledge ofothers’ lives starts affecting a child’s empathic re-sponses. The child can then empathize with peo-ple who are chronically ill, emotionally de-prived, or hopelessly poor—regardless of theothers’ immediate state. If they seem sad, know-ing that their lives are sad may intensify thechild’s empathic distress. If they seem happy, thecontradiction may stop the child short; ratherthan feeling empathic joy, he or she may realizethat a sad life is a more compelling index of well-being and respond with empathic sadness, orwith a mixture of joy and sadness (Szporn,2001). Mature empathy is thus a response to anetwork of cues from others’ behavior, emo-tional expression, immediate situation, and lifecondition.

Empathy for Distressed Groups

It is likely that when children are able to formsocial concepts and classify people, they cancomprehend the plight not only of an individ-

ual but also of an entire group or class of peo-ple (e.g., victims/survivors of chronic illness,poverty, the Holocaust and other instances ofethnic cleansing, hurricanes and other naturaldisasters, war, and terrorism). At empathy’shighest level, then, one can empathize not onlywith an individual’s but also with a group’s dis-tressing life condition (“empathy narrative”).These may go together, as when empathy isgeneralized from a victim to a group—for ex-ample, from the famous picture of a burnedbaby in a fireman’s arms to all of theOklahoma City bombing victims (media-enhanced empathic distress for a group). In-deed, it may be difficult to empathize with amass without first empathizing with individualvictims; then, realizing that others are in thesame boat, one can generalize one’s empathy tothe group.

As with a single victim, one can empathizewith a group’s life condition that contradicts itsbehavior. A student of mine wrote in a term pa-per:

When I read accounts of slaves in America whowere extremely religious and joyful in religiousceremonies, I feel sort of happy that they were do-ing something that gave them a sense of joy, evenecstasy, but I am reminded that they were op-pressed and this is a false sense of joy or hope inthe midst of a distressing, unfair life. I feel happythat they’re happy despite being enslaved, but Ifeel bad for them too because this religious hopeor joy is really a false sense of security. It was a bit-ter irony that they took joy from the promised sal-vation of this religion, given them by the slaveowners whom they wanted to be liberated from.

Empathy for distressed groups has influ-enced U.S. politics and law (Hoffman, 2006).Harriet Beecher Stowe, whose 1852 novel Un-cle Tom’s Cabin humanized and described theliving conditions of slaves in the South, wasmotivated to write it by intense empathic dis-tress (especially for slave mothers forcibly sepa-rated from their children) and by empathic an-ger at the Fugitive Slave Law, as shown in thissummary of an incident from the novel. An af-fluent, politically uninvolved housewife’sdeeply felt empathy for slaves she personallyknows who “have been abused and oppressedall their lives” motivates her to oppose a newlaw against giving food, clothes, or shelter toescaping slaves. She argues with her husband, agovernment official who supports the law, not-ing that the Bible says people should “feed the


hungry, clothe the naked, and comfort the des-olate . . . folks don’t run away when they arehappy, but out of suffering.” She becomes in-tensely opposed to the “shameful, wicked,abominable” law and vows to break it at theearliest opportunity (Stowe, 1852/1938,pp. 99–100).

The book served its purpose: “While no oneshould underestimate the great services of [Wil-liam Lloyd] Garrison [and other abolitionists]in their effort to free the slaves it is truth to saythat all their efforts were but a drop in thebucket compared with the stir and power thatwere in Uncle Tom’s Cabin. . . . Never in hu-man history has a work devoted to a greatcause had such an instantaneous effect” (Ward,1896). This is surely an exaggeration, but his-torians do agree that Stowe’s book, motivatedby empathic distress for slaves, did play a sig-nificant role in preparing the country psycho-logically for emancipation.

Another example is the work of YaleKamisar, a law professor known by his col-leagues as a scholar and enemy of injustice,who combined empathy with a logical, rationalapproach to constitutional law. His intense em-pathy for people accused of crimes and bulliedinto confessing by police interrogation proce-dures led to articles that gave rise to the 1966Miranda v. Arizona decision, which linkedthese procedures to the Fifth Amendment’sclause against self-incrimination and gave theaccused the right to remain silent and have alawyer present during interrogation. He con-tinued supporting Miranda after it was at-tacked by judges and Congressmen for interfer-ing with criminal investigations (Kamisar,2000).

HOW EMPATHY IS SHAPEDBY CAUSAL ATTRIBUTION

Adults spontaneously attribute causality toevents (Weiner, 1985). So do 3-year-olds, whoattribute and ask questions about causality, es-pecially regarding people’s actions, intentions,and feelings (Hickling & Wellman, 2001). Wemay therefore assume that most people attrib-ute causality when witnessing someone in dis-tress. If they blame the victim, empathic dis-tress is reduced. Otherwise, depending on theattribution, empathic distress may be trans-formed in whole or in part into (1) sympatheticdistress when the cause is unknown (as in the

developmental transition I have described ear-lier) or beyond the victim’s control (e.g., illness,accident, loss; Weiner, Graham, Stern, & Law-son, 1982); (2) empathy-based guilt when oneis the cause, when one’s efforts to help have notprevented or alleviated the victim’s distress(Batson & Weeks, 1996), or guilt over inactionwhen one has not tried to help, which wouldallow victims to suffer and which may havemotivated some 1960s civil rights activists(Keniston, 1968) and Germans who saved Jewsfrom Nazis (Oliner & Oliner, 1988); (3)empathic anger when someone else is the cause,even if the victim is distressed and not angry; or(4) empathy over injustice, when there is a dis-crepancy between a victim’s fate and what heor she deserves. The last two of these are cen-tral to morality, law, and society, but have beenlittle researched. They deserve further discus-sion, which now follows.

Empathy over Injusticeand Empathic Anger

Humans seem to have a natural preference forfairness, reciprocity, and equity (Peterson &Cary, 2002). From this preference stem the be-liefs that people should get what they deserveon the basis of performance, effort, gooddeeds, and character; that people’s legal rightsas citizens should be respected; and that pun-ishment should be commensurate with a crime.When one sees others being treated unfairly orunjustly—getting less than they deserve, beingdeprived of their rights, or being punished tooseverely—the preference for fairness is vio-lated, and this may transform empathic distressinto a feeling of injustice, including motivationto right the wrong.

If a perpetrator is involved, one may feelempathic anger toward the perpetrator,whether this is an individual, a group, thelaw, or the state. John Stuart Mill (1861/1952, p. 469) connected empathic anger, em-pathy over injustice, and laws as follows: “. . . the natural feeling of retaliation . . . ren-dered by intellect and sympathy applicable to. . . those hurts that wound us throughwounding others . . . serves as the guardianof justice.” This suggests that empathy overinjustice and empathic anger may be cruciallinks between individuals and laws by provid-ing the voices needed to uphold justice, objectto people and laws that abuse others, and beready to punish or change them.3


An example of empathy over injustice is Su-preme Court Justice Harlan’s lone dissent inPlessy v. Ferguson, which made “separate butequal” education the law in 1896. His dissentexpressed anger at racial discrimination andwhite dissembling about racial motivations. Itwas also suffused with empathic concern forsuffering black Americans, and clearly linkedthis to injustice and the law: “We boast of thefreedom enjoyed by our people. But it is diffi-cult to reconcile that boast with a law which,practically, puts the brand of servitude and deg-radation upon a large class of our fellow citi-zens . . . our equals before the law” (Kluger,1977, p. 82).

SOCIALIZATION

The role of parent has three facets: discipline,model, and nurture. The research, done mostlyin the 1960s through 1990s and reviewed else-where (Hoffman, 2000), is here summarized.The discipline method most likely to foster em-pathy, helping, and guilt over harming others isa type of reasoning called “induction,” oftenused by educated middle-class parents whenchildren harm or are about to harm someone.Inductions direct a child’s attention to theother’s distress, and may thus engage andstrengthen the empathy-arousing modes de-scribed above. By highlighting the child’s rolein causing the other’s distress, inductions alsocontribute to empathy-based guilt. Power-assertive discipline (physical force, threats,commands) is associated with low empathy,helping, and guilt, although it may be needed attimes to get children to attend to and processan induction’s message.

Parents’ behavior outside discipline encoun-ters and their explanations for it provideprosocial models that can reinforce children’sempathic proclivity, encourage causal attribu-tions that foster empathic anger and feelings ofinjustice, and legitimize helping. Examples areempathizing with people in difficult straits (e.g.,homeless persons), linking a television protago-nist’s feeling or situation with the child’s own ex-perience, and pointing up similarities among allhumans (e.g., feeling sad due to separation andloss). Such behaviors can reinforce and broadenthe scope of children’s empathic dispositions, es-pecially when someone needs help. Parentprosocial models also make children more re-ceptive to inductions—and more receptive to the

competing claims of peers, thus allowing moreamicable conflict resolutions. The constructivepeer interaction processes that result, as de-scribed by Piaget (1932) and others, may add anew and important dimension to the prosocialoutcomes of having inductive, nurturant,prosocial models as parents.

GENDER DIFFERENCES

Girls are socialized more than boys to be kindand feel responsible for others’ well-being.Girls are thus expected, increasingly with age,to be more empathic and prosocial and lesslikely to harm others. The research supportsthese expectations, but in varying degrees(Eisenberg & Fabes, 1998; Hoffman, 2001).Girls show more kindness and considerationfor others, and they are more likely to help andshare. They are less likely to harm others andmore apt to feel guilty when they do. They alsoobtain higher empathy scores, but this is clear-est from self-reports (especially questionnairemeasures where it is obvious what is being as-sessed and responses are under subjects’ con-trol). It is not as clear from unobtrusive natu-ralistic observations or physiological indices ofempathy. The empathy findings may thus re-flect subjects’ conceptions of what boys andgirls are supposed to be like, rather than (asusually assumed in self-report research) actualmemories of how they feel and behave in vari-ous situations.

The latter interpretation receives supportfrom several findings not included in previousreviews. In an early study (Hoffman, 1975b), Ifound the following in three large samples offifth and seventh graders and their parents.Girls and their mothers placed a higher valueon considering others than did boys and theirfathers. More important, girls’ and mothers’responses to a story in which a child or an adultwins a contest by cheating and gets away withit showed more concern for the absent victim(the true winner) than did boys and fathers.This suggests that females have more “judicialempathy” (see below) than males. In studies ofnewborn cries in response to a newborn’s cry(see the discussion of global empathic distressabove), both Sagi and Hoffman (1976) andSimner (1971) found that female newbornscried more than males. Although the stimulusin both studies was a female cry, it seems un-likely that this would make a difference, given


Martin and Clark’s (1982) finding that new-borns do not cry more in response to the soundof their own cry. On balance, it seems reason-able to conclude that females are more empath-ic than males; furthermore, if the newborn-cryfindings are replicated, they would suggest apossible biological basis for the difference.

EMPATHY’S LIMITATIONS

Empathy is limited by its fragility, dependenceon salience of distress cues, and observers’ rela-tionship to victims (Hoffman, 1984, 2000).First, it can be trumped by egoistic motives likefear or personal ambition. Second, as discussedabove, it can become so aversive (in EOA) thatbystanders shift attention to their own personaldistress, leave the victim, or think of otherthings to turn off the image of the victim.Third, though people empathize with almostanyone in distress, they empathize more withkin, friends, and their own ethnic group(ingroup or familiarity bias); this may not be aproblem in small homogeneous groups, exceptwhen there are multiple victims and one mustmake a choice, but it could be a serious prob-lem in complex societies when intergroup ri-valry fosters intense empathic anger towardoutgroups. Finally, people are far more likely toempathize with victims who are present thanwith those who are absent (“here-and-now”bias), probably because the preverbal empathy-arousing modes can operate only in response tovictims who are present.

Here-and-now bias has been shown experi-mentally (Batson, Klein, Highberger, & Shaw,1995; Batson, Batson, Todd, & Brummett,1995). Posner (1999) views it as a manifesta-tion of the availability heuristic in the court-room when judges

give too much weight to vivid immediate impres-sions and hence pay too much attention to thefeelings, interest, and humanity of the parties inthe courtroom and too little to absent personslikely to be affected by the decision. . . . You don’tneed much empathy to be moved by a well repre-sented litigant pleading before you. The challengeto the empathic imagination is to be moved bythinking or reading about the consequences of thelitigation for absent—often completely unknownor even unborn—others who will affected by yourdecision.

Posner calls this “judicial empathy.”

An example of here-and-now bias is thehighly publicized 1997 trial of a British nannyLouis Woodward (Hoffman, 2000). When the8-month-old child in Woodward’s care wasshaken to death, there was widespread con-demnation of her and sympathy for the child’sparents. After her trial and conviction theempathic tide shifted in her favor (empathy canalso be fickle). She became the victim and re-cipient of widespread empathic distress, partlybecause of the severe sentence. The retrialjudge let her off, saying, “ . . . let us bring thismatter to a compassionate conclusion.” Theabsent victim was forgotten.

Empathy is important; I view it as the bed-rock of prosocial morality and the glue of soci-ety. It does have the limitations describedabove, however, which must be recognized anddealt with to maximize its contribution tomoral behavior and social life. I have madesuggestions along these lines, which I thinkwould be out of place here (Hoffman, 1987,2000).

NEW DIRECTIONS FOR RESEARCH

Depth of Empathic Feeling

“Depth of empathic feeling” is a concept thathas been overlooked by psychology. It pertainsto empathy’s intensity, its duration, and the ex-tent to which it penetrates one’s motive systemand changes behavior. Kaplan (2005, 2006),working on film- and TV-mediated empathicresponses to trauma, has introduced two con-cepts at the extremes of the depth continuum:“empty empathy” and “witnessing.” Emptyempathy results from brief exposure to traumaimages presented in rapid succession and al-lowing only fleeting empathic responses, eachcancelled by the next and thus devoid of moti-vation to help the victim or victims. In witness-ing, exposure to another’s trauma overwhelmsone with empathic distress; one experiencesEOA, but, instead of turning away, is trans-formed in a prosocial direction—intensely mo-tivated to help, often beyond the immediate sit-uation and at great personal cost. Between thetwo is vicarious trauma, not only in cliniciansbut in anyone. Kaplan views all three as mani-festations of living in a “trauma culture,” con-stantly bombarded with media images of peo-ple worldwide being traumatized by wars andnatural disasters (e.g., any night’s news onCNN).


Kaplan’s witnessing exemplar is SusanSontag’s self-described experience of shock,numbness, and “being changed forever” by im-ages of atrocity at age 12 when first exposed toHolocaust photos. It affected her life work,which often focused on mass suffering, culmi-nating in her last book Regarding the Pain ofOthers (Sontag, 2003). As a second exemplar, Iadd Harriet Beecher Stowe, cited earlier. Herbig change came when her favorite son died:

It was at his dying bed and at his grave when Ilearned what a poor slave mother may feel whenher child is torn away from her. In those depths ofsorrow which seemed to me immeasurable, it wasmy only prayer to God that such anguish mightnot be suffered in vain. . . . I felt I could never beconsoled for it unless this crushing of my ownheart might enable me to work out some greatgood to others. I allude to this here because I haveoften felt that much that is in that book [UncleTom’s Cabin] had its roots in the bitter sorrow ofthat summer. It has left now, I trust, no trace onmy mind except a deep compassion for the sor-rowful, especially mothers who are separatedfrom their children. (Stowe, 1852)

A third exemplar is Lyndon B. Johnson, whohad “deeply felt empathy for the plight ofAfrican-Americans since emancipation andcontinuing through World War II into the1950s,” and for poor “dark-skinned” people ingeneral; his empathy came from personal expe-rience with them (Caro, 2002). At 21, he spenta year teaching Mexican children in SouthTexas brush country. He visited their homes,saw their poverty, and learned that their fatherswere paid slave wages by Anglo farmers.

“I saw hunger in their eyes and pain in their bod-ies. Those little brown bodies had so little andneeded so much. . . . I could never forget the dis-appointment in their eyes and the quizzical ex-pression on their faces . . . they seemed to be ask-ing why don’t people like me? Why do they hateme because I am brown?” (Quoted in Caro, 2002,p. 720)

Besides teaching, he tried to help them (gettingthe school board to buy play equipment, ar-ranging games with other schools), but theirlife circumstances interfered. Johnson’sempathic anger and feelings of injustice com-bined to fuel a promise of future action on theirbehalf: “I swore then and there that if I everhad a chance to help those underprivileged kidsI was going to do it.” That, he said later, was

where his dream began of an America “whererace, religion, color, and language didn’t countagainst you”—long before he was in a positionto act on it.

As a U.S. Senator, Johnson’s empathy (plus,of course, his drive, extraordinary persuasiveskills, and personal ambition) enabled him,against relentless opposition from Southerncolleagues, to get America’s first civil rights leg-islation passed. He later backtracked when hisstrong civil rights stand conflicted with his goalto become President. But even as President heappointed the first African American to the Su-preme Court and won a major addition to thecivil rights laws he had obtained in Congress—the Fair Housing Act, which he hoped wouldsupplement school desegregation and endghettoization of African Americans. His do-mestic programs were, of course, underfundedbecause of the Vietnam War. Still, his accom-plishments illustrate empathy’s potential im-pact on law (and society) when it is deeply andenduringly felt and allied with the egoistic mo-tives of a person in power, as well as its fragilitywhen it conflicts with those same motives andis opposed by powerful social and politicalforces.

Although empty empathy may be a strictlyfilm- and TV-mediated phenomenon, witness-ing and depth of feeling in general, in real aswell as mediated contexts, are worth intensivestudy by psychologists because of their poten-tial influence on law, society, and history. Laub(1995) has made a start. Due to space limita-tions, I can only refer you to Laub’s andKaplan’s interesting analyses of circumstancesfavorable to witnessing. As for witnessing’splace in empathy theory, it can easily be incor-porated into the most advanced stage of empa-thy development: empathy with a distressedgroup; or added as a seventh stage. And depthof feeling should perhaps be incorporated intothe definition of empathy.

Empathy and Legal Judgment

For decades, legal scholars and philosophersfrom Kant to Rawls assumed that the law andits underlying justice principles are, and shouldbe, cleansed of emotion so that reason andlogic can prevail. Recently, however, legalscholars have acknowledged that emotions in-evitably influence not only legal judgments anddecisions by jurors and judges, but at times


law’s very substance. This literature revolvesheavily around empathy—for plaintiffs ordefendants, individuals or groups. Some,including long-time Supreme Court JusticeBlackmun, support empathy’s role in law be-cause it takes people’s needs into account andprovides a humane counterpoint to cold, ab-stract argumentation; furthermore, it can fillgaps and provide information needed to carryout the full intent and spirit of the law (Green-house, 2005; Zipursky, 1990). Others disagreeand argue forcefully that except under certainunusual circumstances, reasonable, just, andpredictable law cannot tolerate empathy(Bandes, 1996).

I have contributed to this literature and haveattempted to broaden empathy’s scope by situ-ating the law in an empathy framework; sug-gesting an empathic base for a sense of justice;giving examples of empathy’s contribution tomaking and changing laws, as well as to judg-ment and decision making in applying laws;and describing individuals who helped pave theway for repealing unjust laws and initiatingfairer ones (Hoffman, 2006). Some of thiswork has been covered in the earlier discus-sions of empathy for distressed groups, empa-thy over injustice, empathy’s here-and-nowbias in courtrooms, and witnesses who helpedprepare the country for passing and changinglaws. I now add a discussion of some empathy-related issues and concepts in Supreme Courtdeliberations.

Empathy Narratives

Empathy contributed importantly to theCourt’s unanimous 1954 decision in Brown v.Board of Education, which overturnedseparate-but-equal doctrine and made desegre-gation of public schools the law. Some writers(Henderson, 1985) see the Brown decision as“traditional legality” clashing with and ulti-mately being transformed by empathy. Here’swhat happened in Brown (which, as heard be-fore the Supreme Court, combined cases fromSouth Carolina, Virginia, Delaware, and Wash-ington, D.C. with the actual Brown case fromKansas). First, the National Association for theAdvancement of Colored People (NAACP) re-lied less on conventional legal argumentationthan on empathy narratives to show how SouthCarolina’s school segregation policy in particu-lar destroyed black children’s self-respect,“stamped them with a badge of inferiority . . .

put up road blocks in their minds.” TheNAACP also employed expert social sciencetestimony, including the famous study of blackchildren who preferred white dolls and labeledblack dolls as “bad,” to describe the nature ofthe humiliation and self-hatred caused by seg-regation. South Carolina’s main argument, onseparate-but-equal legal grounds, was that thestate had made every effort and successfully“wiped out all inequalities between its whiteand colored schools (equal funding, class size,etc) . . . and this ended the matter under thelaw” (Henderson, 1985, p. 1598). Their re-sponse to the empathy narratives was essen-tially to blame the victim: If segregationstamped blacks with feeling of inferiority, thatwas because they chose to construe it that way,and the state lacked the power to deal with psy-chological reactions to segregation.

The Court’s opinion is another example ofempathy over injustice—empathy in this caselinked to the Constitution’s equal protectionclause and to the segregation law’s having theeffect of intensifying harm to victims. Empa-thy narratives helped the judges clarify a legalconcept—namely, showing that the prior, ac-cepted separate-but-equal principle was actu-ally violated. Regarding long-term impact, theimage of children in segregated schools pre-ferring white dolls to black dolls still reso-nates as a lasting symbol of the opinion, de-spite evidence that black children in Northernstates did the same thing. This and otherchallenges have not diminished the dollstudy’s powerful imagery (see Rich, 2004–2005).

The role of empathy gets more complexwhen there are conflicting claimants, as in the1973 case of Roe v. Wade. Amicus briefs andoral arguments were loaded with empathy nar-ratives of the horrible effects of unwanted preg-nancies not only on women, but also on fe-tuses, whom antiabortion lawyers humanizedby calling them “unborn children, human be-ings, the true silent majority that needs some-one to speak for them and protect their rights.”The Justices voted to allow abortion; however,most of them framed the legal issue in termsnot of empathy, but of women’s broad consti-tutional rights and the rights and expertise ofmedical professionals. Subsequent majorityopinions undermined Roe in bits and pieces(e.g., abolishing federal funding for abortionsexcept when a woman’s life was in danger).

Finally, in Thornburg v. American College of


Obstretions (1986), a bare majority staved offan amicus attempt by President Reagan to haveRoe overturned. The Justices were influencedby a National Abortion Rights Action Leaguebrief made up largely of letters by women whoanonymously told stories of their own abortionexperiences, including horror tales of abortionsbefore Roe and narratives of women having toleave jobs, quit school, or marry. These wereempathy narratives, but with a legal equal pro-tection dimension: The right to choose abor-tion would enable women to enjoy, like men,the right to fully use the powers of their mindsand bodies. The Court’s majority opinion ac-knowledged some of these empathy narratives,and its conclusion linked empathy to legal con-cepts:

. . . the Constitution embodies a promise that acertain private sphere of individual liberty will bekept largely beyond the reach of government.That promise extends to women as well as men.Few decisions are more personal and intimate,more properly private, or more basic to individualdignity and autonomy, than a woman’s decision—with the guidance of her physician and within thelimits specified in Roe—whether to end her preg-nancy. A woman’s right to make that decisionfreely is fundamental. Any other result . . . wouldprotect inadequately a central part of the sphereof liberty that our law guarantees equally to all.(Henderson, 1985, pp. 1634–1635)

Empathy narratives may thus have contributedto saving abortion rights.

Victim Impact Statements

Lawyers use victim impact statements in crimi-nal trials to call attention to victims’ suffering. Ihave noted earlier how Yale Kamisar’s empathyhelped criminally accused persons in police in-terrogations. Empathy also helps criminally ac-cused individuals in court when evidence ispresented for their good character and unfortu-nate life circumstances. Arousing empathy forthe victim, however, can do the accused greatharm. Consider this statement by a womanwhose daughter and granddaughter were mur-dered: “He cries for his mom. He doesn’t un-derstand why she doesn’t come home. And hecries for his sister Lacie. He comes to me manytimes during the week and asks, Grandma, doyou miss my Lacie. I tell him yes. He says I’mworried about my Lacie” (quoted in Bandes,1996, p. 361).

Should jurors be allowed to hear suchempathy-arousing testimony? The SupremeCourt in 1991 ruled that they should. Some le-gal scholars weigh the pros and cons, and sayon balance, yes, mainly to counter “the paradeof witnesses who testify to the defendant’scharacter or pressures beyond normal experi-ence that drove him to commit his crime”(Bandes, 1996)—by allowing victims or theirfamilies to present the full reality of human suf-fering the defendant has produced. Others sayno, because victim impact statements may ap-peal to hatred, vengeance, or even bigotry, ormay diminish juries’ ability to process evidencebearing on defendants’ guilt or innocence, andare unnecessary because juries naturally empa-thize with victims. For a discussion of the is-sues, see Blume (2003).

Clearly, there is important work to be doneregarding empathy and the law. Moreover, ifempathy has played a role in emancipation, de-segregation, civil rights, and abortion laws,then it surely has something to say about his-tory, culture, and public policy. Not the lastword, of course—empathy can’t override pow-erful economic and political forces, ethnic divi-sions, natural disasters, or personal ambition—but Harriet Beecher Stowe’s empathy may haveadvanced emancipation of slaves by months,and Lyndon B. Johnson’s empathy may haveadvanced civil rights legislation by years. Legalscholarship now recognizes empathy’s impor-tance; shouldn’t psychology research get onboard?

NOTES

1. Tatiana Freedman helped construct the measures andcollect the data.

2. Clinicians may hide empathic grief to allow patientsto express negative feelings toward the deceased.

3. In multicultural contexts, however, empathic angermay add fuel to interethnic rivalry and contribute toviolence.

REFERENCES

Bandes, S. (1996). Empathy, narrative, and victim im-pact statements. University of Chicago Law Review,63, 361–379.

Bandes, S. A. (1999). The passions of law. New York:New York University Press.

Bandura, A., & Rosenthal, T. L. (1966). Vicarious clas-


sical conditioning as a function of arousal level. Jour-nal of Personality and Social Psychology, 3, 54–62.

Batson, C. D., Batson, J. G., Todd, R. M., & Brummett,B. H. (1995). Empathy and the collective good: Car-ing for one of the others in a social dilemma. Journalof Personality and Social Psychology, 68, 619–631.

Batson, C. D., Early, S., & Salvarani, G. (1997). Per-spective taking: Imagining how another feels versusimagining how you would feel. Personality and So-cial Psychology Bulletin, 23, 751–758.

Batson, C. D., Klein, T. R., Highberger, L., & Shaw, L.L. (1995). Immorality from empathy-induced altru-ism. Journal of Personality and Social Psychology,68, 1042–1054.

Batson, C. D., & Weeks, J. L. (1996). Mood effects ofunsuccessful helping: Another test of the empathy–altruism hypothesis. Personality and Social Psychol-ogy Bulletin, 22, 148–157.

Bischof-Köhler, D. (1991). The development of empathyin infants. In M. Lamb & M. Keller (Eds.), Infant de-velopment: Perspectives from German-speakingcountries (pp. 245–273). Hillsdale, NJ: Erlbaum.

Blass, E. M., Ganchrow, J. R., & Steiner, J. E. (1984).Classical conditioning in newborn humans 2–48hours of age. Infant Behavior and Development, 7,223–235.

Blum, L. A. (1987). In J. Kagan & S. Lamb (Eds.), Theemergence of morality in young children. Chicago:University of Chicago Press.

Blume, J. H. (2003). Ten years of Payne: Victim impactevidence in capital cases. Cornell Law Review, 88,257–281.

Brothers, L. (1989). A biological perspective on empa-thy. American Journal of Psychiatry, 146(1), 10–19.

Caro, R. A. (2002). Master of the Senate. New York:Knopf/Random House.

Eisenberg, N., & Fabes, R. A. (1998). Prosocial devel-opment. In W. Damon (Series Ed.) & N. Eisenberg(Vol. Ed.), Handbook of child psychology: Vol. 3. So-cial, emotional, and personality development (5thed., pp. 701–778). New York: Wiley.

Eisenberg, N., Fabes, R. A., Schaller, M., Carlo, G., &Miller, P. A. (1991). Parental characteristics andpractices, and children’s emotional responding. ChildDevelopment, 62, 1393–1408.

Eisenberg, N., & Miller, P. (1987). Relation of empathyto prosocial behavior. Psychological Bulletin, 101,91–119.

Fabes, R. A., Eisenberg, E., Karbon, M., Troyer, D., &Switzer, G. (1994). The relations of children’s emo-tion regulation to their vicarious emotional responsesand comforting behaviors. Child Development, 65,1678–1693.

Figley, C. R. (1995). Coping with secondary traumaticstress disorder in those who treat the traumatized.New York: Brunner/Mazel.

Gallese, V. (2003). The roots of empathy: The sharedmanifold hypothesis and the neural basis ofintersubjectivity. Psychopathology, 36, 171–180.

Gnepp, J. C., & Gould, M. E. (1985). The development

of personalized inferences: Understanding other peo-ple’s emotional reactions in light of their priorexperiences. Child Development, 56, 1455–1464.

Greenhouse, L. (2005). Becoming Justice Blackmun.New York: Times Books.

Hareli, S., & Weiner, B. (2002). Dislike and envy as an-tecedents of pleasure at another’s misfortune. Moti-vation and Emotion, 26, 257–277.

Hay, D. F., Nash, A., & Pedersen, J. (1981). Responsesof six-month-olds to the distress of their peers. ChildDevelopment, 52, 1071–1075.

Henderson, L. N. (1985). Legality and empathy. Michi-gan Law Review, 85, 1574–1653.

Hickling, A. K., & Wellman, H. M. (2001). Emergenceof children’s causal explanations and theories: Evi-dence from everyday conversation. DevelopmentalPsychology, 37, 668–683.

Hoffman, M. L. (1975a). Developmental synthesis ofaffect and cognition and its implications for altruisticmotivation. Developmental Psychology, 11, 607–622.

Hoffman, M. L. (1975b). Sex differences in moral inter-nalization and values. Journal of Personality and So-cial Psychology, 32, 720–729.

Hoffman, M. L. (1976). Empathy, role-taking, guilt,and the development of altruism. In T. Lickona (Ed.),Morality: Theory, research, and social issues(pp. 124–143). New York: Holt, Rinehart &Winston.

Hoffman, M. L. (1978). Empathy: Its development andprosocial implications. In C. B. Keasey (Ed.), Ne-braska Symposium on Motivation (Vol. 25, pp. 169–218). Lincoln: University of Nebraska Press.

Hoffman, M. L. (1981). Is altruism part of human na-ture? Journal of Personality and Social Psychology,40, 121–137.

Hoffman, M. L. (1984). Empathy, its limitations, and itsrole in a comprehensive moral theory. In J. Gewirtz& W. Kurtines (Eds.). Morality, moral development,and moral behavior (pp. 283–302). New York:Wiley.

Hoffman, M. L. (1987). The contribution of empathy tojustice and moral judgment. In N. Eisenberg & J.Strayer (Eds.), Empathy and its development(pp. 47–80). New York: Cambridge University Press.

Hoffman, M. L. (2000). Empathy and moral develop-ment: Implications for caring and justice. New York:Cambridge University Press.

Hoffman, M. L. (2001). Prosocial behavior and empa-thy: Developmental processes. In N. Smelser & P.Baltes (Eds.). International encyclopedia of the socialand behavioral sciences (pp. 12230–12233). Amster-dam: Elsevier.

Hoffman, M. L. (2002). Empathy and vicarioustraumatization in clinicians. Unpublished manu-script, presented in part to Psychology Department,Simon Fraser University.

Hoffman, M. L. (2006). Empathy, justice, and the law.Unpublished manuscript, presented in part at the


Fullerton State University conference on empathy,June 2006.

Hoffman, M. L. (2007). The origins of empathic moral-ity in toddlerhood. In C. A. Brownell & C. B. Kopp(Eds.), Socioemotional development in the toddleryears (pp. 132–148). New York: Guilford Press.

Hume, D. (1957). An inquiry concerning the principleof morals. New York: Liberal Arts Press. (Originalwork published 1751)

Iacoboni, M., & Lenzi, G. L. (2002). Mirror neurons,the insula, and empathy. Behavioral and Brain Sci-ences, 25, 39–40.

Iacoboni, M., Molnar-Szakacs, I., Gallese, V., Buccino,G., Mazziotta, J. C., & Rizzolatti, G. (2005).Grasping the intentions of others with one’s own mir-ror neuron system. Public Library of Science Biology,3, 529–535.

Jaffe, J., Beebe, B., Feldstein, S., Crown, C. L., &Jasnow, M. D. (2001). Rhythms of dialogue in in-fancy. Monographs of the Society for Research inChild Development, 66.

Johnson, D. B. (1992). Altruistic behavior and the de-velopment of the self in infants. Merrill–PalmerQuarterly, 28, 379–388.

Kamisar, Y. (2000). Can (did) Congress overrideMiranda? Cornell Law Review, 85, 883–955.

Kaplan, E. A. (2005). Trauma culture: The politics ofterror and loss in media and literature. New Bruns-wick, NJ: Rutgers University Press.

Kaplan, E. A. (2006). Empathy, global trauma, andpublic feeling: Viewing images of catastrophe. Un-published manuscript, presented in part at theFullerton State University conference on empathy,June 2006.

Keniston, K. (1968). Young radicals. New York: Har-court.

Kluger, R. (1977). Simple justice: The history of Brownv. Board of Education and Black America’s strugglefor equality. New York: Vintage Books.

Laub, D. (1995). Truth and testimony: The process andthe struggle. In C. Caruth (Ed.), Trauma explorationsin memory (pp. 61–75). Baltimore: Johns HopkinsUniversity Press.

Lewis, M., & Brooks-Gunn, J. (1979). Social cognitionand acquisition of self. New York: Plenum Press.

Lipps, T. (1906). Das wissen von fremden Ichen.PsycholUntersuch, 1, 694–722.

Martin, G. B., & Clark, R. D. (1982). Distress crying inneonates: Species and peer specificity. DevelopmentalPsychology, 18, 3–9.

Mill, J. S. (1952). Utilitarianism. In R. M. Hutchins(Ed.), American state papers (Vol. 43, pp.445–476).Chicago: University of Chicago Press. (Original workpublished 1861)

Neuberg, S. L., Cialdini, R. B., Brown, S. L., Luce, C.,Sagarin, B. J., & Lewis, B. P. (1997). Does empathylead to anything more than superficial helping? Jour-nal of Personality and Social Psychology, 73, 310–316.

Oliner, S. P., & Oliner, P. M. (1988). The altruistic per-sonality. New York: Free Press.

Pasupathi, M., Staudinger, U. M., & Baltes, P. B. (2001).Seeds of wisdom: Adolescents’ knowledge and judg-ment about difficult life problems. DevelopmentalPsychology, 37, 351–361.

Pearlman, L. A., & Saakvitne, K. W. (1995). Traumaand the therapist. New York: Norton.

Peterson, J. M., & Cary, J. (2002). Organizational jus-tice, change anxiety, and acceptance of downsizing.Motivation and Emotion, 26, 83–103.

Piaget, J. (1932). The moral judgment of the child. NewYork: Harcourt.

Posner, R. A. (1999). Emotion vs. emotionalism in law.In S. Bandes (Ed.), The passions of law (pp. 309–329). New York: New York University Press.

Radke-Yarrow, M., & Zahn-Waxler, C. (1984). Roots,motives, and patterns in children’s prosocial behav-ior. In E. Staub, D. Bar-Tal, J. Karylowski, & J.Reykowski (Eds.), Development and maintenance ofprosocial behavior (pp. 81–99). New York: PlenumPress.

Rich, W. J. (2004–2005). Betrayal of the children withdolls: The broken promise of constitutional protec-tion for victims of race discrimination. Cornell LawReview, 90, 419–442.

Ruble, D. N., & Martin, C. L. (1998). Gender develop-ment. In W. Damon (Series Ed.) & N. Eisenberg (Vol.Ed.), Handbook of child psychology: Vol. 3. Social,emotional, and personality development (5th ed.,pp. 933–1016). New York: Wiley.

Sagi, A., & Hoffman, M. L. (1976). Empathic distress inthe newborn. Developmental Psychology, 12, 175–176.

Simner, M. L. (1971). Newborn’s response to the cry ofanother infant. Developmental Psychology, 5, 136–150.

Smith, A. (1976). The theory of moral sentiments. Ox-ford, UK: Clarendon Press. (Original work published1759)

Sontag, S. (2003). Regarding the pain of others. NewYork: Farrar, Straus & Giroux.

Stern, D. N., Hofer, L., Haft, W., & Dore, J. (1985). Af-fect attunement: The sharing of feeling states be-tween mother and infant by means of inter-modalfluency. In T. M. Fields & N. A. Fox (Eds.), Socialperception in infants (pp. 249–268). Norwood, NJ:Ablex.

Stowe, H. B. (1852, December 16). Letter to ElizaCabot Follen. Quote in Harriet Beecher Stowe—mother, reformer. Retrieved from xroads.vir-ginia.edu/~ma97/riedy/hbs.html

Stowe, H. B. (1938). Uncle Tom’s cabin. NewYork: Modern Library. (Original work published1852)

Strayer, J. (1993). Children’s concordant emotions andcognitions in response to observed emotions. ChildDevelopment, 64, 166–201.

Szporn, A. (2001). The impact of life condition informa-


tion on empathy. PhD dissertation, New York Uni-versity.

Valiente, C., Eisenberg, N., Fabes, R. A., Shepard, S.,Cumberland, A., & Losoya, S. H. (2004). Predictionof children’s empathy-related responding from theireffortful control and parents’ expressivity. Develop-mental Psychology, 40, 911–926.

Ward, J. H. (1896, August 1). Instant power of UncleTom’s Cabin. The Washington Post.

Weiner, B. (1985). “Spontaneous” causal thinking. Psy-chological Bulletin, 97, 74–84.

Weiner, B., Graham, S., Stern, P., & Lawson, M. E.(1982). Using affective cues to infer causal thoughts.Developmental Psychology, 18, 278–286.

Williams, C. (1989). Empathy and burnout in male andfemale helping professionals. Research in Nursingand Health, 12, 169–178.

Zahn-Waxler, C., Radke-Yarrow, M., & King, R.(1979). Childrearing and children’s prosocial initia-tions toward victims of distress. Child Development,50, 319–330.

Zahn-Waxler, C., Robinson, J. L., Emde, N. E., &Plomin, R. (1992). The development of empathyin twins. Developmental Psychology, 28, 1038–1047.

Zipursky, B. (1990). Deshaney and the jurisprudence ofcompassion. New York University Law Review, 65,1101–1147.


C H A P T E R 2 8

Social Functions of Emotion

AGNETA H. FISCHER and ANTONY S. R. MANSTEAD

The classic perspective on the functionality ofemotions is that they increase the probability ofan individual’s survival and/or reproductivesuccess. The general argument is that emotionsare functional in the sense that they help the in-dividual to address or overcome problems (e.g.,Frijda, 1986; Lazarus, 1991; Levenson, 1999;Tooby & Cosmides, Chapter 8, this volume).Fear is an obvious example. Fear of predatorsor enemies is clearly adaptive in the sense thatindividuals who have the capacity to experi-ence such fear are more likely to be vigilant andavoidant, and thereby to escape the threat ofpredation or attack (e.g., Tooby & Cosmides,Chapter 8, this volume; Öhman, Chapter 44,this volume).

In this chapter, we shift the attentional focusto functional analyses of emotion that empha-size “social survival”—that is, our human ca-pacity to build social bonds and to address andovercome social problems such as social exclu-sion or loss of power (see, e.g., Barrett, 1995;Fridlund, 1994; Griffiths & Scarantino, inpress; Keltner & Haidt, 1999). The central ar-gument we advance in this chapter is that emo-

tions are important to social survival becausethe emotions we experience and express help us(1) to form and maintain social relationships,and (2) to establish or maintain a social posi-tion relative to others. In developing this argu-ment, we discuss these social functions of emo-tions at two levels: interpersonal and group (cf.Keltner & Haidt, 1999).

The idea that emotions serve social functionsassumes that emotions evolved in a social con-text and should therefore be beneficial for so-cial survival. Social survival is a complex en-deavor because it requires a balance betweencooperation on the one hand, and competitionon the other. We humans are social creatureswho need social bonds in order to thrive(Baumeister & Leary, 1995): We affiliate withothers; work together with others; and seekharmony, closeness, and love. The importanceof social bonds is illustrated by research on so-cial isolation, showing that this leads not onlyto poorer health and well-being, but also to in-hibited development of various social, emo-tional, and cognitive skills (Williams, 2001).However, we also set ourselves goals other than

456

being accepted and liked by others: We want toavoid others who pose a threat to us, to win atthe expense of others, to exert control over oth-ers, or to enhance our social power or socialstanding. These social goals are achieved notmerely by cooperation and affiliation; they typ-ically require distancing ourselves from others,or even competing with others or other groups.

Emotions play an important role in realizingthese two types of goals, and we therefore drawa broad distinction between two general socialfunctions of emotion. The first is helping an in-dividual or group to establish or maintain co-operative and harmonious relations with otherindividuals or other social groups. We refer tothis as the “affiliation function” of emotion.The second social function of emotions is help-ing the individual or group to differentiate theself or group from others and to compete withthese others for social status or power. We referto this second function as the “social distancingfunction” of emotion. Each of these general so-cial functions of emotion operates at both theinterpersonal and group levels.

Perhaps unsurprisingly, an analysis of emo-tion’s social functions encounters many of thesame conceptual problems as does an accountof its general functions (Gross & John, 2002;Oatley & Jenkins, 1992; Parrott, 2001, 2002).Emotions generally have social effects (e.g.,Fischer, Manstead, & Zaalberg, 2003;Fridlund, 1994; Philippot, Feldman, & Coats,1999; Tiedens & Leach, 2004), whether or notthese effects are intended. However, these so-cial effects are not equivalent to social func-tions. We argue that the social functions ofemotion should be inferred from the social-relational goals inherent in the prototypical ap-praisals and action tendencies of a given emo-tion (e.g., Roseman, Wiest, & Swartz, 1994).For example, the social distancing function canbe observed in anger (seeking to change an-other person), contempt (seeking to excludeanother person), or social fear (seeking dis-tance from another person); embarrassment(admitting that one has transgressed), love(wanting to be close to the loved one), happi-ness (sharing positive experiences with others),or sadness (seeking help and support from oth-ers) serve the affiliation function.

Whereas social functions derive from thesesocial-relational goals, the social effects of anemotion may vary with the specific features ofthe social and cultural context (e.g., Fischer etal., 2003; Parkinson, 2005; Parkinson, Fischer,

& Manstead, 2005). Let us take anger as an ex-ample. The social function of anger is toimpose change upon another person and thiscan be achieved through a variety of expres-sions, such as beating, scolding, threatening, orcriticizing. The effects of any given expressionof anger may vary, however, depending on theobject of or the reason for one’s anger. For ex-ample, whereas scolding a friend because he orshe forgot an appointment might elicit an apol-ogy on the part of the friend, the selfsame ex-pression of anger toward a superior who forgotan appointment might evoke an aggressive re-sponse. This does not mean that the socialfunction of anger is to elicit an apology in thefirst case and an aggressive response in the sec-ond, but that the wish to impose change hasdifferent effects in different social contexts. Inthe same vein, the effects of emotion may differwith cultural contexts, depending on what isconsidered a typical or appropriate expressionin a specific situation. A further complication isthat the short-term effects of an emotion ex-pression may differ from its long-term effects.The superior who failed to keep an appoint-ment may become aggressive as an initial reac-tion to one’s anger, but may later regret this re-action and acknowledge that he or she was atfault. Thus the anger may serve its social func-tion of effecting change in the other person inthe long run, even if it does not have this effectin the short run (see Fischer & Roseman,2007).

Finally, the idea that emotions have socialfunctions does not imply that emotions are al-ways socially functional—that is, that they al-ways have the social effects that would be pre-dicted on the basis of their relational goals.Anger, jealousy, and contempt can clearly besocially dysfunctional: Rather than changingothers’ behaviors, they may irreparably dam-age the relationship between individuals or be-tween groups, without achieving anything interms of social control or social standing. Thesame applies to positive emotions such aspride, happiness, or love: Rather than strength-ening social bonds, they may cause others totake exception to what they regard as inappro-priate in the circumstances. Social dysfunc-tionality is especially likely to occur if the socialimpact of one’s emotions is not taken into ac-count or if inappropriate appraisals of the so-cial context are made (e.g., Evers, Fischer,Rodriguez Mosquera, & Manstead, 2004;Manstead & Fischer, 2001; Parrott, 2001).

28. Social Functions of Emotion 457

This is supported by research on emotion regu-lation and emotional intelligence (e.g., Lopes,Salovey, Coté, & Beers, 2005), which hasshown an association between the ability toregulate one’s emotions and the quality of so-cial interactions, as rated by individuals andtheir peers. In other words, the social function-ality of emotion in a particular set of circum-stances is not a given, but rather depends onthe way in which the person assesses his or herconcerns or goals in relation to others’ con-cerns or goals, and regulates his or her emo-tions accordingly. Becoming too jealous, get-ting angry too often, or feeling contempt formany people is unlikely to be socially func-tional. Below we further explore the socialfunctions of emotion at the interpersonal andgroup levels.

INTERPERSONAL RELATIONS

The Affiliation Function of Emotions

Social relations vary in emotional tone fromlove affairs to work relations, but any relation-ship between two persons involves a degree ofemotion. The amount of emotion that is expe-rienced and expressed, however, is closely re-lated to the nature of the relationship, reflect-ing an important function of emotions at theinterpersonal level: Emotions enable us to formand maintain long-term and intimate relation-ships by promoting closeness and harmony,and thus avoiding social isolation. This func-tion can be inferred from the social conse-quences or effects of emotions on the develop-ment of a relationship.

There are various lines of research providingevidence for this function. Research by Clarkand colleagues (see Clark, Fitness, & Brissette,2004, for an overview) has shown that emo-tions are more often experienced and expressedin communal or intimate relations than inother type of relations. In the same vein,Gottman and Levenson (2002) found in theiranalysis of marital conflicts that an absence ofaffect during such conflicts is a predictor ofsubsequent divorce. Emotional involvement ischaracteristic not only of marital or romanticrelationships, but of intimate or communal re-lations in general. For example, Barrett, Robin,Pietromonaco, and Eyssell (1998) found thatthe degree of expression of specific emotions,as reported in a diary, was highly correlatedwith the closeness of the interaction partner.

Research on social sharing has also repeatedlyshown that people share their emotions mostlywith family and friends (Rimé, Philippot, Boca,& Mesquita, 1992). From this we can concludethat the experience, expression, and sharing ofemotion are important ingredients in the devel-opment of intimate relations, although the evi-dence to date does not warrant any conclusionsabout the causal direction of this relationship.

The idea that intimate or communal rela-tions are characterized by positive and negativeemotions can be taken a step further by arguingthat in the context of intimate relations onemight start experiencing emotions similar toone’s partner’s emotions, because one sharesthe emotional perspective of the other and ap-praises the emotional situation in the sameway. This is nicely demonstrated by Anderson,Keltner, and John (2003), who found that dat-ing partners and college roommates becameemotionally more similar over the course of ayear. This emotional convergence effect appliedto both positive and negative emotional reac-tions to events and could not be explained byincreased similarity in personality variables.

Emotional convergence can also be observedat the micro level, as reported in studies of mim-icry and emotional contagion (e.g., Chartrand& Bargh, 1999; Bavelas, Black, Lemery, &Mullett, 1986; Dimberg, 1982; Hatfield,Cacioppo, & Rapson, 1992; Lunquist &Dimberg, 1995; Provine, 1992). “Mimicry” re-fers to the automatic processing of others’ non-verbal displays, resulting in similar nonverbaldisplays by oneself, whereas “emotional conta-gion” refers to a tendency for emotional expe-rience to be influenced by others’ emotions.Hatfield et al. (1992) have suggested that thefunction of mimicry is to synchronize and coor-dinate the interaction, and thereby to facilitatemutual involvement and emotional closeness.Indeed, both mimicry and emotional contagioncan be seen as reflecting the affiliation functionof emotions in relations, especially in view ofthe fact that both increase when individualsidentify with another person, or when they feelmore empathy toward that person (Eisenberg,2000; Fischer, Rotteveel, Evers, & Manstead,2004; Sonnby-Borgström, 2002).

It should be noted that the results of thesevarious studies suggest not only that positiveemotions serve a “social glue” function, butalso that negative emotions can reflect and pos-sibly contribute to the intimacy of the relation-ship. Studies of crying, for example, have


shown that people cry more often in the com-pany of partners or intimates than in the com-pany of strangers (Vingerhoets & Becht, 1997).These data are consistent with the idea that thesocial function of sadness (i.e., crying) is a dis-play of powerlessness and a call for support,because such support is most likely to be forth-coming in relations with intimates. Moreover,studies of anger and aggression have shownthat people, especially women, are most oftenangry with intimates (Fischer, Rodriguez, vanVianen, & Manstead, 2004; Kring, 2000), andthat physical aggression also occurs more oftenwithin intimate relationships, especially on thepart of women (Archer, 2000). Also relevant isthe fact that studies of guilt (Baumeister,Stillwell, & Heatherton, 1994; Tangney, Miller,Flicker, & Barlow, 1996) have shown that thisemotion is mostly expressed within valued rela-tionships in which people have high respect forthe other. Similarly, research on regret showsthat interpersonal regret motivates efforts toundo the harm done in the context of a rela-tionship (Zeelenberg, Van der Pligt, &Manstead, 1998).

Collectively, this research shows that com-munal relations are characterized by the expe-rience and expression of positive and negativeemotions, often leading to emotional conver-gence, which further contributes to the inti-macy of and harmony within the relationship.This is consistent with Reis and Shaver’s (1988)intimacy model, in which intimacy is seen asbeing enhanced by reciprocal self-disclosureand partner responsiveness (see Laurenceau,Barrett, & Pietromonaco, 1998). Again, thisdoes not imply that there are no limits to the in-tensity or frequency with which some negativeemotions can be expressed. For example,whereas Gottman and Levenson (2002) haveshown that an absence of affect predicts di-vorce in the longer term, they have also foundthat the expression of negative emotions be-tween marital partners during conflict is pre-dictive of divorce in the shorter term. Mar-riages characterized by conflicts in which onepartner expresses contempt for the other arealso likely to lead to a divorce. Clearly, emo-tions such as hatred and contempt do not servean affiliation function.

Although emotions generally serve affilia-tion goals in the context of intimate relations,the promotion of these goals can also be ob-served in emotional interactions betweenstrangers. Smiling people, for example, are

more likely to be ascribed positive traits, suchas kindness, humor, intelligence, or honesty,than are their nonsmiling counterparts (Hess,Beaupré, & Cheung, 2002; Reis et al., 1990).Krumhuber and colleagues (Krumhuber &Kappas, 2005; Krumhuber, Manstead, &Kappas, 2007) have shown that these positiveeffects of smiling are moderated by the per-ceived genuineness of the smile. Self-consciousemotions, like guilt, have also been shown toresult in more cooperation on the part of theguilty participant in interactions betweenstrangers. Displays of embarrassment, guilt,shame, and regret serve to show and acknowl-edge that one has acted stupidly, transgressed arule, hurt another person, or made a mistake.Such displays have been shown to evoke sym-pathy, positive evaluations, or helpful behaviorin others (Keltner & Buswell, 1976, 1997;Ketelaar & Au, 2003; Miller, 2004), but alsoincreased perceptions of trustworthiness andconscientiousness—as, for example in studieson blushing, a typical signal of embarrassment(de Jong, 1999). Interestingly, embarrassmentdisplays tend to elicit more positive evaluationson a “warmth” dimension, but not on a “com-petence” dimension (Semin & Manstead,1981), consistent with the notion that thisemotion serves an affiliation function.

Another way in which emotions are sociallyfunctional in the context of interpersonal rela-tions is that the communication of emotionscan direct the behavior of others, by warningthem of impending danger or reassuring themthat the environment is safe. A classic exampleof this function is provided by research on so-cial referencing (e.g., Sorce, Emde, Campos, &Klinnert, 1985), showing that 12-month-oldsare less likely to proceed in an uncertain andpossibly unsafe situation when their mothersdisplay a negative expression, and more likelyto proceed when the mothers display a positiveexpression. Thus emotions have informationalvalue for others, in that they communicate ourconcerns and our appraisals of the current situ-ation. However, we suggest that both the ex-tent to which these signals are sent and the ex-tent to which they are acted on by receivers is afunction of their relational closeness, and that akey outcome of this emotional signaling is tostrengthen the trust and bond between signalerand receiver.

In summary, there is abundant evidence thatthe more intimate a relationship is, the morethe partners feel, express, and share a variety of


emotions. This evidence also suggests that thisemotional disclosure is not simply due to theamount of time spent together or to similaritiesbetween individuals, but rather that expressingand sharing emotions help to increase intimacyin interpersonal relations. This is because one ismore likely to express emotions to a target whois expected to be able to respond appropriatelyto one’s needs. The sharing of such intimate in-formation through the expression of emotionsis likely to result in a motivation to maintain orincrease the social bond with that person (seealso Clark et al., 2004). Furthermore, express-ing an emotion not only conveys intimacy, be-cause it reveals one’s concerns, beliefs, and de-sires, but may also convey an appraisal of thesituation. Thus, in addition to promoting inti-macy, emotions may also serve affiliation func-tions, because they enhance positive interac-tions, cooperation, and trust.

The Social Distancing Function of Emotion

Emotions do not always fuel emotional close-ness and identification. Various social-relational goals of emotions imply the reverserelational movement—namely, increasing dis-tance from others, as in anger, contempt, dis-gust, or fear of another person (e.g., Fridlund,1994; Oatley & Jenkins, 1992). Thus, along-side the affiliation function, we propose a so-cial distancing function. Markus and Kitayama(1991) alluded to a similar notion when theydistinguished between “socially engaging” and“socially disengaging” emotions. The latter aremore socially desirable in cultures where an in-dependent rather than an interdependent self ispromoted. The emotions of anger, contempt,sociomoral disgust, and also pride are assumedto serve this function. We now consider exam-ples of these emotions, and their effects on oth-ers and the self.

Anger is a prime example of an emotion thatserves this function. If directly expressed to theobject of the anger, this emotion confronts theother person with the fact that the angry per-sons wants to change the target’s behavior andexert some control over the target (Fischer &Roseman, 2007; Timmers, Fischer, &Manstead, 1998). This implies that the targetof the anger should apologize, yield, show sub-missiveness, or simply stop doing whatever heor she was doing. Studies by Van Kleef and col-leagues have shown that the verbalization ofanger in a negotiation context leads to more

yielding on the part of the other person (e.g.,Van Kleef, De Dreu, & Manstead, 2004a,2004b). Studies of mimicry have also shownthat although angry faces sometimes lead toangry expressions (e.g., Dimberg & Lunquist,1990), they can also give rise to fear and dis-gust displays (Lunquist & Dimberg, 1995).These studies suggest that anger is related tothe exercise of power or influence over anotherperson. In line with this idea that direct angerexpression is an attempt to regain or maintainstatus or power, or simply to control the other,research by Tiedens (2001) confirms that high-status persons are expected to respond with an-ger (rather than sadness or guilt) to negativeoutcomes, and with pride (rather than appreci-ation) to positive outcomes (see also Kuppens,Van Mechelen, & Meulders, 2004). Moreover,when people show anger and pride they arethought of as being high in status, but whenthey show sadness, guilt, or appreciation theyare thought of as low in status. The underlyingexplanation for these differences, according toTiedens (2001), lies in the appraisals of agencythat are implied by the expression of each ofthese emotions. An expression of anger is seenas reflecting the appraisal of other-blame andcontrollability; by contrast, an expression ofguilt or sadness reflects appraisals of self-blameand uncontrollability. We suggest that these re-sults also support the idea that anger and prideserve a social distancing function, in that theytypically result in greater social distance be-tween the self and the other.

A more extreme example of a social distanc-ing function can be identified in the case ofcontempt. The expression of this emotion, typi-cally in the form of derogation and rejection,often results in the social exclusion of the ob-ject (Fischer & Roseman, 2007). The aim is tomake it clear that the other person is inferiorand worthless, which is a way of boosting one’sown social position or status, either as an indi-vidual or as a group member. Contempt may beelicited in addition to anger if alternative waysof changing others are expected to fail.

Thus anger and contempt typically serve asocial distancing function by blaming anotherperson for a negative outcome and either pres-suring this person into yielding or doing whatone wants (anger), or subjecting him or her tosocial exclusion (contempt). It is worth notingthat although both emotions serve a social dis-tancing function, they differ in the extent towhich they are able also to serve an affiliation


function: Whereas contempt is highly unlikelyto give rise to any improvement in a social rela-tionship, anger often does so, certainly in thelonger term (Fischer & Roseman, 2007). Thissuggests that anger can serve both an affiliationand a social distancing function, but that thesefunctions may be served at different points intime.

To summarize, we have described evidenceshowing that emotions sometimes have the ef-fect of creating distance between oneself andothers in a way that enhances one’s own socialstanding. Expression of the emotion impliesmovement away from, but also above, theother person. Although one could argue thatthis social distancing function is less social innature than is the affiliation function describedearlier, because it appears to serve the selfrather than one’s relations with others, we ar-gue that the protection or enhancement of so-cial standing is a social function. Most obvi-ously, the capacity to constrain the antisocial ordysfunctional behaviors of others by express-ing anger and contempt toward them can besaid to serve the function of protecting the indi-vidual or group from the harmful or dysfunc-tional behavior of others. Likewise, it is some-times functional for individuals and groups toset themselves apart from or to cut themselvesoff from other individuals or groups. Anger felttoward an ex-partner can be helpful in detach-ing oneself from the relationship. Contempt ex-pressed toward others who fail to endorse orlive up to key norms and values can be benefi-cial in protecting those norms and values. Fur-thermore, the success of groups, organizations,and societies can be said to depend to some de-gree on the presence within them of individualswho strive to achieve more than their peers,thereby driving themselves and the groups towhich they belong to better material outcomes.Thus pride in individual or collective achieve-ments may serve a social function, not simplyan individual one.

GROUP AND INTERGROUPRELATIONS

Affiliation and social distancing functions canalso be applied to group emotions. The notionthat groups can “have” emotions may initiallyseem puzzling. Common sense and emotiontheory share the assumption that emotions arestates experienced by individuals. We use the

term “group-based emotion” here to refer tothe fact that members of social groups havemore similar concerns, make more similar ap-praisals, and therefore experience more similaremotions (Smith, 1993) than would be ex-pected by chance (see also Smith & Mackie,Chapter 26, this volume). This idea is based onself-categorization theory (SCT; Turner, Hogg,Oakes, Reicher, & Wetherell, 1987). Accordingto SCT, the salience of one’s social identityshould promote the experience of emotionsthat are driven by appraisals and concerns thatare group-based rather than individual-based.For example, if your personal identity is sa-lient, the success or failure of a given soccerteam is likely to have little impact on your emo-tion; however, if your social identity as a sup-porter of that soccer team is salient, the self-same success or failure of that team may leadto exhilaration or crushing disappointment(Cialdini et al., 1976; see Parkinson et al.,2005, for a review).

This idea of group-based emotions does notnecessarily imply that group members need tobe present, because it is the group-based con-cern that makes the emotion a group emotion.However, if others are present, this may exertinfluence on how one feels and expresses thisfeeling, in which case we refer to “collectiveemotions.” A good illustration of this phenom-enon is provided by Totterdell (2000), who as-sessed the moods of members of two profes-sional sports teams three times a day for 4 daysduring a competitive match. Individual players’moods were more strongly correlated with thecurrent aggregate mood of their own team thanwith the current aggregate mood of the otherteam or with the aggregate mood of their ownteam at other times. These correlations be-tween player mood and team mood were alsofound to be independent of various factors, in-cluding the match situation between the twoteams, effectively ruling out an explanation interms of shared exposure to common situa-tions.

One process that might account for collec-tive emotions is contagion (Hatfield et al.,1992). This phenomenon occurs during face-to-face interaction. If group members engage inmore frequent face-to-face interaction witheach other than with nonmembers, contagionshould be greater among group members thanamong nonmembers. Moreover, if emotionspreads from person to person in this way, theresulting shared behaviors and experiences


should encourage the perception that the af-fected people constitute a group.

Functions of Emotionin Intragroup Relations

The functions of emotion in group settings par-allel those in interpersonal contexts. First, con-sistent with the affiliation function, emotionscan strengthen relations within a social group,enhancing a sense of commitment and belong-ing. The experience of shared emotion ingroups strengthens bonds between group mem-bers and sharpens group boundaries, therebyenhancing loyalty to the group (see also Keltner& Haidt, 1999). In addition, the communica-tion of emotion within a group provides groupmembers with rapid information about groupstructure and the environment (Spoor & Kelly,2004).

The function of emotions in the developmentand constitution of groups may be found in alltypes of groups. For example, emotions insport or work teams may help to create a teamspirit and strengthen the motivation to win as ateam (e.g., Totterdell, 2000). Emotions in workteams may also enhance cooperation by groupmembers (Barsade, 2002; Bartel & Saavedra,2000). Not only do members of work groupsexperience “group moods,” but the extent towhich group members “catch” another per-son’s mood seems to determine levels of coop-eration and conflict (Barsade, 2002). In partic-ular, dispersion of positive emotion leads togreater cooperation and reduces conflict in thegroup, whereas dispersion of negative emotionis associated with the opposite outcomes.George and her colleagues have also shownthat group affective tone is related to prosocialbehavior toward customers (see George, 1990).

Research by Kessler and Hollbach (2005)has also shown that group-based emotions in-fluence ingroup identification. The context ofthis research was the relationship between Eastand West Germans. The East German partici-pants were asked to recall an instance whenthey as East Germans felt either anger or happi-ness toward their own national subgroup or to-ward West Germans. A measure of identifica-tion as East Germans was taken before andafter this procedure. It was found that recalledhappiness toward the ingroup and anger to-ward the outgroup both increased ingroupidentification, whereas anger toward theingroup and happiness toward the outgroup

both decreased ingroup identification. This re-search provides an insight into the emotionaldynamics that may underpin the “basking inreflected glory” phenomenon (Cialdini et al.,1976). The positive affect that supporters pre-sumably experience as a result of their team’ssuccess on the football field should increasetheir ingroup identification. Likewise, Kesslerand Hollbach’s (2005) findings provide a possi-ble emotion-based account for the related phe-nomenon of “cutting off reflected failure”(Snyder, Lassegard, & Ford, 1986), wherebygroup failure leads individuals to distancethemselves from the ingroup in question.

The affiliation function of emotions ingroups can also be illustrated by consideringthe effects of differences in appraisal and emo-tion. Sani and his associates (e.g., Sani, 2005;Sani & Reicher, 1999) have shown that schismsdevelop from perceptions that the positionstaken by other ingroup members undermine orchange in some fundamental way the sharedidentity of the group. Because this shared iden-tity is a core value of the group, any seriousthreat to it is likely to evoke negative emotionssuch as dejection and agitation, and to result indecreased identification with the group andlower perceived cohesiveness. In a field studyof the secession of a subgroup from the Churchof England, Sani (2005) showed that strongnegative emotion aroused by a perceived threatto group identity was a positive predictor of in-tentions to secede from the group.

In summary, emotions experienced in groupsettings have a number of beneficial effects ongroup functioning. To the extent that emotionsare seen to be shared, they help to promotegroup cohesiveness and ingroup cooperationby providing group members with tangible evi-dence of their similarities. The reverse side ofthis coin is that when group members findthemselves in disagreement about issues that inthe eyes of some members threaten the integrityof the group, emotions can play a destructiverole, making it more likely that the group willsplinter into two or more factions.

Functions of Emotionin Intergroup Relations

Emotions not only serve to strengthen thebonds within one’s group, but may also help toimprove or worsen relations between socialgroups. One of the emotions that may serve theintergroup affiliation function is group-based


guilt (Branscombe, Doosje, & McGarty, 2002;Doosje, Branscombe, Spears, & Manstead,1998). Several studies have shown that individ-uals may feel guilty about past behavior oftheir ingroup toward an outgroup, as in thecase of how the Dutch treated Indonesians dur-ing the colonial era (Doosje et al., 1998) orhow European Americans have historicallytreated African Americans (Iyer, Leach, &Crosby, 2003). Group-based guilt implies thatmembers of one group are perceived as respon-sible for disadvantages experienced by anoutgroup. Doosje et al. (1998) found that levelsof group-based guilt among Dutch participantswere higher when past behavior of their na-tional group toward Indonesians during the co-lonial era was presented in negative terms,rather than in positive or mixed terms. More-over, they also showed that participants’ levelsof Dutch national identification was associatedwith the intensity of group-based guilt: Becausehigh levels of identification are likely to lead todefensive denial that any wrong has been per-petrated, high identifiers were less willing toaccept group-based guilt than were low identi-fiers, at least when the evidence concerningtreatment of the outgroup was mixed.

Although these studies seem to show thebeneficial effects of group-based guilt, there arealternative accounts of what follows fromgroup-based guilt. In a study on EuropeanAmericans’ “white guilt” with respect to Afri-can Americans, Iyer et al. (2003) suggested thatgroup-based guilt may involve a focus on theingroup, implying a desire to offer an apologyor engage in reparation with a view to alleviat-ing ingroup distress, rather than a more other-focused desire to offer practical help to mem-bers of the disadvantaged group through pro-grams such as affirmative action. Indeed, Iyeret al.’s studies showed that focusing on Euro-pean American perpetration led to greater feel-ings of guilt, whereas a focus on African Amer-ican suffering led to greater feelings ofsympathy. Guilt predicted support for compen-satory affirmative action, but not equal oppor-tunity affirmative action, whereas sympathypredicted support for equal opportunity affir-mative action and was a marginal predictor ofsupport for compensatory affirmative action.This suggests that although both guilt and sym-pathy serve affiliation functions at the grouplevel, there may be subtle differences in thisfunction that can be inferred from the durationor size of the beneficial effects: Guilt may serve

bonding in order to alleviate the bad con-science of ingroup members, whereassympathy improves bonding in order to helpthe other group.

Whereas group-based guilt or sympathy maylead to an improvement in relations with disad-vantaged outgroups, other group-based emo-tions may function to strengthen the disadvan-taged position of one’s own group and therebyto enhance the social standing of the ingrouprelative to other groups. A study by Gordijn,Yzerbyt, Wigboldus, and Dumont (2006), forexample, shows how a perceiver’s degree ofidentification with the victims of an injusticeaffects not just level of group-based anger, butalso willingness to engage in collective actionto redress the wrong. Here group-based angerhas the effect of mobilizing people who are notthemselves directly affected by the perceived in-justice. In a similar vein, Van Zomeren, Spears,Fischer, and Leach (2004) sought to under-stand why collective disadvantage sometimesdoes and sometimes does not lead to collectiveaction, and hypothesized that this was due todifferent expectations about the type of sup-port that members would receive from theirgroup. These researchers drew a distinction be-tween “instrumental support” (support for ac-tion) and “emotional support” (sharing opin-ions). They found that if members anticipatedinstrumental support, this was enough to pre-dict collective action in the absence of anger;however, if members anticipated emotionalsupport, this led to more group-based angerand thereby to collective action. So here angerserves a function of motivating group membersto engage in collective action.

Emotions may also serve a social distancingfunction, because certain emotions can alsohelp to sharpen group boundaries by stressingdissimilarities with the outgroup, or even bypromoting prejudice. DeSteno, Dasgupta, Bart-lett, and Cajdric (2004) have shown that emo-tional states are capable of creating automaticprejudice toward outgroups. Specifically, angerwas shown to influence automatic negativeevaluations of outgroups, supposedly becauseof its functional relevance to intergroup con-flict and competition. Moreover, studies on in-tergroup Schadenfreude (Leach, Spears,Branscombe, & Doosje, 2003; Spears & Leach,2004) have shown that its intensity is, amongother things, associated with feelings of inferi-ority with respect to the outgroup, suggestingthat intergroup Schadenfreude can be seen as a


way of coping with the lower social status ofone’s own group.

In summary, group-based emotions havepowerful effects on both ingroup and inter-group relations. The emotions that have at-tracted most research attention to date, andthe ones we have focused on here, are guiltand anger. There is consistent evidence thatgroup-based guilt promotes tendencies toapologize and make reparation, just as per-sonal guilt does in the context of interper-sonal relations. On the other hand, group-based anger concerning the maltreatment of agroup is more likely to result in effective po-litical action on behalf of the disadvantagedgroup.

CONCLUSION

Emotions are elicited and evolve in social con-texts, and they help us to deal with the chal-lenges posed by our social environment. Wehave distinguished two general social functionsof emotion: affiliation, which entails the pro-motion of interpersonal and intergroup rela-tionships, on the one hand; and social distanc-ing, which entails separation from and evencontrol or power over others, on the other.Thus emotions can have the effect of reducingor increasing the distance between self and oth-ers, or between one’s own group and othergroups. It should be clear from our discussionthat the social functions of emotion should notbe equated with the social effects of expressingthat emotion in a given setting. The social func-tion of the emotion is relatively independent ofsocial context, and intrinsic to the social-relational goals and prototypical features of thespecific emotional reaction, whereas social ef-fects are contingent upon the way in which theemotion is expressed and upon the specifics ofthe social context. For example, the socialfunction of sadness is to elicit support and con-solation from others, yet the consequence ofcrying in a given situation might be ridiculerather than comfort. This also shows that thenotion that emotions serve social functions isnot equivalent to the idea that emotions are al-ways socially functional. Social functionalitydepends on emotional intelligence (e.g.,Salovey & Mayer, 1990), and on striking anappropriate balance between affiliation and co-operation on the one hand, and social distanc-ing and competition on the other. Social sur-

vival and social success involve a compromisebetween these elementary social goals.

There is a pressing need for more systematicresearch on these issues. Although it is clearthat positive emotions foster positive social in-teractions between individuals, within groups,and between groups, resulting in greater coop-eration and more positive outcomes, the evi-dence concerning negative emotions is lessclear-cut—especially the circumstances underwhich these emotions are less socially func-tional. At the interpersonal level, negative emo-tions are more often expressed in intimate rela-tions. Most of these emotions, like regret,embarrassment, sadness, fear, shame, or guilt,seem to serve an affiliation function, whichleads us to argue that the expression of theseemotions should benefit the relationship. It isobvious, however, that the expression of somenegative emotions will not always improve theintimacy and quality of a relationship. Hatredand contempt seem to serve an individual’sstriving for social distance and social standing,rather than a desire to connect to connect withothers. These emotions appear not to serve anyaffiliation function in an interpersonal setting.

Interestingly, shared hatred and contempt,along with other negative emotions directed atthird persons or outgroups, can serve an affilia-tion function. When negative emotions areshared and expressed toward a third party, thisresults in a stronger tie with the person orgroup with which this negative emotion isshared. When people feel threatened by anoutgroup, for example, one way to controltheir fear is by developing a hatred of theoutgroup and perceiving its members as funda-mentally bad. This strengthens bonds withinthe group and increases the distance betweenone’s own group and other groups, as has beenwitnessed in the aftermath of the 9/11 attacks.Thus the expression of shared fear serves moreof a social distancing function with respect tothe object of one’s fear (moving the group awayfrom the threatening group), but more of an af-filiation function with respect to ingroup mem-bers who face the same threat but who alsoprovide a resource for dealing with the threat(fellow Americans). Thus the social function ofemotions is dependent on the target of the emo-tion. In cases where emotions are shared withingroup members, they are likely to serve an af-filiation function.

At the intergroup level, the affiliation andsocial distancing functions of emotion seem to


parallel those at the interpersonal level, but rel-atively few emotions have been studied in inter-group contexts. The affiliation function of neg-ative emotions has most frequently beenexamined in the context of group-based guilt,which can lead to the tendency to repair thedamage done to an outgroup. Yet some haveargued that there may be limits to how benefi-cial group-based guilt really is, perhaps becauseof a tendency on the part of the advantagedgroup to focus on making itself feel betterrather than on helping the outgroup, and partlybecause highly identified members of advan-taged groups may be reluctant to acknowledgethe harm done to an outgroup. A similar argu-ment may apply to guilt at the interpersonallevel.

When considering other emotions, such asfear and sadness, we can conclude that the con-sequences of groups expressing emotions likefear and sadness may depend on the status orpower relations between the groups involved.For example, the obvious fear and distress of alow-power group (e.g., the refugees in theDarfur region of Sudan) may elicit empathyand help from members of other more power-ful groups. By contrast, the equally obviousfear and distress of a high-power group (e.g.,Americans in the aftermath of 9/11) may beless likely to arouse compassionate reactionson the part of members of other groups, unlessthey perceive a shared identity with the victimgroup.

Much the same applies at the interpersonallevel, in that it seems plausible to suggest thatwe respond differently to fear expressed by ayoung child than to fear expressed by an adultwho is a rival. Here again, it is clear that the so-cial effects of specific emotions depend on thesocial context—that is, on the relationship be-tween the individuals (e.g., communal, ex-change) or groups (e.g., relative power and sta-tus)—but that the social function of theemotion, signaling the need for help, is rela-tively independent of this context. It is evidentthat an emotion may serve a single social func-tion, yet have different social effects in differentsocial settings.

A final point that bears repetition is that al-though we firmly believe that emotions servesocial functions, this does not mean that emo-tions are always socially functional. The poten-tial social functionality of emotions can be in-ferred from their social-relational goals andtypical features. In practice, their functionality

depends on how the individuals or socialgroups involved appraise the social context,and how they regulate their emotions and ex-pressions in a way that is consistent with thosesocial appraisals. Perhaps the ultimate socialfunction of emotions is to persuade others toaccept (at least to some degree) the validity ofone’s own or one’s group’s appraisals.

ACKNOWLEDGMENTS

We thank Lisa Feldman Barrett, Keith Oatley, BrianParkinson, and Jerry Parrott for their helpful com-ments on a previous version of this chapter.

REFERENCES

Anderson, C., Keltner, D., & John, O. P. (2003). Emo-tional convergence between people over time. Jour-nal of Personality and Social Psychology, 84, 1054–1068.

Archer, J. (2000). Sex differences in aggression betweenheterosexual partners: A meta-analytic review. Psy-chological Bulletin, 126, 651–680.

Barrett, K. C. (1995). A functionalist approach toshame and guilt. In J. P. Tangney & K. W. Fischer(Eds.), Self-conscious emotions: The psychology ofshame, guilt, embarrassment, and pride (pp. 25–64).New York: Guilford Press.

Barrett, L. F., Robin, L., Pietromonaco, P. R., & Eyssell,K. (1998). Are women the more emotional sex?: Evi-dence from emotional experiences in social interac-tions. Cognition and Emotion, 12, 555–578.

Barsade, S. G. (2002). The ripple effect: Emotional con-tagion in groups. Administrative Science Quarterly,47, 644–677.

Bartel, C. A., & Saavedra, R. (2000). The collective con-struction of work group moods. Administrative Sci-ence Quarterly, 45, 197–231.

Baumeister, R. F., & Leary, M. R. (1995). The need tobelong: Desire for interpersonal attachments as afundamental human motivation. Psychological Bul-letin, 117, 497–529.

Baumeister, R. F., Stillwell, A. M., & Heatherton, T. F.(1994). Guilt: An interpersonal approach. Psycho-logical Bulletin, 115, 243–267.

Bavelas, J. B., Black, A., Lemery, C. R., & Mullett, J.(1986). “I show how you feel”: Motor mimicry as acommunicative act. Journal of Personality and SocialPsychology, 50, 322–329.

Branscombe, N. R., Doosje, B., & McGarty, C. (2002).Antecedents and consequences of collective guilt. InD. M. Mackie & E. R. Smith (Eds.), From prejudiceto intergroup emotions: Differentiated reactions tosocial groups (pp. 49–66). New York: PsychologyPress.

Chartrand, T. L., & Bargh, J. A. (1999). The chameleon


effect: The perception–behavior link and social inter-action. Journal of Personality and Social Psychology,76, 893–910.

Cialdini, R. B., Borden, R. J., Thorne, A., Walker, M. R.,Freeman, S., & Sloan, L. R. (1976). Basking in re-flected glory: Three (football) field studies. Journal ofPersonality and Social Psychology, 34, 366–375.

Clark, M. S., Fitness, J., & Brissette, I. (2004). Under-standing people’s perceptions of relationships is cru-cial to understanding their emotional lives. In M. B.Brewer & M. Hewstone (Eds.), Emotion and motiva-tion (pp. 21–47). Malden, MA: Blackwell.

de Jong, P. J. (1999). Communicative and remedial ef-fects of social blushing. Journal of Nonverbal Behav-ior, 23, 197–217.

DeSteno, D., Dasgupta, N., Bartlett, M. Y., & Cajdric,A. (2004). Prejudice from thin air: The effect of emo-tion on automatic intergroup attitudes. PsychologicalScience, 15, 319–324.

Dimberg, U. (1982). Facial reactions to facial expres-sions. Psychophysiology, 19, 643–647.

Dimberg, U., & Lunquist, L.-O. (1990). Gender differ-ences in facial reactions to facial expressions. Biolog-ical Psychology, 30, 151–159.

Doosje, B., Branscombe, N. R., Spears, R., &Manstead, A. S. R. (1998). Guilty by association.When one’s group has a negative history. Journal ofPersonality and Social Psychology, 75, 872–886.

Eisenberg, N. (2000). Empathy and sympathy. In M.Lewis & J. M. Haviland-Jones (Eds.), Handbook ofemotions (2nd ed., pp. 677–691). New York:Guilford Press.

Evers, C., Fischer, A. H., Rodriguez Mosquera, P. M., &Manstead, A. S. R. (2004). Anger and social apprais-al: A “spicy” sex difference? Emotion, 5, 258–266.

Fischer, A. H., Manstead, A. S. R., & Zaalberg, R.(2003). Social influences on the emotion process. InW. Stroebe & M. Hewstone (Eds.), European reviewof social psychology (Vol. 14, pp. 171–203). Hove,UK: Psychology Press.

Fischer, A. H., Rodriguez Mosquera, P. M., van Vianen,E. A. M., & Manstead, A. S. R. (2004). Gender andculture differences in emotion. Emotion, 4, 87–94.

Fischer, A. H., & Roseman, I. J. (2007). Beat them orban them: The characteristics and social functions ofanger and concept. Journal of Personality and SocialPsychology, 93, 103–115.

Fischer, A. H., Rotteveel, M., Evers, C., & Manstead, A.S. R. (2004). Emotional assimilation: How we are in-fluenced by others’ emotions. Cahiers de PsychologieCognitive, 22, 223–245.

Fridlund, A. J. (1994). Human facial expression: Anevolutionary view. San Diego, CA: Academic Press.


George, J. M. (1990). Personality, affect, and behaviorin groups. Journal of Applied Psychology, 75, 107–116.

Gottman, J. M., & Levenson, R. W. (2002). A two-factor model for predicting when a couple will di-

vorce: Exploratory analyses using 14-year longitudi-nal data. Family Process, 41, 83–96.

Gordijn, E., Yzerbyt, V. Y., Wigboldus, D., & Dumont,M. (2006). Emotional reactions to harmful inter-group behavior: The impact of being associated withthe victims or the perpetrators. European Journal ofSocial Psychology, 36, 15–30.

Griffiths, P., & Scarantino, A. (in press). Emotions in thewild: The situated perspective on emotion. In P. Rob-bins & M. Aydele (Eds.), Cambridge handbook ofsituated cognition. New York: Cambridge UniversityPress.

Gross, J., & John, O. P. (2002). Wise emotion regula-tion. In L. F. Barrett & P. Salovey (Eds.), The wisdomin feeling: Psychological processes in emotional intel-ligence (pp. 297–319). New York: Guilford Press.

Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1992).Primitive emotional contagion. In M. S. Clark (Ed),Review of personality and social psychology: Vol. 14.Emotion and social behavior (pp. 151–177). Thou-sand Oaks, CA: Sage.

Hess, U., Beaupré, M. G., & Cheung, N. (2002). Who,to whom, and why: Cultural differences and similari-ties in the function of smiles. In M. H. Abel (Ed.), Anempirical reflection on the smile (pp. 187–216).Lewiston, NY: Edwin Mellen Press.

Iyer, A., Leach, C. W., & Crosby, F. J. (2003). Whiteguilt and racial compensation: The benefits and lim-its of self-focus. Personality and Social PsychologyBulletin, 29, 117–129.

Keltner, D., & Buswell, B. N. (1976). Evidence for thedistinctness of embarrassment, shame, and guilt: Astudy of recalled antecedents and facial expressionsof emotion. Cognition and Emotion, 10, 155–171.

Keltner, D., & Buswell, B. N. (1997). Embarrassment:Its distinct form and appeasement function. Psycho-logical Bulletin, 122, 250–270.

Keltner, D., & Haidt, J. (1999). Social functions of emo-tions at four levels of analysis. Cognition and Emo-tion, 13, 505–521.

Kessler, T., & Hollbach, S. (2005). Group based emo-tion as determinants of ingroup identification. Jour-nal of Experimental Social Psychology, 41, 677–685.

Ketelaar, T., & Au, W. T. (2002). The effects of feelingsof guilt on the behavior of uncooperative individualsin repeated social bargaining games: An affect-as-information interpretation of the role of emotion insocial interaction. Cognition and Emotion, 17, 429–455.

Kring, A. M. (2000). Gender and anger. In A. H. Fischer(Ed.), Gender and emotion: Social psychological per-spectives (pp. 211–232). Cambridge, UK: CambridgeUniversity Press.

Krumhuber, E., & Kappas, A. (2005). Moving smiles:The role of dynamic components for the perceptionof the genuineness of smiles. Journal of NonverbalBehavior, 29, 3–24.

Krumhuber, E., Manstead, A. S. R., & Kappas, A.(2007). Temporal aspects of facial displays in personand expression perception: The effects of smile dy-


namics, head-tilt, and gender. Journal of NonverbalBehavior, 31, 39–56.

Kuppens, P., Van Mechelen, I., & Meulders, M. (2004).Every cloud has a silver lining: Interpersonal and in-dividual differences determinants of anger-relatedbehavior. Personality and Social Psychology Bulletin,30, 1550–1564.

Laurenceau, J. P., Barrett, L. F., & Pietromonaco, P. R.(1998). Intimacy as an interpersonal process: The im-portance of self-disclosure, partner disclosure, andperceived partner responsiveness in interpersonal ex-changes. Journal of Personality and Social Psychol-ogy, 74, 1238–1251.


Leach, C. W., Spears, R., Branscombe, N. R., & Doosje,E. J. (2003). Malicious pleasure: Schadenfreude atthe suffering of another group. Personality and SocialPsychology, 84, 932–943.

Levenson, R. W. (1999). Intrapersonal functions ofemotion. Cognition and Emotion, 13, 481–504.

Lopes, P. N., Salovey, P., Coté, S., & Beers, M. (2005).Emotion regulation abilities and the quality of socialinteraction. Emotion, 5, 113–118.

Lunquist, L.-O., & Dimberg, U. (1995). Facial expres-sions are contagious. Journal of Psychophysiology, 9,203–211.

Manstead, A. S. R., & Fischer, A. H. (2001). Socialappraisal: The social world as object of and influenceon appraisal processes. In K. R. Scherer, A. Schorr, &T. Johnstone (Eds.), Appraisal processes in emotion:Theory, methods, research (pp. 221–232). NewYork: Oxford University Press.

Markus, H. R., & Kitayama, S. (1991). Culture and theself: Implications for cognition, emotion, and moti-vation. Psychological Review, 98, 224–253.

Miller, R. S. (2004). Emotion as adaptive interpersonalcommunication: The case of embarrassment. In L. Z.Tiedens & C. W. Leach (Eds.), The social life of emo-tions (pp. 87–105). Cambridge, UK: Cambridge Uni-versity Press.

Oatley, K., & Jenkins, J. M. (1992). Human emotions:Function and dysfunction. Annual Review of Psy-chology, 43, 55–85.

Parkinson, B. (2005). Do facial movements expressemotions or communicate motives? Personality andSocial Psychology Review, 9, 278–311.

Parkinson, B., Fischer, A. H., & Manstead, A. S. R.(2005). Emotions in social relations: Cultural, group,and interpersonal processes. New York: PsychologyPress.

Parrott, W. G. (2001). Implications of dysfunctionalemotions for understanding how emotions function.Review of General Psychology, 5, 180–186.

Parrott, W. G. (2002). The functional utility of negativeemotions. In L. F. Barrett & P. Salovey (Eds.), Thewisdom in feeling: Psychological processes in emo-tional intelligence (pp. 341–362). New York:Guilford Press.

Philippot, P., Feldman, R. S., & Coats, E. J. (Eds.).

(1999). The social context of nonverbal behavior.New York: Cambridge University Press.

Provine, R. R. (1992). Contagious laughter: Laughter isa sufficient stimulus for laughs and smiles. Bulletin ofthe Psychonomic Society, 30, 1–4.

Reis, H. T., McDougal Wilson, I., Monestere, C.,Bernstein, S., Clark, K., Seidl, E., et al. (1990). Whatis smiling is beautiful and good. European Journal ofSocial Psychology, 20, 259–267.

Reis, H. T., & Shaver, P. R. (1988). Intimacy as an inter-personal process. In S. Duck, D. F. Hay, S. E. Hobfoll,W. Ickes, & B. M. Montgomery (Eds.), Handbook ofpersonal relationships: Theory, research, and inter-ventions (pp. 367–389). Chichester, UK: Wiley.

Rimé, B., Philippot, P., Boca, S., & Mesquita, B. (1992).Long-lasting cognitive and social consequences ofemotion: Social sharing and rumination. In W.Stroebe & M. Hewstone (Eds.), European review ofsocial psychology (Vol. 3, pp. 225–258). Chichester,UK: Wiley.

Roseman, I. J., Wiest, C., & Swartz, T. S. (1994). Phe-nomenology, behaviors and goals differentiate dis-crete emotions. Journal of Personality and SocialPsychology, 67, 206–221.

Salovey, P., & Mayer, J. D. (1990). Emotional intelli-gence. Imagination, Cognition, and Personality, 9,185–211.

Sani, F. (2005). When subgroups secede: Extending andrefining the social psychological model of schisms ingroups. Personality and Social Psychology Bulletin,31, 1074–1086.

Sani, F., & Reicher, S. (1999). Identity, argument andschism: Two longitudinal studies of the split in theChurch of England over the ordination of women tothe priesthood. Group Processes and Intergroup Re-lations, 2, 279–300.

Semin, G. R., & Manstead, A. S. R. (1981). The socialimplications of embarrassment displays and restitu-tion behaviour. European Journal of Social Psychol-ogy, 12, 367–377

Smith, E. R. (1993). Social identity and social emotions:Toward new conceptualizations of prejudice. In D.M. Mackie, D. Hamilton, & D. Lewis (Eds.), Affect,cognition, and stereotyping: Interactive processes ingroup perception (pp. 297–315). San Diego: Aca-demic Press.

Sonnby-Borgström, M. (2002). Automatic mimicry re-actions as related to differences in emotional empa-thy. Scandinavian Journal of Psychology, 43, 433–443.

Snyder, C. R., Lassegard, M., & Ford, C. E. (1986). Dis-tancing after group success and failure: Basking in re-flected glory and cutting off reflected failure. Journalof Personality and Social Psychology, 51, 382–388.

Sorce, J. F., Emde, R. N., Campos, J. J., & Klinnert, M.D. (1985). Maternal emotional signaling: Its effect onthe visual cliff behavior of 1-year-olds. Developmen-tal Psychology, 21, 195–200.

Spears, R., & Leach, C. W. (2004). Intergroupschadenfreude: Conditions and consequences. In L.


Z. Tiedens & C. W. Leach (Eds.), The social life ofemotions (pp. 336–355). New York: Cambridge Uni-versity Press.

Spoor, J. R., & Kelly, J. R. (2004). The evolutionary sig-nificance of affect in groups: Communication andgroup bonding. Group Processes and Intergroup Re-lations, 7, 401–415.

Tangney, J. P., Miller, R. S., Flicker, L., & Barlow, D. H.(1996). Are shame, guilt, and embarrassment distinctemotions? Journal of Personality and Social Psychol-ogy, 70, 1256–1264.

Tiedens, L. Z. (2001). Anger and advancement versussadness and subjugation: The effect of negativeemotion expressions on social status conferral.Journal of Personality and Social Psychology, 80,86–94.

Tiedens, L. Z., & Leach, C. W. (Eds.). (2004). The sociallife of emotions. New York: Cambridge UniversityPress.

Timmers, M., Fischer, A. H., & Manstead, A. S. R.(1998). Emotion expression as a function of genderand social context. Personality and Social Psychol-ogy Bulletin, 24, 974–985.

Totterdell, P. (2000). Catching moods and hitting runs:Mood linkage and subjective performance in profes-sional sports teams. Journal of Applied Psychology,85, 848–859.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D.,

& Wetherell, M. S. (1987). Rediscovering the socialgroup: A self-categorization theory. Oxford:Blackwell.

Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S.R. (2004a). The interpersonal effects of anger andhappiness in negotiations. Journal of Personality andSocial Psychology, 86, 57–76.

Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S.R. (2004b). The interpersonal effects of emotions innegotiations: A motivated information processingapproach. Journal of Personality and Social Psychol-ogy, 87, 510–528.

Van Zomeren, M., Spears, R., Fischer, A. H., & Leach,C. W. (2004). Put your money where your mouth is!:Explaining collective action tendencies throughgroup-based anger and group efficacy. Journal ofPersonality and Social Psychology, 87, 649–664.

Vingerhoets, A. J. J. M., & Becht, M. C. (1997). TheISAC study: Some preliminary findings. Poster pre-sented at the annual meeting of the American Psycho-somatic Society, Santa Fe, NM.

Williams, K. D. (2001). Ostracism: The power of si-lence. New York: Guilford Press.

Zeelenberg, M., Van der Pligt, J., & Manstead, A. S. R.(1998). Undoing regret on Dutch television: Apolo-gizing for interpersonal regrets involving actions orinactions. Personality and Social Psychology Bulle-tin, 24, 1113–1119.


PA R T V

PERSONALITY ISSUES

C H A P T E R 2 9

Subjective Well-Being

RICHARD E. LUCAS and ED DIENER

Psychological health, like physical health, canbe judged on a variety of dimensions. Yet inboth realms, it is difficult to say which of thesedimensions are essential for overall well-being.Can a woman say that she is in good physicalshape because she is free from disease, or mustshe also have an abundance of energy and agreat deal of strength? Does a man have psy-chological well-being if he is free from depres-sion, or must he have a positive opinion of him-self and his life? Throughout the history of thefield, psychologists have tended to focus on thenegative end of the psychological spectrum,while often ignoring the positive (Myers &Diener, 1995). Partly in response to this unbal-anced focus, researchers have become increas-ingly interested in positive psychological out-comes. One area where this shifting focus isvery salient is the field of “subjective well-being” (SWB).

As its name implies, the field of SWB focuseson people’s own evaluations of their lives. Al-though measures such as crime statistics, healthindices, and indicators of wealth are surely re-lated to quality of life, these external indicators

cannot capture what it means to have a subjec-tive sense that one’s life is good. An activistfor society’s downtrodden, for example, maythrive in a high-crime, low-income neighbor-hood. The sense of meaning and personal satis-faction this individual gains from helping oth-ers may not be reflected in traditional socialindicators. People evaluate conditions differ-ently, depending on their expectations, values,and previous experiences. SWB researchers as-sign importance to this subjective element andassess individuals’ thoughts and feelings abouttheir lives.

DEFINING SWB

There are numerous ways in which an individ-ual can evaluate the quality of his or her life.Perhaps the most basic of these relates to themoment-to-moment affect that a person feelsover time. Emotions and moods are ubiquitousphenomena; they give a subjective feeling tovirtually all of one’s waking moments. In anexperience-sampling study of emotional expe-

471

rience, Diener, Sandvik, and Pavot (1991)found that people reported some affect virtu-ally all of the time. Furthermore, emotional ex-periences have valence. One of the most basicfeatures of these experiences is that a personcan easily tell whether they are positive or neg-ative (Kahneman, 1999). Given the ubiquity ofemotion and its valenced nature, it is not sur-prising that when people evaluate their ownwell-being, the pleasantness of their affect ap-pears to play a central role. Larsen (1989)found that scores on self-report SWB scales fallat the end of the pleasantness dimension of theemotion circumplex. Thus we begin by suggest-ing that the experience of SWB reflects a pre-ponderance of pleasant rather than unpleasantaffect in one’s life over time.

It is important to realize, however, that thisrelatively simple starting point belies the con-siderable complexity that is involved in defin-ing and measuring SWB. First, although it istempting to group all affective experience to-gether, research shows that pleasant and un-pleasant affect are not polar opposites. In per-haps the first investigation of this issue,Bradburn (1969) reported that pleasant andunpleasant affect were empirically separable,and that each correlated with a distinct set ofpersonality traits. Watson and his colleagues(e.g., Watson, Clark, & Tellegen, 1988; Watson& Tellegen, 1985) showed that positive andnegative affect formed distinct and orthogonalfactors, and the two-factor structure that theyidentified has been replicated often (Diener &Emmons, 1985; Warr, Barter, & Brownbridge,1983; Zevon & Tellegen, 1982).

Theorists on the other side of the indepen-dence debate suggest that pleasant and un-pleasant affect are really two poles of a singledimension, and that empirical evidence fortheir independence results from a variety of ar-tifacts. Green, Goldman, and Salovey (1993),for example, suggested that when measurementerror is controlled via structural equation mod-eling, the correlation between pleasant and un-pleasant affect becomes much stronger than itis when single-method observed variables areassessed. Similarly, Russell (1980) proposed acircumplex model of affect in which emotionscan be described in terms of their pleasantness–unpleasantness and in terms of their level ofarousal. According to this model, if arousal isheld constant, pleasant and unpleasant affectwill exhibit strong negative correlations.

Schimmack (2008) has provided a compre-hensive review that addresses most of the issues

surrounding the independence debate. Henotes that confusion about independence oftenarises because multiple types of independenceare sometimes discussed as if they were thesame thing. Citing Bradburn’s (1969) initialwork in this area, Schimmack distinguishes“structural independence,” “momentary inde-pendence,” and “causal independence.” Struc-tural independence refers to the extent towhich long-term levels of positive affect are in-dependent from long-term levels of negative af-fect. Schimmack’s review shows that even whenthe best multimethod techniques are used, andeven when positive and negative affect itemsare matched for levels of arousal, the correla-tion between positive and negative affect ismuch lower than 1.00. Thus there is supportfor the idea that there is some degree of struc-tural independence.

Momentary independence refers to the ques-tion of whether people can experience positiveand negative affect simultaneously. In contrastto the question of structural independence, itappears that there is quite a bit of support formomentary independence (Schimmack, 2008).Although people can experience low levels ofpositive affect at the same time they experiencelow levels of negative affect, the simultaneousexperience of more intense positive and nega-tive affect is rare (Diener & Iran-Nejad, 1986;Schimmack, 2001). Thus people’s intuition thatpositive and negative affect cannot be indepen-dent may be due to the fact that they are think-ing about momentary independence ratherthan structural independence.

Causal independence refers to the questionof whether positive and negative affect resultfrom different underlying causal processes(Schimmack, 2008). Interestingly, it is possiblethat even if positive and negative affect showstrong momentary correlations, they may stillbe causally independent (Larsen, McGraw, &Cacioppo, 2001). For instance, if independentunderlying systems are mutually inhibitory, thismay result in strong negative correlations at thelevel of a person’s experience. Personality re-searchers have often emphasized the fact thatdifferent personality traits are differentiallyassociated with positive and negative affect(e.g., Carver, Sutton, & Scheier, 2000; Elliot& Thrash, 2002; Tellegen, 1985). This pro-vides some initial evidence for causal indepen-dence of these two affective dimensions (seeSchimmack, 2008, for a more detailed review).

Finally, it is important to point out that theevaluation of one’s life does not end with emo-

472 V. PERSONALITY ISSUES

tional experience itself. As Kahneman and Riis(2005) put it, in addition to living their lives,people can also think about their lives. Andthese retrospective, cognitive evaluations candiffer from the average emotions that the per-sons experience over time. It is likely that peo-ple use their affective experiences as one sourceof information about their overall well-being.However, other sources will also play a role.For instance, it is possible that people may bewilling to endure high levels of negative affectif the situational factors leading to these affec-tive experiences ultimately contribute to theattainment of important life goals. Thus theassociations between affective and cognitivecomponents may vary across individuals. Infact, Suh, Diener, Oishi, and Triandis (1998)showed that different cultures use affective in-formation differently when judging overall lifesatisfaction. Among participants from individ-ualist nations, affective experience is stronglycorrelated with life satisfaction, whereasamong participants from collectivist nations,the associations are somewhat weaker. Thus,although some people have suggested that psy-chologists place too much emphasis on hedo-nism by focusing on SWB, the separability oflife satisfaction from affective feelings allowsfor a variety of paths to the subjectively goodlife. SWB does not result solely or inevitablyfrom hedonistic pleasures.

Because the various components of SWB areonly moderately intercorrelated (and becausethe size of these correlations may vary acrossdifferent populations), SWB researchers recom-mend assessing at least four different con-structs to get a relatively complete view of aperson’s subjective quality of life. Specifically,we (Diener, Suh, Lucas, & Smith, 1999) haverecommended assessing positive affect, nega-tive affect, life satisfaction, and satisfactionwith specific life domains (e.g., one’s health,one’s romantic partner, and one’s job). Al-though the various affective and cognitive com-ponents are separable (Lucas, Diener, & Suh,1996), they are often interrelated, suggestingthe existence of a higher-order construct ofSWB.

MEASURING SWB

SWB has often been measured by means of sim-ple one-time self-reports. These self-reportsmay consist of single-item or multiple-itemscales that ask respondents to reflect on how

happy they are or how happy they have beenover a circumscribed period of time. The evi-dence to date indicates that such self-reports ofhappiness are reliable and valid: Most instru-ments show impressive internal consistency,temporal stability, convergence with non-self-report measures of well-being, and acceptablelevels of criterion validity.

For instance, Fujita and Diener (2005) useddata from a 17-year nationally representativepanel study from Germany to show that year-to-year stabilities of a single-item life satisfac-tion measure averaged about .55. Schimmackand Oishi (2005) conducted a meta-analysis ofstability coefficients for life satisfaction mea-sures; they arrived at similar 1-year stabilitiesfor single-item measures and somewhat higherestimates (~.65) for multiple-item measures.Lucas and Donnellan (2007) analyzed two na-tionally representative longitudinal datasets(including the one that Fujita and Diener ana-lyzed), using latent state–trait models. Thesemodels allowed them to estimate the extent towhich stable trait and transient state varianceaffected life satisfaction judgments. Lucas andDonnellan found that approximately 30–37%of the variance was unchanging trait variancethat was stable across all waves of the study.An additional 30–33% of the variance was ac-counted for by a moderately stable autoregres-sive component that led to increased stabilityfrom one year to the next and reduced stabilityover time. Importantly, across the two studies,only about 35% of the variance in these single-item life satisfaction measures could be ac-counted for by an occasion-specific state com-ponent. Because this state component includedboth random measurement error and trueoccasion-specific variance, this figure providesa lower-bound estimate of the reliability ofsingle-item measure of life satisfaction. Thesestudies show that at least 65% (and probablymore) of the variance in these simple measuresis reliable.

Evidence for the validity of these measures isalso strong. In a study examining the conver-gent validity of well-being measures, Sandvik,Diener, and Seidlitz (1993) found strong evi-dence of convergence between self-reports ofSWB on the one hand, and interview ratings,peer reports, the average daily ratio of pleasantto unpleasant moods, and memory for pleasantminus unpleasant events on the other. Similarly,we (Lucas et al., 1996) showed that when mul-tiple measures of various well-being constructswere assessed, different measures of the same

29. Subjective Well-Being 473

construct tended to correlate more stronglythan did different constructs assessed with thesame method. This evidence of convergent anddiscriminant validity shows that SWB measuresassess valid variance, and that they do not sim-ply reflect social desirability or other irrelevantresponse styles and response sets.

Evidence of convergent and discriminant va-lidity is encouraging because it suggests thatexisting measures tap into something real.However, to be sure that these measures assesswhat psychologists think they measure, it isalso important to determine whether they be-have in ways that SWB theories would suggest.Measures of SWB tend to be sensitive to exter-nal conditions and responsive to changing lifecircumstances. For instance, Dijkers (1997)showed that individuals with spinal cord inju-ries tend to have lower levels of subjective qual-ity of life (see also Putzke, Richards, Hicken, &DeVivo, 2002). In addition, in a 6-year longitu-dinal study, Headey and Wearing (1989) foundthat positive and negative life events led toconcomitant increases and decreases in SWB.More recently, we have used nationally repre-sentative panel studies to show that whenmajor life events such as marriage, divorce,widowhood, and unemployment occur, life sat-isfaction scores change in response (Lucas,Clark, Georgellis, & Diener, 2003, 2004;Lucas, 2005). Similar results have also beenfound in national surveys: Reports from na-tions in turmoil show low well-being (e.g., Eu-rope after World War II, Eastern Europe in1991), suggesting that the measures are sensi-tive to external conditions in theoreticallymeaningful ways (Veenhoven, 1994).

Although these studies show that SWB mea-sures have some degree of validity, other re-searchers have suggested that they are alsosusceptible to the influence of irrelevant con-textual factors. In the most programmatic se-ries of studies, Schwarz, Strack, and their col-leagues found that the type of response scaleused, the response options, and the order andpresentation of questions can all influence thelevels of SWB that individuals report (seeSchwarz & Strack, 1999, for a review). For in-stance, in one of the most famous studies,Schwarz and Clore (1983) showed that peoplereported greater life satisfaction on sunny daysthan on cold rainy days. Presumably, theweather influenced mood, and people simplyrelied on their mood when constructing a lifesatisfaction judgment.

Although mood and other context effectssurely occur, research shows that these contex-tual factors often have only small effects on thevalidity of SWB measures. For instance, one fa-mous context effect that Schwarz and Strack(1999) reviewed concerned the order in whichspecific items were presented. Strack, Martin,and Schwarz (1988) found that participantswho were asked about their satisfaction withdating before being asked about their satisfac-tion with life showed stronger correlations be-tween the two questions than did participantswho received the questions in the reverse order.Presumably, the experimenters made partici-pants’ satisfaction with dating salient by askingabout this domain, and this increased salienceled participants to incorporate this informationinto their overall judgments of life satisfaction.Although such context effects may influencewell-being judgments, Schimmack and Oishi(2005) showed that chronically accessible fac-tors tend to have a much larger impact on lifesatisfaction judgments than do transient fac-tors made salient by situational manipulations.The fact that people tend to rely on the sameinformation when making their satisfactionjudgments results in considerable stability overtime.

Similarly, Eid and Diener (2004) showedthat mood effects on SWB ratings are oftenweak, and these transient factors tend to beoverwhelmed by stable variance. To demon-strate this, Eid and Diener assessed globalSWB, along with current mood, three timesover the course of a 3-month period. They usedstructural equation modeling techniques to de-termine how much variability in SWB could beexplained by stable trait factors versus tran-sient state factors. They showed that the major-ity of the SWB variance was stable over time.Perhaps more importantly, the occasion-specific associations between current moodand global SWB were weak and inconsistent.Thus mood appears to play a small role inglobal SWB judgments.

Nonetheless, Diener (1994) recommendsthat multimethod batteries be used to assessemotional well-being (see also Eid & Diener,2004, for a full discussion of multimethodtechniques). A major limitation of self-report isthat it relies exclusively on people’s cognitivelabels of their feelings. But the affective experi-ence that forms one major part of SWB is rec-ognized to be a multichannel phenomenon thatincludes physiological, facial, nonverbal, cog-


nitive, behavioral, and experiential compo-nents. In addition, there may be individual- orgroup-level differences in self-presentationalstyle, and these differences may affect self-reports of well-being. In order to obtain a com-plete picture of a person’s SWB, it is desirableto use informant reports, coding of nonverbalbehavior, and other techniques in order to as-sess the full range of cognitive and affectiveresponses about the person’s life. When themeasures converge, one will obtain greaterconfidence in the results. When the measuresdiverge, one will gain more complex knowl-edge of the emotional well-being of the individ-uals or groups being compared.

CAN HAPPINESS CHANGE?

When evaluating the quality of SWB measures,psychologists often point to their stability asevidence that they have some degree of reliabil-ity and validity. As discussed above, this pieceof evidence shows that contextual factorsdo not completely overwhelm these measures.Furthermore, because external circumstancesare not expected to change that frequently, thestability of SWB measures also agrees with in-tuitive theories about the way external circum-stances affect well-being. However, the stabilityof SWB measures also raises questions aboutthe extent to which happiness can change. Doexternal circumstances matter, or do SWBjudgments reflect an unchanging and perhapsbiologically determined outlook on life?

One important piece of evidence concerningthis question comes from studies that attemptto estimate the heritability of SWB by compar-ing the similarity of identical twins to that offraternal twins. In perhaps the first such study,Tellegen et al. (1988) used a separated-twin de-sign to assess the heritability of various well-being measures (all of which were derived fromthe Multidimensional Personality Question-naire). They showed that these heritabilitiesranged from .40 (for global positive emotional-ity) to .55 (for global negative emotionality).Thus about half of the variance could be ac-counted for by one’s genes. In addition, grow-ing up in the same household played almost norole in the similarity of twins; those who wereraised together were no more similar thanthose who were raised apart. These broadheritability estimates have been replicated of-ten by researchers using a variety of samples

and different measures of SWB (e.g., Nes,Roysamb, Tambs, Harris, & Reichborn-Kjennerud, 2006; Roysamb, Harris, Magnus,Vitterso, & Tambs, 2002; Roysamb, Tambs,Reichborn-Kjennerud, Neale, & Harris, 2003;Stubbe, Posthuma, Boomsma, & De Geus,2005).

Although the broad heritability of single-occasion measures of SWB is no longer contro-versial, some of the implications and exten-sions of this research can be. For instance, inan extension of the basic twin study design,Lykken and Tellegen (1996) examined the ex-tent to which SWB is stable over time and theextent to which the stable component is herita-ble. Consistent with previous studies, Lykkenand Tellegen found stability estimates near .50.In addition, they showed that the cross-twin,cross-time correlations were approximately.40. This means that about 80% of the stablevariance can be accounted for by genes. Theauthors concluded from this finding that “try-ing to be happier [may be] as futile as trying tobe taller” (p. 189). Although at first glance thispessimistic perspective appears to be supportedby Lykken and Tellegen’s data, there are threereasons why psychologists should hold outhope that happiness can in fact change.

First, it is true that strong heritability esti-mates suggest that inborn biological factorsplay a role in SWB. Yet, until the precise pro-cess is understood, it cannot be determinedwhether happiness is changeable or not. For in-stance, inborn traits could affect specific be-haviors, which in turn could affect happiness.If so, then interventions could be designed tochange the behaviors that people with thesepersonality traits tend to exhibit.

Second, Lykken and Tellegen’s (1996) studywas based on a relatively small sample of rela-tively young twins. Although the estimatesfrom this study have been replicated with onelarger sample of young twins from Norway(Nes et al., 2006), a separate study that as-sessed an older sample found much lower esti-mates for the heritability of the stable compo-nent of SWB (Johnson, McGue, & Krueger,2005). Thus it is unclear whether additionalfactors such as the age of the participants canmoderate the effects that Lykken and Tellegenreported.

Finally, additional research that directly ex-amines the effect of major life events showsthat external circumstances can have lasting ef-fects on SWB. For instance, we (Lucas et al.,


2003) found that although people eventuallyadapted to the death of a spouse, this adapta-tion process took approximately 8 years—a substantial proportion of the respondents’lives. Other events, such as divorce (Lucas,2005), unemployment (Lucas et al., 2004), andthe onset of a long-term disability (Lucas,2007), had even stronger and longer-lasting ef-fects. For example, the onset of a serious dis-ability was associated with more than a fullstandard deviation’s drop in life satisfaction(Lucas, 2007). Importantly, this longitudinalstudy showed that very little adaptation oc-curred following this important life event.Finally, all of these studies showed considerableindividual differences in reaction and adapta-tion to life events. Even where participantsadapted to events on average, some partici-pants experienced lasting changes from theirown personal baseline. Thus studies of lifeevents show that life circumstances can have animportant impact on SWB.

CORRELATES OF SWB

In 1967 Wilson reviewed the limited empiricalevidence regarding the “correlates of avowedhappiness.” He concluded that the happy per-son is a “young, healthy, well-educated, well-paid, extroverted, optimistic, worry-free, reli-gious, married person with high self-esteem,job morale, modest aspirations, of either sex,and of a wide range of intelligence” (p. 294). Inthe 40 years since Wilson’s review, thousandsof studies have been conducted, and we nowknow much more about the correlates of SWB(see Diener et al., 1999, for a comprehensivereview).

A number of Wilson’s conclusions havestood the test of time. Most significantly, Wil-son was correct about (and probably underesti-mated the importance of) personality. Re-searchers consistently find that the personalitytraits of extraversion, neuroticism, optimism,and self-esteem correlate with measures ofemotional well-being (Diener & Lucas, 1999).However, we must caution that the pattern ofrelations may vary across cultures. Diener andDiener (1995) found that the size of the corre-lation between self-esteem and life satisfactionwas greater in individualistic nations than incollectivistic nations, perhaps because the for-mer place greater emphasis on autonomy andinternal feelings. Another study (Lucas, Diener,

Grob, Suh, & Shao, 1999) found that extra-version was correlated less strongly withpleasant affect in collectivistic nations than inindividualistic nations (though the correlationwas strong in both). A third study (Oishi,Diener, Suh, & Lucas, 1999) suggested thatvalues and goals may mediate the relationsamong personality traits and well-being con-structs. Although the exact cause for the rela-tions between personality and SWB is unclear,the relation itself is robust.

Wilson (1967) was also correct that incomeplays a role in SWB, though this relation ismore complex than Wilson could have known.Studies of personal wealth, income change, na-tional wealth (i.e., gross national product), andstudies of very wealthy individuals consistentlyfind significant positive correlations betweenwealth and SWB (Diener & Biswas-Diener,2002). These correlations are much larger atthe national level than at the individual level.However, even the relatively small effects thatare found at the individual level can have prac-tical importance. Furthermore, the relation isoften larger among those at the lowest levels ofincome (Diener, Sandvik, Seidlitz, & Diener,1993), and even very small differences in abso-lute wealth can have important implicationsfor happiness among those with very few re-sources (Biswas-Diener & Diener, 2001).

Marriage and religion also show consistentcorrelations with SWB. A number of studies inthe United States (e.g., Glenn, 1975; Gove &Shin, 1989) and other countries (e.g.,Mastekaasa, 1995; White, 1992) support Wil-son’s (1967) conclusion that married individu-als tend to be happier than divorced or unmar-ried individuals. However, this association maynot be due to the causal effect of marriage onhappiness. Although divorce is associated withlasting changes in happiness (Lucas, 2005),marriage is not (Lucas et al., 2003). Instead, itappears that single people who will eventuallymarry are actually happier than average, evenbefore the marriage occurs.

Studies on religion show that religious peo-ple are often happier than nonreligious individ-uals. More specifically, SWB correlates with re-ligious certainty (Ellison, 1991), strength ofone’s relationship with the divine (Pollner,1989), prayer experiences (Poloma &Pendleton, 1991), and participation in religiousactivities (Ellison, Gay, & Glass, 1989; Gart-ner, Larson, & Allen, 1991). This effect holdseven after other demographic factors (such as


age, income, and education) are controlled for(Diener et al., 1999).

Although Wilson’s (1967) description of thehappy individual was accurate in a number ofrespects, a few of Wilson’s conclusions havebeen overturned by subsequent research. Mostsignificantly, SWB researchers now questionthe popular notion that people become moreunhappy as they age. In a study that examinednational probability samples from 40 nations,Diener and Suh (1998) found that althoughpleasant affect tended to decline with age, therewere no significant trends in life satisfaction orunpleasant affect. Furthermore, decreases inpositive affect may be due to the fact that moststudies measure aroused types of pleasant emo-tions. If less aroused emotions such as “con-tentment” and “affection” are examined, agedeclines may not be found (Diener & Suh,1998). Of course, the precise patterns ofchange that are found within a single study de-pend somewhat on the age range examined andthe type of design used (Pinquart, 2001). Forinstance, Mroczek and Spiro (2005) found thatlife satisfaction increased from age 40 to 65,but then declined after that (particularly as re-spondents approached death).

Recent research also suggests that the associ-ation between health and SWB may be morecomplicated than Wilson realized. Althoughhealth is positively associated with well-being,this relation holds tends to be stronger for self-reports than for more objective measures ofhealth (e.g., George & Landerman, 1984;Larson, 1978; Okun, Stock, Haring, & Witter,1984). When more objective health data (e.g.,physicians’ ratings) are examined, the correla-tion sometimes weakens considerably or evendisappears (e.g., Watten, Vassend, Myhrer, &Syversen, 1997).

One possible explanation for this discrep-ancy is that self-reports of health reflect one’slevel of emotional adjustment as well as one’sobjective physical condition (Hooker & Siegler,1992; Watson & Pennebaker, 1989). There-fore, the relation between health and SWB maybe artificially inflated by this emotional com-ponent. However, an alternative explanation isthat objective measures of health are them-selves flawed. Nelson et al. (1983), for exam-ple, examined the correspondence between pa-tients’ and doctors’ reports of health. Therewere numerous disagreements between the re-ports, and when Nelson et al. asked doctorsabout the reasons for these discrepancies, the

doctors admitted that 44% of the disagree-ments were due to clinician error. Anadditional 12% were due to insufficient knowl-edge of the patients. This study demonstratesthat “objective” reports can be wrong, and thatpatients’ subjective ratings may actually bemore valid. In fact, a number of studies havenow shown that subjective reports of healthcan predict outcomes such as longevity, evenafter objective reports of health are controlledfor (Ganz, Lee, & Siau, 1990; McClellan,Anson, Birkeli, & Tuttle, 1991; Mossey &Shapiro, 1982; Rumsfeld et al., 1999). Thusthe association between subjective ratings ofhealth and SWB may reflect more than justshared method variance.

THEORIES OF THE CAUSES OF SWB

Much of the research summarized above an-swers the question “What external conditionsare necessary for happiness to ensue?” Do peo-ple need to have a lot of money to be happy, orcan they have a modest income? Do they needto be married, or can they be single? Thesequestions reflect the underlying theoretical as-sumption that people have universal needs, andthat the degree to which these needs are met byexternal circumstances and personal resourcesdetermines happiness (Wilson, 1967). Unfortu-nately, the effects of these variables are oftensmall, leading some researchers to complain ofa lack of theoretical progress in the field (e.g.,Ryff, 1989; Wilson, 1967). Fortunately, re-searchers have attempted to balance this em-phasis on external factors and demographicswith a greater focus on the psychological vari-ables that moderate the effects of life circum-stances. People’s needs and resources must beexamined in the context of individual lives,goals, values, and personalities. Although nocomplete theoretical formulation is available,we (Diener et al., 1999) have suggested somecomponents that such a theory must include.

Dispositional Influences

One important moderator of situational effectson happiness is personality. We (Diener &Lucas, 1999) have reviewed findings from tem-perament studies, heritability studies, longitu-dinal studies, and cross-situational consistencystudies that suggest the existence of stable emo-tional styles of responding to events and cir-


cumstances. Although people may often re-spond similarly to similar events, the intensityand duration of those responses are likely to beinfluenced by people’s personality. Although anumber of traits exhibit reliable correlationswith SWB (including self-esteem, optimism,agreeableness, and conscientiousness), re-searchers often focus on the widely replicatedassociations between extraversion and positiveaffect on the one hand, and neuroticism andnegative affect on the other (Diener & Lucas,1999).

McCrae and Costa (1991) suggested thatthere are at least two general classes of expla-nations for the associations between traits likeextraversion and neuroticism and SWB vari-ables like positive and negative affect. Accord-ing to instrumental explanations, personalityaffects well-being indirectly through choice ofsituations or through differential propensitiesto experience positive and negative life events.For instance, extraverts may be happier thanintroverts because they spend more time in so-cial situations, and this increased social activitymay lead to greater positive affect. Support forsuch instrumental models comes from studiesshowing that extraversion and neuroticismprospectively predict positive and negative lifeevents, respectively (e.g., Headey & Wearing,1989; Magnus, Diener, Fujnita, & Pavot,1996).

In contrast to these instrumental theories,temperament theorists posit that there is a di-rect path from personality traits to affective ex-perience. For instance, Gray (1981) proposedthat much of the variability in personality canbe traced to three fundamental systems thatregulate reactions to signals of conditioned re-ward (the “behavioral activation system” orBAS), signals of conditioned punishment (the“behavioral inhibition system” or BIS), andsignals of unconditioned punishment (the“fight–flight system” or FFS). Extraversion ap-pears to be linked to individual differences inthe BAS, whereas neuroticism appears to belinked to individual differences in the BIS(Tellegen, 1985). Thus extraverts should bemore sensitive to signals of reward than are in-troverts, and this reward sensitivity should takethe form of enhanced information processingand increased positive emotions when extra-verts are exposed to rewarding stimuli. Simi-larly, individuals who are high in neuroticismshould be more sensitive to signals of punish-ment than are more stable individuals. Re-

searchers have tested these theories by examin-ing individual differences in attention to (e.g.,Derryberry & Reed, 1994), memory for (e.g.,Rusting, 1998), and reaction to positive andnegative stimuli (e.g., Canli et al., 2001; Larsen& Ketelaar, 1991; Lucas & Baird, 2004).

It is important to note that these explana-tions of the personality–SWB association arenot mutually exclusive. Extraverts may attendmore and respond more strongly to rewardingstimuli, while simultaneously engaging in be-haviors that make rewarding circumstancesmore likely to occur. A theory of SWB musttake these dispositional influences into consid-eration when evaluating the effects of externalinfluences on well-being. Although the loss of aloved one may cause sadness and grief in mostindividuals, the intensity, duration, and meansof coping with this tragedy may all be influ-enced by one’s personality.

Goals

A number of theoretical approaches to SWBcan be grouped together as “telic theories”(Diener, 1984). These approaches posit thathappiness occurs when a person arrives atsome end state. The end state may be set by in-nate biological drives, as in some need theories(e.g., Maslow, 1954); by psychological needs ormotives (Murray, 1938); or by conscious goals(Emmons, 1986). These models are similar tothe dispositional explanation of well-being, inthat they recognize that different events andcircumstances affect individuals differently, de-pending on the context of their lives. Becauseneeds and goals depend on learning, life cycle,and biological factors, SWB may result fromquite different telic states for different people atdifferent times in their lives.

At their simplest, goal theories suggest thatprogress toward goals and attainment ofgoals lead to increased positive affect (e.g.,Brunstein, 1993; Emmons, 1986). However,more complex theories of the relations be-tween goals and SWB have been proposed.Emmons (1986), for example, finds evidencethat certain characteristics of goals (ratherthan progress toward goals) affect happiness.The value that one places on goals and thedegree of effort that the goal requires tend tobe associated with positive affect, whereasconflict between goals and ambivalence to-ward goals tends to be associated with nega-tive affect. Other researchers (e.g., Ryan,


Sheldon, Kasser, & Deci, 1996) argue thatcertain types of goals are more likely to fosterwell-being than are others. More specifically,Kasser and Ryan (1993) found that self-acceptance, affiliation, and community goalswere positively associated with higher SWB,whereas financial success, social recognition,and physical attractiveness goals were associ-ated with lower SWB. Kasser and Ryan arguethat the former goals reflect intrinsic needssuch as autonomy, relatedness, and compe-tence, and that the fulfillment of these needsis most important for SWB.

Many goal theories have the ability to ex-plain the small effects of resources and otherbottom-up factors: Only those factors that re-late to one’s goals (which can vary across indi-viduals) should influence well-being. Cantorand Sanderson (1999), for example, have ar-gued that the importance of certain goalschanges across individuals, cultures, and devel-opmental phases. According to their theory,SWB when an individual chooses and is able toattain goals that are valued by the individualand culture, and that are appropriate to the in-dividual’s developmental phase. We (Oishi etal., 1999) found that achievement in certaindomains influenced overall well-being onlywhen those domains were related to respon-dents’ goals. Thus goals provide an importantcontext for the events and circumstances thatindividuals experience. We must understandthis individual context before we can under-stand the impact of events.

Culture

An additional contextual factor that is relatedto theories of goals and SWB is the culture inwhich individuals are immersed. Althoughgoals can vary among individuals within a cul-ture, certain goals may be more prevalent incertain cultures. These differences may result indifferential levels of happiness in different cul-tures. However, the impact of culture on indi-viduals’ well-being extends beyond the goalsthat they hold. Culture also affects the weight-ing of different sources of pleasure and pain.Thus SWB may mean different things in differ-ent places.

For example, Suh et al.’s (1998) research oncross-cultural variation in the determinants oflife satisfaction suggests that cultural differ-ences in one’s outlook can have important im-plications for theories of SWB. Considerable

research suggests that people living in individu-alist cultures tend to view the self as anautonomous, self-sufficient entity (Markus &Kitayama, 1991). Consequently, feelings andemotions weigh heavily as determinants ofbehavior. Individuals living within collectivistcultures, on the other hand, stress harmonywith family and friends rather than stressingone’s autonomy from these people. Feelingsabout the self weigh less heavily in these cul-tures. This difference in outlook can explainwhy emotions were a stronger determinant ofglobal life satisfaction in individualist culturesthan they were in collectivist cultures in Suh etal.’s study. Similarly, Diener and Diener (1995)found that feelings about the self (i.e., self-esteem) were more highly related to life satis-faction in individualist nations than in collec-tivist nations. Just as individuals may have dif-ferent goals across cultures, they may havedifferent ideas of what it means to be happy.

Social Comparison

It has been hypothesized that a person’s level ofSWB is in part determined by comparisons heor she makes with standards (Michalos, 1985).Often people’s standards come from observingpeople around them or from rememberingwhat they themselves were like in the past. Ifpeople exceed these standards, they should behappy and satisfied, but if they fall short oftheir standards, they might experience low lev-els of well-being. Although this process seemsstraightforward, the process of making socialcomparisons is more subtle than originally be-lieved (Diener & Fujita, 1997). People do notsimply look around and judge their happinessby their distance above or below their friendsand neighbors on relevant domains. Instead,people choose their targets for comparison, theinformation to which they attend, and the waythey use this information in complex ways.

Wood (1996) stated that social comparisonis simply “the process of thinking about infor-mation about one or more other people in rela-tion to the self” (p. 520). The targets forcomparison can be proximate individuals, indi-viduals people see on television, or even indi-viduals that people construct in their minds(Wood, Taylor, & Lichtman, 1985). People canpay attention to similarities, dissimilarities, orboth when comparing themselves to a target.And, finally, people do not always contrastthemselves with the target (i.e., they do not al-


ways feel unhappy when they are worse off,and they do not always feel happy when theyare better off). Some people may find compari-sons with more successful individuals motivat-ing, while others may simply be reminded oftheir own lack of success. Diener and Fujita(1997) suggested that social comparison mayactually be used as a coping strategy and maybe influenced by personality and SWB. Thechoice of comparison target, the type of infor-mation to which an individual attends, and theway he or she uses this information may be aresult rather than a cause of SWB.

Adaptation, Adjustment, and Coping

Just as certain events may affect people differ-ently, depending on their goals and personali-ties, reactions to events are markedly different,depending on the amount of time that has pro-gressed since the event occurred. The processesthat can account for these differences are animportant part of theories of SWB. However,theories of adaptation have changed consider-ably in recent years. Adaptation effects wereonce thought to be so strong that some sug-gested that people were stuck on a “hedonictreadmill” (Brickman & Campbell, 1971). Al-though the treadmill theory allowed for short-lived reactions to events, many psychologistsbelieved that emotional reactions would even-tually subside and that people would inevitablyreturn to hedonic neutrality. Although this the-ory has played an important role in SWB re-search, more recent evidence shows that adap-tation processes are not as strong as oncethought.

For instance, we (Diener, Lucas, & Scollon,2006) have recently reviewed the evidence foradaptation and developed five important revi-sions to the treadmill theory. First, althoughsome adaptation occurs, it does not bring peo-ple back to hedonic neutrality. Instead, mostpeople are happy most of the time (Diener &Diener, 1996). Second, if SWB set points exist,they vary across individuals. Personality re-search shows that there are considerable indi-vidual differences in long-term levels of happi-ness. Third, adaptation occurs in differentways for different components of SWB. Al-though one’s affective experiences may returnto baseline levels after an important event hasoccurred, it is possible that levels of life satis-faction will not (see, e.g., Lucas, 2007). Fourth,adaptation does not inevitably occur. As re-

viewed above, some life events do seem toaffect SWB permanently. Finally, there are im-portant individual differences in adaptation,with some people quickly returning to baselinelevels and others experiencing permanent ef-fects.

Although psychologists know much moreabout adaptation than they used to know, anumber of important questions remain. For ex-ample, it is unclear how the nature of the eventaffects adaptation. Do people adapt to worsen-ing conditions, or only to stable conditions orone-time events? Do unexpected events lead toless adaptation than predictable events do(Frederick & Loewenstein, 1999)? Similarly, itis unclear what individual-difference variablespredict adaptation to life events (though seeBonanno, 2004, for a discussion of some possi-bilities). Finally, it is unclear whether adapta-tion represents a decrease in emotional re-sponse, an adjustment in goals and strategiesfor coping with the event, or both. A theory ofsubjective SWB must more clearly explain theprocesses responsible for adaptation: Whendoes it occur, what processes are responsiblefor adaptation, and what are the limits to indi-viduals’ abilities to adapt?

THE EFFECTS OF HAPPINESS

Although people tend to think about happinessas an end state that results from the accom-plishment of desired goals, psychologists havebegun to ask whether happiness itself can playa role in helping people attain those goals. Forinstance, Fredrickson (1998) developed onewell-known model of the function of positiveemotions. She suggested that joy and otherpleasant emotions broaden one’s thinking andhelp to build resources. The evidence she re-viewed shows that positive emotions can fostercreativity and can lead people to try newthings. This broadened approach to the worldallows happy people to develop new strategiesfor overcoming challenges, new social skills,and even new physical resources. Together,these resources can lead to positive outcomes inpeople’s lives.

More recently, Lyubomirsky, King, andDiener (2005) have compiled an extensive listof domains in which happier people do betterthan less happy people. Their meta-analytic re-view (which covered both correlational and ex-perimental research) showed that happy people


are more likely to get married, to have morefriends, to make more money in the future, toperform better at work, to have better health,and perhaps even to live longer than unhappypeople. Thus, although the stereotype of thehappy person is someone who is somewhatlazy and unthinking, a growing body of evi-dence suggests that happiness can be functionaland may promote positive outcomes in people’slives.

FUTURE DIRECTIONSFOR SWB RESEARCH

SWB reflects the extent to which people thinkand feel that their lives are going well. In the fu-ture, psychologists will need to continue toidentify the processes that promote high levelsof this values characteristic. However, some ofthe most exciting developments are takingplace outside the traditional boundaries of thefield. Because subjective measures capture as-pects of the good life that cannot be assessedvia traditional social indicators, psychologistsand economists have increasingly called for na-tional and local governments to track the SWBof populations over time (Diener & Seligman,2004; Kahneman, Kruger, Schkade, Schwartz,& Stone, 2003). If environmental factors andpolicy decisions reliably affect the SWB of thesepopulations, then this information could beused to guide future policy decisions. Ulti-mately, this would allow governments to usethe results of SWB research to increase overallquality of life. Thus, although SWB researchwas initially developed within the context ofexisting psychological theory, it has importantimplications for research, theory, and applica-tion in a wide variety of domains.

REFERENCES

Biswas-Diener, R., & Diener, E. (2001). Making the bestof a bad situation: Satisfaction in the slums of Cal-cutta. Social Indicators Research, 55, 329–352.

Bonanno, G. A. (2004). Loss, trauma, and human resil-ience: Have we underestimated the human capacityto thrive after extremely aversive events? AmericanPsychologist, 59, 20–28.

Bradburn, N. M. (1969). The structure of psychologicalwell-being. Chicago: Aldine.

Brickman, P., & Campbell, D. T. (1971). Hedonic rela-tivism and planning the good society. In M. H.Appley (Ed.), Adaptation level theory: A symposium(pp. 287–302). New York: Academic Press.

Brunstein, J. C. (1993). Personal goals and subjectivewell-being: A longitudinal study. Journal of Personal-ity and Social Psychology, 65, 1061–1070.

Canli, T., Zhao, Z., Desmond, J. E., Kang, E., Gross, J.,& Gabrieli, J. D. E. (2001). An fMRI study of per-sonality influences on brain reactivity to emotionalstimuli. Behavioral Neuroscience, 115, 33–42.

Cantor, N., & Sanderson, C. A. (1999). Life task partic-ipation and well-being: The importance of takingpart in daily life. In D. Kahneman, E. Diener, & N.Schwarz (Eds.), Well-being: The foundations ofhedonic psychology (pp. 230–243). New York: Rus-sell Sage Foundation.

Carver, C. S., Sutton, S. K., & Scheier, M. F. (2000). Ac-tion, emotion, and personality: Emerging conceptualintegration. Personality and Social Psychology Bulle-tin, 26, 741–751.

Derryberry, D., & Reed, M. A. (1994). Temperamentand attention: Orienting toward and away from posi-tive and negative signals. Journal of Personality andSocial Psychology, 66, 1128–1139.

Diener, E. (1984). Subjective well-being. PsychologicalBulletin, 95, 542–575.

Diener, E. (1994). Assessing subjective well-being: Prog-ress and opportunities. Social Indicators Research,31, 103–157.

Diener, E., & Biswas-Diener, R. (2002). Will money in-crease subjective well-being?: A literature review andguide to needed research. Social Indicators Research,57, 119–169.

Diener, E., & Diener, C. (1996). Most people are happy.Psychological Science, 7, 181–185.

Diener, E., & Diener, M. (1995). Cross-cultural corre-lates of life satisfaction and self-esteem. Journal ofPersonality and Social Psychology, 68, 653–663.

Diener, E., & Emmons, R. A. (1985). The independenceof positive and negative affect. Journal of Personalityand Social Psychology, 47, 1105–1117.

Diener, E., & Fujita, F. (1997). Social comparisons andsubjective well-being. In B. Buunk & R. Gibbons(Eds.), Health, coping, and social comparison(pp. 329–357). Mahwah, NJ: Erlbaum.

Diener, E., & Iran-Nejad, A. (1986). The relationship inexperience between various types of affect. Journal ofPersonality and Social Psychology, 50, 1031–1038.

Diener, E., & Lucas, R. E. (1999). Personality and sub-jective well-being. In D. Kahneman, E. Diener, & N.Schwarz (Eds.), Well-being: The foundations ofhedonic psychology (pp. 213–229). New York: Rus-sell Sage Foundation.

Diener, E., Lucas, R. E., & Scollon, C. N. (2006). Be-yond the hedonic treadmill: Revisions to the adapta-tion theory of well-being. American Psychologist, 61,305–314.

Diener, E., Sandvik, E., & Pavot, W. (1991). Happinessis the frequency, not the intensity, of positive versusnegative affect. In F. Strack, M. Argyle, & N.Schwarz (Eds.), Subjective well-being: An interdisci-plinary perspective (pp. 119–139). Elmsford, NY:Pergamon Press.


Diener, E., Sandvik, E., Seidlitz, L., & Diener, M.(1993). The relationship between income and subjec-tive well-being: Relative or absolute. Social Indica-tors Research, 28, 195–223.

Diener, E., & Seligman, M. E. P. (2004). Beyond money:Toward an economy of well-being. Psychological Sci-ence in the Public Interest, 5, 1–31.

Diener, E., & Suh, E. M. (1998). Subjective well-beingand age: An international analysis. Annual Review ofGerontology and Geriatrics, 17, 304–324.

Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. (1999).Subjective well-being: Three decades of progress.Psychological Bulletin, 125, 276–302.

Dijkers, M. (1997). Quality of life after spinal cord in-jury: A meta analysis of the effects of disablementcomponents. Spinal Cord, 35, 829–840.

Eid, M., & Diener, E. (2004). Global judgments of sub-jective well-being: Situational variability and long-term stability. Social Indicators Research, 65, 245–277.

Elliot, A. J., & Thrash, T. M. (2002). Approach–avoidance motivation in personality: Approach andavoidance temperaments and goals. Journal of Per-sonality and Social Psychology, 82, 804–818.

Ellison, C. G. (1991). Religious involvement and subjec-tive well-being. Journal of Health and Social Behav-ior, 32, 80–99.

Ellison, C. G., Gay, D. A., & Glass, T. A. (1989). Doesreligious commitment contribute to individual lifesatisfaction? Social Forces, 68, 100–123.

Emmons, R. A. (1986). Personal strivings: An ap-proach to personality and subjective well-being.Journal of Personality and Social Psychology, 51,1058–1068.

Frederick, S., & Loewenstein, G. (1999). Hedonic adap-tation. In D. Kahneman, E. Diener, & N. Schwarz(Eds.), Well-being: The foundations of hedonic psy-chology (pp. 302–329). New York: Russell SageFoundation.

Fredrickson, B. L. (1998). What good are positive emo-tions? Review of General Psychology, 2(3), 300–319.

Fujita, F., & Diener, E. (2005). Life satisfaction setpoint: Stability and change. Journal of Personalityand Social Psychology, 88, 158–164.

Ganz, P. A., Lee, J. J., & Siau, J. (1991). Quality of lifeassessment: An independent prognostic variable forsurvival in lung cancer. Cancer, 67, 3131–3135.

Gartner, J., Larson, D. B., & Allen, G. D. (1991). Reli-gious commitment and mental health: A review ofthe empirical literature. Journal of Psychology andReligion, 19, 6–25.

George, L. K., & Landerman, R. (1984). Health andsubjective well-being: A replicated secondary analy-sis. International Journal of Aging and Human De-velopment, 19, 133–156.

Glenn, N. D. (1975). The contribution of marriage tothe psychological well-being of males and females.Journal of Marriage and Family Relations, 37, 594–600.

Gove, W. R., & Shin, H. (1989). The psychological well-

being of divorced and widowed men and women.Journal of Family Issues, 10, 122–144.

Gray, J. A. (1981). A critique of Eysenck’s theory of per-sonality. In H. J. Eysenck (Ed.), A model for person-ality (pp. 246–276). New York: Springer-Verlag.

Green, D. F., Goldman, S., & Salovey, P. (1993). Mea-surement error masks bipolarity in affect ratings.Journal of Personality and Social Psychology, 64,1029–1041.

Headey, B., & Wearing, A. (1989). Personality, lifeevents, and subjective well-being: Toward a dynamicequilibrium model. Journal of Personality and SocialPsychology, 57, 731–739.

Hooker, K., & Siegler, I. C. (1992). Separating applesfrom oranges in health ratings: Perceived health in-cludes psychological well-being. Behavior, Health,and Aging, 2, 81–92.

Johnson, W., McGue, M., & Krueger, R. F. (2005). Per-sonality stability in late adulthood: A behavioral ge-netic analysis. Journal of Personality, 73(2), 523–551.

Kahneman, D. (1999). Objective happiness. In D.Kahneman, E. Diener, & N. Schwarz (Eds.), Well-be-ing: The foundations of hedonic psychology (pp. 3–25). New York: Russell Sage Foundation.

Kahneman, D., Krueger, A. B., Schkade, D., Schwarz,N., & Stone, A. (2003). Toward national well-beingaccounts. American Economic Review, 94, 429–434.

Kahneman, D., & Riis, J. (2005). Living and thinkingabout it: Two perspectives on life. In F. A. Huppert, N.Baylis, & B. Keverne (Eds.), The science of well-being(pp. 285–304). New York: Oxford University Press.

Kasser, T., & Ryan, R. M. (1993). A dark side of theAmerican dream: Correlates of financial success as acentral life aspiration. Journal of Personality and So-cial Psychology, 65, 410–422.


Larsen, R. J. (1989, August). Personality as an affectdispositional system. In L. A. Clark & D. Watson(Chairs), Emotional bases of personality. Symposiumconducted at the meeting of the American Psycholog-ical Association, New Orleans.

Larsen, R. J., & Ketelaar, T. (1991). Personality and sus-ceptibility to positive and negative emotional states.Journal of Personality and Social Psychology, 61,132–140.

Larson, R. (1978). Thirty years of research on the sub-jective well-being of older Americans. Journal ofGerontology, 33, 109–125.

Lucas, R. E. (2005). Time does not heal all wounds: Alongitudinal study of reaction and adaptation to di-vorce. Psychological Science, 16, 945–950.

Lucas, R. E. (2007). Long-term disability is associatedwith lasting changes in subjective well-being: Evi-dence from the two nationally representative longitu-dinal studies. Journal of Personality and Social Psy-chology, 92, 717–730.


Lucas, R. E., & Baird, B. M. (2004). Extraversion andemotional reactivity. Journal of Personality and So-cial Psychology, 86, 473–485.

Lucas, R. E., Clark, A. E., Georgellis, Y., & Diener, E.(2003). Re-examining adaptation and the set pointmodel of happiness: Reactions to changes in maritalstatus. Journal of Personality and Social Psychology,84, 527–539.

Lucas, R. E., Clark, A. E., Georgellis, Y., & Diener, E.(2004). Unemployment alters the set-point for lifesatisfaction. Psychological Science, 15, 8–13.

Lucas, R. E., Diener, E., Grob, A., Suh, E. M., & Shao,L. (2000). Cross-cultural evidence for the fundamen-tal features of extraversion. Journal of Personalityand Social Psychology, 79, 452–468.

Lucas, R. E., Diener, E., & Suh, E. M. (1996). Discrimi-nant validity of well-being measures. Journal of Per-sonality and Social Psychology, 71, 616–628.

Lucas, R. E., & Donnellan, M. B. (2006). How stable ishappiness?: Using the STARTS model to estimate thestability of life satisfaction. Journal of Research inPersonality, 41, 1091–1098.

Lykken, D., & Tellegen, A. (1996). Happiness is a sto-chastic phenomenon. Psychological Science, 7, 186–189.

Lyobomirsky, S., King, L., & Diener, E. (2005). Thebenefits of frequent positive affect: Does happinesslead to success? Psychological Bulletin, 131, 803–855.

Magnus, K. B., Diener, E., Fujita, F., & Pavot, W.(1993). Extraversion and neuroticism as predictorsof objective life events: A longitudinal analysis. Jour-nal of Personality and Social Psychology, 65, 316–330.

Markus, H. R., & Kitayama, S. (1991). Culture and theself: Implications for cognition, emotion, and moti-vation. Psychological Review, 98, 224–253.

Maslow, A. H. (1954). Motivation and personality. NewYork: Harper.

Mastekaasa, A. (1995). Age variations in the suiciderates and self-reported subjective well-being of mar-ried and never-married persons. Journal of commu-nity and Applied Social Psychology, 5, 21–39.

McClellan, W. M., Anson, C., Birkeli, K., & Tuttle, E.(1991). Functional status and quality of life: Predic-tors of early mortality among patients entering treat-ment for end stage renal disease. Journal of ClinicalEpidemiology, 44, 83–89.

McCrae, R. R., & Costa, P. T. (1991). Adding Liebe undArbeit: The full five-factor model and well-being.Personality and Social Psychology Bulletin, 17, 227–232.

Michalos, A. C. (1985). Multiple discrepancies the-ory (MDT). Social Indicators Research, 16, 347–413.

Mossey, J. M., & Shapiro, E. (1982). Self-rated health:A predictor of mortality among the elderly. AmericanJournal of Public Health, 72, 800–808.

Mroczek, D. K., & Spiro, A. (2005). Change in life sat-isfaction during adulthood: Findings from the Vet-

erans Affairs Normative Aging Study. Journal ofPersonality and Social Psychology, 88, 189–202.

Murray, H. A. (1938). Explorations in personality. NewYork: Oxford University Press.

Myers, D. G., & Diener, E. (1995). Who is happy? Psy-chological Science, 6, 10–19.

Nelson, E., Conger, B., Douglass, R., Gephart, D., Kirk,J., Page, R., et al. (1983). Functional health statuslevels of primary care patients. Journal of the Ameri-can Medical Association, 249, 3331–3338.

Nes, R. B., Roysamb, E., Tambs, K., Harris, J. R., &Reichborn-Kjennerud, T. (2006). Subjective well-be-ing: Genetic and environmental contributions to sta-bility and change. Psychological Medicine, 36(7),1033–1042.

Oishi, S., Diener, E., Suh, E. M., & Lucas, R. E. (1999).Values as a moderator in subjective well-being. Jour-nal of Personality, 67, 157–184.

Okun, M. A., Stock, W. A., Haring, M. J., & Witter, R.A. (1984). Health and subjective well-being: A meta-analysis. International Journal of Aging and HumanDevelopment, 19, 111–132.

Pinquart, M. (2001). Age differences in perceived posi-tive affect, negative affect, and affect balance in mid-dle and old age. Journal of Happiness Studies, 2,375–405.

Pollner, M. (1989). Divine relations, social relations,and well-being. Journal of Health and Social Behav-ior, 30, 92–104.

Poloma, M. M., & Pendleton, B. F. (1991). The effectsof prayer and prayer experiences on measures of gen-eral well-being. Journal of Psychology and Theology,29, 71–83.

Putzke, J. D., Richards, J. S., Hicken, B. L., & DeVivo,M. J. (2002). Predictors of life satisfaction: A spinalcord injury cohort study. Archives of Physical Medi-cine and Rehabilitation, 83, 555–561.

Roysamb, E., Harris, J. R., Magnus, P., Vitterso, J., &Tambs, K. (2002). Subjective well-being: Sex-specificeffects of genetic and environmental factors. Person-ality and Individual Differences, 32(2), 211–223.

Roysamb, E., Tambs, K., Reichborn-Kjennerud, T.,Neale, M. C., & Harris, J. R. (2003). Happiness andhealth: Environmental and genetic contributions tothe relationship between subjective well-being, per-ceived health, and somatic illness. Journal of Person-ality and Social Psychology, 85(6), 1136–1146.

Rumsfeld, J. S., MaWhinney, S., McCarthy, M., Jr.,Shroyer, A. L. W., VillaNueva, C. B., O’Brien, M., etal. (1999). Health-related quality of life as a predic-tor of mortality following coronary artery bypassgraft surgery. Journal of the American Medical Asso-ciation, 281, 1298–1303.


Rusting, C. L. (1998). Personality, mood, and cognitiveprocessing of emotional information: Three concep-tual frameworks. Psychological Bulletin, 124, 165–196.


Ryan, R. M., Sheldon, K. M, Kasser, T., & Deci, E. L.(1996). All goals are not created equal: An organis-mic perspective on the nature of goals and their regu-lation. In P. M. Gollwitzer & J. A. Bargh (Eds.), Thepsychology of action: Linking cognition and motiva-tion to behavior (pp. 7–26). New York: GuilfordPress.

Ryff, C. D. (1989). Happiness is everything, or is it?: Ex-plorations on the meaning of psychological well-be-ing. Journal of Personality and Social Psychology,57, 1069–1081.

Sandvik, E., Diener, E., & Seidlitz, L. (1993). Subjectivewell-being: The convergence and stability of self-report and non-self-report strategies. Journal of Per-sonality, 61, 317–342.

Schimmack, U. (2001). Pleasure, displeasure, and mixedfeelings?: Are semantic opposites mutually exclusive?Cognition and Emotion, 15, 81–97.

Schimmack, U. (2008). The structure of subjective well-being. In M. Eid & R. J. Larsen (Eds.), The science ofsubjective well-being (pp. 97–123). New York:Guilford Press.

Schimmack U., & Oishi, S. (2005). Chronically accessi-ble versus temporarily accessible sources of life satis-faction judgments. Journal of Personality and SocialPsychology, 89, 395–406.

Schwarz, N., & Clore, G. L. (1983). Mood, misattri-bution, and judgments of well-being: Informativeand directive functions of affective states. Jour-nal of Personality and Social Psychology, 45, 513–523.

Schwarz, N., & Strack, F. (1999). Reports of subjectivewell-being: Judgmental processes and their method-ological implications. In D. Kahneman, E. Diener, &N. Schwarz (Eds.), Well-being: The foundations ofhedonic psychology (pp. 61–84). New York: RussellSage Foundation.

Strack, F., Martin, L. L., & Schwarz, N. (1988).Priming and communication: Social determinants ofinformation use in judgments of life satisfaction.European Journal of Social Psychology, 18, 429–442.

Stubbe, J. H., Posthuma, D., Boomsma, D. I., & DeGeus, E. J. C. (2005). Heritability of life satisfactionin adults: A twin–family study. Psychological Medi-cine, 35(11), 1581–1588.

Suh, E. M., Diener, E., Oishi, S., & Triandis, H. (1998).The shifting basis of life satisfaction judgments

across cultures. Journal of Personality and SocialPsychology, 70, 1091–1102.

Tellegen, A. (1985). Structures of mood and personalityand their relevance to assessing anxiety, with an em-phasis on self-report. In A. H. Tuma & J. D. Maser(Eds.), Anxiety and the anxiety disorders (pp. 681–706). Hillsdale, NJ: Erlbaum.

Tellegen, A., Lykken, D. T., Bouchard, T. J., Wilcox, K.J., Segal, N. L., & Rich, S. (1988). Personality simi-larity in twins reared apart and together. Journal ofPersonality and Social Psychology, 54(6), 1031–1039.

Veenhoven, R. (1994). Correlates of happiness: 7836findings from 603 studies in 69 nations: 1911–1994.Unpublished manuscript, Erasmus University, Rot-terdam, The Netherlands.

Warr, P., Barter, J., & Brownbridge, G. (1983). On theindependence of positive and negative affect. Journalof Personality and Social Psychology, 44, 644–651.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Devel-opment and validation of brief measures of positiveand negative affect: The PANAS scales. Journal ofPersonality and Social Psychology, 54, 1063–1070.

Watson, D., & Pennebaker, J. W. (1989). Health com-plaints, stress, and distress: Exploring the central roleof negative affectivity. Psychological Review, 96,234–254.

Watson, D., & Tellegen, A. (1985). Towards a consen-sual structure of mood. Psychological Bulletin, 98,219–235.

Watten, R. G., Vassend, D., Myhrer, T., & Syversen, J.L. (1997). Personality factors and somatic symptoms.European Journal of Personality, 11, 57–68.

White, J. M. (1992). Marital status and well-being inCanada. Journal of Family Issues, 13, 390–409.

Wilson, W. (1967). Correlates of avowed happiness.Psychological Bulletin, 67, 294–306.

Wood, J. V. (1996). What is social comparison and howshould we study it? Personality and Social Psychol-ogy Bulletin, 22, 520–537.

Wood, J. V., Taylor, S. E., & Lichtman, R. R. (1985). So-cial comparison in adjustment to breast cancer. Jour-nal of Personality and Social Psychology, 49, 1169–1183.

Zevon, M. A., & Tellegen, A. (1982). The structure ofmood change: An idiographic/nomothetic analysis.Journal of Personality and Social Psychology, 43,111–112.


C H A P T E R 3 0

Temperament and Emotion

JOHN E. BATES, JACKSON A. GOODNIGHT, and JENNIFER E. FITE

Temperament provides ways to talk about indi-vidual differences in emotion and how they de-velop. Temperament and emotion can be sepa-rated for the sake of theoretical analysis, butthey correspond in so many ways that it is clearthat they are part of the same whole (an im-pression supported by the 1,680,000 Googlelistings for the phrase “temperament and emo-tion” and 4,548 published sources identifiedby PsycINFO). As theory about emotion ad-vances, theory about temperament advances inthe same directions.

Since the chapter on temperament in the sec-ond edition of this handbook (Bates, 2000) ap-peared, the literature on emotion has grown,especially on the topic of self-regulation; corre-spondingly, this has been an especially impor-tant theme of work on temperament in thattime (Rothbart & Bates, 2006). Conversely, in-dividual differences and their development arecrucial parts of emotional phenomena. Closelyconnected to interest in individual differencesand development, and also providing a frame-work for understanding the basic phenomenaof emotions, is interest in social adaptation.Emotions serve crucial functions in socialadaptation, and individuals’ characteristic pat-

terns of emotions influence and are influencedby variations in social adaptations. Variationsin social adaptations are of interest in both ba-sic and applied research.

There are direct linkages between tempera-ment dimensions and social adjustment. Anumber of studies show that early childhoodtendencies to be distressed by novelty predictlater internalizing behaviors, that tendenciesto be impulsive or unmanageable predictexternalizing behaviors, and that tendenciesto be high in negative emotionality predictboth internalizing and externalizing behaviors(Bates, 1989; Rothbart & Bates, 2006). How-ever, the predictive links between tempera-ment and adjustment are of a moderate size,even if we allow for attenuated relations dueto error of measurement. This suggests that agiven temperament does not by itself produceadjustment. Developmental-systems theoriespropose many other causal factors. The mostprominent kinds of additional factors arefamily characteristics, such as parenting style.One would expect that temperament andfamily qualities might additively predict childadjustment outcomes. Empirical findings insupport of this expectation are surprisingly

485

sparse (Bates & Pettit, 2007). This may bebecause relatively few studies have evaluatedsuch additive models of social development.However, it may also be due to unreportedfailures to find such effects. If the latter istrue, it could be due to overlap in variancebetween child temperament and parenting,due to common measurement sources (oftenparent report) or due to shared genetic rootsfor both child and parent traits. Another pos-sible explanation is that temperament andfamily environment interact in shaping socialdevelopment (Chess & Thomas, 1984;Wachs, 2000). A temperament trait becomesa behavior problem only in some kinds offamilies, or a parenting characteristic pro-motes behavior problems only for childrenwith certain temperaments. Since the secondedition of this handbook (Lewis & Haviland-Jones, 2000) was published, there has beenremarkable growth in the literature on howtemperament and environmental characteris-tics interact with one another in shaping so-cial adaptations. This chapter considers tem-perament, emotion, and the development ofsocial adaptations, with special emphases onself-regulation and the interaction betweenemotional dispositions and environmentalqualities in shaping social development.

CONCEPTS OF EMOTIONAND TEMPERAMENT

Emotion Concepts

We think of “emotion” as a set of hypotheticalconstructs representing processes that rangeacross all levels of psychological theory—including genetics, biochemistry, neural sys-tems, perceptual and cognitive processes (bothconscious and nonconscious), motivation, af-fective displays, behavior, and social systems.We are most interested in the aspects of emo-tion constructs that help to describe socialtransactions. Although emotion constructsshare properties that tie them to the generalconcept of emotion, specific emotions (such asjoy, interest, fear, or anger) are the main foci ofscientific study, not emotion as an umbrellaconcept. More detailed consideration of the ba-sic emotion constructs guiding the presentchapter can be found in Bates (2000). Recentwritings have further refined appreciation ofthe neural and cognitive underpinnings of regu-lation of emotion (e.g., Lewis, 2005).

Temperament Concepts

We think of “temperament” as a set of hy-pothetical constructs describing individualdifferences in reactivity and self-regulation(Rothbart & Bates, 2006). One kind of reactiv-ity dimension includes negative reactions, suchas general distress and more specific fear andfrustration. For example, some infants becomeangry faster and more intensely than other in-fants when their arms are held down, prevent-ing them from reaching an attractive toy.Another kind of reactivity includes positive re-actions, such as interest and joy. For example,some infants show more excitement when theysee a toy and are more eager to reach it thanother infants. An emotion in one such situationdoes not define a temperament trait. Tempera-ment is the pattern of responses in a given typeof incentive condition across many occasions.

Such behavior patterns are rooted in biologi-cal traits, at both genetic and neural levels.They appear relatively early, befitting a conceptof core personality, but they are not fullyformed at birth. As brain and behavioral reper-toire develop, temperamental traits emerge andstabilize. The key example of this is effortfulcontrol of attention, which depends on matura-tion of the relevant frontal lobe structures ofthe infant’s brain, beginning late in the firstyear and continuing throughout childhood(Reuda, Posner, & Rothbart, 2004; Rothbart,Derryberry, & Posner, 1994). Any personalitytrait is by definition stable, and it would be fit-ting for temperament traits, since they are corein personality, to be highly stable. However,measures of temperament are typically foundto be only moderately stable.

Part of the limitation in stability of tempera-ment traits is theoretically appropriate, andpart is due to error of measurement of the tem-perament phenotype. The theoretically appro-priate part of the limited stability concerns thefundamental nature of temperament and howthis might be measured. Temperament can beconceived at three different levels: genetic, neu-ral, and behavioral. One might presuppose thatthe action of genes underlying temperamentwould be highly stable, but in theory it shouldonly be relatively stable. We may eventuallyhave specific genetic configurations that under-lie temperament. This seems possible as a prod-uct of today’s extremely vigorous efforts to usemolecular genetics tools in developmental sci-ence. For example, genes coding for dopamine


functions may influence attraction to novelty,and genes controlling for serotonin functionsmay influence distress tendencies (Auerbach,Faroy, Ebstein, Kahana, & Levine, 2001;Lakatos et al., 2003; Rothbart & Bates, 2006).However, even as we look forward to suchmarkers, it is important to keep in mind thatgenes do not operate like a fixed program.Multiple genes interact with one another andwith other factors in the biological environ-ment, and thus there still may be some instabil-ity in the functioning of genes that code fortemperament-relevant processes. Likewise, al-though there has been enormous progress indescribing the anatomy and functioning of thenervous system, we do not have measures ofthe neural patterns that constitute tempera-ment. And even when such markers are de-fined, we would expect that individuals wouldshow changes in neural patterns, reflectingboth experience and maturation. For example,a child might develop new emotional apprais-als of situations by forming new associationsvia synaptic changes in structures such as theamygdala and hippocampus (Lewis, 2005;Posner & Rothbart, 2000). Finally, to considerobservable behavior (the top level of a temper-ament construct), all measures of behaviorpatterns—whether based on self-report, care-giver report, or direct observation—have limi-tations in reliability and validity (Rothbart &Bates, 1998). However, even if ideal measureswere possible, there would still be changes inthe behavior patterns due to changes in individ-uals’ life circumstances and the correspondingchanges in the incentive conditions for activat-ing motivations and emotions. Nevertheless,despite these considerations, temperamentmeasures do show sufficient levels of stability,even across many years of development, to re-tain their usefulness as markers of traits (Caspi,1998; Kagan, 1998; Rothbart & Bates, 2006).

Emotion versus Emotion Regulation

A particularly interesting discussion has con-cerned the distinction between “emotion” and“emotion regulation.” This conceptual distinc-tion echoes the distinction between reactiveand regulatory aspects of temperament. Cam-pos, Frankel, and Camras (2004) argue thatemotion and emotion regulation are function-ally inseparable. There is no “pure” emotion,in the sense of an event that exists in a measur-able way separate from regulation. Emotion

and emotion regulation are concurrent pro-cesses, and an emotional response to a givenincentive stimulus depends on preexisting regu-latory processes, such as cortical inhibition orthe way the stimulus is interpreted. Neverthe-less, although Campos and his colleagues dosee emotion and emotion regulation as inextri-cably part of the same whole and reject a for-mulation in which an emotion occurs and issubsequently regulated, they also admit to aconceptual, analytical distinction between thetwo intimately related concepts. In a somewhatdifferent view, Cole, Martin, and Dennis(2004) argue that emotion regulation can beconceptually defined as separate from the emo-tions that are regulated and empirically definedby systematically measuring changes in emo-tions, even though they also agree that “emo-tions are inherently regulatory” (p. 319). Coleet al. review several areas of research thatcan be interpreted as showing the emotion–regulation distinction, such as studies in whichhigh levels of regulatory responses—both at thelevel of directly observable behavior and at thelevel of psychophysiological indices such asvagal tone—are associated with low levels ofnegative affect. In summary, although it seemsthat emotion and emotion regulation are con-ceptually part of the same set of processes,there are also theoretical reasons to distinguishthe two separate components in at least someconcrete ways.

Related to the issue of how to distinguish be-tween emotion and emotion regulation is theissue of how to distinguish between reactiveand effortful or executive forms of regulation(Nigg, 2000). It is likely that regulatory behav-ior differences often involve components ofboth reactive and effortful control (Eisenberg,Spinrad, & Smith, 2004). Nevertheless, thetheoretical distinction can be useful. Gray(1993) has described two neural systems. Thebehavioral inhibition system (BIS) processesstimuli that may signify danger or loss, such asnovelty, and produces inhibition as a primarymotor output. Fear and anxiety are the emo-tional products of activity in these circuits. Thebehavioral activation system (BAS) processesstimuli that may signify potential rewards, andproduces approach actions as a primary motoroutput. Interest, excitement, and joy are theemotional products. Suppose that it is the be-ginning of the year in a preschool class, and theteacher has asked the children to stand in aline. Most of the children will comply, but there

30. Temperament and Emotion 487

may be different temperamental–emotionalbases for their compliance. Some will be veryinterested in the room, and would otherwise beexploring; however, the teacher’s commandcues their effortful self-regulation, and they fo-cus their attention away from the exciting toysand onto the markers on the floor that theteacher has asked them to stand next to.Some—discomfited by the general novelty ofthe situation, and thus reactively or fearfullyinhibited—are predisposed to stand still any-way. Some will comply with difficulty becausethey are so positively aroused by the situation,and others will comply with ease because theyare relatively uninterested. Among the few whohave to be helped to comply, there will be atleast one who is joyfully pursuing fun (e.g.,showing a new friend how he or she can jump,or trying to look more closely at the collectionof toy cars). And there may even be one whohas reacted with anger to the restriction. Em-pirical measurement of effortful control sepa-rate from reactive control is difficult. Casualobservation may not reveal the components ofthe behavior, but detailed and extensive obser-vation of the children may help in distinguish-ing the underlying systems. The distinctionsmay not matter for some purposes, but may ul-timately be important in understanding devel-opment of individual differences in social adap-tation (Eisenberg, Spinrad, et al., 2004; Fox &Calkins, 2003; Stifter, Spinrad, & Braungart-Rieker, 1999). In addition to having measuresof behavior patterns across different kinds ofincentive conditions and at different stages ofdevelopment, it will also be useful to have mea-sures that can more clearly refer to the effortfulforms of emotion regulation.

Emotion Regulation and Social Adaptation

Promising measures of core emotion regulationprocesses are being developed. Some of thesemeasures have implications for the study of in-dividual differences in social development, andwe mention several examples here. Lewis andSteiben (2004) argue that emotion regulationcan be measured via event-related potentials inthe prefrontal cortex. They review past find-ings on two event-related potentials that arethought to reflect the activity of the anteriorcingulate cortex. The inhibitory N2 response,occurring between 200 and 350 millisecondsfollowing a challenging stimulus, and the error-related negativity (ERN) response, occurringabout 80 milliseconds after an error such as

pushing the wrong button, both reflect aspectsof cognitive control. Individuals low on N2and ERN responses have been found to be rela-tively high on impulsivity and aggression.Lewis and Steiben (2004) also report an in-triguing preliminary result from a study of chil-dren in a go/no-go task. In a go/no-go task, aprepotent “go” response (typically pushing abutton upon seeing a particular symbol on thecomputer screen) is established via rewards,but also, less frequently, a stimulus calls for anondominant nonresponse, which requireseffortful control. The Lewis and Steiben ver-sion of the task included a block of trials inwhich the experimenter increased the rate oferrors via faster presentation of stimuli,thereby creating a failure experience. Lewisand Steiben found that children who were highon internalizing problems showed bigger N2and ERN responses than other children, and inthe block of failure trials showed a significantincrease in size of the ERN, which the otherchildren did not. This suggests that tendenciesto be highly prepared to detect and regulate er-rors at speeds faster than conscious processingare associated with symptoms of anxiety anddepression.

A second possible index of emotional self-regulation ability is the time it takes to stopa response that has been triggered by astimulus—an executive ability that is mediatedby the orbito-frontal region of the prefrontalcortex. The stop signal reaction time measureof response inhibition has been previously usedfor understanding adaptive difficulties such asattention-deficit/hyperactivity disorder (Logan,Schachar, & Tannock, 1997). In a smallexperience-sampling study with preadolescentchildren, Hoeksma, Osterlaan, and Schipper(2004) found that variability of anger over theperiod of self-observation was positively corre-lated with stop signal reaction time. The longerit took a child to inhibit a motor response in alaboratory test, the more variability the childshowed in naturally occurring levels of anger.Basing their argument on the assumptions thatfeelings of anger reflect emotional process andare cues for cognitive regulation (because un-regulated anger is maladaptive), Hoeksma etal. interpreted their obtained correlation asshowing that the capacity for quick responseinhibition leads to more effective anger regula-tion.

Attentional measures, such as sustained at-tention and tendency to direct attention awayfrom frustrating objects, are associated with


tendencies to modulate negative emotions andultimately with social adjustment (Hill &Braungart-Rieker, 2002; Stifter & Spinrad,2002). An interesting index of an attention-related regulatory process is the suppression ofheart rate variability associated with breathing(respiratory sinus arrhythmia). For example, inone study, infants who showed lower levels ofsuppression of respiratory sinus arrhythmia inan attention task (they were given a block toexamine) showed higher levels of frustration intasks where they were prevented from reachinggoals (Calkins, Dedmon, Gill, Lomax, & John-son, 2002). It appears that the angry infantswere less able to focus their attention onnonfrustrating aspects of a situation. Similarpatterns of association between ability to focusattention and lower tendencies to express nega-tive affect have been found in a number ofother studies (Eisenberg, Smith, Sadovsky, &Spinrad, 2004).

Finally, more traditional measures of earlychildhood effortful control are found in numer-ous response inhibition measures, based onlaboratory tasks in which a child is asked to dosuch things as walk a line slowly or delay be-fore consuming a piece of candy, and in ques-tionnaires asking caregivers about children’sability to focus and shift attention and to in-hibit responses where appropriate (Rothbart &Bates, 2006). Such measures are successful inforecasting children’s adjustment in later child-hood, even after early childhood adjustment iscontrolled for. Children who are more success-ful on response inhibition tasks are less aggres-sive, are less noncompliant, and show morepositive social qualities in childhood and ado-lescence than children who are less successfulat inhibiting behaviors on demand (Denhamet al., 2003; Eisenberg, Smith, et al., 2004;Eisenberg et al., 2005; Olson, Sameroff, Kerr,Lopez, & Wellman, 2005; Olson, Schilling, &Bates, 1999; Rothbart & Bates, 2006). Thereare also a few studies that show that childrenwho are very high on regulation in laboratorytasks show correspondingly high levels of inter-nalizing problems (e.g., Murray & Kochanska,2002). However, it seems likely that such pat-terns reflect overcontrol that is based in reac-tive, anxious inhibition rather than effortful,executive control (Rothbart & Bates, 2006).

In summary, the distinctions between reac-tive and regulatory aspects of temperamentclosely parallel similar distinctions betweenemotion and emotion regulation. It is likelythat in nature, emotion and its regulation and

their corresponding personality traits consti-tute parts of the same complex processes.Nevertheless, at this time, it does seem theoreti-cally helpful to keep them as separate concepts.It is possible that the distinction will help de-scribe important details in how children attaintheir unique social adaptations. At this point, afair amount of research has suggested that atemperamental tendency toward wariness fore-casts later, conceptually analogous internaliz-ing problems. The same tendency also some-times serves to retard the growth of conductproblems, which can be seen as a function ofsensitivity to the punishment that externalizingbehavior often brings (Bates, Pettit, & Dodge,1995; Keiley, Lofthouse, Bates, Dodge, &Pettit, 2003). Other kinds of reactive regula-tion, such as the quick detection of errors(Lewis & Steiben, 2004), could also play a rolein the development of anxiety and preventionof aggression. Research has also suggested thattemperamental tendencies toward unmanage-ability or impulsivity, including low effortfulcontrol, forecast conceptually analogous ten-dencies toward externalizing behavior. Andtemperamental negative emotionality—whichcan reflect irritability due to sensitivity to un-pleasant stimuli, sensitivity to frustration of ef-forts to approach or control, and weak abilityto divert attention from such stimuli—turnsout to predict both internalizing and ex-ternalizing problems. This differential patternof linkage between temperament and adjust-ment provides a basis for appreciating the com-plex but theoretically important recent trend inresearch to explore how children’s adjustmentoutcomes develop as an interactive function ofboth temperament and environmental charac-teristics.

ADJUSTMENT AS THE INTERACTIVEPRODUCT OF TEMPERAMENTAND ENVIRONMENT

Developmental theory has for a long time em-phasized complex relations between child andenvironment as determinants of adjustmentdifferences between children. A given environ-ment may have different implications for chil-dren with different temperament or other char-acteristics, such as intelligence; or a given childcharacteristic may have different implications,depending on the environment within which itis expressed (Collins, Maccoby, Steinberg,Hetherington, & Bornstein, 2000; Wachs,


2000). As recently as 10 years ago, therewas barely a smattering of studies showingtemperament–environment interaction effects.Theory expected interaction effects, but fewhad been found. As we have discussed in previ-ous reviews of this topic (e.g., Rothbart &Bates, 2006), this might have been a functionof the methodological difficulties of findingsuch interaction effects (McClelland & Judd,1993). However, temperament–environmentinteraction effects have been reported at suchan accelerating rate in recent years that we nowentertain the possibility that the previous pau-city of findings was due less to methodologicalobstacles than to the research Zeitgeist in thatera.

The emerging research, despite being quiteencouraging, still has methodological short-comings. For example, not all interaction effectfindings adequately control for main effects,and not all findings fully evaluate the effects ofdifferent distributions of predictor and out-come variables at the contrasting levels of themoderator variable (Bates & McFadyen-Ketchum, 2000). Nor are we totally sanguineabout the adequacy of measurement of any ofthe key constructs. The temperament measuresin most of the relevant studies are behavioral,which in effect means that it is often difficult todistinguish between emotional reactivity andemotion regulation traits. For example, highlevels of frustration reactivity could reflect tem-perament substrates of strong tendencies to ap-proach, weak ability to turn attention awayfrom a frustrating stimulus, or both. However,the overall pattern of results allows us to seethe glass as half full. We believe that the tem-perament constructs do have some meaning. Asmentioned earlier, there is a differential patternof linkages between temperament and adjust-ment dimensions, with a temperament lowin regulation predicting high levels ofexternalizing problems and a temperamenthigh in novelty distress predicting high levelsof internalizing problems. Furthermore, thereare emerging patterns of temperament–environment interaction effects in the develop-ment of children’s adjustment. There is a sub-stantial pattern of temperament–environmentinteraction results for three categories of tem-perament: novelty distress, negative emotional-ity, and impulsivity–effortful control (Bates &Pettit, 2007).

In the novelty distress category, two patternsstand out, each with some but not yet adequatereplication. First, for young children with high

novelty distress, gentle maternal control ap-pears to promote the development of prosocialbehavior better than harsh control does,whereas for low-fear children the harshness ofmaternal control does not seem related to de-velopment of prosocial behavior (Kochanska,1997). Second, children who are reactive inways that predispose them to novelty distressare less likely to develop full-blown behavioralinhibition and anxiety problems if they arehandled in more directive and challenging ways(e.g., curtly or even in an irritated way) than ifthey are handled in less directive and challeng-ing ways (Arcus, 2001; Bates & Pettit, 2007).

In the negative emotionality category, thereis a pattern in which negative parentingappears to amplify the links between childnegative emotionality and the development ofexternalizing behavior problems—or the con-verse, in which child negative emotionality am-plifies the links between negative parenting andbehavior problems (Bates & Pettit, 2007). Andthere is also the pattern referred to in connec-tion with novelty distress: Young infants whoare negatively emotional in a challenging labo-ratory situation are at risk for the behavioralinhibition pattern, but they appear to be lesslikely to develop this pattern if treated in morerather than less challenging ways (Arcus, 2001;Bates & Pettit, 2007). Negative emotionality,especially when assessed in younger infants,may signify tendencies toward frustration andanger, or it may signify tendencies toward fear-ful distress. These are not well differentiateduntil later in infancy. In any event, the patternsare at this point suggestive, rather than fully es-tablished.

In the impulsivity–effortful control category,there are suggestions (again needing further rep-lication), for two kinds of patterns. First, chil-dren who are low in self-regulation—that is,children high in unmanageability or impulsivity,and low in effortful control of attention—aremore likely to develop externalizing problems iftheir parents are negative and lacking in warmthand responsiveness, whereas parenting does notmatter so much for children who are high in self-regulation (Bates & Pettit, 2007). The secondpattern is that children’s early unmanageabil-ity tendencies are more likely to becomeexternalizing behavior problems when mothersare less rather than more controlling of the chil-dren’s misbehavior (Bates & Pettit, 2007; Bates,Pettit, Dodge, & Ridge, 1998). Because our em-phasis here is on advances in understanding self-regulation, the remainder of this chapter ex-


pands on the role of children’s self-regulatorycharacteristics in interaction with characteris-tics of their environments.

Developmental Process and Self-Regulation

Consistent with the preceding paragraph’ssummary of some relevant studies, we assumethat children who are well regulated—that is,able to control their attention and to inhibitdominant responses for the sake of performingsubdominant responses—will be less likelythan poorly self-regulated children to behave indisruptive ways when attracted to potential re-wards in a situation where approach or domi-nance behaviors are not appropriate. We alsoassume that they will be less likely to behave inneurotic ways when frightened by potentialpunishers in a situation where pursuit of re-wards is more appropriate than withdrawal.These assumptions are essentially definitional,however. They do not explain how behavioraladjustment develops. What happens as a resultof a child’s successful adaptation in a particularsituation? What happens in instances where theadaptation is unsuccessful? How are these ex-periences in some way carried forward? Inmodern, systems-based theories, it is assumedthat the child learns something or fails to learnsomething, but that in addition, the child’stransactions in some way transform the envi-ronment. Over many iterations, tendenciesmay become increasingly solidified.

Initial probabilities of well-regulated chil-dren’s self-regulative responses should not haveto be markedly higher than those of poorly reg-ulated children for them to contribute eventu-ally to a substantial difference in adaptive out-comes. Over the multiple steps of a real-lifeevent, a well-regulated child will experience asuccess that may condition, in some small way,dispositions to expect good results. And then inthe future, the child will be a little more likelyto approach a similar task with positive emo-tion or a minimum of negative emotion. If weassume that caregivers will feel similarly andwill be more likely to provide more advancedopportunities, the child will take further stepstoward competence. Repeated over many suchoccasions, small initial differences between awell-regulated child and a less well-regulatedyoung child may grow to socially significantdifferences in the developmental trajectories ofthese children.

A recent model of such processes builds onthe widely used and well-supported coercive

family process model of Patterson and his col-leagues (e.g., Patterson, Reid, & Dishion,1992): In this new model, Granic and Patterson(2006) provide a nonlinear dynamic-systemsaccount of how children and their social sys-tems may develop toward clinically significantantisocial behavior problems. For example, toconsider one kind of situation, a boy at risk forconduct problems and his parent experiencecognitive–emotional–behavioral feedback pro-cesses that frequently resolve toward an “at-tractor state” of mutual hostility. Essentially,the probability that a hostile act by one partywill be followed by a hostile act by the otherparty increases in a nonlinear way, and thisstate, among all the other possible ones, comesto be especially likely. Alternatively, sometimesthe same system resolves toward a child’s beingsullen or defiant and the parent’s being permis-sive or even submissive toward the child; andsometimes it resolves toward fun and mutuallysatisfying states. As the transactions of a givenevent proceed, small factors (e.g., the actors’initial moods and motives, their skills for self-regulation, or the presence of other actors) mayprobabilistically influence the dyad’s move-ments through the various mutual state possi-bilities, shaping the likelihood that one attrac-tor state, such as repeated hostile exchanges,will dominate. After many such experiences,such a coercive state may be achieved in just anexchange or two, and from there hostility mayescalate even to violence. Over time, these ex-periences shape development. In a cascadingseries, initial constraints on the phase transi-tions and attractors of the system, such as childtemperament or parent stress, produce furtherconstraints, such as rigid patterns of coerciveparent–child interaction and rigid patterns ofchild approach to various opportunities (e.g.,learning and peer interaction opportunities inpreschool); these patterns lead to key conflictsand, fairly abruptly, to aggressive behavior.These difficulties are constraints that in turncreate further constraints, such as peer rejec-tion, teacher negativity, and so on.

The emerging findings on how temperamentand environment moderate one another’s influ-ence in shaping social development are interest-ing from a number of perspectives, but mostparticularly in the context of the Granic andPatterson (2006) formulation, because they canprovide plausible leads in the search for controlparameters that explain how a developmentalsystem may turn one way versus another. It isimportant to learn about moderators of risk


conditions, such as child temperament—bothfor the sake of further development of theoriessuch as Granic and Patterson’s, and for theguidance of practical experiments, such as im-proved treatments for child behavior problems.Self-regulation differences may be particularlycritical parameters in the developmental sys-tem, and ones whose implications influence theimplications of other parameters and viceversa. We therefore now turn to a discussion ofa handful of recent studies of regulation × envi-ronment interactions in the development ofchildren’s adjustment.

Recent Examples of Self-Regulation× Environment Interactions

Decades of research have firmly established anassociation of generally moderate strength be-tween negative parenting (or the absence ofpositive parenting) and child behavior prob-lems (Rothbaum & Weisz, 1994). This negativeparenting risk factor also interacts, it now ap-pears, with child temperament in the produc-tion of behavior problems. As mentioned, onepattern in the recent literature is that childrenwith tendencies to be dysregulated are morelikely to have concurrent or future conductproblems when they experience hostile or oth-erwise dysregulated parenting. One good ex-ample of this is the study by Rubin, Burgess,Dwyer, and Hastings (2003), which predictedlater adjustment from a composite of labora-tory and mother-rated measures of child self-regulation of frustration and inhibition of pro-hibited behavior at age 2, and from a compos-ite of lab observations and self-report measuresof intrusive and hostile parenting, also at age 2.The study showed a stronger prediction fromchildren’s poor self-regulation of frustrationand poor inhibition of prohibited behavior atage 2 to mother-rated externalizing problemsat age 4 when their mothers were highly intru-sive and hostile at age 2 than when they werelow on negative parenting. In other words, tod-dlers’ emotional and behavioral self-regulationdeficits were less likely to become conductproblems if their mothers were measured andpositive in their dealings with the children, pre-sumably not building a mutual hostility “at-tractor.” Thus, we can speculate, a more be-nign system allowed the children to learn moreadaptive, less rigid patterns of conflict resolu-tion and to develop more positive expectanciesabout social interactions in general. A mother

characteristic moderated the developmentalimplications of a toddler characteristic. In aconceptually related way, but from anotherperspective, Rubin et al. (2003) also found astronger relation between children’s aggressive-ness observed in play with a peer at age 2 andmother-rated externalizing problems at age 4when the children were low on the self-regulation index than when they were high onit. In other words, children’s self-regulationcharacteristics moderated the developmentalimplications of their early aggressiveness. Per-haps the children’s ability to regulate frustra-tion and restrain actions in the face of a prohi-bition allowed adaptive responses to thenegative feedback that often follows aggressiveacts, and thus the children learned socially ac-ceptable ways to express assertive (or low-levelaggressive) motives.

Aside from negative parenting, another well-established risk factor for child behavior prob-lems is a family’s socioeconomic status (SES)(e.g., Dodge, Pettit, & Bates, 1994). Of course,SES sums a wide variety of differences in theexperiences of children, including multiplequalities of parenting (Dodge et al., 1994);neighborhood qualities (Beyers, Bates, Pettit,& Dodge, 2003); and major stressors anddaily hassles, including poverty (McLeod &Shanahan, 1996). In a recent, large-scale,cross-sectional study, Veenstra, Lindenberg,Oldehinkel, DeWinter, and Ormel (2006)found that the predictable relation of SES topreadolescents’ antisocial behavior (a compos-ite of parent and youth report) held only foryouths who were described as high on frustra-tion and low on effortful control (also mea-sured by both parent and youth report). Thesefindings suggest that child emotional andbehavioral self-regulation may be important indetermining whether negative experiences inthe family and neighborhood cascade towardstable patterns of externalizing behavior.

Similar processes may influence the develop-ment of internalizing problems. A study byCrockenberg and Leerkes (2006) illustrates thisin a particularly important way. Not onlydoes this study consider a temperament–environment interaction; it also considers aninteraction between two temperament charac-teristics, one concerning reactivity and theother concerning regulation. Recall that dis-tress to novelty has been theoretically and em-pirically linked to later anxiety behaviors, andthat looking away from something that might


cause distress is viewed as an early marker ofself-regulation ability (Rothbart & Bates,2006). Infants who were described by theirmothers as high in distress to novelty at 6months, but who also were observed to lookaway from a novel toy presented to them (andlook at something else), showed less mother-reported anxiety at 30 months than those whodid not show this attention-regulating ability.Importantly, however, this effect was foundwhen mothers were observed at 6 months to below in sensitivity and engagement (in situationsaway from the novel toy presentation), and notwhen they were high in sensitivity and engage-ment. It appears that sensitive, engaged moth-ers may have been able to provide a com-pensating form of regulation to preventexperiences of novelty distress from growinginto more generalized anxious behavior. Thisillustrates a theoretical point often madein developmental psychopathology (e.g.,Cummings, Davies, & Campbell, 2000): De-velopmental systems typically allow more thanone way to reach a socially important outcome.If one mechanism is lacking, such as early-developing self-regulation ability, anothermechanism, such as appropriately supportiveparenting, can help maintain adaptive develop-ment.

Two recent studies from our own lab providefurther examples of ways in which self-regulatory tendencies may interact with envi-ronment and other characteristics of youths toshape the development of externalizing behav-iors. We (Goodnight, Bates, Newman, Dodge,& Pettit, 2006) found that adolescents’ perfor-mance in a laboratory game (the Card PlayingTask; Newman, Patterson, & Kosson, 1987)moderated the often-observed association be-tween their friends’ deviance and their ownexternalizing behaviors. Youths who continuedto pursue rewards in the Card Playing Task, de-spite experiencing an increasingly higher ratioof punishment to reward, showed a strongerassociation between their friends’ and theirown externalizing behaviors 2 years later. Thiswas so even when their own, earlier levels ofexternalizing behavior were controlled for.Friends’ deviance may reflect the degree towhich youths’ peer groups provide positive re-inforcement for antisocial behavior (Dishion,Spracklen, Andrews, & Patterson, 1996). It isalso important to note that the findings indi-cate that high-impulsivity, reward-dominantyouths who had nondeviant peers showed re-

duced development of conduct problems rela-tive to other youths, perhaps because their peergroups provided reinforcement for normativeand achievement-related behavior rather thanantisocial behavior. In other words, youths’strong response to rewarding events or failureto modulate their pursuit of rewards in the faceof punishment was found to predict higher-than-typical growth of conduct problems whentheir friends engaged in antisocial activities,but lower-than-typical growth of conductproblems when their friends were not antiso-cial. Thus the same temperamental, regulatorycharacteristic could be either a risk factor forantisocial or an aid to prosocial development,depending on qualities of the peers.

In a second study (Fite, Goodnight, Bates,Dodge, & Pettit, in press), we have found thata temperament-like tendency toward impulsiveaction can moderate the extent to which one’sstyle of social information processing, as mea-sured by responses to hypothetical vignettes(Crick & Dodge, 1994), predicts actual behav-ior. We think of social information processingas potentially influenced by self-regulationabilities. For example, highly regulated chil-dren may tell themselves to stop and think be-fore responding to some ambiguous provo-cation (e.g., someone’s cutting in line), thusallowing them to fully process the relevant de-tails of the situation and to generate multipleresponse options. However, impulsivity is notthe same as social information processing, sowe were able to ask whether the two kinds oftendency might interact in predicting adaptivebehavior. We reasoned that highly impulsiveyouths might be more likely to act withoutfully processing real-world social situations,making them more likely to display their de-fault or most characteristic patterns of socialinformation processing, rather than to let theactual situations guide their behavior. Wefound for highly impulsive youths, as rated bytheir teachers at ages 11–13, that the youths’endorsement of aggressive responses in a socialinformation-processing assessment at age 13was significantly predictive of their later,parent-reported aggressive behavior at ages14–17, even when earlier levels of aggressionwere controlled for; however, the hypotheticalendorsement of aggression did not predictbehavior for low-impulsivity youths. Thus ado-lescents’ impulsivity increased the likelihoodthat a cognitive processing default would betranslated into behavior in real life.


CONCLUSION

Temperament concepts are a subset of emotionconcepts that aid in thinking about individualdifferences in emotion and how they develop.This chapter has summarized ways in whichtemperament accounts for individual differ-ences in social development. Self-regulation hasbeen of increasing interest in the area of tem-perament, in the same way as in the broaderarea of emotion, so the chapter has emphasizedregulatory traits and how they pertain to socialdevelopment. The conceptual distinction be-tween emotion and emotion regulation appearssomewhat useful, even though it also seemsclear that in terms of psychological processthey are closely intertwined elements of a largerphenomenon. The same may be true of the con-ceptual distinction between temperamental re-activity and self-regulation. Nevertheless, therehave been some interesting empirical efforts tohighlight regulatory processes against a back-drop of other emotion processes—for instance,careful descriptions of changes in emotion fol-lowing regulatory responses such as attentionand the activity of brain centers such as the an-terior cingulate gyrus. Although there havebeen some direct links between regulatorytraits and behavioral adjustment outcomes, thelinks are of moderate size. Theorists in devel-opmental psychopathology have long arguedthat individual child characteristics should af-fect outcomes to a large extent in interactionwith characteristics of the environment. How-ever, until very recently, there was almost noempirical evidence of such processes. In theyears since the second edition of this handbookappeared, there has been an explosion of re-search that shows temperament–environmentinteraction effects upon social developmentoutcomes. As exciting as this emerging researchhas been, further research is needed. Method-ological advances will be important. For exam-ple, one methodological issue is how to mea-sure the different emotional components of atemperament trait separately. It is often hard toknow to what extent a measure of a youth’simpulsive, reward-seeking behavior reflects astrong positive emotion/approach system, aweak fear/inhibition system, or a poor execu-tive self-control system. One promising avenueis cognitive modeling of individuals’ trial-by-trial decision making in a task such as a go/no-go task, as in another study from our lab(Yechiam et al., 2006). This study derived pa-rameters for sensitivity to reward versus pun-

ishment, attention to feedback from the mostrecent trials versus earlier trials, and rate of im-provement in performance on the task.

As the work on temperament–environmentinteraction effects proceeds, we will also wantto begin learning how the moderator effects oc-cur. That is, what mediates the moderator ef-fects? Perhaps they are mediated by observabletransactions, as in models like that of Granicand Patterson (2006). For example, the par-ent’s irritability exacerbates the child’s irritabil-ity, and the product of mutual irritability cre-ates obstacles to learning crucial self-regulatoryskills. Or perhaps they are mediated by geneticfactors. For example, if parental responsivenessmoderates the relation between child frustra-tion and later behavior problems, perhaps theparental responsiveness may mark a trait alsoinherited by the child that comes online later indevelopment and that allows for improvedchild self-regulation. Obviously, these are ad-vanced questions requiring new methodolo-gies, such as detailed observations over criticalperiods of development; perhaps these shouldinclude observations of not only parent–childinteractions, but also brain functioning, alongwith precisely targeted measurements of con-figurations of genetic polymorphisms whoseemotional functions will have become well de-scribed. However, technologies are advancingrapidly, so we can envision efforts to addresssuch questions. In the meanwhile, we arepleased and encouraged by the advances in de-velopmental research, including sharper con-cepts of emotion and regulation traits, and thebeginnings of richer empirical models of howthese traits act in developmental systems toshape children’s social adaptations.

REFERENCES

Arcus, D. (2001). Inhibited and uninhibited children:Biology in the social context. In T. D. Wachs & G. A.Kohnstamm (Ed.), Temperament in context (pp. 43–60). Mahwah, NJ: Erlbaum.

Auerbach, J. G., Faroy, M., Ebstein, R., Kahana, M., &Levine, J. (2001). The association of the dopamineD4 receptor gene (DRD4) and the serotonin trans-porter promoter gene (5-HTTLPR) with tempera-ment in 12-month-old infants. Journal of Child Psy-chology and Psychiatry, 42(6) 772–783.

Bates, J. E. (1989). Applications of temperament con-cepts. In G. A. Kohnstamm, J. E. Bates, & M. K.Rothbart (Eds.), Temperament in childhood(pp. 321–355). Chichester, UK: Wiley.

Bates, J. E. (2000). Temperament as an emotion con-struct: Theoretical and practical issues. In M. Lewis


& J. M. Haviland-Jones (Eds.), Handbook of emo-tions (2nd ed., pp. 382–396). New York: GuilfordPress.

Bates, J. E., & McFadyen-Ketchum, S. (2000). Tempera-ment and parent–child relations as interacting factorsin children’s behavioral adjustment. In V. J. Molfese& D. L. Molfese (Eds.), Temperament and personal-ity across the life span (pp. 141–176). Mahwah, NJ:Erlbaum.

Bates, J. E., & Pettit, G. S. (2007). Temperament, par-enting, and socialization. In J. Grusec & P. Hastings(Eds.), Handbook of socialization (pp. 153–177).New York: Guilford Press.

Bates, J. E., Pettit, G. S., & Dodge, K. A. (1995). Familyand child factors in stability and change in children’saggressiveness in elementary school. In J. McCord(Ed.), Coercion and punishment in long-term per-spectives (pp. 124–137). New York: Cambridge Uni-versity Press.

Bates, J. E., Pettit, G. S., Dodge, K. A., & Ridge, B.(1998). Interaction of temperamental resistance tocontrol and restrictive parenting in the developmentof externalizing behavior. Developmental Psychol-ogy, 34(5), 982–995.

Beyers, J. M., Bates, J. E., Pettit, G. S., & Dodge, K. A.(2003). Neighborhood structure, parenting pro-cesses, and the development of youths’ externalizingbehaviors: A multilevel analysis. American Journal ofCommunity Psychology, 31(1–2), 35–53.

Calkins, S. D., Dedmon, S. E., Gill, K. L., Lomax, L. E.,& Johnson, L. M. (2002). Frustration in infancy: Im-plications for emotion regulation, physiological pro-cesses, and temperament. Infancy, 3(2), 175–197.

Campos, J. J., Frankel, C. B., & Camras, L. (2004). Onthe nature of emotion regulation. Child Develop-ment, 75(2), 377–394.

Caspi, A. (1998). Personality development across thelife course. In W. Damon (Series Ed.) & N. Eisenberg(Vol. Ed.), Handbook of child psychology: Vol. 3. So-cial, emotional, and personality development (5thed., pp. 311–388.). New York: Wiley.

Chess, S., & Thomas, A. (1984). Origins and evolutionof behavior disorders: From infancy to early adultlife. New York: Brunner/Mazel.

Cole, P. M., Martin, S. E., & Dennis, T. A. (2004). Emo-tion regulation as a scientific construct: Methodolog-ical challenges and directions for child developmentresearch. Child Development, 75(2), 317–333.

Collins, W. A., Maccoby, E. E., Steinberg, L., Hethering-ton, E. M., & Bornstein, M. H. (2000). Contempo-rary research on parenting: The case for nature andnurture. American Psychologist, 55(2), 218–232.

Crick, N. R., & Dodge, K. A. (1994). A review and re-formulation of social information-processing mecha-nisms in children’s social adjustment. PsychologicalBulletin, 115, 74–101.

Crockenberg, S. C., & Leerkes, E. M. (2006). Infant andmaternal behavior moderate reactivity to novelty topredict anxious behavior at 2.5 years. Developmentand Psychopathology, 18, 17–34.

Cummings, E. M., Davies, P. T., & Campbell, S. B.

(2000). Developmental psychopathology and familyprocess: Theory, research, and clinical implications.New York: Guilford Press.

Denham, S. A., Blair, K. A., DeMulder, E., Levitas, J.,Sawyer, K., Auerbach-Major, S., et al. (2003). Pre-school emotional competence: Pathway to socialcompetence? Child Development, 74(1), 238–256.

Dishion, T. J., Spracklen, K. M., Andrews, D. W., &Patterson, G. R. (1996). Deviancy training in maleadolescent friendships. Behavior Therapy, 27, 373–390.

Dodge, K. A., Pettit, G. S., & Bates, J. E. (1994). Social-ization mediators of the relation between socioeco-nomic status and child conduct problems. Child De-velopment, 65, 649–665.

Eisenberg, N., Sadovsky, A., Spinrad, T., Fabes, R. A.,Losoya, S. H., Valiente, C., et al. (2005). The rela-tions of problem behavior status to children’s nega-tive emotionality, effortful control, and impulsivity:Concurrent relations and prediction of change. De-velopmental Psychology, 41(1), 193–211.

Eisenberg, N., Smith, C. L., Sadovsky, A., & Spinrad, T.L. (2004). Effortful control: Relations with emotionregulation, adjustment, and socialization in child-hood. In R. F. Baumeister & K. D. Vohs (Eds.),Handbook of self-regulation: Research, theory, andapplications (pp. 259–282). New York: GuilfordPress.

Eisenberg, N., Spinrad, T. L., & Smith, C. L. (2004)Emotion-related regulation: Its conceptualization, re-lations to social functioning, and socialization. In P.Philippot, & R. S. Feldman (Eds.), The regulation ofemotion (pp. 277–306). Mahwah, NJ: Erlbaum.

Fite, J. E., Goodnight, J. A., Bates, J. E., Dodge, K. A.,& Pettit, G. S. (in press). Adolescent aggression andsocial cognition in context: Impulsivity as a modera-tor of predictions from social information process-ing. Aggressive Behavior.

Fox, N. A., & Calkins, S. D. (2003). The developmentof self-control of emotion: Instrinsic and extrinsic in-fluences. Motivation and Emotion, 27(1), 7–26.

Goodnight, J. A., Bates, J. E., Newman, J. P., Dodge, K.A., & Pettit, G. S. (2006). The interactive influencesof friend deviance and reward dominance on the de-velopment of externalizing behavior during middleadolescence. Journal of Abnormal Child Psychology,34, 573–583.

Granic, I., & Patterson, G. R. (2006). Toward a com-prehensive model of antisocial development: A dy-namic systems approach. Psychological Review,113(1), 101–131.

Gray, J. A. (1993). The neuropsychology of tempera-ment. In J. Strelau & A. Angleitner (Eds.), Explora-tions in temperament: International perspectives ontheory and measurement (pp. 105–128). New York:Plenum Press.

Hill, A. L., & Braungart-Rieker, J. M. (2002). Four-month attentional regulation and its prediction ofthree-year compliance. Infancy, 3(2), 261–273.

Hoeksma, J. B., Oosterlaan, J., & Schipper, E. M.(2004). Emotion regulation and the dynamics of feel-


ings: A conceptual and methodological framework.Child Development, 75(2), 354–360.

Kagan, J. (1998). Biology and the child. In W. Damon(Series Ed.) & N. Eisenberg (Vol. Ed.), Handbook ofchild psychology: Vol. 3. Social, emotional, and per-sonality development (5th ed., pp. 177–235). NewYork: Wiley.

Keiley, M. K., Lofthouse, N., Bates, J. E., Dodge, K. A.,& Pettit, G. S. (2003). Differential risks of covaryingand pure components in mother and teacher reportsof externalizing and internalizing behavior acrossages 5 to 14. Journal of Abnormal Child Psychology,31, 267–283.

Lakatos, K., Nemoda, Z., Birkas, E., Ronai, Z., Kovacs,E., Ney, K., et al. (2003). Association of D4 dopa-mine receptor gene and serotonin transporter pro-moter polymorphisms with infants’ response to nov-elty. Molecular Psychiatry, 8, 90–97.

Kochanska, G. (1997). Multiple pathways to consciencefor children with different temperaments: Fromtoddlerhood to age 5. Developmental Psychology,33(2), 228–240.

Lewis, M., & Haviland-Jones, J. M. (Eds.). Handbookof emotions (2nd ed.). New York: Guilford Press.

Lewis, M. D. (2005). Bridging emotion theory and neu-robiology through dynamic systems modeling.Behavioral and Brain Sciences, 28, 169–245.

Lewis, M. D., & Stieben, J. (2004). Emotion regulationin the brain: Conceptual issues and directions for de-velopmental research. Child Development, 75(2),371–376.

Logan, G. D., Schachar, R. J., & Tannock, R. (1997).Impulsivity and inhibitory control. Psychological Sci-ence, 8, 60–64.

McClelland, G. H., & Judd, C. M. (1993). Statisticaldifficulties of detecting interactions and moderatoreffects. Psychological Bulletin, 114, 376–390.

McLeod, J. D., & Shanahan, M. J. (1996). Trajectoriesof poverty and children’s mental health. Journal ofHealth and Social Behavior, 37, 207–220.

Murray, K. T., & Kochanska, G. (2002). Effortful con-trol: Factor structure and relation to externalizingand internalizing behaviors. Journal of AbnormalChild Psychology, 30(5), 503–514.

Newman, J. P., Patterson, C. M., & Kosson, D. S.(1987). Response perseveration in psychopaths.Journal of Abnormal Psychology, 96, 145–148.

Nigg, J. T. (2000). On inhibition/disinhibition in devel-opmental psychopathology: Views from cognitiveand personality psychology and a working inhibitiontaxonomy. Psychological Bulletin, 126(2), 220–246.

Olson, S. L., Sameroff, A. J., Kerr, D. C. R., Lopez, N.L., & Wellman, H. M. (2005). Developmental foun-dations of externalizing problems in young children:The role of effortful control. Development and Psy-chopathology, 17, 23–45.

Olson, S. L., Schilling, E. M., & Bates, J. E. (1999).Measurement of impulsivity: Construct coherence,longitudinal stability, and relationship with

externalizing problems in middle childhood and ado-lescence. Journal of Abnormal Child Psychology,27(2), 151–165.

Patterson, G. R., Reid, J. B., & Dishion, T. J. (1992).Antisocial boys. Eugene, OR: Castalia.

Posner, M. I., & Rothbart, M. K. (2000). Developingmechanisms of self-regulation. Development andPsychopathology, 12, 427–441.

Reuda, M. R., Posner, M. I., & Rothbart, M. K. (2004).Attentional control and self-regulation. In R. F.Baumeister & K. D. Vohs (Eds.), Handbook of self-regulation: Research, theory, and applications(pp. 283–300). New York: Guilford Press.

Rothbart, M. K., & Bates, J. E. (1998). Temperament.In W. Damon (Series Ed.) & N. Eisenberg (Vol. Ed.),Handbook of child psychology: Vol. 3. Social, emo-tional, and personality development (5th ed.,pp. 105–176). New York: Wiley.

Rothbart, M. K., & Bates, J. E. (2006). Temperament.In W. Damon (Series Eds.) & N. Eisenberg (Vol. Ed.),Handbook of child psychology: Vol. 3. Social, emo-tional, and personality development (6th ed., pp. 99–166). Hoboken, NJ: Wiley.

Rothbart, M. K., Derryberry, D., & Posner, M. I.(1994). A psychobiological approach to the develop-ment of temperament. In J. E. Bates & T. D. Wachs(Eds.), Temperament: Individual differences at the in-terface of biology and behavior (pp. 83–116). Wash-ington, DC: American Psychological Association.

Rothbaum, F., & Weisz, J. R. (1994). Parental caregiv-ing and child externalizing behavior in nonclinicalsamples: A meta-analysis. Psychological Bulletin,116, 55–74.

Rubin, K. H., Burgess, K. B., Dwyer, K. M., & Hastings,P. D. (2003). Predicting preschoolers’ externalizingbehaviors from toddler temperament, conflict, andmaternal negativity. Developmental Psychology,39(1), 164–176.

Stifter, C. A., & Spinrad, T. L. (2002). The effect of ex-cessive crying on the development of emotion regula-tion. Infancy, 3(2), 133–152.

Stifter, C. A., Spinrad, T. L., & Braungart-Rieker, J. M.(1999). Toward a developmental model of child com-pliance: The role of emotion regulation in infancy.Child Development, 70(1), 21–32.

Veenstra, R., Lindenberg, S., Oldehinkel, A. J.,DeWinter, A. F., & Ormel, J. (2006). Temperament,environment, and antisocial behavior in a populationsample of preadolescent boys and girls. InternationalJournal of Behavioral Development, 30(5), 422–432.

Wachs, T. D. (2000). Necessary but not sufficient: Therespective roles of single and multiple influences onindividual development. Washington, DC: AmericanPsychological Association.

Yechiam, E., Goodnight, J. A., Bates, J. E., Busemeyer, J.R., Dodge, K. A., Pettit, G. S., et al. (2006). A formalcognitive model of the go/no-go discrimination task:Evaluation and implications. Psychological Assess-ment, 18(3), 239–249.


C H A P T E R 3 1

Emotion Regulation

JAMES J. GROSS

Have you ever gotten so angry that you’vedone something really spiteful? Or felt so muchlove for your child that you’ve bored someoneto tears by recounting your child’s exploits? Orbeen so sad that life has temporarily lost itsmeaning? If your answer to any of these ques-tions is “yes”—or if anyone you care aboutwould answer “yes”—then this chapter is foryou.

The focus of this chapter is emotion regula-tion, which refers to how we try to influencewhich emotions we have, when we have them,and how we experience and express these emo-tions (Gross, 1998b). Although the topic ofemotion regulation is a relatively late additionto the field of emotion, a concern with emotionregulation is anything but new. Emotion regu-lation has been a focus in the study of psycho-logical defenses (Freud, 1926/1959), stress andcoping (Lazarus, 1966), attachment (Bowlby,1969), and self-regulation (Mischel, Shoda, &Rodriguez, 1989).

What is new are the theoretical and empiri-cal advances that have been made in recentyears, thanks to a dramatic increase in atten-

tion to this topic (Gross, 2007). Until the early1990s, there were just a few citations a yearcontaining the phrase “emotion regulation.”For example, in 1990, PsycINFO listed 4 cita-tions containing the phrase “emotion regula-tion.” Since this time, there has been an as-tonishing increase in citations: In 2005, forinstance, the PsycINFO citation count was671. Although citation counts are an imperfectmetric, the 150-fold-plus increase in citationsover this 15-year period clearly reflects thegrowing popularity of this topic.

Popularity is a wonderful thing, but despitethis increased attention, there remains an un-fortunate degree of confusion about what emo-tion regulation is (and isn’t), and what effects(if any) emotion regulation has on importantoutcomes. My goal in this chapter is to providea conceptual map and readable introductionuseful to anyone who is interested in emotionregulation. In the first section, I provide an ori-entation to emotion and emotion regulation,and sketch a process model of emotion regula-tion that my colleagues and I have found usefulin our work. In the second section, I describe

497

the five major families of processes that popu-late our conception of emotion regulation. Inthe third section, I consider three excitinggrowth points for the field: (1) an emerging un-derstanding of the way particular beliefs en-courage or discourage specific forms of emo-tion regulation; (2) an increasing appreciationof automatic (as opposed to effortful) forms ofemotion regulation; and (3) a growing sense ofthe implications the field of emotion regulationhas for the diagnosis and treatment of psycho-pathology.

EMOTION ANDEMOTION REGULATION

Contemporary emotion theories emphasize theways emotions facilitate adaptation by ready-ing behavioral responses (Tooby & Cosmides,Chapter 8, this volume), enhancing mem-ory for important events (Phelps, 2006), andguiding interpersonal interactions (Keltner &Kring, 1998). However, emotions are by nomeans always helpful. They can hurt us as wellas help us (Parrott, 1993). They do so whenthey are of the wrong type, when they come atthe wrong time, or when they occur at thewrong intensity level. At times such as these,we may be highly motivated to try to regulateour emotions. To understand how emotions areregulated (or become dysregulated), wefirst must consider the target of emotionregulation—namely, emotion itself.

What Is Emotion?

As with many of the terms we use in psychol-ogy, “emotion” was lifted from everyday dis-course. For this reason, it has fuzzy boundariesrather than classical edges, and it refers to anastonishing array of happenings—from themild to the intense, the brief to the extended,the simple to the complex, and the private tothe public. Irritation when a shoelace breakscounts. So do amusement at a joke, anger atpolitical oppression, surprise at a friend’s new“look,” grief at the death of a parent, and guiltover a moral lapse. This incredible diversity hasled many theorists to despair of ever deriving atidy classical definition of emotion—one thatlists the necessary and sufficient conditions forsomething to qualify as a “real” emotion. In-stead, they have begun to think of emotion in

prototype terms, and have identified three keyfeatures.

One commonly described feature has to dowith what gives rise to emotions. Emotions arethought to arise when an individual attends toa situation and understands it as being relevantto his or her current goals (Lazarus, 1991a). It’simportant to appreciate that these goals may beenduring and central to the person’s self-concept (wanting to be trustworthy) or tran-sient and peripheral (wanting the last slice ofcake). They may be conscious and complicated(wanting to survive the rigors of graduateschool) or unconscious and simple (wanting todistance oneself from a snake). They may bewidely shared and understood in a given cul-ture (wanting to be a good son/daughter) or id-iosyncratic and somewhat mysterious to others(wanting to travel on a UFO). Whatever thegoal, and whatever meaning a situation has forthe individual, it is this meaning that gives riseto emotion. As either the goal or meaningchange over time (due to a change in the per-son, the situation, or the meaning the situationholds for an individual), the emotion will alsochange.

A second commonly described feature ofemotion has to do with its constituent ele-ments. Emotions are generally conceptualizedas multifaceted, embodied phenomena that in-volve loosely coupled changes in the domainsof subjective experience, behavior, and periph-eral physiology (Mauss, Levenson, McCarter,Wilhelm, & Gross, 2005). The experiential as-pect of emotion—or what it feels like when wehave an emotion—is so tightly bound up withwhat we mean by emotion that in everydayspeech, the terms “emotion” and “feeling” areoften used interchangeably. However, surpris-ingly little is known about the psychologicaland biological underpinnings of emotion expe-rience (Barrett, Mesquita, Ochsner, & Gross,2007), and there are many contexts in whichthere are dissociations between emotion expe-rience and other aspects of an emotion (e.g.,Bonanno, Keltner, Holen, & Horowitz, 1995).In addition to giving rise to subjective feelings,emotions also often make us more likely to dosomething (e.g., approach others, say some-thing mean, cry) than we otherwise would havebeen (Frijda, 1986). These impulses to act incertain ways (and not to act in others) are asso-ciated with autonomic and neuroendocrinechanges that both anticipate the associated


behavioral responses (thereby providing meta-bolic support for the action) and follow it, of-ten as a consequence of the motor activity asso-ciated with the emotional response. As themany chapters of this volume attest, there re-mains considerable debate about which ofthese aspects of emotion should be prioritized,and how these aspects of emotion co-occurduring emotion. For our purposes, it is enoughto note that emotions often involve changes ineach of these response domains.

A third commonly described feature of emo-tion has to do with its malleability. Emotionspossess an imperative quality, in that they caninterrupt what we are doing and force them-selves upon our awareness (Frijda, 1988).However, emotions must compete with otherresponses occasioned by the situations we arein, and therefore do not automatically trumpother possible responses to the situation. Themalleability of emotion has been emphasizedsince William James (1884), who viewed emo-tions as response tendencies that may be modu-lated in a large number of ways. It is this thirdaspect of emotion that is most crucial for ananalysis of emotion regulation, because it isthis feature that gives rise to the possibility forregulation.

The “Modal Model” of Emotion

Because these three core features of emotionare emphasized in many different theories ofemotion, I have found it useful to refer to themas constituting a consensual model or “modalmodel” of emotion (Barrett, Ochsner, & Gross,2007; Gross, 1998a). According to this model,emotion arises in the context of a person–situation transaction that compels attention,has a particular meaning to an individual, andgives rise to a coordinated yet malleablemultisystem response to the ongoing person–situation transaction. I believe that it is no acci-dent that this heuristic “modal model” under-lies lay intuitions about emotion, and also rep-resents crucial points of convergence amongresearchers and theoreticians concerned withemotion.

In Figure 31.1, I present in schematic formthe situation–attention–appraisal–response se-quence specified by the modal model of emo-tion (with the organismal “black box” inter-posed between situation and response). Thissequence begins with a psychologically relevant

situation, which is often external and hencephysically specifiable. This situation is at-tended to in various ways, giving rise to ap-praisals that constitute the individual’s assess-ment of—among other things—the situation’sfamiliarity, valence, and value relevance(Ellsworth & Scherer, 2003). As noted above,the emotional responses generated by apprais-als are thought to involve changes in experien-tial, behavioral, and physiological responsesystems. It is important to keep in mind thatthese responses often change the situation thatgave rise to the response in the first place. Forexample, when someone appears embarrassedafter committing a faux pas, others see this em-barrassment, and are then more likely to for-give the social lapse (Keltner, 1995). One wayto depict this recursive aspect of emotion is byan arrow that shows the response feeding backto (and modifying) the situation. The key ideahere is that emotions can and often do changethe environment, thereby altering the probabil-ity of subsequent instances of emotion.

What Is Emotion Regulation?

With this schematic conception of emotion inview, we are ready to turn to emotion regula-tion. It will come as no great surprise that like“emotion,” the concept of “emotion regula-tion” is a slippery one. This is partly becausethe concept inherits all of the complexities thatare inherent in the term “emotion.” But theconstruct is confusing in a second way, in thatit isn’t clear whether it refers to how emotionsregulate something else, such as blood pres-sure, memory, or parent–child interactions(regulation by emotions) or to how emotionsare themselves regulated (regulation of emo-tions). Both usages have currency, but the prob-lem with the first usage (regulation by emo-tions) is that one of the functions of emotion isthe coordination of diverse response systems

31. Emotion Regulation 499

FIGURE 31.1. The “modal model” of emotion.From Gross and Thompson (2007). Copyright2007 by The Guilford Press. Reprinted by permis-sion.

(Levenson, 1999). Thus emotion regulation inthis first sense is redundant with emotion, inthat all instances of emotion would constituteemotion regulation. I therefore find the secondusage more sensible (regulation of emotions),in which emotion regulation refers to the heter-ogeneous set of processes by which emotionsare themselves regulated.

Another point of confusion is whether emo-tion regulation refers to intrinsic processes(Amy regulates her own emotions: regulationin self), to extrinsic processes (Amy regulatesbaby Bob’s emotions: regulation in other), or toboth. In general, researchers in the adult lit-erature typically focus on intrinsic processes,whereas researchers in the developmental liter-ature focus more on extrinsic processes (Gross& Thompson, 2007). In my view, it makessense to include both forms of regulation, andto use the qualifiers “intrinsic” and “extrinsic”whenever clarification is needed, such as whenAmy helps Bob to regulate his anger (extrinsicemotion regulation) in order to be able to calmdown herself (intrinsic emotion regulation).

Putting aside for a moment the complexityassociated with intrinsic versus extrinsic emo-tion regulation, what are people trying to ac-complish when they regulate emotions? Whenwe think of emotion regulation, many of the in-stances that leap to mind—at least in a Westerncultural context—involve turning down (de-creasing) the experiential and/or behavioral as-pects of negative emotions such as anger, fear,and sadness (Gross, Richards, & John, 2006).This is not to say that positive emotions aren’tregulated; they certainly are, as when we try tolook less happy than we are about winning ahard-fought tennis game, or when we try to de-crease feelings of attraction that (for whateverreason) we find objectionable. It’s also impor-tant to note that emotion regulation needn’tinvolve down-regulation. It can also involvemaintaining or increasing emotion, as when weshare good news with others, thereby prolong-ing its effects (Langston, 1994), or even—in thecontext of negative emotion—when bill collec-tors try to increase their anger to help collectdelinquent accounts (Sutton, 1991).

Many of these emotion regulation goals arereadily understood in hedonistic terms: Peopleare motivated to avoid pain and seek pleasure.But if emotion regulation involves increasing/initiating and decreasing/stopping negative orpositive emotions, it is not clear how we canexplain the “odd” cells (increasing negative

emotion and decreasing positive emotion) onthe basis of short-term hedonic considerations.Tamir (2005) has argued that hedonic consid-erations can sometimes be trumped by otherconsiderations, such as whether a given emo-tion will help a person achieve his or her imme-diate objectives. One example is when indi-viduals high (vs. low) in neuroticism try toincrease their levels of negative emotion in or-der to maximize their performance on a de-manding cognitive task. This finding suggeststhat emotions are regulated with a view to bothhow they feel and what they help us to do.

EMOTION REGULATION STRATEGIES

If “emotion regulation” refers to the processesby which we influence which emotions wehave, when we have them, and how we experi-ence and express these emotions, we face anembarrassment of riches. Many processes areinvolved in decreasing, maintaining, or increas-ing one or more aspects of emotion. Indeed,relevant processes range from changing one’sjob to calling one’s mother to keeping a stiff up-per lip. How should we conceptualize the po-tentially overwhelming number of processesinvolved in regulating our own or others’ emo-tions?

My approach has been to undertake a con-ceptual analysis of the processes underlying di-verse emotion regulatory acts. Using the modalmodel of emotion shown in Figure 31.1 as astarting point, I have argued that emotion regu-latory acts may be seen as having their primaryimpact at different points in the emotion gener-ative process (Gross, 2001). In particular, Ihave suggested that the modal model specifies asequence of processes involved in emotion gen-eration, each of which is a potential target forregulation. In Figure 31.2, I have redrawn themodal model, highlighting five points at whichindividuals can regulate their emotions. Thesefive points represent five families of emotionregulation processes: situation selection, sit-uation modification, attentional deployment,cognitive change, and response modulation.

Two complementary points should be madeabout this process model of emotion regula-tion. First, although this model makes a five-way distinction among emotion regulation pro-cesses, there are higher-order commonalities.For example, for some purposes, the first fouremotion regulation families may be considered


“antecedent-focused,” in that they occur be-fore appraisals give rise to full-blown emo-tional responses. These may be contrasted with“response-focused” emotion regulation, whichoccurs after the responses are generated (Gross& Munoz, 1995; Gross & Thompson, 2007).The second point about these distinctions isthat what someone does in everyday life to reg-ulate emotions—such as going fishing with abuddy to cool down after a big fight witha spouse—often involves multiple regulatoryprocesses. Nonetheless, I believe that this pro-cess model provides a conceptual frameworkuseful for understanding the causes, conse-quences, and mechanisms underlying variousforms of emotion regulation.

In the following sections, I selectively reviewresearch relevant to each of the five families ofemotion regulation processes. My focus is onemotion regulation processes in adults (forconsiderations of developmental issues, seeCharles & Carstensen, 2007; Eisenberg &Morris, 2002; Gross & Thompson, 2007;Thompson, 1994; for considerations of indi-vidual differences in emotion regulation, seeJohn & Gross, 2004, 2007).

Situation Selection

The first type of emotion regulation we’ll con-sider is situation selection, which I’ve placed atthe leftmost point in Figure 31.2 because it af-fects the situation to which a person is exposed,and thus shapes the emotion trajectory fromthe earliest possible point. Situation selectioninvolves taking actions to make it more likelythat we’ll be in a situation we expect will giverise to the emotions we’d like to have (or lesslikely that we’ll be in a situation that will giverise to emotions we’d prefer not to have).

Of course, many of our decisions aboutwhich appointments to keep, where to go tolunch, whom to spend time with, and what todo after work have implications for how we’lllater feel, but these decisions are not alwaysshaped by our estimates of which emotionsthese situations will engender. “Situation selec-tion” refers to the subset of these choices thatare taken with a view, at least in part, to the fu-ture consequences of our actions for our emo-tional responses. Often we are aware of the tra-jectory our emotions are likely to take during agiven period of time (e.g., a day) if we don’ttake steps to influence our emotions. Thisawareness may motivate us to take steps to al-ter the default emotional trajectory via situa-tion selection. Thus we may try hard to avoidsituations we know will bring us face to facewith an ex-spouse or ex-lover, or we may ac-tively seek out situations that will provide uswith contact with friends when we need achance to vent and/or share positive emotions.

These examples make situation selectionsound like a rather simple calculus. It is not. In-deed, there is a growing appreciation of justhow difficult it is either to remember how weused to feel, or to predict how we will feel.When we look backward in time, there is a pro-found gap between what might be called the“experiencing self” and the “rememberingself” (Kahneman, 2000). In one of the morecolorful illustrations of this gap, Redelmeierand Kahneman (1996) studied patients whowere undergoing colonoscopies (a decidedlyunpleasant procedure in which a probe is in-serted into one’s innermost recesses) and pro-vided pain ratings at regular intervals through-out the procedure. They found that even whenthe procedure was longer (and thus gave rise tomore “units” of experienced pain), participants


FIGURE 31.2. A process model of emotion regulation that highlights five families of emotion regulationstrategies. From Gross and Thompson (2007). Copyright 2007 by The Guilford Press. Reprinted by per-mission.

later expressed a preference for the longer pro-cedure when it ended with lower levels of pain.

This “duration neglect” is also evident in af-fective forecasting, when we look forwardrather than backward in time. In one illustra-tion of this phenomenon, Gilbert, Pinel, Wil-son, Blumberg, and Wheatley (1998) askedparticipants how they would feel if they brokeup with a partner or were denied academic ten-ure. They found that participants did a goodjob of figuring out what they would feel.Where participants miscalculated was in figur-ing out how long they would feel that way. Inparticular, participants dramatically overesti-mated how long their negative responses wouldlast. These backward- and forward-looking bi-ases hint at the complexity and the fallibility ofthe judgments involved in using situation selec-tion.

Even if we had perfect information regardingpast and future emotional responses to situa-tions, there would remain the thorny issue ofhow to appropriately weigh short-term benefitsof emotion regulation versus longer-term costs.For example, take Harold, a mild-manneredperson who hates angry confrontations. If he isinterested in maximizing short-term psycholog-ical comfort, it seems obvious that he shouldavoid situations in which angry confrontationswill occur. But is this the best long-term strat-egy? What if his avoidance of conflict is givingimplicit permission to others to bully him, andto behave in generally unreasonable and toxicways to him? For his long-term (rather thanshort-term) happiness, it might be better toseek out an opportunity for a confrontation—even an angry one—if this meant that his worksituation were changed in ways that made it abetter place for him. Because of the complexityof these tradeoffs, situation selection often re-quires the perspective of others, whether par-ents, friends, or therapists.

Situation Modification

Potentially upsetting situations—such as mak-ing a social gaffe or seeing the family televisiongo dead just before a favorite show is to start—do not inevitably lead to negative emotional re-sponses. After all, one can always make a jokeof one’s social lapse or play a family game in-stead of watching television. Such efforts tomodify the situation directly so as to alter itsemotional impact constitute a second form of

emotion regulation, shown next in line in Fig-ure 31.2. In the stress and coping tradition, thistype of emotion regulation is referred to as“problem-focused coping” (Lazarus & Folk-man, 1984) or “primary control” (Rothbaum,Weisz, & Snyder, 1982).

What forms may situation modificationtake? When a romantic interest comes over fordinner, it may take the form of mood lighting,music, and the strategic excision of unflatteringmemorabilia. Situation modification may alsotake the form (with children) of laying outgames in a way that will ensure a smooth playdate, helping with scaffolding that will allowthem to solve a difficult problem, (partially)absenting oneself when their friends come over,or reinforcing one’s limits via clear emotion ex-pressions. The last case is particularly interest-ing theoretically, because it’s a case in whichemotion expressions themselves can be a po-tent extrinsic form of emotion regulation. Thisis because emotional expressions have impor-tant social consequences: If one’s partner sud-denly looks sad, this can shift the trajectory ofan angry interaction as one pauses to expressconcern, apologize, or offer support.

Given the vagueness of the term “situation,”it is sometimes difficult to draw a bright linebetween situation selection and situation modi-fication. This is because efforts to modify a sit-uation may effectively call a new situation intobeing. Also, although I have previously empha-sized that situations can be external or internal,situation modification (as I mean it here) has todo with modifying external physical environ-ments. I consider efforts at modifying “inter-nal” environments (i.e., cognitions) in the sec-tion on cognitive change below.

Attentional Deployment

The first two forms of emotion regulation—situation selection and situation modification—both help to shape the situation to which an in-dividual will be exposed. However, it is alsopossible to regulate emotions without actuallychanging the environment. Situations havemany aspects, and attentional deployment re-fers to influencing emotional responding by re-directing attention within a given situation.Attentional deployment is thus an internal ver-sion of situation selection, in that attention isused to select which of many possible “internalsituations” are active for an individual at any


point in time. In Figure 31.2, attentional de-ployment comes after situation modification inthe emotion trajectory.

In one form or another, attentional deploy-ment is used from infancy through adulthood,particularly when it is not possible to change ormodify one’s situation (Rothbart, Ziaie, &O’Boyle, 1992). It is used, for example, by chil-dren who are waiting for delayed rewards,and spontaneous use of attentional deploymentpowerfully affects success during delay of grati-fication (Mischel et al., 1989). Attentionaldeployment may also include physical with-drawal of attention (e.g., covering the eyes orears), internal redirection of attention (e.g.,through distraction), and responding to exter-nal redirection of attention (e.g., a parent’s re-direction of a hungry child by telling thechild an interesting story). Two of the best-researched forms of attentional deployment aredistraction and rumination.

“Distraction” involves a shift in attention ei-ther away from emotional aspects of the situa-tion or away from the situation altogether, suchas when an infant shifts its gaze during anoverly intense emotional interaction (Stifter &Moyer, 1991). Distraction may also involve achange in internal focus, such as when an indi-vidual invokes thoughts or memories that areinconsistent with the undesirable emotionalstate. Distraction has often been studied in thecontext of pain, where it leads to increased ac-tivation of brain regions associated with cogni-tive control (such as lateral prefrontal corticalregions) and diminished activation of brain re-gions associated with pain generation (such asthe insula) (Ochsner & Gross, 2005).

“Rumination” refers to a perseverative focuson thoughts and feelings associated with anemotion-eliciting event. Rumination on sad orangry events increases the duration and in-tensity of negative emotion (Bushman, 2002;Morrow & Nolen-Hoeksema, 1990; Ray, Wil-helm, & Gross, in press) and is associated withgreater levels of depressive symptoms (Nolen-Hoeksema, Morrow, & Fredrickson, 1993;Spasojevic & Alloy, 2001). Unlike distraction,rumination involves a sustained focus onemotion-eliciting stimuli. Another point of dif-ference is that while distraction can take theform of attention directed outwards, to com-peting stimuli, or inwards, to thoughts, rumi-nation typically involves an inflexibility ininner-directed attention.

Cognitive Change

Even after a potentially emotion-eliciting situa-tion has arisen and been attended to, emotiondoes not necessarily follow. This is because anemotion further requires that the individual im-bue the situation with a certain kind of mean-ing. As noted above, emotion theorists havedelineated the different appraisals that arethought to lead to different emotions (Scherer,Schorr, & Johnstone, 2001). Cognitive change(shown fourth in line in Figure 31.2) refers tochanging one or more of these appraisals in away that alters the situation’s emotional signifi-cance, by changing how one thinks eitherabout the situation itself or about one’s capac-ity to manage the demands it poses.

One form of cognitive change that has re-ceived particular attention is reappraisal(Gross, 2002). “Reappraisal” involves chang-ing a situation’s meaning in such a way thatthere is a change in the person’s emotional re-sponse to that situation. For example, take asituation in which an acquaintance breezes byus in the hall and seems to ignore our smile andwave of greeting. For many, a natural responsein such a situation is to feel hurt or angry at thisperceived snub. In this case, cognitive changemay take the form of thinking about the ac-quaintance as distracted, or perhaps preoccu-pied with his or her own problems. Such aninterpretation of the situation—whether objec-tively correct or not—can profoundly affect thequality (which emotion) as well as the quantityor intensity (how much emotion) of the subse-quent emotional response.

To date, studies of reappraisal have focusedon quantitative changes in emotion, particu-larly decreases in negative emotion. These stud-ies have provided evidence that reappraisalleads to decreased negative emotion experienceand expressive behavior (Dandoy & Goldstein,1990; Gross, 1998a). Reappraisal has alsobeen shown to lead to decreased startle re-sponses (Dillon & LaBar, 2005; Jackson,Malmstadt, Larson, & Davidson, 2000), de-creased neuroendocrine responses (Abelson,Liberzon, Young, & Khan, 2005), and de-creased autonomic responses (Stemmler, 1997;but see Gross, 1998a). Importantly, compara-ble effects have been observed when researchparticipants spontaneously use reappraisal, ei-ther in a negative-emotion-eliciting situationin the lab (Egloff, Schmukle, Burns, &


Schwerdtfeger, 2006), or in everyday life(Gross & John, 2003). These findings suggestthat studies manipulating emotion regulationhave ecological validity: They provide insightsinto reappraisal as it naturally occurs in every-day life.

Consistent with these behavioral and physi-ological findings, reappraisal in the service ofemotion down-regulation is associated withdecreased activation in subcortical emotion-generative regions (such as the insula andamygdala), as well as increased activation indorsolateral and medial prefrontal regions as-sociated with cognitive control (Levesque etal., 2003; Ochsner, Bunge, Gross, & Gabrieli,2002; Ochsner et al., 2004). When reap-praisal is used in the service of emotion up-regulation, similar prefrontal regions are acti-vated, but in this context (as one might ex-pect), there are increases rather than de-creases in activation of emotion-generativestructures such as the amygdala (Ochsner &Gross, 2004; Schaefer et al., 2002). As ourprocess model would predict, activations inprefrontal regions associated with the top-down control of emotion seem to occur rela-tively early (in the first few seconds), whereasthe downstream consequences of decreasedexperience and behavior seem to last consid-erably longer (Goldin, McRae, Ramel, &Gross, in press).

If reappraisal occurs relatively early in theemotion-generative process, we might expectthat using reappraisal would not interfere withother ongoing cognitive processes. This is justwhat we’ve found in a series of studies thathave tested whether reappraisal impairs subse-quent memory for information presented dur-ing the reappraisal period (Richards & Gross,1999, 2000, 2006). Findings from these stud-ies, which have used slides or films to elicitemotion, and have used a variety of techniquesto probe incidental memory, show that reap-praisal does not compromise later memory formaterial presented while a participant was en-gaging in reappraisal (relative to not using re-appraisal). We have also found that when unac-quainted pairs of participants interactedsocially, there are no signs of social disruptionwhen one member of a dyad is covertly in-structed to engage in cognitive reappraisal dur-ing the interaction (Butler et al., 2003). Takentogether, these findings suggest that reappraisalintervenes early in the emotion-generative pro-cess, and alters the experiential, behavioral,

and physiological components of the emotionalresponse without incurring any appreciablecosts.

Response Modulation

Response modulation, the last of the emotionregulation families, is shown on the right sideof Figure 31.2. As this placement indicates, itoccurs late in the emotion-generative process,after response tendencies have been initiated.“Response modulation” refers to influencingphysiological, experiential, or behavioral re-sponses relatively directly. For example, exer-cise and relaxation may be used to decreasephysiological and experiential aspects of nega-tive emotions.

One of the best-researched forms of responsemodulation is “expressive suppression,” whichrefers to attempts to decrease ongoingemotion-expressive behavior (Gross, 2002).Examples of suppression abound, includingour efforts to hide the anger we feel toward aboss, the anxiety we feel during an interview, orthe amusement we feel at a coworker’s decid-edly politically incorrect joke. One reason sup-pression has attracted interest is that there aretwo opposing ideas about what happens whenemotions are suppressed (Gross & Levenson,1993). One idea is that behavioral expressionsof emotion constitute a channel for dischargingemotion. According to this “hydraulic” model,if emotions are denied expression, they willleak out elsewhere—for example, as increasedphysiological responses. A second idea, how-ever, leads to opposite conclusions about the ef-fects of suppression. According to this “facialfeedback” model, behavioral expressions ofemotion (such as facial expressions) actuallyserve to amplify the emotional response; thus ifthey are inhibited, the emotion itself will bemuted.

Empirical studies of expressive suppressionhave yielded findings that conform neatly toneither of these two models. On the one hand,participants who have been instructed to sup-press their emotions (during emotion-elicitingslides, films, or conversations) have shown in-creases in sympathetic activation of the cardio-vascular system, as indexed, for example, bymeasures that reflect blood pressure (Demareeet al., 2006; Gross, 1998a; Gross & Levenson,1993, 1997; Harris, 2001). On the other hand,when asked to suppress their emotions, partici-pants report feeling either comparable or de-


creased levels of emotion (with decreases oc-curring more commonly for positive emotion)(Goldin et al., in press; Gross, 1998a; Gross &Levenson, 1993, 1997; McCanne & Anderson,1987; Stepper & Strack, 1993; Strack, Martin,& Stepper, 1988).

Neurally, only one study to date has beenconducted on expressive suppression (Goldinet al., in press). In this study, participants wereasked to suppress their ongoing emotion-expressive behavior in the scanner during 15-second film segments that elicited intense levelsof disgust. Findings indicated that suppressionled to robust increases in the activation of dor-sal and medial prefrontal regions associatedwith cognitive control, as well as to increasedactivation in emotion-generative regions suchas the amygdala. Importantly, as the processmodel of emotion regulation would predict,these activations were evident late in the induc-tion period, suggesting that suppression wasassociated with ongoing cognitive activity asthe participants effortfully tried to manageeach emotional impulse as it arose throughoutthe course of each film.

If this conception of expressive suppressionis correct, we might expect that unlike reap-praisal, suppression should have clear cognitiveand social costs. In a series of studies, this isprecisely what we have found. In studies ofmemory, we have repeatedly found that sup-pression (compared to no emotion regulation)leads to worse memory for material presentedduring the suppression period (Richards &Gross, 1999, 2000). Indeed, the degree ofmemory impairment associated with suppres-sion was as large as when we instructed partici-pants to distract themselves as much as possi-ble during the presentation of information(Richards & Gross, 2006). In studies of socialinteractions in the laboratory, we have simi-larly found that suppression is associated withsignificant social costs: Partners of suppressorsreport less comfort and ease with their interac-tion partners (Butler et al., 2003). As with thereappraisal findings, we would note that thecosts that have been associated with instructedsuppression in the laboratory also seem to beevident when suppression is used spontane-ously in the laboratory (Egloff et al., 2006;Richards & Gross, 2006); during an importantlife transition, the transition to college(Srivastava, Tamir, McGonigal, John, & Gross,2008); and in everyday life (Gross & John,2003).

EMERGING DIRECTIONS IN THESTUDY OF EMOTION REGULATION

Now that we have reviewed the emotion-regulation processes shown in Figure 31.2, wecan step back and take stock of where this fieldis now and where it is going. Clearly, this is atime of unmatched excitement for the field ofemotion regulation. There has never beforebeen such a focused scientific effort to examineemotion regulation processes, nor has therebeen such a variety of perspectives brought tobear. Because emotion regulation lies at the in-tersection of the major subareas of psychology,it benefits from—and contributes to—developments in biological, cognitive, develop-mental, personality, social, and clinical areas(Gross, 1998b, 2007).

In the following sections, I consider threepromising new directions for research in emo-tion regulation, each of which seems likely tobroaden and extend the way we think aboutemotion regulation. The first concerns the cog-nitive antecedents of emotion regulation; thesecond concerns the boundaries of emotionregulation; the third concerns the implicationsemotion regulation research may have for un-derstanding psychopathology. Although by nomeans exhaustive, these three selections exem-plify the promise of emotion regulation re-search.

Beliefs and Emotion Regulation

One intriguing puzzle is why people use oneemotion regulation strategy rather than an-other. If some strategies are associated withgenerally beneficial consequences (such as re-appraisal), while others are associated withgenerally harmful consequences (such as sup-pression), why doesn’t everyone use reappraisaland not suppression? One possibility, ofcourse, is that people differ in their emotionregulation goals. Thus some people may wantto increase high-arousal positive emotions,whereas others may want to decrease theseemotions. This possibility is consistent withgrowing evidence of cultural differences inemotion regulation goals (Mesquita & Albert,2007). For example, in individualistic culturalcontexts, people generally seek out high-arousal positive emotional states; incollectivistic cultural contexts, people generallyseek out low-arousal positive emotional states(Tsai, Knutson, & Fung, 2006).


But what about when emotion regulationgoals are shared (e.g., when people wish to de-crease their sadness)? How can we explain whypeople differ even when they are pursuing thesame emotion regulation goal? One interestingpossibility is that people may differ in their be-liefs regarding emotion and emotion regula-tion, and these differences may in turn shapewhether people try to regulate their emotions,and (when they do so) which emotion regula-tion strategies they employ. This idea derivesfrom the “lay theories” perspective, a perspec-tive that draws inspiration from the social-cognitive approach to personality (Molden &Dweck, 2006). The lay theories perspectiveholds that people differ in the assumptions theymake about themselves and the social world(these constitute their “lay theories”), and itseeks to determine whether and how such laytheories influence important life outcomes.

One particular focus of the work on lay orimplicit theories has been the distinction be-tween “entity theories” (which hold that attrib-utes such as personality and intelligence arefixed and stable) and “incremental theories”(which hold that such attributes are dynamicand malleable) (Dweck, 1986, 1999; Dweck,Chiu, & Hong, 1995). Individuals who holdincremental beliefs make flexible, contextualinterpretations of events; when challenged,they make assertive attempts at self-regulation,increasing the chances of successful behavior.In contrast, individuals who hold entity beliefsview attributes as fixed and impossible to con-trol; when challenged, they make fewer at-tempts at self-regulation, leading to self-regulation failure.

Prior work on lay theories has focused on in-telligence, but we (Tamir, John, Srivastava, &Gross, 2007) wondered whether extending thiswork to the domain of emotion might help tounravel the mystery of why people differ sodramatically in their use of successful emotionregulation strategies. We (1) hypothesized thatpeople differ in whether they believe emotionsare generally malleable (incremental theorists)or fixed (entity theorists), and (2) suggestedthat incremental theorists should be morelikely than entity theorists to use antecedent-focused emotion regulation strategies such asreappraisal. To test these hypotheses, we de-vised a measure of implicit beliefs regardingemotion, and administered it to students facinga crucial life transition—namely, the transitionto college.

Findings revealed that participants did differin their beliefs about emotion, and that partici-pants with incremental as opposed to entityviews of emotion reported greater emotion reg-ulation self-efficacy and greater use of reap-praisal. By the end of freshman year, partici-pants with incremental views of emotionreported greater levels of positive emotions,lesser levels of negative emotions, higher levelsof well-being, and lower levels of depression.Incremental participants also had higher levelsof social adjustment and lower levels of loneli-ness.

These findings indicate that participants’naïve beliefs concerning their emotions—as ei-ther fixed or malleable—influenced how theyregulated their emotions, and how they fared inan important life transition. Although this find-ing clearly must be replicated in other samplesand in the context of other transitions, one im-portant emerging direction for research in thisarea is the study of the role played by beliefsabout emotion and emotion regulation in shap-ing emotion regulation choices and success.

Automatic versus EffortfulEmotion Regulation

Many of the examples of emotion regulationthat come to mind—and the majority of exam-ples offered so far in this chapter—involveeffortful and conscious attempts to down-regulate negative emotion. As we have dis-cussed, however, emotion regulation can occuranywhere in the 2 × 2 matrix formed by cross-ing negative and positive emotion (say, as col-umns) with up- and down-regulation (say, asrows). Each of these dimensions can be furtherfleshed out, too. Additional columns can beadded for those who prefer to think in discrete-emotion terms (e.g., pride, amusement, sad-ness, disgust), and additional rows for thosewanting to do fuller justice to the complexitiesof the temporal dynamics of emotion (e.g.,maintaining emotion). This sounds compli-cated enough—even before we recall the manyfamilies of regulation strategies that are used toachieve each of these types of change in emo-tion described by our 2 (or more) × 2 (or more)matrix—and we may be tempted to stop here.But one other dimension of variation has re-cently begun to be explored systematically, andthis concerns variation in whether a givenepisode of emotion regulation is relativelyeffortful and conscious or relatively automatic


and unconscious (Bargh & Williams, 2007;Mauss, Bunge, & Gross, 2007).

Just what does “automatic” mean in thiscontext? Contemporary dual-process modelscontrast “automatic” (also called “noncon-scious,” “implicit,” or “impulsive”) processeswith “deliberate” (also called “controlled,”“conscious,” “explicit,” or “reflective”) pro-cesses (e.g., Strack & Deutsch, 2004). Whereasdeliberate processes require attentional re-sources, are volitional and conscious, and aregoal-driven, automatic processes require nei-ther attention nor intention, occur outside ofawareness, and are stimulus-driven. Althoughoften framed as clear opposites, many research-ers think that these processes are located on acontinuum from conscious, effortful, and con-trolled regulation to unconscious, effortless,and automatic regulation (Shiffrin & Schnei-der, 1977).

The notion that relatively high-level self-regulatory processes such as emotion regula-tion can be performed automatically mayseem counterintuitive (Bargh, 2004). How-ever, research on automatic goal pursuit haschallenged the notion that “higher-level” pro-cesses can only take place in a deliberatefashion, and it appears that the full sequenceof goal pursuit—from goal setting to thecompletion of the goal—can proceed outsideof conscious awareness. In a series of studies,Bargh and colleagues have shown that goalscan indeed be activated and executed withoutthe intervention of conscious awareness. Forexample, they implicitly primed goals such asthe intention to cooperate with others or toperform well on a cognitive task, and foundthat subsequently participants behaved inagreement with these goals, without knowingwhy or even that they were acting this way(Bargh, Gollwitzer, Lee-Chai, Barndollar, &Trötschel, 2001).

To see whether emotion regulation couldalso operate automatically, we (Mauss, Cook,& Gross, 2007) manipulated automatic emo-tion regulation by priming emotion controlversus emotion expression with an adaptationof the Sentence Unscrambling Task (e.g.,Bargh, Chen, & Burrows, 1996; Bargh et al.,2001; Srull & Wyer, 1979). This task unobtru-sively exposed participants to words relating toemotion control or expression, thereby implic-itly activating (priming) related concepts andgoals. Participants were then instructed by an“unfriendly” and “arrogant” experimenter to

repeatedly perform a boring yet cognitivelystraining task.

As expected, most participants became an-gry during the task. Of particular interest,however, was the finding that participantsprimed with emotion control reported less an-ger than did participants primed with emotionexpression. These results have been corrobo-rated by an individual-difference study thatemployed an emotion regulation implicit asso-ciation test; this study showed that participantswith positive implicit associations with emo-tional control felt less angry when provoked,and exhibited an adaptive challenge response(rather than a maladaptive threat response),characterized by greater sympathetic activa-tion, greater cardiac output, and lower totalperipheral resistance (Mauss, Evers, Wilhelm,& Gross, 2006).

Although these initial studies are promising,it’s important to note that it is too early to con-clude that all forms of automatic emotion regu-lation are benign or helpful. The venerable clin-ical literature on repression (e.g., Freud, 1936)has long cautioned that certain forms of auto-matic emotion regulation—such as whensomeone struggles to keep anxiety out ofawareness—may have maladaptive conse-quences ranging from personality disturbancesto psychosomatic illnesses. One of the chal-lenges we face in understanding automaticemotion regulation is developing methods forassessing and manipulating different automaticemotion regulation processes. Difficult as thischallenge is, work in this area is badly neededto clarify the types and timing of automaticemotion regulation processes that are helpfulversus unhelpful.

Emotion Regulation and Psychopathology

Inappropriate emotional responses are impli-cated in a large number of forms of psycho-pathology (Gross & Levenson, 1997; Thoits,1985). Indeed, more than half of the Axis Iclinical disorders (such as the anxiety disordersand mood disorders), and all of the Axis II per-sonality disorders (such as borderline personal-ity disorder), involve problematic emotionalresponses (American Psychiatric Association,2000). What’s proven more difficult than onemight expect, however, has been moving frombroad statements such as these to specific em-pirically grounded insights concerning how dif-ferences in emotional reactivity and/or emotion


regulation contribute to different forms of psy-chopathology (Rottenberg & Gross, 2003;Rottenberg & Johnson, 2007), and how thera-peutic interventions might be used to correctdysregulated emotion (Moses & Barlow,2006).

Take major depressive disorder. This dis-order is a devastating psychiatric conditionwhose definition includes increased negativeaffect and anhedonia (diminished positive af-fect). From this definition, it might seem obvi-ous that depression leads to disrupted emotionregulation (Gross & Munoz, 1995). However,there are no fewer than three competing viewsof how depression disrupts emotional respond-ing, and without clarity about the nature of theproblematic emotions, it is very difficult todraw conclusions about the role of emotionregulation (Rottenberg, Gross, & Gotlib,2005). The first view is that depression in-volves diminished emotional reactivity to posi-tive situations. In support of this “positive at-tenuation” view, convincing evidence from avariety of induction contexts suggests that indi-viduals who are depressed respond with lesspositive emotion than individuals who are notdepressed. The second view is that depressioninvolves increased negative emotional reactiv-ity. Like the positive attenuation hypothesis,the “negative potentiation” view seems to fol-low directly from the very definition of depres-sion, as well as from major theories of depres-sion (e.g., Beck, Rush, Shaw, & Emery, 1979).However, the preponderance of empirical evi-dence actually suggests that individuals whoare depressed show lesser rather than greateremotional reactivity.

These findings suggest a third view—namely,the “emotion context insensitivity” view(Rottenberg et al., 2005). This view derives fromevolutionary accounts of depression as charac-terized by disengagement (Nesse, 2000), andsees emotional responses (whether negative orpositive) as involving energetic engagement withthe environment. In this view, depression leadsto pervasive disengagement, and hence to dimin-ished levels of both positive and negative emo-tional reactivity. Consistent with this third view,Rottenberg and colleagues (Rottenberg, Kasch,Gross, & Gotlib, 2002; Rottenberg & Johnson,2007) have presented studies showing that rela-tive to either formerly depressed or never-depressed participants, depressed individualsshowed less reactivity tohappyandsad stimuli.

Are the challenges associated with specifyingprecise emotion and emotion regulation defi-cits unique to depression? It appears that theyare not. Take social anxiety disorder (so-cial phobia), another common and debilitatingpsychiatric condition, which by definition in-cludes high levels of anxiety in social contexts.Given this definition, it seems obvious that so-cial anxiety involves heightened levels of ex-periential, behavioral, and physiological re-sponses in social contexts. To test thishypothesis, individuals who were either high orlow in social anxiety were asked to give aspeech on a difficult topic. Participants ratedhow anxious they felt at several points duringthe session. They also rated their physiologicalresponses (e.g., how much their hearts wereracing), and objective physiological measureswere taken throughout the study (Mauss, Wil-helm, & Gross, 2004).

As might be expected, compared to low-trait-anxiety participants, high-trait-anxietyparticipants said that they were feeling moreanxious, and that their bodies were respondingmuch more violently. Intriguingly, however,there were no differences in the observed physi-ological responses between the high- and low-anxiety participants. Participants in bothgroups showed expected increases in variousindicators of sympathetic nervous system re-sponding, but there was no difference betweenthe groups, either in the magnitude of their re-sponses to the initial speech or in their re-sponses to a second speech (Mauss, Wilhelm,& Gross, 2003). Although it is possible thatthese findings are specific to nonclinical sam-ples, the available evidence does not suggestthis is so. Like the findings from major depres-sive disorder, these findings from social anxietyhint at the complexities that lie ahead as we tryto discern the ways in which emotion and emo-tion regulation are disrupted in various formsof psychopathology.

SUMMARY

Emotions have been said to represent the “wis-dom of the ages” (Lazarus, 1991b, p. 820), andfunctionalist approaches to emotion haverightly emphasized the many adaptive benefitsof emotion. But even the wisest guides havetheir limits, and since the early 1990s there hasbeen a dramatic increase in research attention


to how emotions can be regulated so as to helppeople benefit from what is useful about them,but avoid what is not useful.

In this chapter, I have used the “modalmodel” of emotion to highlight the idea thatemotions arise in the context of person–situation transactions that compel attention,have particular meaning to an individual, andgive rise to coordinated yet flexible sets of ex-periential, behavioral, and physiological re-sponses to the ongoing person–situation trans-actions. Using the modal model as a jumping-off point, I have described a process model ofemotion regulation that my colleagues and Ihave found useful, and have argued that thismodel provides a valuable conceptual frame-work for organizing and directing researchon emotion regulation processes. Within thisframework, I have distinguished five families ofemotion regulation processes that have theirprimary impact at different points in theemotion-generative process. For each of thesefamilies of processes, I have selectively sampledrecent research findings. Taken together, thesefindings suggest that different emotion regula-tion processes have different consequences;what seems crucial, therefore, is using a strat-egy that matches one’s goals.

One reason why research in this area is socompelling is that we all come face to face withemotion regulation issues in our lives—whether in handling our own emotions orthose of family members, friends, or work as-sociates. Emotions matter, and when emotionsgo wrong, we want to do something about it.Another reason why emotion regulation re-search is attracting so much attention is that itis a “poster child” for two broad scientifictrends: multilevel/multispecialty collaboration,and the bidirectional interplay between basicresearch and clinical application. Of the manygrowth points in this field, I have identifiedthree as particularly exciting: (1) the role of be-liefs in shaping when and how we try to regu-late our emotions; (2) the largely unexploredrealm of automatic emotion regulation pro-cesses; and (3) the bridges that are beginning tobe built between basic research on emotion andemotion regulation on the one hand, and clini-cal science and practice on the other. Findingsfrom these research areas, and others, promiseto transform how we think about the intricatedance in which we at once regulate and are reg-ulated by our emotions.

ACKNOWLEDGMENTS

I would like to thank Lisa Feldman Barrett, IrisMauss, Kateri McRae, Jon Rottenberg, and MayaTamir for their helpful comments on this chapter.Work on this chapter was supported by National In-stitutes of Health Grants R01 MH66957, R01MH58147, and R01 MH76074.

REFERENCES

Abelson, J. L., Liberzon, I., Young, E. A., & Khan, S.(2005). Cognitive modulation of the endocrine stressresponse to a pharmacological challenge in normaland panic disorder subjects. Archives of General Psy-chiatry, 62, 668–675.

American Psychiatric Association. (2000). Diagnosticand statistical manual of mental disorders (4th ed.,text rev.). Washington, DC: Author.

Bargh, J. A. (2004). Being here now: Is consciousnessnecessary for human freedom? In J. Greenberg, S. L.Koole, & T. Pyszczynski (Eds.), Handbook of experi-mental existential psychology (pp. 385–397). NewYork: Guilford Press.

Bargh, J. A., Chen, M., & Burrows, L. (1996).Automaticity of social behavior: Direct effects of traitconstruct and stereotype activation on action. Jour-nal of Personality and Social Psychology, 71, 230–244.

Bargh, J. A., Gollwitzer, P. M., Lee-Chai, A., Barndollar,K., & Trötschel, R. (2001). The automated will:Nonconscious activation and pursuit of behavioralgoals. Journal of Personality and Social Psychology,81, 1014–1027.

Bargh, J. A., & Williams, L. E. (2007). The non-conscious regulation of emotion. In J. J. Gross (Ed.),Handbook of emotion regulation (pp. 429–445).New York: Guilford Press.

Barrett, L. F., Mesquita, B., Ochsner, K. N., & Gross, J.J. (2007). The experience of emotion. Annual Reviewof Psychology, 58, 373–403.

Barrett, L. F., Ochsner, K. N., & Gross, J. J. (2007). Onthe automaticity of emotion. In J. Bargh (Ed.), Socialpsychology and the unconscious: The automaticity ofhigher mental processes (pp. 173–217). New York:Psychology Press.

Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G.(1979). Cognitive therapy of depression. New York:Guilford Press.

Bonanno, G. A., Keltner, D., Holen, A., & Horowitz,M. J. (1995). When avoiding unpleasant emotionsmight not be such a bad thing: Verbal–autonomic re-sponse dissociation and midlife conjugal bereave-ment. Journal of Personality and Social Psychology,69, 975–989.


Bushman, B. J. (2002). Does venting anger feed or extin-


guish the flame?: Catharsis, rumination, distraction,anger and aggressive responding. Personality and So-cial Psychology Bulletin, 28, 724–731.

Butler, E. A., Egloff, B., Wilhelm, F. W., Smith, N. C.,Erickson, E. A., & Gross, J. J. (2003). The social con-sequences of expressive suppression. Emotion, 3, 48–67.

Charles, S. T., & Carstensen, L. L. (2007). Emotion reg-ulation and aging. In J. J. Gross (Ed.), Handbook ofemotion regulation (pp. 307–327). New York:Guilford Press.

Dandoy, A. C., & Goldstein, A. G. (1990). The use ofcognitive appraisal to reduce stress reactions: A repli-cation. Journal of Social Behavior and Personality, 5,275–285.

Demaree, H. A., Schmeichel, B. J., Robinson, J. L., Pu,J., Everhart, D. E., & Berntson, G. G. (2006). Up-and down-regulating facial disgust: Affective, vagal,sympathetic, and respiratory consequences. Biologi-cal Psychology, 71, 90–99.

Dillon, D. G., & LaBar, K. S. (2005). Startle modulationduring conscious emotion regulation is arousal-dependent. Behavioral Neuroscience, 119, 1118–1124.

Dweck, C. S. (1986). Motivational processes affectinglearning. American Psychologist, 41, 1040–1048.

Dweck, C. S. (1999). Self-theories: Their role in motiva-tion, personality, and development. New York: Psy-chology Press.

Dweck, C. S., Chiu, C. Y., & Hong, Y. Y. (1995). Im-plicit theories and their role in judgments and reac-tions: A world from two perspectives. PsychologicalInquiry, 6, 267–285.

Egloff, B., Schmukle, S. C., Burns, L. R., &Schwerdtfeger, A. (2006). Spontaneous emotion reg-ulation during evaluated speech tasks: Associationswith negative affect, anxiety expression, memory,and physiological responding. Emotion, 6, 356–366.

Eisenberg, N., & Morris, A. S. (2002). Children’semotion-related regulation. Advances in Child De-velopment, 30, 189–229.

Ellsworth, P. C., & Scherer, K. R. (2003). Appraisal pro-cesses in emotion. In R. J. Davidson, K. R. Scherer, &H. H. Goldsmith (Eds.), Handbook of affective sci-ences (pp. 572–595). New York: Oxford UniversityPress.

Freud, A. (1936). The ego and the mechanisms of de-fense. New York: International Universities Press.

Freud, S. (1959). Inhibitions, symptoms and anxiety (A.Strachey, Trans., & J. Strachey, Ed.). New York:Norton. (Original work published 1926)


Frijda, N. H. (1988). The laws of emotion. AmericanPsychologist, 43, 349–358.

Gilbert, D. T., Pinel, E. C., Wilson, T. D., Blumberg, S.J., & Wheatley, T. P. (1998). Immune neglect: Asource of durability bias in affective forecasting.Journal of Personality and Social Psychology, 75,617–638.

Goldin, P. R., McRae, K., Ramel, W., & Gross, J. J. (inpress). The neural bases of emotion regulation duringreappraisal and suppression of negative emotion. Bi-ological Psychiatry.

Gross, J. J. (1998a). Antecedent- and response-focusedemotion regulation: Divergent consequences for ex-perience, expression, and physiology. Journal of Per-sonality and Social Psychology, 74, 224–237.

Gross, J. J. (1998b). The emerging field of emotion reg-ulation: An integrative review. Review of GeneralPsychology, 2, 271–299.

Gross, J. J. (2001). Emotion regulation in adulthood:Timing is everything. Current Directions in Psycho-logical Science, 10, 214–219.

Gross, J. J. (2002). Emotion regulation: Affective, cog-nitive, and social consequences. Psychophysiology,39, 281–291.

Gross, J. J. (Ed.). (2007). Handbook of emotion regula-tion. New York: Guilford Press.

Gross, J. J., & John, O. P. (2003). Individual differ-ences in two emotion regulation processes: Implica-tions for affect, relationships, and well-being. Jour-nal of Personality and Social Psychology, 85, 348–362.

Gross, J. J., & Levenson, R. W. (1993). Emotional sup-pression: Physiology, self-report, and expressivebehavior. Journal of Personality and Social Psychol-ogy, 64, 970–986.

Gross, J. J., & Levenson, R. W. (1997). Hiding feelings:The acute effects of inhibiting positive and negativeemotions. Journal of Abnormal Psychology, 106, 95–103.

Gross, J. J., & Munoz, R. F. (1995). Emotion regulationand mental health. Clinical Psychology: Science andPractice, 2, 151–164.

Gross, J. J., Richards, J. M., & John, O. P. (2006). Emo-tion regulation in everyday life. In D. K. Snyder, J. A.Simpson, & J. N. Hughes (Eds.), Emotion regulationin couples and families: Pathways to dysfunction andhealth (pp. 13–35). Washington, DC: American Psy-chological Association.

Gross, J. J., & Thompson, R. A. (2007). Emotion regu-lation: Conceptual foundations. In J. J. Gross (Ed.),Handbook of emotion regulation (pp. 3–24). NewYork: Guilford Press.

Harris, C. R. (2001). Cardiovascular responses of em-barrassment and effects of emotional suppression ina social setting. Journal of Personality and SocialPsychology, 81, 886–897.

Jackson, D. C., Malmstadt, J. R., Larson, C. L., &Davidson, R. J. (2000). Suppression and enhance-ment of emotional responses to unpleasant pictures.Psychophysiology, 37, 515–522.


John, O. P., & Gross, J. J. (2004). Healthy and un-healthy emotion regulation: Personality processes, in-dividual differences, and lifespan development. Jour-nal of Personality, 72, 1301–1334.

John, O. P., & Gross, J. J. (2007). Individual differences


in emotion regulation. In J. J. Gross (Ed.), Handbookof emotion regulation (pp. 351–372). New York:Guilford Press.

Kahneman, D. (2000). Experienced utility and objectivehappiness: A moment-based approach. In D.Kahneman & A. Tversky (Eds.), Choices, values, andframes (pp. 673–692). Cambridge, UK: CambridgeUniversity Press.

Keltner, D. (1995). Signs of appeasement: Evidence forthe distinct displays of embarrassment, amusement,and shame. Journal of Personality and Social Psy-chology, 68, 441–454.

Keltner, D., & Kring, A. (1998). Emotion, social func-tion, and psychopathology. Review of General Psy-chology, 2, 320–342.

Langston, C. A. (1994). Capitalizing on and copingwith daily-life events: Expressive responses to posi-tive events. Journal of Personality and Social Psy-chology, 67, 1112–1125.

Lazarus, R. S. (1966). Psychological stress and the cop-ing process. New York: McGraw-Hill.

Lazarus, R. S. (1991a). Emotion and adaptation. Ox-ford: Oxford University Press.

Lazarus, R. S. (1991b). Progress on a cognitive–motivational–relational theory of emotion. AmericanPsychologist, 46, 819–834.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisaland coping. New York: Springer.

Levenson, R. W. (1999). The intrapersonal functions ofemotion. Cognition and Emotion, 13, 481–504.

Levesque, J., Fanny, E., Joanette, Y., Paquette, V.,Mensour, B., Beaudoin, G., et al. (2003). Neural cir-cuitry underlying voluntary suppression of sadness.Biological Psychiatry, 53, 502–510.

Mauss, I. B., Bunge, S. A., & Gross, J. J. (2007). Auto-matic emotion regulation. Socil and Personality Psy-chology Compass, 1, 146–167.

Mauss, I. B., Cook, C. L., & Gross, J. J. (2007). Auto-matic emotion regulation during an anger provoca-tion. Journal of Experimental Social Psychology, 43,698–711.

Mauss, I. B., Evers, C., Wilhelm, F. H., & Gross, J.J.(2006). How to bite your tongue without blowingyour top: Implicit evaluation of emotion regulationpredicts affective responding to anger provocation.Personality and Social Psychology Bulletin, 32, 589–602.

Mauss, I. B., Levenson, R. W., McCarter, L., Wilhelm, F.H., & Gross, J. J. (2005). The tie that binds?: Coher-ence among emotion experience, behavior, and physi-ology. Emotion, 5, 175–190.

Mauss, I. B., Wilhelm, F. H., & Gross, J. J. (2003). Au-tonomic recovery and habituation in social anxiety.Psychophysiology, 40, 648–653.

Mauss, I. B., Wilhelm, F. W., & Gross, J. J. (2004). Isthere less to social anxiety than meets the eye? Emo-tion experience, expression, and bodily responding.Cognition and Emotion, 18, 631–662.

McCanne, T. R., & Anderson, J. A. (1987). Emotionalresponding following experimental manipulation of

facial electromyographic activity. Journal of Person-ality and Social Psychology, 52, 759–768.

Mesquita, B., & Albert, D. (2007). The cultural regula-tion of emotions. In J.J. Gross (Ed.), Handbook ofemotion regulation (pp. 486–503). New York:Guilford Press.

Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). De-lay of gratification in children. Science, 244, 933–938.

Molden, D. C., & Dweck, C. S. (2006). Finding “mean-ing” in psychology: A lay theories approach to self-regulation, social perception, and social develop-ment. American Psychologist, 61, 192–203.

Morrow, J., & Nolen-Hoeksema, S. (1990). Effects ofresponses to depression on the remediation of depres-sive affect. Journal of Personality and Social Psychol-ogy, 58, 519–527.

Moses, E. B., & Barlow, D. H. (2006). A new unifiedtreatment approach for emotional disorders based onemotion science. Current Directions in PsychologicalScience, 15, 146–150.

Nesse, R. M. (2000). Is depression an adaptation? Ar-chives of General Psychiatry, 57, 14–20.

Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L.(1993). Response styles and the duration of episodesof depressed mood. Journal of Abnormal Psychology,102, 20–28.

Ochsner, K. N., Bunge, S. A., Gross, J. J., & Gabrieli, J.D. E. (2002). Rethinking feelings: An fMRI study ofthe cognitive regulation of emotion. Journal of Cog-nitive Neuroscience, 14, 1215–1229.

Ochsner, K. N., & Gross, J. J. (2004). Thinking makes itso: A social cognitive neuroscience approach to emo-tion regulation. In R. F. Baumeister & K. D. Vohs(Eds.), Handbook of self regulation: Research, the-ory, and applications (pp. 229–255). New York:Guilford Press.

Ochsner, K. N., & Gross, J. J. (2005). The cognitivecontrol of emotion. Trends in Cognitive Sciences, 9,242–249.

Ochsner, K. N., Ray, R. R., Cooper, J. C., Robertson, E.R., Chopra, S., Gabrieli, J. D. E., & Gross, J. J.(2004). For better or for worse: Neural systems sup-porting the cognitive down- and up-regulation ofnegative emotion. NeuroImage, 23, 483–499.

Parrott, W. G. (1993). Beyond hedonism: Motives forinhibiting good moods and for maintaining badmoods. In D. M. Wegner & J. W. Pennebaker (Eds.),Handbook of mental control (pp. 278–305).Englewood Cliffs, NJ: Prentice-Hall.

Phelps, E. A. (2006). Emotion and cognition: Insightsfrom studies of the human amygdala. Annual Reviewof Psychology, 57, 27–53.

Ray, R. D., Wilhelm, F. H., & Gross, J. J. (in press).All in the mind’s eye?: Anger rumination and reap-praisal. Journal of Personality and Social Psychol-ogy.

Redelmeier, D. A., & Kahneman, D. (1996). Patients’memories of painful medical treatments: Real-time


and retrospective evaluations of two minimally inva-sive procedures. Pain, 66, 3–8.

Richards, J. M., & Gross, J. J. (1999). Composure atany cost?: The cognitive consequences of emotionsuppression. Personality and Social Psychology Bul-letin, 25, 1033–1044.

Richards, J. M., & Gross, J. J. (2000). Emotion regula-tion and memory: The cognitive costs of keepingone’s cool. Journal of Personality and Social Psychol-ogy, 79, 410–424.

Richards, J. M., & Gross, J. J. (2006). Personality andemotional memory: How regulating emotion impairsmemory for emotional events. Journal of Research inPersonality, 40, 631–651.

Rothbart, M. K., Ziaie, H., & O’Boyle, C. G. (1992).Self-regulation and emotion in infancy. In N.Eisenberg & R. A. Fabes (Eds.), Emotion and its reg-ulation in early development (pp. 7–23). San Fran-cisco: Jossey-Bass.

Rothbaum, F., Weisz, J. R., & Snyder, S. S. (1982).Changing the world and changing the self: A two-process model of perceived control. Journal of Per-sonality and Social Psychology, 42, 5–37.

Rottenberg, J., & Gross, J. J. (2003). When emotiongoes wrong: Realizing the promise of affective sci-ence. Clinical Psychology: Science and Practice, 10,227–232.

Rottenberg, J., Gross, J. J., & Gotlib, I. H. (2005). Emo-tion context insensitivity in major depressive disor-der. Journal of Abnormal Psychology, 114, 627–639.

Rottenberg, J., & Johnson, S. L. (2006). Emotion andpsychopathology: Bridging affective and clinical sci-ence. Washington, DC: American Psychological As-sociation.

Rottenberg, J., Kasch, K. L., Gross, J. J., & Gotlib, I.H. (2002). Sadness and amusement reactivity dif-ferentially predict concurrent and prospective func-tioning in major depressive disorder. Emotion, 2,135–146.

Schaefer, S. M., Jackson, D. C., Davidson, R. J.,Aguirre, G. K., Kimberg, D. Y., & Thompson-Schill, S. L. (2002). Modulation of amygdalar ac-tivity by the conscious regulation of negative emo-tion. Journal of Cognitive Neuroscience, 14, 913–921.

Scherer, K. R., Schorr, A., & Johnstone, T. (Eds.).(2001). Appraisal processes in emotion: Theory,methods, research. New York: Oxford UniversityPress.

Shiffrin, R. M., & Schneider, W. (1977). Controlled andautomatic human information processing: II. Percep-tual learning, automatic attending, and a general the-ory. Psychological Review, 84, 127–190.

Spasojevic, J., & Alloy, L. B. (2001). Rumination as a

common mechanism relating depressive risk factorsto depression. Emotion, 1, 25–37.

Srivastava, S., Tamir, M., McGonigal, K. M., John, O.P., & Gross, J. J. (2008). The social consequences ofemotional suppression: A prospective study of thetransition to college. Manuscript submitted for publi-cation.

Srull, T. K., & Wyer, R. S. (1979). The role of categoryaccessibility in the interpretation of informationabout persons: Some determinants and implications.Journal of Personality and Social Psychology, 37,1660–1672.

Stemmler, G. (1997). Selective activation of traits:Boundary conditions for the activation of anger. Per-sonality and Individual Differences, 22, 213–233.

Stepper, S., & Strack, F. (1993). Proprioceptive determi-nants of emotional and nonemotional feelings. Jour-nal of Personality and Social Psychology, 64, 211–220.

Stifter, C. A., & Moyer, D. (1991). The regulation ofpositive affect: Gaze aversion activity during mother–infant interaction. Infant Behavior and Develop-ment, 14, 111–123.

Strack, F., & Deutsch, R. (2004). Reflective and impul-sive determinants of social behavior. Personality andSocial Psychology Review, 8, 220–247.


Sutton, R. I. (1991). Maintaining norms about ex-pressed emotions: The case of bill collectors. Admin-istrative Science Quarterly, 36, 245–268.

Tamir, M. (2005). Don’t worry, be happy?: Neuroti-cism, trait-consistent affect regulation, and perfor-mance. Journal of Personality and Social Psychology,89, 449–461.

Tamir, M., John, O. P., Srivastava, S., & Gross, J. J.(2007). Implicit theories of emotion: Affective and so-cial outcomes across a major life transition. Journal ofPersonality and Social Psychology, 92, 731–744.

Thoits, P. A. (1985). Self-labeling processes in mental ill-ness: The role of emotional deviance. American Jour-nal of Sociology, 91, 221–249.

Thompson, R. A. (1994). Emotion regulation: A themein search of definition. In N. A. Fox (Ed.), The devel-opment of emotion regulation: Biological and behav-ioral considerations. Monographs of the Society forResearch in Child Development, 59(2–3, Serial No.240), 25–52).

Tsai, J. L., Knutson, B., & Fung, H. H. (2006). Culturalvariation in affect valuation. Journal of Personalityand Social Psychology, 90, 288–307.


C H A P T E R 3 2

Emotional Complexity

KRISTEN A. LINDQUIST and LISA FELDMAN BARRETT

Imagine someone who is “emotionally com-plex,” and a number of characteristics mightcome to mind: the ability to see the good andthe bad in all things; the ability to describe feel-ings with detail and precision; the ability tospecifically and reliably anticipate which feel-ings will arise in a given situation; or the ten-dency to remember experiencing many emo-tions at once. You might also imagine someonewho tends to characterize himself or herself asan emotionally complex person. It appears thatthere are myriad ways in which a person can beconsidered “emotionally complex.” In fact, theconcept of emotional complexity is similarlyvaried in the psychological literature. In thischapter, we review this literature with a focuson three main formulations of emotional com-plexity as (1) dialecticism and precision inpeople’s self-reports of emotion experiences;(2) explicit, propositional knowledge aboutemotion in situations; and (3) people’s self-characterizations of their degree of complexity.

Before we begin our review of the emotionalcomplexity literature, it seems apropos toclearly define just what we think people are be-

ing complex about. There are two general ap-proaches to defining the nature of emotion.A “natural-kinds” perspective (e.g., Ekman,1972; Izard, 1994; Tomkins, 1962; Panksepp,2000; Roseman, 1984) views a select set ofemotions (e.g., anger, sadness, fear, anger, dis-gust, and happiness) as biologically given andfixed categories (for recent reviews of thenatural-kinds perspective, see Barrett, 2006a;Barrett, Lindquist, et al., 2007). In this perspec-tive, complexity in self-reports of online experi-ences of emotion results when more than oneemotion circuit fires at a given point in time, oris caused by variations in the accuracy withwhich people translate experience into words.In this sense, emotion categories are perceptualcategories that are either hard-coded into thebrain at birth or learned by inducing statisticalregularities in the environment. According tothis perspective, people gain complexity inpropositional knowledge of emotion when theylearn to associate the consequences of the firingof a given emotion circuit with certain environ-mental conditions. Complexity in self-characterizations of experience derives from

513

differences in people’s ability to characterizetheir experiences as complex. In this sense,complexity in self-characterizations of experi-ence exists because some people merely overlaycomplexity onto what is really a fixed and sta-ble system.

The second approach to the nature of emo-tion takes a “psychological constructionist”perspective (e.g., Barrett, 2006b; Barrett,Lindquist, et al., 2007; Russell, 2003; Mandler,1975; Schacter & Singer, 1962). In this view,complexity is not a conceptual overlay; it is in-trinsic to the neurobiological and psychologicalsystems that constitute emotional experience.From this point of view, a discrete emotionalevent emerges in consciousness (i.e., a person“has an emotion”) when an instance of a morebasic core affective state is automatically andimplicitly categorized as having emotionalmeaning. Core affect is an ongoing, ever-changing psychologically primitive state thathas both valenced, and to some extent, arousal-based properties (see Barrett, 2004; Russell,2003; Russell & Barrett, 1999). The eventsthat people call (in English) anger, sadness, fear,and so on, result when core affect is catego-rized using the “conceptual system for emo-tion.” This term refers to what people knowabout emotion and how that knowledge is rep-resented in emotion categories (see Barrett,2006b). Categorizing an instance of core affectproceeds efficiently and automatically to pro-duce a state that is at once affective and con-ceptual, where internal sensory informationfrom the body and external sensory informa-tion from the world are bound together in amoment in time. A person experiences an emo-tion like anger, for example, when a state of un-pleasant affect is categorized as having beencaused by an event that blocked a person’sgoals. Categorizing core affect bounds it as adiscrete experience: it allows core affective ex-perience to pop out in consciousness and givesit meaning. Categorization transforms core af-fect into an intensional state, allowing a personto make inferences about what caused the af-fective change, what to do next, and to com-municate that state to others in an effective andefficient manner. From this perspective, emo-tional complexity is the direct result of the con-ceptual system for emotion.

In the context of this review, we explore howvarious forms of emotional complexity resultfrom the form and function of the conceptualsystem for emotion. The idea that emotional

and conceptual complexity are related is notnew. Lane and colleagues have proposed acognitive-developmental model of emotionalawareness (e.g., Lane & Garfield, 2005; Lane& Pollermann, 2002) that argues for concep-tual development as the core determinant ofemotional complexity. The view that we pro-pose in this chapter is distinct from thecognitive-developmental model on severalpoints. First, we propose that emotionalcomplexity is grounded in a highly flexi-ble, context-dependent conceptual system ofsituated representations of emotion (seeNiedenthal, Chapter 36, this volume), ratherthan a system composed of schemas or fixedprototypic or script-like concepts. Second, weemphasize the possibility that conceptual com-plexity has a hand in psychologically construct-ing each episode that we call “emotion” duringthe online categorization of affective (pleasantor unpleasant) events (cf. Barrett, 2006b).Finally, our perspective assumes that complex-ity in the conceptual system derives not onlyfrom which categories of emotion populate aperson’s conceptual system (i.e., whether or nota person possesses categories such as “sad,”“happy,” “fear,” etc.), but also fromidiographic variations in the content of cate-gory knowledge (e.g., what exactly a personknows about happiness, sadness, or fear), inthe representational format of category knowl-edge (i.e., how that category knowledge is con-stituted in memory and during online use ofcategory information), and in the resources touse category knowledge to construct the expe-rience of emotion (i.e., whether people canreadily access and manipulate what they knowduring online experience). In this chapter, weargue that greater complexity in the structure,content, and representational format of theconceptual system, and deftness in wieldingsuch knowledge, contribute to greater com-plexity in the psychological events that we call“emotion.” The conceptual system for emotionis a unifying factor in producing emotionalcomplexity in its various forms.

COMPLEXITY IN SELF-REPORTEDEXPERIENCES OF EMOTION

The best way to assess the properties of an ex-perience (such as its complexity) is to ask peo-ple how they feel and to examine the content ofwhat they answer. Although self-report meth-


ods have obvious drawbacks for assessing theprocesses that might produce emotional com-plexity, they can tell scientists a good dealabout the contents of what people feel (cf.Barrett, 2004; Barrett, Mesquita, Ochsner, &Gross, 2007; Conner, Barrett, Bliss-Moreau,Lebo, & Kashub, 2003). “Emotional complex-ity” refers to two types of contents in emotionself-reports: dialecticism and granularity in theexperience of emotion. Implicit in both is theidea that self-reports are verbal behaviors thatcan be analyzed in a way that unearths thestructure of emotion experience.

Dialecticism in Self-Reported Experienceof Emotion

As a form of emotional complexity, “dialecti-cism” refers to the experience of pleasant andunpleasant states in a coincidental or tempo-rally related fashion (Bagozzi, Wong, & Yi,1999). The term is derived from Confucianphilosophy (see Peng & Nisbett, 1999) andwas first used in the context of cross-culturalresearch (Bagozzi et al., 1999). In the contextof this chapter, “dialecticism” refers to all stud-ies assessing the relation in reported experi-ences of positive and negative emotions.

Cross-Cultural Variation in Dialectic Experienceof Emotion

There appears to be general scientific consen-sus that individuals from Eastern cultures suchas China, Korea, or Japan are more likely tohave dialectic experiences of emotion than peo-ple from Western cultures such as the UnitedStates or Europe are (Bagozzi et al., 1999;Kitayama, Markus, & Kurokawa, 2000;Scollon, Oishi, Diener, & Biswas-Diener, 2005;Shimmack, Oishi, & Diener, 2002). There isalso some evidence for a Yiddish form of dia-lecticism (see Peng & Nisbett, 1999), but mostresearch focuses on the East–West dichotomy.The psychological dimensions typically usedto describe the difference between Easternand Western cultures (e.g., individualism–collectivism) fail to account for the cross-cultural variance in the dialectic experience ofemotions, however (cf. Shimmack et al., 2002).Instead, differences in dialecticism are thoughtto derive from emotion regulation strategiesthat are promoted by the philosophical tradi-tions within each culture. Eastern dialecticismhas been linked to a philosophical tradition

that promotes balance and the acceptance ofcontradiction (Peng & Nisbett, 1999, althoughsee Lee, 2000), where opposites are conceptu-alized as being intrinsically related to one an-other. In this tradition, the concept “good”overlaps with “bad” such that something canbe both good and bad at the same time. Thisorientation leads people to enlist in strategiesthat promote more affectively balanced lives.In contrast, Western experience is grounded inAristotelian philosophy that favors an “either–or” type of reasoning, where opposites are bi-polar. In this tradition, the concept “good” isthe antithesis of “bad,” so something can neverbe both at the same time. This orientation leadspeople to enlist in strategies that promote max-imally pleasant (at the expense of unpleasant)experience (Heine, Lehman, Markus, &Kitayama, 1999). The philosophical differ-ences that characterize the East–West dividetranslate into different conceptions of whatconstitutes “ideal affect” (the affective statesthat people deem most valuable and desire tofeel most; Tsai, 2007). What may lie at theheart of the cross-cultural differences in the di-alectic experience of emotion is a difference inthe value that cultures place on the experienceof unpleasant emotion.

The claims for broad cross-cultural variationin dialecticism may hide individual variabilityin dialectic experience within each cultural tra-dition, however. “Dialecticism” is sometimesoperationally defined as a negative correlationbetween reported pleasant and unpleasantemotions (e.g., Bagozzi et al., 1999), but it hasalso been defined as a zero correlation (e.g.,Diener & Emmons, 1985) or any decrease inthe magnitude of negative correlation betweenthe two (Shimmack et al., 2002; Carstensen,Pasupathi, Mayr, & Nesselroade, 2000).1 Incross-sectional studies, the presence of a near-zero correlation between reports of pleasantand unpleasant experience indicates that thereis no systematic covariation between reports ofthese states within a group of people. Whereassome respondents report feeling both morepleasant and more unpleasant emotion whencompared to each respective group mean, otherindividuals report greater pleasant emotionwith an absence of negative emotion (i.e.,pleasant scores fall above the group mean forpleasure, but unpleasant scores fall below themean for displeasure), or vice versa. In reality,then, correlations near zero mask individualdifferences in the dialectic experience of emo-

32. Emotional Complexity 515

tion that appear to be present in both Easternand Western contexts, although perhaps at dif-ferent base rates.

Individual Variation in Dialectic Experienceof Emotion

Several lines of research point to individual dif-ferences in the dialectic experience of emotion.As people age, their experience becomes moredialectic. In an experience-sampling study ofAmerican participants ranging in age from 18to 94, younger participants experienced pleas-ant and unpleasant emotions as inversely re-lated, but this correlation diminished with age(Carstensen et al., 2000). Greater cognitivecomplexity promotes dialectic thought (i.e., theability to conceptualize contradiction), therebyproducing more dialectic experience. Individ-uals greater in cognitive complexity did notdemonstrate a systematic relationship betweenpleasant and unpleasant experiences, as com-pared to individuals low in complexity, whosepleasant and unpleasant experiences were in-versely related (Davis, Zautra, & Smith, 2004;Reich, Zautra, & Potter, 2001).

Women demonstrate exaggerated versions ofthe dialecticism patterns associated with theirculture (e.g., Bagozzi et al., 1999; Shimmack etal., 2002). In one cross-cultural study, Ameri-can women had less dialectic experience (i.e.,larger negative correlations between experienceof pleasant and unpleasant emotions), but Chi-nese women had more dialectic experience (i.e.,larger positive correlations between experi-ences of pleasant and unpleasant emotions),than their respective male counterparts(Bagozzi et al., 1999). These findings stand incontrast to the stereotype that women are themore emotionally complex sex (for a discus-sion, see Barrett, Robin, Pietromonaco, &Eyssell, 1998), and call into question thegeneralizability of findings that Westernwomen have more complex emotional aware-ness than Western men do (Barrett, Lane,Sechrest, & Schwartz, 2000).

Sources of Knowledge Sampled

The strongest evidence for dialecticism comesfrom responses that are more likely to be in-fused with culturally embedded beliefs, such asjudging how a hypothetical scenario might feel(e.g., Leu et al., 2007), recalling prior experi-ences (e.g., Kitayama et al., 2000; Reich et al.,

2001; Oishi, 2002; Ong & Bergeman, 2004;Carstensen et al., 2000), or summarizing expe-riences across a period of time to produce a re-sponse (e.g., Larsen, McGraw, & Cacioppo,2001; Larsen, McGraw, Mellers, & Cacioppo,2004; for discussions, see Barrett, 1997; Rob-inson & Clore, 2002; Ross, 1989). Momentaryexperiences of emotion (“How happy are youright now?”) are less belief-based and corre-spondingly fail to yield evidence for dialecti-cism (e.g., Vansteelandt, Van Mechelen, &Nezlek, 2005), even in cultures typically char-acterized as more prone to dialectic experiencesof emotion (e.g., Chinese culture; Scollon et al.,2005).

Granularity in Self-Reported Experienceof Emotion

A second form of complexity in self-reports ofemotion experience is the ability to verballycharacterize such experiences with precision,referred to as “emotional granularity” (Barrett,1998, 2004; Barrett, Gross, Christensen, &Benvenuto, 2001; Feldman, 1995; Tugade,Barrett, & Gross, 2007).2 Individuals who areemotionally granular use emotion adjectives(such as “sad,” “contented,” “angry,”“afraid,” “joyful,” etc.) to represent discreteand qualitatively different experiences. Thoselower in granularity use these same wordsin a less precise way to represent broad,global affective states, such as pleasantness–unpleasantness or arousal–quiescence.

Determining Emotional Granularity

Emotional granularity is determined by assess-ing the relatedness in emotion experiences asthey are represented through people’s endorse-ment of emotion adjectives during the self-report process. Typically, participants are givena set of adjectives (“happy,” “anxious,” “an-noyed,” etc.) and rate, on a Likert scale, howclosely each adjective described their emotionalstate at a given measurement moment; this isdone across a series of measurement instances(e.g., in an experience-sampling paradigm).The relatedness between ratings is then calcu-lated by using person-correlations (e.g.,Barrett, 1998; Barrett et al., 2001; Feldman,1995) or intraclass correlations (Tugade et al.,2007). A strong positive correlation betweentwo such ratings is evidence for low granular-ity, meaning that an individual uses emotion


words (e.g., “angry” and “sad”) in a nonspeci-fic fashion to represent what those two feelingstates have in common (e.g., displeasure). Aweak (or zero) correlation, or a strong negativecorrelation, indicates high granularity, meaningthat an individual uses two emotion words torepresent two qualitatively different states(e.g., “angry” as a different feeling from“sad”). (A correlation of zero between ratingsof “angry” and “sad” indicates high granular-ity, because it reflects the fact that a person usesthe words differently across measurement oc-casions. In some instances, “angry” and “sad”are rated above a person’s own mean [indicat-ing that both intense anger and sadness are be-ing felt], whereas in other instances, “angry” israted higher and “sad” rated lower, or viceversa [indicating that one is being felt in the ab-sence of the other].)

Individuals from the United States vary tre-mendously in their degree of emotional granu-larity, even when verbal ability is controlled for.Estimates from one study (Barrett, 1998) putthe granularity for unpleasant experiences be-tween .16 and .89 (M = .52, SD = .24). Thegranularity for pleasant experiences ranged be-tween .51 and .96 (M = .77, SD = .28). Peoplewho represent their negative states in a granu-lar way also typically report their positivestates in a granular way (Barrett, 1998).

Processes Underlying Emotional Granularity

THE CONCEPTUAL SYSTEM FOR EMOTION

When people report on their experiences, theymust represent and communicate those experi-ences in words. Differences in the structure ofthe conceptual system for emotion may be onesource of variation in how people describe theirexperiences of emotion, leading to variation inemotional granularity. For example, the wordsthat correspond to basic-level categories may in-fluence how people use emotion words to repre-sent experience during the self-report process.Categories for emotion can be thought of as hi-erarchically organized, ranging from categoriesthat represent the most general feelings to cate-gories that represent the most specific feelings(e.g., “pleasant” and “unpleasant”) (e.g., “frus-tration,” “aggravation,” “irritation,” etc., forthe category of “anger”). Basic-level categoriesrepresent the level of conceptualization thatpeople prefer when parsing a domain (Murphy,2002; Rosch, Mervis, Gray, Johnson, & Boyes-

Braem, 1976), and words that correspond to ba-sic-level categories are used most frequently byparents when naming objects and events fortheir children (Brown, 1957). Most researchersassume that the categories corresponding to thewords “anger,” “sadness,” “fear,” and so forthare basic-level, and cross-sectional researchbears this out (e.g., Alonso-Arbiol et al., 2006;Fehr & Russell, 1984; Shaver, Schwartz, Kirson,& O’Connor, 1987; Shaver, Murdaya, & Fraley,2001). It is possible, however, that significantand important individual differences exist inwhich categories are used as basic. We proposethat individuals lower in granularity usenomothetically superordinate category knowl-edge as the modal means of categorizing theirexperiences (e.g., they categorize as “unpleas-ant” those feelings that might normatively becategorized as “sadness,” “anger,” or “fear,”and categorize as “pleasant” those feelings thatwould normatively be categorized as “joy,”“happiness,” or “interest”). Categorizing an af-fective state gives it meaning, so that a personcan communicate it to others, make inferencesabout it, and make predictions about how to act(cf. Barrett, 2006b). If a person usessuperordinate categories of emotion (e.g.,“pleasant” and “unpleasant”) as the modalmeans of categorization, then that person notonly will communicate affective experience in abroad manner, but may also experience thosestates as broad and undifferentiated. Alter-nately, individuals higher in granularity may beexperts in emotion and use subordinate catego-ries as basic (e.g., they may use “frustration,”“annoyance,” and “rage” as basic, rather than“anger”). These individuals may report and ex-perience affect as discrete and nuanced emo-tional events, much in the same way that expertsin reading X-rays (Christensen et al., 1981), sex-ing chickens (Biederman & Shiffrar, 1987), andjudging wine (Solomon, 1990, 1997), can per-ceive important differences that novices cannot(for a similar point, see Lane, 2000, p. 348).

Although no studies have yet explicitly ex-amined the relation between conceptualbasicness and emotional granularity, existingfindings suggest the plausibility of such a link.Two-year-old children typically use the word“sad” to refer to anything unpleasant (suchas faces depicting “anger,” “sadness,” and“fear”), and the word “happy” to refer to any-thing pleasant (see Widen & Russell, Chapter21, this volume); that is, their basic level of cat-egorization is “pleasant–unpleasant.” This pat-


tern of response is very similar to how low-granularity adults use emotion adjectives tocommunicate only the most global or generalaffective states (Barrett, 1998, 2004; Feldman,1995). As children’s conceptual systems be-come more differentiated, they learn to distin-guish reliably between exemplars of otheremotion categories (e.g., “anger,” “fear,” and“sadness”). As children develop, their ability tocorrectly identify faces depicting these emo-tions begins to mirror that of adults high ingranularity, and who use emotion words in aprecise way to represent emotional experience.We would expect, then, that a child’s experi-ence of emotion becomes more granular as hisor her conceptual system becomes more differ-entiated. Even more tempting are the implica-tions for emotional complexity in adulthood.Complexity in the content of a person’s concep-tual system for emotion might intrinsicallyshape the complexity of emotion experience inmuch the same way that conceptual complexityshapes children’s perceptions of emotion acrossearly stages of development.

In addition to variation in structure, the con-tent of the conceptual system for emotion maybe linked to emotional granularity. Typically,emotion concepts (e.g., the concept for “an-ger”) have been conceived of as single, stablerepresentations of information that are orga-nized classically (e.g., Johnson-Laird & Oatley,1989; Ortony, Clore, & Foss, 1987), as proto-types (Russell, 1991), as schemas (Lane &Pollermann, 2002), or as theories (Clore &Ortony, 1991) that can be retrieved from long-term memory when needed. More recently,emotion concepts have been understood asflexible constructions deriving from a moregenerative system of heterogeneous and variedsituated conceptualizations (see Niedenthal,Chapter 36, this volume). From the vantagepoint of this more recent view, emotion con-cepts are examples of what Edelman (1989)has termed the “remembered present,” whereneural representations of an emotion exemplarcombines information from the current situa-tion with information stored from previous ex-perience. The result contributes to the repre-sentational structure that is utilized duringlater online categorizations of percepts fromthat domain.

To date, no studies have examined idio-graphic variability in the richness of conceptualcontent. Studies assessing the conceptual sys-tem for emotion primarily ask individuals to

list words for emotion categories (e.g., Fehr &Russell, 1984, Study 1), to rate the similaritybetween words (e.g., Alonso-Arbiol et al.,2006, Study 2; Barrett, 2004; Shaver et al.,1987, Study 1), or to rate the prototypicality ofemotion words (e.g., Alonso-Arbiol et al.,2006, Study 1; Fehr & Russell, 1984, Study 3;Shaver et al., 1987, Study 1). Emotion wordsare not synonymous with emotion concepts,however, so tests of how people use and thinkabout emotion words may not sufficiently mapthe variability and detail in what people knowabout emotion. Words can be processed by us-ing shallow, quick, associative methods thatfail to activate deeper representations of cate-gory information (Barsalou, Santos, Simmons,& Wilson, in press). Even studies that ask par-ticipants to write narratives for episodes of“anger,” “sadness,” and “fear,” and then codethose narratives for their prototypical features(Shaver et al., 1987, Study 2), may not be suit-ably sensitive to capture individual differences,because people can construct a prototype of acategory even if conceptual knowledge is notstored that way.

There is some evidence to suggest that so-cialization contributes to variability in theconceptual system for emotion. People mayhave a richer conceptual base for emotionavailable to them if they are exposed to awider range of emotion categories, are taughta more varied emotion vocabulary, and learnto represent their experiences with greater de-tail and complexity. Children learn aboutemotion categories through formal, rule-based instruction, where parents make ex-plicit links among affective feelings, a situa-tional context, and emotion words (e.g.,when a child throws a toy at Jimmy, a parentmight say, “You’re feeling angry right now,because Jimmy grabbed your toy withoutasking”). Parents also reminisce about emo-tional memories in a way that helps childrenlearn about a particular emotion category—by discussing the feelings that were involvedduring the experience of that emotion, the in-terpersonal consequences of expressing thatemotion, or coping strategies relevant to thatemotion (e.g., Fivush, Berlin, Sales, Mennuti-Washburn, & Cassidy, 2003; Miller &Sperry, 1988). Indeed, children whose parentsspeak to them about emotion know moreabout emotion (e.g., Harris, 2006a, 2006b;Harris, de Rosnay, & Pons, 2005; de Rosnay,Pons, Harris, & Morell, 2004).


Children may also acquire emotion conceptsvia inductive, associative learning, such aswhen an emotion word is used (but not explic-itly paired) with features of an emotional epi-sode (e.g., a child hears a parent refer to feelingdepressed while the parent is slumped at the ta-ble with a defeated expression, discussing a re-cent job loss). The role of associative learningin the acquisition of emotion category knowl-edge remains to be empirically tested, however.

COGNITIVE RESOURCES

Differences in granularity may stem not onlyfrom what people know about emotion, butalso from how well they use what they know.For example, a person’s working memory ca-pacity (WMC) probably shapes his or her abil-ity to attend to affective states and to accessand efficiently wield conceptual knowledgeduring emotion representation. WMC is theability to control attention for the purposes ofprocessing information in contexts where thereare competing demands (Barrett, Tugade, &Engle, 2004). Working memory is requiredduring the self-report process, where respon-dents must hold an experience in mind as theydescribe that feeling, using emotion adjectivespresented to them in a serial format. Witheffortful, controlled processing, people can re-ject adjectives that are not characteristic of thecurrent state: The resulting correlations be-tween emotion adjectives will be substantiallylower, resulting in high emotional granularity.WMC may also have an impact on emotionalexperience itself, because people higher inWMC will be able to hold more informationabout the current affective state in mind as con-ceptual knowledge about emotion (i.e., contentspecific to a certain emotion category such ascontextual information, arousal content, lin-guistic representations) is retrieved to aid in on-line categorization that feeling. This should re-sult not only in the report of more discreteemotional experiences, but also in the experi-ence of more discrete emotional states.

Summary

In this section, we have discussed how emo-tional complexity is observed in verbal repre-sentations of experience. In studies of both dia-lecticism and emotional granularity, evidenceof emotional complexity is derived from thestructure of people’s self-reports of emotion ex-

perience; self-reports are treated as verbal be-haviors, and the degree of complexity in thestructure of those behaviors is examined. Wehave suggested that conceptual knowledgeabout emotion seems to play a role in bothforms of complexity, but not merely becauseself-reports require language. Rather, there isgood reason to hypothesize that conceptualknowledge about emotion plays an intrinsicrole in the complexity of emotions as they areexperienced. Individual differences in the con-tent of the conceptual system for emotion andin executive function, as well as the ways inwhich these differences shape experience, haveyet to be explored and constitute a new frontierof research in emotional complexity. In thenext section, we examine what is known aboutcomplexity in a particular form of conceptualknowledge: propositional knowledge for emo-tion. We then explore the experiential ramifica-tions of deficits in this type of knowledge, toshed further light on the potential mechanismsunderlying emotional complexity.

COMPLEXITY IN PROPOSITIONALKNOWLEDGE OF EMOTION

The Levels of Emotional Awareness Scale

To date, complexity in conceptual knowledgeof emotion has been assessed with measuresthat ask individuals to make explicit, proposi-tional statements about emotion experience(e.g., “If I didn’t win a contest, then I wouldfeel disappointed”). The Levels of EmotionalAwareness Scale (LEAS; Lane, Quinlan,Schwartz, Walker, & Zeinlin, 1990) is the mostfrequently used measure and is based on the as-sumption that emotion experience occurs whenfeelings of emotional arousal are conceptual-ized using knowledge about emotion. Individ-ual differences in emotional complexity, asmeasured by the LEAS, reflect variations in thedegree of differentiation and integration ofa person’s “emotional schemas” (Lane &Pollermann, 2002; Lane & Schwartz, 1987;Lane et al., 1990). According to Lane and col-leagues, the development of emotional schemasis facilitated both by language and by individu-als’ ability to encode and represent past experi-ences of emotional arousal. Individuals who fo-cus on proprioceptive information duringemotional arousal will be more likely to encodeand later represent emotion knowledge in vis-ceral or action-oriented terms. Individuals who


abstract relational meaning from feelings ofemotional arousal (such as how different feel-ings relate, how they are coordinated, etc.) willbe more likely to encode and later representemotion knowledge in a nuanced and differen-tiated fashion.

The LEAS measures the complexity of prop-ositional knowledge of emotion by asking indi-viduals to describe the types of emotional expe-riences that would occur during hypotheticalemotion-eliciting situations (e.g., “You andyour friend are competing for a prize and yourfriend wins. How would you feel? How wouldyour friend feel?”). Responses are coded for theextent to which they make reference to bodilysensations, specific behavioral responses or ac-tion tendencies, or discrete-emotion words,each of which is taken as evidence of increasingemotional complexity (see, Lane et al., 1990).LEAS scoring criteria can be used on any narra-tive of emotional experience, whether it de-scribes a current experience or a prior, remem-bered experience (e.g., Bliss-Moreau, Barrett,Connor, & McCarthy, 2007). In such cases, thescoring represents a person’s tendency to repre-sent his or her experience as emotional (ratherthan somatic or affective), and to characterizethat experience by using multiple emotion con-cepts (rather than reflecting anything about thedetail and situated nature of the conceptualcontent).

Complexity in propositional knowledge foremotion predicts more normative identificationof emotion cues in others and the surroundingcontext. People who score higher on the LEASare better able to identify emotional content inother people’s facial behaviors and in the envi-ronment, using Rau’s (1993) Perception of Af-fect Test (Lane, Sechrest, Reidel, Shapiro, &Kazniak, 2000; Lane et al., 1996). HigherLEAS scores are also associated with greaterright cerebral hemisphere dominance duringperception of emotional face stimuli (Lane,Kivley, Du Bois, Shamasundara, & Schwartz,1995)—a phenomenon that is thought to indexheightened sensitivity to external emotion cues(see Lane et al., 1995).

Sex Differences in Propositional Knowledgeof Emotion

There are robust and consistent sex differencesin complexity of propositional knowledge foremotion. Women from Western contexts rou-tinely outperform men on the LEAS (e.g.,

Conway, 2000; Bliss-Moreau et al., 2007),even when factors such as verbal intelligence,language of origin, or socioeconomic status(SES) are controlled for (Barrett et al., 2000).These sex differences are even present in chil-dren: Girls outperform boys on the children’sversion of the LEAS (LEAS-C; Bajgar,Ciarrochi, Lane, & Deane, 2005). At facevalue, these findings suggest that males and fe-males routinely differ in their use of emotionknowledge, but this may not actually reflectdifferences in emotional aptitude. Sex differ-ences in LEAS performance disappear whenitems on the LEAS are made particularly self-relevant (Bliss-Moreau et al., 2007), suggestingthat men and women may possess the samerange of knowledge, but use it differently underthe testing circumstances imposed by the stan-dard LEAS administration (i.e., a controlledlab situation with few contextual or socialcues). Under such circumstances, responses tohypothetical scenarios (as in the LEAS) are verylikely to draw on culturally infused genderedbeliefs about emotion. Consistent with thisidea is the finding that both male and femalerespondents who describe themselves as moremasculine perform more poorly on the LEASthan do those who describe themselves as lessmasculine (Conway, 2000).

Cultural stereotypes about emotion willhave both distal and proximal effects on LEASresponses. The distal effects of stereotypes be-gin in childhood, where parents transmit ste-reotypes about sex roles to children via implicitmeans (talking about emotion differently withboys and girls) or explicit means (teaching boysand girls to behave in different ways duringemotional situations). For example, middle-class European American mothers elaborateand evaluate emotional memories more and aremore likely to discuss the interpersonal con-texts of those memories, when speaking totheir daughters as opposed to their sons(Fivush et al., 2003). Proximally, stereotypesmay have an impact on LEAS performance be-cause they affect the motivation to respond in acomplex manner, or because they act as a filterthrough which participants retrieve conceptualknowledge about emotion.

Developmental Differences in PropositionalKnowledge of Emotion

Individual differences in propositional knowl-edge for emotion also exist among children


with the LEAS-C (Bajgar et al., 2005). Consis-tent with research using the LEAS in someadult samples (e.g., Lane et al., 1990), perfor-mance on the LEAS-C was associated with in-creased verbal skill and vocabulary. Althoughlinguistic and cognitive capacity may driveLEAS-C performance, the evidence is also con-sistent with the idea that children with greaterlanguage skills have a better understanding ofemotion (e.g., de Rosnay & Harris, 2002; Pons& Harris, 2005; Pons, Lawson, Harris, & deRosnay, 2003; Widen & Russell, Chapter 21,this volume). LEAS-C scores increase with age(Bajgar et al., 2005), suggesting that proposi-tional knowledge about emotion increases incomplexity as people learn about emotion.

Alexithymia: An Absence of ConceptualKnowledge of Emotion

Deficits in propositional knowledge of emotioncharacterize an emotional disturbance knownas “alexithymia.” The term “alexithymia” (lit-erally meaning “absence of words for emo-tion”) was first coined by Sifneos (1973) todescribe patients who appeared to lack concep-tual knowledge of emotion, resulting in an im-paired ability to symbolically represent their af-fective feelings as emotional. Alexithymicindividuals’ apparent lack of conceptualknowledge results in difficulties expressingemotion, imagining, and socializing, and in-creases the likelihood that emotion will be ex-perienced as somatic symptoms (Haviland &Reise, 1996; Rieffe, Oosterveld, & MeerumTerwogt, 2006; see Lane, Ahern, Schwartz, &Kazniak, 1997, for a discussion of thealexithymia construct). As might be expected,individuals high in alexithymia consistentlyperform poorly on the LEAS (e.g., Lane et al.,1996) and on measures of emotional intelli-gence (e.g., Lumley, Gustavson, Partridge, &Labouvie-Vief, 2005; Parker, Taylor, & Bagby,2001). Alexithymia occurs most frequently inolder individuals (although see Rieffe et al.,2006), men, individuals of lower SES, and per-sons with fewer years of education (Lane,Sechrest, & Riedel, 1998). Alexithymia is alsoseen in patients with various psychological andsomatic disorders.

It is widely believed that alexithymia resultswhen individuals fail to develop the cognitiveresources needed to consciously represent emo-tional states (Berenbaum & James, 1994;Frawley & Smith, 2001; Lane & Schwartz,

1987; Taylor, 2000; Taylor, Bagby, & Parker,1997). This rudimentary conceptual system foremotion results in a form of “affectiveblindsight,” where individuals experience sub-jective bodily sensations or “background feel-ings” but fail to experience them as emotional(Lane et al., 1997; Lane & Garfield, 2005); theanalogy is to patients with classic blindsight,who have behavioral awareness of visual ob-jects in the absence of the conscious experienceof seeing those objects (e.g., de Gelder,Vroomen, Pourtois, & Weiskrantz, 1999).

Deficits in Emotion Experience Associatedwith Alexithymia

Alexithymic individuals report less intense ex-perience of emotion (e.g., Luminet, Rimé,Bagby, & Taylor, 2004; Mantani, Okomoto,Shirao, Okada, & Yamawaki, 2005; Stone &Nielson, 2001), and use fewer emotion wordsto describe their emotional states (e.g., Luminetet al., 2004; Roedema & Simons, 1991) thando their nonalexithymic counterparts.Alexithymic individuals also demonstrate de-creased capacity for coping with emotion(Parker, Taylor, & Bagby, 1998), supportingthe idea that such individuals lack the complexconceptual system that contributes to success-ful regulatory strategies.

Even as the experience of emotion is dis-rupted, the experience of core affect (i.e., psy-chologically primitive states of pleasure anddispleasure; see discussion on p. 514) is not.Alexithymic and non-alexithymic individualsmake similar ratings when they report pleasureand displeasure in response to stimuli (Aftanas,Varlamov, Reva, & Pavlov, 2003; Berthozet al., 2002; McDonald & Prkachin, 1990).Alexithymic persons have difficulty describingtheir affective states as discrete instances ofemotion, but report experiencing many so-matic symptoms (e.g., Joergen Grabe, Spitzer,& Juergen Freyberger, 2004), consistent withreporting a basic experience of feeling “good”or “bad.” Some studies find that alexithymicindividuals actually experience more intensenegative affect (e.g., Friedlander, Lumley,Farchione, & Doyal, 1997), and enlist morebehaviors designed to reduce the experience ofnegative affect (e.g., Troisi et al., 2000), thando their nonalexithymic counterparts; thesefindings suggest that alexithymic individualsdo in fact have preserved experience of core af-fect.


Alexithymic individuals also have preserved(and sometimes greater) physiological arousalto evocative stimuli when compared to thosewithout alexithymia. For example, individualswith and without alexithymia did not differin skin conductance rate or heart rate whilewatching an unpleasant, highly arousing video(Stone & Nielson, 2001). Although they hadintact physiological reactions to the movie, thealexithymic participants reported less intensefeelings (disgust, concern, depression, surprise,fear, etc.) than the nonalexithymic participantsdid (Stone & Nielson, 2001). In some studies,individuals with alexithymia actually experi-ence greater physiological arousal in reactionto emotionally evocative stimuli (Byrne &Ditto, 2005; Infrasca, 1997; Luminet et al.,2004; Martínez-Sánchez, Ortiz-Soria, & Ato-García, 2001; Wehmer, Brejnak, Lumley, &Stettner, 1995).

Taken together, these findings suggest thatpeople with alexithymia feel core affect (again,defined as psychologically primitive pleasantand unpleasant states; Barrett, 2006a; Russell,2003), perhaps as somatic symptoms, but thatthey fail to translate this core affective stateinto a mental representation of emotion (i.e.,the experience of “having” an emotion) (seeBarrett, 2006b; Barrett, Mesquita, et al.,2007). Existing neuroscientific evidence is con-sistent with the idea that alexithymic individu-als suffer from a deficit in the ability to experi-ence their affective states as emotional. Studiesfind that during emotional experience, individ-uals with alexithymia show less activation inseveral areas within medial prefrontal cortex(MPFC) (e.g., Berthoz et al., 2002) and withinanterior cingulate cortex (ACC) (Berthoz et al.,2002; Lane et al., 1998) than individuals with-out alexithymia do. These areas seem to playan important function in representing core af-fective states as experiences of emotion (seeBarrett, Mesquita, et al., 2007; Wager et al.,Chapter 15, this volume).

Although the majority of studies find deficitsin alexithymic persons’ tendency to report dis-crete experiences of emotion, those that do notare particularly illuminating. Studies that usemore complex cues to induce emotion, such asvideo clips or imagery, tend to find evidence ofemotion experience deficits in alexithymic par-ticipants (e.g., Luminet et al., 2004; Mantani etal., 2005; Stone & Nielson, 2001; although seeAftanas & Varlamov, 2004), whereas studiesthat use less complex cues, such as pictures

(e.g., International Affective Picture System),do not (e.g., Aftanas et al., 2003; McDonald &Prkachin, 1990). Sufficiently simple cues mayfacilitate alexithymic individuals’ ability to ac-cess a rudimentary cache of emotion knowl-edge, and to make “normal” reports aboutemotion experience.

Deficits in Emotion Expression Associatedwith Alexithymia

People who suffer from alexithymia not onlyhave deficits in the experience of emotion, butalso exhibit low levels of spontaneous expres-sive behavior. Alexithymic individuals are typi-cally rated as less expressive during social inter-actions (e.g., Luminet et al., 2004) and morehumorless during clinical interviews (Lumley etal., 2005) than their nonalexithymic counter-parts are. In one study, individuals with alexi-thymia produced significantly less intense andmore ambiguous spontaneous emotional facialbehaviors when viewing unpleasant stimulithan those without alexithymia did, althoughboth groups were equally capable of posingemotional expressions when prompted (Mc-Donald & Prkachin, 1990).

This decrease in emotional expressivity oc-curs alongside greater somatic manifestationsof emotion. Alexithymic persons report greaterbodily concerns and somatic complaints (e.g.,Nakao, Barsky, Kumano, & Kuboki, 2002;Taylor, Parker, Bagby, & Acklin, 1992; for ameta-analytic review, see de Gucht & Heiser,2003). They have also been found to have moretension headaches (e.g., Yücel et al., 2002), andto demonstrate higher levels of hypertension(e.g., Todarello, Taylor, Parker, & Fanelli,1995) and cortisol reactivity during stress (e.g.,Lindholm, Lehtinen, Hyyppa, & Puukka,1990). With an impoverished conceptual sys-tem for emotion, alexithymic individuals maybe more likely to experience free-floating affectas somatic. This would explain why the linkbetween alexithymia and somatic complaintsdisappears when factors such as trait anxiety,depression, or experience of negative affect arecontrolled for (de Gucht, Fischler, & Heiser,2004; Lundh & Simonsson-Sarnecki, 2001).

Deficits in Emotion Perception Associatedwith Alexithymia

Individuals suffering from alexithymia havedifficulty not only with perceiving their own af-


fective states as emotion, but also with perceiv-ing emotion depicted in other people’s faces(e.g., Lane et al., 1995, 1996, 2000; Parker,Taylor, & Bagby, 1993; Vermeulen, Luminet,& Corneille, 2006). This effect extends to pro-cessing emotional content in other stimuli, suchas sentences about emotion, pictures depictingemotional situations (Lane et al., 1996, 2000),and words that are emotional in content(Luminet, Vermeulen, Demaret, Taylor, &Bagby, 2006; Suslow & Junghanns, 2002).Neuroscientific evidence supports the idea thatalexithymic individuals have difficulty concep-tualizing external cues such as faces and wordsas emotional. Alexithymic as compared tononalexithymic individuals exhibited decreasesin brain activity in a number of brain regions—such as right inferior frontal gyrus (IFG)(Brodmann’s area [BA] 44–45), orbito-frontalcortex (OFC) (BA 11), middle frontal gyrus(MFG) (BA 9), anterior cingulate cortex (ACC)(BA 24), and cerebellum—while viewing emo-tional facial behaviors (Kano et al., 2003).Meta-analyses of the imaging literature suggestthat these regions are consistently implicated intasks involving the perception and categoriza-tion of emotion (see Wager et al., Chapter 15,this volume).

SELF-CHARACTERIZATIONSOF COMPLEXITY

Thus far, we have examined emotional com-plexity as displayed in self-report behaviors(where individuals verbally represent their ex-periences of emotion) and in propositional con-ceptual knowledge of emotion (where indi-viduals make explicit, propositional statementsabout emotion). In both domains, complexityis assessed using performance measures: scien-tists abstract information about a person’semotional complexity from patterns of re-sponses or the ability to perform emotionallycomplex operations. The third and final way ofconceptualizing emotional complexity is asself-characterization, where respondents areasked to describe their own degree of emo-tional complexity.

A number of existing scales tap an individ-ual’s beliefs about his or her own emotionalcomplexity. Scales typically ask individuals tocharacterize how aware they are of their ownaffective states (e.g., the Mood AwarenessScale; Swinkels & Giuliano, 1995), as well as

the degree to which they attend to, can distin-guish between, and are capable of repairingthose states (e.g., the Trait Meta-Mood Scale;Mayer & Stevens, 1994). Perhaps the best ex-ample of a self-characterization measure ofcomplexity is the Range and Differentiation ofEmotional Experience Scale (RDEES; Kang &Shaver, 2005), which assesses the degree towhich people believe that they experience abroad range of emotional states with subtle dis-tinctions between them. Recently, self-characterization scales of emotional complex-ity have been developed for use in childsamples (e.g., the Emotion AwarenessQuestionnaire; Rieffe et al., 2007; theAlexithymia Questionnaire for children;Rieffe et al., 2006). The various self-characterization measures sometimes correlatestrongly with one another (e.g., Gohm &Clore, 2000), but often do not (e.g., Kang &Shaver, 2005).

More importantly, studies comparing self-characterizations and performance measures ofemotional complexity reveal a disconnectionbetween the two. For example, the RDEES isonly moderately correlated with the LEAS (r =.30) and the Toronto Alexithymia Scale (Bagby,Parker, & Taylor, 1994) (r’s ranged from .36 to.38 across three studies; Kang & Shaver, 2005).The RDEES also failed to correlate with indicesof emotional granularity computed over a 1-month period (Barrett, unpublished data). Itwould be tempting to assume that people’s be-liefs about themselves do not match theirbehavior, were it not for the fact that the per-formance measures themselves often fail to cor-relate. Granularity is essentially unrelated toperformance on the LEAS (Barrett, 2001).Taken together, these findings suggest thatemotional complexity may not be a single, ho-mogeneous construct, but rather a heteroge-neous construct encompassing many differentaspects of complexity.

The link between greater conceptual com-plexity and complexity in people’s self-characterizations of emotion experience is atpresent speculative, but we suggest that thereis good reason to assume the existence ofsuch a link. When people complete self-characterization scales of emotional complex-ity, they are typically asked to rate the fre-quency, intensity, or differentiation with whichthey typically experience discrete emotions. Todescribe themselves on such questionnaires, re-spondents must remember, summarize, and


integrate their past experiences into a consis-tent set of responses to the questionnaire items.Recalling information is a reconstructive pro-cess, however. Those people who have a greaterstore of emotion knowledge to draw from mayfind it easier to construct a response during theself-report process. In essence, people who pos-sess more complex conceptual systems foremotion should exhibit greater fluency of emo-tion knowledge during the self-characterizationprocess, resulting in a higher degree of self-characterized complexity.

OUTSTANDING ISSUES

Measurement and Conceptualizationof Emotional Complexity

Emotional complexity is a broad and variedconstruct. In this review, we have seen thatemotional complexity has been conceptualizedas dialecticism or emotional granularity in theexperience of emotion, as the ability to usemultiple emotion adjectives in propositional re-sponses to emotion-evoking scenarios, and asself-characterizations of complexity. If thesedifferent forms of complexity really tap a com-mon construct, then we would expect them tobe strongly correlated, such that one type ofcomplexity could stand in for another. Yet thistype of coherence does not describe the state ofthe literature. One possibility is that the vari-ous forms of complexity fail to cohere as a sin-gle construct because the measures that havebeen used to assess them are flawed.

A second, perhaps more likely possibility isthat the measures are causal indicators of emo-tional complexity, so that complexity is a con-struct that emerges from its measured parts. Incausal indicator models, a latent construct (inthis case, emotional complexity) is a linearcombination of its essentially uncorrelatedcauses (or measures) (Bollen & Lennox, 1991).The validity of the emergent construct cannotbe judged on the basis of covariation among itsindicators, but is instead determined by itsability to predict an externally measuredcriterion—such as well-being, interpersonaladaptability, emotion regulation, prosocialbehavior, or perhaps another person’s percep-tion of the target person’s emotional complex-ity in a given instant.

A third possibility is that various conceptual-izations of complexity are causally relatedunder certain (but perhaps not all) conditions.

We have suggested that richness and detail inthe conceptual system for emotion may serve acommon function in the various forms of emo-tional complexity. This perspective calls for amore detailed, idiographically sensitive map-ping of the conceptual system for emotion.Studies assessing the nature of the conceptualsystem for emotion would determine whichemotion categories people know and use, whatthe content of those categories are, and howthat content is represented.

Variability in Emotional Complexity

A more comprehensive examination of sex anddevelopmental differences in the various facetsof emotion complexity is warranted. Clear sexdifferences exist in LEAS performance (Barrettet al., 2000) and in self-characterizations ofcomplexity (see Gohm & Clore, 2000, for a re-view), but there are no consistent sex differ-ences in emotional granularity. To gain an un-derstanding of sex differences in complexity,future research must flesh out when and whysex differences appear. Mothers make more fre-quent references to emotion when discussingemotional memories with their daughters thanthey do with their sons (Adams, Keubli, Boyle,& Fivush, 1995), and they also discuss the in-terpersonal content and connotations of nega-tive emotional memories more with daughters(Fivush et al., 2003). This socialization processshould have implications for the richness ofchildren’s conceptual systems for emotion.There are also developmental changes in emo-tional complexity (indicated by age differencesin LEAS-C performance, differences in self-reported complexity in children, and impliedby observed increases in facility with emotionperception in early childhood). Little is knownhowever, about how these differences might ex-tend to the experience of emotion. Children’sability to precisely identify other people’s emo-tional behavior mirrors the stages of concep-tual development (see Widen & Russell, Chap-ter 21, this volume), and as we have notedpreviously, it is possible that those with morewell-defined conceptual systems may more pre-cisely represent their own experiences of emo-tion. At the other end of the lifespan, develop-mental differences also exist. Dialecticismincreases in old age, and more recent workdemonstrates age differences in the organiza-tion of positive and negative information inmemory, indicating that older adults organize


and represent emotional memories in a moredialectic fashion than younger adults do(Ready, Robinson, & Weinberger, 2006). Theremay also be age-related increases in granular-ity, although this hypothesis has not been ex-plicitly tested. Carstensen et al. (2000) foundthat older adults used emotion words in a moredifferentiated manner to represent online expe-rience, whereas younger adults used emotionwords in a less differentiated manner.

Finally, person factors such as WMC may in-fluence emotional complexity, because greaterWMC facilitates the creation of a complex con-ceptual system for emotion and allows formore efficient use of that system. WMC influ-ences the construction of mental representa-tions that support new learning (Cantor &Engle, 1993) and shapes how individuals usealready existing information to support the en-coding of new knowledge (Daily, Lovett, &Reder, 2001; Hambrick & Engle, 2002; but forevidence of additive effects, see Rukavina &Daneman, 1996). Thus, all other factors beingequal, individuals higher in WMC may havemore complex or idiographic detail in theirconceptual systems for emotion, merely be-cause they are better at learning and storingcategory content. During online experience,WMC may affect how individuals wield con-ceptual knowledge because it facilitates easier,more efficient access to conceptual knowledgeof emotion (see Barrett et al., 2004), allowingindividuals to make more specific, situationallytuned categorizations of their affective experi-ences as they occur.

CONCLUSION

Regardless of how emotional complexity isconceptualized, one thing is clear: Emotionalcomplexity is advantageous. Greater dialecti-cism, for example, is associated with greater re-silience and lower stress in Western contexts(Davis et al., 2004; Ong & Bergeman, 2004),particularly in older individuals (Carstensen etal., 2000; Charles, 2005). Greater emotionalgranularity confers more frequent and flexibleemotion regulation (Barrett et al., 2001;Tugade, Fredrickson, & Barrett, 2004; Tugadeet al., 2007). Complexity in propositionalknowledge of emotion is also loosely associ-ated with greater psychological well-being, asindicated by the fact that complexity in propo-sitional knowledge of emotion increases after

psychodynamic treatment in a patient popula-tion (Subic-Wrana, Bruder, Thomas, Lane, &Köehle, 2005). Children who are better at iden-tifying and verbalizing their emotions experi-ence less worry and depression than thosechildren who have difficulty distinguishing orcommunicating their emotions (e.g., Rieffe etal., 2007). Finally, self-characterized complex-ity is related to greater interpersonal adaptabil-ity (e.g., Kang & Shaver, 2005). It would ap-pear that greater emotional complexity confersgreater capacity to navigate and cope with theemotional world.

ACKNOWLEDGMENTS

This work was supported by a National ScienceFoundation (NSF) Graduate Research Fellowship toKristen A. Lindquist; by NSF Grants (No. BCS9727896 and No. SES 0074688); and by a NationalInstitute of Mental Health Independent Scientist Re-search Award (K02 MH001981) to Lisa FeldmanBarrett.

NOTES

1. The absence of dialecticism has also been defined as azero correlation (e.g., Scollon et al., 2005).

2. Emotional granularity is not defined in terms of va-lidity, because it is currently not possible to verify ob-jectively that a certain emotional event is present orabsent (i.e., there is no empirically justifiable accu-racy criterion that is independent of an observer;Barrett, 2006b). It might be possible to examinewhether observers’ perceptions of emotion agreewith persons’ ratings of their own experiences, butestimates of self–other agreement address a differentquestion.

REFERENCES

Adams, S., Keubli, J., Boyle, P. A. & Fivush, R. (1995).Gender differences in parent–child conversationsabout past emotions: A longitudinal investigation.Sex Roles, 33, 309–324.

Aftanas, L. I., & Varlamov, A. A. (2004). Associationsof alexithymia with anterior and posterior activationasymmetries during evoked emotions: EEG evidenceof right hemisphere “electrocortical effort.” Interna-tional Journal of Neuroscience, 114, 1443–1462.

Aftanas, L. I., Varlamov, A. A., Reva, N. V., & Pavlov, S.V. (2003). Disruption of early event-related thetasynchronization of human EEG in alexithymics view-ing affective pictures. Neuroscience Letters, 340, 57–60.

Alonso-Arbiol, I., Shaver, P. R., Fraley, C., Oronoz, B.,


Unzurrunzaga, E., Ruben, U. (2006). Structure of theBasque emotion lexicon. Cognition and Emotion, 20,836–865.

Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (1994).The twenty-item Toronto Alexithymia Scale: I. Itemselection and cross-validation of the factor structure.Journal of Psychosomatic Research, 38, 23–32.

Bagozzi, R. P., Wong, N., & Yi, Y. (1999). The role ofculture and gender in the relationship between posi-tive and negative affect. Cognition and Emotion, 13,641–672.

Bajgar, J., Ciarrochi, J., Lane, R., & Deane, F. P. (2005).Development of the Levels of Emotional AwarenessScale for Children (LEAS-C). British Journal of De-velopmental Psychology, 3, 569–586.

Barrett, L. F. (1997). The relationship among momen-tary emotion experiences, personality descriptionsand retrospective ratings of emotion. Personality andSocial Psychology Bulletin, 23, 1100–1114.

Barrett, L. F. (1998). Discrete emotions or dimensions?:The role of valence focus and arousal focus. Cogni-tion and Emotion, 12(4), 579–599.

Barrett, L. F. (2001). [Granularity and Emotional Com-plexity]. Unpublished raw data.

Barrett, L. F. (2004). Feelings or words?: Understandingthe content in self-report ratings of experienced emo-tion. Journal of Personality and Social Psychology,87, 266–281.

Barrett, L. F. (2006a). Emotions as natural kinds? Per-spectives on Psychological Science, 1, 28–58.

Barrett, L. F. (2006b). Solving the emotion paradox:Categorization and the experience of emotion. Per-sonality and Social Psychology Review, 10, 20–46.

Barrett, L. F., Gross, J., Christensen, T. C., &Benvenuto, M. (2001). Knowing what you’re feelingand knowing what to do about it: Mapping the rela-tion between emotion differentiation and emotionregulation. Cognition and Emotion, 15(6), 713–724.

Barrett, L. F., Lane, R., Sechrest, L., & Schwartz, G.(2000). Sex differences in emotional awareness. Per-sonality and Social Psychology Bulletin, 26, 1027–1035.

Barrett, L. F., Lindquist, K. A., Bliss-Moreau, E.,Duncan, S., Gendron, M., Mize, J., et al. (2007). Ofmice and men: Natural kinds of emotions in themammalian brain? A response to Panksepp andIzard. Perspectives on Psychological Science, 2, 297–312.

Barrett, L. F., Mesquita, B., Ochsner, K. N., & Gross, J.J. (2007). The experience of emotion. Annual Reviewof Psychology, 58, 373–403.

Barrett, L. F., Robin, L., Pietromonaco, P. R., & Eyssell,K. M. (1998). Are women the “more emotionalsex”?: Evidence from emotional experiences in socialcontext. Cognition and Emotion, 12, 555–578.

Barrett, L. F., Tugade, M. M., & Engle, R. W. (2004). In-dividual differences in working memory capacity anddual-process theories of the mind. Psychological Bul-letin, 130, 553–573.

Barsalou, L. W., Santos, A., Simmons, W. K., & Wilson,C. D. (in press). Language and simulation in concep-tual processing. In M. de Vega, A. M. Glenberg, & A.C. Graesser (Eds.), Symbols, embodiment, and mean-ing. Oxford: Oxford University Press.

Berenbaum, H. H., & James, T. T. (1994). Correlatesand retrospectively reported antecedents of alexi-thymia. Psychosomatic Medicine, 56, 353–359.

Berthoz, S., Artiges, E., Van De Moortele, P. F., Poline, J.B., Rouquette, S., Consoli, S. M., et al. (2002). Effectof impaired recognition and expression of emotionson frontocingulate cortices: An fMRI study of menwith alexithymia. The American Journal of Psychia-try, 159, 961–967.

Biederman, I., & Shiffrar, M. M. (1987). Sexing day-oldchicks: A case study and expert systems analysis of adifficult perceptual-learning task. Journal of Experi-mental Psychology: Learning, Memory and Cogni-tion, 13, 640–645.

Bliss-Moreau, E., Barrett, L., Connor, T., & McCarthy,K. (2007). Are men really from Mars? Understandingsex differences in emotional awareness. Manuscriptin preparation.

Bollen, K., & Lennox, R. (1991). Conventional wisdomon measurement: A structural equation perspective.Psychological Bulletin, 110, 305–314.

Brown, R. W. (1957). Linguistic determinism and thepart of speech. Journal of Abnormal and Social Psy-chology, 55, 1–5.

Byrne, N., & Ditto, B.. (2005). Alexithymia, cardiovas-cular reactivity, and symptom reporting during blooddonation. Psychosomatic Medicine, 67, 471–475.

Cantor, J. J., & Engle, R. W. (1993). Working-memorycapacity as long-term memory activation: Anindividual-differences approach. Journal of Experi-mental Psychology: Learning, Memory, and Cogni-tion, 19, 1101–1014.

Carstensen, L. L., Pasupathi, M., Mayr, U., &Nesselroade, J. R. (2000). Emotional experience ineveryday life across the adult life span. Journal ofPersonality and Social Psychology, 79, 644–655.

Charles, S. T. (2005). Viewing injustice: Age differencesin emotional experience. Psychology and Aging, 20,159–164.

Christensen, E. E., Murry, R. C., Holland, K., Reynolds,J., Landay, M. J., & Moore, J. G. (1981). The effectof search time on perception. Radiology, 138, 361–365.


Conner, T., Barrett, L. F., Bliss-Moreau, E., Lebo, K., &Kashub, C. (2003). A practical guide to experience-sampling procedures. Journal of Happiness Studies,4, 53–78.

Conway, M. (2000). On sex roles and representations ofemotional experience: Masculinity, femininity, andemotional awareness. Sex Roles, 43, 687–698.

Daily, L. Z., Lovett, M. C., & Reder, L. M. (2001).


Modeling individual differences in working memoryperformance: A source activation account. CognitiveSciences, 25, 315–353.

Davis, M. C., Zautra, A. J., & Smith, B. W. (2004).Chronic pain, stress, and the dynamics of affectivedifferentiation. Journal of Personality, 72, 133–1159.

de Gelder, B., Vroomen, J., Pourtois, G., & Weiskrantz,L. (1999). Non-conscious recognition of affect in theabsence of striate cortex. NeuroReport, 10, 3759–3763.

de Gucht, V., Fischler, B., & Heiser, W. (2004). Neuroti-cism, alexithymia, negative affect, and positive affectas determinants of medically unexplained symptoms.Personality and Individual Differences, 36, 1655–1667.

de Gucht, V., & Heiser, W. (2003). Alexithymia andsomatisation: quantitative review of the literature.Journal of Psychosomatic Research, 54, 425–434.

de Rosnay, M., & Harris, P. L. (2002). Individual differ-ences in children’s understanding of emotion: Theroles of attachment and language. Attachment andHuman Development, 4, 39–54.

de Rosnay, M., Pons, F., Harris, P., & Morrell, J. (2004).A lag between understanding false belief and emotionattribution in young children: Relationships with lin-guistic ability and mothers’ mental-state languageBritish Journal of Developmental Psychology, 22,197–218.

Diener, E., & Emmons, R. A. (1985). The independenceof positive and negative affect. Journal of Personalityand Social Psychology, 47, 1105–1117.

Edelman, G. M. (1989). The remembered present: A bi-ological theory of consciousness. New York: BasicBooks.

Ekman, P. (1972). Universals and cultural differences infacial expressions of emotion. 1971. In J. R. Cole(Ed.), Nebraska Symposium on Motivation (Vol. 19,pp. 207–283). Lincoln: University of Nebraska Press.

Fehr, B., & Russell, J. A. (1984). Concept of emotionviewed from a prototype perspective. Journal of Ex-perimental Psychology: General, 113, 464–486.

Feldman, L. A. (1995). Valence focus and arousal focus:Individual differences in the structure of affective ex-perience. Journal of Personality and Social Psychol-ogy, 69, 153–166.

Fivush, R., Berlin, L. J., Sales, J. M., Mennuti-Washburn, J., & Cassidy, J. (2003). Functions ofparent–child reminiscing about emotionally negativeevents. Memory, 11, 179–192.

Frawley, W., & Smith, R. N. (2001). A processing the-ory of alexithymia. Journal of Cognitive Systems Re-search, 2, 189–206.

Friedlander, L., Lumley, M. A., Farchione, T., & Doyal,G. (1997). Testing the alexithymia hypothesis: Physi-ological and subjective responses during relaxationand stress. Journal of Nervous and Mental Disorders,185, 233–239.

Gohm, C. L., & Clore, G. L. (2000). Individual differ-

ences in emotional experience: Mapping availablescales to processes. Personality and Social Psychol-ogy Bulletin, 26, 679–697.

Hambrick, D. Z., & Engle, R. W. (2002). Effects of do-main knowledge, working memory capacity, and ageon cognitive performance: An investigation of theknowledge-is-power hypothesis. Cognitive Psychol-ogy, 44, 339–387.

Harris, P. L. (2006a). Use your words. British Journal ofDevelopmental Psychology, 24, 253–261.

Harris, P. L. (2006b). It’s probably good to talk.Merrill–Palmer Quarterly, 52, 158–169.

Harris, P. L., de Rosnay, M., & Pons, F. (2005). Lan-guage and children’s understanding of mental states.Current Directions in Psychological Science, 14, 69–73.

Haviland, M. G., & Reise, S. P. (1996). A California Q-set alexithymia prototype and its relationship to ego-control and ego-resiliency. Journal of PsychosomaticResearch, 41, 597–407.

Heine, S. J., Lehman, D. R., Markus, H. R., &Kitayama, S. (1999). Is there a universal need forpositive self-regard? Psychological Review, 106,766–794.

Infrasca, R. R. (1997). Alexithymia, neurovegetativearousal and neuroticism: An experimental study. Psy-chotherapy and Psychosomatics, 66, 276–280.

Izard, C. E. (1994). Innate and universal facial expres-sions: Evidence from developmental and cross-cultural research. Psychological Bulletin, 115, 288–299.

Joergen Grabe, H., Spitzer, C., & Juergen Freyberger, H.(2004). Alexithymia and personality in relation to di-mensions of psychopathology. American Journal ofPsychiatry, 161, 1299–1301.

Johnson-Laird, P. N., & Oatley, K. (1989). The lan-guage of emotions: An analysis of a semantic field.Cognition and Emotion, 3, 81–123.

Kang, S., & Shaver, P. R. (2005). Individual differencesin emotional complexity: Their psychological impli-cations. Journal of Personality, 72, 687–726.

Kano, M., Fukudo, S., Gyoba, J., Kamachi, M.,Tagawa, M., Mochizuki, H., et al. (2003). Specificbrain processing of facial expressions in people withalexithymia: An H2 15O-PET study. Brain, 126,1474–1484.

Kitayama, S., Markus, H. R., & Kurokawa, M. (2000).Culture, emotion, and well-being: Good feelings inJapan and the United States. Cognition & Emotion,14, 93–124.

Lane, R. D. (2000). Neural correlates of conscious emo-tional experience. In R. D. Lane & L. Nadel (Eds.),Cognitive neuroscience of emotion (pp. 345–370).New York: Oxford University Press.

Lane, R. D., Ahern, G. L., Schwartz, G. E., & Kaszniak,A. W. (1997). Is alexithymia the emotional equiva-lent of blindsight? Biological Psychiatry, 42, 834–844.

Lane, R. D., & Garfield, D. A. S. (2005). Becoming


aware of feelings: Integration of cognitive-developmental, neuroscientific, and psychoanalyticperspectives. Neuro-Psychoanalysis, 7, 5–30.

Lane, R. D., Kivley, L. S., Du Bois, M. A.,Shamasundara, P., & Schwartz, G. E. (1995). Levelsof emotional awareness and the degree of right hemi-spheric dominance in the perception of facial emo-tion. Neuropsychologia, 33, 525–538.

Lane, R. D., & Pollermann, B. Z. (2002). Complexity ofemotion representations. In L. F. Barrett & P. Salovey(Eds.), The wisdom in feeling: Psychological pro-cesses in emotional intelligence (pp. 271–293). NewYork: Guilford Press.

Lane, R. D., Quinlan, D. M., Schwartz, G. E., Walker, P.A., & Zeitlin, S. B. (1990). The Levels of EmotionalAwareness Scale: A cognitive-developmental measureof emotion. Journal of Personality Assessment, 55,124–134.

Lane, R. D., & Schwartz, G. E. (1987). Levels of emo-tional awareness: A cognitive-developmental theoryand its application to psychopathology. AmericanJournal of Psychiatry, 144, 133–143.

Lane, R. D., Sechrest, L., & Riedel, R. (1998).Sociodemographic correlates of alexithymia. Com-parative Psychiatry, 39, 377–385.

Lane, R. D., Sechrest, L., Riedel, R., Shapiro, D. E., &Kaszniak, A. W. (2000). Pervasive emotion recogni-tion deficit common to alexithymia and the repres-sive coping style. Psychosomatic Medicine, 62(4),492–501.

Lane, R. D., Sechrest, L., Reidel, R. G., Weldon, V.,Kaszniak, A. W., & Schwartz, G. E. (1996). Im-paired verbal and nonverbal emotion recognition inalexithymia. Psychosomatic Medicine, 58, 203–210.


Larsen, J. T., McGraw, A. P., Mellers, B. A., &Cacioppo, J. T. (2004). The agony of victory andthrill of defeat: Mixed emotional reactions to disap-pointing wins and relieving losses. Psychological Sci-ence, 15, 325–330.

Lee, Y. (2000). What is missing in Chinese–Western dia-lectical reasoning? American Psychologist, 55, 1065–1067.

Leu, J., Mesquita, B., Ellsworth, P. C., Zhiyong, Z.,Huijian, Y., Buchtel, E., et al. (2007). Cultural mod-els of emotion regulation in the East and West: “Dia-lectical” emotions vs. the pursuit of pleasantness.Manuscript in preparation.

Lindholm, T., Lehtinen, V., Hyyppa, M. T., & Puukka,P. (1990). Alexithymic features in relation to thedexamethasone suppression test in a Finnish popula-tion sample. American Journal of Psychiatry, 147,1216–1219.

Luminet, O., Rimé, B., Bagby, R. M., & Taylor, G. J.(2004). A multimodal investigation of emotional re-

sponding in alexithymia. Cognition and Emotion,18, 741–766.

Luminet, O., Vermeulen, N., Demaret, C., Taylor, G. J.,& Bagby, R. M. (2006). Alexithymia and levels ofprocessing: Evidence for an overall deficit in remem-bering emotion words. Journal of Research in Per-sonality, 40, 713–733.

Lumley, M. A., Gustavson, B. J., Partridge, R. T., &Labouvie-Vief, G. (2005). Assessing alexithymia andrelated emotional ability constructs using multiplemethods: Interrelationships among measures. Emo-tion, 5, 329–342.

Lundh, L., & Simonsson-Sarnecki, M. M. (2001).Alexithymia, emotion, and somatic complaints. Jour-nal of Personality, 69, 483–510.

Mandler, G. (1975). Mind and emotion. New York:Wiley

Mantani, T., Okamoto, Y., Shirao, N., Okada, G., &Yamawaki, S. (2005). Reduced activation of poste-rior cingulate cortex during imagery in subjects withhigh degrees of alexithymia: A functional magneticresonance imaging study. Biological Psychiatry, 57,982–990.

Martínez-Sánchez, F., Ortiz-Soria, B., & Ato-García, M.(2001). Subjective and autonomic stress responses inalexithymia. Psicothema, 13, 57–62.

Mayer, J. D., & Stevens, A. A. (1994). An emerging un-derstanding of the reflective (meta-)experience ofmood. Journal of Research in Personality, 28, 351–373.

McDonald, P. W., & Prkachin, K. M. (1990). The ex-pression and perception of facial emotion inalexithymia: A pilot study. Psychosomatic Medicine,52, 199–210.

Miller, P. J., & Sperry, L. L. (1988). Early talk about thepast: The origins of conversational stories of personalexperience. Journal of Child Language, 15, 293–315.

Murphy, G. L. (2002). The big book of concepts. Cam-bridge, MA: MIT Press.

Nakao, M., Barsky, A. J., Kumano, H., & Kuboki, T.(2002). Relationship between somatosensory ampli-fication and alexithymia in a Japanese psychosomaticclinic. Psychosomatics, 43, 55–60.

Oishi, S. (2002). The experiencing and remembering ofwell-being: A cross-cultural analysis. Personality andSocial Psychology Bulletin, 28, 1398–1406.

Ong, A. D., & Bergeman, C. S. (2004). The complexityof emotions in later life. Journals of Gerontology: Se-ries B: Psychological Sciences and Social Sciences, 59,117–122.

Ortony, A., Clore, G. L., & Foss, M. A. (1987). The psy-chological foundations of the affective lexicon. Jour-nal of Personality and Social Psychology, 53, 751–766.

Panksepp, J. (2000). Emotions as natural kinds withinthe mammalian brain. In M. Lewis & J. M.Haviland-Jones (Eds.), Handbook of emotions (2nded., pp. 137–156)). New York: Guilford Press.

Parker, J. D., Taylor, G. J., & Bagby, R. M. (1993).


Alexithymia and the recognition of facial expressionsof emotion. Psychotherapy and Psychosomatics, 59,197–202.

Parker, J. D., Taylor, G. J., & Bagby, R. M. (1998).Alexithymia: Relationship with ego defense and cop-ing styles. Comprehensive Psychiatry, 39, 91–98.

Parker, J. D., Taylor, G. J., & Bagby, R. M. (2001). Therelationship between emotional intelligence andalexithymia. Personality and Individual Differences,30, 107–115.

Peng, K., & Nisbett, R. E. (1999). Culture, dialecticsand reasoning about contradiction. American Psy-chologist, 54, 741–754.

Pons, F., & Harris, P. L. (2005). Longitudinal changeand longitudinal stability of individual differences inchildren’s emotion understanding. Cognition andEmotion, 19, 1158–1174.

Pons, F., Lawson, J., Harris, P. L., & de Rosnay, M.(2003). Individual differences in children’s emotionunderstanding: Effects of age and language. Scandi-navian Journal of Psychology, 44, 347–353.

Rau, J. C. (1993). Perception of verbal and nonverbalaffective stimuli in complex partial seizure disorder.Dissertation Abstracts International, 54, 506B.

Ready, R. E., Robinson, M. D., & Weinberger, M.(2006). Age differences in the organization of emo-tion knowledge: Effects involving valence and timeframe. Psychology and Aging, 21, 726–736.

Reich, J. W., Zautra, A. J., & Potter, P. T. (2001). Cogni-tive structure and the independence of positive andnegative affect. Journal of Social and Clinical Psy-chology, 20, 99–115.

Rieffe, C., Meerum Terwogt, M., Petrides, K. V., Cow-an, R., Miers, A. C., & Tolland, A. (2007). Psycho-metric properties of the Emotion Awareness Ques-tionnaire for children. Personality and IndividualDifferences, 43, 95–105.

Rieffe, C., Oosterveld, P., & Meerum Terwogt, M.(2006). An Alexithymia Questionnaire for children:Factorial and concurrent validation results. Personal-ity and Individual Differences, 40, 123–133.

Robinson, M. D., & Clore, G. L. (2002). Belief andfeeling: Evidence for an accessibility model of emo-tional self-report. Psychological Bulletin, 128, 934–960.

Roedema, T. M., & Simons, R. F. (1999). Emotion-processing deficit in alexithymia. Psychophysiology,36, 379–387.

Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M.,& Boyes-Braem, P. (1976). Basic objects in naturalcategories. Cognitive Psychology, 8, 382–439.

Roseman, I. J. (1984). Cognitive determinants of emo-tion: A structural theory. Review of Personality andSocial Psychology, 5, 11–36.

Ross, M. (1989). Relation of implicit theories to theconstruction of personal histories. Psychological Re-view, 96, 341–357.

Rukavina, I., & Daneman, M. (1996). Integration andits effect on acquiring knowledge about competing

scientific theories for text. Journal of EducationalPsychology, 88, 272–287.

Russell, J. A. (1991). Culture and the categorization ofemotions. Psychological Bulletin, 110, 426–450.




Scollon, N. C. Diener, E., Oishi, S., & Biswas-Diener, R.(2005). An experience sampling and cross cultural in-vestigation of the relation between pleasant and un-pleasant affect. Cognition and Emotion, 19, 27–52.

Shaver, P. R., Murdaya, U., & Fraley, C. (2001). Struc-ture of the Indonesian emotion lexicon. Asian Jour-nal of Social Psychology, 4, 201–224.

Shaver, P. R., Schwartz, J., Kirson, D., & O’Connor, C.(1987). Emotion knowledge: Further exploration ofa prototype approach. Journal of Personality and So-cial Psychology, 52, 1061–1086.

Shimmack, U., Oishi, S., & Diener, E. (2002). Culturalinfluences on the relation between pleasant emotionsand unpleasant emotions: Asian dialectic philoso-phies or individualism–collectivism? Cognition andEmotion, 16, 705–719.

Sifneos, P. E. (1973). The prevalence of ‘alexithymic’characteristics in psychosomatic patients. Psych-therapy and Psychosomatics, 22, 255–262.

Solomon, G. E. A. (1990). Psychology of novice and ex-pert wine talk. American Journal of Psychology, 103,495–517.

Solomon, G. E. A. (1997). Conceptual change and wineexpertise. Journal of the Learning Sciences, 6, 41–60.

Stone, L. A., & Nielson, K. A. (2001). Intact physiologi-cal response to arousal with impaired emotional rec-ognition in alexithymia. Pscyhotherapy and Psycho-somatics, 70, 92–102.

Subic-Wrana, C., Bruder, S., Thomas, W., Lane, R. D.,& Köhle, K. (2005). Emotional awareness deficits ininpatients of a psychosomatic ward: a comparison oftwo different measures of alexithymia. Psychoso-matic Medicine, 67, 483–489.

Suslow, T., & Junghanns, K. (2002). Impairments ofemotion situation priming in alexithymia. Personal-ity and Individual Differences, 32, 541–550.

Swinkels, A., & Guiliano, T. A. (1995). The measure-ment and conceptualization of mood awareness:Monitoring and labeling one’s mood states. Personal-ity and Social Psychology Bulletin, 21, 934–949.

Taylor, G. J. (2000). Recent developments in alexi-thymia theory and research. Canadian Journal ofPsychiatry, 45, 134–142.

Taylor, G. J., Bagby, R. M., & Parker, J. D. (1997). Dis-orders of affect regulation: Alexithymia in medical


and psychiatric illness. Cambridge, UK: CambridgeUniversity Press.

Taylor, G., Bagby, R. M., Ryan, D. P., Parker, J. D.,Doody, K. F., & Keefe, P. P. (1988). Criterion validityof the Toronto Alexithymia Scale. PsychosomaticMedicine, 50, 500–509.

Taylor, G. J., Parker, J. D., Bagby, R. M., & Acklin, M.W. (1992). Alexithymia and somatic complaints inpsychiatric out-patients. Journal of PsychosomaticResearch, 36, 417–424.

Todarello, O. O., Taylor, G. J., Parker, J. D., &Fanelli, M. M. (1995). Alexithymia in essentialhypertensive and psychiatric outpatients: A com-parative study. Journal of Psychosomatic Research,39, 987–994.

Tomkins, S. S. (1962). Affect, imagery, consciousness:Vol. 1. The positive affects. New York: Springer.

Troisi, A., Belsanti, S., Bucci, A. R. Mosco, C., Sinti, F.,& Verucci, M. (2000). Affect regulation in alexi-thymia: An ethological study of displacement behav-ior during psychiatric interviews. Journal of Nervousand Mental Disease, 188, 13–18.

Tsai, J. L. (2007). Ideal affect: Cultural causes andbehavioral consequences. Perspectives on Psycholog-ical Science, 2, 242–259.

Tugade, M., Barrett, L. F., & Gross, J. (2007). Distinc-tions that matter: Emotional granularity and emo-tion regulation profiles. Manuscript submitted forpublication.

Tugade, M. M., Fredrickson, B. L., & Barrett, L. F.(2004). Psychological resilience and emotional gran-ularity: Examining the benefits of positive emotionson emotion regulation and health. Journal of Person-ality, 72, 1161–1190.

Vansteelandt, K., Van Mechelen, I., & Nezlek, J. B.(2005). The co-occurrence of emotions in daily life: Amultilevel approach. Journal of Research in Person-ality, 39, 325–335.

Vermeulen, N., Luminet, O., & Corneille, O. (2006).Alexithymia and the automatic processing of affec-tive information: Evidence from the affective primingparadigm. Cognition and Emotion, 20, 64–91.

Wehmer, F., Brejnak, C., Lumley, M., & Stettner, L.(1995). Alexithymia and physiological reactivity toemotion-provoking visual scenes. Journal of Nervousand Mental Disease, 183, 351–357.

Yücel, B., Kora, K., Ozyalçín, S., Alçalar, N., Ozdemir,O., & Yücel, A. (2002). Depression, automaticthoughts, alexithymia, and assertiveness in patientswith tension-type headache. Headache, 42, 194–199.


PA R T V I

COGNITIVE FACTORS

C H A P T E R 3 3

Emotional Intelligence

PETER SALOVEY, BRIAN T. DETWEILER-BEDELL, JERUSHA B.DETWEILER-BEDELL, and JOHN D. MAYER

Basic research in emotion has proliferated overthe past few decades, and although there is stilla great deal to be learned, a consistent view ofemotion has begun to emerge. Affective phe-nomena constitute a unique source of informa-tion for individuals about their surrounding en-vironment and prospects, and this informationinforms their thoughts, actions, and subse-quent feelings. The essential assumption in ourwork has been that individuals differ in howskilled they are at perceiving, understanding,regulating, and utilizing this emotional infor-mation, and that a person’s level of “emotionalintelligence” contributes substantially to his orher intellectual and emotional well-being andgrowth. In this chapter, we review our model ofemotional intelligence and the competencies ithighlights, including a discussion of measure-ment issues. We then describe how emotionalintelligence influences important psychologicalphenomena.

Personality psychology most often empha-sizes differences in the way people typicallythink, feel, and act. However, as Mischel(1990) has pointed out, people relate to the

world in a manner that is much more flexiblethan the terms “dispositions” and “styles”might alone suggest. According to Mischel,cognitive psychology’s essential lesson for per-sonality psychologists is that individuals selec-tively construct their experiences of reality, andthat the result of this process depends heavilyon individuals’ construction competencies (i.e.,information-processing abilities that determinethe range of potential thoughts, feelings, andbehaviors the individual can enlist within andacross situations). To understand the person,we must augment our study of dispositionswith an appreciation for these competencies.The investigation of such competencies, asidefrom general intelligence, has focused largelyon social problem-solving skills and other prac-tical abilities and has been referred to as “socialintelligence,” among other labels (Cantor &Kihlstrom, 1985, 1987; Gardner, 1983; Stern-berg, 1985, 1988; Sternberg & Smith, 1985;Thorndike, 1920; Thorndike & Stein, 1937;Wagner & Sternberg, 1985).

We believe that emotional competencies arefundamental to social intelligence (Mayer,

533

Caruso, & Salovey, 2004; Salovey & Mayer,1990). This is because social problems and sit-uations are laden with affective information.Moreover, emotional competencies apply notonly to social experiences, but also to experi-ences within the individual. Indeed, some in-vestigators have argued that self-knowledgeand the individual’s inner life are characterizedmost saliently by emotional experiences (e.g.,Showers & Kling, 1996). Thus emotional intel-ligence is more focused than social intelligence,in that it pertains specifically to emotional phe-nomena; yet it can be applied directly to abroad range of emotional problems embeddedin both interpersonal and intrapersonal experi-ence (Epstein, 1998; Mayer & Salovey, 1997;Saarni, 1990, 1997; Salovey & Mayer, 1990).It is this efficient, parsimonious nature of theemotional intelligence framework that we findso compelling. Indeed, we find it puzzling thatpsychology has taken so long to recognize theimportance of emotional competencies.

EMOTION: FROM DISINFORMATIONTO INFORMATION

Emotion, historically, has taken a back seat tocognition. Philosophers and scientists (psychol-ogists included) have relied on and glorified an-alytic intelligence throughout much of Westernhistory. Aristotle (384–322 B.C.E.), for exam-ple, argued that the human intellect is “thehighest thing in us, and the objects that it ap-prehends are the highest things that can beknown” (1976, p. 328; see also Aristotle,1947). At the same time, emotion has been re-garded as an inferior, often disruptive elementof human nature. The passions are fallibleguides for action. Anger often leads to unjustacts of violence; fear often leads to debilitatingcowardice. This sentiment toward feeling wasthe impetus for Malebranche’s unequivocalprescription: “Impose silence on your senses,your imagination, and your passions, and youwill hear the pure voice of inner truth” (quotedin James, 1890/1950, p. 10).

As a result of the historic mistrust of emo-tion, many psychologists have taken the posi-tion that the intellect and passions are at cross-purposes (e.g., Schaffer, Gilmer, & Schoen,1940; Woodworth, 1940). In this view, the in-tellect provides accurate information, whereasemotion clouds our minds with disinformation.Young (1936) even went so far as to say that

emotions have no conscious purpose and cause“a complete loss of cerebral control” (pp. 457–458). It is quite understandable, then, thatearly conceptions of intelligence in the field ofpsychology were decidedly rational. Termanstated, “An individual is intelligent in propor-tion as he is able to carry on abstract thinking”(1921, p. 128). Being emotional was not con-sidered smart.

Contemporary psychology has moved awayfrom this view that reason is superior to emo-tion, and toward an emphasis on the function-ality of emotions. This shift originated in thephilosophy of David Hume and the ethologicalobservations of Charles Darwin. Hume (1739/1948), an early 18th-century empiricist philos-opher, argued that emotional impulses moti-vate all action. He believed that reason doesnothing more than consider facts and generateinferences about the world relevant to achiev-ing and prioritizing the agendas set by the pas-sions. Freud (1923/1962) held a somewhatsimilar position. He emphasized the primacy ofthe id—the seat of the self’s emotionality andpsychic energy—and maintained that the otheraspects of the self are derivative. Freud, muchlike Hume, put reason in the service of emo-tion.

Although Hume challenged the position thatreason is superior to emotion, it was not untilDarwin published The Expression of the Emo-tions in Man and Animals (1872/1965) that thefunctional purpose of emotion was established.Through his intensive ethological observationsof animal life, Darwin revealed that emotionserves at least two highly advantageous func-tions. First, emotion energizes adaptive behav-iors such as flight (fear) and procreation (loveor lust). Second, emotion gives rise to a signal-ing and communication system that confers asignificant survival advantage on entire speciesas well as individual organisms (e.g., a singledeer’s fear response upon seeing a predatorquickly informs other deer of the threat). By at-tributing these functions to emotion, Darwinbrought attention to the adaptive, seeminglyintelligent nature of emotional expression.

Today great emphasis is placed on the psy-chological importance of emotion, and it isgenerally accepted that emotions augmentrather than interfere with other cognitive ca-pacities. Emotions certainly have the signalingfunction identified by Darwin (e.g., Ekman,1984). Moreover, there is wide agreement thatemotions are the primary sources of motiva-

534 VI. COGNITIVE FACTORS

tion: They arouse, sustain, and direct humanaction (e.g., Izard, 1971; Frijda, 1986; Leeper,1948; Tomkins, 1962). Finally, many emotionresearchers have adopted a broad affect-as-information view, according to which internalemotional experiences provide individuals withimportant information about their environ-ment and situation. This information shapesthe individual’s judgments, decisions, priorities,and actions (e.g., Schwarz, 1990; Schwarz &Clore, 1983, 2003).

A FRAMEWORK

When we first began developing the concept ofemotional intelligence (e.g., Salovey & Mayer,1990), our intention was to draw closer atten-tion to the cooperative relationship betweenemotion and reason (Mayer, Caruso, &Salovey, 1999). Humans are not, in any practi-cal sense, predominantly rational beings, norare they predominantly emotional beings. Theyare both. Thus people’s abilities to adapt andcope in life depend on the integrated function-ing of their emotional and rational capacities.As Tomkins has said, “Out of the marriage ofreason with affect there issues clarity with pas-sion. Reason without affect would be impo-tent, affect without reason would be blind”(1962, p. 112). Success in life depends on one’sability to reason about emotional experiencesand other affect-laden information, and to re-spond in emotionally adaptive ways to the in-ferences drawn by reason about one’s situation,prospects, and past.

Generally, we have described emotional in-telligence as the ability to perceive and expressemotions, to understand and use them, and tomanage emotions so as to foster personalgrowth (Mayer & Salovey, 1997; Mayer et al.,2001, 2004; Salovey & Grewal, 2005; Salovey& Mayer, 1990). More formally, however, wedefine emotional intelligence by the specificcompetencies it encompasses, including theability to perceive emotion accurately; the abil-ity to access and generate feelings when they fa-cilitate cognition; the ability to understandaffect-laden information and make use of emo-tional knowledge; and the ability to manage orregulate emotions in oneself and others to pro-mote emotional and intellectual growth andwell-being. Our model of emotional intelli-gence is presented in Figure 33.1. The model iscomposed of four branches, each of which rep-

33. Emotional Intelligence 535

Perception, Appraisal, and Expressionof Emotion

• Ability to identify emotion in one’s physicaland psychological states.

• Ability to identify emotion in other peopleand objects.

• Ability to express emotions accurately, andto express needs related to those feelings.

• Ability to discriminate between accurate andinaccurate, or honest and dishonest,expressions of feelings.

Emotional Facilitation of Thinking• Ability to redirect and prioritize one’s thinking

based on the feelings associated withobjects, events, and other people.

• Ability to generate or emulate vivid emotionsto facilitate judgments and memoriesconcerning feelings.

• Ability to capitalize on mood swings to takeadvantage of multiple points of view; abilityto integrate these mood-inducedperspectives.

• Ability to use emotional states to facilitateproblem solving and creativity.

Understanding and Analyzing EmotionalInformation

• Ability to understand how different emotionsare related.

• Ability to perceive the causes andconsequences of feelings.

• Ability to interpret complex feelings, such asemotional blends and contradictory feelingstates.

• Ability to understand and predict likelytransitions between emotions.

Regulation of Emotion• Ability to be open to feelings, both those that

are pleasant and those that are unpleasant.• Ability to monitor and reflect on emotions.• Ability to engage, prolong, or detach from an

emotional state depending upon its judgedinformativeness or utility.

• Ability to manage emotion in oneself andothers.

FIGURE 33.1. The emotional intelligence frame-work. Adapted from Mayer and Salovey (1997).Copyright 1997 by Peter Salovey and David J.Sluyter. Adapted by permission of Basic Books, amember of Perseus Books, L.L.C.

resents a class of skills. The subskills of eachbranch are organized according to their com-plexity, such that the more sophisticatedsubskills of each branch are more likely to de-pend on skills from the other branches of themodel (Mayer & Salovey, 1997).

Perceiving Emotions

The first branch of emotional intelligence, “per-ceiving emotions,” is the ability to detect anddecipher emotions in faces, pictures, voices, andcultural artifacts. It also includes the ability toidentify one’s own emotions. Perceiving emo-tions may represent the most basic aspect ofemotional intelligence, as it makes all other pro-cessing of emotional information possible.

Individuals can be more or less skilled at at-tending to, appraising, and expressing theirown emotional states. These competencies arebasic information-processing skills in whichthe relevant information consists of feelingsand mood states. For example, some individu-als are unable to express their emotions ver-bally, presumably because they have difficultyidentifying their feelings; this condition iscalled “alexithymia” (Apfel & Sifneos, 1979;Lumley, Stettner, & Wehmer, 1996). These ba-sic emotional competencies are important, be-cause those who can quickly and accurately ap-praise and express their emotions are betterable to respond to their environment and toothers. There is some evidence, for instance,that individuals who can communicate theiremotions skillfully are more empathic and lessdepressed than those who are unable to do so(Mayer, DiPaolo, & Salovey, 1990; Notarius &Levenson, 1979; Prkachin, Craig, Papageorgis,& Reith, 1977).

Individuals also must appraise the emotionsof others. Again, there are individual differ-ences in people’s ability to perceive accurately,understand, and empathize with others’ emo-tions (reviewed in Buck, 1984; Barrett &Salovey, 2002), and individuals who are bestable to do so may be better able to respond totheir social environment and build a supportivesocial network (Salovey, Bedell, Detweiler, &Mayer, 1999).

Using Emotions to Facilitate Thinking

The second branch of emotional intelligence,“using emotions,” is the ability to harness emo-

tions to facilitate various cognitive activities,such as thinking and problem solving. Emo-tional states and their effects can be harnessedby individuals toward a number of ends. Forexample, positive moods make positive out-comes appear more likely, whereas negativemoods make negative outcomes appear morelikely (Johnson & Tversky, 1983; Mayer,Gaschke, Braverman, & Evans, 1992). Thusaddressing a problem while in different moodsmay enable individuals to consider a widerrange of possible actions and outcomes (Mayer& Hanson, 1995).

Likewise, a number of investigators (e.g.,Isen, 1987; Magai & Havliand-Jones, 2002;Palfai & Salovey, 1993; Schwarz, 1990) haveargued that emotions create different mentalsets that are more or less adaptive for solv-ing certain kinds of problems. That is, dif-ferent emotions create different information-processing styles. Happy moods facilitate amental set that is useful for creative tasks inwhich one must think intuitively or expansivelyin order to make novel associations (e.g., Isen,Daubman, & Nowicki, 1987). Sad moods gen-erate a mental set in which problems are solvedmore slowly, with particular attention to detail,and via more focused and deliberate strategies.Palfai and Salovey (1993) argued that thesetwo different information-processing styles(i.e., intuitive and expansive vs. focused anddeliberate) should be effective for two differentkinds of problem-solving tasks: inductive prob-lems like analogical reasoning, and deductivelogical tasks, respectively.

Emotionally intelligent individuals may alsobe able to harness the motivational qualities ofemotion. For example, a student may focuspurposefully on the negative consequences offailing to submit a term paper on time, in orderto self-induce a state of fear that will spur himto get an early start on the paper. Another stu-dent may remind herself of all her successes be-fore sitting down to write the paper. The self-induced positive mood that results bolsters herconfidence in writing the paper, and she may bemore likely to persevere when faced with a par-ticularly challenging section of it.

Understanding Emotional Information

A third branch of emotional intelligence con-cerns essential knowledge about the emotionalsystem. “Understanding emotions” is the abil-


ity to comprehend emotion language and toappreciate complicated relationships betweenemotions. The most fundamental competenciesat this level are the abilities to label emotionswith words and to recognize the relationshipsamong exemplars of the affective lexicon. Theemotionally intelligent individual is able to rec-ognize that the terms used to describe emotionsare arranged into families, and that groups ofemotion terms form fuzzy sets (see Ortony,Clore, & Collins, 1988). For instance, individ-uals learn that words such as “rage,” “irrita-tion,” and “annoyance” can be grouped to-gether as terms associated with anger. Perhapsmore importantly, the relations among theseterms are deduced—that annoyance and irrita-tion can lead to rage if the provocative stimulusis not eliminated; or that envy is often experi-enced in contexts that evoke jealousy, but jeal-ousy is less likely to be part of envy-provokingsituations (Salovey & Rodin, 1986, 1989).

To understand the emotions, individualsmust learn what emotions convey about rela-tionships. Lazarus (1991), for example, de-scribes how core relational themes—the centralharm or benefit in adaptational encounters thatunderlies each emotion—are associated withdifferent kinds of feelings. Anger results from ademeaning offense against the self, guilt fromtransgressing a moral imperative, and hopefrom facing the worst but yearning for better(see Table 3.4 in Lazarus, 1991).

Increased complexity in this domain of emo-tional intelligence is represented by knowledgethat emotions can combine in interesting andsubtle ways. At a high school reunion, nostal-gic conversation can give rise to wistfulness, ablend of both joy and sorrow. Startled surpriseat the wonders of the universe combined withfear about one’s insignificant place in it maygive rise to awe.

Finally, understanding and analyzing emo-tions include the ability to recognize transitionsamong emotions. For example, Tangney andher colleagues have written extensively abouthow shame but not guilt can turn quickly torage. Individuals can literally be shamed intoanger (Tangney, Wagner, Fletcher, & Gramzow,1992; Tangney & Salovey, 1999).

Managing Emotion

Emotional knowledge also contributes to thefourth component of emotional intelligence,

“emotion regulation.” However, individualsmust develop further competencies in order toput their knowledge into action. They mustfirst be open to the experience of mood andemotion, and then practice and become adeptat engaging in behaviors that bring about de-sired feelings in themselves and others. Theseemotion regulation skills enable individuals toengage in mood maintenance and mood repairstrategies, such as avoiding unpleasant activi-ties or seeking out activities that they typicallyfind rewarding. Individuals who are unable tomanage their emotions are more likely to expe-rience negative affect and remain in poor spirits(Erber, 1996).

Through the self-reflective experience ofemotion, individuals acquire knowledge of thecorrelates and causes of their emotional experi-ences. Knowledge of emotion thus enables indi-viduals to form theories of how and why emo-tions are elicited by different situations. Thisability to understand and analyze emotionalexperiences translates into the ability to under-stand oneself and one’s relation to the environ-ment better, which may foster effective emo-tional regulation and greater well-being. In thepsychotherapy literature, this has been termed“emotional literacy” (Steiner & Perry, 1997;see also Maurer & Brackett, 2004, for an ap-plication to education).

Although our work has focused primarily onreflective metamood abilities (i.e., thoughtsabout moods; see Mayer & Gaschke, 1988), itis worth noting that emotional intelligence maymanifest itself in a second way. Individuals of-ten react emotionally toward their direct expe-riences of emotion, and these metaemotionalexperiences can either facilitate or impedefunctioning (Gottman, 1997). For example, aperson can feel ashamed for having felt orexpressed anger toward a loved one. The meta-emotion in this case is shame, which takes as itsobject the individual’s direct experience of an-ger, and it may motivate the individual to in-hibit anger or at least suppress angry behaviorin the future. This type of learned emotional re-straint can be highly advantageous betweenparents and children, between lovers, and inmost other social relationships. To date, therehave been very few investigations of meta-emotion (although see Gottman, 1997), in partbecause studying emotional responses to directemotional experiences is a complex affair.However, metaemotion is a fascinating in-


stance of how humans take themselves andtheir experiences as objects and respond tothese objects in a higher-order manner.

The ability to help others enhance or repairtheir moods is also an important skill. Individ-uals often rely on their social networks to pro-vide not just a practical but an emotionalbuffer against negative life events (for a review,see Stroebe & Stroebe, 1996). Moreover, indi-viduals appear to derive a sense of efficacy andsocial worth from helping others feel better andfrom contributing to the joy of their lovedones. The ability to manage others’ emotionalexperiences also plays a significant role inimpression management and persuasion. Al-though this skill is sometimes employed un-scrupulously by sociopaths, cult leaders, andsome advertisers, impression management andpersuasion are often employed prosocially aswell. Thus individuals who are effective at reg-ulating the emotions of others are better able toact prosocially and to build and maintain solidsocial networks.

One might wonder what skills, beyond emo-tional knowledge, undergird competence atemotion regulation. We cannot answer thisquestion definitively, because very little re-search has explored the distinction betweenemotional knowledge and emotion regulation.However, this distinction can be illustratedquite easily. Consider the embarrassment manyof us experience while dancing. Those of uswho enjoy dancing are able to lose ourselves inthe music and motion of dance. Unfortunately,this delightful state is elusive when we are anx-ious about being evaluated by others. The real-ity, however, is that few others typically carehow we dance. Other dancers are either too en-grossed or, ironically, too embarrassed them-selves to notice us; spectators pass over those ofus whose dancing is at worst a little boring, andare attracted instead to those whose dancing ismarked by skill and elegance. Interestingly,those of us who find dancing embarrassing areoften aware that no one else is paying attentionto our dancing. We even understand that ourfear of being evaluated is itself our greatestimpediment on the dance floor. This is meta-emotional knowledge. However enlighteningthis knowledge is, it enables only a few of us toactually overcome our embarrassment. This isbecause emotional regulation is distinct fromthe metaemotional knowledge it presupposes.That is, regulatory skills are needed in order toput metaemotional knowledge into action.

Summary

Much remains to be learned about each of thecomponents of emotional intelligence. As a re-sult, our conception of emotional intelligence isstill evolving. Nonetheless, our understandingof emotional intelligence has already translatedinto some research fundings. We next turn tomeasures of emotional intelligence, after whichwe discuss some of the findings and implica-tions that have stemmed from the use of thesemeasures and the theory more generally.

MEASURING EMOTIONALINTELLIGENCE

The introduction of the concept of emotionalintelligence (Salovey & Mayer, 1990) had im-mediate intuitive and popular appeal (e.g.,Goleman, 1995a; Segal, 1997), and this ideahas been used to organize much past and con-temporary research. Attempts to operationalizeand directly measure this construct were inevi-table. Guided by the original framework ofemotional intelligence, Mayer and Saloveyinitially examined the metaexperience, or re-flective experience, of mood (e.g., Mayer &Gaschke, 1988; Salovey, Mayer, Goldman, Tur-vey, & Palfai, 1995). Two self-report scales toassess metamood cognitions have been em-ployed: a trait scale (Salovey et al., 1995) and astate scale (Mayer & Stevens, 1994). The for-mer, for example, is the Trait Meta-Mood Scale(TMMS), which taps into people’s beliefs abouttheir propensity to attend with clarity totheir own mood states and to engage inmood repair. The items of this measure arestraightforward—for instance, “I pay a lot ofattention to how I feel” (Attention), “I cannever tell how I feel” (Clarity, reverse-scored),and “I try to think good thoughts no matterhow badly I feel” (Repair). The psychometricproperties of the TMMS are quite good, andsome empirical findings have been generatedfrom its use (Goldman, Kraemer, & Salovey,1996; Salovey et al., 1995; Salovey, Stroud,Woolery, & Epel, 2002). Nevertheless, this traitscale has its limitations. First, its factor struc-ture consists of only three dimensions (i.e., At-tention, Clarity, and Repair), representing onlya few of the emotional competencies outlinedin our framework. Moreover, the TMMS, likeother self-report measures (see, e.g., Bar-On,1997), essentially asks individuals whether or


not they are emotionally intelligent; it does notrequire individuals to demonstrate their emo-tional competencies. We believe that a morevalid measure of core emotional intelligence re-quires a test that relies on tasks and exercisesrather than on self-report.

More recent task-based attempts to measureemotional intelligence have focused on com-prehensive aptitude-type tests that rely on theassessment of relevant skills. The MultifactorEmotional Intelligence Scale can be adminis-tered with pencil and paper or on a computer(Mayer, Caruso, & Salovey, 1998). It is orga-nized into four main branches, reflecting ourcurrent framework for understanding emo-tional intelligence: perceiving emotions, usingemotions, understanding emotions, and man-aging emotions. Within each of these fourbranches, we designed a series of subtests to as-sess various competencies.

More recently, we have published a shorterand better-normed ability-based test of emo-tional intelligence, called the Mayer–Salovey–Caruso Emotional Intelligence Test (MSCEIT;Mayer, Salovey, & Caruso, 2002). TheMSCEIT is a 40-minute battery that may becompleted either on paper or on a computer. Bytesting a person’s abilities on each of the fourbranches of emotional intelligence, it generatesfive scores: one for each of the four branches aswell as a total score. Central to the four-branchmodel is the idea that emotional intelligence re-quires attunement to social norms. Therefore,the scoring of the MSCEIT is based on consen-sual norms, with higher scores indicating theamount of overlap between an individual’s an-swers and those provided by a broad sample ofthousands of respondents. In addition, theMSCEIT can be expert-scored, so that theamount of overlap is calculated between an in-dividual’s answers and those provided by agroup of 21 emotion researchers. Importantly,both methods are reliable and yield similarscores, indicating that both laypeople andexperts possess shared social knowledgeabout emotions (Mayer, Salovey, Caruso, &Sitarenios, 2003).

Creating an assessment battery that success-fully tests a construct as broad as emotional in-telligence is challenging, but it appears that theMSCEIT is an appropriate starting point.Scores on each of the four branches (perceiv-ing, using, understanding, and managing emo-tions) correlate modestly with one another, andthe subscales and overall measure are reliable

(Mayer et al., 2003). Lopes et al. (2003) foundsmall positive correlations between scores onthe MSCEIT and the traits of agreeablenessand conscientiousness. The MSCEIT appearsto test emotional abilities rather than personal-ity traits, however, and does not correlate withsocial desirability scales.

Although the construct of emotional intelli-gence and its measurement have generated con-siderable interest, validity data are only begin-ning to appear. There is a converging senseamong researchers, although not necessarilyamong those working in more applied settings,of what emotional intelligence is (a set of com-petencies concerning the appraisal and expres-sion of feelings, the use of emotions to facilitatecognitive activities, knowledge about emo-tions, and the regulation of emotion) and whatemotional intelligence is not (good character,optimism, delay of gratification, or persistence;see Mayer, Salovey, & Caruso, 2000, for acomparison of our model of emotional intelli-gence to those popularized by others). Yet thereis considerably less consensus on how best tomeasure emotional intelligence. Although wehave argued for the advantages of task-basedand behavioral assessment (Mayer et al., 1990,2000), various self-assessments have also ap-peared that may measure important aspects ofindividuals’ perceptions of their competenciesin this and related domains (e.g., Bar-On,1996, 1997; Petrides & Furnham, 2003;Schutte et al., 1998; Tett, Fox, & Wang, 2005).There is a concern, however, that such self-report tests correlate quite highly with tradi-tional measures of personality and well-being(Brackett & Mayer, 2003; Brackett, Rivers,Shiffman, Lerner, & Salovey, 2006). Other self-report tests that have been repackaged underthe emotional intelligence rubric appear tohave little to do with this construct (e.g.,Simmons & Simmons, 1997).

Findings concerning the correlates of emo-tional intelligence as operationalized by theMSCEIT are now being reported. Emotionalintelligence is negatively associated with devi-ant behavior in male adolescents (Brackett,Mayer, & Warner, 2004). College-age maleswho scored lower on the MSCEIT reported en-gaging in more recreational drug use andgreater alcohol consumption. In addition, theseparticipants reported having more unsatisfyingrelationships with their friends. Even when theeffects of participants’ personality and analyticintelligence were controlled, the findings in-


volving emotional intelligence remained signifi-cant. Similarly, Lopes, Salovey, and Straus(2003) administered the MSCEIT to a sampleof college students, along with questionnairesthat assessed self-reported satisfaction withsocial relationships. Participants who scoredhigher on the MSCEIT were more likely to re-port having positive relationships with others,including greater perceived support from theirparents and fewer negative interactions withtheir close friends.

A limitation of the two studies describedabove is that they used the MSCEIT to predictthe self-reported quality of social relationships.Lopes et al. (2004), however, examined the re-lationship between emotional intelligence andpeer reports of one’s attributes. American col-lege students took the MSCEIT and were askedto have two of their close friends rate their per-sonal qualities. The students who scored higheron the MSCEIT received more positive ratingsfrom their friends. The friends also reportedthat students high in emotional intelligencewere more likely to provide them with emo-tional support in times of need. In anotherstudy, German students were asked to keep dia-ries of their daily social interactions. Thosestudents who scored higher on the MSCEITreported greater success in their social interac-tions with members of the opposite sex. For ex-ample, they were more likely to report thatthey had come across in a competent/attractivemanner and that their opposite-sex partnersperceived them as having desirable qualities,such as intelligence and friendliness.

Emotional intelligence may also help peoplein relationships with their partners andspouses. One study examined the emotional in-telligence of 180 college-age couples (Brackett,Cox, Gaines, & Salovey, 2005). They com-pleted the MSCEIT and then answered ques-tions about the quality of their relationships.The couples were classified by how well matchedthe partners were in emotional intelligence.The couples in which both individuals scoredlow on the MSCEIT reported the greatest un-happiness with their relationship, as comparedto the happiness ratings of the other twogroups. The couples in which both partnerswere emotionally intelligent were very happy.Furthermore, couples in which only one part-ner had high emotional intelligence tended tofall in between the other groups in happiness.

Emotional intelligence also may matter atwork. A sample of employees of a Fortune 500

insurance company, who worked in smallteams each headed by a supervisor, completedthe MSCEIT. All employees were asked to rateeach other on the qualities they displayed atwork, such as handling stress and conflict welland displaying leadership potential. Supervi-sors were also asked to rate their employees.Employees with higher scores on the MSCEITwere rated by their colleagues as easier to dealwith and more responsible for creating a posi-tive work environment. Their supervisors ratedthem as more interpersonally sensitive, moretolerant of stress, more sociable, and havinggreater potential for leadership. Moreover,higher scores on the MSCEIT were related tohigher salaries and larger annual raises (Lopes,Grewal, Kadis, Gall, & Salovey, 2006).

Despite what has been learned about themeasurement of emotional intelligence over thepast few years, research on the validity, in par-ticular, of most emotional intelligence tests isstill in its adolescence. Thus Boring’s (1923)suggestion that “intelligence is what the teststest” (p. 35) is especially misleading in this areaof study. Nonetheless, we are excited that emo-tional intelligence as measured with theMSCEIT seems to moderate robust (and whathave been thought to be relatively universal)psychological phenomena, such as biases in af-fective forecasting (Dunn, Brackett, Ashton-James, Schneiderman, & Salovey, 2007). As wesettle on a clear conceptual understanding ofemotional intelligence, we will continue to re-fine and expand measures of emotional intelli-gence to reflect this understanding, and espe-cially to look for ways to measure emotionalintelligence with information-processing tasks.With this in mind, we now turn our attentiontoward promising findings pertaining to emo-tional intelligence and what these findingssuggest with respect to particularly relevantdomains—coping with stressors and other as-pects of self-regulation.

EMOTIONALLY INTELLIGENTCOPING AND SELF-REGULATION

We believe, with others (e.g., Izard, 1971;Tomkins, 1962), that emotion is the wellspringof human motivation, the “primary provider ofblueprints for cognition, decision and action”(Tomkins, 1962, p. 22). This view implies thatemotional intelligence amounts to motivationalor self-regulatory intelligence. To the extent


that one has highly developed emotionalknowledge and competencies, successful andefficient self-regulation toward desired endsshould be facilitated.

Because past events cannot be changed, cop-ing with a previous traumatic experience is amatter of understanding the event and reinter-preting it in a more meaningful way. The im-portance of emotional competencies is perhapsmost evident in these cases, in that individualsare forced to respond to the powerful emotionselicited by memories of past events. Elsewhere,we have argued that successful coping dependson the integrated operation of many emotionalcompetencies, and we have suggested that defi-ciencies in such basic emotional competenciesas emotion perception and expression will in-terfere with the development and implementa-tion of more complex coping skills, such asemotion regulation (Salovey, Bedell, Detweiler,& Mayer, 1999). The relevance of the emo-tional intelligence framework to specific meth-ods of coping such as rumination, disclosure,and distraction is reviewed below.

“Rumination” is defined as “passively andrepetitively focusing on one’s symptoms of dis-tress and the circumstances surrounding thosesymptoms” (Nolen-Hoeksema, McBride, &Larson, 1997, p. 855). Nolen-Hoeksema(1991; Treynor, Gonzalez, & Nolen-Hoeksema,2003), who has investigated rumination exten-sively, regards it as a particular style of re-sponding to stressful events that tends to inten-sify and lengthen periods of depressed mood.Following the 1989 Loma Prieta earthquake inCalifornia, for example, Nolen-Hoeksema andMorrow (1991) found that people who hada more ruminative response style before theearthquake exhibited higher levels of depres-sion 10 days after the event. Similarly, newlybereaved men identified as ruminators prior totheir loss experienced longer and more severeperiods of depression after their partners’deaths (Nolen-Hoeksema et al., 1997). Thedeleterious effects of ruminative coping havebeen corroborated in a number of laboratorystudies as well (e.g., Nolen-Hoeksema & Mor-row, 1993).

Although ruminating about a negative ex-perience exacerbates one’s negative mood,Pennebaker has demonstrated that disclosingemotional traumas in writing, even anony-mously, has numerous beneficial effects (seePennebaker, 1997, for a review). In general, webelieve that individuals who can identify how

they are feeling, understand the implications ofthese feelings, and effectively regulate theiremotional experiences will cope more suc-cessfully with negative experiences than lessemotionally intelligent individuals will. At abasic level, those who are unable to perceiveand appraise their own emotional states accu-rately may fail to recognize the origin of theirtroubles. If so, the coping process will stall,precluding effective emotional disclosure. Inprevious work from our own lab, individualsopen to emotional experience (even when thiswas negative) reported lower levels of depres-sion than those who claimed to fight the feelingor asserted that everything is okay (Mayer,Salovey, Gomberg-Kaufman, & Blainey, 1991).

In addition, emotionally intelligent individu-als should be able to recognize and pursue themost effective means of coping. For instance,Nolen-Hoeksema and colleagues have arguedthat one of the most effective approaches fordisengaging from a ruminative coping cycle isdistraction (Morrow & Nolen-Hoeksema,1990; Nolen-Hoeksema & Morrow, 1993).When people use pleasant activities to relievetheir moods, they show better problem-solving skills and fewer negative thoughts(Lyubomirsky & Nolen-Hoeksema, 1995).One of the most advanced skills within the re-flective regulation of emotion is the abilityto ameliorate negative emotions and promotepleasant emotions (Mayer & Salovey, 1997).Thus we would argue that individuals who areskilled at regulating emotions should be betterable to move to repair their emotional states byusing pleasant activities as a distraction fromnegative affect.

Engaging in distraction is different, however,from avoiding negative affect altogether.Wegner’s work on ironic processes (e.g.,Wegner, Erber, & Zanakos, 1993) has dem-onstrated that attempts to avoid negativethoughts and feelings altogether are doomed tofail. The failure of sheer mental willpower oc-curs because the suppressed thoughts and feel-ings are maintained as markers of how success-fully the person has avoided them. This work isconsistent with the model of cognitive changes,suggesting that a negative experience will con-tinue to challenge one’s thoughts and feelingsuntil it is resolved and is thus no longer some-thing that needs to be avoided. A further com-ponent of reflective regulation is the ability tounderstand emotions without exaggeratingor minimizing their importance (Mayer &


Salovey, 1997). As a result, emotionally intelli-gent individuals should be able to strike ahealthy balance between pleasant distractionsand coming to terms with their feelings. Sometypes of distractions may even facilitate activecoping. For instance, we would expect emo-tionally intelligent individuals to seek out thecompany of others in an effort to be remindedby them that life is good. Individuals naturallyturn to others in order to discuss and makesense of negative life events (Pennebaker &Harber, 1993), and the availability of high-quality social support may prevent individualsfrom ruminating (Nolen-Hoeksema, 1991).

Emotionally intelligent individuals shoulddisclose their emotional experiences more of-ten, because they are more apt to recognize thatsharing is an efficient means of organizing andthus regulating one’s emotions. Moreover, thelinguistic features characterizing effective emo-tional disclosure (i.e., insight, causal thinking,and a balance of emotion) reflect one’s abilityto understand, analyze, and actively regulateemotion. Thus individuals with strong emo-tional competencies should be able to (1) rec-ognize their emotional responses to a trauma asnatural, (2) see the trauma and their emotionsin the broader context of their lives, and (3)make positive attributions about the traumaand their emotions.

Emotional intelligence should also lead tomore adaptive coping when goals are blocked.Obviously, failing to attain a goal is unpleas-ant. Even moving more slowly than expectedtoward a goal can be aversive (Carver &Scheier, 1990; Hsee & Abelson, 1991). How-ever, not everyone reacts similarly when a goalis blocked. Some people are able to disengagegracefully when they realize a goal is out ofreach. Others react more negatively, ruminat-ing about their failure and its broader implica-tions.

McIntosh (1996) has proposed a model ex-plaining why certain individuals react stronglyto goal nonattainment. The model begins withthe observation that goals are structured hier-archically, which restates the notion that proxi-mal day-to-day goals (such as going for a run)are instrumental in achieving more general, dis-tal strivings (such as being healthy or slim). Oc-casionally a particular lower-order goal mustbe achieved in order to reach a higher-ordergoal, but typically there are a number of waysto accomplish goals higher up the goal hierar-chy (e.g., eating properly also promotes health

and weight control). Nonetheless, some indi-viduals tend to view particular lower-ordergoals as necessary even when they are not.McIntosh refers to these individuals as “link-ers,” because they interpret failure to attainlower-order goals as failure to attain their moredistal goals. Linkers believe more strongly thanothers that their happiness depends on the ac-complishment of goals, both big and small,proximal and distal. When a proximal goal islinked to an intermediate goal in the hierarchy,the person may become distressed if the initialgoal is blocked. For instance, not running to-day may be taken as evidence that one willnever routinely exercise, which may be dis-heartening. More dramatically, the proximalgoal may be linked to a goal at the very top ofthe hierarchy. Not running today may be takenas evidence that one will never be healthy, slim,attractive, happy, or a good person. In this way,a minor failure can lead to a depressed or evenhopeless state.

In support of the linking model, McIntoshhas shown that linkers are more likely thannonlinkers to ruminate about a current unreal-ized goal (Martin, Tesser, & McIntosh, 1993),and linkers report less happiness and more neg-ative feelings as a result of their propensity toruminate (McIntosh, 1996). In a short-termprospective study, McIntosh, Harlow, andMartin (1995) asked college students to com-plete measures of linking, stress, rumination,depression, and physical symptoms at an initialsession and then again after a 2-week interval.Students who tend to link the attainment oflower-order goals to the attainment of higher-order goals reported more rumination, depres-sion, and physical symptoms overall than theirnonlinking peers did. Moreover, linkers whoreported experiencing stressful events at thetime of the first session were the most likely toreport depression and physical symptoms atthe second session. In contrast, high stress atthe first session failed to predict later depres-sion and physical symptoms among nonlinkers.This parallels Nolen-Hoeksema’s (e.g., Nolen-Hoeksema, 2000; Nolen-Hoeksema & Mor-row, 1991, 1993) finding that ruminators re-port elevated levels of depression after experi-encing stress, whereas individuals who do notruminate are more resilient. Indeed, McIntosh’sstudies may explain in part why some peopleruminate whereas others do not. People wholink lower-order goals to higher-order goals ap-pear to ruminate more when they experience


stress or failure, and rumination mediates theinfluence of goal linking on subsequent dissat-isfaction and negative affect.

Emotional intelligence may make it easier totake stock of various goals; it enables the indi-vidual to sense the personal importance of eachgoal and to use this information in reasoningabout competing goals as well as alternatemeans of achieving long-term pursuits. Thisanalysis enables emotionally intelligent individ-uals to invest themselves wisely in specific ac-tivities. If a setback or failure occurs, these in-dividuals experience a loss, but the loss is welldefined and assessable with respect to othermeans of moving forward. This stabilizes self-regulation in much the same way that effectiveemotional disclosures facilitate coping: by clar-ifying the individual’s situation and averting ru-minative thinking and paralysis.

FINAL THOUGHTSON EMOTIONAL INTELLIGENCE

Most constructs in personality and social psy-chology mature for decades before they find apopular audience. The fate of emotional intelli-gence, however, was quite different. Some timeafter our initial work on the subject was pub-lished (e.g., Mayer et al., 1990; Mayer &Salovey, 1993; Salovey & Mayer, 1990), a pop-ular book on emotional intelligence appearedand skyrocketed up the best-seller list(Goleman, 1995a). With this book, emotionalcompetencies went almost overnight from a setof abilities worthy of further study (our view)to a wealth of personal assets capable of deter-mining a person’s character, life achievements,and health (Goleman’s view). Truly extraordi-nary claims on behalf of emotional intelligencebecame commonplace—for instance, “Havinggreat intellectual abilities may make you a su-perb fiscal analyst or legal scholar, but a highlydeveloped emotional intelligence will make youa candidate for CEO or a brilliant lawyer”(Goleman, 1995b, p. 76). In proposing theframework of emotional intelligence, it wouldseem that we stumbled upon a panacea for in-dividual and society alike without even know-ing it!

Despite the popularization of the constructin the 1990s, serious empirical research onemotional intelligence has begun to emerge inthe present decade. The problematic issues inthis area of work are not surprising, given the

relative immaturity of this research domain(see, e.g., Matthews, Roberts, & Zeidner,2004; Matthews, Zeidner, & Roberts, 2003).For one thing, the term “emotional intelli-gence” is used to represent various aspects ofthe human condition. We prefer to focus nar-rowly on specific abilities and competenciesconcerned with appraising, understanding, andregulating emotions, and using them to facili-tate cognitive activities. However, others havedefined emotional intelligence in terms of moti-vation (persistence, zeal), cognitive strategies(delay of gratification), and even character (be-ing a good person). Emotional intelligence maycontribute to persistence, delay, and character,but they are not one and the same thing. A conartist may be especially skilled at reading andregulating the emotions of other people, butmay have little of what is commonly thought tobe good character.

Second, this area of research will not proveto be productive unless the abilities that makeup emotional intelligence can be measured reli-ably, and unless these abilities are related to im-portant, real-world outcomes. We think less at-tention should be focused on the issue ofwhether a monolithic “emotional intelligencequotient” has utility, or on gathering togethervarious measures of social competence andcalling them “emotional intelligence,” andmore on the development of tasks that can beused to assess actual emotion-related skills. Weare not confident that self-reported abilities inthis domain will prove any more useful thanthey have in the measurement of traditional,analytic intelligence. Finally, we urge educatorsand business managers to maintain their pres-ent interest in emotional intelligence, but to beskeptical of “quick-fix” programs. Althoughemotional intelligence research may challengeus to reconsider our notions of what it meansto be smart (and what it means to be in touchwith feelings), it will not, at the end of the day,be the key to reducing international conflict,fighting the war on drugs, or terminating theglobal plague of AIDS. Grandiose claims to thecontrary serve only as palliatives to the publicand as suppressors of scientific inquiry.

ACKNOWLEDGMENTS

Preparation of this chapter was facilitated by grantsfrom the National Cancer Institute (No. R01-CA68427) and the National Institute of MentalHealth (No. P01-MH/DA56826).


REFERENCES

Apfel, R. J., & Sifneos, P. E. (1979). Alexithymia: Con-cept and measurement. Psychotherapy and Psycho-somatics, 3, 180–190.

Aristotle. (1947). On the soul (J. A. Smith, Trans.). In R.McKeon (Ed.), Introduction to Aristotle (pp. 145–235). New York: Random House.

Aristotle. (1976). Ethics (J. A. K. Thomson, Trans.).London: Penguin Books.

Bar-On, R. (1996, August). The era of the EQ: Definingand assessing emotional intelligence. Paper presentedat the 104th Annual Convention of the AmericanPsychological Association, Toronto.

Bar-On, R. (1997). EQ-I: Bar-On Emotional QuotientInventory. Toronto: Multi-Health Systems.

Barrett, L. F., & Salovey, P. (Eds.). (2002). The wisdomin feeling: Psychological processes in emotional intel-ligence. New York: Guilford Press.

Boring, E. G. (1923, June 6). Intelligence as the tests testit. New Republic, pp. 35–37.

Brackett, M. A., Cox, A., Gaines, S. O., & Salovey, P.(2005). Emotional intelligence and relationship qual-ity among heterosexual couples. Manuscript submit-ted for publication.

Brackett, M. A., & Mayer, J. D. (2003). Convergent,discriminant, and incremental validity of competingmeasures of emotional intelligence. Personality andSocial Psychology Bulletin, 29, 1147–1158.

Brackett, M. A., Mayer, J. D., & Warner, R. M. (2004).Emotional intelligence and the prediction of be-havior. Personality and Individual Differences, 36,1387–1402.

Brackett, M. A., Rivers, S. E., Shiffman, S., Lerner, N.,& Salovey, P. (2006). Relating emotional abilities tosocial functioning: A comparison of self-report andperformance measures of emotional intelligence.Journal of Personality and Social Psychology, 91,780–795.

Buck, R. (1984). The communication of emotion. NewYork: Guilford Press.

Cantor, N., & Kihlstrom, J. F. (1985). Social intelli-gence: The cognitive basis of personality. Review ofPersonality and Social Psychology, 6, 15–33.

Cantor, N., & Kihlstrom, J. F. (1987). Personality andsocial intelligence. Englewood Cliffs, NJ: PrenticeHall.

Carver, C. S., & Scheier, M. F. (1990). Origins and func-tions of positive and negative affect: A control-process view. Psychological Review, 97, 19–35.


Dunn, E. W., Brackett, M. A., Ashton-James, C.,Schneiderman, E., & Salovey, P. (2007). On emotion-ally intelligent time travel: Individual differences inaffective forecasting ability. Personality and SocialPsychology Bulletin, 33, 85–93.

Ekman, P. (1984). Expression and the nature of emo-

tion. In K. Scherer & P. Ekman (Eds.), Approaches toemotion (pp. 319–344). Hillsdale, NJ: Erlbaum.

Epstein, S. (1998). Constructive thinking: The key toemotional intelligence. New York: Praeger.

Erber, R. (1996). The self-regulation of moods. In L. L.Martin & A. Tesser (Eds.), Striving and feeling: Inter-actions among goals, affect, and self-regulation(pp. 251–275). Mahwah, NJ: Erlbaum.

Freud, S. (1962). The ego and the id (J. Strachey, Ed., &J. Riviere, Trans.). New York: Norton. (Originalwork published 1923)


Gardner, H. (1983). Frames of mind: The theory of mul-tiple intelligences. New York: Basic Books.

Goldman, S. L., Kraemer, D. T., & Salovey, P. (1996).Beliefs about mood moderate the relationship ofstress to illness and symptom reporting. Journal ofPsychosomatic Research, 41, 115–128.

Goleman, D. (1995a). Emotional intelligence. NewYork: Bantam.

Goleman, D. (1995b). EQ: What’s your emotional intel-ligence quotient? The Utne Reader, 72, 74–76.

Gottman, J. M. (1997). Meta-emotion: How familiescommunicate emotionally. Mahwah, NJ: Erlbaum.

Hsee, C. K., & Abelson, R. P. (1991). Velocity relation:Satisfaction as a function of the first derivative ofoutcome over time. Journal of Personality and SocialPsychology, 41, 1–15.

Hume, D. (1948). A treatise of human nature. In H. D.Aiken (Ed.), Hume: Moral and political philosophy(pp. 1–169). New York: Hafner Press. (Originalwork published 1739)

Isen, A. M. (1987). Positive affect, cognitive processes,and social behavior. In L. Berkowitz (Ed.), Advancesin experimental social psychology (Vol. 20, pp. 203–253). New York: Academic Press.

Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987).Positive affect facilitates creative problem solving.Journal of Personality and Social Psychology, 52,1122–1131.


James, W. (1950). The principles of psychology (Vol. 2).New York: Dover. (Original work published 1890)

Johnson, E. J., & Tversky, A. (1983). Affect, generaliza-tion, and the perception of risk. Journal of Personal-ity and Social Psychology, 15, 294–301.


Leeper, R. W. (1948). A motivational theory of emotionsto replace “emotions as disorganized response.” Psy-chological Review, 55, 5–21.

Lopes, P. N., Brackett, M. A., Nezlek, J. B., Schütz, A.,Sellin, I., & Salovey, P. (2004). Emotional intelligenceand social interaction. Personality and Social Psy-chology Bulletin, 30, 1018–1034.

Lopes, P. N., Grewal, D., Kadis, J., Gall, M., & Salovey,P. (2006). Evidence that emotional intelligence is re-


lated to job performance and affect and attitudes atwork. Psicothema, 18, 132–138.

Lopes, P. N., Salovey, P., & Straus, R. (2003). Emotionalintelligence, personality, and the perceived quality ofsocial relationships. Personality and Individual Dif-ferences, 35, 641–658.

Lumley, M. A., Stettner, L., & Wehmer, F. (1996). Howare alexithymia and physical illness linked?: A reviewand critique of pathways. Journal of PsychosomaticResearch, 41, 505–518.

Lyubomirsky, S., & Nolen-Hoeksema, S. (1995). Effectsof self-focused rumination on negative thinking andinterpersonal problem solving. Journal of Personalityand Social Psychology, 69, 176–190.

Magai, C., & Haviland-Jones, J. (2002). The hidden ge-nius of emotion. New York: Cambridge UniversityPress.

Martin, L. L., Tesser, A., & McIntosh, W. D. (1993).Wanting but not having: The effects of unattainedgoals on thoughts and feelings. In D. M. Wegner & J.W. Pennebaker (Eds.), Handbook of mental control(pp. 552–572). Englewood Cliffs, NJ: Prentice Hall.

Matthews, G., Roberts, R. D., & Zeidner, M. (2004).Seven myths about emotional intelligence. Psycho-logical Inquiry, 15, 179–196.

Matthews, G., Zeidner, M., & Roberts, R. D. (2003).Emotional intelligence: Science and myth. Cam-bridge, MA: MIT Press.

Maurer, M., & Brackett, M. A. (2004). Emotional liter-acy in the middle school. Port Chester, NY: Dude.

Mayer, J. D., Caruso, D., & Salovey, P. (1998). TheMultifactor Emotional Intelligence Scale: MEIS. (Un-published manual available from John D. Mayer, De-partment of Psychology, University of New Hamp-shire, Conant Hall, Durham, NH 03824).

Mayer, J. D., Caruso, D., & Salovey, P. (1999). Emo-tional intelligence meets traditional standards for anintelligence. Intelligence, 27, 267–298.

Mayer, J. D., DiPaolo, M., & Salovey, P. (1990). Per-ceiving the affective content in ambiguous visualstimuli: A component of emotional intelligence. Jour-nal of Personality Assessment, 54, 772–781.

Mayer, J. D., & Gaschke, Y. N. (1988). The experienceand meta-experience of mood. Journal of Personalityand Social Psychology, 55, 102–111.

Mayer, J. D., Gaschke, Y. N., Braverman, D. L., & Ev-ans, T. W. (1992). Mood-congruent judgment is ageneral effect. Journal of Personality and Social Psy-chology, 63, 119–132.

Mayer, J. D., & Hanson, E. (1995). Mood-congruentjudgment over time. Personality and Social Psychol-ogy Bulletin, 21, 237–244.

Mayer, J. D., & Salovey, P. (1993). The intelligence ofemotional intelligence. Intelligence, 17, 433–442.

Mayer, J. D., & Salovey, P. (1997). What is emotionalintelligence? In P. Salovey & D. J. Sluyter (Eds.),Emotional development and emotional intelligence(pp. 3–31). New York: Basic Books.

Mayer, J. D., Salovey, P., & Caruso, D. (2000). Models

of emotional intelligence. In R. J. Sternberg (Ed.),Handbook of human intelligence (2nd ed., pp. 396–420). New York: Cambridge University Press.

Mayer, J. D., Salovey, P., & Caruso, D. (2002). TheMayer–Salovey–Caruso Emotional Intelligence Test(MSCEIT). Toronto: Multi-Health Systems.

Mayer, J. D., Salovey, P., & Caruso, D. (2004). Emo-tional intelligence: Theory, findings, and implica-tions. Psychological Inquiry, 15, 197–215.

Mayer, J. D., Salovey, P., Caruso, D., & Sitarenios, G.(2001). Emotional intelligence as a standard intelli-gence. Emotion, 1, 232–242.

Mayer, J. D., Salovey, P., Caruso, D., & Sitarenios, G.(2003). Measuring emotional intelligence with theMSCEIT V2.0. Emotion, 3, 97–105.

Mayer, J. D., Salovey, P., Gomberg-Kaufman, S., &Blainey, K. (1991). A broader conception of moodexperience. Journal of Personality and Social Psy-chology, 60, 100–111.

Mayer, J. D., & Stevens, A. (1994). An emerging under-standing of the reflective (meta-)experience of mood.Journal of Research in Personality, 60, 100–111.

McIntosh, W. D. (1996). When does goal non-attainment lead to negative emotional reactions, andwhen doesn’t it?: The role of linking and rumination.In L. L. Martin & A. Tesser (Eds.), Striving and feel-ing: Interactions among goals, affect, and self-regulation (pp. 53–77). Mahwah, NJ: Erlbaum.

McIntosh, W. D., Harlow, T. F., & Martin, L. L. (1995).Linkers and nonlinkers: Goal beliefs as a moderatorof the effects of everyday hassles on rumination, de-pression, and physical complaints. Journal of Ap-plied Social Psychology, 25, 1231–1244.

Mischel, W. (1990). Personality dispositions revisitedand revised: A view after three decades. In L. A.Pervin (Ed.), Handbook of personality: Theory andresearch (pp. 111–134). New York: Guilford Press.

Morrow, J., & Nolen-Hoeksema, S. (1990). Effects ofresponses to depression on the remediation of depres-sive affect. Journal of Personality and Social Psychol-ogy, 58, 519–527.

Nolen-Hoeksema, S. (1991). Responses to depressionand their effects on the duration of depressive epi-sodes. Journal of Abnormal Psychology, 100, 569–582.

Nolen-Hoeksema, S. (2000). The role of rumination indepressive disorders and mixed anxiety/depressivesymptoms. Journal of Abnormal Psychology, 109,504–511.

Nolen-Hoeksema, S., McBride, A., & Larson, J. (1997).Rumination and psychological distress among be-reaved partners. Journal of Personality and SocialPsychology, 72, 855–862.

Nolen-Hoeksema, S., & Morrow, J. (1991). A prospec-tive study of depression and posttraumatic stresssymptoms after a natural disaster: The 1989 LomaPrieta earthquake. Journal of Personality and SocialPsychology, 61, 115–121.

Nolen-Hoeksema, S., & Morrow, J. (1993). Effects of


rumination and distraction on naturally occurringdepressed mood. Cognition and Emotion, 7, 561–570.

Notarius, C. I., & Levenson, R. W. (1979). Expressivetendencies and physiological response to stress. Jour-nal of Personality and Social Psychology, 37, 1201–1204.

Ortony, A., Clore, G. L., & Collins, A. (1988). The cog-nitive structure of emotions. Cambridge, UK: Cam-bridge University Press.

Palfai, T. P., & Salovey, P. (1993). The influence of de-pressed and elated mood on deductive and inductivereasoning. Imagination, Cognition, and Personality,13, 57–71.

Pennebaker, J. W. (1997). Writing about emotional ex-periences as a therapeutic process. Psychological Sci-ence, 8, 162–166.

Pennebaker, J. W., & Harber, K. D. (1993). A socialstage model of collective coping: The Loma Prietaearthquake and the Persian Gulf War. Journal of So-cial Issues, 49, 125–145.

Petrides, K. V., & Furnham, A. (2003). Trait emotionalintelligence: Behavioural validation in two studies ofemotion recognition and reactivity to mood induc-tion. European Journal of Personality, 17, 39–57.

Prkachin, K. N., Craig, K. B., Papageorgis, D., & Reith,G. (1977). Nonverbal communication deficits and re-sponse to performance feedback in depression. Jour-nal of Abnormal Psychology, 86, 224–234.

Saarni, C. (1990). Emotional competence: How emo-tions and relationships become integrated. In R. A.Thompson (Ed.), Nebraska Symposium on Motiva-tion (Vol. 36, pp. 115–182). Lincoln: University ofNebraska Press.

Saarni, C. (1997). Emotional competence and self-regulation in childhood. In P. Salovey & D. J. Sluyter(Eds.), Emotional development and emotional intelli-gence: Educational implications (pp. 35–66). NewYork: Basic Books.

Salovey, P., Bedell, B. T., Detweiler, J. B., & Mayer, J. D.(1999). Coping intelligently: Emotional intelligenceand the coping process. In C. R. Snyder (Ed.),Coping: The psychology of what works (pp. 141–164). New York: Oxford University Press.

Salovey, P., & Grewal, D. (2005). The science of emo-tional intelligence. Current Directions in Psychologi-cal Science, 14, 281–285.

Salovey, P., & Mayer, J. D. (1990). Emotional intelli-gence. Imagination, Cognition and Personality, 9,185–211.

Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., &Palfai, T. P. (1995). Emotional attention, clarity, andrepair: Exploring emotional intelligence using theTrait Meta-Mood Scale. In J. W. Pennebaker (Ed.),Emotion, disclosure, and health (pp. 125–154).Washington, DC: American Psychological Associa-tion.

Salovey, P., & Rodin, J. (1986). Differentiation ofsocial-comparison jealousy and romantic jealousy.

Journal of Personality and Social Psychology, 50,1100–1112.

Salovey, P., & Rodin, J. (1989). Envy and jealousy inclose relationships. Review of Personality and SocialPsychology, 10, 221–246.

Salovey, P., Stroud, L. R., Woolery, A., & Epel, E. S.(2002). Perceived emotional intelligence, stress reac-tivity, and symptom reports: Further explorations us-ing the Trait Meta-Mood Scale. Psychology andHealth, 17, 611–627.

Schaffer, L. F., Gilmer, B., & Schoen, M. (1940). Psy-chology. New York: Harper.

Schutte, N. S., Malouff, J. M., Hall, L. E., Haggerty, D.,Cooper, J. T., Golden, C. J., et al. (1998). Develop-ment and validation of a measure of emotional intel-ligence. Personality and Individual Differences, 25,167–177.

Schwarz, N. (1990). Feelings as information: Informa-tional and motivational functions of affective states.In E. T. Higgins & E. M. Sorrentino (Eds.), Hand-book of motivation and cognition (Vol. 2, pp. 527–561). New York: Guilford Press.

Schwarz, N., & Clore, G. L. (1983). Mood, mis-attribution, and judgments of well-being: Informa-tive and directive functions of affective states. Jour-nal of Personality and Social Psychology, 45, 513–523.

Schwarz, N., & Clore, G. L. (2003). Mood as informa-tion: Twenty years later. Psychological Inquiry, 14,294–301.

Segal, J. (1997). Raising your emotional intelligence: Apractical guide. New York: Holt.

Showers, C. J., & Kling, K. C. (1996). The organizationof self-knowledge: Implications for mood regula-tion. In L. L. Martin & A. Tesser (Eds.), Strivingand feeling: Interactions among goals, affect, andself-regulation (pp. 151–173). Mahwah, NJ:Erlbaum.

Simmons, S., & Simmons, J. C. (1997). Measuring emo-tional intelligence. Arlington, TX: Summit.

Steiner, C., & Perry, P. (1997). Achieving emotional lit-eracy: A personal program to increase your emo-tional intelligence. New York: Avon.

Sternberg, R. J. (1985). Beyond IQ: A triarchic theoryof human intelligence. Cambridge, UK: CambridgeUniversity Press.

Sternberg, R. J. (1988). The triarchic mind: A new the-ory of human intelligence. New York: Viking.

Sternberg, R. J., & Smith, C. A. (1985). Social intelli-gence and decoding skills in nonverbal communica-tion. Social Cognition, 3, 168–192.

Stroebe, W., & Stroebe, M. (1996). The social psychol-ogy of social support. In E. T. Higgins & A. W.Kruglanski (Eds.), Social psychology: Handbook ofbasic principles (pp. 597–621). New York: GuilfordPress.

Tangney, J. P., & Salovey, P. (1999). Problematic socialemotions: Shame, guilt, jealousy, and envy. In R. M.Kowalski & M. R. Leary (Eds.), The social psychol-


ogy of emotional and behavioral problems: Inter-faces of social and clinical psychology (pp. 167–195).Washington, DC: American Psychological Associa-tion.

Tangney, J. P., Wagner, P. E., Fletcher, C., & Gramzow,R. (1992). Shamed into anger?: The relation ofshame and guilt to anger and self-reported aggres-sion. Journal of Personality and Social Psychology,62, 669–675.

Terman, L. M. (1921). Second contribution to “Intelli-gence and its measurement: A symposium.” Journalof Educational Psychology, 12, 127–133.

Tett, R. P., Fox, K. E., & Wang, A. (2005). Developmentand validation of a self-report measure of emotionalintelligence as a multidimensional trait domain.Personality and Social Psychology Bulletin, 31, 859–888.

Thorndike, E. L. (1920). Intelligence and its uses.Harper’s Magazine, 140, 227–235.

Thorndike, R. L., & Stein, S. (1937). An evaluation of

the attempts to measure social intelligence. Psycho-logical Bulletin, 34, 275–284.


Treynor, W., Gonzalez, R., & Nolen-Hoeksma, S.(2003). Rumination reconsidered: A psychometricanalysis. Cognitive Therapy and Research, 27, 247–259.

Wagner, R. K., & Sternberg, R. J. (1985). Practical intel-ligence in real-world pursuits: The role of tacitknowledge. Journal of Personality and Social Psy-chology, 50, 737–743.

Wegner, D. M., Erber, R., & Zanakos, S. (1993). Ironicprocesses in the mental control of mood and mood-related thought. Journal of Personality and SocialPsychology, 65, 1093–1104.

Woodworth, R. S. (1940). Psychology (4th ed.). NewYork: Holt.

Young, P. T. (1936). Motivation of behavior. New York:Wiley.


C H A P T E R 3 4

Some Ways in Which PositiveAffect Influences Decision

Making and Problem Solving

ALICE M. ISEN

In the years since the publication of the secondedition of this handbook, interest in the topicof affect or emotion has grown tremendously,both among researchers and in the popularpress. Not only has this growth been evident inseveral subareas of psychology, and in relatedapplied domains such as consumer behaviorand organizational behavior; it has also cap-tured attention in fields such as economics andfinance.

Still, it remains true that, with a few notableexceptions, understanding of emotion and therole that common, everyday mild affect (feel-ings, emotion) plays in people’s thinking andbehavior is still relatively rudimentary. Mostpeople seem to have a sense that affect can in-fluence their decisions and thought processes,but it is still usually assumed that affect’s influ-ence is something irregular or unusual; thatonly strong and infrequent feelings will havesuch influence; and that most often only basicnegative feelings (such as anger, sadness, orfear) will have an impact on thinking processes.Furthermore, most people assume that whenaffect—even positive affect—plays a role in

their decision processes, such influences aredisruptive and tend to make their decisions “ir-rational” and less appropriate than otherwise.

The research that has been carried out, how-ever, indicates that even mild and even positiveaffect can markedly influence everyday thoughtprocesses, and does so regularly. Some of theearliest work in the field showed that mild pos-itive feelings cue positive material in memory,making access to such thoughts easier, and thusmaking it more likely that positive material will“come to mind,” while at the same time notimpairing access to negative material when thatis relevant (e.g., Isen, Shalker, Clark, & Karp,1978; Teasdale & Fogarty, 1979).

This effect on memory reflects the fact thatmaterial in mind is organized and accessible interms of its positive affective tone, and thatpeople spontaneously use positive affect as away to organize their thoughts (Isen, 1987).The same is not true for negative affect such assadness, though, and I will say more about theasymmetry later in this chapter. Thus the evi-dence indicates that, far from being an infre-quent influence on thought processes, common

548

positive feelings are fundamentally involved incognitive organization and cognitive process-ing.

However, mild positive affect has even morefar-reaching effects on thinking and decisionmaking than are implied by its effects on mem-ory, and even beyond its effects on cognitive or-ganization. Positive affect has also been foundto promote flexibility in thinking—which inturn has been shown to facilitate problem solv-ing (including creative problem solving) andinnovation, as well as both efficiency and thor-oughness in decision making, and to enableimproved thinking, especially where tasksare complex (e.g., Carnevale & Isen, 1986;Derryberry, 1993; Estrada, Isen, & Young,1997; Greene & Noice, 1988; Fredrickson,2001; Isen, Daubman, & Nowicki, 1987; Isen,Rosenzweig, & Young, 1991; Kazen & Kuhl,2005; Kuhl & Kazen, 1999; Roehm &Sternthal, 2001; Staw & Barsade, 1993). Theseeffects have been found in a wide range of set-tings and populations, ranging from youngadolescents (e.g., Greene & Noice, 1988) topracticing physicians engaging in diagnosticreasoning (Estrada et al., 1997), and in the lit-erature on consumer behavior (e.g., Barone,Miniard, & Romeo, 2000; Roehm &Sternthal, 2001) and on behavior in organiza-tions (e.g., Amabile, Barsade, Mueller, & Staw,2005; Staw & Barsade, 1993). Furthermore,recent studies in the coping literature are find-ing that even when people must cope withadverse events, positive affect is helpful, facili-tating effective coping and reducing defensive-ness (e.g., Aspinwall & MacNamara, 2005;Aspinwall & Taylor, 1997; Fredrickson, 2001;Trope & Pomerantz, 1998).

Thus the influence of mild positive feelingson thinking and decision making has beenfound to be not only substantial but facilita-tive under many circumstances, leading to im-proved decision making and problem solving.How can these two views—people’s intuitionsand the findings of such studies—be recon-ciled?

First, it is likely that when people typicallythink about the effects of emotion, they tendnot to pay a lot of attention to mild affect, andtherefore are less likely to attribute influence toit. Second, they may be especially unlikely topay attention to common, mild positive feel-ings. Third, the influence of positive affect isusually facilitative, but we often take positiveoutcomes of everyday processes more or less

for granted, not searching for causes or connec-tions unless things go wrong or performance isimpaired (e.g., Weiner, 1985). Thus we may notnotice facilitative effects of positive feelings,but may pay more attention when we have rea-son to think that happy feelings have impairedpeople’s judgment or problem solving.

All of these phenomena most probably con-tribute to people’s underestimation of positiveaffect as a facilitating factor in decision makingand problem solving. In addition, results ofsome studies have also indicated that there aretimes when positive affect actually does atleast appear to impair or interfere with prob-lem solving or thinking (e.g., Bless, Bohner,Schwarz, & Strack, 1990; Bodenhausen,Kramer, & Susser, 1994; Mackie & Worth,1989; Melton, 1995). However, most studiesreporting such effects find specific conditionsunder which impairment is observed, ratherthan general impairment; they also tend to in-volve situations in which participants may notbe engaging in the task, or not fully engaging init, on account of aspects of the task and theway it is presented. For example, Bodenhausenet al. (1994) reported that people in whom pos-itive affect had been induced were more likelyto use a stereotype—a process assumed to im-ply superficial processing—in making judg-ments about another person, but only when thetask was not relevant to them. By contrast, inBodenhausen et al.’s Study 4, when peopleknew that they would have to account for theirjudgment, those in positive affect did notengage in more stereotyping than controls,and even tended to stereotype less than con-trols.

Similarly, the majority of studies addressingpositive affect’s influence on thinking do not re-port impaired performance, but rather indi-cate more flexibility in information-processingstrategies, and generally enhanced thinking andproblem solving whenever a person has reasonto engage a problem fully. This point was morerecently noted by Bodenhausen, Mussweiler,Gabriel, and Moreno (2001), in their summaryof the influence of positive affect on stereotyp-ing, when they concluded, “Claims that happypeople are generally unable or unwilling to en-gage in systematic thinking appear to be inac-curate. Rather, happy people appear to be flexi-ble in their information-processing strategies(cf. Isen, 1993)” (Bodenhausen et al., 2001,p. 334). Thus one important question will be totry to understand what factors play a role in

34. Positive Affect, Decision Making, and Problem Solving 549

determining the influence of positive affect onperformance in different situations, and, moreimportantly, what processes underlie these in-fluences. This chapter argues that flexibility,rather than superficiality or “playfulness”(both of which have been suggested in the liter-ature), is a better way to characterize the effectof positive affect on thinking and decisionmaking.

In order to understand affect’s influence ondecisions, it is necessary to consider its impacton processes underlying decision making, suchas cognitive organization (the way material isthought about and related to other material),cognitive flexibility (the ability to think aboutmaterial in multiple ways or to change the wayone is thinking about the material as needed),focus of attention, and motivation. This is be-cause organization, context, and motivation,including a person’s goals and the way he orshe is thinking about the situation, play crucialroles in decision making (e.g., Simonson, 1989;Tversky & Kahneman, 1981). In fact, as wasevident in considering the findings reported byBodenhausen et al. (1994) some of the reasonsfor the discrepancy among the studies, men-tioned above, may lie in these other effects ofpositive feelings, such as motivation and senseof freedom to behave as one thinks appropriateor to pursue one’s goals. In addition, manyother effects of positive affect—including somerelating to expectancy motivation, intrinsicmotivation, self-control, integration of context,and aspects of cognitive flexibility such as flexi-bility in focus of attention, are being found tohave basic influences on thinking and behavior,and need to be integrated into consideration ofthe decision-making process. Finally, the chap-ter briefly mentions a theoretical account at theneurological level (the dopamine hypothesis;Ashby, Isen, & Turken, 1999) that sheds lighton why and how positive affect may have its ef-fects on thinking and decision making, as wellas on cognitive flexibility, the process that is socentral to its influence on thinking and behav-ior.

To address all of these topics in this singlechapter, consideration of each must be brief. Ihope that the chapter will serve as a spring-board for thought, even though it cannot be ex-haustive or provide a detailed review of all ofthe literature in this rapidly growing field. I be-gin, then, by considering the influence of posi-tive affect on cognitive flexibility and organiza-tion.

POSITIVE AFFECT,COGNITIVE FLEXIBILITY,AND COGNITIVE ORGANIZATION

The Flexibility Hypothesis

There is growing evidence that positive affectleads people to be more flexible thinkers and de-cision makers—able to consider multiple as-pects of situations and switch attention amongthem, and thus to respond effectively to com-plex or changing circumstances. This evidencecomes from studies showing that positive affectpromotes such processes as elaboration, respon-siveness to context, creative problem solving,and flexible focus of attention, all of which re-sult in changes in cognitive organization and theability to see things in new or multiple wayswithout losing sight of the usual ways. Note that“flexible” does not mean unconstrained orloose, and is not a synonym for “creative” or“easygoing”; even though flexibility can lead toinnovation, it is based on assessment of the situ-ation, responsiveness to the context, and thusappropriateness and effectiveness in the situa-tion. Theoretical support for the idea that thiskind of flexibility results from positive affect co-mes from the dopamine hypothesis, notedabove; this hypothesis suggests that positive af-fect involves release of dopamine in the brain,activating brain regions that include frontal ar-eas responsible for high-level thinking, execu-tive processes such as working memory, andprocesses such as ability to switch attention andresolve conflicting stimuli. This increased flexi-bility seems responsible for many of the effectsof positive affect, including improved problemsolving and decision making.

Elaboration

Considerable research indicates that positiveaffect influences cognitive organization—theway in which stimuli or ideas are thoughtabout or related to other ideas in mind. The ef-fects are thought to occur through processesof elaboration and ability to switch flexiblyamong elaborated ideas. But this process ofelaboration has been found to depend on thenature of the materials.

Range of Associations, Categorization,and Similarity–Difference Judgment

Studies have shown that people in whom posi-tive affect has been induced in any of several


simple ways—such as watching 5 minutes ofa nonsexual, nonaggressive comedy film; re-ceiving (not eating) a cute, small bag of 10wrapped pieces of hard candy, given as a tokenof appreciation; or responding with word asso-ciations to positive words—have a broaderrange of word associates, and more diverse as-sociates, to neutral material (Isen, Johnson,Mertz, & Robinson, 1985). Similarly, people insuch conditions are able to categorize materialmore flexibly, seeing ways in which nontypicalmembers of categories can fit into or be viewedas members of these categories, while not los-ing the understanding that typical members ofthe categories are better exemplars of them(Isen & Daubman, 1984; Isen, Niedenthal, &Cantor, 1992; Kahn & Isen, 1993). This hasbeen found for items in natural categories, forproducts in the mildly pleasant class of snackfoods, and for person types in positive—butnot in negative—person categories. Thus posi-tive affect enables people to see more similari-ties or connections among items (and they havealso been found to see more differences, iflooking for differences; e.g., Murray, Sujan,Hirt, & Sujan, 1990).

The process that underlies these effects, assuggested by the word association findings,may be that people experiencing positive affectengage in greater elaboration about the mate-rial and reason flexibly about it, seeing bothsimilarities and differences. Thus they see moreaspects of the items and have more thoughtsabout them and associations to them. Then, asexplained by Tversky and Gati (1978) forknowledge about material in general (not affec-tive material or induced affect specifically), thecontext supplied by the task (looking for simi-larities vs. differences) determines whether thisgreater elaboration (greater knowledge, in thework by Tversky & Gati) results in a judgmentof greater similarity or greater difference.

Context: Nature of the Materials,Importance of the Task,Potential Benefit–Cost

Valence of the Materials

One aspect of the context that is critically im-portant in determining the impact of positiveaffect is the nature, including the valence, ofthe materials being considered. In accord withresults of studies showing that positive affectcues positive material in memory and enablesreadier access to such material and a broader

range of associates, presumably people in posi-tive affect will see more positive aspects ofrelatively neutral material, and will rate suchmaterial more positively than at another time.But nothing in their positive state would be ex-pected to prompt their rating negative stimuli(about which there are few if any positive as-pects to be cued) more positively. Similarly,they would not be expected to rate mildly nega-tive material as more negative, as if just judgingeverything as more extreme. Likewise, with al-ready positive material, one would not expectpositive affect to lead to higher evaluation (be-cause the positives are already clear to peoplein all conditions of the experiment).

These expectations have been upheld in sev-eral studies. One study, for example, showedthat positive affect influenced the categoriza-tion of relatively neutral person types intopositive categories, but did not change the per-ceived category fit of neutral types into nega-tive categories (Isen et al., 1992). For anotherexample, a study of the impact of positive af-fect on job perceptions and satisfaction foundthat people in whom positive affect had beeninduced perceived an interesting task that theyhad been assigned, but not a meaningless one,as richer and more satisfying (Kraiger, Billings,& Isen, 1989). Again, this can be seen as re-flecting an ability of those experiencing posi-tive affect to see additional associations and as-pects of interesting things, not a global effecton ratings. Another paper reporting resultssupportive of this point (Isen & Reeve, 2005)investigated the influence of positive affect onintrinsic motivation and self-control. In twoexperiments, induced positive affect led to in-creased liking for an enjoyable task, but did notaffect liking for a dull, routine work task.Finally, in two studies looking at the influenceof positive affect on components of expectancymotivation, results showed that people in posi-tive affect rated the attractiveness of moder-ately attractive, but not unattractive or ex-tremely attractive, outcomes as more positive(Erez & Isen, 2002). Thus both theory and em-pirical evidence suggest that one cannot expectpositive affect to influence the rating, the cate-gorization, or the perceived similarity (or dif-ference) of all stimuli in the same way, regard-less of the valence of the materials and thesituation. Although an underlying process (in-creased elaboration) is postulated to occur, thisprocess is expected to have different effects ondifferent kinds of material.


Importance or Usefulness of the Task

Besides valence, other aspects of the situationplay a role in the specific effect that positive af-fect will have. As discussed in the context of thework by Bodenhausen et al. (1994) on stereo-type use, the importance or relevance of thetask has been shown to play a role in the im-pact of positive affect on judgment and deci-sions. However, beyond that, this aspect of thesituation has even been found to influence theeffect that valence itself will have. For example,whether the context is one of danger versus en-joyment, and whether the task is or is notimportant or useful, both influence the waypeople in positive affect respond to negativematerial. If the task to be done, or material tobe evaluated, is boring or negative, and if thereis no purpose or benefit to paying attention toit, then positive affect may result in people’s notdealing with the material or not dealing with itcarefully. This tendency may be more notableamong people in positive affect because—allelse being equal—they may be seeking to main-tain that positive state (e.g., Isen & Simmonds,1978; Wegener, Petty, & Smith, 1995), andthey may feel freer to behave as they think ap-propriate (Forest, Clark, Mills, & Isen, 1979).Whether such a reaction occurs, however, willdepend on the importance of the task. Wherethe task is important, but requires people to fo-cus on possible meaningful loss or to cope withdifficult situations, the evidence shows thatpeople in positive affect face the situation di-rectly and deal with it. For example, in a high-risk situation, those in positive affect have beenfound to have more thoughts about losing thando controls, and to behave more conservativelyso as to protect themselves from the loss (e.g.,Isen & Geva, 1987; Isen, Nygren, & Ashby,1988). This is also compatible with data show-ing that positive affect promotes effective cop-ing in negative or stressful situations that haveto be dealt with or in which there is some bene-fit to be obtained from attending to the task(e.g., Aspinwall & MacNamara, 2005;Aspinwall & Taylor, 1997), and that positiveaffect also reduces “defensiveness” (e.g.,Aspinwall, 1998; Trope & Pomerantz, 1998).The way positive affect influences response tonegative material, then, depends on other as-pects of the context and situation.

One point illustrated in this discussion isthat although people in positive affect may befound to ignore negative material if there is no

reason in the situation to attend to it, positiveaffect does not lead to distortion of negativematerial, and it does not lead to ignoring ofnegative material if attending to that materialcan benefit the person in either the long term orshort term. In addition, the interactions be-tween positive affect and type of task or situa-tion illustrate that a substantive process relatedto elaboration and flexible thinking character-izes the influence of positive affect and under-lies the observed effects—not an artifact suchas response bias, “seeing things through rose-colored glasses,” over-inclusiveness in categori-zation, or generally poor (nonsystematic, im-paired, or sloppy) processing. Another point il-lustrated here is that use of negative material instudies of the influence of positive affect (suchas occurred in attitude change studies) canmake for complexity in formulating predic-tions.

Creative Problem Solving

An ability that is greatly desired by most peo-ple, organizations, and society at large iscreativity—and especially the capacity for in-novation or creative problem solving. This is aset of skills that evidence shows is fostered bymild, everyday positive affect. As we have seen,positive affect facilitates innovative but sensi-ble responding in measures of word associa-tion, and in categorization of plausible cate-gory members that are typically not seen asmembers of the category (e.g., “elevator” in thecategory “vehicle”). Such responding can beseen as involving cognitive flexibility or theability to put ideas together in new ways, andperhaps it reflects not only a change in cogni-tive organization but also in creativity. Com-patibly with that suggestion, another seriesof studies (e.g., Isen, Daubman, & Nowicki,1987) indicates more specifically that positiveaffect does enable creative problem solving, asmeasured by performance on tasks such asDuncker’s (1945) candle problem and items ofthe Remote Associates Test (RAT; Mednick,Mednick, & Mednick, 1964).

It is important to note that these tasks—theDuncker task and the RAT, which are usuallyconsidered to require innovative respondingand to reflect creativity—are not simply mea-sures of fluency or expansive thinking, butrequire specific answers to specific, difficultproblems. In the Duncker task, the participantis given a book of matches, a box of tacks, and


a candle, and is told to affix the candle to thewall (a cork board) so that the candle can be litand will not drip wax on the table or floor. Theproblem can be solved if the tacks are removedfrom the box and the box is then tacked to thewall and used as a holder for the candle. In theRAT, which involves word problems in eachitem, the person is given three words andneeds to think of a fourth word that fits witheach of the other three (e.g., for the stem“mower, atomic, foreign,” the correct answeris “power”). These tasks are very difficult: Forexample, on the candle task, the rate of correctsolution in the control conditions is about only11–16%; that rate rises to about 58–75% inthe positive affect conditions.

Thus it is not correct to think of creativeproblem solving as involving only divergent,expansive thinking or unconstrained, loosethinking. Some authors have done that, andthus have dismissed the importance of successon these tasks as shedding light on the basic ef-fect that positive affect has on thinking and de-cision making. Those authors maintain thatpositive affect, when it affects thinking andbehavior, reduces thinking to superficial, care-less, non-detail-oriented, schematic thought,and they argue that the findings with regard tocreative problem solving can be explainedas involving only playfulness and superficialthinking (e.g., Forgas, 2002; Schwarz, 2002;Schwarz & Bless, 1991). In actuality, successfulsolutions to these kinds of problems requirenot only inventiveness in coming up with possi-ble solutions, but also convergent, focusedthinking, to bring the new ideas to bear accu-rately on the task that needs to be accom-plished. If expansive thinking is required, soalso are inhibition of that process and con-strained thinking, in a reciprocal process thatapplies the novel ideas to the current situationand utilizes them to solve the specific problem.Neither is there reason to interpret perfor-mance on either of these tasks (or others shownto be facilitated by positive affect, such as effec-tive negotiation between parties with conflict-ing interests; e.g., Carnevale & Isen, 1986) asevidence of having taken a “playful” approach(e.g., Schwarz, 2002).

It should also be noted that in the workshowing that positive affect facilitates creativeproblem solving, the effects have been foundwith both verbal (e.g., the RAT items) and spa-tial (the candle problem) tasks as assessmentsof creative problem solving. This, then, is not

compatible with the hypothesis put forth byGray (2001) that positive affect facilitates per-formance on verbal tasks but impairs perfor-mance on spatial tasks. The reasons for the dis-crepancy between Gray’s finding and that ofthe work on creative problem solving are notclear at present, but perhaps they will becomeso in the near future.

Focus of Attention

Recently there has been increased interest inthe influence of affect on focus of attention—particularly on the ability to switch among rel-evant stimuli, and on broad versus narrow fo-cus, or focus on global versus local aspects ofthe context (e.g., Baumann & Kuhl, 2005;Derryberry, 1993; Fredrickson, 2001; Fried-man & Forster, 2005; Gasper & Clore, 2002;Isen & Shmidt, 2007). This topic is relevant todecision making because it deals not only withthe aspects of stimuli or situations that peoplenotice first in a situation (which could influ-ence their decisions), but also with whetherpeople can consider multiple aspects of situa-tions as they proceed in decision making.

It has been known for a long time that nega-tive affect in the form of anxiety or arousal(sometimes referred to as “emotion”), or “highdrive” (in the language of the learning theoriesof the 1940s and 1950s), narrows the focus ofattention, impairs cue utilization, and canimpair performance as a result (e.g., seeEasterbrook, 1959; Isen, 2008). More recently,affect, arousal, and motivation have been dis-tinguished and studied separately, as have spe-cific emotions and positive and negative affect(although there continues to be some confusionamong them). Some models of emotion distin-guish between affect and arousal (see, e.g.,Russell & Carroll, 1999; Watson & Tellegen,1999, for discussions of these models), andsome research has investigated valence andarousal separately (e.g., Mano, 1994; Sanbon-matsu & Kardes, 1988). Similarly, some workhas contrasted effects of positive and negativeaffect and affectless arousal (induced by simplephysical activity; e.g., Isen, Daubman, &Nowicki, 1987).

Thus, as positive and negative affect havebeen distinguished and studied independently,there have been new investigations of the influ-ence of positive and negative affect on focus ofattention, to see whether positive affect (in con-trast with negative) might actually broaden fo-


cus. The idea that positive affect might result inthe ability to broaden attention was suggestedin an early study of affect’s influence on helpingbehavior (see Isen, 2008, for a discussion). Par-ticipants were asked a set of questions about aperson and activities that had unexpectedly oc-curred in the room where they were waiting forthe experimenter to return. Not only did peo-ple experiencing negative affect (created by re-port of failure on a task) perform significantlymore poorly on the surprise question about thesurroundings; those in whom positive affecthad been induced (by report of success on thetask) did not have narrowed attention, as peo-ple in the negative affect condition did, but ac-tually reported significantly more correct in-formation about what had happened aroundthem.

Results compatible with these were reportedby Derryberry (1993), using a different taskand materials. He found that success resultedin the ability to attend to low-value targetswithout missing high-value targets, whereasfailure reduced the ability to attend to the low-value targets while the person was attending tothe high-value targets. Thus it appears thatpositive affect (success) facilitates broader de-ployment of attention with no loss of accuracyor speed of processing, and that negative affect(failure) reduces breadth of attention in orderto remain accurate on a subset of the items.Derryberry (1993, p. 84) summarized the workon this effect as best described as showing thatstressful states “narrow” focus rather than“concentrating” it, because they impaired per-formance on the secondary information with-out improving performance on the primarytargets. With regard to positive affect, he con-cluded that attention broadens, but that thisbroadening is not accompanied by dilution ofresources devoted to the primary focus.

More recently, authors have begun to inves-tigate the question of focus of attention dur-ing positive and negative affect in a differentway—by looking at matters such as, literally,whether people, when shown a display, tend touse its broad features (“global focus”) or itsmore narrowly focused features (“local focus”)to understand what they are seeing (e.g.,Gasper & Clore, 2002). For example, somestudies use a task similar to the one developedby Navon (1977) or Kimchi and Palmer(1982), in which a stimulus is presented to par-ticipants that is a shape constructed out ofsmaller building blocks of another shape. In

these studies, people are asked to identify theshape that is presented, and responding interms of the superordinate shape is taken as ev-idence of having a global focus; responding interms of the building-block shape is regardedas reflecting a local focus. According to thoseauthors, most people respond based on thebroader focus, indicating a hierarchical organi-zation of perception. But the point of the re-search on affect and focus has been to see howpositive and negative affect influence focus ofattention. This research has reported that peo-ple in negative affect tend more to process thenarrow dimension initially, and that people inpositive affect tend to use the global dimensioninitially, just as controls do (Gasper & Clore,2002). Some researchers assume that thismeans that people in positive affect can onlyprocess broadly, that they cannot focus nar-rowly, or that they will have difficulty (or showimpaired processing and distractibility) whenrequired to focus narrowly.

Baumann and Kuhl (2005) addressed thisquestion specifically, by having the experimen-tal participants solve a problem that requiredusing both the broader dimension and the nar-rower dimension in order to obtain the correctsolutions. Their data show that when a focuson the narrow dimension was necessary in or-der to solve the problem, people in positive af-fect opted for that focus and performed well, asreflected by both reaction time and error rates.These findings then, support the view that posi-tive affect enables flexible deployment of atten-tion, rather than merely broadening attentionor limiting people to the one perspective. Theyare thus also compatible with the work re-ported by Derryberry (1993), showing thatpositive affect leads to broad or narrow focusas needed (attention to secondary targets with-out cost in terms of the processing of primarytargets)—that is, to more flexible processing.

It must be noted that what Gasper and Clore(2002) actually found was that the positive af-fect condition and the control condition didnot differ, and that the participants experienc-ing negative affect were more narrowly focusedthan the controls. The normal focus for peopleis the broader one, and, as noted above, it isthought to reflect the normal hierarchical or-ganization of perception. Although in placesGasper and Clore (2002) are careful to refer tothe difference between the positive and nega-tive conditions in terms of negative affect’scausing a narrower focus, they do imply that


their results indicate that positive mood fostersa global focus—which they did not actuallyfind. Furthermore, they assert the points, madein more detail elsewhere (e.g., Bless et al.,1996; Schwarz, 2002), that positive affect canbe assumed to prompt less careful or less de-tailed processing (less detailed processing, theybelieve, would correspond to the more typicalglobal focus), and that negative affect can beassumed to foster more intensive processing(which they argue would correspond to themore local focus). Gasper and Clore acknowl-edge that their findings are not supportive ofsuch views: They did not find, for example, evi-dence that negative affect increased processingor facilitated performance of any kind; nor didthey find that positive affect led to less carefulprocessing. However, these authors invite con-fusion by their repeated statements of those un-supported hypotheses, as if their findings hadconfirmed them.

Flexibility in Focus of Attention

Consequently, it is useful to consider the find-ings of other authors, such as those presentedabove—Baumann and Kuhl (2005), Derryberry(1993), and others—to provide more contextfor the paper by Gasper and Clore (2002). Assuggested by Derryberry and colleagues, nar-row focus may not indicate more extensive ordetailed processing, and the broad focus shownby the positive feedback group in his studies cor-responds to more (not less) extensive, detailed,and effective processing. This interpretation fitswith the flexibility hypothesis—which predictsimproved ability to change focus as appropriateand switch among stimuli, as a result of positiveaffect, and which is also compatible with thedata described earlier showing that positive af-fect broadens the range of material thoughtabout, resulting in more diverse associations,flexible categorization, and more effective waysof solving problems.

Another program of research dealing withthe question of focus, but in a different way, isthat of Fredrickson, who postulates that peoplein positive affect “broaden and build” theirskills, knowledge, and resources and then canuse them in multiple ways, and who describes“upward spirals” of well-being as resultingfrom positive feelings (e.g., Fredrickson, 2001;Fredrickson & Branigan, 2005; Fredrickson &Joiner, 2002). That is, not only do people inpositive affect look at the bigger picture, but

they develop broader thought–action reper-toires and build their resources as they do so.Although Fredrickson and her colleagues donot address flexibility directly, this work wouldbe compatible with the interpretation (andfindings described above) that people in posi-tive affect are flexible and not limited to thewider perspective, even though they may beable to access it and use it more readily thanthose in comparison affect states.

Applications of Broadened Perspective

Another study that may relate to positive af-fect’s broadening of focus or perspective is onementioned earlier, showing that positive affectcan facilitate the process of negotiation and re-sult in improved outcomes in an integrativebargaining situation (Carnevale & Isen, 1986).This may be an application of the broaden-and-build hypothesis, and an example of theintegration of context into a situation, becauseit shows that the broadened perspective is usedto put information together to solve a problemand come up with an integrative solution. It iscertainly an example of the flexibility that ac-companies positive affect, because people expe-riencing positive affect, in contrast with con-trols, were better able to figure out the otherperson’s payoff matrix and integrate informa-tion to come up with good solutions for theproblem (see Isen, 2000, for a discussion).

Does Broadened Focus Imply Increasein Distractibility?

A related topic is whether this broadened fo-cus, or increased ability to switch among as-pects of a stimulus display or tasks, implies de-creased ability to remain focused on a focaltask that one is trying to perform. Some au-thors (e.g., Dreisbach & Goschke, 2004) havehypothesized, and found, that positive affectreduces perseveration (i.e., increases ability toswitch when the context calls for it), but that italso increases distractibility (i.e., reduces abil-ity to stay on target on a focal task when thereare potentially distracting stimuli present). Incontrast, work by others indicates no such lossof ability to perform well on the main targets(e.g., Derryberry, 1993; Friedman & Forster,2005; Isen & Shmidt, 2007; Kazen & Kuhl,2005; Kuhl & Kazen, 1999). In addition, theflexibility hypothesis proposes that people ex-periencing positive affect should not become


stuck on a broadened perspective, or unable tofocus on the task they are trying to perform;rather, they should be able to switch attentionflexibly between the broad and the narrow, asneeded in the situation. This also would bemore compatible with Fredrickson’s (e.g.,2001) broaden-and-build hypothesis, becausein that view one uses the information fromthe broadened perspective to build something.Some additional evidence on this point mayalso be available in the experiments on creativeproblem solving, where the innovative orbroadened responses were successfully appliedto specific problems that needed to be solved.Thus the topic of affect and distractibility islikely to attract continuing research interest.

Affect and False Memory?

Still another question possibly related tobreadth of focus is whether positive affect maylead to an increase in false memory. A full con-sideration of this topic is beyond the scope ofthis chapter, but a few words on the subjectseem appropriate (see Roediger, McDermott,& Robinson, 1998, for a review, and Isen,2008, for a fuller discussion of this topic in thecontext of affect). False memory is studied inexperimental psychology with a procedure inwhich participants are presented with multiplelists of semantically associated words (e.g.,“snow,” “winter,” “ice”) that are also relatedto a critical item (e.g., “cold”) that is not pre-sented. Then they are given a recognition testthat presents words that were previously pre-sented and words that were not previously pre-sented, including the critical word (the one thatis related to the theme of the presented wordsbut was not itself presented), and asks partici-pants to say whether each word in the test waspresented before or not. The test for false mem-ory is whether the critical word is erroneouslyidentified as having been presented previously.Results in the cognitive psychology literatureindicate that false memory measured in thisway is very common, with people generallymaking this error as often as they correctlyidentify words that had been presented (seeRoediger et al., 1998).

It is easy to see how this task might bethought to relate to the effects of positiveaffect—either from the point of view of theflexibility hypothesis (which would expect noincrease in false memory from positive affect,and possibly even a decrease, if something in

the situation led people to realize the potentialfor such errors); or from the point of view ofthe researchers who believe that positive affectleads to superficial reliance on schematic pro-cessing, dependence on habitual or routineconcepts/structures that come to mind mosteasily, and failure to attend to detail. Those re-searchers presumably would expect increasedfalse memory with positive affect.

In fact, there have recently been some at-tempts to address this question, but thus farpositive affect has not been found to increasefalse memory (e.g., Storbeck & Clore, 2005,despite the title of their paper; Yang, Ceci, &Isen, 2006). In fact, one series of studies hasfound that when participants were told aheadof time about the possibility of such errors,people experiencing positive affect made fewerfalse-memory judgments than controls (Yang etal., 2006).

Summary

Thus, to summarize this section on positive af-fect, cognitive flexibility, and cognitive organi-zation, positive affect appears to influence theway in which cognitive material is organized—how ideas are related to one another in themind. In particular, positive affect has beenfound, in most situations, to give rise to greaterelaboration in response to neutral or positivestimulus material (but not negative or tediousmaterial unless there is a reason to engage it)and to a richer cognitive context, which in turnpromotes flexibility in thinking.

This means that in a task dealing with mate-rial of neutral to positive valence, and under-taken while a person is feeling happy, oneshould expect unusual and innovative (thoughreasonable and logical) thoughts and re-sponses. It is a mistake to assume that peopleexperiencing positive affect will think onlythose arguments and thoughts about the exper-imental materials suggested by the experiment-ers. Based on the research reviewed here, weshould expect people in positive affect to thinkabout the materials in a more elaborated,extensive, flexible, responsible way—providedthat the materials are not negative or boring, orprovided that there is sufficient reason for thesepeople to consider the materials carefully, evenif they are negative or boring.

In the case of negative material, it is moredifficult to predict the behavior of people inpositive affect. If there is not a clear reason to


focus on the negative material, we would notexpect these people to elaborate the negativematerial more than controls do. On some tasks(e.g., categorization, word association), thiswill result in their responses’ not differing fromthose of control subjects (e.g., Isen et al., 1985,1992); on other tasks, we would expect partici-pants in positive affect to avoid or show cau-tion with the materials, if either is possible inthe situation. This might mean that they wouldbe slower to engage in the task, or appear to beimpaired in ability to perform the task, com-pared with controls (e.g., Melton, 1995). How-ever, in situations where people need to dealwith the negative material, or where it wouldbe to the persons’ benefit (immediate or long-term) to do so, the research indicates that peo-ple in positive affect will show greater elabora-tion and enhanced coping even with nega-tive materials or problematic situations (e.g.,Aspinwall & Taylor, 1997; Isen & Geva, 1987;Trope & Pomerantz, 1998).

We have also seen that positive affect leadsto more flexibility in focus of attention, as wellas to broader thinking and ability to processand integrate more aspects of situations andstimuli. For now, the evidence is mixed with re-gard to whether there is a cost for the ability toengage this broader focus: Some authors reportthat positive affect does lead people to be moreeasily distracted, but others indicate that peo-ple in positive affect do not suffer increased dis-tractibility (as would be reflected by impairedperformance on the focal task when partici-pants are asked to perform a second task simul-taneously). Regarding the related hypothesisthat positive affect may lead to more falsememory, there is no evidence supporting this asyet, despite repeated attempts to find it; there isalso some evidence that positive affect im-proves monitoring and avoidance of such er-rors when people are made aware of the possi-bility of them.

POSITIVE AFFECT AND MOTIVATION

The basic principle to be kept in mind inconsidering motivation is that behavior ismultidetermined and results from the resolu-tion of many motives and factors, all poten-tially operative at the same time. A person’schoice of action depends on how he or she re-solves all of the competing influences, in thecontext of his or her goals, strategies, concerns,

and constraints in the given situation. Positiveaffect has several potential influences on moti-vation, not just one, and the motives that willgovern behavior, decisions, and choice in anygiven situation will depend on what the personbelieves is the most appropriate behavior underthose circumstances. It is not appropriate toisolate one motive that is engendered by posi-tive affect and assume that it, above all others,will necessarily determine behavior when a per-son is feeling happy.

For example, some authors, based on a lim-ited reading of the affect literature, single outmaintenance of positive affect as the primary(or only) motive of people in positive affect,and assume that the desire to stay happy willalways govern behavior and choice whensomeone is feeling happy. What the researchshows, however, is that this is far from true. Aswe have seen, people in positive affect are morelikely than controls to pay attention to negativeor threatening material if it is useful in eitherthe long term or short term, or if there is an-other reason to attend to it (e.g., Aspinwall,1998; Aspinwall & Taylor, 1997; Isen &Geva, 1987; Isen & Reeve, 2005; Trope &Pomerantz, 1998; see Isen, 2003, for a discus-sion). Thus a better way to think about the af-fect maintenance motive in positive affect isthat it will be considered, and, all else beingequal, people who are feeling happy will be-have in such a way as to remain so.

Not only do the various motivations associ-ated with positive affect need to be integrated inorder to predict resultant behavior, but the ef-fects of positive affect on cognitive processingalso must be taken into account at the same time.For example, positive affect leads to more likelyintegration of context, and greater flexibility infocus and interpretation of contextual factors;these processes and interpretations may also in-fluence what the person is motivated to do, orwhat strategies seem appropriate, in a situation.Flexibility seems a key to understanding how thevarious motives activated by positive affect willbe integrated and/or resolved.

Regarding a Global Increaseor a Global Decrease in Motivation

Two kinds of possible effects of feelings onglobal motivation have been suggested in theliterature: trying harder (as would follow fromthe view of emotion as arousal), and, ironically,trying less hard (as would follow from the


mood-as-information or “cognitive tuning”view).

Global Increase in Motivation

The suggestion that positive affect will increasemotivation in general, or will lead a person totry harder, seems intuitively correct to manypeople. All of the work that indicates that posi-tive affect promotes enjoyment and enrichmentof potentially enjoyable, though work-related,tasks suggests that positive affect may influencetask motivation (because richer tasks and taskenjoyment are also more motivating; e.g.,Kanfer, 1990). However, no evidence as yetsuggests that positive affect simply raises efforton all tasks, as the concept of arousal has beenthought to do. Rather, it has been found thatpositive affect leads people to try harder onsome tasks or in some situations—those inwhich there is a reason to try or it seems thattheir effort will make a difference—but not onother tasks or situations in which there is nopoint in trying (e.g., Erez & Isen, 2002; Isen &Reeve, 2005).

General Decrease in Motivation

Another possibility that has been suggested isthat positive affect will lead people to try lesshard—an implication of the mood-as-information or cognitive tuning view that posi-tive affect reduces motivation and leads to lesseffort and/or care, because it signals that all iswell (e.g., Schwarz, 2002; Schwarz & Bless,1991). This effect on motivation seems un-likely, in view of the evidence (summarized inthis chapter and elsewhere) showing that peo-ple in positive affect often do voluntarily tryhard and perform well, when the task is eitherinteresting or important, with no sign that suchperformance has taken special effort and ex-hausted them, required a sense of ill ease,or spoiled their mood (Ashby et al., 1999;Bodenhausen et al., 1994, 2001; Erez & Isen,2002; Isen, 1993). Given the lack of evidencefor either type of global effect of positive affecton motivation, it seems more promising to askabout specific aspects of tasks and task motiva-tions that seem to result from positive feelings.

Helping

One of the first effects of positive affect discov-ered has been an increased tendency, under

normal circumstances, to be helpful, generous,socially responsible, friendly, and kind (e.g.,Cunningham, 1979; Isen & Levin, 1972; seeIsen, 1987, for a review and discussion). Theevidence suggests that these effects are not theresult of simple compliance or giving in to oth-ers’ wishes or demands, but rather result froma person’s flexibility in thinking and his or herresultant decision in the situation. For exam-ple, these effects include helping others and do-nating to charity, all else being equal—but notwhen the helping task would cause discomfortto a third party (e.g., Isen & Levin, 1972), notwhen it would threaten the helper’s affectivestate (and the other’s need is not pressing; Isen& Simmonds, 1978), and not when the causeto be helped would benefit a disliked group(Forest et al., 1979). In those situations, peoplein positive affect were less likely than controlsto help. Moreover, these effects may themselvesdepend on such factors as the amount of harmthat might come to the person in need if the po-tential helper did not help, and so on (see Isen,1987, for a discussion).

The point here is that in general, positiveaffect promotes not selfishness and notcompliance—but a tendency to be kind and fairto both self and others, social connectedness, re-sponsibility, the ability to see situations from an-other person’s perspective as well as one’s own,and the ability to navigate among the conflictingneeds and goals to reach a satisfying solution tothe dilemma. Thus the evidence shows that peo-ple who are feeling happy are more motivated tobe helpful and socially responsible, and they arealso more flexible in their thinking with the re-sult that they can adopt a problem-solving ap-proach where there are conflicting needs of peo-ple with differing concerns, in a way that isappropriately responsive to the situation.

Maintenance of Positive Affect

There is evidence that a motive to maintain thepositive state is one of the motives engenderedby positive affect. For example, as noted in thediscussion of affect’s influence on helping, peo-ple in positive affect were less likely than con-trols to help a stranger when the helping taskwas portrayed as virtually certain to make theparticipants feel depressed, and when the otherperson’s need was not pressing (Isen & Sim-monds, 1978). The results of that study wereinterpreted as indicating that, at least undersome circumstances, positive affect engenders a


motive to avoid loss of the positive state. Sucha motive has also been theorized to relate topositive affect subjects’ relative risk aversion,which has been observed under certain circum-stances, and which is discussed further in thesection on decision making. In addition, workby Petty and colleagues is compatible with theidea that positive affect fosters an affect main-tenance motive, which then also plays a role inother choices and behavior (e.g., Wegener etal., 1995).

Besides these effects, this preference toavoid loss of the positive state may causepeople who are feeling happy to tend to leavemore negative topics for another time, or atleast to consider doing so, when that is possi-ble. Consequently, positive affect may influ-ence responses, performance, or latency of re-sponding on tasks involving negativematerial. Thus positive affect maintenancewill be a factor when all else is equal, orwhen the situation is such that attention toone’s own affect maintenance is the most ap-propriate focus. However, as noted earlier, re-search has shown that it will not be the only,or even the primary, concern of a person whois feeling happy.

To some extent, then, this motive has beenoveremphasized in the literature on positive af-fect, with some authors assuming that main-taining positive affect will be the only or theprimary motive engendered by such affect,and will govern behavior and style of cogni-tive processing—causing not only “defensive”blind spots, but also avoidance of careful,effortful thought (e.g., Erber & Erber, 2000;Loewenstein & Lerner, 2002; Schwarz, 2003;see Isen, 2003, for a discussion). As notedabove, the evidence does not actually supportsuch assumptions.

Thus the motive to maintain positive affectshould not be expected to be absolute in its ef-fects; it has not been found to result in blind, ir-rational bias, or in distortion of negative stim-uli or tasks. In fact, far from leading people toignore or distort needed negative information,positive affect has been shown to enable ef-fective coping and reduce “defensiveness” inproblematic situations that must be dealt with.Thus it appears that positive affect promoteseffective thinking about even negative material,if doing so is useful or necessary, even though itleads people to sidestep unnecessary consider-ation of unpleasant material when that is ap-propriate.

Expectancy Motivation

Although I have noted that positive affect doesnot simply raise general motivation or effort onall tasks, through something akin to general“arousal” or activation, recent work suggeststhat it may influence the cognitive processes re-lated to expectancy motivation, which isthought to underlie motivation in specific cir-cumstances. Thus positive affect may increasemotivation through cognitive processes in cer-tain situations.

The theory of expectancy motivation pro-poses that work motivation is a function of theconnection a person sees among his or her ef-fort, performance, and the outcomes he or shereceives, as well as the desirability of those out-comes (see Kanfer, 1990, for a fuller descrip-tion). Basically, if a person sees relatedness be-tween his or her effort and performance, andbetween his or her successful performance andthe ultimate obtaining of a desired reward, andif the reward is seen as desirable, the personwill be motivated to try; if one of those compo-nents is missing or low, the person will not bemotivated to work or try.

This theory is especially relevant to the roleof positive affect in motivation, because posi-tive affect is known to have the potential to in-fluence the perceived desirability of desirablerewards, as well as to influence the ability tosee connections or relationships among compo-nents of situations. Thus positive affect may in-crease motivation through raising the expec-tation that one’s effort will lead to goodperformance, and that good performance willresult in obtaining a desirable reward. How-ever, if all of the details proposed regardingpositive affect are correct, then this effect ofpositive affect should be true primarily withincertain ranges of valence, performance, and ef-fort, and it should hold only where outcome isnot chance-dependent. That is, for very high orvery low levels of effort, positive affect shouldnot increase one’s expectation that one will per-form well; and for very high or very low levelsof performance, positive affect should not in-crease the sense that one will get the reward.Only in the moderate range of effort, and per-formance, should positive affect have an influ-ence. Regarding the desirability of the reward,as noted in the context of affect’s influence onevaluation of other stimuli, positive affectshould increase the perceived desirability ofmoderately attractive outcomes, but not ex-


tremely attractive or extremely unattractiveoutcomes. Finally, if these effects result from arational decision by a person in positive affect,as held in this chapter, then positive affectshould not have an effect when the results aredescribed as dependent on chance.

This is exactly what was found in two stud-ies investigating these possibilities (Erez &Isen, 2002). Positive affect, induced in either oftwo ways, increased each of the components ofexpectancy motivation, but only in the moder-ate ranges, and only when the outcome was notdependent on chance. In a moderate range ofexpected performance (but not at extremelyhigh or extremely low expected performancelevels), people in positive affect saw more con-nection between how hard they tried and howwell they would do; this resulted in greater mo-tivation, more actual task persistence, and im-proved actual performance (see Erez & Isen,2002, for more details). This is not just a mat-ter of affect’s influencing motivation globally inthe sense of “trying harder.” It reflects how ef-fective a person expects to be (and is) if he orshe tries moderately hard in a situation whereeffort can have influence.

These findings provide strong additional evi-dence that people in positive affect do not loseor lack motivation, provided that it makessense to try in the situation. Furthermore, theyare not superficial in their processing or consid-eration of the task, and they pay careful atten-tion to the details and make fine distinctionsamong the characteristics of the situations. Thefindings also add support to the point raisedearlier that positive affect does not function asa lens or filter—that is, does not improve theestimates of all concepts or stimuli consideredwhile people are feeling happy.

Intrinsic Motivation and Variety Seeking

Intrinsic Motivation

In one series of studies, it was found that peo-ple in whom positive affect had been inducedshowed more intrinsic motivation than con-trols, but also responded well to extrinsic moti-vation when there was a work task that neededto be done (Isen & Reeve, 2005). Intrinsic mo-tivation was measured, as is standard in thatliterature (e.g., Deci & Ryan, 1985), in a free-choice situation by (1) choice of a more inter-esting task over a dull task (which, in this case,had a very small chance of paying a small

amount of money; and (2) reported enjoymentof the pleasant task while engaging in it. How-ever, when participants were informed thatthere was some work that needed to be done,and again had the choice of activity during thetime period, they chose to do the work thatneeded to be done (thus significantly reducingthe amount of time they spent on the enjoyabletask)—but they spent the rest of the time withthe more enjoyable task, and thus still spentmore time than controls on the liked task (eventhough they got the work done with few errors,just as the control group did). They again likedthe enjoyable task more than controls (but didnot differ from controls in evaluation of the te-dious work task). Thus they displayed self-control in the sense in which that term is nor-mally used: They did not like the work task asmuch as the puzzle, but voluntarily engaged init to complete it well. Furthermore, there wasno evidence of any kind of superficial process-ing or ignoring of details, and this was trueeven though they were deferring playing withthe puzzle in favor of the work that needed tobe done.

These results show that people in positive af-fect are more intrinsically motivated, but at thesame time show that positive affect will notcause people to avoid work tasks or more bor-ing or unpleasant or difficult tasks, if it is clearthat those tasks need to be done, or if there issome potential benefit in doing them. Peoplewho feel good prefer pleasant things, and enjoythem more when they do them; however, rela-tive to control subjects, they do not shirk, irra-tionally “defend against,” or irresponsibly re-fuse to engage in, less pleasant tasks. Moreover,they do not rate the tedious work task as morepleasant, showing again that the influence ofpositive affect is not like a lens or filter, color-ing all stimuli. It is also of interest, in view ofattribution theory or other type of “signaling”theory, that when the people in positive affectworked on the tedious task (which they didvoluntarily in both of these studies), they didnot inflate their evaluation of that task.

Variety Seeking

Another type of motive that appears to be in-duced by positive affect is variety seekingamong safe, enjoyable options (Kahn & Isen,1993). Three studies reported that when peoplein whom positive affect had been induced weregiven the opportunity to make several choices


in a food category such as snacks, they showedmore switching among alternatives than con-trols did and included a broader range of itemsin their choice sets, as long as unpleasant ornegative potential features of the items werenot salient. In contrast, when a negative butnot risky feature (such as the possibility that alow-salt product would taste less good than theregular) was salient, there was no difference be-tween the groups in variety seeking. Thus thereis evidence that positive affect promotes prefer-ence for variety and a wider range of optionsconsidered. This can be seen as safe stimulationseeking, but it is not risk seeking. The empiricaldifference in affect’s influence on variety seek-ing and risk seeking will become clear when wediscuss the results of studies on positive affect’sinfluence on risk taking.

The observed effects of positive affect on in-trinsic motivation and variety seeking relate toa developing area of decision making focusedon what has been called the “demotivating” ef-fects of “too much” variety on choosing and onsatisfaction with the choice (e.g., Iyengar &Lepper, 2000). Several studies show that hav-ing a large variety can result in people’s defer-ring the choice or feeling less satisfied withtheir ultimate choice. Space limitations pre-clude detailed consideration of this topic here,but of interest in the present context is thatpositive affect has been found to counter theeffects of this “choice overload” (Isen &Spassova, 2008), leading to more satisfactionwith both the choice process and the item se-lected from the very large variety.

Summary

Thus, in summary, positive affect appears toproduce behavior that may be seen as resultingnot only from the cognitive effects that havebeen discussed (such as increased flexibility,elaboration and access to positive material inmemory, increased integration of concepts andability to see connections among ideas, etc.),but also from apparent motivational changes.The work on expectancy motivation showsthat people in positive affect work diligently ontasks, if there is reason to believe that doing somakes sense; indeed, they may be even moremotivated than people in neutral affect, be-cause they are more likely to see a connectionbetween how hard they try and the outcomesthey achieve. As has been found in many otherexperiments reviewed in this chapter, there is

no evidence in these studies that people in posi-tive affect typically ignore or miss details. Infact, they persist longer and perform betterthan controls, as long as the goal is at leastmoderately desirable and achievable throughtheir effort.

Similarly, the work on intrinsic motivationand self-control indicates that people in posi-tive affect show more intrinsic motivation andenjoy pleasant tasks more than controls, butthat they voluntarily readily and accuratelyperform work tasks that need to be done, evenif doing so requires reducing time on a pleasur-able task. As illustrated, the desire to maintainthe pleasant affective state—one of the motivesengendered by positive affect that was identi-fied several years ago—does not necessarilytake precedence over other motives fostered byhappy feelings, other motives in the situation,or other effects of positive affect such as care-fulness and flexibility in thinking.

POSITIVE AFFECTAND DECISION MAKING

In the years since the second edition of thishandbook was published, interest in the role ofemotion has increased notably within the fieldof cognitive psychology and decision making,as it also has in some related fields such as eco-nomics and finance (e.g., Hirshleifer & Shum-way, 2003; Lerner, Small, & Loewenstein,2004; Slovic, Finucane, Peters, & MacGregor,2004, just to name a few of the many, many pa-pers appearing in the past several years fromthese and related domains). Some of these pa-pers still focus on negative affect and still retainthe assumption that affect, when it has influ-ence on thinking, disrupts systematic thinking.But some, like those mentioned above, arereaching beyond those positions.

A few representative examples of the variousapproaches that now characterize the treat-ment of affect in these various fields relatedto decision making include the following:Hirshleifer and Shumway (2003) looked at theeffect of sunshine, which is known to be relatedto positive affect and to produce some ofthe same effects as positive affect (e.g.,Cunningham, 1979), on stock market invest-ment decisions; they reported that when othervariables were controlled, sunshine was relatedto better stock returns. Djamasbi (2007) exam-ined the role of positive affect in utilization of a


decision support system, and found positive af-fect to increase the number of cues used and theaccuracy of the judgment made.

Lerner et al. (2004) looked at the effects oftwo different negative affects, disgust and sad-ness, and reported different effects on the salesprices and buying prices found acceptableunder those conditions in an “endowment ef-fect” paradigm taken from the behavioral eco-nomics literature. Those authors argued for theimportance of considering the effects of indi-vidual emotions, rather than just consideringthe general valence of the emotion. This sug-gestion comes out of the discrete-emotions tra-dition, but it happens to be also compatiblewith the view that the influence of affect de-pends on what the affect leads the person tothink about in the situation (e.g., Isen, 1987,2000)—which can be different not only for dis-crete emotions, but also for any given emotionin different situations, or when the person hasdifferent goals. Unlike negative affect, positiveaffect of many different kinds has been foundto have similar effects on such behaviors ashelping, problem solving, and so on (e.g., Isen,1987). However, the point here may still applyto positive affect, in that it would mean thatone might expect different effects of such affectin different situations or with differing goals ofthe person as well—which, of course, is whathas been found in many studies, as notedthroughout this chapter.

Slovic et al. (2004) discuss what has beencalled a “two-systems” approach: They pro-pose that decision making can take place bothby calculation and by intuitive reaction, andthat the two may often have something to con-tribute to one another. They do assume that theintuitive system will be the one that involves af-fect, and that the rational system does not in-volve affect (assumptions that may not be sup-ported), but they allow for the integration ofthe two by the means just described.

Other scholars in decision making have pro-posed models by which the affect is a part ofthe decision-making model itself, in that the ex-pected enjoyment or pleasure at the outcome isconsidered part of the calculus by which a ra-tional decision is reached (e.g., Mellers, 2000).This idea is compatible with other work thatalso views affect as part of a rational decision-making process, in the sense that affect’s influ-ence on the processes that are involved in thecalculation, or on thoughts about the situation,plays a role in the ways the components are

viewed and weighted (see Isen, 2003, for a dis-cussion). Hermalin and Isen (in press) have alsoproposed an economic model of some of the ef-fects of affect in which the influence of affect isrepresented as part of the model. These are justa few of the initiatives that have begun in thepast few years to integrate affect into decisionmaking from the decision-making or economicside, and it seems likely that these and otherswill be developed further in the near future.

It should be clear by now that consideringthe influence of affect on decision making in-volves recognizing the interacting roles of af-fect, the valence or interestingness of the task,the importance or utility of the task, the fram-ing of the situation, the person’s goals and mo-tives in the situation, and other aspects of thecontext. Furthermore, the processes describedare seen as depending on people’s interpreta-tions of the situation. This suggests that pro-cesses such as decision making and problemsolving may be hierarchically organized: Beforethe problem is actually addressed, some com-mand or executive decisions (attention deploy-ment decisions) or evaluations may be made re-garding how important the task is, what itsutility might be, how dangerous or safe the sit-uation is, or whether the person has any con-trol over its eventual outcome, as well as whatits hedonic consequences might be, and otherpossible interpretations of the situation (e.g.,Martin, Ward, Achee, & Wyer, 1993). Thesedecisions may influence the way in which theproblem is framed or addressed. Furthermore,the person may also reevaluate such decisionswhile solving the task, using incoming informa-tion as he or she proceeds.

That is, perhaps the person makes a series ofdecisions in deciding or solving a problem, andperhaps an early one relates to the domain ofthe task, with regard to both valence and im-portance. A helpful way of viewing this firstlevel of decision may be in terms of the framingof the problem. The person may derive a senseof whether this is a situation that he or she canenjoy (gain something, share, etc.) or a situa-tion in which he or she must be concerned notto be harmed (not to lose what he or she al-ready has or needs), and a sense of what his orher options are.

A conceptualization in terms of the framingof the situation has been useful in the decision-making and risk-taking literature, where differ-ences have been found in people’s decisions, de-pending on whether a problem was framed as a


potential gain or as a potential loss (e.g.,Tversky & Kahneman, 1981). The parallel tothis work is not exact, because, in the well-known Asian disease problem, for example, thesituation still involves danger and death,whether it is framed negatively (percentage oflives to be lost) or positively (percentage oflives to be saved). However, framing as studiedpreviously in this literature may nonethelessbear some relevance to the two types ofmotives and actions (exploration vs. self-protection) resulting from positive affect thatare under discussion here. Issues related to theframing of the experimental situation, and tothe possibility of a kind of hierarchical evalua-tion or decision process (especially as this inter-acts with affect), may be explored more fully inthe years to come.

Risk Preference

One kind of decision that has been studied as afunction of positive affect is risk preference. Inthese studies on positive affect and risk, peopleexperiencing positive affect were more risk-averse than controls, but only when the risk sit-uation about which they were reasoning was arealistic one, involving potential for a real, sub-stantial, meaningful loss.

For example, when betting chips represent-ing their credit for participating in the study,those in positive affect bet less (Isen & Patrick,1983), and required a higher probability ofwinning before agreeing to a substantial bet(but not a trivial bet), and also showed morethoughts about losing in a thought-listing taskfollowing this assessment (Isen & Geva, 1987).However, when asked about taking a chanceon a hypothetical risk, without having to risktheir own resources, people in positive affectresponded as if they would be more willing totake this chance (Isen & Patrick, 1983). Simi-larly, when people were asked, without an af-fect induction, to estimate what effect theythought positive affect would have on their riskpreference, they intuited that it would increasetheir willingness to take the risk. This suggeststhat it may not be wise to rely on people’s hy-pothetical estimates of preferences and choicesin assessing the influence of positive affect onrisk.

The relative risk aversion observed in posi-tive affect subjects considering real risks mayrelate to affect maintenance: People who arefeeling happy risk losing that state, as well as

any tangible stake, if they lose the gamble.Therefore, perhaps because they have more tolose if facing a consequential risk, they are risk-averse relative to controls. This interpretationis supported by results of a study that exam-ined the utility associated with various out-comes and found that people in positive affectdisplayed a greater negative utility for a sub-stantial loss than did controls (Isen et al.,1988). That is, the same potential loss seemedworse to people in positive affect—and thismay reflect a motive to maintain the positivestate.

Some authors have assumed that affectmaintenance, when it plays a role in a situationsuch as risk preference, influences decisionsand behavior by causing a person to ignore de-tails of the situation. For example, Loewensteinand Lerner (2003) assume that affect mainte-nance implies that people in positive affect willnot pay attention to details and will oversim-plify tasks. Furthermore, some researchers—also thinking that affect maintenance domi-nates all the effects of positive affect—may mis-understand the effects of positive affect morebroadly, assuming that in order to maintainpositive affect, a person in a positive state willalways engage in superficial cognitive process-ing. The context of this work on risk percep-tion and risk-related behavior, then, is a goodplace to point out once again that positive af-fect maintenance does not imply either thatpeople in positive affect will not pay attentionto details, or that they will oversimplify thetask or perform poorly on it. The details of therisk studies show instead that people in positiveaffect consider the options in detail and quitecarefully; for example, they differentiateamong gambles and even have significantlymore thoughts than controls do about a poten-tial substantial loss. Desire for affect mainte-nance, then, when it is a factor in determiningbehavior in these risk situations, is theorized toplay a role—not through careless informationprocessing—but through avoiding the possibil-ity of loss by reducing betting or risky behavior.

The results described above indicate thatpositive affect increases the negative utility of areal, meaningful potential loss. At the sametime, positive affect has been found to increasethe subjective probability of success in situa-tions of risk assessment (e.g., Johnson &Tversky, 1983; Nygren, Isen, Taylor, & Dulin,1996). Thus it seems that the two componentsof risk assessment, probability and utility, are


influenced in functionally opposite ways bypositive affect: Although the subjective proba-bility of winning is increased, the negative util-ity or perceived danger of the potential loss isalso increased. And the resultant behavior—relative risk avoidance by people in positiveaffect—suggests that the utility information ismore influential in such people’s decisions thanthe probability information (Nygren et al.,1996). These findings regarding positive affectand risk illustrate, as noted in other contexts,that the influence of positive affect on risk per-ception is not simple; it depends on, and inter-acts with, task and setting conditions.

Complex Decision Making

Another type of decision making that has beenstudied as a function of positive affect is whatmight be called “complex decision making,” inwhich people are asked to choose the best itemfrom among several alternatives varying on anumber of dimensions, or to solve a complexproblem (e.g., making a medical diagnosis).Research has shown that positive affect, in-duced by either an unrelated success or a smallbag of candy, leads to decision making that isboth more efficient and more thorough, if thetask lends itself to increased effort or care.

For example, in one study (Isen et al., 1991),the participants were advanced medical stu-dents whose task was to choose, from amongsix descriptions of patients varying with regardto each of nine health-relevant factors, the pa-tient most likely to have lung cancer. Resultsshowed that people in positive affect solved(correctly) the assigned task more efficientlythan controls, but then went on to do morethan was required, such as suggesting treat-ments or offering diagnoses for the other pa-tients. Thus they were both more efficient andmore thorough than controls. In addition, thepeople in positive affect showed significantlyless confusion and a significantly greater ten-dency to integrate the material with which theywere working.

The results of this study are compatible withthose of other research on affect and decisionmaking, such as research on hypothetical carchoice; however, particular measures (such astotal amount of time and amount of redun-dancy in looking at the materials) produceddifferent results, because of the contextual dif-ferences that made continuing to work with thematerials after the assigned task was completed

sensible and appropriate (see Isen, 2000, for adiscussion).

Another study, examining physicians’ diag-nostic processes, reported that doctors inwhom positive affect had been induced (thistime by receipt of a small bag of candy) identi-fied the domain of the medical problem signifi-cantly earlier in their protocols and were moreopen to information, being significantly lesslikely than physicians in a control group to dis-tort or ignore information that did not fit witha diagnostic hypothesis they were considering(Estrada at al., 1997). Thus this study confirmsthat positive affect does facilitate integration ofinformation for decision making, and also fos-ters openness to information. At the same time,this study found no evidence that positive af-fect promotes premature closure, superficialthinking, jumping to conclusion without suffi-cient evidence, or any other indication of anyimpairment or carelessness in thinking. Therewas also no evidence of increased reliance onestablished schemas in preference to incomingnew information; in fact, just the opposite wasfound.

The findings of the studies on medical deci-sion making suggest that positive affect maypromote not only more efficiency but alsomore thoroughness, as well as openness to in-formation and the tendency to check hypothe-ses and tentative answers against additional in-formation. It is likely that such an effect may beobserved only where the materials allow forthat possibility, and only where there is somereason to engage the task. It is clear, however,that such tasks include ones requiring complexconsideration of serious topics of interest to thesubjects; they are not limited to stereotypically“positive” or fun topics, and not to so-called“insight” or creativity problems. Moreover, thefact that people in positive affect were open toinformation disconfirming their initial hypoth-eses, and checked their work carefully, speaksagainst the hypothesis that positive affect leadsto a careless, superficial, or “playful” ap-proach, or to reliance on a person’s establishedmental schemas rather than incoming new in-formation.

Positive Affect and Use of Heuristics

Does Positive Affect Lead to Careless Thinking?

Several studies have addressed the issue ofwhether positive affect, rather than facilitating


careful thinking and performance, impairscareful thinking because it takes up cognitivecapacity (e.g., Mackie & Worth, 1989, 1991)and/or undermines the motivation to thinkcarefully (e.g., Bless et al., 1990; Schwarz,2002; Schwarz & Bless, 1991). The theoreticalpositions underlying the latter work hold thatpositive affect, by its nature, provides a “sig-nal” that leads to careless, superficial thinkingand use of heuristics, as contrasted with sys-tematic cognitive processing (and thus often re-sults in impaired processing). As we have seenin the preceding discussion of the decision-making studies, however, there are also find-ings that are not compatible with such formu-lations. How can the differences be under-stood? Perhaps specific methods are playing arole.

One approach to demonstrating the viewthat positive affect typically promotes use ofheuristics uses an attitude change paradigmand infers nonsystematic processing from thepatterns of arguments that are successful inbringing about attitude change. For example,attitude change in response to relatively weak(though not irrational) arguments is taken asindication of nonsystematic processing; there-fore, if people in positive affect show as muchattitude change when weaker arguments arepresented as they do when strong argumentsare presented, the conclusion is that positive af-fect interferes with systematic processing (e.g.,Mackie & Worth, 1989).

As suggested in the earlier editions of thischapter, one problem with this inference, how-ever, is that attitude change could be reportedfor reasons unrelated to processing of the mes-sage: People in positive affect may want to bemore agreeable (which is known from the so-cial psychology literature to be associated withpositive affect, all else being equal). Or theymay think of additional good arguments oftheir own, and this may lead them to displaymore change in attitude, independent of thestrength of the arguments presented in the task.

The need for alternative interpretations suchas these is also suggested by the fact that partic-ipants experiencing positive affect did not dif-fer from other participants in their recognitionthat the “weak” arguments were weaker thanthe “strong” arguments, or in their ability torecall the message content, when that was mea-sured directly (Bless et al., 1990; Mackie &Worth, 1989). In addition, in some instancesthe materials used in the studies focused on

negative, upsetting topics, and the situation didnot provide much justification for working onthe task. As noted earlier, it is becoming appar-ent that without a reason to work on an un-pleasant task—that is, if the task is both un-pleasant and unimportant—people in positiveaffect may not engage the task. Thus it is farfrom clear, from such studies, that lack of ei-ther cognitive capacity or motivation to “thinkstraight” is what causes any difference betweengroups in reported attitude change.

In another approach that has been taken,researchers reasoned that use of stereotypescould be employed as an indication of heuristicor nonsystematic processing, and on this basispositive affect was investigated as a determi-nant of stereotype use (e.g., Bodenhausen et al.,1994, to mention one of several). The paper byBodenhausen et al. (1994) illustrates the com-plexity of the findings in that area: Althoughthree of four studies in that paper indicatedthat positive affect did result in increased use ofa stereotype, the fourth study showed that thisdifference was not found—and the results evenexhibited a marginally significant tendency to-ward the reverse, if participants were given areason to pay more attention to the task (e.g., ifthey were to be accountable for their decision).The fact that the effect disappeared when theimportance of the task was increased under-mines the suggestion that, as a rule, positiveaffect leads to use of stereotypes and to non-systematic processing generally; moreover, thefact that the people in positive affect respondedappropriately to the different condition inStudy 4 indicates their flexibility, rather thaninability or unwillingness to process more care-fully and individually. As mentioned earlier,these points were noted by Bodenhausen et al.(2001) in their more recent summary of the in-fluence of positive affect on stereotyping, whenthey pointed out that happy people are flexiblein their information-processing strategies,rather than either unable or unwilling to en-gage in systematic thinking.

Recent evidence reported by Johnson andFredrickson (2005) runs counter to the inter-pretation that any increased use of stereotypesthat is observed occurs because positive affectimpairs the ability or motivation to process de-tails, or promotes use of established schemas.These authors found that positive affect, rela-tive to neutral affect and to fear, significantlyimproved white people’s ability to recognize in-dividual black persons’ faces, and reduced the


“own-race-bias effect” (the relative inability todistinguish among faces of individuals of a raceother than one’s own), which is usually takento be part of the tendency to stereotype. Theseresults, then, add to the evidence that positiveaffect does not necessarily promote use of ste-reotypes. More importantly in the present con-text, they suggest that people in positive affectcan and do, pay close attention and process de-tail rather than relying on their establishedschemas.

Finally with regard to stereotype use, as isevident from the examples here, any evidencethat studies of this type might provide for thesuggestion that positive affect interferes withsystematic processing is not compelling, be-cause it would be indirect and open to alterna-tive interpretation. This is similar to the pointmade with regard to using the attitude changeparadigm to try to show that positive affect im-pairs systematic thinking: Stereotype use couldoccur for reasons other than nonsystematicthinking, impaired processing, or desire fornoneffortful thinking.

In another approach to the question ofwhether positive affect impairs systematic pro-cessing, one paper has reported impaired rea-soning performance among people in whompositive affect was induced (e.g., Melton,1995). As noted above, in the context of the at-titude change studies, it is possible that findingsof seemingly impaired processing may resultfrom the materials and context of the particu-lar experiments that show the findings. Thismay be the case in the paper by Melton (1995)as well. A recent preliminary study has foundthat materials like those used in that paper areactually tedious and annoying to experimentalparticipants, and it is known that use of thiskind of material can lead—especially amongparticipants in positive affect—to less engage-ment with the task unless there is a reason forthe participants to engage the task carefully(e.g., Isen & Reeve, 2005). Preliminary resultsindicate that equally difficult material of thesame kind (logical reasoning), but more inter-estingly formatted, actually reveals improvedrather than impaired performance among par-ticipants in positive affect (Isen, Erez, Nester,& Shmidt, 2007). In addition, it has recentlybeen proposed that positive affect may give riseto people’s being able to notice that infor-mation needed for drawing a conclusion ormaking a decision is actually missing (Mantelet al., 2007).

Thus the evidence that positive affect per sepromotes superficial processing per se is not asclear as is often assumed. In cases where it ap-pears to do so, this may result from a lack ofmotivation for, or interest in, the task pre-sented; however, this kind of motivational ef-fect is different from one that postulates gen-eral interference with motivation overall, orwith motivation to process carefully or to payattention to detail. On the contrary, as shownin the section on motivation, the evidence sug-gests that positive affect can facilitate motiva-tion by increasing the perceived link betweeneffort and outcome, and there is much evidencethat people in positive affect regularly pay at-tention to detail.

Use of Established Cognitive Structures

In addition, Schwarz, Clore, Bless, and theircolleagues—long-time proponents of the viewthat positive affect interferes with systematicprocessing because it undermines the motiva-tion to think carefully—have reported resultsincompatible with that view, and have pro-moted an alternative but similar view (Bless etal., 1996). In the 1996 paper, they acknowl-edge that “the evidence that heuristic process-ing is due to the hypothesized motivational orcapacity deficits is less conclusive than is oftenassumed” (p. 665). They report results of stud-ies showing that although people in positive af-fect demonstrated a greater tendency, relativeto controls, to use a script to organize theirlearning and memory of a story, those partici-pants also performed better than controls on asecondary task that had to be performed simul-taneously. In the words of those authors, thisfinding “is incompatible with the assumptionthat happy moods decrease either cognitive ca-pacity or processing motivation in general,which would predict impaired secondary-taskperformance” (Bless et al., 1996, p. 665).

As an outgrowth of that thinking, those re-searchers have proposed that positive affectleads to reliance on established cognitive struc-tures (such as schemas or scripts) and habitualroutines, but they still propose that positive af-fect leads to heuristic, superficial, nonsystematiccognitive processing (e.g., Bless et al., 1996;Gasper & Clore, 2002; Schwarz, 2002; Storbeck& Clore, 2005). In some places, there is fleetingacknowledgment that use of establishedschemas and routines says nothing about sys-tematic processing, that it may even free up ca-


pacity, and/or that positive affect may enableflexibility and enhanced overall performance.Nevertheless, their primary argument is still thatpositive affect is expected to lead to overrelianceon global, general knowledge structures, and todiscourage careful processing or attention to de-tail, often to the detriment of performance. Thusthis position still promotes the view that positiveaffect will result in superficial, careless cognitiveprocessing, with lack of attention to detail (andthat negative affect will promote careful, de-tailed processing)—even though their own dataoften do not confirm those predictions (e.g.,Gasper & Clore, 2002, pp. 37, 39; Storbeck &Clore, 2005).

They maintain this stance even though it alsoignores the fact that, as pointed out in thischapter, a great many data show that positiveaffect is compatible with careful, detailedthinking under most conditions. In addition,the position that positive affect will lead to reli-ance on one’s “usual routines” that come tomind most easily (e.g., Schwarz, 2002) is atvariance with the wide range of findings in theliterature for quite some time, showing thatpositive affect increases the innovativeness ofpeople’s responses and thinking, their opennessand flexibility in thinking, and their preferencefor variety (as well as their careful attention todetail and their improved performance in awide range of tasks and problems). These find-ings regarding creative problem solving and va-riety seeking do not fit with the notion thatpositive affect promotes only (or even primar-ily) use of existing general knowledge struc-tures and old, “routine,” or habitual ways oflooking at things.

Flexibility as an Alternative Interpretation

Those authors contrast the process of usinggeneral knowledge structures with that of rely-ing on data and engaging in learning, as if thesetwo processes will not occur together. In con-trast, the flexibility hypothesis suggests thatpeople in positive affect should be able to uti-lize both kinds of information—informationthat is new in the situation, and informationfrom established schemas, past experiences,and past learning—and that they should bebetter at switching among these varioussources and integrating them and the informa-tion available from them. These approaches(flexibility vs. schema-based) differ in manypredictions that would follow from them. It is

also worth noting that the kind of integrationproposed by the flexibility hypothesis—integration of new and established materialthrough processes of assimilation (of new ma-terial into existing schemas) and accommoda-tion (change of existing schemas in response tothe new material)—is at the heart of the pro-cess of learning, according to several theoriesof education (e.g., see Isen, Daubman, &Gorgoglione, 1987, for consideration of thepotential role of positive affect in this and re-lated processes of learning).

The findings reviewed in the present chapteralso support the suggestion made in the earliereditions of this handbook chapter that heuristicand systematic cognitive processes may notnecessarily be alternatives to one another (as isoften assumed), but may occur together. Per-haps, also, chunking, integration, or someother method of simplifying or organizing of acomplex set of data actually frees up capacityor resources for use on other tasks. Thus pro-cessing in positive affect that appears “simpli-fied” may actually result from more elabo-rated, differentiated thinking and from betterunderstanding of the issues (see, e.g., Isen,1993, for a fuller discussion).

Summary

The evidence that as a rule mild positive affectdisrupts systematic thinking or leads to superfi-cial processing is actually not compelling. Thisis because the studies attempting to show thisare difficult to interpret—either because severallayers of inference are required in order toreach that conclusion (as in the attitude-changeand stereotyping studies); because they did notinclude a control group (positive affect andnegative affect were contrasted); or becausethey used materials or tasks allowing for alter-native interpretation, such as ones involvingunimportant/tedious tasks, negative topics, ortopics over which participants had no control.

In contrast, for more than a decade now,studies have been confirming that positive af-fect enables improved performance, even whensimplifying devices are also used. In addition,people in positive affect seem to be able to usesimplifying devices and systematic processingtogether, rendering their processing both moreefficient and more thorough. These findingssuggest that the consequence of positive affectis flexibility in modes of thinking and decisionmaking—attention to new data and detail,


integrated with existing knowledge structures,rather than exclusive reliance on existingknowledge structures and inability or reluc-tance to attend to detail.

A NEUROPSYCHOLOGICAL THEORYOF POSITIVE AFFECT’S INFLUENCEON COGNITION

As noted at the outset of this chapter, researchon the neuropsychology of emotion is growingat a rapid pace. Because of space limitations, Ican only touch the surface of this topic, brieflymentioning one theory that may help to under-stand how positive affect has its effects on flex-ibility, thinking, and behavior—the dopaminehypothesis (Ashby et al., 1999). This theory fo-cuses on the role of the neurotransmitter dopa-mine and proposes that many of the observedeffects of positive affect on cognition may re-sult from increased levels of dopamine in cer-tain brain regions. Noting that dopamine isknown to be associated with reward, this the-ory proposes that it is also present at increasedlevels during positive affect situations. Al-though other neurotransmitters no doubt playa role as well (see, e.g., Katz, 1999), and mayeven act in concert with dopamine to determinespecific nuances of behavior, the evidence isstrong that brain levels of dopamine play animportant role in mediating many of the cogni-tive effects of positive affect that have been ob-served.

For example, because dopamine release intoportions of the anterior cingulate region of thebrain is thought to be involved in flexible selec-tion of cognitive perspective and switchingamong alternative perspectives, it is likely thatrelease of dopamine in such frontal brain areasmediates the increased flexibility observedunder conditions of positive affect. This, then,would implicate dopamine in many of the ef-fects of positive feelings that have been ob-served: creative problem solving, openness toinformation, exploration, integration of ideas,ability to focus on important negative informa-tion when that is needed, and ability to keepothers’ perspectives as well as one’s own inmind (which may also play a role in many ofthe social effects of positive affect that havebeen established over the years, such as cooper-ativeness, social responsibility, improved nego-tiation skills, and generosity to both self andothers), just to name a few.

In addition, because there are many excit-atory dopamine receptors in frontal brain areasand prefrontal cortex, this theory also predictsthat positive affect will facilitate processes in-fluenced by those brain regions, such as atten-tion deployment, working memory, and mem-ory consolidation. Evidence is accumulatingthat supports this theory (e.g., Isen & Shmidt,2007; Kazen & Kuhl, 2005; Kuhl & Kazen,1999), although it is too soon to draw firmconclusions. Much remains to be explored re-lating to the dopamine hypothesis, but its pre-dictions provide more specific targets than wehave had in the past, and ideally these will addto our understanding.

This neurological theory adds to our toolsfor studying the impact of affect; however, itcannot replace studies conducted on the cogni-tive and behavioral levels. Rather than viewingthese levels of analysis (e.g., neurological vs.behavioral) as opposing ways to advance un-derstanding, it is possible instead to attempt tobring them together and integrate work fromthese multiple levels of analysis. Insights fromthe neurological level can help to inform andguide our search for effects and determinantsof feelings on the behavioral and cognitive lev-els as well. Similarly, research on the behavioraland cognitive levels can point to neurologicalprocesses that may be critical. Indeed, studieson the behavioral and cognitive levels of analy-sis led to our understanding that positive affectpromotes flexibility, problem solving, innova-tion, and improved attention deployment, andthus to the hypothesis that dopamine mightmediate some of the cognitive and behavioraleffects of positive affect.

CONCLUSION

In summary of the work presented in this chap-ter, it seems appropriate to emphasize onceagain that the influence of affect depends onwhat it makes the person think about, and thatthis is determined not by the affective statealone, but by this state in conjunction with sev-eral aspects of the situation that together influ-ence the person’s motives or goals, judgments,expectations, and choices. Brain regions acti-vated by neurotransmitters (particularly dopa-mine) associated with positive affect play a rolein the effects of feelings, but this is only one ofseveral influences that together determine cog-nitive processes and behavior.


Despite the complex interaction of factorsthat determines affect’s ultimate impact onthought processes, decision making, andbehavior, a few general conclusions can be of-fered. All else being equal, mild positive affecttends to promote exploration and enjoyment ofnew ideas and possibilities, and new ways oflooking at things, especially in safe, enjoyablesituations. Thus people in positive affect maybe alert to possibilities and may solve problemsboth more efficiently and more thoroughlythan controls. However, in situations of poten-tial harm, people who are feeling good respondappropriately cautiously and are not likely totake chances. Nonetheless, where there is a rea-son to think about possible losses or difficul-ties, they will elaborate and consider the nega-tive possibilities thoroughly and effectively.Thus positive affect facilitates coping and focuson long-term welfare as well.

The majority of existing data, then, suggestthat mild positive affect enables cognitive flexi-bility and thus fosters improved considerationof situations and problems, and improved per-formance on a wide range of tasks, as long asthere is reason for the person to engage thetask. People who are feeling mildly happy arebetter able to think about multiple aspects ofsituations and to see situations and stimuli inmultiple ways, seemingly simultaneously.

Compatibly with both the neurological in-terpretation of the underpinnings of these find-ings, and the cognitive interpretation (in termsof range of concepts activated), there is no rea-son to expect the influence of positive affect tobe linear, or to think that more intense positiveaffect will generate more (or even as much) orbetter problem solving. Everything that hasbeen said in this chapter applies to mild posi-tive affect, and there is no reason to think thatit would also be true of intense positive affect.That is, once dopamine is released or extensivecognitive context is cued, the effects are en-abled, and more intense affect would not neces-sarily cause more of those effects. In fact, in-tense positive feelings or very good news mayeven interfere with performance rather than fa-cilitating it. But the reason may be that thenews changes what people want to thinkabout—not that positive affect itself drainstheir capacity, distracts them, or signals themnot to process carefully. In other words, in thisregard, the positive event would not be differ-ent from any other interesting topic that causespeople to refocus their attention.

Regarding positive affect’s influence on sys-tematic thinking, the popular idea that positiveaffect undermines the ability or motivation tothink carefully, pay attention to detail, or takein new information fails to account for a verylarge and growing literature examining the per-formance of people in positive affect. Many re-searchers have now concluded that that idea isnot correct. Similarly, the suggestion that posi-tive affect leads people to use their establishedcognitive structures rather than taking in newinformation—which, again, is proposed to becaused by and/or to result in lack of attentionto detail—is not itself well supported by dataand also fails to account for many of the otherexisting data. Therefore, the “signaling” andspeculative “cognitive tuning” metaphor de-scribed by Schmarz (2002) as applying to posi-tive and negative emotion seem overused andmisleading in the context of positive affect.

This chapter has examined some of the evi-dence regarding ways in which positive affectinfluences decision making, including somecognitive and motivational processes involved.However, more work is needed to explore theserelationships, the circumstances under whichthey occur, and the processes that are involvedin producing them. Simple answers seem not tobe supported; for example, positive affect hasbeen found to give rise to elaboration and awide range of cognitive associations in re-sponse to neutral stimulus material, while notreducing association or thought about othermaterial, and not reducing the ability to stayfocused on a task a person is trying to perform.Feeling happy does not function as a lens or fil-ter, improving evaluation of everything in theperson’s line of vision or thought, but insteadincreases liking only for likable things. It in-creases preference for variety and acceptance ofa broader range of options into people’s choicesets, when the choice is among safe, enjoyablealternatives, but it does not promote risk tak-ing in situations of genuine risk. Happy feelingcan lead to efficient and thorough decisionmaking; it stimulates enjoyment of enjoyabletasks and perception of interesting tasks aseven more enriched (but not at the cost ofworking on less interesting things if they needto be done). When less interesting, or annoy-ing, tasks are presented, those things may bedeferred if deferring them is an option, but willbe addressed effectively if it is not. Positive feel-ings seem to promote activities that foster en-joyment and maintenance of those feelings, but


in rational, responsible, adaptive ways—not asif affect maintenance were of primary or singu-lar importance. Socially, of course, positive af-fect is known to promote the very importantprocesses of generosity, helpfulness, responsi-bility, and taking the perspective of anotherperson as well as one’s own, under most cir-cumstances, but it does not lead to mere com-pliance. Thus happy feelings provide manybenefits, but the processes involved are com-plex. Given the importance of positive affect inour lives, and the great advantages to socialbehavior and problem solving that result fromfeeling happy, positive affect seems an impor-tant topic for continued investigation.

REFERENCES

Amabile, T. M., Barsade, S. G., Mueller, J. S., & Staw, B.M. (2005). Affect and creativity at work. Adminis-trative Science Quarterly, 50, 367–403.

Ashby, F. G., Isen, A. M., & Turken, A. U. (1999). Aneuropsychological theory of positive affect and itsinfluence on cognition. Psychological Review, 106,529–550.

Aspinwall, L. G. (1998). Rethinking the role of positiveaffect and self-regulation. Motivation and Emotion,22(1), 1–32.

Aspinwall, L. G., & MacNamara, A. (2005). Takingpositive changes seriously: Toward a positive psy-chology of cancer survivorship and resilience. Can-cer, 104(11, Suppl.), 2549–2556.

Aspinwall, L. G., & Taylor, S. E. (1997). A stitch intime: Self-regulation and proactive coping. Psycho-logical Bulletin, 121, 417–436.

Barone, M. J., Miniard, P. W., & Romeo, J. B. (2000).The influence of positive mood on brand extensionevaluations. Journal of Consumer Research, 26,386–400.

Baumann, N., & Kuhl, J. (2005). Positive affect andflexibility: Overcoming the precedence of global overlocal processing of visual information. Motivationand Emotion, 29, 123–134.

Bless, H., Bohner, G., Schwarz, N., & Strack, F. (1990).Mood and persuasion: A cognitive response analysis.Personality and Social Psychology Bulletin, 16, 331–345.

Bless, H., Clore, G. L., Schwarz, N., Golisano, V., Rabe,C., & Wolk, M. (1996). Mood and the use of scripts:Does a happy mood really lead to mindlessness?Journal of Personality and Social Psychology, 71,665–679.

Bodenhausen, G. V., Kramer, G. P., & Susser, K. (1994).Happiness and stereotypic thinking in social judg-ment. Journal of Personality and Social Psychology,66, 621–632.

Bodenhausen, G. V., Mussweiler, T., Gabriel, S., &

Moreno, K. N. (2001). Affective influences on stereo-typing and intergroup relations. In J. P. Forgas (Ed.),Handbook of affect and social cognition (pp. 319–343). Mahwah, NJ: Erlbaum.

Carnevale, P. J. D., & Isen, A. M. (1986). The influenceof positive affect and visual access on the discovery ofintegrative solutions in bilateral negotiation. Organi-zational Behavior and Human Decision Processes,37, 1–13.

Cunningham, M. R. (1979). Weather, mood, and help-ing behavior: Quasi-experiments in the sunshine Sa-maritan. Journal of Personality and Social Psychol-ogy, 37, 1947–1956.

Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivationand self-determination in human behavior. NewYork: Plenum Press.

Derryberry, D. (1993). Attentional consequences ofoutcome-related motivational states: Congruent, in-congruent, and focusing effects. Motivation andEmotion, 17, 65–89.

Djamasbi, J. (2007). Does positive affect influence theeffective usage of a decision support system? Deci-sion Support Systems, 43, 1707–1717.

Dreisbach, G., & Goschke, T. (2004). How positive af-fect modulates cognitive control: Reducedperseveration at the cost of increased distractibility.Journal of Experimental Psychology: Learning,Memory, and Cognition, 30, 343–353.

Duncker, K. (1945). On problem-solving. PsychologicalMonographs, 58(Whole No. 5).

Easterbrook, J. A. (1959). The effect of emotion on cueutilization and the organization of behavior. Psycho-logical Review, 66, 183–201.

Erber, R., & Erber, M. W. (2000). The self-regulation ofmoods: Second thoughts on the importance of happi-ness in everyday life. Psychological Inquiry, 11, 142–148.

Erez, A., & Isen, A. M. (2002). The influence of positiveaffect on components of expectancy motivation.Journal of Applied Psychology, 89, 1055–1067.

Estrada, C. A., Isen, A. M., & Young, M. J. (1997). Pos-itive affect facilitates integration of information anddecreases anchoring in reasoning among physicians.Organizational Behavior and Human Decision Pro-cesses, 72, 117–135.

Forest, D., Clark, M., Mills, J., & Isen, A. M. (1979).Helping as a function of feeling state and nature ofthe helping behavior. Motivation and Emotion, 3,161–169.

Forgas, J. P. (2002). Feeling and doing: Affective influ-ences on interpersonal behavior. Psychological In-quiry, 13, 1–29.

Fredrickson, B. L. (2001). The role of positive emotionsin positive psychology: The broaden-and-build the-ory of positive emotions. American Psychologist, 56,218–226.

Fredrickson, B. L., & Branigan, C. (2005). Positiveemotions broaden the scope of attention andthought–action repertoires. Cognition and Emotion,19, 313–332.


Fredrickson, B. L., & Joiner, T. (2002). Positive emo-tions trigger upward spirals toward emotional well-being. Psychological Science, 13, 172–175.

Friedman, R. S., & Forster, J. (2005). The influence ofapproach and avoidance cues on attentional flexibil-ity. Motivation and Emotion, 29, 69–81.

Gasper, K., & Clore, G. L. (2002). Attending to the bigpicture: Mood and global versus local processing ofvisual information. Psychological Science, 13, 34–40.

Gray, J. R. (2001). Emotion modulation of cognitivecontrol: Approach–withdrawal states double-dissociate spatial from verbal two-back task perfor-mance. Journal of Experimental Psychology: Gen-eral, 130, 436–452.

Greene, T. R., & Noice, H. (1988). Influence of positiveaffect upon creative thinking and problem solving inchildren. Psychological Reports, 63, 895–898.

Hermalin, B. E., & Isen, A. M. (in press). A model of theeffect of affect on economic decision making. Quan-titative Marketing and Economics.

Hirshleifer, D., & Shumway, T. (2003). Good day sun-shine: Stock returns and the weather. Journal of Fi-nance, 58, 1009–1032.

Isen, A. M. (1987). Positive affect, cognitive processes,and social behavior. In L. Berkowitz (Ed.), Advancesin experimental social psychology (Vol. 20, pp. 203–253). New York: Academic Press.

Isen, A. M. (1993). Positive affect and decision mak-ing. In M. Lewis & J. M. Haviland (Eds.), Hand-book of emotion (pp. 261–277). New York:Guilford Press.

Isen, A. M. (2000). Positive affect and decision making.In M. Lewis & J. M. Haviland-Jones (Eds.), Hand-book of emotions (2nd ed., pp. 417–435). New York:Guilford Press.

Isen, A. M. (2003). Positive affect as a source of humanstrength. In L. G. Aspinwall & U. Staudinger (Eds.),A psychology of human strengths (pp. 179–195).Washington, DC: American Psychological Associa-tion.

Isen, A. M. (2008). Positive affect and decision pro-cesses: Some recent theoretical developments withpractical implications. In C. Haugdvedt, P. M. Herr,& F. R. Kardes (Eds.), Handbook of consumer psy-chology (pp. 273–296). New York: Psychology Press.

Isen, A. M., & Daubman, K. A. (1984). The influence ofaffect on categorization. Journal of Personality andSocial Psychology, 47, 1206–1217.

Isen, A. M., Daubman, K. A., & Gorgoglione, J. M.(1987). The influence of positive affect on cognitiveorganization: Implications for education. In R. E.Snow & J. M. Farr (Eds.), Aptitude, learning, andinstruction (Vol. 3, pp. 143–164). Hillsdale, NJ:Erlbaum.


Isen, A. M., Erez, A., Nester, M. A., & Schmidt, E.(2007). Positive affect facilitates recognition of insuf-

ficient information in logical reasoning. Unpublishedmanuscript, Cornell University.

Isen, A. M., & Geva, N. (1987). The influence of posi-tive affect on acceptable level of risk: The personwith a large canoe has a large worry. OrganizationalBehavior and Human Decision Processes, 39, 145–154.

Isen, A. M., Johnson, M. M. S., Mertz, E., & Robinson,G. F. (1985). The influence of positive affect on theunusualness of word associations. Journal of Person-ality and Social Psychology, 48, 1413–1426.

Isen, A. M., & Levin, P. F. (1972). Effect of feeling goodon helping: Cookies and kindness. Journal of Person-ality and Social Psychology, 21, 384–388.

Isen, A. M., Niedenthal, P., & Cantor, N. (1992). An in-fluence of positive affect on social categorization.Motivation and Emotion, 16, 65–78.

Isen, A. M., Nygren, T. E., & Ashby, F. G. (1988). Theinfluence of positive affect on the subjective utility ofgains and losses: It is just not worth the risk. Journalof Personality and Social Psychology, 55, 710–717.

Isen, A. M., & Patrick, R. (1983). The effect of positivefeelings on risk-taking: When the chips are down.Organizational Behavior and Human Performance,31, 194–202.

Isen, A. M., & Reeve, J. (2005). The influence of posi-tive affect on intrinsic and extrinsic motivation: Fa-cilitating enjoyment of play, responsible work behav-ior, and self-control. Motivation and Emotion, 29,297–325.

Isen, A. M., Rosenzweig, A. S., & Young, M. J. (1991).The influence of positive affect on clinical problemsolving. Medical Decision Making, 11, 221–227.

Isen, A. M., Shalker, T., Clark, M. S., & Karp, L.(1978). Affect, accessibility of material and behavior:A cognitive loop? Journal of Personality and SocialPsychology, 36, 1–12.

Isen, A. M., & Shmidt, E. (2007, January). Positive af-fect: Broadened focus without increased distractibili-ty. Paper presented at the Emotions preconference tothe annual meeting of the Society for Personality andSocial Psychology, Memphis, TN.

Isen, A. M., & Simmonds, S. F. (1978). The effect offeeling good on a helping task that is incompatiblewith good mood. Social Psychology Quarterly, 41,345–349.

Isen, A. M., & Spassova, G. (2008, February). Positiveaffect reduces the demotivating effects of choice over-load. Paper presented at the annual meeting of theSociety for Personality and Social Psychology, Albu-querque, NM.

Iyengar, S. S., & Lepper, M. R. (2000). When choice isdemotivating: Can one desire too much of a goodthing? Journal of Personality and Social Psychology,79, 995–1006.

Johnson, E., & Tversky, A. (1983). Affect, generaliza-tion and the perception of risk. Journal of Personalityand Social Psychology, 45, 20–31.

Johnson, K. J., & Fredrickson, B. L. (2005). “We alllook the same to me”: Positive emotions eliminate


the own-race bias in face recognition. PsychologicalScience, 16, 875–881.

Kahn, B. E., & Isen, A. M. (1993). The influence of pos-itive affect on variety-seeking among safe, enjoyableproducts. Journal of Consumer Research, 20, 257–270.

Kanfer, R. (1990). Motivation theory and industrial andorganizational psychology. In M. D. Dunnette & L.M. Hough (Eds.), Handbook of industrial and orga-nizational psychology (Vol. 1, pp. 75–170). PaloAlto, CA: Consulting Psychologists Press.

Katz, L. D. (1999). Dopamine and serotonin: Inte-grating current affective engagement with longer-term goals. Behavioral and Brain Sciences, 22, 527.

Kazen, M., & Kuhl, J. (2005). Intention memory andachievement motivation: Volitional facilitation andinhibition as a function of affective contents of need-related stimuli. Journal of Personality and Social Psy-chology, 89, 426–448.

Kimchi, R., & Palmer, S. E. (1982). Form and texture inhierarchically constructed patterns. Journal of Ex-perimental Psychology, 8, 521–535.

Kraiger, K., Billings, R. S., & Isen, A. M. (1989). The in-fluence of positive affective states on task perceptionsand satisfaction. Organizational Behavior and Hu-man Decision Processes, 44, 12–25.

Kuhl, J., & Kazen, M. (1999). Volitional facilitation ofdifficult intentions: Joint activation of intentionmemory and positive affect removes Stroop interfer-ence. Journal of Experimental Psychology: General,128, 382–399.

Lerner, J., Small, D. A., & Loewenstein, G. (2004).Heart strings and purse strings: Carryover effects ofemotions on economic decisions. Psychological Sci-ence, 15, 337–341.

Loewenstein, G., & Lerner, J. S. (2003). The role of af-fect in decision making. In R. J. Davidson, K. R.Scherer, & H. H. Goldsmith (Eds.), Handbook of af-fective sciences (pp. 619–642). New York: OxfordUniversity Press.

Mackie, D. M., & Worth, L. T. (1989). Processing defi-cits and the mediation of positive affect in persua-sion. Journal of Personality and Social Psychology,57, 27–40.

Mackie, D. M., & Worth, L. T. (1991). Feeling good butnot thinking straight: The impact of positive mood onpersuasion. In J. P. Forgas (Ed.), Emotion and socialjudgment (pp. 201–220). Oxford: Pergamon Press.

Mano, H. (1994). Risk taking, framing effects, and af-fect. Organizational Behavior and Human DecisionProcesses, 57, 38–58.

Mantel, S. P., Kardes, F. R., & Isen, A. M. (2007). Ef-fects of positive affect on omission detection in themultiattribute evaluation and Ellsberg Paradigms.Unpublished manuscript, Indiana University.

Martin, L. M., Ward, D. W., Achee, J. W., & Wyer, R.A. (1993). Mood as input: People have to interpretthe motivational implications of their moods. Jour-nal of Personality and Social Psychology, 64, 317–326.

Mednick, M. T., Mednick, S. A., & Mednick, E. V.(1964). Incubation of creative performance and spe-cific associative priming. Journal of Abnormal andSocial Psychology, 69, 84–88.

Mellers, B. A. (2000). Choice and the relative pleasureof consequences. Psychological Bulletin, 126, 910–924.

Melton, R. J. (1995). The role of positive affect in syllo-gism performance. Personality and Social PsychologyBulletin, 21, 788–794.

Murray, N., Sujan, H., Hirt, E. R., & Sujan, M. (1990).The influence of mood on categorization: A cognitiveflexibility interpretation. Journal of Personality andSocial Psychology, 59, 411–425.

Navon, D. (1977). Forest before trees: The precedenceof global features in visual perception. Cognitive Psy-chology, 9, 353–383.

Nygren, T. E., Isen, A. M., Taylor, P. J., & Dulin, J.(1996). The influence of positive affect on the deci-sion rule in risk situations: Focus on outcome (espe-cially avoidance of loss) rather than probability.Organizational Behavior and Human Decision Pro-cesses, 66, 59–72.

Roediger, H. L., III, McDermott, K. B., & Robinson, K.J. (1998). The role of associative process in creatingfalse memories. In M. A. Conway, S. E. Gathercole,& C. Cornoldi (Eds.), Theories of memory II (pp.187–245). Hove, UK: Psychological Press.

Roehm, M., & Sternthal, B. (2001). The moderating ef-fects of knowledge and resources on the persuasiveimpact of analogies. Journal of Consumer Research,28, 257–272.

Russell, J. A., & Carroll, J. M. (1999). On the bipolarityof positive and negative affect. Psychological Bulle-tin, 125, 3–30.

Sanbonmatsu, D. M., & Kardes, F. R. (1988). The ef-fects of physiological arousal on information pro-cessing and persuasion. Journal of Consumer Re-search, 15, 379–385.

Schwarz, N. (2002). Situated cognition and the wisdomof feelings: Cognitive tuning. In L. F. Barrett & P.Salovey (Eds.), The wisdom in feeling (pp. 144–166).New York: Guilford Press.

Schwarz, N., & Bless, H. (1991). Happy and mindless,but sad and smart? The impact of affective states onanalytic reasoning. In J. P. Forgas (Ed.), Emotion andsocial judgment (pp. 55–71). Oxford: PergamonPress.

Simonson, I. (1989). Choice based on reasons: The caseof attraction and compromise effects. Journal ofConsumer Research, 16, 158–174.

Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D.G. (2004). Risk as analysis and risk as feelings. RiskAnalysis, 24, 311–322.

Staw, B. M., & Barsade, S. G. (1993). Affect and mana-gerial performance: A test of the sadder-but-wiser vs.happier-and-smarter hypotheses. Administrative Sci-ence Quarterly, 38, 304–331.

Storbeck, J., & Clore, G. L. (2005). With sadness comesaccuracy, with happiness, false memory: Mood and


the false memory effect. Psychological Science, 16,785–791.

Teasdale, J. D., & Fogarty, S. J. (1979). Differential ef-fects of induced mood on retrieval of pleasant andunpleasant events from episodic memory. Journal ofAbnormal Psychology, 88, 248–257.

Trope, Y., & Pomerantz, E. M. (1998). Resolving con-flicts among self-evaluative motives: Positive experi-ences as a resource for overcoming defensiveness.Motivation and Emotion, 22, 53–72.

Tversky, A., & Gati, I. (1978). Studies of similarity. In E.Rosch & B. B. Lloyd (Eds.), Cognition and categori-zation (pp. 79–98). Hillsdale, NJ: Erlbaum.

Tversky, A., & Kahneman, D. (1981). The framing ofdecisions. Science, 211, 453–458.

Watson, D., & Tellegen, A. (1999). Issues in the dimen-

sional structure of affect: Effects of descriptors, mea-surement error, and response formats-Comment onRussell and Carroll (1999). Psychological Bulletin,125, 601–610.

Wegener, D. T., Petty, R. E., & Smith, S. E. (1995). Posi-tive mood can increase or decrease message scrutiny:The hedonic contingency view of mood and messageprocessing. Journal of Personality and Social Psy-chology, 69, 5–15.

Weiner, B. (1985). “Spontaneous” causal thinking. Psy-chological Bulletin, 97, 74–84.

Yang, H., Ceci, S., & Isen, A. M. (2006, January). Posi-tive affect increases monitoring and does not increasefalse memory. Poster presented at the annual meetingof the Society for Personality and Social Psychology,Palm Springs, CA.


C H A P T E R 3 5

Advances in Modeling Emotionand Thought

The Importance of Developmental, Online,and Multilevel Analyses

NANCY L. STEIN, MARC W. HERNANDEZ, and TOM TRABASSO

How do we process, represent, and understandemotional experience? How can we best de-scribe the memory, thinking, and evaluationprocesses that continually accompany the ex-perience of emotion? How do memory andthinking processes interleave with the physio-logical and neurophysiological processes ofemotion? What are the developmental originsof this representational process? Do the repre-sentation and expression of emotional experi-ence change as a function of development, orare stable representational structures present atbirth, only to be fine-tuned by environmentaland physiological factors? These are the ques-tions that motivate the theoretical overviewpresented in this chapter.

In the last 20 years, serious attempts havebeen made to describe the thinking and physio-logical processes that accompany emotionalexperience. These theoretical approaches(Frijda, 1987, 1988, 2005; Lazarus, 1982,1984, 1991; Oatley & Johnson-Laird, 1987;

Sandler, Grandjean, & Scherer, 2005; Scherer,1984; Roseman, 1991; Stein & Jewett, 1986;Stein & Levine, 1987; Stein, Trabasso, &Liwag, 1993, 2000; Zajonc, 1980), despitetheir different perspectives, derive from an ef-fort to describe the ways in which thinking,planning, and action interact with the physio-logical and neurological processes that occurwhen an emotion is experienced. The goal ofthis chapter is to discuss some of the difficul-ties associated with modeling emotion andthought, and to bring evidence to bear on a setof issues that has influenced our models andunderstanding of emotion and thought.

One feature that distinguishes our approachfrom other models of emotion and thought isthat we study emotion and thought in childrenas well as adults. Our approach is further dif-ferentiated from others because we examinethe online thinking and appraisal processesthat precede, accompany, and follow the expe-rience of emotion. Thus we focus on the entire

574

emotion episode, not just individual compo-nents of an emotion episode (such as physicalemotion expression, physiological change, orverbal reports). Our analysis is grounded in atheory of goal-based action that elucidates theattention, thinking, and appraisal processes oc-curring during the course of emotional experi-ence and intentional action (Stein et al., 1993,2000; Stein & Hernandez, 2007). We also ex-amine the physiological processes that accom-pany goal-based action, especially those thatprecede, accompany, and follow emotional ex-perience. Heart rate, blood pressure, the activa-tion of the autonomic nervous system (ANS),changes in galvanic skin response, and changesin cortisol levels are all physiological measuresthat we consider (Stein & Boyce, 2005).

Goals are central to our theory of emotion,because the monitoring of personally signifi-cant goals, in terms of their failure and success,set the conditions for the experience and evoca-tion of emotion. The evocation and experienceof emotion occur when unexpected changes inpersonally significant goals are perceived. Theepisode that surrounds an emotion beginswhen a precipitating event occurs and alerts aperson to some type of change in a personallysignificant goal. Once a person perceives thischange, all cognitive effort is focused on deter-mining the nature of the change and the impactthe change will have on personally relevantgoals. We define an “emotion episode” as a se-quence of events that includes the precipitatingevent; appraisals of the change in the status of agoal; the physiological and neurophysiologicalreactions that occur in relation to the change;the emotional reaction itself; and subsequentappraisal, planning, and behavior sequencescarried out to cope with the impact of the goalchange.

An emotion is said to occur when four di-mensions are present: (1) when unexpectedchanges occur in personal goal states; (2) whensubsequent changes occur in the ANS to shiftattention to the changes in goal states; (3) whencognitive evaluations have been carried out toassess the impact of these changes on the per-son’s well-being; and (4) when beliefs about theprobability of attaining, maintaining, avoiding,or preventing a goal state are activated.

“Beliefs” are organized forms of knowledgethat carry expectations about the state of someaspect of the world. They often carry a valuejudgment (e.g., good or bad) or a preference(e.g., like or dislike) for the aspect of the object

being evaluated. Beliefs reflect what peoplethink is, was, or could be true about theirworld (Folkman & Stein, 1997; Stein &Hernandez, 2007; Stein & Levine, 1987, 1990,1999; Stein et al., 1993, 2000). Prior knowl-edge in the form of organized belief systems isactivated to evaluate and understand the per-sonal impact of an event on well-being. In acti-vating belief systems, people are able to gener-ate precise inferences about how an event hasalready affected them and how it will affectthem in the future.

The ways in which goal appraisal processesoperate during emotion understanding and re-membering correspond to a series of five ques-tions: (1) What happened? (2) How do I feelabout it? (3) What can I do about it? (4) Whatdid I do about it? and (5) What were the conse-quences of carrying out a plan of action? (Stein& Hernandez, 2007). Each question reflects aclass of inferences made in relation to evaluat-ing the status of a goal that has been affectedby a precipitating event. We use these five ques-tions as a pedagogical device to organize ourinquiry into the sequences of goal appraisalprocesses that occur online during an emo-tional episode. The questions correspond to thetemporal and causal unfolding of an emotionalepisode and allow us to specify the exact natureof goal appraisal processes involved in under-standing and remembering emotional events.

The questions were derived by describing intoto all of the information contained in over1,000 different narratives generated as childrenand adults actually experienced an emotion(Stein & Hernandez, 2007; Stein, Trabasso, &Albro, 2001) or recalled a past experience andattempted to reenact the scenario (Liwag &Stein, 1995; Stein & Albro, 2001). Using theseonline procedures enables us to describe andunfold the dynamic properties and sequence ofbehavioral, cognitive, and physiological pro-cesses that interleave as emotions are experi-enced, expressed, and enacted. The language,emotion expressions, and emotional behaviorsare videotaped before, during, and after the on-line experience of emotion, enabling us to de-termine subjects’ focus of attention when emo-tions are experienced and expressed (Ross,Ross, Stein, & Trabasso, 2007; Stein & Boyce,2005; Stein & Hernandez, 2007; Stein,Sheldrick, & Broaders, 1999).

Using all of these techniques, we are able todocument the content and natural languageused during and after an emotion episode. We

35. Advances in Modeling Emotion and Thought 575

determine whether it matches the emotion ex-pression on the face, test how it corresponds tochanges in heart rate and blood pressure, judgethe accuracy of the memory for the emotion ep-isode, and assess how aware family membersare of each other’s emotions. Finally, we discussdata that show how emotions, appraisals, andplans affect and predict psychological well-be-ing.

We have found that the talk accompanyingemotional experience is so constrained that wehave created a computer program that encodesand analyzes much of the online talk of emo-tion and coping (Stein & Hernandez, 2007).Although some researchers are wary of the in-finite variation in and unreliability of talksurrounding emotion (see LeDoux, 2002;Mandler, 1975, 1984, for examples), we havefound just the opposite to be the case. The con-straints on the talk of emotional understandingoccur over development and across cultures.Describing the online thinking, planning, andtalk about emotion has allowed us to outlinethe temporal and causal sequences accompany-ing emotional experience.

The ability to outline the temporal andcausal sequences of the cognitive processing as-sociated with emotion experience is a uniqueadvantage of our methodology. We fully sup-port and use physiological measures and neu-roscience technology to advance our theory ofemotion, but we are also aware of the limi-tations that the technology of neuroscienceplaces on the temporal description of emotionprocesses. Thus we use those modalities in con-junction with paradigms that allow us to probemore deeply into the processes of emotion ex-perience, and we then try to map the thinking,talk, and behaviors of emotion onto the physi-ology and neurophysiology of emotion.

When researchers fail to unfold the thinkingprocesses involved in emotion experience, seri-ous theoretical errors are made. Specifically,the amount and type of thinking and decisionmaking that occur during emotional experienceare underestimated. Descriptions of the organi-zation of the temporal and causal sequence ofemotional experience are incorrect. The speedwith which cognitive processes are carried outis sorely underestimated. Finally, descriptionsof unconscious and conscious processing arelacking in accuracy, in terms of when peopleare consciously aware of the events surround-ing emotion and when they use unconsciousprocessing to encode and interpret the meaningof an event.

MODELING EMOTIONAND THOUGHT

Throughout this chapter, we argue that bothchildren and adults use conscious and uncon-scious cognitive processes to understand, eval-uate, plan, and respond to events that provokeemotion. From the very beginning of life, ex-periencing and expressing emotion are goal-based, expressive, and action-oriented. Experi-encing emotion involves continual monitoringof personally relevant goals, and involves aconstant appraisal of the value and worth ofevents, people, objects, activities, ideas, inter-nal states, and anything else that impinges onthe psychological and physical well-being ofthe person involved. The monitoring of goalsrequires both unconscious and conscious pro-cessing, and the evocation of emotion is the re-sult of both types of processing. One of themain questions surrounding the description ofthe thinking that accompanies emotion is whenand under what conditions emotional process-ing results in conscious awareness, as opposedto remaining an unconscious process.

As we have shown in almost all of our workon children and adults over the last 25-plusyears, everyone, including 3-year-old children,is highly aware of the causes and consequencesof emotional experience (Stein & Albro, 2001;Stein & Levine, 1989, 1999; Trabasso, Stein,& Johnson, 1981). In talking about the causesof their emotions, people can verbalize thegoals that have changed, those that are atstake, and those that have failed. They can dis-tinguish between their desires for action andthe actions that actually get carried out, andthey are keenly aware of the consequences oftheir actions. Although many current lines ofresearch are aimed at studying the unconsciousprocesses involved in the experience of emo-tion, it is important that we describe the con-scious processes as well as the unconscious pro-cesses.

The same should hold for the study of devel-opmental processes. Although the emphasis indevelopment is on the nature and type ofchange that occur over the lifespan, similaritiesin functioning and understanding must also beconsidered. The fact that emotion is central tosurvival and is central to a theory of evolutionsuggests that the very young may experiencerapid growth and understanding of emotionand its impact on behavior.

The belief that goal-directed action is centralto emotional processing is not a new idea. Al-


most all current theories of emotion speak tothe necessity of goals, values, desires, concerns,or motives as one of the critical components“driving” appraisal processes, which interactwith and regulate emotional experience. Thewords that theorists use for goals are different,but the concepts and final results are the same.Much of the goal-based theory used in modelsof emotion theory derives from Miller,Galanter, and Pribram’s (1960) seminal workon the importance of goals and planning in allof human action and interaction.

Magda Arnold (1960a, 1960b) influencedmany emotion researchers (see Frijda, 1987,1988; Lazarus, 1991; Stein & Oatley, 1992;Scherer, 1984) because of her sensitivity to andnecessity for characterizing the importance ofappraisal and evaluation. However, in order toadvance such a theory, a model of the onlinetemporal and causal properties of emotionneeds to be formulated. The necessary step inbuilding a theory of emotion and thinking, es-pecially one that focuses on the dynamic un-folding and expression of emotion and behav-ior, is to describe the content and organizationof the appraisal and goal-based thinking pro-cesses that occur from the beginning to the endof an emotion episode (Stein & Hernandez,2007; Stein et al., 1993; Trabasso & Stein,1993). The theory of Miller et al. (1960) hasbeen critical in this endeavor. These researchersmodeled the ways in which goals regulate theplanning process and constrain the actions thatare undertaken, as well as the outcomes associ-ated with goal attainment. Our theory buildson this model and so do other appraisal theo-ries, especially those of Frijda (1987), Oatleyand Johnson-Laird (1987), Scherer (1984), andRoseman (1984, 1991).

Constraints on the outcomes of goals are im-portant in mapping patterns of appraisal pro-cesses onto emotions. The pursuit of a goal canend in only three ways: People can be success-ful in achieving a goal; they can fail at achiev-ing a goal; or the attainment of a goal can re-main uncertain, with neither success nor failureoccurring. The limited number of possible goaloutcomes constrains the experience of emo-tion, even across development and culture. Theachievement of goals is responsible for positiveemotion, while the failure or blockage of goalsis responsible for negative emotion. These out-comes result in the familiar valenced nature ofemotion. Despite the frustration of philoso-phers in terms of using valence to classify emo-tion (Charland, 2005), the specific combina-

tions of goals and outcomes associated witheach of the five basic emotions remain constantacross situations and people. Thus understand-ing the definition of a goal and the ways goalsconstrain outcomes is essential for understand-ing emotion and testing an appraisal theory ofemotion and goal-directed action.

In our theory of emotion, a “goal” is definedas a desire to go from one state to another or asa desire to maintain a specific state. A goal mayinvolve any valued object, activity, or state thata person wants to attain or maintain. Goals of-ten occur in conjunction with statements ofpreference, which almost always include somereference to liking, loving, missing, hating,avoiding, or disliking: “I really like it when wecan just spend some quiet time without fight-ing,” “ I really hate the way he leaves his rooma mess in the morning.” A preference may im-ply a goal, but preference by itself is not a goal.Preferences encode an evaluation of the posi-tive or negative value of something, but neednot include the desire to change from one stateto another. Preferences reflect the pleasure anddispleasure associated with anything of value,but do not necessitate a desire to change fromone state to another. Appraising the value ofsomething is necessary but not sufficient for theemergence and pursuit of a goal. People need todecide, either consciously or unconsciously, topursue a goal that would result in a state con-cordant with a given preference.

Linguistically, goals can be identified byverbs such as “wish,” “want,” “decide,” “go-ing to do,” “try to do,” or “must do”: “I reallywish she didn’t interfere so much,” “I wish Icould make him understand what I want,” “Ireally don’t want her to eat so much junk foodafter school,” “I wish he would have a morepositive attitude toward helping around thehouse.” Goals can also be identified as thestated reasons or purposes for actions thatwere taken. Goals or reasons motivate actionand are often connected to them by words orphrases such as “to,” “for,” “so that,” or “inorder to”: For example, “I read some of her fa-vorite books so that she would go to sleep ather proper bedtime.”

Even in infancy, at least from the age of 3months on, children understand emotion-ladenevents in an organized and systematic way,much as adults do. The basis of their under-standing, like that of adults, is dependent upona continual evaluation and appraisal of incom-ing information, especially information that isrelated to personally meaningful goals and to


the success or failure of a goal. The critical taskis to show that monitoring of a goal occursvery early in infancy, and that emotions occurbecause of perceived changes in importantgoals. The use of the word “goal” has beenproblematic for many researchers, because theyenvision a goal as something abstract and inac-cessible to an infant. However, goals can bevery concrete and very simple. The desire toeliminate noxious stimulation (e.g., the desireto stop someone who is restraining an infant’smovement), the desire to get out of a state ofhunger, the desire to get out of a state of wet-ness, and the desire for stimulation all havevery concrete aspects of goal-directed action at-tached to them.

The work of Stenberg and Campos (1990),Willatts (1990), and Woodward (1998) allshow strong support for infants’ sensitivity togoals. Furthermore, the work of Lewis (1990)and his colleagues (Alessandri et al., 1990)shows that very young infants can and do ha-bituate to incoming information, that they arequite good at detecting novelty and discrepan-cies, and that they respond differently to theonset of different events. As both Sroufe (1979)and Emde (1980) have noted, if we take thetime to watch and film infants in dynamic ac-tion, infants engage in quite a bit of looking,gaze shifting, and eye contact before they re-spond emotionally and with action. Their emo-tional responses do not occur spontaneously orautomatically in response to most stimuli.Many emotion expressions occur after pro-longed looking or after a switch of attentionand more looking. Stenberg and Campos(1990) showed that this is the case when in-fants express anger at being physically con-strained. Thus, in describing the unfolding ofaction over time, characterizing the looking, at-tention, and shifts in gaze is critical to deter-mining the nature and sequence of thoughtsand actions during the experience of emotion.

EMOTIONAL VERSUS AFFECTIVERESPONSES

Both infants and adults respond automaticallyand affectively to certain types of stimuli. Au-tomatic physical responses are generated in afixed action pattern, without evaluating in-formation. The startle response is one suchexample. The eye-blinking response is another.These types of responses are seen frequently in

children and to some extent less frequently inadults, due to adults’ ability to cognitivelydampen or inhibit these responses. Affective re-sponses can also include feelings of general dis-tress and are evoked when the speed, intensity,and duration of the stimulus (usually sensorystimulation) are at a high level. Affective re-sponses are evolutionary primitive signals thatcommunicate a state of change in physiologicalreactivity. The ANS is aroused, with little cog-nitive activity preceding the arousal.

The important point we wish to make is thatautomatic responses to sensory input requirelittle or no cognitive evaluation. They are re-flexive in nature. Affective feelings require littlecognitive evaluation and play an important in-formational role in indicating the changes thatoccur either in the environment or in the body.In our theory, however, emotional responsesare distinct from automatic, affective, andphysical responses, in that emotions requirehigher-order cognitive processes and encode aplan of action. Affective responses do not re-quire higher-order cognitive responses and donot encode a plan of action.

In our online studies, affective responses pre-cede emotional responses, with the affective re-sponses occurring rapidly in response to an ini-tiating event and resulting in a “stunned”demeanor, followed by evaluation of the in-coming information, an emotional response,and a plan of action. A pause almost always oc-curs between the affective and emotional re-sponses. Appraisals and planning occur after aswell as during the experience of emotion (Stein& Boyce, 2005).

Failure to distinguish between affective andemotional responses is one of the most frequentreasons for different stances on the relationshipbetween emotion and thought. A failure to un-derstand how thought gets carried out over thecourse of emotional experience, where talk isinterleaved with nonverbal components ofemotional experience, is a second difference inthe various emotion and appraisal theories thatexist. Researchers like LeDoux (2002) interpretthe freezing response that goes along with in-tense stimulation or the introduction of noveltyas an innate action that gets carried out in theservice of protecting an animal or human be-ing.

In our theory, freezing is a response that au-tomatically occurs without cognitive evalua-tion or planning. Freezing may occur becausethe incoming stimulation has triggered an in-


tense ANS response that momentarily inhibitsaccess to any thinking and planning. As the in-tensity of the physiological response declines,however, access to planning and thinking is re-instated. If animals and humans engage in be-haviors such as “flight or fight,” we posit thatsome form of evaluation or goal-directed plan-ning precedes the flight or fight response of ananimal. The timing and unfolding of the flightor fight response are critical, as is the measure-ment of thinking.

Despite the fact that many situations engen-der very rapid action responses in animals andhumans, rapid cognitive evaluations can pre-cede almost any directed action pattern. Thespeed with which cognitive processes get car-ried out is extremely fast—just as fast as, orfaster than, access to a motor plan of action.Many cognitive processes are parallel in na-ture, and are in part unconscious and auto-matic. To withdraw from a painful stimulus re-quires the retrieval of a plan of action forwithdrawal. Without the plan, extreme distressoccurs in reaction to pain, along with bodywrithing, but no withdrawal action will be car-ried out. Similarly, approaching something re-quires an evaluation of the object being ap-proached, as well as a plan of action. The focusof attention needs to be shifted toward the ob-ject being approached, and the object needs tobe constantly monitored in order to keep thetarget on the path of approach. Gaze behavior,a shift in attention, and the speed of attack allsuggest that rapid evaluation of the situationhas been carried out. The important point is toacknowledge the speed with which cognitiveprocesses get carried out, especially since manyof the cognitive processes are automatic andunconscious.

To determine whether emotion precedesthinking and evaluation in any critical fashion,we need to focus on unfolding in a dynamicfashion all of the behaviors that get carried outin reaction to an initiating event that begins thesequence of emotional experience. We mustalso determine whether distinctions betweenemotional and affective responding are valid.Distress, startle, surprise, and stun responsesare not considered emotions in our theory, be-cause a plan of action is not encoded in any ofthese responses. As we have illustrated in bothour nonverbal and verbal data, distress re-sponses almost always precede emotional re-sponses and goal-directed action. Furthermore,the expression of startle or surprise always pre-

cedes the expression of fear (Stein & Boyce,2005), if startle and surprise occur in the sameemotion sequence.

When children experienced a fire alarm go-ing off, Stein and Boyce (2005) were able tomonitor talk, emotion experience, and expres-sion, as well as heart rate and blood pressure,throughout a 10-minute session. They werealso able to collect retrospective memory ac-counts of the fire alarm experience, so that ac-curacy for emotion experience and the eventssurrounding the fire alarm could be evaluated.Stein and Boyce found that of the 24 childrenparticipating in their study, 22 expressed sometype of emotion in reaction to the fire alarm’sgoing off. Of those children who expressed anemotion, 8 out of 22 expressed the affectivestate of surprise, followed by either fear or hap-piness. Surprise was always followed by the ex-pression of an emotional state.

Furthermore, children who were classified as“low reactors” (i.e., their heart rate and bloodpressure decreased as a function of the firealarm’s going off) were more likely to be the oneswho experienced an affective state coupled withan emotion. Seventy-five percent of the low reac-tors expressed startle, compared to 25% of the“high reactors.” A second difference betweenthe high and low reactors was in the length oftime that affective states and emotions were ex-pressed. The average length of time that an affec-tive or emotion state was expressed by low reac-tors was 5 seconds, compared to 11.5 secondsfor high reactors. Thus the expressive time forhigh reactors was twice as long as that for low re-actors. After the fire alarm was turned off, signif-icant differences between high and low reactorswere still apparent. Over half of the low reactorsexpressed no emotion after the alarm went off,whereas 84% of the high reactors continued toexpress an emotion state and did not return to aneutral stance.

The affective response of startle was alwaysfollowed by more scanning of the environment,and then fear responses were expressed. Thefixed nature of this sequence suggests stronglythat affective responses are mechanisms thatfurther alert the cognitive and physiologicalsystems that something has changed. However,further appraisal and scanning were necessaryto encode exactly what had changed, afterwhich an emotion was expressed. The length oftime varied in terms of the duration of emo-tion, and half of the children had difficulty re-turning to a baseline stance.


These findings underscore the fact the signif-icant individual differences are expressed in theability to think and plan during emotional ex-perience. Although emotions do encode plansof action, and very rapid action can be takenwhen emotions are experienced, the person ex-periencing the emotion may not be able tosolve the problem or cope with the emotion sit-uation in a constructive fashion. The fact thatarousal remained high for half of the childrenin the fire alarm study, especially for high reac-tors, suggests that more reflective planningmay not be available until arousal returns tonormal.

UNFOLDING APPRAISAL PROCESSESFOR EMOTIONS

Since all current models of emotion (see Frijda,2005, and Sandler et al., 2005, for examples ofother appraisal models) argue for the use of anappraisal system that regulates emotional re-sponses, it becomes imperative that appraisalsbe clearly defined, described, and identified, sothat tests of an appraisal theory can be madeand compared to theories that do not posit thenecessity of appraisals. As we discuss ourmodel of the appraisal process, we highlightthe differences between our model and those ofother researchers.

Our model includes descriptions of the con-tent of appraisals, as well as descriptions ofhow children and adults are similar to and dif-ferent from one another. We require that pro-cesses of emotion be described as they unfoldover time, in real situations. We need to deter-mine the sequence of processes that get carriedout, which ones occur in parallel with others,and which ones occur in a temporal sequence.Without these data, we cannot test the crucialhypotheses associated with different theories ofemotion and appraisal.

Elsewhere, we and our associates have de-scribed the appraisal and planning processesthat evoke emotions such as happiness,sadness, anger, fear, and disgust (Stein &Hernandez, 2007; Stein & Jewett, 1986; Stein& Levine, 1987, 1989, 1990; Stein &Trabasso, 1992; Stein et al., 1993, 2000). Thespecific content of the appraisal process thatleads to different emotional experiences indi-cates how appraisals of each emotion areunique in relation to other emotions. UnlikeOrtony and Turner (1990) and Johnson-Lairdand Oatley (1989, 1992), who believe that sin-

gle components of the appraisal process canelicit an emotion, we believe that at least threegoal outcome evaluations must be inferred andnecessary for any one specific emotion to be ex-perienced. First, an evaluation must be madeabout the certainty or uncertainty of goal at-tainment (did a goal fail, succeed, or remainuncertain/threatened?). Second, an evaluationmust be made about whether a goal state canbe reinstated, maintained, avoided, or pro-tected. Third, an evaluation of the outcomeand of the future possibilities for goal successor failure is crucial for the experience of anemotion.

The presence of a novel or unexpected ele-ment in an event must also be detected.Novelty signals that some component of theemotion situation is unexpected and requiresimmediate attention and immediate processing.The necessity of novelty or an unexpected ele-ment in an ongoing event was seminal to bothMandler’s (1975, 1984) and Gray’s (1971;1990) descriptions of emotion processing.Scherer (1984) picked up on the importance ofnovelty in emotion processes, but never gave anexplanatory analysis of why he chose to incor-porate novelty into his theory of appraisals.Frijda (1988) argued for the necessity of con-sidering habituation (repeated processing ofthe same event) in relation to emotion and nov-elty, but never provided empirical evidence toanswer his question of what happens to think-ing and emotion when an event is repeated overtime. Frijda concluded that one habituates toevents that cause joy and pleasure, but does notget used to aversive events, harassment, or hu-miliation (i.e., the law of hedonic asymmetry).

The data from our repeated-event study(Stein & Trabasso, 1989) showed that somechildren habituated to both painful and pleas-ant events. The repetition of a painful event al-lows an increase in tolerance for a particulartype of stimulation, so that the meaning and se-riousness of the painful event change. Trainingsoldiers for war is a good example of people’sbecoming habituated to events that are initiallypainful, aversive, and noxious, but over timebecome less painful, even when intense horrorand pain are initially perceived. The key towhether habituation or nonhabituation occurslies in the meaning given to an event, the con-text in which the event is interpreted, and theintensity of pain that the event causes.

During habituation, events that were ini-tially interpreted as horrific can be appraised asacceptable and necessary to the final achieve-


ment of an important goal. The dropping of anuclear bomb is a good example. So is killing aperson at short range. Indeed, a major problemthat the armed forces of any country must con-front is the challenge of debriefing soldiers andproviding them with the necessary psychologi-cal and physical training to desist in the use ofcertain appraisal strategies that are appropriatefor coping with harm on the battlefield, but donot work in civil society.

In terms of testing whether the perception ofnovelty is required for an emotion to be experi-enced, however, a repeated-event paradigm isan excellent way to assess how thought relatesto emotion. Emotions do not occur withoutsome type of internal or external event’s im-pinging on an organism, both psychologicallyand physiologically. If novelty is a necessarycomponent of emotion, then over time, as aperson gets used to and encodes the basic struc-ture of a presented event, one of two thingsshould happen. If the person adds no newmeaning to the incoming information from arepeated event, then the person should desistfrom responding emotionally as the event ispresented repeatedly. A neutral stance shouldemerge without any emotion. If, however, newmeaning is inferred from observing a repeatedevent, so that the event is now connected tonew information not accessed previously, anemotional response should continue to be ex-perienced and expressed.

This is exactly what happened in the Steinand Trabasso (1989) study. When repeatedlypresented with an event that ended in a positiveor negative outcome, some children gave emo-tional responses of lessening intensity, whileother children stopped expressing emotion al-together. Some children, however, continued togive emotional responses that were as strongas, or stronger than, their reaction the first timethe event was presented. When asked why theyresponded with such strength, all children men-tioned the focus of attention on a novel elementthat for them was unexpected. That is, theylinked the information from the focal event tonew information that had consequences forpersonally relevant goals, not previously notedon their first exposure to the event. Thus newappraisals were made about an event despitethe repeated nature of the event. Whether ornot habituation occurred depended upon theinterpretation of the event and the goals theperceiver brought to the situation. Further-more, it was possible to habituate to negativeevents as well as positive events, and it was also

possible to interpret a repeated positive or neg-ative event in a new light, despite the frequencyof presentation. The crucial element was themeaning attributed to an event.

Although emotions such as happiness, anger,sadness, and fear by no means encompass thespectrum of possible emotions, we have usedthem as examples because, along with disgust,these five emotions represent a set of emotionsfound in some form in every known culture.Each of these emotions corresponds to a partic-ular goal–outcome combination in terms ofwhether or not valued goals have failed or beenfulfilled (Stein et al., 1993, 2000). Although theexact label given to each of these five emotionstates may vary across development and cul-ture (e.g., “angry,” “mad,” “furious”), the goaloutcomes, initial desires, and types of apprais-als associated with each of the five emotions donot vary.

CONSTRAINTS ON THE MENTALREPRESENTATIONS AND LANGUAGEOF EMOTION

The mapping of specific types of goal outcomesonto each emotion is hard and fast. Emotioncategories do not have fuzzy boundaries,as initially suggested by Shaver, Schwartz,O’Connor, and Kirson (1987). When anger isexperienced, the event under consideration hasto have led to the failure of an important goalor the inability to avoid an aversive state. Fur-thermore, the person experiencing anger has tobelieve that the focal goal can somehow be re-instated, that an aversive state can be removed,or that the source of goal failure can be re-moved. These types of inferences remain neces-sary over age and culture, despite variation inthe specific events under consideration or thesociety’s approval or disapproval in regard tothe actions carried out in the service of a goal.

The language associated with emotional ex-perience is also highly constrained. We (Stein& Hernandez, 2007) have illustrated that peo-ple use a very restricted range of emotion ter-minology when they freely narrate a past event.The constraints hold for adults (Stein et al.,2001) as well as for children (Stein &Hernandez, 2007). In analyzing over 1,000narrative accounts from both adults and chil-dren, we found that the range and frequency ofnegative emotions were focused on 12 specificemotion states. Four emotion terms and theirvariants (“sad,” “angry,” “afraid,” “anxious”)


accounted for 95% of all expressed negativeemotions during talk about stressful, traumat-ic, or unexpected events. Similarly, three emo-tion terms and their variants (“happy,” “hope-ful,” “relieved”) accounted for over 97% of allexpressed positive emotions during such talk.These data provide strong evidence that a smallset of emotion terms is used to describe alltypes of stressful experiences.

Most of the time, the emotional terminologyof adults in describing ongoing or past events ishighly similar to the language of children.Adults produce more mental state and emotionlanguage than children do, but the same cate-gories of words are used. Adults’ and children’semotion narratives, in terms of the presence ofmental state beliefs, are also similar. Three-year-old children produced statements in everymental and internal state category that wasproposed in the Stein and Hernandez (2007)study. Thus, before we conclude that emotionalexperience is infinitely varied over develop-ment and time, we need data from real-worldexperiences, where adults and children actuallyuse emotion and mental state language that hasnot been constrained by any particular instru-ment or theory.

Finally, the analysis of children’s emotionalaccounts elicited during an ongoing event(Stein & Hernandez, 2007), or told retrospec-tively right after the event has occurred (Stein& Boyce, 2005), show that 3-year-old childrenhave a far greater understanding of events thanhas been evident in more experimental studies,where children are presented with events thatare somewhat unfamiliar to them or events thatthey have not actually experienced. When veryyoung children talk about events they have ac-tually experienced (Levine, Stein, & Liwag,1999; Stein & Albro, 2001), they narrate withthe precision and complexity of older children,and sometimes with far more detail than theirparents do. Furthermore, when children are inconflict with their parents, and are asked togive both their account as well as their parents’account of the conflict, children are often moreaccurate about their parents’ emotions andmental states than parents are about their chil-dren’s mental states (Stein & Hernandez,2007).

The rapid development of emotional under-standing, in terms of the mental representa-tions used to understand and respond to emo-tion events (Stein & Albro, 1997), stronglysuggests that much of the structure needed to

understand emotional interaction is well inplace by late infancy, before children are usinglanguage freely to communicate what theywant and know about emotion situations. Thedata also suggest that theories of emotion for-mulated by testing only adults will need testingon children to determine whether the processesof emotion posited in the theory also hold forchildren.

Learning and Emotional Experience

As a function of attending to novel informa-tion, learning almost always results duringemotional experience. Some type of violationof expectation always occurs before the experi-ence of emotion, and the violation signals thatcurrently held beliefs are incorrect (Stein &Jewett, 1986; Stein & Levine, 1987, 1989). Inan effort to understand the consequences ofsucceeding or failing to attain their goals, peo-ple are forced to revise and update their beliefsabout two dimensions: (1) what the probabilityis of maintaining their goals under a new set ofcircumstances; and (2) whether or not they cangenerate a plan that will lead to successful goalmaintenance, reinstatement, or revision. Forexample, when people fail to maintain or attaina desired goal state, they attend to the condi-tions that prevented them from being success-ful, the consequences of their failure, and theprobability of generating a plan of action toovercome goal failure. This sequence charac-terizes both adults’ and children’s emotional re-sponses (Capatides, 1989; Stein & Levine,1989; Stein & Hernandez, 2007), as well asparents’ descriptions and responses to theirchildren’s emotional behaviors (Capatides,1989; Levine et al., 1999).

In updating knowledge about viable condi-tions that lead to goal success or failure, peoplemay change the value they have imputed to aparticular set of goals. That is, when a personattains or fails to attain a goal, the value associ-ated with that and other goals may increase ordecrease in strength. As a result, people oftenforfeit their goals, or they may intensify theirefforts to achieve them, depending on the newvalue that these goals now possess.

In privileging intentional action as a corepart of our theory, we are assuming that peoplehave a built-in mechanism that allows themto represent goal–action–outcome sequencesin relationship to the maintenance of goals


(Gallistel, 1985; Piaget, 1981; Stein & Levine,1987, 1990; Stein et al., 1993, 2000). Even in-fants are able to infer and represent the causalconditions that produce actions resulting incertain outcomes, and they are able to use thisknowledge to achieve goals (Woodward,1998). Thus, when a change in goal status oc-curs and emotions are experienced, a revisedgoal is accessed and a plan is developed. Cre-ating or accessing a plan that specifies the con-ditions necessary to achieve a goal provides anopportunity for coping with goal failure. Simi-larly, constructing and carrying out a plan en-able the maintenance of a goal, once the goalhas been achieved.

There are situations, however, where littleplanning occurs or where plans are not easilyavailable and retrieved. Intense emotional ex-perience often precludes access to certain typesof information (Stein & Boyce, 1997, 2005).Under these conditions, critical inferencesabout the emotion situation are not made.Thus the intensity of an emotional experienceand the associated sensory and physiologicalfeedback become important variables in pre-dicting the thinking, decision making, andquality of planning that take place during anemotional experience.

MEMORY FOR EMOTIONAL EVENTS

Although there are reports that emotional in-volvement enhances the accuracy of memoryfor events (Burke, Heuer, & Reisberg, 1992;Goodman, 1991), this is not always the case(Peters, 1991; Rachman, 1978). Dependingupon the intensity of the emotional responseand the necessity to take action, memory forthe external events surrounding the emotioncan be diminished or facilitated. The relation-ship among memory, thinking, and emotion,then, must be stipulated according to the spe-cific circumstances that constrain the experi-ence.

Stein and Boyce (1997, 2005) reported thatmemory for a stressful event (a real-life firealarm experience) is regulated in part by thephysiological reactivity of the individual childinvolved. Those children who responded to thestressful event with an increase in heart rateand blood pressure had poorer memories forthe event than those children who respondedwith a decrease in heart rate and blood pres-sure. Children’s attention and emotional re-

sponses, as well as their memory, were also afunction of whether their heart rate and bloodpressure increased or decreased in response to astressful event. Children whose heart rate in-creased as a function of the fire alarm had lessaccurate memory for their emotional responseand less accurate memory for the events thatoccurred before and after the fire alarm wentoff.

Three types of memory were evaluated: thenumber of clauses occurring in free recall, thenumber of clauses occurring in probed recall,and the number of explanatory statementsgiven throughout the recall of the fire alarm ex-perience. Children who were identified as lowreactors recalled more clauses in both sponta-neous and probed recall, compared to childrenlabeled as high reactors. Furthermore, childrenlabeled as low reactors also gave more explana-tions or sought more explanations throughoutthe fire alarm experience than children labeledas high reactors.

The fact that low- and high-reacting childrenhad different quantities of memory, and thefact that these two groups of children also ex-perienced emotions for different lengths oftime, serve as evidence for the absolute neces-sity of beginning to link physiological and neu-rological processes to cognitive activity duringthe experience of an emotion. The fact thatlow-reacting children rarely lost sight of newincoming information, compared to high-reacting children (who were often stunned andexperienced a startle response), may indicatethat emotional events are processed differentlyas a function of reactivity, right from the begin-ning of an event. Stein and Boyce found thatonce heart rate and blood pressure were raised,they stayed raised for about 45 minutes and didnot decrease or increase quickly. These changesindicate that an intense emotional experiencemay regulate behavior far in advance.

CONCLUSION

The appraisal model that we advocate requiresthat we try to unfold the temporal properties ofthinking and emotion as they occur in real timethroughout an emotional experience. The fail-ure to use and present online data for psycho-logical, physiological, and neurological pro-cesses severely limits the accuracy of what issaid about the relationship between emotionand thought, as well as our ability to link the


psychological, physiological, and neurologicalprocesses that get carried out during emotionalexperience. The fact that the experience ofemotion is not recognized as a basic com-ponent associated with learning also limitshow much we understand about the functionof emotion. If belief violations occur everytime an emotion is experienced (current dataon emotion strongly supports this viewpoint),then we need to be focused on ways in whichemotion regulates and affects learning and theability to change.

The data that we have presented in thischapter show that the thinking and learningprocesses, as well as the affective and physio-logical processes surrounding emotion experi-ence, are far more complex than most emotionresearchers have described. Cognitive process-ing occurs much more rapidly than many re-searchers think, and the temporality of cogni-tive and physiological processes has yet to beelucidated in the necessary depth. Unfoldingthe timing of cognitive and affective variablesfurther shows the need to acknowledge themultilayered properties of both processes.What we have argued for is the existence oftwo very different types of affective processes.The first occurs in an automatic fashion, as afunction of incoming information and a changein ANS levels. These responses need little or nocortical input, and are not considered emo-tional responses. They have no plans of actionattached to them. Emotional responses includeaffective responses, but they occur because ofthe appraisal of personally meaningful goals.They also encode a plan of action.

We have further shown that despite the vari-ation in the language used to verbally expressdifferent emotions, everyday emotion languageused to describe personally meaningful eventsis constrained to the five basic emotions plus orminus a few others. These constraints are im-portant, because they reflect the constraints puton problem solving, planning, and behavior.Certain types of appraisals and plans will occuronly with certain types of emotions, and thisspecificity is seen in children as young as 3years, as well as in mature adults.

Finally, the link between physiological indica-tors of reactivity (heart rate and blood pressure)and memory has been shown to be significant.Although most researchers would predict thatthe experience of emotion is always correlatedwith a rise in blood pressure and heart rate, thisis not the case. Bauer, Quas, and Boyce (2002)

and Jacobson and Gottman (1998) have shownthat both increases and decreases in heart rateoccur in response to emotion-eliciting informa-tion. The Stein and Boyce (1997) data further in-dicate that the control of attention and subse-quent encoding processes are affected bywhether an increase or decrease occurs in heartrate and blood pressure. Linking these differentphysiological responses to brain activity duringthe experience of emotion would allow us to un-derstand better the variation in responses toemotion-eliciting stimuli. It may be that plan-ning and thinking are not as greatly disruptedduring emotional experience in low reactors asthey are in high reactors. This would explain thedifferent reports and beliefs about the salience ofemotion-laden memories in terms of their accu-racy. It would also allow us to better understandthe connection between thought and emotion, interms of how emotion regulates thought and viceversa.

REFERENCES

Alessandri, S. M., Sullivan, M. W., & Lewis, M. (1990).Violation of expectancy and frustration in early in-fancy. Developmental Psychology, 26(5), 738–744.

Arnold, M. B. (1960a). Emotions and personality: Vol.1. Psychological aspects. New York: Columbia Uni-versity Press.

Arnold, M. B. (1960b). Emotions and personality: Vol.2. Neurological and physiological aspects. NewYork: Columbia University Press.

Bauer, A. M., Quas, J. A., & Boyce, W. T. (2002). Asso-ciations between physiological reactivity and chil-dren’s behavior: Advantages of a multisystem ap-proach. Journal of Developmental and BehavioralPediatrics, 23(2), 102–113.

Burke, A., Heuer, F., & Reisberg, D. (1992). Remem-bering emotional events. Memory and Cognition,20(3), 277–290.

Capatides, J. (1989). Mothers’ socialization of children’saffect expression. Unpublished doctoral dissertation,Columbia University, New York.

Charland, L. C. (2005). The heat of emotion: Valenceand the demarcation problem. Journal of Conscious-ness Studies, 12(8–10), 82–102.

Emde, R. (1980). Levels of meaning in infant develop-ment. In W. A. Collins (Ed.), Minnesota Symposiumon Child Psychology (Vol. 13, pp. 1–38). Hillsdale,NJ: Erlbaum.

Folkman, S., & Stein, N. L. (1997). The analysis of be-lief and goal processes during reports of stressfulevents by caregivers of men with AIDS. In N. L.Stein, P. A. Ornstein, B. Tversky, & C. Brainerd(Eds.), Memory for everyday and emotional events(pp. 113–138). Hillsdale, NJ: Erlbaum.


Frijda, N. H. (1987). Emotion, cognitive structure, andaction tendency. Cognition and Emotion, 1, 115–143.

Frijda, N. (1988). The laws of emotion. American Psy-chologist, 43(5), 349–358.

Frijda, N. (2005). Emotion experience. Cognition andEmotion, 19(4), 473–497.

Gallistel, C. R. (1985). Motivation, intention, and emo-tion: Goal-directed behavior froma cognitiveneuroethological perspective. In M. Frese & J. Sabini(Eds.), Goal-directed behavior: The concept of actionin psychology (pp. 48–66). Hillsdale, NJ: Erlbaum.

Goodman, G. S. (1991). Commentary: On stress and ac-curacy in research in children’s testimony. In J. Doris(Ed.), The suggestibility of children’s recollections(pp. 77–82). Washington, DC: American Psychologi-cal Association.

Gray, J. A. (1971). The psychology of fear and stress.London: Weidenfeld & Nicolson.

Gray, J. A. (1990). Psychobiological aspects of relation-ships between emotions and cognition [Special issue].Cognition and Emotion, 4.

Jacobson, N., & Gottman, J. (1998). When men batterwomen. New York: Simon & Schuster.

Johnson-Laird, P. N., & Oatley, K. (1989). The lan-guage of emotions: An analysis of a semantic field.Cognition and Emotion, 3(2), 81–123.

Johnson-Laird, P. N., & Oatley, K. (1992). Basic emo-tions, rationality, and folk theory. Cognition andEmotion, 6(3–4), 201–223.

Lazarus, R. S. (1982). Thoughts on the relations be-tween emotion and cognition. American Psycholo-gist, 37, 1019–1024.

Lazarus, R. S. (1984). On the primacy of cognition.American Psychologist, 39(2), 124–129.


LeDoux, J. (2002). Synaptic self. New York: Viking.Levine, L., Stein, N. L., & Liwag, M. (1999). Remem-

bering children’s emotions: Sources of concordantand discordant accounts between parents and chil-dren. Developmental Psychology, 5(3), 210–230.

Lewis, M. (1990). The development of intentionalityand the role of consciousness. Psychological Inquiry,1(3), 231–247.

Liwag, M. D., & Stein, N. L. (1995). Children’s memoryfor emotion episodes: The importance of emotion en-actment cues. Journal of Experimental Child Psy-chology, 60, 2–31.


Mandler, G. (1984). Mind and body: Psychology ofemotion and stress. New York: Norton.

Miller, G. A., Galanter, E., & Pribram, K. H. (1960).Plans and the structure of behavior. New York: Holt,Rinehart & Winston.

Oatley, K., & Johnson-Laird, P. N. (1987). Towards acognitive theory of emotions. Cognition and Emo-tion, 1(1), 29–50.

Ortony, A., & Turner, T. J. (1990). What’s basic aboutbasic emotion? Psychological Review, 97, 315–331.

Peters, D. (1991). The influence of stress and arousal onthe child witness. In J. Doris (Ed.), The suggestibilityof children’s recollections (pp. 60–76). Washington,DC: American Psychological Association.

Piaget, J. (1981). Intelligence and affectivity. Palo Alto,CA: Annual Reviews.

Rachman, S. J. (1978). Fear and courage. San Francisco:Freeman.

Roseman, I. (1984). Cognitive determinants of emotion:A structural theory. Review of Personality and SocialPsychology, 11–36.

Roseman, I. (1991). Appraisal determinants of discreteemotions. Cognition and Emotion, 5(3), 161–200.

Ross, H., Ross, M., Stein, N. L., & Trabasso, T. (2007).How siblings negotiate and resolve their conflicts.Child Development, 77(6), 1730–1745.

Sandler, D., Grandjean, D., & Scherer, K. R. (2005). Asystems approach to appraisal mechanisms in emo-tion. Neural Networks, 18, 317–352.

Scherer, K. R. (1984). On the nature and function ofemotion: A component process approach. In K. R.Scherer & P. Ekman (Eds.), Approaches to emotion(pp. 293–318). Hillsdale, NJ: Erlbaum.

Shaver, P., Schwartz, J., O’Connor, C., & Kirson, D.(1987). Emotion knowledge: Further explanations ofa prototype approach. Journal of Personality and So-cial Psychology, 52, 1016–1086.

Sroufe, A. (1979). Socioemotional development. In J.Osofsky (Ed.), Handbook of infant development(pp. 462–516). New York: Wiley.

Stein, N. L., & Albro, E. R. (1997). The emergence ofnarrative understanding: Evidence for rapid learningin personally relevant contexts. Contemporary Issuesin Education, 60, 83–98.

Stein, N. L., & Albro, E. R. (2001). The origins and na-ture of arguments: Studies in conflict understanding,emotion, and negotiation. Discourse Processes,32(2&3), 113–133.

Stein, N. L., & Boyce, W. T. (1997). The role of individ-ual differences in reactivity and attention in account-ing for on-line and retrospective memory. Paper pre-sented at the biennial meeting of the Society forResearch in Child Development, Washington, DC.

Stein, N. L., & Boyce, W. T. (2005). The unfolding ofchildren’s emotional experience in response to astressful event: How reactivity influences emotion.Unpublished manuscript, University of Chicago.

Stein, N. L., & Hernandez, M. W. (2007). Assessingemotional understanding in narrative on-line inter-views: The use of the Narcoder. In J. A. Coan & J. J.B. Allen (Eds.), Handbook of emotion elicitation andassessment (pp. 298–317). New York: Oxford Uni-versity Press.

Stein, N. L., & Jewett, J. (1986). A conceptual analysisof the meaning of negative emotions: Implications fora theory of development. In C. Izard & P. Read(Eds.), Measuring emotions in infants and children(Vol. 2, pp. 238–267). Cambridge, UK: CambridgeUniversity Press.

Stein, N. L., & Levine, L. (1987). Thinking about feel-


ings: The development and organization of emo-tional knowledge. In R. Snow & M. Farr (Eds.), Ap-titude, learning, and instruction (Vol. 3, pp. 165–197). Hillsdale, NJ: Erlbaum.

Stein, N. L., & Levine, L. (1989). The causal organiza-tion of emotional knowledge: A developmental study.Cognition and Emotion, 3(4), 343–378.

Stein, N. L., & Levine, L. (1990). Making sense out ofemotion: The representation and use of goal struc-tured knowledge. In N. L. Stein, B. Leventhal, & T.Trabasso (Eds.), Psychological and biological ap-proaches to emotion (pp. 45–73). Hillsdale, NJ:Erlbaum.

Stein, N. L., & Levine, L. J. (1999). The early emergenceof emotional understanding and appraisal. Implica-tions for theories of development. In T. Dalgleish &M. Power (Eds.), Handbook of cognition and emo-tion (pp. 383–410). Chicester: Wiley.

Stein, N. L., & Oatley, K. (1992). Basic emotions: The-ory and measurement. Cognition and Emotion, 6(3–4), 161–168.

Stein, N. L., Sheldrick, R., & Broaders, S. (1999). Pre-dicting psychological well-being from beliefs andgoal appraisal processes during the experience ofemotional events. In S. Goldman, P. L. Van denBroek, & A. Graesser (Eds.), Essays in honor of TomTrabasso (pp. 279–302). Mahwah, NJ: Erlbaum.

Stein, N. L., & Trabasso, T. (1989). Children’s under-standing of changing emotion states. In C. Saarni &P. L. Harris (Eds.), Children’s understanding of emo-tion (pp. 50–77). New York: Cambridge UniversityPress.

Stein, N. L., & Trabasso, T. (1992). The organization ofemotional experience: Creating links among emo-tion, thinking and intentional action. Cognition andEmotion, 6(3–4), 225–244.

Stein, N. L., Trabasso, T., & Albro, E. R. (2001). Under-standing and organizing emotional experience: Auto-

biographical accounts of traumatic events. EnglishStudies of America, 19(1), 111–130.

Stein, N. L., Trabasso, T., & Liwag, M. (1993). The rep-resentation and organization of emotional experi-ence. Unfolding the emotional episode. In M. Lewis& J. Haviland (Eds.), Handbook of emotions(pp. 279–300). New York: Guilford Press.

Stein, N. L., Trabasso, T., & Liwag, M. D. (2000). Agoal appraisal theory of emotional understanding:Implications for development and learning. In M.Lewis & J. M. Haviland-Jones (Eds.), Handbook ofemotions (2nd ed., pp. 436–457). New York:Guilford Press.

Stenberg, C. R., & Campos, J. J. (1990). The develop-ment of anger expressions in infancy. In N. L. Stein,B. Leventhal, & T. Trabasso (Eds.), Psychologicaland biological approaches to emotion (pp. 247–282).Hillsdale, NJ: Erlbaum.

Trabasso, T., & Stein, N. L. (1993). How do we repre-sent both emotional experience and meaning?: A re-view of Richard Lazarus’ Emotion and Adaptation,New York: Oxford University Press, 1991. Psycho-logical Inquiry, 41(4), 326–333.

Trabasso, T., Stein, N. L., & Johnson, L. R. (1981).Children’s knowledge of events: A causal analysis ofstory structure. In G. Bower (Ed.), Learning and mo-tivation (Vol. 15, pp. 237–282). New York: Aca-demic Press.

Willatts, P. (1990). The development of problem-solvingstrategies in infancy. In D. Bjorklund (Ed.),Children’s strategies: Contemporary views of cog-nitive development (pp. 23–66). Hillsdale, NJ:Erlbaum.

Woodward, A. (1998). Infants selectively code the goalobject of an actor’s reach. Cognition, 69, 1–34.

Zajonc, R. (1980). Feeling and thinking: Preferencesneed no inferences. American Psychologist, 35, 151–175.


C H A P T E R 3 6

Emotion Concepts

PAULA M. NIEDENTHAL

“Concepts” are mental representations of cate-gories of entities (natural and artifactual), situ-ations, experience, and action. Cognitive scien-tists study concepts because they are used inmost acts of cognition—including high-levelprocesses such as thinking, reasoning, and lan-guage use, and also lower-level processes suchas perception, attention, and recognition. Sup-porting the vast literature in cognitive psychol-ogy, the social-psychological literature reportsevidence that concepts facilitate encoding,memory retrieval processes (e.g., Cohen,1981), and the ability to make inferences aboutnever-before-seen entities (e.g., Cantor &Mischel, 1977). When, in a crowded train sta-tion, I suddenly “see” my husband carrying ourchild, I am relying on my concepts of my hus-band and child to categorize the input as theseparticular people. Other experiences with thesocial world also involve concept use. Under-standing another person’s emotions and know-ing how the emotions have come about andwhat we might do to alter or celebrate themalso involves the use of concepts. So does per-ceiving a facial display as an expression of dis-

gust or contempt. These phenomena are sup-ported by emotion concepts—the topic of in-terest in the present chapter.

Emotion concepts are not only fundamentalfor an understanding of the social world; theyare fundamental to the development of an indi-vidual’s behavioral repertoire. One of the mostcompelling examples of this (an example towhich I will return) is that of instructed fearlearning. If I tell my son that a certain object orevent would be painful or frightening to en-counter, he can, even at a relatively early pointin cognitive development, avoid that object orevent without ever having to experience thepain or fear with which it is (held to be) associ-ated. The example of instructed fear learning isnoteworthy, because it demonstrates that indi-viduals’ concepts of “fear” or “pain” are suffi-ciently powerful to determine future behavior.This observation suggests that the understand-ing of emotion concepts is vital for a full ac-count of human behavior.

It should be noted in starting that in order tostudy emotion concepts at all, theorists havehad to decide on the set of categories that they

587

are working with. What are the categories ofemotion? How can we specify the “naturalkinds” to which emotion concepts refer? Orcan we (Barrett, 2006)? Because scientists areable to point to cats and trees and furniture inthe perceivable world, and even to measuretheir physical properties, they can be in reason-able agreement about the existence of and la-bels for these and many other natural-kind andartifactual categories (within a culture). Theycan then ask in experimental research: Whatare the properties of such categories that arepreserved in individuals’ concepts; how are theproperties and related concepts structured inmemory; and how are the concepts used to un-derstand newly encountered category instances(e.g., Lamberts & Shanks, 1997)? The same isnot true of emotions. Some parts of an emotioncan be seen, such as a facial expression or aposture, but others cannot. Theorists and lay-people disagree about the categories that accu-rately cut emotional experience at its joints. Sohow do scientists proceed in the study of emo-tion concepts when there is no a priori agree-ment about the categories that constitute thedomain of interest?

There are two ways in which this decisionhas been made by scientists interested in emo-tion concepts. One is an empirical method.In this case, many things that could be “anemotion” are studied (usually represented bywords), and then the underlying conceptualstructure is empirically derived (e.g., Ortony,Clore, & Foss, 1987; Shaver, Schwartz, Kirson,& O’Connor, 1987). The most fully developedaccount of emotion concepts that relies on thismethod is the dimensional approach, in which,as we shall see, the underlying structure is de-rived by the application of various scalingmethods to judgments about emotional statesand emotional objects. Other approaches, suchas semantic-primitives and the semantic-network accounts, rely on evidence in favor ofthe existence of certain irreducible, perhaps bi-ologically basic emotions (evidence concernedwith possible structures for the set of emotionconcepts is reviewed in Barrett, 2006, andNiedenthal, Krauth-Gruber, & Ric, 2006, aswell as in several chapters of the present hand-book). These approaches then try to define theconceptual content for what people appear toexplicitly “know” about a set of categoriesso chosen, rather than what they apparentlyknow, as revealed by scaling studies.

In their explicit knowledge about emotion,

individuals seem to know about at least threeclasses of information. First, people’s conceptscontain information about the situational ante-cedents or elicitors of emotions. Individualsknow that seeing an oncoming car lose controland head directly at one’s own car most oftenelicits fear (though it can also elicit exhilara-tion, for example). Second, such concepts con-tain information about the actions that arelikely to be taken when a given emotion is ex-perienced. Thus people know that fear in hu-mans is often associated with an avoidance or aflight response from the situation. Third, con-cepts contain information about the introspec-tive states that constitute the “hot” componentof emotions. So people know, for instance, thatfear is associated with very negative and highlyaroused feelings, with a high heart rate, andoften with sweating and trembling (Rimé,Philippot, & Cisamolo, 1990). A completemodel of emotion concepts must be able to ac-count for the representation of all three typesof knowledge (most desirably in a parsimoni-ous way), as well as for the differences betweenwhat people say they know about emotion andwhat is revealed by objective measures of partsof emotional experience (Philippot, 1991).

To preview what follows, the first part ofthis chapter reviews ways in which emotionconcepts have been characterized in the litera-ture. These approaches include dimensional,semantic-primitive, prototype, and semantic-network models. After briefly describing theseaccounts, I discuss some of the assump-tions about representation upon which they aremore or less explicitly based. By and large,these accounts view the conceptual system asencapsulated from perceptual (input) and mo-tor (output) systems, and thus as a disembod-ied system of amodal, abstract symbols. An al-ternative account, an “embodied-simulation”account, is then presented. Evidence in favor ofthe idea that concepts in general and emotionconcepts in particular are embodied (and whatthat means) is discussed in greater detail.

EMOTION CONCEPTS ASCOMPOSED OF IRREDUCIBLEDIMENSIONS OF MEANING

As mentioned, one way to determine how emo-tions are conceptualized is to find the basic di-mensions underlying the ways in which indi-viduals make judgments about different aspects


of emotions, including the co-occurrences ofemotions and the perceived similarity in theirvarious components. This approach does notdeny that emotion concepts can be morenuanced and contain further information (as inthe prototype account discussed later). But theconcern is to uncover the fundamental or irre-ducible features of emotion concepts. Such di-mensions have been referred to as the “primarydimensions of meaning,” and they were origi-nally thought to include “evaluation,” “activ-ity,” and “potency” (Osgood & Suci, 1955).

More recently, researchers taking this ap-proach have endorsed a two-dimensional ac-count of emotion knowledge (e.g., Barrett &Russell, 1999; Larsen & Diener, 1992; Mayer& Gaschke, 1988; Russell & Barrett, 1999;Watson & Tellegen, 1985). The two dimen-sions, with some differences in how they arebelieved to be related to each other, correspondto the degree to which a state is “pleasant” ver-sus “unpleasant” and the degree to which astate is experienced as “activated” versus “de-activated.” The evidence in favor of a two-dimensional structure of emotion concepts isinterpreted as meaning that states that are la-beled as “fear” and “anger” are understood interms of the degree of pleasure and activationthat typically characterize them. For example,“anger” is conceptualized as highly unpleasantand moderately activated. And many states of“fear” are conceptualized as moderately un-pleasant and highly activated (see, e.g., Russell& Barrett, 1999, and Russell & Mehrabian,1977, for specific empirical demonstrations ofthese descriptions).

Although relevant research has repeatedlyrevealed a two-dimensional structure, themeaning of the dimensions and the relation-ships between them have not been interpretedin precisely the same way by different theorists.Figure 36.1 depicts the ways that the two di-mensions have been interpreted in three theo-ries. As the figure illustrates, although theterms that have been employed differ some-what, Russell (1980; Barrett & Russell, 1999)and Larsen and Diener (1992) both considerthe two underlying dimensions of emotion tobe something akin to “pleasantness” and “acti-vation.” In contrast, although Watson andTellegen (1985) also find a two-dimensionalstructure, these researchers propose that a rota-tion of the axis of the observed factors by 45degrees constitutes the best characterization ofit. Specifically, while two dimensions—“pleas-

antness” versus “unpleasantness” and whatthey call “engagement” versus “disengage-ment” (which can be considered a reinterpre-tation of the activation dimension)—emergefrom their data, they hold that the dimensionsof theoretical interest lie 45 degrees betweenthose axes, and should be labeled “negative ac-tivation” (high to low) and “positive activa-tion” (high to low) (Watson, Wiese, Vaidya, &Tellegen, 1999). In the original Watson andTellegen paper, these dimensions were called“positive affect” and “negative affect.” In thisview, fundamental emotion concepts containthe notion of being engaged in an experiencewhile feeling unpleasant and highly activated(“negative activation”) and the notion of beingengaged in an experience and feeling pleasantand highly activated (“positive activation”).

In support of their interpretation, Tellegen,Watson, and their colleagues note that most ofthe terms that individuals use in daily life to la-bel their emotions seem to cluster in the partsof the dimensional space corresponding to the45-degree rotation, so that characterizing emo-tion structure with a focus on these parts of thespace is important. Furthermore, they are mostinterested in high negative activation and highpositive activation, in part because they do notbelieve that words such as “sleepy” that anchorthe low-activation ends of the dimensions nec-essarily refer to actual emotional states.

The type of methodological and analyticstrategy applied to judgments of and percep-tions of the subjective experience of emotionand emotional objects tells us something aboutthe concepts that underlie individuals’ percep-tion of the subjective experience of emotion. Itdoes not reveal or account for the antecedentand situational knowledge about emotion thatindividuals apparently possess. The next twoapproaches, the semantic-primitives and theprototype approaches, try to readdress theseshortcomings.

DEFINING EMOTION CONCEPTSWITH SEMANTIC PRIMITIVES

There is debate about the meaning of the struc-ture revealed by multidimensional and factoranalyses of individuals’ perceptions of the rela-tions between emotional states. According tosome theorists, the resulting dimensional struc-tures reveal the dimensions that are most im-portant for building emotion concepts, but do

36. Emotion Concepts 589


FIGURE 36.1. Three descriptive models of experienced affect. Dimensions of theoretical interest aregiven in capital letters. From Russell and Barrett (1999). Copyright 1999 by the American PsychologicalAssociation. Reprinted by permission.

not solve the problem of characterizing thecontent of emotion concepts. Wierzbicka(1992), for instance, suggests that the concept“pleasant,” while perhaps more inclusive andeven basic than the concept “happy,” is not inany way better defined; it is probably as com-plex an abstract concept as “happy.” There-fore, while “pleasantness” may be a more fun-damental component of experience, this onlypushes the need to define emotion concepts to adifferent level. Now we have to ask what con-stitutes the concept “pleasant.”

One possibility is that emotions, perhapssome set of biologically basic ones, possessclassical definitions. Classical theories of con-cepts call for sets of necessary and sufficientfeatures that characterize all members of aclass. A number of arguments can and havebeen leveled against the classical view as use-ful to define emotion concepts (Russell,1991). Because concepts appear to varyacross individuals and even within individualsacross situations (Barsalou, 2005), it seemsthat the possibility of deriving classical defini-tions for any natural kind is unlikely. An al-ternative approach to defining emotion con-cepts has been to isolate a set of “semanticprimitives” and then to examine the possibil-ity that such concepts can be constructedfrom this limited set. The construction of listsof semantic primitives is a bootstrapping,bottom-up activity that involves the genera-tion of possibilities and the attempt to defineas many concepts as possible, independent of

a specific language, and without addingmore. Thus, according to proponents of thisapproach, while the words “anger” and “sad-ness” are culture-bound and language-specific, semantic primitives such as “good”and “bad,” and “want” and “happen,” arenot. Wierzbicka (1992) proposed the primi-tives listed in Table 36.1.

These primitives can describe some of thebasic themes that characterize emotion (John-son-Laird & Oatley, 1989). For example, it hasbeen noted that emotions involve good and badthings that happen to ourselves and other peo-ple, and that we ourselves and other people do.They also involve other people’s and our ownevaluations of ourselves and our actions, andthe relationships that can be constructed on thebases of these evaluations and reactions. Whensemantic primitives are used to build thesethemes, they seem to provide enough nuancesto characterize many different emotions. Forexample, the emotions “happiness” and “fear”can be defined as follows:

HappinessX feels happy.X feels something.Sometimes a person thinks something like this:

Something good happened to me.I wanted this.I don’t want other things.

Because of this, this person feels somethinggood.

X feels like this.


TABLE 36.1. A List of Semantic Primitives Proposed by Wierzbicka (1992)

Category Item

Substantives I, you, someone, something, people

Determiners and qualifiers this, the same, other, one, two, many/much, all

Mental predicates thinking (about), say, know (about), feel, want (to)

Actions and events do, happen (to)

Evaluative good, bad

Descriptive big, small

Time and place when, where, after/before, under/above

Metapredicates no/negation, because, if/would, can/may

Intensifier very

Taxonomy/partonomy kind of, part of

Hedge/prototype like

Note. From Wierzbicka (1992). Copyright 1992 by the Cognitive Science Society, Inc. Reprintedby permission.

FearX feels frightened.Sometimes a person thinks something like this:

Something bad can happen.I don’t want this.Because of this, I would want to do some-

thing.I don’t know what I can do.

Because of this, this person feels somethingbad.

X feels like this.

Despite its appeal, the semantic-primitivesapproach also has some shortcomings. Al-though the definitions seem to contain infor-mation about the antecedents of and situationsfor emotions, the “hot” or bodily aspect of theemotion, except for its being good or bad, isnot contained in the definition. Presumably thebodily experience can be derived from locatinga good or bad feeling in the specific semanticcontext, but it is not clear just how. Some ofthese problems could be handled by the prop-osition made by Johnson-Laird and Oatley(1989) that a set of what might be called basicemotions—fear, anger, happiness, sadness, anddisgust—are themselves semantic primitives.Then all of the information is consolidated intoone symbol for a complex state involving per-ception, interoception, and action. However,neither use of the semantic-primitives approachaddresses the way in which semantic primitivesare represented and processed. Although theassumption must be that the primitives are in-nate, it is still not clear what is being used whenthey are activated. By default to what has beencalled “first-generation” representational mod-els in cognitive science (Gallese & Lakoff,2005), the primitives may be abstract symbolsbuilt into the system.

EMOTION CONCEPTSAS PROTOTYPES

As with the search for semantic primitives, theproposal that emotion concepts are defined interms of probabilistic features was motivatedby opposition to classical theories of concepts.Rosch (1973) was an early proponent of twoimportant features of conceptual structure.One is that concepts are organized hierar-chically—that is, people know about the fea-tures of abstract categories, such as “furni-ture”; their more specific exemplars, such as

“chair”; and even more specific categories,such as “kitchen chair”—and the hierarchiespossess certain structural properties, such asnested features and graded structure. Anotherimportant idea was that at any given level ofabstraction, the represented category (the con-cept) is a fuzzy one defined by a set of probablefeatures, but not necessary and sufficient ones,that overlap with closely related categories.Thus the boundaries between concepts repre-senting related categories are not strict or im-penetrable.

Several researchers have tested the applica-bility of Rosch’s theory of conceptual structurein the domain of emotion knowledge, in orderto learn more about how individuals use theirknowledge of categories of emotion (e.g., Fehr& Russell, 1984; Shaver et al., 1987). For in-stance, in Shaver et al.’s (1987) work, experi-mental participants were supplied with 135cards, each containing the English name of oneemotion or affective state. The participantssorted the cards into piles that represented, forthem, groups of words whose meanings wenttogether or were similar. The card sorts wereanalyzed with another statistical technique—hierarchical cluster analysis, which identifiesclusters of variables (emotion words, in thiscase) and provides information about their hi-erarchical relations.

Consistent with the work of Rosch, the find-ings revealed three levels in the structure. Themost abstract contained the categories of “neg-ative” and “positive” emotions. The next levelcontained what appeared to be five or six basiccategories. Shaver and colleagues labeled these“love,” “joy,” “anger,” “sadness,” and “fear.”Although they also found a possible “surprise”category, they were not in favor of allocating itthe status of a basic-emotion category. Finally,there were many subordinate categories thatreflected fine gradations of the five or six basiccategories. For example, the category “fear”could be further broken down into somethinglike “horror/panic” and “nervousness/dread.”The authors noted that the five basic categoriesrevealed by the hierarchical cluster analysis arethe same as those shown by Bretherton andBeeghly (1982) in their study of emotion termslearned in early childhood. In addition, thesefive basic categories correspond to the emo-tions most often proposed to be biologicallybasic ones in the various lists of basic emotions(e.g., Ekman, 1984). Additional analyses oflanguage report similar findings (e.g., Conway


& Bekerian, 1987; Johnson-Laird & Oatley,1989; Russell & Bullock, 1986).

Analyses of the explicit content of emotionconcepts, as revealed by property generationtasks, have also provided support for Rosch’sprototype theory of natural concepts (e.g., Fehr& Russell, 1984; Keltner & Haidt, 2003;Shaver et al., 1987). Knowledge about thebasic-emotion categories appears to containthe three types of features described previously(i.e., knowledge of antecedents, situations, andthe bodily characteristics of the emotion). Rus-sell (1991) further characterizes the prototypesas scripts (e.g., Abelson, 1981), arguing thatprototypes are to objects what scripts are toevents. For Russell, emotions are events havinga causal and temporal structure, and not ob-jects, and so the notion of a temporally struc-tured script best captures the representation ofan emotion (see Table 36.2 for a possible scriptfor anger, based on Lakoff’s [1987] analysis).Consistent with prototype theory, moreover,the script contains features that are probabilis-tic and not “all or none” in nature. Russell(1991) notes that “the features that constituteemotion concepts describe the subevents thatmake up the emotion: causes, beliefs, feelings,physiological changes, desires, overt actions,and vocal and facial expressions. . . . To knowthe sense of a term like anger, fear, or jealousyis to know the script for that emotion” (p. 39).

The prototype account of emotion conceptscan show and has shown how the three types ofinformation are contained and fit together inan emotion concept. The notion of a prototype,or a script, largely makes claims about how

something emotional (a facial expression, asubjective experience, a situation, or a behav-ior) is classified as constituting an instance of aparticular emotion. Or, as Clore and Ortony(1991) have noted, “Prototypes seem, there-fore, not to be concerned with the function of‘defining the concept’ but with the function ofidentifying instances” (p. 50). The semantic-network models, discussed last, represent per-haps the only approach to modeling conceptsthat is not agnostic to the representation andprocessing of emotional information.

SEMANTIC-NETWORK MODELSOF EMOTION CONCEPTS

The single explicit class of representationalmodels of emotion concepts in the literature todate consists of the semantic-network modelsof emotion (Bower, 1981, 1991; Ingram, 1984;Lang, 1984; Teasdale, 1983). The variations onthese models hold that knowledge is repre-sented in a semantic network of units of repre-sentation sometimes called “nodes,” or al-ternatively “concepts,” “categories,” “traces,”“processors,” or “units” (Collins & Loftus,1975; Anderson & Bower, 1973). Nodes storeand transform information in propositionalform. They are linked by connecting pathwaysthat reflect the strength of semantic associa-tions among them. A particular idea comes tomind, or enters consciousness, when its node isactivated above some critical threshold. A nodecan be excited by the spread of activation fromneighboring nodes, or through direct sensorystimulation.

The semantic-network models of emotion allpropose that emotions impose a fundamentalorganizational structure on information storedin the semantic network. Each emotion or af-fective state is conceptualized as a central, or-ganizing node. Nodes that represent beliefs,antecedents, and physiological patterns associ-ated with given emotions are linked to thenodes corresponding to those nodes in mem-ory. Of course, one has to ask this question:Which emotions impose a fundamental organi-zation? Bower (1981) proposed that the set ofso-called “basic emotions” imposed such orga-nization, although his later writings suggestedthat the network was organized accordingto valence (e.g., Bower, 1991). Research andtheoretical considerations (see Niedenthal,Setterlund, & Jones, 1994, for a discussion)


TABLE 36.2. An Anger Script

Step Subevent

1 The person is offended. The offense isintentional and harmful. The person isinnocent. An injustice has been done.

2 The person glares and scowls at the offender

3 The person feels internal tension andagitation, as if heat and pressure wererapidly mounting inside. He feels his heartpounding and his muscles tightening.

4 The person desires retribution.

5 The person loses control and strikes out,harming the offender.

Note. From Russell (1991). Copyright 1991 by the Ameri-can Psychological Association. Reprinted by permission.

provide strong arguments in favor of the 1981model.

A categorical or discrete-emotions version ofthe model makes the straightforward predic-tion that when an emotion unit (e.g., the unitthat represents “sadness”) is activated abovesome threshold, activation spreads through-out the network to associated information.Autonomic reactions, expressive behaviors,emotion-related events, and personal memoriesare thereby excited and may enter conscious-ness. For instance, when one is feeling happy,the material in memory related to happinessbecomes activated. As a consequence, one mayexperience an increase in heart rate and inblood pressure, an activation of the zygo-maticus major muscle, and a heightened acces-sibility to the words and memories associatedwith happiness. In some versions of this model,the nodes that represent “opposite” states (e.g.,perhaps happiness and sadness) are connectedby inhibitory links, such that the activation ofone emotion node leads to the inhibition of theother one (Bower, 1981). For instance, activat-ing happiness is expected to inhibit the activa-tion of sadness.

This model has been applied to account for“emotion congruence” and to generate otherpredictions about emotion and informationprocessing. The emotion congruence hypothe-sis states that the processing of emotional in-formation that has an emotional quality con-gruent with the quality of the emotional stateexperienced by the individual/perceiver will bemore efficient than the processing the process-ing of information that has an incongruentemotional quality. For example, when appliedto perception, the emotion congruence hy-pothesis states that objects and events that havethe same emotion significance as the individ-ual’s current affective state are perceived bythat individual with greater efficiency thanother stimuli, such as neutral or emotion-incongruent stimuli. Findings supportive of thishypothesis have been reported (e.g., Nieden-thal, Halberstadt, & Setterlund, 1997; Nieden-thal & Setterlund, 1994; see Niedenthal et al.,1994, for a discussion). Other demonstrationsof an emotion congruence effect as predictedby a semantic-network model have been re-ported for retrieval from long-term memory(e.g., Bower, Gilligan, & Monteiro, 1981; Eich,Macaulay, & Ryan, 1994; Ehrlichman &Halpern, 1988; Fiedler & Stroehm, 1986;

Mayer, McCormick, & Strong, 1995) andjudgment (e.g., Abele & Petzold, 1994;DeSteno, Petty, Wegener, & Rucker, 2000;Forgas, 1992, 1993, 1995; Lerner & Keltner,2001).

The semantic-network models are powerfulfor accounting for and generating some hy-potheses regarding the structure and content ofemotion concepts (see Niedenthal et al., 1994,for a discussion). However, in these models,each node is connected to many other nodes,and each node is itself defined entirely in termsof its relation to the other nodes. The problem,then, is that tracing out relations between un-defined nodes does not result in meaning.Therefore, as in the other models describedthus far, it is not clear what grounds emotionconcepts. Some underlying assumptions of theapproaches described thus far are consideredexplicitly in the next section.

FIRST-GENERATION MODELSOF EMOTION REPRESENTATION

Although their purposes were not always totest this assumption, or even to be explicitabout it, all the accounts of emotion conceptsdescribed thus far represent concepts asredescriptions of the input from the sensorysystem into an abstract language (Barsalou,1999; Gallese & Lakoff, 2005). Thus extantand explicit representational models of emo-tion are based on a general view of cognition inwhich higher-order mental content is repre-sented in an abstract, language-like code (e.g.,Fodor, 1975), and the symbols used in higher-level cognitive processes are “amodal.” Anamodal representation does not preserve any-thing analogical about the perceptual experi-ence of the object, event, or state, but is ab-stracted and abstract in format (e.g., Ortony,Clore, & Collins, 1988).

Whether the resulting representation takes aform something like a word, a feature list, orvectors on which different values can be posi-tioned, the assumption is that the representa-tion and the initial perception do not take placein the same system. The dominant approach torepresenting emotion knowledge thus rests onthe “transduction” principle (Bower, 1981;Johnson-Laird & Oatley, 1989; Ortony et al.,1987). According to this principle, knowledgestructures are taken from emotional experience


and then redescribed to represent emotion con-cepts. Furthermore, representing knowledge ofan emotion in the absence of experiencing it in-volves activating the appropriate amodal repre-sentation (e.g., in the case of the semantic-network models, an emotion node). Because itdescribes different parts of the events and expe-riences relevant to the emotion, when activatedthe knowledge structure can support inferencesabout it.

In evaluating the strength of such models, we(Niedenthal et al., 1994) noted:

[Although] some emotion theorists view proposi-tional codes as sufficient for representing emo-tional stimuli, meaning, and responses[,] it is pos-sible that other types of code preserve the visual,motor, and somatovisceral aspects of such experi-ence. That we recognize a subjective feeling as“what it is like to be in love” is neither cold nortrivial. Rather, this fact means that there exists amemory of the bodily feeling of an emotion thathas been associated with a verbal label. Thus, thegeneral idea that emotions are stored in and orga-nize memory in an associative way does not haveto do away with the “hot” aspects of emotion; anexclusive focus on propositional representationmay do so. (p. 106)

EMBODIED-SIMULATION THEORY

The social psychology and emotion literaturesare filled with evidence (see a review inNiedenthal, Barsalou, Ric, & Krauth-Gruber,2005), now supported also by findings fromneuroimaging studies, that there is a differentway to model emotion concepts—one that fol-lows from the concerns cited in the quotationabove (Barsalou, 1999; Gallese & Lakoff,2005). In theories of “embodied cognition,”the modality-specific states that represent per-ception, action, and introception when one isactually in interaction with a particular entity,or having a particular subjective experience,represent these same stimuli and events whenthe original entity or experience is not actuallypresent. Put otherwise, in this view, usingknowledge involves simulations that are re-activations in the sensorimotor system. For ex-ample, retrieving the memory of a landscapeinvolves reactivating parts of the visual statesthat were active while the person was originallyperceiving it. In the same manner, thinkingabout the movements involved in riding a bi-

cycle involves partially activating the motorstates that support the activity.

What having a concept is, then, is having theability to reenact being with an instance of acategory, or having the ability to simulate it.Concepts in this approach are therefore alsocalled “simulators” (Barsalou, 2003) or “em-bodied simulations” (Gallese & Lakoff, 2005).According to Barsalou’s account, a simulatordevelops for any object, event, or aspect of ex-perience that has been repeatedly attended to.Due to its exquisite flexibility, attention can beallocated to different parts of our overall expe-rience. Across development, a number of simu-lators are established in long-term memory torepresent these different experiences. After asimulator is established, it can be used to reen-act aspects of experience, thus supporting thecapacity to perform conceptual tasks.

Representing Emotion Concepts Modally

Extending the embodied-simulation accountto emotion knowledge holds that modality-specific states represent the content of conceptsof emotion. In considering the three domainsof emotion knowledge mentioned earlier—antecedent situations, actions or action tenden-cies, and introceptive states—the embodied-simulation account says that reenactments ofmodality-specific states, rather than amodalsymbols, represent the conceptual content inthese domains. So reenactments of visually per-ceiving smiles on other people’s faces belong tothe situational knowledge that triggers “happi-ness,” as do the motor and somatosensoryexperiences of smiling oneself. Similarly,reenactments of valence and arousal states rep-resent these introspective aspects of emotionconcepts, rather than amodal symbols thatstand in for them.

In such a view, then, knowledge of an emo-tion concept is not a separate description of therespective emotion. Instead, knowledge of theemotion is grounded in actual emotional states,some conscious and some unconscious. Al-though these states may not constitute full-blown emotions, they may usually containenough information about the original states tofunction as representations of them conceptu-ally. Importantly, these partial reenactmentsconstitute the core knowledge of emotionalconcepts. Embodied states are not merely pe-ripheral events that trigger emotion concepts,


or that result from the activation of such con-cepts. (This latter description does character-ize accurately how a semantic-network modelwould link conceptual knowledge and bodilymanifestations of emotion.) Instead, embodiedstates represent the core conceptual content ofan emotion.

Empirical Supportfor Embodied-Emotion Concepts

Evidence in favor of simulation in the concep-tual processing of emotion was recentlydemonstrated in two studies (Mondillon,Niedenthal, Vermeulen, & Winkielman, 2007).In the first of those studies, individuals had tomake judgments about whether words refer-ring to concrete objects (e.g., “vomit”) were as-sociated with an emotion (they did not have tosay which emotion; they provided simply a“yes” or “no” response). The list of concepts towhich the experimental participants were ex-posed included concepts that were associatedwith joy, disgust, and anger, as well as no par-ticular emotion. While the participants wereexposed to the concepts and making their judg-ments, the activation of four facial muscles wasmeasured with electromyographic recording.Two of the muscles, the orbicularis (around theeyes) and the zygomaticus major (around themouth) muscles, are typically activated whenan individual is smiling with happiness. Thecorrugator (over the eyebrows) is typically acti-vated when an individual is frowning with an-ger. And the levator muscle is activated whenan individual makes the grimace of disgust.

According to the amodal representationalmodels, the judgment that, for example, theword “slug” is associated with disgust does notrequire the simulation of being there with aslug. That it engenders disgust is another fea-ture of slugs that is represented in a feature listby an amodal representation. It can be accessedwithout recourse to the emotion itself. On theother hand, the embodied-simulation modelpredicts that the judgment is based on a simula-tion of being there with a slug. Consequently,the amodal models do not predict that judg-ments about whether an object is associatedwith an emotion are accompanied by the spe-cific emotional experience (a simulation, whichcan be detected by activation of facial muscles),whereas the embodied-simulation model doespredict this.

The results of the study just described, as wellas a second study in which the words to bejudged were abstract emotion words (e.g., “en-raged,” “delighted,” and “disgusted”), sup-ported predictions of a modal account of rep-resentation. Specifically, in both studies,judgments about words that refer to objects elic-iting joy or that are synonyms for “joy” were ac-companied by specific activation of theorbicularis and the zygomaticus major muscles;judgments about words that refer to objects elic-iting anger or that are synonyms for “anger”were accompanied by specific activation of thecorrugator muscle; and judgments about wordsthat refer to objects eliciting disgust or that aresynonyms for “disgust” were accompanied byactivation of the levator muscle. Thus the find-ings support a proposed process by which con-ceptual processing involves simulation of theconcept in sensorimotor systems.

Another type of specific evidence comesfrom an extension of research on “switchingcosts” to the area of emotion. Researchers inperception have known for a while that shiftingattention from processing in one sensory mo-dality, such as vision, to another, such as audi-tion, involves temporal processing costs (e.g.,Posner & DiGirolamo, 2000; Spence, Nicholls,& Driver, 2000). Interestingly, similar costs arealso found when participants engage in apurely conceptual task. For example, Pecher,Zeelenberg, and Barsalou (2003) found thatparticipants were slower in verifying propertiesof a concept from one modality after they hadjust verified a concept from another modality(e.g., “bomb–loud” followed by “lemon–tart”), once again suggesting involvement ofperceptual processes in conceptual representa-tion (see also Kan, Barsalou, Solomon, Minor,& Thompson-Schill, 2003, for neuroimagingevidence).

We (Vermeulen, Niedenthal, & Luminet,2007) examined switching costs in verifyingproperties of positive and negative conceptssuch as “triumph” and “victim.” Properties ofthese concepts were taken from vision, audi-tion, and the affective system. Parallel to theswitching costs observed for neutral concepts,the study showed that for positive and negativeconcepts, verifying properties from differentmodalities produced costs, such that reactiontimes were longer and error rates were higherthan if no modality switching was required.Importantly, this effect was observed when par-


ticipants had to switch from the affective sys-tem to sensory modalities, and vice versa. Inother words, verifying that a “victim” can be“stricken” was slower and less accurate if theprevious trial involved verifying that a “spider”can be “black” than if the previous trial in-volved verifying that an “orphan” can be“hopeless.” And verifying that a “spider” canbe “black” was less efficient when preceded bythe judgment that an “orphan” can be “hope-less” than that a “wound” can be “open.” Thisresearch provides evidence that affective prop-erties of concepts are simulated in the emo-tional system when the properties are the sub-ject of active thought.

Recently, we (Niedenthal, Barsalou, Ric, &Krauth-Gruber, 2005) reviewed many addi-tional findings from the emotion literature tofurther reinforce this view of how emotion con-cepts are grounded. The additional findings in-dicate that individuals embody other people’semotional gestures and postures; that embod-ied emotional gestures and postures can pro-duce corresponding emotional states in an indi-vidual; that imagining other people and eventsalso produces embodied emotions and corre-sponding feelings; and that embodied emotionsmediate cognitive responses. Taken together,then, the logical and experimental data in favorof this modal account of emotion concepts arevery strong and motivate many importantquestions for the field.

EMOTION CATEGORIESAND CONCEPTUAL CONTENT

If knowledge acquisition occurs through em-bodiment (Niedenthal, Barsalou, Winkielman,Krauth-Gruber, & Ric, 2005), then this ac-count suggests ways to address two importantconcerns that have been raised in this chapter.One concerns what emotion concepts reallycorrespond to. In the embodied-simulation ac-count of emotion, concepts are simulationsthat are used online for the purposes of per-forming conceptual tasks. So, as we have seen,in order to know that a “slug” is associatedwith an emotion, we simulate the sight of a slugand our affective response (if any) to it. Or, ifwe need to list (for Shaver et al. or for Russell)the typical features of a state of anger, we cansimulate it and describe that (re)experience inwords. This means that there is little difference

between what we know about emotion and theprocess of having an emotion. The situated na-ture of knowledge about emotion, moreover,makes the link between a concept and a specific“natural-kind” (or other) category a mootpoint (Barrett, 2006).

Thus the account can address a secondconcern—namely, individual and cultural dif-ferences in emotion concepts. As we havelearned, acquiring emotion knowledge is inpart determined by the allocation of selectiveattention to parts of experience or incoming in-formation (such as facial expressions or otheremotional gestures). That is, even if many pro-cesses operate automatically when an emotionis evoked and experienced, this does not meanthat a residue of all such processes is present ina representation in long-term memory of theantecedent events or of the experience of theemotion. Over different experiences with anemotional state, selective attention can be allo-cated to different aspects of the embodied emo-tion (because much of it is potentially availableto consciousness, including changes in heartrate, breathing, and muscular tension), and thissupports nuanced individual and cultural dif-ferences in the content of emotion concepts.

Strategies for characterizing these differencesare already established. Neuroimaging studiesof the brain subsystems that support condi-tioned learning (e.g., conditioned fear learn-ing), observational learning, and instructedlearning can help us better understand the dif-ferences in the role of experience versus con-cepts in knowledge about emotional events.Suggestive evidence shows that there are im-portant similarities, and fewer but also impor-tant differences, in neural activation duringthese three types of fear learning (e.g., Phelps etal., 2001). Future studies will be able to linkthe role of attention to specific features of theinitial experienced situation that are simulatedin concept use.

CONCLUSION

In this chapter, I have reviewed some of thedominant models of emotion concepts. I havenoted that the models are not competing witheach other, as they actually attempt to do quitedifferent things. Dimensional analyses try todefine the irreducible structure and content foremotion concepts. A semantic-primitives ap-


proach has a similar aim for defining the con-tent, although not the structure, of emotionconcepts. A prototype approach says quite a bitabout the content of emotion concepts andhow it is used to identity instances. And onlythe semantic-network models address represen-tation and process. Recent findings in the cog-nitive, social, and emotion psychology litera-tures indicate that there is a closer relationshipbetween sensorimotor experience of and withentities in the world and the knowledge wepossess about them than the semantic-networkmodels would suggest. In the present chapter,therefore, I have tried to summarize the princi-ples of and the utility of the more recent theorythat the conceptual content for emotion knowl-edge is grounded in the sensorimotor states oc-curring in interaction with emotional stimuliand in the experience of emotional states. Thismodel will not suffice for accounting for allcognitive phenomena that we observe, and I donot make that argument here (see Solomon &Barsalou, 2004). I believe, however, that thistheory should be viewed as providing a prioriaccounts of embodied phenomena that tradi-tionally have been difficult to explain. I believethat an embodied-simulation account of emo-tion concepts can provide emotion psycholo-gists with powerful new ways of theorizingabout representations and the mechanisms thatprocess emotions.

REFERENCES

Abele, A., & Petzold, P. (1994). How does mood oper-ate in an impression formation task?: An informationintegration approach. European Journal of SocialPsychology, 24, 173–184.

Abelson, R. P. (1981). Psychological status of thescript concept. American Psychologist, 36, 715–729.

Anderson, J. R., & Bower, G. H. (1973). Human asso-ciative memory. Washington, DC: Winston.

Barrett, L. F. (2006). Emotions as natural kinds? Per-spectives on Psychological Science, 1, 28–58.

Barrett, L. F., & Russell, J. A. (1999). Structure of cur-rent affect. Current Directions in Psychological Sci-ence, 8, 10–14.

Barsalou, L. W. (1999). Perceptual symbol system.Behavioral and Brain Sciences, 22, 577–660.

Barsalou, L. W. (2003). Situated simulation in the hu-man conceptual system. Language and CognitiveProcesses, 18, 513–562.

Barsalou, L. W. (2005). Situated conceptualization. InH. Cohen & C. Lefebvre (Eds.), Handbook of cate-

gorization in cognitive science (pp. 619–650). Am-sterdam: Elsevier.

Bower, G. H. (1981). Mood and memory. AmericanPsychologist, 36, 129–148.

Bower, G. H. (1991). Mood congruity in social judg-ments. In J. P. Forgas (Ed.), Emotion and social judg-ments (pp. 31–53). Oxford: Pergamon Press.

Bower, G. H., Gilligan, S. G., & Monteiro, K. P. (1981).Selectivity of learning caused by affective states. Jour-nal of Experimental Psychology: General, 110, 451–473.

Bretherton, I., & Beeghly, M. (1982). Talking about in-ternal states: The acquisition of an explicit theory ofmind. Developmental Psychology, 18, 906–912.

Cantor, N., & Mischel, W. (1977) Traits as prototypes:Effects on recognition memory. Journal of Personal-ity and Social Psychology, 35, 38–48.


Cohen, C. E. (1981). Person categories and social per-ception: Testing some boundaries of the processingeffects of prior knowledge. Journal of Personalityand Social Psychology, 40, 441–452.

Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychologi-cal Review, 82, 407–428.

Conway, M. A., & Bekerian, D. A. (1987). Situationalknowledge and emotions. Cognition and Emotion, 1,145–191.

DeSteno, D., Petty, R. E., Wegener, D. T., & Rucker, D.D. (2000). Beyond valence in the perception of likeli-hood: The role of emotion specificity. Journal of Per-sonality and Social Psychology, 78, 397–416.

Ehrlichman, H., & Halpern, J. N. (1988). Affect andmemory: Effects of pleasant and unpleasant odorson retrieval of happy and unhappy memories. Jour-nal of Personality and Social Psychology, 55, 769–779.

Eich, E., Macaulay, D., & Ryan, L. (1994). Mood-dependent memory for events of the personal past.Journal of Experimental Psychology: General, 123,201–215.

Ekman, P. (1984). Expression and the nature of emo-tion. In P. Ekman & K. Scherer (Eds.), Approaches toemotion (pp. 319–343). Hillsdale, NJ: Erlbaum.

Fehr, B., & Russell, J. A. (1984). Concept of emotionviewed from a prototype perspective. Journal of Ex-perimental Psychology: General, 113, 464–486.

Fiedler, K., & Stroehm, W. (1986). What kind of moodinfluences what kind of memory? Memory and Cog-nition, 14, 181–188.

Fodor, J. (1975). The language of thought. Cambridge,MA: Harvard University Press.

Forgas, J. P. (1992). On mood and peculiar people: Af-fect and person typicality in impression formation.Journal of Personality and Social Psychology, 62,863–875.

Forgas, J. P. (1993). On making sense of odd couples:


Mood effects on the perception of mismatched rela-tionships. Personality and Social Psychology Bulle-tin, 19, 59–71.

Forgas, J. P. (1995). Strange couples: Mood effects onjudgments and memory about prototypical and atyp-ical targets. Personality and Social Psychology Bulle-tin, 21, 747–765.

Gallese, V., & Lakoff, G. (2005). The brain’s concepts:The role of the sensory–motor system in reason andlanguage. Cognitive Neuropsychology, 22, 455–479.

Ingram, R. E. (1984). Toward an information process-ing analysis of depression. Cognitive Therapy andResearch, 8, 443–478.

Johnson-Laird, P. N., & Oatley, K. (1989). The meaningof emotions: Analysis of a semantic field. Cognitionand Emotion, 3, 81–123.

Kan, I. P., Barsalou, L. W., Solomon, K. O., Minor, J. K.,& Thompson-Schill, S. L. (2003). Role of mental im-agery in a property verification task: fMRI evidencefor perceptual representations of conceptual knowl-edge. Cognitive Neuropsychology, 20, 525–540.

Keltner, D., & Haidt, J. (2003). Approaching awe, amoral, spiritual, and aesthetic emotion. Cognitionand Emotion, 17, 297–314.

Lakoff, G. (1987). Women, fire, and dangerous things:What categories reveal about the mind. Chicago:University of Chicago Press.

Lamberts, K., & Shanks, D. (Eds.). (1997). Knowledge,concepts and categories. Cambridge, MA: MIT Press.

Lang, P. J. (1984). Cognition in emotion: Cognition inaction. In C. E. Izard, J. Kagan, & R. B. Zajonc(Eds.), Emotions, cognition, and behavior (pp. 192–226). New York: Cambridge University Press.

Larsen, R. J., & Diener, E. (1992). Promises and prob-lems with the circumplex model of emotion. Reviewof Personality and Social Psychology, 13, 25–59.

Lerner, J. S., & Keltner, D. (2000). Beyond valence: To-ward a model of emotion specific influences on judg-ment and choice. Cognition and Emotion, 14, 473–493.

Mayer, J. D., & Gaschke, Y. N. (1988). The experienceand meta-experience of mood. Journal of Personalityand Social Psychology, 55, 102–111.

Mayer, J. D., McCormick, L. J., & Strong, S. E. (1995).Mood-congruent memory and natural mood: Newevidence. Personality and Social Psychology Bulletin,21, 736–746.

Mondillon, L., Niedenthal, P. M., Winkielman, P., &Vermeulen, N. (2007). Embodiment of emotion con-cepts: Evidence from EMG measures. Manuscriptsubmitted for publication.

Niedenthal, P. M., Barsalou, L. W., Ric, F., & Krauth-Gruber, S. (2005). Embodiment in the acquisitionand use of emotion knowledge. In L. F. Barrett, P. M.Niedenthal, & P. Winkielman (Eds.), Emotion andconsciousness (pp. 21–50). New York: GuilfordPress.

Niedenthal, P. M., Barsalou, L. W., Winkielman, P.,Krauth-Gruber, S., & Ric, F. (2005). Embodiment in

attitudes, social perception, and emotion. Personalityand Social Psychology Review, 9, 184–211.

Niedenthal, P. M., Halberstadt, J. B., & Setterlund, M.B. (1997). Being happy and seeing “happy”: Emo-tional state mediates visual word recognition. Cogni-tion and Emotion, 11, 403–432.

Niedenthal, P. M., Krauth-Gruber, S., & Ric, F. (2006).Psychology of emotion: Interpersonal, experiential,and cognitive approaches. New York: PsychologyPress.

Niedenthal, P. M., & Setterlund, M. B. (1994). Emotioncongruence in perception. Personality and Social Psy-chology Bulletin, 20, 401–411.

Niedenthal, P. M., Setterlund, M. B., & Jones, D. E.(1994). Emotional organization of perceptual mem-ory. In P. M. Niedenthal & S. Kitayama (Eds.), Theheart’s eye: Emotional influences in perception andattention (pp. 87–113). San Diego, CA: AcademicPress.

Ortony, A., Clore, G. L., & Collins, A. (1988). The cog-nitive structure of emotions. Cambridge, UK: Cam-bridge University Press.

Ortony, A., Clore, G. L., & Foss, M. A. (1987). The ref-erential structure of the affective lexicon. CognitiveScience, 11, 341–364.

Osgood, C. E., & Suci, G. J. (1955). Factor analysis ofmeaning. Journal of Experimental Psychology, 50,25–338.

Pecher, D., Zeelenberg, R., & Barsalou, L. W. (2003).Verifying different-modality properties for conceptsproduces switching costs. Psychological Science,14(2), 119–124.

Phelps, E. A., O’Connor, K. J., Gateby, J. J., Grillon, C.,Gore, J. C., & Davis, M. (2001). Activation of theamygdala by cognitive representations of fear. Na-ture Neuroscience, 4, 437–441.

Philippot, P. (1991). Reported and actual physiologicalchanges in emotion. In A. J. W. Boelhouwer & C. H.M. Brunia (Eds.), Proceedings of the First EuropeanPhysiological Conference (p. 132). Tilburg, TheNetherlands: Tilburg Press.

Posner, M. I., & DiGirolamo, G. J. (2000). Cognitiveneuroscience: Origins and promise. PsychologicalBulletin, 126, 873–889.

Rimé, B., Philippot, P., & Cisamolo, D. (1990). Socialschemata of peripheral changes in emotion. Journalof Personality and Social Psychology, 59, 38–49.

Rosch, E. H. (1973). Natural categories. Cognitive Psy-chology, 4, 328–350.


Russell, J. A. (1991). In defense of a prototype approachto emotion concepts. Journal of Personality and So-cial Psychology, 60, 37–47.



Russell, J. A., & Bullock, M. (1986). Fuzzy conceptsand the perception of emotion in facial expressions.Social Cognition, 4, 309–341.

Russell, J. A., & Mehrabian, A. (1977). Evidence for athree-factor theory of emotions. Journal of Researchin Personality, 11, 273–294.

Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C.(1987). Emotion knowledge: Further exploration ofa prototype approach. Journal of Personality and So-cial Psychology, 52, 1061–1086.

Solomon, K. O., & Barsalou, L. W. (2004). Perceptualsimulation in property verification. Memory andCognition, 32, 244–259.

Spence, C., Nicholls, M. E., & Driver, J. (2001). Thecost of expecting events in the wrong sensory mo-dality. Perception and Psychophysics, 63(2), 330–336.

Teasdale, J. D. (1983). Negative thinking in depression:Cause, effect or reciprocal thinking? Advances in Be-haviour Research and Therapy, 5, 3–25.

Vermeulen, N., Niedenthal, P. M., & Luminet, O.(2007). Switching between sensory and affective sys-tems incurs processing costs. Cognitive Science, 31,183–192.

Watson, D., & Tellegen, A. (1985). Toward a consen-sual structure of mood. Psychological Bulletin, 98,219–235.


Wierzbicka, A. (1992). Defining emotion concepts.Cognitive Science, 16, 539–581.


C H A P T E R 3 7

Memory and Emotion

ELIZABETH A. KENSINGER and DANIEL L. SCHACTER

Although the concept of memory has existedfor thousands of years, its systematic study waslaunched in the 1880s by the seminal experi-ments of the German philosopher HermannEbbinghaus. Through careful assessments ofhis own memory, Ebbinghaus forged the wayfor the field of memory research by demon-strating that humans’ ability to retain informa-tion over time could be studied scientifically. Itis telling that Ebbinghaus’s studies involved theintentional memorization of nonsense sylla-bles: He believed that to understand memoryprocesses, one should study retention of infor-mation that is void of meaning or personalimportance. Although memory researchersseemed to embrace Ebbinghaus’s views on thisissue for nearly a century, over the past coupleof decades there has been increased emphasison examining memory for personally impor-tant experiences and for events that evoke emo-tional reactions.

Throughout this chapter, we use terms like“emotional stimuli” as shorthand to denote in-formation in the environment that elicits achange in the internal affective state of the or-ganism. The focus of this chapter is on howthese internal changes affect memory. Be-havioral examinations of explicit (conscious)memory for emotional experiences have re-vealed three broad influences of emotion onmemory: on the number (quantity) of events re-membered, on the subjective vividness (quality)of the remembered events, and on the amountof accurate detail remembered about prior ex-periences. This chapter explores these threelines of investigation, highlighting both thegeneral conclusions that have emerged fromthe research and the open questions that re-main. We conclude with a brief discussion ofrecent research suggesting an effect of emotionon implicit (unconscious) memory. In additionto presenting the behavioral data and cognitive

601

theories of emotional memory, this chapter in-cludes discussion of relevant neuroimaging andneuropsychological research that has been in-fluential in examining the extent to whichmemory for emotional experiences is sup-ported by processes distinct from those sup-porting memory for nonemotional events.

EMOTION’S INFLUENCEON THE QUANTITYOF REMEMBERED INFORMATION

Individuals often remember more emotionalevents than nonemotional ones. Within the lab-oratory, recall rates are higher for positive andnegative stimuli than for neutral stimuli(reviewed by Buchanan & Adolphs, 2002;Hamann, 2001). This finding has been docu-mented with a variety of stimuli, includingwords, sentences, pictures, and narrated slideshows (e.g., Bradley, Greenwald, Petry, &Lang, 1992; Cahill & McGaugh, 1995;Kensinger, Brierly, Medford, Growdon, &Corkin, 2002). Similar effects have been notedwithin the autobiographical memory literature.For example, when individuals are asked togenerate memories in response to cue words,the retrieved memories are often rated as per-sonally significant and emotional (e.g.,Conway, 1990; Rubin & Kozin, 1984). Thereare also many instances in which positive andnegative events are more likely to be recognizedthan neutral ones, although the recognitionmemory advantage is seen less consistentlythan the recall advantage (reviewed byChristianson, 1992)—a finding that we will re-turn to later in this chapter.

A topic of ongoing investigation is the extentto which the “valence” of an event (i.e.,whether it elicits positive or negative affect) in-fluences the likelihood that the event is remem-bered, and virtually every conceivable outcomehas been observed. Often the boost in recall orrecognition is comparable for positive andnegative stimuli (e.g., Bradley et al., 1992;Kensinger et al., 2002). However, in some stud-ies, particularly those assessing memory forverbal or pictorial stimuli presented within alaboratory setting, negative items are morelikely to be recalled than positive ones (e.g.,Charles, Mather, & Carstensen, 2003; Ortony,Turner, & Antos, 1983). Yet other studies, gen-erally those assessing memory for autobio-graphical experiences or information encoded

in reference to the self, have revealed the op-posite pattern: a greater tendency to recallpositive events than negative ones (e.g.,D’Argembeau, Comblain, & van der Linden,2005; Linton, 1975; Matt, Vasquez, & Camp-bell, 1992; White, 2002).

Some of these conflicting findings with re-gard to the effect of valence on the likelihoodof remembering information may be explainedby the proposal that memory mechanisms haveevolved to facilitate the encoding and retrievalof the affective information that is most rele-vant to one’s goals (Lazarus, 1991; LeDoux,1996). Remembering a negative experience of-ten may be relevant to survival or well-being,because reexperiencing the event will help aperson plan for (or avoid) its future recurrence(LeDoux, 1996). In these instances, more at-tention may be paid to the negative item,thereby enhancing memory for this negative in-formation. However, there are probably in-stances in which positive events are just as rele-vant, or more relevant, to one’s goals asnegative events. Indeed, when positive and neg-ative stimuli are equally related to one’s currentconcerns, they show similar capture of atten-tion (Riemann & McNally, 1995). Further-more, there is some evidence that individuals(e.g., older adults) who seek positive goal statesshow enhanced memory for positive as com-pared to negative events (reviewed by Mather& Carstensen, 2005).

Researchers have focused intensively onwhether this mnemonic benefit for personallyrelevant and emotional information resultsfrom the engagement of processes that are re-lated specifically to the processing of emotionalinformation, or whether the memory booststems from engagement of the same processesthat allow accurate remembering of neutral in-formation. Although parsimony favors the hy-pothesis that the same processes are recruitedto remember emotional and neutral informa-tion, lesion and neuroimaging studies suggestthat the amygdala (an almond-shaped region ofthe medial temporal lobe) is specifically relatedto memory for emotional, but not for neutral,information. Patients with damage to theamygdala do not show a memory boost foremotional information: They are no morelikely to remember positive or negative eventsthan they are to remember neutral ones. Theabsence of the emotional memory enhance-ment has been reported in patients with focalamygdala damage (e.g., Adolphs, Cahill, Schul,


& Babinsky, 1997; Brierley, Medford, Shaw, &David, 2004; Cahill, Babinsky, Markowitsch,& McGaugh, 1995; Markowitsch et al., 1994)and in individuals with amygdala atrophycaused by Alzheimer’s disease (e.g., Abrisqueta-Gomez, Bueno, Oliveira, & Bertolucci, 2002;Kensinger et al., 2002; Kensinger, Anderson,Growdon, & Corkin, 2004). In these latter in-dividuals, the amount of the amygdala damagecorresponds with the degree of blunted emo-tional memory enhancement (Mori et al.,1999).

Although these neuropsychological studieshave demonstrated the necessary contributionof the amygdala to emotion-mediated memoryenhancements, they do not allow investigationof the memory stage(s) during which theamygdala exerts its influence. Neuroimagingmethods provide a way to address this issue.Researchers can examine the neural processesassociated with the successful encoding of emo-tional information (by comparing brain activ-ity while participants are encoding items thatwill later be remembered vs. items that willlater be forgotten) or with the successful re-trieval of emotional information (by compar-ing brain activity associated with correct en-dorsements vs. misses). These neuroimagingstudies have demonstrated that the amygdalaplays a fundamental role during the encodingof emotional information. Individuals whoshow the greatest amygdala activity during theviewing of emotional items are those who showthe greatest emotional memory enhancement(Cahill et al., 1996). Moreover, for a particularindividual, those emotional items that elicit thegreatest amygdala activity during encoding arethose that are most likely to be remembered(reviewed by Hamann, 2001; Phelps, 2004;Kensinger, in press). Although the vast majorityof neuroimaging studies have examined theamygdala’s role in encoding negative informa-tion, a recent neuroimaging study indicatedthat the amygdala was equally active duringthe successful encoding of positive and negativehigh-arousal items (Kensinger & Schacter,2006a). This finding is consistent with researchdemonstrating that the amygdala is importantfor processing reward-related informationas well as threat-related information (seeDavidson & Irwin, 1999; Baas, Aleman, &Kahn, 2004), and that it may primarily re-spond based on the arousal, and not thevalence, of information (Anders, Lotze, Erb,Grodd, & Girbaumer, 2004; Anderson et al.,

2003; Garavan, Pendergrass, Ross, Stein, &Risinger, 2001; Hamann, Ely, Hoffman, &Kilts, 2002; Kensinger & Schacter, 2006c;Royet et al., 2000; Small et al., 2003; however,see Buchanan, Tranel, & Adolphs, 2006, forevidence that the right amygdala may be moreinvolved in memory for negatively valencedthan positively valenced information).

The amygdala, of course, does not act in iso-lation. It has been proposed that once theamygdala is activated during the processing ofemotional information, it can modulate thefunctioning of sensory cortices to assure thatthe information is attended to (reviewed byDolan & Vuilleumier, 2003), and to enhancemnemonic consolidation processes in the hip-pocampal formation to increase the likelihoodthat emotional information is retained in a sta-ble memory trace (reviewed by McGaugh,2004; Phelps, 2004). The amygdala is wellsuited for these modulatory functions, as it isone of the most extensively connected sub-cortical regions of the brain, with links to nu-merous cortical and subcortical regions(Amaral, Price, Pitkanen, & Carmichael, 1992;Amaral, 2003).

Recent neuroimaging studies have providedstrong evidence for these modulatory effects ofthe amygdala. In one study investigating thelinks between amygdala activity and visual at-tention, patients with varying amounts ofamygdala damage were scanned while theyperformed a task in which they had to attend tofearful or neutral faces (Vuilleumier, Richard-son, Armony, Driver, & Dolan, 2004). Individ-uals with intact amygdalas showed enhancedactivity in the fusiform gyrus (a visual process-ing region) when they attended to fearful facesas compared to neutral faces. Patients with ex-tensive amygdala damage did not show thispattern: They showed equivalent fusiform ac-tivity for neutral and fearful faces. Moreover,the amount of amygdala preservation corre-sponded with the amount of fusiform modula-tion based on the emotional content of the at-tended faces. These results suggest that theamygdala can modulate visual processing inhumans, increasing the likelihood that an emo-tional item in the environment is detected andattended to.

In addition to these influences on sensoryprocesses, a number of neuroimaging studieshave provided evidence for amygdalar modula-tion of mnemonic processes, suggesting that in-teractions between the amygdala and the hip-

37. Memory and Emotion 603

pocampus serve a critical role in modulatingthe memory enhancement for emotional infor-mation in humans (reviewed by McGaugh,2004). In healthy individuals, there are strongcorrelations between the amount of activity inthe amygdala and in the hippocampus duringthe encoding of emotional information (e.g.,Dolcos, LaBar, & Cabeza, 2004; Hamann, Ely,Grafton, & Kilts, 1999; Kensinger & Corkin,2004; Kensinger & Schacter, 2005a). Althoughthese correlations cannot speak to the direc-tion of modulation, a neuroimaging study ofencoding-related neural activity in patientswith varying amounts of amygdala and hippo-campal damage provided evidence for the im-portance of reciprocal connections. While inthe scanner, patients were asked to encode a se-ries of emotionally aversive and neutral words.Outside the scanner, they performed a recogni-tion task, and the encoding trials were sortedon a post hoc basis into those words that werelater remembered and those that were later for-gotten. The critical findings from the studywere that the extent of amygdala atrophy cor-related negatively with the magnitude of activ-ity in the hippocampus during the encoding ofemotional information, and that the amount ofhippocampal atrophy was also inversely re-lated to amygdala activity (Richardson,Strange, & Dolan, 2004). Thus bidirectionalconnections between the amygdala and the hip-pocampus may be important for modulatingthe encoding of emotional information (seealso Kilpatrick & Cahill, 2003).

In contrast to the extensive literature exam-ining the amygdala’s role during encoding, rela-tively few studies have considered its role dur-ing episodic retrieval. It is clear that theamygdala is engaged during retrieval of emo-tional items (e.g., Dolan, Lane, Chua, &Fletcher, 2000; Taylor et al., 1998) and is moreactive during retrieval of information learnedin emotional contexts compared to nonemo-tional ones (e.g., Maratos, Dolan, Morris,Henson, & Rugg, 2001; Smith, Henson,Dolan, & Rugg, 2004; Sterpenich et al., 2006).The strongest test for the role of the amygdalain retrieval, however, is a demonstration of aninteraction between emotional content andsuccessful retrieval—that is, a stronger relationto successful retrieval (as compared to retrievalfailures) for emotional items than for neutralitems. Some recent studies have provided evi-dence for such an interaction, underscoring the

potential importance of amygdala engagementduring the retrieval process (Dolcos, LaBar, &Cabeza, 2005; Kensinger & Schacter, 2005b;Sergerie, Lepage, & Armony, 2006).

As this section has highlighted, a tremendousamount has been learned about the effects ofemotion on memory by focusing on emotion-induced enhancements in the likelihood of re-membering information. These studies havedemonstrated that positive and negative eventsoften are more likely to be remembered thannonemotional ones are, and that the amygdalaappears to be critical for this quantitative mem-ory boost. Through its interactions with othercortical and subcortical regions, the amygdalacan modulate sensory and mnemonic func-tions, increasing the likelihood that emotionalinformation is perceived and retained in a sta-ble memory trace.

QUALITY-BASED INFLUENCES

Although quantitative assessments of memoryhave been instrumental in laying the ground-work for investigations of emotion–memoryinteractions, they may underestimate the influ-ence of emotion. Not all remembrances are cre-ated equal. Sometimes we feel transported intime as we reexperience a prior event, and ourmemory seems to include a tremendousamount of detail about where and when theevent occurred. Other times we recognize thatwe’ve seen something before, but our memorydoes not include information about the contextof the prior encounter: We recognize an indi-vidual in the airport but do not know where wemet this person, or we recognize a stretch ofroad without knowing when we have driven itbefore (reviewed by Gardiner & Java, 1993;Yonelinas, 2002).

Many of emotion’s effects on memory be-come apparent only when the quality of amemory is considered. As noted in the sectionabove, there are many instances in which indi-viduals are no more likely to correctly recog-nize an emotional item than a nonemotionalone. One potential reason for this null effect ofemotion is that successful recognition perfor-mance does not require a vivid memory: Al-though we can recognize an item because wevividly remember its prior presentation, simplyknowing that we’ve seen something before issufficient. Indeed, effects of emotion are more


likely to occur when the vividness of a memoryis considered. For example, when individualsare asked not only whether they recognize anitem but also whether they vividly “remember”the item’s prior presentation, rates of “remem-bering” tend to be much higher for emotionalpictures or words than for nonemotional ones(Dewhurst & Parry, 2000; Kensinger &Corkin, 2003; Ochsner, 2000; Sharot,Delgado, & Phelps, 2004). This boost in theability to vividly remember emotional informa-tion often occurs even when overall recognitionrates are equivalent for emotional and neutralinformation (e.g., Ochsner, 2000; Sharot et al.,2004).

In many instances, enhanced vividness de-fines an emotional memory. An extreme exam-ple is a “flashbulb memory”: As the term im-plies, individuals sometimes believe that theyhave maintained an almost photographic-quality memory of a highly emotional and con-sequential event (Brown & Kulik, 1977). Forexample, people claim to remember where theywere and what they were doing when theylearned of the assassination of President John F.Kennedy (Brown & Kulik, 1977; Christianson,1989; Winograd & Killinger, 1983); the Sep-tember 11, 2001, terrorist attacks (Budson etal., 2004; Paradis, Solomon, Florer, & Thomp-son, 2004; Pezdek, 2003; Smith, Bibi, &Sheard, 2003); or the explosions of the spaceshuttles Challenger and Columbia (Bohannon,1988; Kensinger, Krendl, & Corkin, 2006;Neisser & Harsch, 1986). Although such ex-tremely vivid memories are formed only rarely,studies of autobiographical memory have con-firmed that individuals often remember emo-tional experiences in a particularly vivid man-ner (e.g., Conway, 1990; Rubin & Kozin,1984; Schaefer & Philippot, 2005). In all ofthese examples, what is noteworthy about thememories is not that the individuals rememberthe experience, but rather that they reexperi-ence it with tremendous vividness.

In this section, we examine whether the emo-tional qualities of events can affect how vividlythey are remembered. In particular, we will dis-cuss whether the arousal (the degree of excita-tion or pacification) or valence (the degree ofdispleasure or contentment; e.g., Russell, 1980;Lang, Greenwald, Bradley, & Hamm, 1993)elicited by an event influences the vividnesswith which a person remembers those experi-ences.

Effects of Arousal on Memory’s Vividness

The vast majority of studies examining the ef-fects of emotion on memory’s vividness havefocused on stimuli that elicit high arousal. Forthese stimuli, mnemonic influences appear tooccur via interactions between the amygdalaand the hippocampus. For example, Kensingerand Corkin (2004) compared encoding-relatedactivity for words that participants later vividlyremembered and for words that participantslater forgot. For the arousing words, interac-tions between the amygdala and the hippocam-pus were found to be critical: Activity in bothof these regions corresponded with the likeli-hood that a participant would later vividly re-member a negative arousing word, and theamount of activity in the two regions was cor-related strongly. For nonarousing words, hip-pocampal activity predicted the vividness of amemory, but amygdala activity did not. Thusamygdala activity during encoding relates notonly to an increased likelihood of rememberingan emotional item (as discussed in the priorsection), but also to an increased likelihood ofremembering such an item vividly (see alsoDolcos et al., 2004, for evidence that amygdalaactivity during encoding leads to vivid memo-ries of emotionally arousing information).

These mnemonic influences appear to occurrelatively automatically. Kensinger and Corkin(2004) asked participants to study words eitherwith full attention devoted toward the encod-ing task, or with attention divided between theencoding task and a secondary (sound discrimi-nation) task. The addition of the secondarytask impaired the vividness with which thenonarousing words were remembered, whereasit did not have a large effect on the vividness ofmemories for the arousing words (see also Bush& Geer, 2001). This finding is consistent withproposals that emotional information is privyto prioritized or relatively automatic process-ing (reviewed by Dolan & Vuilleumier, 2003;Pessoa, Kastner, & Ungerleider, 2003).

Although these studies demonstrate thatemotional arousal can be a critical factor con-tributing to the emotional memory enhance-ment effect (see also Cahill & McGaugh, 1995;McGaugh, 2004), an arousal response is notrequired for emotional modulation of mem-ory’s vividness. Items that evoke changes in va-lence, but not changes in arousal, also canbe remembered with enhanced vividness


(Kensinger & Corkin, 2003; Ochsner, 2000).The processes that lead to the mnemonic en-hancements for nonarousing items, however,seem to be distinct from those mechanisms en-gaged for arousing information. In particular,the mnemonic boost for nonarousing stimuliappears to stem from controlled and elabora-tive processing of the stimuli (see Kensinger,2004, for further discussion). Thus, in contrastto the minimal effect of divided attention onparticipants’ memories for arousing items(Bush & Geer, 2001; Kensinger & Corkin,2004), divided attention has a large detrimen-tal effect on the likelihood that participantswill vividly remember negative nonarousingitems. In fact, when participants’ attention isdivided during encoding, the mnemonic en-hancement for negative nonarousing wordsdisappears (Kensinger & Corkin, 2004).

There also are distinct neural signatures asso-ciated with the successful encoding of arousingand nonarousing words (Kensinger & Corkin,2004). In contrast to arousing items, which ap-pear to be vividly remembered due to amygdalaengagement, successful encoding ofnonarousing items is associated with increasedactivity in the prefrontal cortex and hippocam-pus (and see LaBar & Phelps, 1998, for evidencethat memory for negative nonarousing stimuliwas not impaired in a patient with amygdaladamage). The prefrontal cortex and hippocam-pus support later memory for neutral items aswell, but the strength of their correspondencewith later memory is greater for the negativenonarousing items (Kensinger & Corkin, 2004).This overlap in the neural processes engaged toremember neutral and negative nonarousingwords is consistent with the conclusion that par-ticipants are remembering the negativenonarousing words because of increased en-gagement of the same types of cognitive and neu-ral processes that lead to a vivid memory fornonemotional information. This finding empha-sizes that simply showing a memory benefit foremotional stimuli does not necessitate that a dis-tinct mnemonic mechanism be postulated (andsee Talmi & Moscovitch, 2004).

Thus, although many researchers have fo-cused on amygdala-mediated effects of emo-tion on memory (thought to arise due to the ac-tion of stress hormones; McGaugh, 2000;McGaugh & Roozendaal, 2002), emotional in-formation that does not elicit amygdalar mod-ulation of memory can also be more vividly re-membered than neutral information. These

effects may arise because individuals are morelikely to elaborate on the material during en-coding or to rehearse the information (for dis-cussions of effects of elaboration and rehearsalon memory for emotional information, seeBohannon, 1988; Brown & Kulik, 1977;Christianson & Engelberg, 1999; Isen, 1999).

Effects of Valence on the Qualityof a Memory

To assess the effects of valence on memory, re-searchers have contrasted memory for positivearousing and negative arousing stimuli. Ifhighly arousing positive and negative experi-ences have different memory characteristics,then it cannot be only the arousal of the eventsthat influences memory quality. A number ofstudies have suggested that the valence elicitedby the event does influence the subjective vivid-ness of the memory. Within the laboratory, neg-ative events are often remembered with agreater sense of vividness than positive eventsare (e.g., Ochsner, 2000; Dewhurst & Parry,2000). Positive stimuli, in contrast, are oftenremembered with only a feeling of familiarity,or with general (nonspecific) information (e.g.,Ochsner, 2000; Bless & Schwarz, 1999). Thiseffect of valence on memory for detail can holdeven in individuals (e.g., older adults) who tendto focus more on positive information than onnegative, and it can exist even when overall rec-ognition rates are equated for negative andpositive information (Kensinger, O’Brien,Swanberg, Garoff-Eaton, & Schacter, in press;Kensinger, Garoff-Eaton, & Schacter, 2007).Positive mood also has been associated withmore memory reconstruction errors than nega-tive mood has, probably because individuals ina happy mood rely on gist-based informationor on heuristics, while individuals in a negativemood are more likely to focus on the specificdetails of information (e.g., Bless et al., 1996;Storbeck & Clore, 2005).

It has been unclear to what extent these lab-oratory findings extend to real-life events in-fused with emotional importance. Research onautobiographical memory has often supportedthe opposite conclusion from laboratory re-search: that positive memories are morevivid than negative ones (e.g., D’Argembeau,Comblain, & van der Linden, 2003; Schaefer& Philippot, 2005). For example, Schaefer andPhilippot (2005) asked participants to recallpositive, negative, and neutral events, and for


each one to rate the number of sensory, seman-tic, temporal, and contextual associations re-trieved about the memory (using the MemoryCharacteristics Questionnaire; Johnson, Foley,Suengas, & Raye, 1988). They found that par-ticipants’ ratings were higher for positive thanfor negative memories, indicating greater re-trieval of contextual detail for positive events.However, some studies suggest little effect ofvalence on memory vividness, and instead havefound intensity to be the primary predictor ofautobiographical memory characteristics (e.g.,Talarico, LaBar, & Rubin, 2004).

A difficulty in these studies is finding posi-tive and negative events that are comparableacross a range of dimensions (e.g., duration ofevent, public or private nature of event,amount of media coverage or rehearsal). Twoprior studies have attempted to circum-vent many of these difficulties by examiningwhether a person’s response to an event out-come (finding it positive or negative) affectswhat the person remembers about the event(Levine & Bluck, 2004; Kensinger & Schacter,2006e). Levine and Bluck (2004) asked partici-pants to indicate whether particular events hadoccurred during the verdict decision in the O. J.Simpson trial. We (Kensinger & Schacter,2006e) examined what Red Sox fans and Yan-kees fans remembered about the final game ofthe 2004 American League Championship Se-ries, in which the Red Sox overcame a surpris-ing 0–3 setback in the series to win the champi-onship. Consistent with Talarico et al. (2004),both studies found that the overall amount ofdetail remembered about the event was not in-fluenced by the event’s valence. Valence did,however, affect some memory characteristics:Levine and Bluck (2004) found that individualswho were happy about the Simpson verdict re-membered the event more vividly and weremore liberal in accepting that something hadoccurred. Similarly, we (Kensinger & Schacter,2006e) found that Red Sox fans, who foundthe outcome positive, showed more memoryinconsistencies and were more likely to beoverconfident in their memories than wereYankees fans. In both studies valence did notaffect the quantity of remembered information,but it did influence the qualitative nature of theretrieved memories. These results emphasizethe need to examine the effects of emotion notonly on the likelihood of remembering infor-mation, but also on the quality of rememberedinformation.

EFFECTS OF EMOTION ON MEMORYFOR DETAIL

The prior section emphasizes that emotion canaffect not only the likelihood of rememberingan event, but also the subjective vividness withwhich it is remembered. Individuals can saythat they vividly remember an item for a vari-ety of reasons, however. They may rememberspecific perceptual details of the event; theymay recollect specific contextual details, suchas something they thought of during the event;or they may be biased to say that they remem-ber the prior experience vividly. In this sectionwe first examine whether emotion primarily in-flates people’s confidence in their memories, orwhether it also has beneficial effects on mem-ory for detail. We then discuss the neural pro-cesses that mediate the effects of emotion onmemory for detail, as well as the types of de-tails that are most likely to be rememberedabout an emotional event.

Inflated Confidence or Enhanced Detail?

When Brown and Kulik (1977) coined the term“flashbulb memory” to refer to the vivid recol-lection of a surprising and consequential event,they believed that it reflected a separate mem-ory mechanism, immune to memory distortionor disruption. Numerous studies have nowdemonstrated that these memories are in factprone to significant distortions over time. Indi-viduals often report high confidence in so-called flashbulb memories despite low consis-tency in their reports over time, and there is of-ten little or no correlation between how confi-dent individuals are about their memories andhow accurate or consistent their memories are(Neisser & Harsch, 1992; Schmidt, 2004;Schmolck, Buffalo, & Squire, 2000; Talarico &Rubin, 2003). Clearly, emotional events do notleave indelible traces. Nevertheless, the ques-tion remains of whether enhanced confidencein memories of emotional experiences is justi-fied: Are individuals just biased to believe thatthey have retained a detailed memory of anemotional experience, or are these events re-trieved with more accurate detail than non-emotional events are?

On the one hand there is some evidence thatemotion may primarily bias people to believethat they remember a prior experience: Acrossa number of recognition paradigms, emotionalitems (and particularly negative ones) have


been more likely to be falsely recognized thannonemotional ones (Budson et al., 2006; Ehlerset al., 1988; Windmann & Kruger, 1998;Windmann & Kutas, 2001). Because the in-creased false alarms occur not only for relatedlures but also for unrelated emotional words,these results suggest that individuals may be bi-ased to believe that they have encountered anemotional item previously. One likely contribu-tor to this bias is the fact that emotional items(and negative ones in particular) tend to beprocessed more fluently than neutral items(Bargh, Chaiken, Govender, & Pratto, 1992;Kitayama, 1990; Öhman, 1988). Because itemsthat have been studied recently tend to be pro-cessed more fluently than items that have notbeen encountered recently (Whittlesea, 1993;Whittlesea & Williams, 2000), participantsmay misinterpret the fluid processing of anemotional item on a recognition task as evi-dence that they have recently studied the item.Thus enhanced fluency could lead individualsto falsely endorse more nonpresented negativeitems than nonpresented neutral items. An-other possible contributor to the bias is the se-mantic relatedness of most emotional items(see Talmi & Moscovitch, 2004). Because the-matically associated lures tend to be falselyendorsed more often than unrelated lures(Roediger, Watson, McDermott, & Gallo,2001; Stradler, Roediger, & McDermott,1999), emotional items may be falsely recog-nized more frequently because of their seman-tic cohesion.

Sharot et al. (2004) have suggested thatamygdala activity at retrieval may primarilyserve to inflate a person’s estimate of a mem-ory’s vividness and level of detail. They foundthat amygdala activity at retrieval corre-sponded with an individual’s belief that anemotional item was vividly remembered (seealso Dolcos et al., 2005), whereas activity inthe parahippocampal gyrus (associated with vi-sual memory) corresponded with vivid remem-bering of neutral information. Because the en-hancement in subjective vividness for theemotional items occurred in the absence of anemotion-related boost in the ability to discrimi-nate “old” from “new” items, Sharot et al. in-terpreted their findings as indicating thatamygdala engagement at retrieval leads indi-viduals to feel that they have a vivid memorynot because they remember perceptual detail(as is the case with neutral items), but ratherbecause of the feeling of arousal and perceptual

fluency that accompanies the remembrance ofemotional items.

On the other hand, however, there are in-stances in which emotion can increase the like-lihood that details are remembered about anitem. In a couple of studies (Kensinger, Garoff-Eaton, & Schacter, 2006; Kensinger et al.,2007), participants were shown a series of neg-ative and neutral objects (e.g., a spider, ablender) and later were asked to distinguish“same” objects (those that were identical to astudied item) from “similar” objects (those thathad the same verbal label as a studied item, butthat differed in any number of visual details).Individuals were more accurate at discriminat-ing “same” from “similar” negative items thanthey were at distinguishing “same” from “sim-ilar” neutral items. Importantly, this en-hancement in memory specificity occurred evenwhen the ability to recognize that a particularitem type had been presented (e.g., to remem-ber that a snake or a blender had been studied)was not influenced by emotion. These resultsparallel the self-report data (Ochsner, 2000;Sharot et al., 2004), with negative emotionalcontent enhancing not the ability to recognizean item from a study episode, but rather theability to vividly remember its presentation.According to our results (Kensinger, Garoff-Eaton, & Schacter, 2006), the pattern of self-report data does not necessitate that individu-als are biased to believe that they vividly re-member an emotional item. Rather, emotioncan affect the amount of detail rememberedabout a studied item, while not affecting theoverall proportion of items remembered.

Although these studies focused on memoryfor perceptual detail, emotion may have abroader effect on the ability to remember con-textual (or “source”) information. “Sourcememory” is frequently defined as any contex-tual aspect (e.g., perceptual, spatiotemporal,affective) present when an encoding eventoccurred (Johnson, Hashtroudi, & Lindsay,1993). Source memory can be contrasted with“item memory”—the ability to recall or recog-nize that an item was previously encountered,without the ability to retrieve details about itsencoding context. In a number of paradigmsemotion has been found to enhance memoryfor such details as the color in which a wordwas presented (Doerksen & Shimamura, 2001;Kensinger & Corkin, 2003; D’Argembeau &van der Linden, 2004; MacKay et al., 2004),the spatial location of a word (D’Argembeau


& van der Linden, 2004; MacKay & Ahmet-zanov, 2005), or whether words or objectswere imagined or visually presented (Kensinger& Schacter, 2006d). Contextual informationpresented in a sentence may also be more likelyto be remembered if the sentence is emotionallynegative than if it is neutral (Kensinger et al.,2002, 2004). These results indicate that eventswith emotional meaning are often more likelyto be remembered with detail than are eventsvoid of emotional importance.

Moreover, amygdala activity at retrieval cancorrespond with retrieval of accurate detail. We(Kensinger & Schacter, 2005b) asked partici-pants to view photographs of some objects, butonly to imagine viewing others. At retrieval,participants were required to indicate whichitems they had imagined viewing and whichones they had actually seen (i.e., to make areality-monitoring decision). The critical find-ing was that activity in the orbito-frontal cor-tex, amygdala, and hippocampus, was greaterfor emotional items accurately attributed topresentation or to imagination than for emo-tional items misattributed. Convergent findingswere revealed in an elegant study by Smith,Stephan, Rugg, and Dolan (2006). Participantsstudied neutral objects in either neutral or emo-tional scenes. Some of the neutral and emo-tional scenes contained people, while othersdid not. Participants were then asked to reportthe context in which objects had been studied.In one condition, the options were “emotionalcontext” or “neutral context”; in another con-dition, the options were “context with people”or “context without people.” The data re-vealed enhanced hippocampus–amygdala con-nectivity whenever individuals retrieved infor-mation studied in an emotional context,regardless of whether the task required report-ing of the emotional context (i.e., in both the“emotional” vs. “neutral” and “people” vs.“no-people” conditions). When the retrievalof the emotional information was critical tosuccessful memory performance (i.e., in the“emotional” vs. “neutral” condition), hippo-campus–amygdala connectivity increasedbidirectionally, modulated by the orbitofrontalcortex. Taken together, these two studies pro-vide strong evidence that limbic engagementduring memory retrieval does not relate only toan inflated confidence in the subjective richnessof a memory; rather, increased limbic activitycan correspond with the retrieval of details re-garding an item’s presentation.

How Emotion Influences Memoryfor Detail

Although these studies indicate that at leastsome details are more likely to be rememberedfor items with negative emotional content thanfor items with neutral content, they do not clar-ify whether this increased memory accuracy forthe negative arousing information is specifi-cally related to the processing of the emotionalinformation, or whether it stems from engage-ment of the same processes that lead to accu-rate memory for neutral information. The in-creased memory accuracy for negative arousinginformation could result from domain-generalfactors, such as the greater semantic related-ness (Talmi & Moscovitch, 2004) or thegreater distinctiveness of emotional items(Schmidt, 2002); or the enhanced accuracycould stem from processes specific to emo-tional processing.

To adjudicate between these alternatives, weconducted a series of studies to examine the ef-fect of emotion on reality-monitoring ability(the ability to distinguish what has been per-ceived from what has been imagined; Johnson& Raye, 1981). Accurate reality-monitoringattributions are thought to rely on an individ-ual’s ability to encode, and later to retrieve, de-tails of the encoding episode. Memories forperceived events typically include more sensoryand contextual information, whereas memoriesfor imagined events often include more infor-mation about the cognitive operations thatsupported the internal generation of informa-tion (Johnson & Raye, 1981). Thus retrieval ofperceptual information will tend to correspondwith attribution of a memory to a presentedsource, whereas retrieval of information linkedto self-referential processing is related to an at-tribution of a memory to an internal source(Gonsalves & Paller, 2000; Kensinger &Schacter, 2006b).

We demonstrated that negative arousingitems were more often accurately attributedthan were neutral items (Kensinger & Schacter,2006d). This enhanced discrimination oc-curred both when encoding was incidental andwhen it was intentional, and the effect waspresent both for verbal stimuli and for singleobjects. Using functional magnetic resonanceimaging, we investigated the neural processesthat led to accurate memory attributions(Kensinger & Schacter, 2005a). Most notably,whereas enhanced encoding-related activity in


the posterior hippocampus was related to accu-rate memory attributions for all items (nega-tive arousing and nonemotional), enhancedencoding-related activity in the amygdala andthe orbito-frontal cortex corresponded witha reduction in the likelihood of memorymisattributions specifically for the negativearousing items. Activity in these limbic regions,often engaged during the processing of emo-tional information (e.g., Bechara, Damasio, &Damasio, 2000; Phan et al., 2005; Zald, 2003),showed no correspondence to memory accu-racy for the neutral items. Thus the enhancedaccuracy for negative arousing items did notstem solely from the additional engagement ofdomain-general processes that enhanced accu-racy for all items. Rather, domain-specific pro-cesses (processing of emotional information inthe amygdala and orbito-frontal cortex) in-creased the likelihood of accurate memory at-tributions for the emotional items.

It is important to note, however, that part ofthe effect of these emotion-specific processesappeared to have been exerted via their interac-tions with regions that promote accurate en-coding of both emotional and nonemotionalitems. In particular, activity in the amygdalawas highly correlated with activity in the hip-pocampus during the encoding of negativearousing items that were accurately attributedlater. Many studies have demonstrated the crit-ical role of the hippocampus in binding to-gether a nonemotional item and its context:Patients with hippocampal lesions have diffi-culties remembering the context in which anitem was studied (Giovanello, Verfaellie, &Keane, 2003; Shoqeirat & Mayes, 1991); olderadults who exhibit medial temporal lobe dys-function have deficits on these types of bindingtasks as well (Chalfonte & Johnson, 1996;Collie, Myers, Schnirman, Wood, & Maruff,2002); and neuroimaging studies have im-plicated the hippocampus specifically in theability to learn associative information, ascompared to nonassociative information (e.g.,Davachi & Wagner, 2002; Giovanello, Schnyer,& Verfaellie, 2004; Jackson & Schacter, 2004;but see Stark & Squire, 2001). Thus it appearsthat memory for the details of an emotionalevent can be enhanced not because individualsbring online an entirely distinct set of processesto help them remember the information, butrather because of limbic modulation of thesame processes (e.g., hippocampal bindingmechanisms) that are typically recruited to re-

member the details of nonemotional informa-tion.

In this reality-monitoring paradigm, mem-ory accuracy was higher for items with nega-tive content. However, not all types of contex-tual details seem to be enhanced by emotionalcontent. For example, emotion conferred noadvantage when individuals were asked to re-member whether they rated an item asanimate–inanimate or as common–uncommon(Kensinger & Schacter, 2006a). Importantly, inthis paradigm, amygdala activity correspondedonly with memory for the item (i.e., knowingwhether something was “old” or “new”), butnot with memory for the task performed withthe item (i.e., the “source”). A related findinghas been reported by Adolphs and colleagues(reviewed by Buchanan & Adolphs, 2002):They have found that although emotion en-hances the ability to remember the “gist” (gen-eral semantic theme) of scenes or stories, it canreduce memory for specific details. They havealso provided evidence that this emotion-related effect on memory may be mediated bythe amygdala: Patients with damage to theamygdala do not show the enhanced memoryfor gist or the impaired memory for detail(Adolphs, Tranel, & Buchanan, 2005).

These results highlight the fact that the roleof the amygdala during encoding of event de-tails may depend on the particular type of de-tail that is assessed. Emotion does not enhancememory for all aspects of an encoding episode,and amygdala engagement at encoding doesnot ensure that all details will be accurately re-membered. Although additional research willbe needed to clarify the circumstances in whichamygdala activity does or does not relate to theencoding of event details, one possible explana-tion is that amygdala activity guides encodingof details that are intrinsic to an item (e.g., itsphysical appearance or its gist), but does notenhance encoding of attributes that are extrin-sic to an item (e.g., the task performed with theitem; and see Mather et al., 2006, for evidencethat the neural processes supporting workingmemory for spatial location are disrupted foritems that elicit an arousal response). Thus re-call of such details as a word’s font may be en-hanced by emotion (Doerksen & Shimamura,2001; Kensinger & Corkin, 2003; MacKay etal., 2004) because those details are processedas intrinsic item attributes: A vivid memory ofa word’s presentation is likely to include thecolor or location of the word. Similarly, indi-


viduals may be more likely to remember neu-tral words occurring in an emotional sentence(Kensinger et al., 2002, 2004) because individ-uals process the entire sentence as a single stim-ulus rather than as a series of individual words.In contrast, focusing on the emotional itemmay actually preclude the processing of detailsextrinsic to that item. This hypothesis would beconsistent with the proposal that the effects ofemotion on memory may best be characterizedas tradeoffs: Some aspects of an event arebetter remembered because of the event’s emo-tional salience, whereas other aspects are morelikely to be forgotten (see reviews by Buchanan& Adolphs, 2002, and Reisberg & Heuer,2004). Of course this proposal leaves opena number of questions, including how theamygdala can exert these selective effectson memory. One possibility is that, just asamygdala–hippocampus interactions at re-trieval are guided by the orbito-frontal cortex(Smith et al., 2006), so does orbito-frontal ac-tivity during encoding guide the stimulus at-tributes that are attended and remembered.

EMOTION’S EFFECTSON IMPLICIT MEMORY

Although we have focused exclusively on emo-tion’s influence on conscious retrieval of previ-ously learned information, emotion also modu-lates implicit memory, altering how recentexperiences for which we do not have con-scious memory affect our behavior (but seeSteidl, Mohiuddin, & Anderson, 2006, for evi-dence that emotion can have dissociable effectson some forms of implicit memory comparedto explicit memory). Implicit memory can takemany forms, but here we focus on emotion’s ef-fects on “priming”—changes in a person’s per-ception or response to a stimulus due to itsprior exposure (for a recent review on priming,see Schacter, Dobbins, & Schnyer, 2004; for re-views of emotion’s effects on other forms of im-plicit memory, such as fear conditioning, seeMaren, 2001; Lavond, Kim, & Thompson,1993; LeDoux, 1996). A tremendous amountof research has indicated that prior exposure toa stimulus can influence people’s affective re-sponse toward it: Even when individuals arenot aware that they have encountered a stimu-lus previously, they often will have a preferencefor the previously seen item as compared to anovel one (the “mere-exposure effect”; for re-

views, see Bornstein, 1989; Harrison, 1977;Zajonc, 2001).

Although this research has focused on howunconscious processing of information can in-fluence a person’s affective response to a stimu-lus, more recent research has begun to addresswhether emotional content of information in-fluences responses within nonaffective domains(see Butler & Berry, 2004, for discussion ofsimilarities and differences between the mere-exposure effect and repetition priming). For ex-ample, are individuals faster to perceptuallyidentify or to make decisions about previouslyencountered information with emotional con-tent, as compared to previously encounteredinformation lacking emotional meaning? Theliterature in this area is relatively sparse, butthe studies to date suggest that emotional con-tent of information confers an advantage onpriming tasks, just as it does on tasks of ex-plicit memory: Emotional stimuli show largerenhancements on perceptual and conceptualpriming tasks than do nonemotional stimuli(Burton et al., 2004; Collins & Cooke, 2005;LaBar et al., 2005; Michael, Ehlers, &Halligan, 2005). Future studies will be requiredto elucidate the range of tasks for which emo-tion confers these priming benefits; the extentto which emotional stimuli are sensitive to ma-nipulations that alter the magnitude of primingfor nonemotional information (e.g., changes inmodality of presentation, of stimulus appear-ance, or of task instructions); and the degree towhich the affective quality of the stimuli (e.g.,the valence or arousal elicited, or the discreteemotion evoked; see Burton et al., 2005) influ-ences the strength of the priming. Future stud-ies would also do well to investigate the neuralcorrelates of emotional influences on priming.Neuroimaging studies of nonemotional infor-mation have consistently revealed that primingis accompanied by decreased activity in a num-ber of cortical regions (for reviews, see Grill-Spector, Henson, & Martin, 2006; Schacter &Buckner, 1998; Wiggs & Martin, 1998). Initialneuroimaging evidence suggests that emotioncan modulate the magnitude of such decreases(Ishai, Pessoa, Bikle, & Ungerleider, 2004), butfurther studies of the issue are required.

CONCLUSIONS

In this chapter, we have reviewed three ap-proaches to examining the effects of emotion


on memory: those focused on understandingwhat makes individuals more likely to remem-ber emotional experiences than nonemotionalones; those examining why emotional memo-ries often are reexperienced with tremendousvividness; and those investigating the types ofdetails that are more likely to be rememberedabout emotional experiences than aboutnonemotional ones. Each of these lines of re-search has provided behavioral evidence ofemotion–memory interactions, and neuroimag-ing and neuropsychological studies have dem-onstrated that the effects of emotion are notdue to domain-general processes that boostmemory for all information, but rather stemfrom engagement of emotion-specific process-ing (particularly in the amygdala and theorbito-frontal cortex). The modulatory influ-ence of these regions is apparent during mem-ory encoding, consolidation, and retrieval;during each of these memory phases, limbic ac-tivity serves to modulate perceptual and mne-monic function to increase the likelihood thatinformation is attended to and that at leastsome details are retained. Thus, at least in part,emotional information is remembered betterthan nonemotional information not because ofthe engagement of processes unique to memoryfor emotional information, but rather becauseof limbic modulation of the same processesthat are typically recruited to remember non-emotional information. Although increasingbehavioral evidence indicates that this limbicmodulation does not boost memory for all de-tails of an encoding episode, future studies willbe required to delineate the types of details thatare better remembered for emotional than forneutral information, as well as the neuralmechanisms that allow emotion to exert selec-tive effects on memory for some item attributesbut not for others.

REFERENCES

Abrisqueta-Gomez, J., Bueno, O. F., Oliveira, M. G., &Bertolucci, P. H. (2002). Recognition memory foremotional pictures in Alzheimer’s disease. ActaNeurologica Scandinavica, 105, 51–54.

Adolphs, R., Cahill, L., Schul, R., & Babinsky, R.(1997). Impaired declarative memory for emotionalmaterial following bilateral amygdala damage in hu-mans. Learning and Memory, 4, 291–300.

Adolphs, R., Tranel, D., & Buchanan, T. W. (2005).Amygdala damage impairs emotional memory for

gist but not details of complex stimuli. Nature Neu-roscience, 8, 512–518.

Amaral, D., Price, J., Pitkanen, A., & Carmichael, S.(1992). Anatomical organization of the primateamygdaloid complex. In J. Aggleton (Ed.), Theamygdala: Neurobiological aspects of emotion,memory, and mental dysfunction (pp. 1–67). NewYork: Wiley-Liss.

Amaral, D. G. (2003). The amygdala, social behavior,and danger detection. Annals of the New York Acad-emy of Sciences, 1000, 337–347.

Anders, S., Lotze, M., Erb, M., Grodd, W., &Girbaumer, N. (2004). Brain activity underlyingemotional valence and arousal: A response-relatedfMRI study. Human Brain Mapping, 23, 200–209.

Anderson, A. K., Christoff, K., Stappen, I., Panitz, D.,Ghahremani, D. G., Glover, G., et al. (2003). Dissoci-ated neural representations of intensity and valencein human olfaction. Nature Neuroscience, 6, 196–202.

Baas, D., Aleman, A., & Kahn, R. S. (2004).Lateralization of amygdala activation: A systematicreview of functional neuroimaging studies. Brain Re-search Reviews, 45, 96–103.

Bargh, J. A., Chaiken, S., Govender, R., & Pratto, F.(1992). The generality of the attitude activation ef-fect. Journal of Personality and Social Psychology,62, 893–912.

Bechara, A., Damasio, H., & Damasio, A. R. (2000).Emotion, decision making and the orbitofrontal cor-tex. Cerebral Cortex, 10, 295–307.

Bless, H., Clore, G. L., Schwarz, N., Golisano, V., Rabe,C., & Wolk, M. (1996). Mood and the use of scripts:Does a happy mood really lead to mindlessness?Journal of Personality and Social Psychology, 71,665–679.

Bless, H., & Schwarz, N. (1999). Sufficient and neces-sary conditions in dual-process models: The case ofmood and information processing. In S. Chaiken &Y. Trope (Eds.), Dual-process theories in social psy-chology (pp. 423–440). New York: Guilford Press.

Bohannon, J. N. (1988). Flashbulb memories for thespace shuttle disaster: A tale of two theories. Cogni-tion, 29, 179–196.

Bornstein, R. F. (1989). Exposure and affect: Overviewand meta-analysis of research, 1968–1987. Psycho-logical Bulletin, 106, 265–289.

Bradley, M. M., Greenwald, M. K., Petry, M. C., &Lang, P. J. (1992). Remembering pictures: pleasureand arousal in memory. Journal of Experimental Psy-chology: Learning, Memory, and Cognition, 18,379–390.

Brierley, B., Medford, N., Shaw, P., & David, A. S.(2004). Emotional memory and perception in tempo-ral lobectomy patients with amygdala damage. Jour-nal of Neurology, Neurosurgery and Psychiatry, 75,593–599.

Brown, R., & Kulik, J. (1977). Flashbulb memories.Cognition, 5, 73–99.


Buchanan, T. W., & Adolphs, R. (2002). The role of thehuman amygdala in emotional modulation of long-term declarative memory. In S. Moore & M.Oaksford (Eds.), Emotional cognition: From brain tobehavior (pp. 9–34). Amsterdam: Benjamins.

Buchanan, T. W., Tranel, D., & Adolphs, R. (2006).Memories for emotional autobiographical events fol-lowing unilateral damage to medial temporal lobe.Brain, 129, 115–127.

Budson, A. E., Simons, J. S., Sullivan, A. L., Beier, J. S.,Solomon, P. R., Scinto, L. F., et al. (2004). Memoryand emotions for the 9/11/01 terrorist attacks in pa-tients with Alzheimer’s disease, mild cognitive im-pairment, and healthy older adults. Neuropsycholo-gy, 18, 315–327.

Budson, A. E., Todman, R. W., Chong, H., Adams, E.H., Kensinger, E. A., Krandel, T. S., et al. (2006).False recognition of emotional word lists in agingand Alzheimer’s disease. Cognitive and BehavioralNeurology, 19, 71–78.

Burton, L. A., Rabin, L., Vardy, S. B., Frohlich, J.,Wyatt, G., Dimitri, D., et al. (2004). Gender differ-ences in implicit and explicit memory for affectivepassages. Brain and Cognition, 54, 218–224.

Burton, L. A., Rabin, L., Wyatt, G., Frohlich, J.,Bernstein Vardy, S., & Dimitri, D. (2005). Priming ef-fects for affective vs. neutral faces. Brain and Cogni-tion, 59, 322–329.

Bush, S. I., & Geer, J. H. (2001). Implicit and explicitmemory of neutral, negative emotional, and sexualinformation. Archives of Sexual Behavior, 30, 615–631.

Butler, L. T., & Berry, D. C. (2004). Understanding therelationship between repetition priming and mere ex-posure. British Journal of Psychology, 95, 467–487.

Cahill, L., Babinsky, R., Markowitsch, H. J., &McGaugh, J. L. (1995). The amygdala and emotionalmemory. Nature, 377, 295–296.

Cahill, L., Haier, R. J., Fallon, J., Alkire, M. T., Tang,C., Keator, D., et al. (1996). Amygdala activity at en-coding correlated with long-term, free recall of emo-tional information. Proceedings of the NationalAcademy of Sciences USA, 93, 8016–8021.

Cahill, L., & McGaugh, J. L. (1995). A novel demon-stration of enhanced memory associated with emo-tional arousal. Consciousness and Cognition, 4, 410–421.

Chalfonte, B. L., & Johnson, M. K. (1996). Featurememory and binding in young and older adults.Memory and Cognition, 24, 403–416.

Charles, S. T., Mather, M., & Carstensen, L. L.(2003). Aging and emotional memory: The forget-table nature of negative images for older adults.Journal of Experimental Psychology: General, 132,310–324.

Christianson, S.-A. (1989). Flashbulb memories: spe-cial, but not so special. Memory and Cognition, 17,435–443.

Christianson, S.-A. (1992). Emotional stress and eyewit-

ness testimony: A critical review. Psychological Bulle-tin, 112, 284–309.

Christianson, S.-A., & Engelberg, E. (1999). Organiza-tion of emotional memories. In T. Dalgleish & M.Power (Eds.), Handbook of cognition and emotion(pp. 211–227). Chichester, UK: Wiley.

Collie, A., Myers, C., Schnirman, G., Wood, S., &Maruff, P. (2002). Selectively impaired associativelearning in older people with cognitive decline. Jour-nal of Cognitive Neuroscience, 14, 484–492.

Collins, M. A., & Cooke, A. (2005). A transfer appro-priate processing approach to investigating implicitmemory for emotional words in the cerebral hemi-spheres. Neuropsychologia, 43, 1529–1545.

Conway, M. A. (1990). Conceptual representation ofemotions: The role of autobiographical memories. InK. J. Gilhooly, M. T. G. Keane, R. H. Logie, & G.Erdos (Eds.), Lines of thinking: Reflections on thepsychology of thought. Vol. 2. Skills, emotion, cre-ative processes, individual differences and teachingthinking (pp. 133–143). Chichester, UK: Wiley.

D’Argembeau, A., Comblain, C., & van der Linden, M.(2003). Phenomenal characteristics of autobiograph-ical memories for positive, negative, and neutralevents. Applied Cognitive Psychology, 17, 281–294.

D’Argembeau, A., Comblain, C., & van der Linden, M.(2005). Affective valence and the self-reference effect:Influence of retrieval conditions. British Journal ofPsychology, 96, 457–466.

D’Argembeau, A., & van der Linden, M. (2004). Influ-ence of affective meaning on memory for contextualinformation. Emotion, 4, 173–188.

Davachi, L., & Wagner, A. D. (2002). Hippocampalcontributions to episodic encoding: Insights fromrelational and item-based learning. Journal ofNeurophysiology, 88, 982–990.

Davidson, R. J., & Irwin, W. (1999). The functionalneuroanatomy of emotion and affective style. Trendsin Cognitive Sciences, 3, 11–20.

Dewhurst, S. A., & Parry, L. A. (2000). Emotionality,distinctiveness, and recollective experience. Euro-pean Journal of Cognitive Psychology, 12, 541–551.

Doerksen, S., & Shimamura, A. (2001). Source memoryenhancement for emotional words. Emotion, 1, 5–11.

Dolan, R. J., Lane, R., Chua, P., & Fletcher, P. (2000).Dissociable temporal lobe activations during emo-tional episodic memory retrieval. NeuroImage, 11,203–209.

Dolan, R. J., & Vuilleumier, P. (2003). Amygdalaautomaticity in emotional processing. Annals of theNew York Academy of Sciences, 985, 348–355.

Dolcos, F., LaBar, K. S., & Cabeza, R. (2004). Interac-tion between the amygdala and the medial temporallobe memory system predicts better memory for emo-tional events. Neuron, 42, 855–863.

Dolcos, F., LaBar, K. S., & Cabeza, R. (2005). Remem-bering one year later: Role of the amygdala and themedial temporal lobe memory system in retrieving


emotional memories. Proceedings of the NationalAcademy of Sciences USA, 102, 2626–2631.

Ehlers, A., Margraf, J., Davies, S., & Roth, W. T.(1988). Selective processing of threat cues in subjectswith panic attacks. Cognition and Emotion, 2, 201–219.

Garavan, H., Pendergrass, J. C., Ross, T. J., Stein, E. A.,& Risinger, R. C. (2001). Amygdala response to bothpositively and negatively valenced stimuli. Neuro-Report, 12, 2779–2783.

Gardiner, I. M., & Java, R. I. (1993). Recognizing andremembering. In A. F. Collins, S. E. Gathercole, M.A. Conway, & P. E. Morris (Eds.), Theories of mem-ory (pp. 163–188). Hove, UK: Erlbaum.

Giovanello, K. S., Schnyer, D. M., & Verfaellie, M.(2004). A critical role for the anterior hippocampusin relational memory: Evidence from an fMRI studycomparing associative and item recognition. Hippo-campus, 14, 5–8.

Giovanello, K. S., Verfaellie, M., & Keane, M. M.(2003). Disproportionate deficit in associative recog-nition relative to item recognition in global amnesia.Cognitive, Affective, and Behavioral Neuroscience,3, 186–194.

Gonsalves, B., & Paller, K. A. (2000). Neural events thatunderlie remembering something that never hap-pened. Nature Neuroscience, 3, 1316–1321.

Grill-Spector, K., Henson, R., & Martin, A. (2006).Repetition and the brain: Neural models of stimulus-specific effects. Trends in Cognitive Sciences, 10, 14–23.

Hamann, S. (2001). Cognitive and neural mechanismsof emotional memory. Trends in Cognitive Sciences,5, 394–400.

Hamann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. D.(1999). Amygdala activity related to enhanced mem-ory for pleasant and aversive stimuli. Nature Neuro-science, 2, 289–293.

Hamann, S. B., Ely, T. D., Hoffman, J. M., & Kilts, C.D. (2002). Ecstasy and agony: Activation of the hu-man amygdala in positive and negative emotion. Psy-chological Science, 13, 135–141.

Harrison, A. A. (1977). Mere exposure. In L. Berkowitz(Ed.), Advances in experimental social psychology(Vol. 10, pp. 39–83). New York: Academic Press.

Isen, A. M. (1999). Positive affect. In T. Dalgleish & M.Powers (Eds.), Handbook of cognition and emotion(pp. 75–94). Chichester, UK: Wiley.

Ishai, A., Pessoa, L., Bikle, P. C., & Ungerleider, L. G.(2004). Repetition suppression of faces is modulatedby emotion. Proceedings of the National Academy ofSciences, 101, 9827–9832.

Jackson, O., & Schacter, D. L. (2004). Encoding activityin anterior medial temporal lobe supports associativerecognition. NeuroImage, 21, 456–464.

Johnson, M. K., Foley, M. A., Suengas, A. G., & Raye,C. L. (1988). Phenomenal characteristics of memo-ries for perceived and imagined autobiographicalevents. Journal of Experimental Psychology: Gen-eral, 117, 371–376.

Johnson, M. K., Hashtroudi, S., & Lindsay, D. S.(1993). Source monitoring. Psychological Bulletin,114, 3–28.

Johnson, M. K., & Raye, C. L. (1981). Reality monitor-ing. Psychological Review, 88, 67–85.

Kensinger, E. A. (2004). Remembering emotional expe-riences: The contribution of valence and arousal. Re-views in the Neurosciences, 15, 241–252.

Kensinger, E. A. (in press). Neuroimaging the formationand retrieval of emotional memories. In Brain map-ping: New research. Hauppauge, NY: Nova Science.

Kensinger, E. A., Anderson, A., Growdon, J. H., &Corkin, S. (2004). Effects of Alzheimer disease onmemory for verbal emotional information. Neuro-psychologia, 42, 791–800.

Kensinger, E. A., Brierley, B., Medford, N., Growdon, J.H., & Corkin, S. (2002) The effect of normal agingand Alzheimer’s disease on emotional memory. Emo-tion, 2, 118–134.

Kensinger, E. A., & Corkin, S. (2003). Memory en-hancement for emotional words: Are emotionalwords more vividly remembered than neutral words?Memory and Cognition, 31, 1169–1180.

Kensinger, E. A., & Corkin, S. (2004). Two routes toemotional memory: Distinct neural processes for va-lence and arousal. Proceedings of the National Acad-emy of Sciences of the USA, 101, 3310–3315.

Kensinger, E. A., Garoff-Eaton, R. J., & Schacter, D. L.(2006). Memory for specific visual details can be en-hanced by negative arousing content. Journal ofMemory and Language, 54, 99–112.

Kensinger, E. A., Garoff-Eaton, R. J., & Schacter, D. L.(2007). Effects of emotion on memory specificity inyoung and older adults. Journal of Gerontology, 62,208–215.

Kensinger, E. A., Garoff-Eaton, R. J., & Schacter, D. L.(2007). Effects of emotion on memory specificity:Memory trade-offs elicited by negative visuallyarousing stimuli. Journal of Memory and Language,56, 575–591.

Kensinger, E. A., Krendl, A. C., & Corkin, S. (2006).Memories of an emotional and a nonemotionalevent: Effects of aging and delay interval. Experimen-tal Aging Research, 32, 23–45.

Kensinger, E. A., O’Brien, J., Swanberg, K., Garoff-Eaton, R. J., & Schacter, D. L. (in press). The effectsof emotional content on reality-monitoring perfor-mance in young and older adults. Psychology andAging.

Kensinger, E. A., & Schacter, D. L. (2005a). Emotionalcontent and reality-monitoring ability: FMRI evi-dence for the influence of encoding processes. Neuro-psychologia, 43, 1429–1443.

Kensinger, E. A., & Schacter, D. L. (2005b). Retrievingaccurate and distorted memories: Neuroimaging evi-dence for effects of emotion. NeuroImage, 27, 167–177.

Kensinger, E. A., & Schacter, D. L. (2006a). Amygdalaactivity is associated with the successful encoding ofitem, but not source, information for positive and


negative stimuli. Journal of Neuroscience, 26, 2564–2570.

Kensinger, E. A., & Schacter, D. L. (2006b). Neural pro-cesses underlying memory attribution on a reality-monitoring task. Cerebral Cortex, 16, 1126–1133.

Kensinger, E. A., & Schacter, D. L. (2006c). Processingemotional pictures and words: Effects of valence andarousal. Cognitive, Affective, and Behavioral Neuro-science, 6, 110–127.

Kensinger, E. A., & Schacter, D. L. (2006d). Realitymonitoring and memory distortion: Effects of nega-tive, arousing content. Memory and Cognition, 34,251–260.

Kensinger, E. A., & Schacter, D. L. (2006e). When theRed Sox shocked the Yankees: Comparing negativeand positive memories. Psychonomic Bulletin andReview, 13, 757–763.

Kilpatrick, L., & Cahill, L. (2003). Amygdala modula-tion of parahippocampal and frontal regions duringemotionally influenced memory storage. Neuro-Image, 20, 2091–2099.

Kitayama, S. (1990). Interaction between affect andcognition in word perception. Journal of Personalityand Social Psychology, 58, 209–217.

LaBar, K. S., & Phelps, E. A. (1998). Arousal-mediatedmemory consolidation: Role of the medial temporallobe in humans. Psychological Science, 9, 490–493.

LaBar, K. S., Torpey, D. C., Cook, C. A., Johnson, S. R.,Warren, L. H., Burke, J. R., et al. (2005). Emotionalenhancement of perceptual priming is preserved inaging and early-stage Alzheimer’s disease. Neuro-psychologia, 43, 1824–1837.

Lang, P. J., Greenwald, M. K., Bradley, M. M., &Hamm, A. O. (1993). Looking at pictures: Affective,facial, visceral, and behavioral reactions. Psycho-physiology, 30, 261–273.

Lavond, D. G., Kim, J. J., & Thompson, R. F. (1993).Mammalian brain substrates of aversive classicalconditioning. Annual Review of Psychology, 44,317–342.


LeDoux, J. E. (1996). The emotional brain: The myste-rious underpinnings of emotional life. New York: Si-mon & Schuster.

Levine, L. J., & Bluck, S. (2004). Painting with broadstrokes: Happiness and the malleability of eventmemory. Cognition and Emotion, 18, 559–574.

Linton, M. (1975). Memory for real-world events. In D.A. Norman & D. E. Rumelhart (Eds.), Explorationsin cognition (pp. 376–404). San Francisco: Freeman.

MacKay, D. G., & Ahmetzanov, M. V. (2005). Emotion,memory and attention in the taboo Stroop paradigm:An experimental analogue of flashbulb memories.Psychological Science, 16, 25–32.

MacKay, D. G., Shafto, M., Taylor, J. K., Marian, D. E.,Abrams, L., & Dyer, J. R. (2004). Relations betweenemotion, memory, and attention: Evidence from ta-boo Stroop, lexical decision, and immediate memorytasks. Memory and Cognition, 32, 474–488.

Maratos, E. J., Dolan, R. J., Morris, J. S., Henson, R.N., & Rugg, M. D. (2001). Neural activity associatedwith episodic memory for emotional context. Neuro-psychologia, 39, 910–20.

Maren, S. (2001). Neurobiology of Pavlovian fear con-ditioning. Annual Review of Neuroscience, 24, 897–931.

Markowitsch, H. J., Calabrese, P., Wurker, M., Durwen,H. F., Kessler, J., Babinsky, R., et al. (1994). Theamygdala’s contribution to memory: A study on twopatients with Urbach–Wiethe disease. NeuroReport,5, 1349–1352.

Mather, M., & Carstensen, L. L. (2005). Aging and mo-tivated cognition: The positivity effect in attentionand memory. Trends in Cognitive Sciences, 9, 496–502.

Mather, M., Mitchell, K. J., Raye, C. L., Novak, D. L.,Greene, E. J., & Johnson, M. K. (2006). Emotionalarousal can impair feature binding in working mem-ory. Journal of Cognitive Neuroscience, 18, 614–625.

Matt, G. E., Vazquez, C., & Campbell, W. C. (1992).Mood-congruent recall of affectively toned stimuli: Ameta-analytic study. Clinical Psychology Review, 12,227–255.

McGaugh, J. L. (2004). The amygdala modulates theconsolidation of memories of emotionally arousingexperiences. Annual Review of Neuroscience, 27, 1–28.

McGaugh, J. L., & Roozendaal, B. (2002). Role of adre-nal stress hormones in forming lasting memories inthe brain. Current Opinion in Neurobiology, 12,205–210.

Michael, T., Ehlers, A., & Halligan, S. L. (2005). En-hanced priming for trauma-related material in post-traumatic stress disorder. Emotion, 5, 103–112.

Mori, E., Ikeda, M., Hirono, N., Kitagaki, H.,Imamura, T., & Simomura, T. (1999). Amygdalarvolume and emotional memory in Alzheimer’s dis-ease. American Journal of Psychiatry, 156, 216–222.

Neisser, U., & Harsch, N. (1992). Phantom flashbulbs:False recollections of hearing the news about Chal-lenger. In E. Winograd & U. Neisser (Eds.), Affectand accuracy in recall: Studies of ‘flashbulb’ memo-ries (pp. 9–31). Cambridge, UK: Cambridge Univer-sity Press.

Ochsner, K. N. (2000). Are affective events richly “re-membered” or simply familiar? The experience andprocess of recognizing feelings past. Journal of Ex-perimental Psychology: General, 129, 242–261.

Öhman, A. (1988). Preattention processes in the genera-tion of emotions. In V. Hamilton, G. H. Bower, & N.H. Frijda (Eds.), Cognitive perspectives on emotionand motivation (Vol. 44, pp. 127–143). Norwell,MA: Kluwer Academic.

Ortony, A., Turner, T. J., & Antos, S. J. (1983). A puzzleabout affect and recognition memory. Journal of Ex-perimental Psychology: Learning, Memory, and Cog-nition, 9, 725–729.

Paradis, C. M., Solomon, L. Z., Florer, F., & Thompson,


T. (2004). Flashbulb memories of personal events of9/11 and the day after for a sample of New York Cityresidents. Psychological Reports, 95, 304–310.

Pessoa, L., Kastner, S., & Ungerleider, L. G. (2003).Neuroimaging studies of attention: From modulationof sensory processing to top-down control. Journalof Neuroscience, 23, 3990–3998.

Pezdek, K. (2003). Event memory and autobiographicalmemory for the events of September 11, 2001. Ap-plied Cognitive Psychology, 17, 1033–1045.

Phan, K. L., Fitzgerald, D. A., Nathan, P. J., Moore, G.J., Uhde, T. W., & Tancer, M. E. (2005). Neural sub-strates for voluntary suppression of negative affect: Afunctional magnetic resonance imaging study. Bio-logical Psychiatry, 57, 210–219.

Phelps, E. A. (2004). Human emotion and memory: In-teractions of the amygdala and hippocampal com-plex. Current Opinion in Neurobiology, 14, 198–202.

Reisberg, D., & Heuer, F. (2004). Remembering emo-tional events. In D. Reisberg & P. Hertel (Eds.),Memory and emotion (pp. 3–41). New York: OxfordUniversity Press.

Richardson, M. P., Strange, B., & Dolan, R. J. (2004).Encoding of emotional memories depends on theamygdala and hippocampus and their interactions.Nature Neuroscience, 7, 278–285.

Riemann, B., & McNally, R. J. (1995). Cognitive pro-cessing of personally relevant information. Cognitionand Emotion, 9, 325–340.

Roediger, H. L., III, Watson, J. M., McDermott, K. B.,& Gallo, D. A. (2001). Factors that determine falserecall: A multiple regression analysis. PsychonomicBulletin and Review, 8, 385–407.

Royet, J.-P., Zald, D., Versace, R., Costes, N., Lavenne,F., Koenig, O., et al. (2000). Emotional responses topleasant and unpleasant olfactory, visual, and audi-tory stimuli: A positron emission tomography study.Journal of Neuroscience, 20, 7752–7758.

Rubin, D. C., & Kozin, M. (1984). Vivid memories.Cognition, 16, 63–80.


Schacter, D. L., & Buckner, R. L. (1998). Priming andthe brain. Neuron, 20, 185–195.

Schacter, D. L., Dobbins, I. G., & Schnyer, D. M.(2004). Specificity of priming: A cognitive neurosci-ence perspective. Nature Reviews Neuroscience, 5,853–862.

Schaefer, A., & Philippot, P. (2005). Selective effects ofemotion on the phenomenal characteristics of auto-biographical memories. Memory, 13, 148–160.

Schmidt, S. R. (2002). Outstanding memories: The posi-tive and negative effects of nudes on memory. Journalof Experimental Psychology: Learning, Memory, andCognition, 28, 353–361.

Schmidt, S. R. (2004). Autobiographical memories forthe September 11th attacks: Reconstructive errors

and emotional impairment of memory. Memory andCognition, 32, 443–454.

Schmolck, H., Buffalo, E. A., & Squire, L. R. (2000).Memory distortions develop over time: Recollectionsof the O. J. Simpson trial verdict after 15 and 32months. Psychological Science, 11, 39–45.

Sergerie, K., Lepage, M., & Armony, J. L. (2006). Aprocess-specific functional dissociation of the amyg-dala in emotional memory. Journal of Cognitive Neu-roscience, 18, 1359–1367.

Sharot, T., Delgado, M. R., & Phelps, E. A. (2004).How emotion enhances the feeling of remembering.Nature Neuroscience, 12, 1376–1380.

Shoqeirat, M. A., & Mayes, A. R. (1991). Dispropor-tionate incidental spatial-memory and recall deficitsin amnesia. Neuropsychologia, 29, 749–769.

Small, D. M., Gregory, M. D., Mak, Y. E., Gitelman,D., Mesulam, M. M., & Parrish, T. (2003). Disso-ciation of neural representation of intensity and af-fective valuation in human gustation. Neuron, 39,701–711.

Smith, A. P., Henson, R. N., Dolan, R. J., & Rugg, M.D. (2004). fMRI correlates of the episodic retrieval ofemotional contexts. Neuroimage, 22, 868–878.

Smith, A. P., Stephan, K. E., Rugg, M. D., & Dolan, R.J. (2006). Task and content modulate amygdala–hippocampal connectivity in emotional retrieval.Neuron, 49, 631–638.

Smith, M. C., Bibi, U., & Sheard, D. E. (2003). Evidencefor the differential impact of time and emotion onpersonal and event memories for September 11,2001. Applied Cognitive Psychology, 17, 1047–1055.

Stark, C. E., & Squire, L. R. (2001). Simple and associa-tive recognition memory in the hippocampal region.Learning and Memory, 8, 190–197.

Steidl, S., Mohiuddin, S., & Anderson, A. (2006). Ef-fects of emotional arousal on multiple memory sys-tems: Evidence from declarative and procedurallearning. Learning and Memory, 13, 650–658.

Sterpenich, V., D’Argembeau, A., Desseilles, M.,Balteau, E., Albouy, G., Vandewalle, G., et al. (2006).The locus ceruleus is involved in the successful re-trieval of emotional memories in humans. Journal ofNeuroscience, 26, 7416–7423.

Storbeck, J., & Clore, G. L. (2005). With sadness comesaccuracy; with happiness, false memory: Mood andthe false memory effect. Psychological Science, 16,785–791.

Stradler, M. A., Roediger, H. L., III, & McDermott, K.B. (1999). Norms for word lists that create falsememories. Memory and Cognition, 27, 494–500.

Talarico, J. M., LaBar, K. S., & Rubin, D. C. (2004).Emotional intensity predicts autobiographical mem-ory experience. Memory and Cognition, 32, 1118–1132.

Talarico, J. M., & Rubin, D. C. (2003). Confidence, notconsistency, characterizes flashbulb memories. Psy-chological Science, 14, 455–461.


Talmi, D., & Moscovitch, M. (2004). Can semantic re-latedness explain the enhancement of memory foremotional words? Memory and Cognition, 32, 742–751

Taylor, S. F., Liberzon, I., Fig, L. M., Decker, L. R.,Minoshima, S., & Koeppe, R. A. (1998). The effectof emotional content on visual recognition memory:A PET activation study. NeuroImage, 8, 188–197.

Vuilleumier, P., Richardson, M. P., Armony, J. L., Driver,J., & Dolan, R. J. (2004). Distinct influences ofamygdala lesion on visual cortical activation duringemotional face processing. Nature Neuroscience, 7,1271–1278.

White, R. T. (2002). Memory for events after twentyyears. Applied Cognitive Psychology, 16, 603–612.

Whittlesea, B. W. A. (1993). Illusions of familiarity.Journal of Experimental Psychology: Learning,Memory, and Cognition, 19, 1235–1253.

Whittlesea, B. W. A., & Williams, L. D. (2000). Thesource of feelings of familiarity: The discrepancy-attribution hypothesis. Journal of Experimental Psy-chology: Learning, Memory, and Cognition, 26,547–565.

Wiggs, C. L., & Martin, A. (1998). Properties andmechanisms of perceptual priming. Current Opinionin Neurobiology, 8, 227–233.

Windmann, S., & Kruger, T. (1998). Subconscious de-tection of threat as reflected by an enhanced responsebias. Consciousness and Cognition, 7, 603–633.

Windmann, S., & Kutas, M. (2001). Electrophysio-logical correlates of emotion-induced recognitionbias. Journal of Cognitive Neuroscience, 13, 577–592.

Winograd, E., & Killinger, W. A., Jr. (1983). Relatingage at encoding in early childhood to adult recall: De-velopment of flashbulb memories. Journal of Experi-mental Psychology: General, 112, 413–422.

Yonelinas, A. P. (2002). The nature of recollection andfamiliarity: A review of 30 years of research. Journalof Memory and Language, 46, 441–517.

Zajonc, R. B. (2001). Mere exposure: A gateway to thesubliminal. Current Directions in Psychological Sci-ence, 10, 224–228.

Zald, D. H. (2003). The human amygdala and the emo-tional evaluation of sensory stimuli. Brain Research:Brain Research Reviews, 41, 88–123.


C H A P T E R 3 8

A Framework for RepresentingEmotional States

MARVIN MINSKY

We often describe personalities in terms of con-trasting pairs of traits (e.g., solitary vs. socia-ble, tranquil vs. anxious, or joyous vs. sorrow-ful). But sorrow is not the mere absence of joy,nor is anxiety the reverse of tranquility. Simi-larly, we often describe how a person is think-ing in terms of opposing characteristics (e.g.,rational vs. intuitive, intellectual vs. emotional,conscious vs. unconscious, or deliberate vs.spontaneous). Indeed, such pairwise divisionsseem indispensable for everyday social commu-nication, by providing us with compact expres-sions for depicting a person’s mental condition.However, it seems to me that when we usethose distinctions in psychology, this retardsour technical progress by making it very diffi-cult to see inside our mental processes. Thischapter suggests a way to describe our minds interms of larger numbers of smaller parts, whichmight make better ingredients for us to use toconstruct the kinds of theories we need.1

Many attempts to classify emotions begin bydescribing some of them as “primary”—andthen try to show how these might combine to

produce more complex states of mind. Forexample, we might start with the likes ofpleasure, anger, fear, and disgust—and thencombine these to represent such conditions ashappiness, grief, and anxiety. This approachwas successful in physical science; once we dis-covered the chemical elements and how thesecombined to form molecules, this soon led todeep understandings of the properties of com-mon substances. So why not use the same ap-proach to show how our high-level feelings andtraits could come from combining simpler,more basic ones?

But what should we choose for those basicingredients? It is tempting to use the familiarterms that we find in our everyday folk psy-chology, because the meanings of words suchas “thinking,” feeling,” “emotion,” and “con-sciousness” seem so natural, clear, and direct tous. But then we find it extremely hard to an-swer questions like these “Can you have feel-ings without being conscious?” or “Can youthink without having emotions?” The troubleis that each of those common words refers to so

618

many different mental processes that they sim-ply don’t serve well as “elements” for makinggood theories about how minds work.

What “elements” could we use instead as abasis for psychology? One answer comes fromanatomy. We know, both from neuropathologyand from the new scanners we use today, thateach human brain has hundreds of parts, eachof which does different kinds of jobs. Some ofthem recognize situations; others tell musclesto execute actions; and yet other regions of thebrain accumulate and go on to exploit enor-mous bodies of common-sense knowledge.And although our present-day methods don’tyet have enough resolution (in space or in time)to tell us how those processes work, they cer-tainly show solid evidence that our present the-ories don’t have enough parts. So now let’s con-sider a view of the mind that does, at least inprinciple, provide enough room to include themany sorts of processes that support our hu-man resourcefulness.

A METATHEORY OF EMOTIONS

What are emotions, and how do they work?Perhaps the most popular view of emotions isthat they somehow add feelings, values, andother dimensions to ordinary rationalthinking—just as artists add colors to black-and-white drawings. But that does not helpmuch to explain how those processes producetheir complex effects.

Another popular approach is to try to clas-sify emotions by starting with some “primary”ones (e.g., discomfort, pleasure, anger, and joy)and using these as coordinates that generatesimple two- or four-dimensional spaces; thenwe can try to represent higher-level emotions asregions in a metrical space. This can yield someclinically useful distinctions, but does not pro-vide good explanations. We can’t understanda complex machine just from distinguishingsome of its parts; we also need ideas about howeach part works and how they are all connect-ed. Those popular diagrams and charts simplydo not tell us enough.

Furthermore, it seems to me that the lan-guage of common-sense psychology makes ithard for us to formulate good theories aboutthe great networks of processes inside ourbrains because, to understand that machinery,we’ll need much more expressive terms thanwords like “feelings,” “emotion,” and “con-

sciousness.” For example, Chapter 4 of TheEmotion Machine (Minsky, 2006) argues that“consciousness” refers to at least 20 differentkinds of processes! If so, we’ll need to replaceeach traditional term with a substantial collec-tion of technical concepts.

At first, replacing some ideas that seem sim-ple at first with new ones that seem much moredifficult might appear to make everythingworse. However, on a larger scale, this increasein complexity can actually make our job easier.For once we split each old idea into parts, wethen can replace each old unsolvable mysterywith several new and smaller ones—each ofwhich still presents obstacles, but no longerseems unsolvable. So, to start dividing thoseold big problems into smaller ones, this newtheory begins by portraying a typical brain ascontaining a great many parts, each of whichcan do some particular kind of job (see Figure38.1). I’ll use the term “resource” to refer toany such structure or function inside the brain,from the lowest-level reactions to the highest-level mental activities.2

We can use this image whenever we want toexplain some mental condition or activity (suchas anger or fear) by trying to show how thatstate of mind might result from activating acertain collection of mental resources. For ex-ample, the state called “anger” appears toarouse resources that make us react with un-usual speed and strength, while suppressing re-sources that we might otherwise use to planand act more prudently; thus anger replacescautiousness with aggressiveness and changessympathy into hostility. Similarly, the conditioncalled “fear” appears to engage resources inways that cause us to retreat.3

CITIZEN: I sometimes find myself in a statewhere everything seems cheerful and bright.

38. A Framework for Representing Emotional States 619

FIGURE 38.1. An image of the typical humanbrain containing many parts. Adapted fromMinsky (2006).

At other times (although nothing haschanged), all my surroundings seem drearyand dark, and my friends describe me as be-ing “depressed.” Why do I have such moods,or feelings, or dispositions—and what causesall their strange effects?

Some popular answers to this person’s ques-tion are “Those states are caused by chemicalsin the brain,” or “They result from an excess ofstress,” or “They come from thinking depress-ing thoughts.” However, such statements saylittle about how those processes work, whereasthe idea of selecting a set of resources can sug-gest more specific ways in which thinking canchange, as in the case of another phenomenon:

When a person has fallen in love, it’s almostas though someone new has emerged—a per-son who thinks in other ways, with alteredgoals and priorities. It’s almost as though aswitch has been thrown and a different pro-gram has started to run.

What could happen inside a brain to makesuch changes in how it thinks? Here is an ap-proach that we could take: Each of our major“emotional states” results from turning certainresources on while turning certain others off—and thus changing some ways that our brainsbehave. But what activates such sets of re-sources? The new theory suggests that ourbrains are equipped with resources that I shallcall “Critics”—each of which is specialized torecognize a certain condition, and then to acti-vate some particular set of other resources.Some of our Critics are built in from birth, toactivate certain “instinctive” reactions thatevolved to help our ancestors survive (see Fig-ure 38.2). Thus anger and fear evolved for de-fense and protection, while hunger and thirstevolved for nutrition.

The first few chapters of The Emotion Ma-chine also try to show how this kind of machin-

ery could activate states of mind that we mightidentify with feelings like attachment, grief,and depression. However, as we learn andgrow, we also develop new sets of higher-levelresources—and this leads to the types of mentalstates that we regard as more “intellectual”than “emotional.” For example, whenever aproblem seems hard to you, your mentalCritics may start to switch among different setsof resources. One such set may help you to di-vide the problem into smaller parts; anothermay start to search through your memories tofind suggestive analogies; and yet another set ofresources may lead you to ask some other per-son for help. In effect, your collection of mentalCritics may switch you into many different“Ways to Think.” Each of our major Ways toThink results from turning certain resources onwhile turning certain others off—and thuschanging some ways that our brains behave(see Figure 38.3).

CITIZEN: It seems strange that you’ve given thesame description both for emotions and forregular thinking. But thinking is basicallyrational—dry, detached, and logical—whereas emotions enliven our ways to thinkby adding irrational feelings and biases.

I think it’s a myth that there’s any such thingas purely logical, rational thinking, because ourthoughts are always affected by our assump-tions, biases, values, and goals. Furthermore,emotions don’t always add features tothoughts; indeed, many so-called emotionalstates appear to suppress the use of certain re-sources, as in this case of infatuation:

“I’ve just fallen in love with a wonderful per-son. I scarcely can think about anything else.My sweetheart is unbelievably perfect—ofindescribable beauty, flawless character, andincredible intelligence. There is nothing Iwould not do for her.”


FIGURE 38.2. Examples of built-in critics.Adapted from Minsky (2006).

FIGURE 38.3. Examples of Ways to Think.Adapted from Minsky (2006).

This speaker sees all this as positive, and itmakes him feel happy and more productive.But note that there’s something strange aboutthose superlatives: Most of those phrases ofpositive praise use syllables like “un-,” “-less,”and “in-”—which show that they really arenegative statements describing the personwho’s saying them!

Wonderful. Indescribable.(I can’t figure out what attracts me toher.)

I scarcely can think about anything else.(Most of my mind has stopped working.)

Unbelievably perfect. Incredible.(No sensible person believes such things.)

She has a flawless character.(I’ve abandoned my critical faculties.)

There is nothing I would not do for her.(I’ve forsaken most of my usual goals.)

This example suggests that some of the pleas-ant effects of infatuation may result from sup-pressing some mental resources that a personotherwise uses to recognize faults and defectsin somebody else. In other words, what we’reseeing here is the turning off of some of thisspeaker’s mental Critics!

CITIZEN: I still think that your view of emo-tions ignores too much. For example, emo-tional states like fear and disgust involve thebody as well as the brain, as when we feeldiscomfort in the chest or gut, or when wesense palpitations of the heart, or when wefeel faint or tremble or sweat.

We can make our model account for theways in which emotions involve the body’s con-dition by regarding our body parts as addi-tional resources that our brains can use tochange (or maintain) their mental states—aswhen we make ourselves stick to a plan bymaintaining a certain facial expression.

THE “CRITIC–SELECTOR” MODELOF MIND

Some early behaviorist theories described ananimal’s mind as mainly composed of a simplecollection of “If → Do” rules, where each “If”describes an external situation, and its “Do”describes a way to react to it (see Figure 38.4).Now let us try to extend this idea to include

mental reactions to mental situations. In thismore general view of the mind, a mental Criticcan react to recent mental events by making alarge-scale change in how we think. It does thisby changing the set of resources in use. So we’llsuppose that each Critic has learned to recog-nize a certain type of mental situation, so thatthen, whenever that condition occurs, thatCritic will try to activate a state of mind thathas been useful in the past for dealing with thattype of situation. For example, if you fail toachieve some goal—and if one of your Criticscan diagnose what has gone wrong—then itcan activate a different Way to Think thatmight help you to deal with that problem (seeFigure 38.5).

Here are a few useful ways in which thiscould work; several chapters of The EmotionMachine (Minsky, 2006) describe these inmuch greater detail.

• Analogy: If a problem seems familiar, try us-ing a method that has worked on some simi-lar problems.

• Reformulation: If the problem seems unfa-miliar, change how you are describing it byusing a different kind of representation.

• Planning: If the problem seems too difficult,try to divide it into several smaller parts,which you can use as separate goals orsteppingstones.

• Simplification: If a problem seems too com-plex, ignore some aspects that seem


FIGURE 38.4. “If–Do” rules. Adapted fromMinsky (2006).

FIGURE 38.5. Selecting from a cloud of re-sources. Adapted from Minsky (2006).

difficult—and then try to restore them, oneat a time.

• Contradiction: Try to prove that your prob-lem cannot be solved, and then find a flaw inthat argument.

• Simulation to anticipate: Mental experi-ments in virtual worlds can help you safelymake predictions.

• Emotion: A flash of impatience or anger cancut what seems like a hopelessly tangledknot. Each “emotional state” is a somewhatdifferent way to think.

• Collaboration: If none of your usual meth-ods work, ask some other person to assist, inmuch the same way as an infant will cry forhelp.

To develop this view of the mind further, sev-eral chapters of The Emotion Machine discusswhat sorts of mental resources a person mighthave. Some resources act as Critics that canrecognize ways in which you get stuck and thenact to select which other resources will be en-gaged next. Thus, when you have trouble withsolving a problem, a Critic may turn some ofyour active resources off, while turning on oth-ers that may help to split that problem intoparts. Other Critics may use other tricks—suchas to remember a similar task from the past, orto change how you’re describing that problem.Some such selections may change your mentalstate so much that, in effect, you’ll be using aquite different Way to Think.

STUDENT: Why should you ever turn off a re-source? Why not keep them all working allthe time?

Indeed, certain resources are never switchedoff—such as those involved with vital functionslike respiration, balance, and posture, or thosethat constantly keep watch for certain particu-lar types of danger. However, if all your re-sources were active at once, they would too of-ten get into conflicts. You can’t make yourbody both walk and run, or move in two differ-ent directions at once. So when you have sev-eral goals that conflict (say, because they com-pete for the same energy, time, or space), thenyou need some machinery that helps to selectwhich resources to activate, and with howmuch intensity or priority.

One of the functions that mental Criticscould serve could be to assess progress towardone of your current goals—and if there’s no

movement in that direction, a Critic could tryto replace it by one or more other goals. Buthow could you represent a goal so that it takesthe form of some sort of resource? One way todo this would be to say that when you “want”a certain condition or thing, this means thatyou’re using a mental process that works to re-duce the difference between your present situa-tion and a condition in which you possess thatthing (Newell & Simon, 1963). Figure 38.6 is adiagram that shows a way to embody that pro-cess as a machine.

For example, every baby is born with multi-ple systems for maintaining normal body tem-perature. Such a goal is aroused when the in-fant is too hot, and causes it to sweat, pant,stretch out, or vasodilate. However, when thebaby is too cold, this goal will cause it to curlup, shiver, vasoconstrict, or raise its metabolicrate. And whereas infants do these by using in-nate reactions, they eventually learn to usehigher-level, more deliberate methods (see Fig-ure 38.7).

A SIX-LEVEL MODEL OF MIND

A particular Critic might observe that twosubgoals are incompatible, that the system hasbeen repeating itself, or that some other prob-lem needs more urgent attention. If just one ofyour Critics is thus aroused, then it is clearwhat to do next. But what happens when sev-eral Critics are aroused, but disagree abouthow to proceed? You could sometimes solvesuch conflicts by arranging for those Critics tocompete in some sort of marketplace, wherethey each have different weights, priorities, orsome other sorts of currencies. However, whensimple schemes like that don’t work, thenyou’ll need higher-level Critics that can recog-nize this conflict itself as being another prob-


FIGURE 38.6. The process of wanting. Adaptedfrom Minsky (2006).

lem to solve. For example, if you have no sim-ple way to choose between two conflictingalternatives (such as whether to fight or toflee), then it might make sense to “look ahead”by imagining, and then comparing, the likelyconsequences of both alternatives. (Note thatsimple “If → Do” rules cannot make such pre-dictions; one would need more powerful “If →Do → Then” rules to envision effects of imag-ined actions.) This suggests that such a systemwould need a second layer of machinery to crit-icize and supervise the collection of simpler, re-active behaviors (see Figure 38.8).

To see how such a system might work, con-sider this pair of episodes in which a young in-fant got upset. At first, that change seemed asquick as the flip of a switch:

A certain infant could not bear frustration,and would react to each setback by throwinga tantrum. He’d hold his breath and his backwould contract, so that he’d fall backwardon his head.

Yet several weeks later, that behavior hadchanged:

No longer completely controlled by his rage,he could also add ways to protect himself, sothat when he felt this coming on, he’d run tocollapse on some soft, padded place.

This example suggests that the infant brain isnot at first able to resolve the kinds of conflictswe face in our later lives. This led our humanbrains to evolve higher-level systems that madeus much more able to “control ourselves.”Then some instincts that formerly were distinctcould now become increasingly mixed.

In adult human beings, these processes haveevolved to a high degree, and The EmotionMachine concludes that they extend to includeleast six different levels (see Figure 38.9). Thelowest level of this figure includes the kinds ofinstinctive reactions with which all mammalianbrains are equipped from birth. Then, in child-hood, we go on to develop several more levelsof methods that we use to deal with variouskinds of problems, situations, and conflicts(this includes much of what we call “common-


FIGURE 38.7. The process of maintaining normal body temperature. Adapted from Minsky (2006).

FIGURE 38.8. A system with three levels ofCritics and Selectors. Adapted from Minsky(2006).

FIGURE 38.9. The six-level model of mind.Adapted from Minsky (2006).

sense thinking”), and eventually we each de-velop systems that support the sorts of con-cepts that we call by names like “ethics” or“values.” (I don’t mean to suggest that theselevels are formed by some uniform, general-purpose learning scheme; instead, I suspect thatthose layers are formed by specific systems ofhighly evolved genetic machinery. That’s whyyou cannot teach calculus to your cat.)

We can see some aspects of how such a sys-tem might work in this example of an everydayhuman event:

Joan is starting to cross the street on the wayto deliver a finished report. While thinkingabout what to say at the meeting, she hears asound and turns her head —and sees aquickly oncoming car. Uncertain whether tocross or retreat, but uneasy about arrivinglate, Joan decides to sprint across the road.She later remembers her injured knee and re-flects upon her impulsive decision: “If myknee had failed, I could have been killed—and what would my friends have thought ofme?”

Let’s try to describe the events in Joan’s mind inrelation to the six levels of activities, rangingfrom her simplest instinctive reactions up to theprocesses that engage her highest ideals andpersonal goals:

• Innate reactions: Joan hears a sound andturns her head. Some reactions evolved to helpus to survive, by responding in particular waysto certain kinds of events in the world, in thebody, or inside the mind.

• Learned reactions: Joan sees a quickly on-coming car. She has had to learn that certainconditions demand specific new ways to react.

• Deliberative thinking: Joan is consideringwhat to say at the meeting. She attempts topredict and compare the effects of somealternatives—just as an adult who encounters athreat need not just react instinctively, but in-stead could deliberate on whether to retreat orattack. Indeed, it can make sense to make one-self angry deliberately, if one believes that thiswill intimidate an adversary.

• Reflective thinking: Joan reflects on herdecision. Here she reviews her recent selectionsof methods and actions, and wonders whetherthose decisions were good.

• Self-reflective thinking: Joan is uneasyabout arriving late. Now she sees that in the

plan she made, she may not have allowedenough time to be sure she could actually carryit out.

• Self-conscious thinking: “What would myfriends have thought of me?” Here Joan won-ders how well her actions agreed with her ide-als, and appraises her abilities and goals incomparison with those of other people sheknows.

At the lowest levels, these Critics and Selec-tors are almost the same as the “Ifs” and“Thens” of simple reactions. But at the higherreflective levels, these Critics and Selectors cancause so many changes that, in effect, theyswitch us to different Ways to Think. I shouldnote that this model agrees with some ofSigmund Freud’s ideas, in which he recognizedthat human thinking does not proceed in anysingle, uniform way, but as a host of diverse ac-tivities that often lead to conflicts and inconsis-tencies (see Figure 38.10).

PSYCHOLOGIST: I do not see clear distinctionsamong the six levels of your model—especially among the three at the top, be-cause they seem to use similar thinking tech-niques. Why do you think we should treatthem as separate—and do you really need somany levels? Most physicists would be in-clined to insist that we should never makemore assumptions than we need.

I agree that those boundaries are indistinct.However, as Freud might suggest, we shouldn’tsuccumb to “physics envy.” It is true that thescience of physics has progressed amazinglywell by trying to find the simplest possible ex-planations. However, we already know thatour brains contain hundreds of structuralparts, and when we know that our theories are


FIGURE 38.10. A model of some Freudian con-cepts. Adapted from Minsky (2006).

incomplete (as is clearly the case in psychol-ogy), then it is risky to use a model that leavesno room for other ideas that we’re likely toneed.

Viewing the mind as a cloud of resourcescould be a powerful tool to help understandmany kinds of mental processes. However, todevelop this scheme further, we’ll need manyfar more specific ideas about which resourcesare engaged for each particular mentalfunction—along with some suggestions abouthow each particular such resource works, andwhich others it gets connected to. The EmotionMachine proposes several such schemes, butthere is not enough space here to describe theirdetails, so below I’ll just outline a few of them(readers will find a final draft of that book atwww.emotionmachine.net, and several relatedessays at web.media.mit.edu/~minsky).

OUTLINE OF THEORIESIN

The goal of The Emotion Machine is to try toexplain the diversity that distinguishes us hu-mans from most other animals—and from allthe machines that we’ve built in the past. Soeach chapter discusses a different collection ofsources of our uniquely human resourceful-ness.

Chapter 1 suggests that our mental stateschange when we alter our set of active re-sources. Such a transition may start when oneresource gains enough control to turn on sev-eral more, while also suppressing a number ofothers. Then, if some of those newly arousedresources start to affect yet other ones, this canlead to a spreading cascade that may greatlymodify the way in which we think aboutthings. Of course, many resources will remainunchanged, so that we will still be the same inmany respects. We will still be able to see, hear,and speak—but now we’ll perceive things indifferent ways, and select different subjects tothink about. We still will have access to most ofour knowledge—but now we will have differ-ent attitudes. We may still hold some of thesame plans and goals—but now they’ll havedifferent priorities. Yet, despite all thesechanges, our self-reflective processes may insistthat we retain the same “identity”— a conceptdiscussed in Chapter 9.

Chapter 2 is concerned with questions abouthow human children acquire their goals. Many

traditional theories of learning attribute our ac-quisition of knowledge to a process called“reinforcement,” in which we learn from beingrewarded for each successful performance. Butwhat we perceive as a failure or success de-pends on the goals that we then are pursuing;for instance, food won’t reward a sated animal.Each of us is born with some “basic” goals thatevolved for nutrition, health, comfort, and de-fense; then later, we learn to connect newsubgoals to these. However, this connectingdoes not explain how we can learn new kindsof goals (such as the ones we call “values” or“ideals”) that are not connected to existinggoals—so we also need some theories abouthow we acquire new kinds of goals. In otherwords, we need some ideas about how we learnwhat we “ought” to want.

Accordingly, this chapter conjectures thatthe special emotions called “pride” and“shame” may play unique and peculiar roles inhow we humans develop new values, ideals,and goals. When children are praised or re-jected by the persons to whom they’ve becomeattached, they don’t simply just feel pleasure ordissatisfaction, which would merely reinforceor extinguish their currently active wants; in-stead, these would be elevated or depressed. Inother words, whereas trial and error teach usnew ways to achieve the goals we alreadymaintain, attachment-related shame and praiseteach us which goals we should discard or re-tain. If so, then we need a theory about whatsorts of brain machinery could be involvedwith exploiting those special social relation-ships. To illustrate the power of such emotions,here is how Michael Lewis (1995) depicts someeffects of shame:

Shame results when an individual judges his or heractions as a failure in regard to his or her stan-dards, rules and goals and then makes a global at-tribution. The person experiencing shame wishesto hide, disappear or die. It is a highly negativeand painful state that also disrupts ongoingbehavior and causes confusion in thought and aninability to speak. The body of the shamed personseems to shrink, as if to disappear from the eye ofthe self or others. Because of the intensity of thisemotional state, and the global attack on the self-system, all that individuals can do when presentedwith such a state is to attempt to rid themselves ofit. (pp. 68–78)

Chapter 3 asks how the sensation called“pain” can lead to the kinds of conditions that


we describe as “hurting,” “distress,” and “suf-fering.” I conjecture that prolonged or chronicpain (no matter whether mental or physical)can start to suppress so many goals and pro-cesses that we become imprisoned from havingso few remaining alternatives. I also discusswhat would happen when too many Critics be-come suppressed or aroused (which could leadto manic or depressive states), as well as the ex-tent to which our higher levels could learn tocontrol or exploit such conditions.

Chapter 4 discusses the extent to which vari-ous parts of a brain can know, represent, orcomprehend the events in other parts of thatbrain. I argue that when we use terms like“conscious” or “self-aware,” these do not referto some single thing, but to some 20-odd dif-ferent sorts of processes—so, naturally, theirmeanings will seem controversial and mysteri-ous. I try to resolve this by observing that mostmental processes work in ways that do not leadus to reflect on why and how we are doingthem. However, when those processes don’tfunction well, or when they encounter obsta-cles, this can cause our Critics to initiate atleast four kinds of higher-level activities—which (1) make use of our most recent memo-ries, (2) tend to be more serial and less parallel,(3) tend to use symbolic descriptions, and (4)involve the models we make of ourselves. I ar-gue that such processes could help to accountfor many properties of what we like to call“conscious thought.”

Chapter 5 develops more details of how thesix-level model of mind can help to explain ourabilities to “imagine” things—that is, howparts of our higher-level machinery can enableus to make predictions about what would hap-pen if we were to perform some sequence ofmental or physical actions. The chapter focusesmainly on visual perceptions and their internalrepresentations.

Chapter 6 is about common-sense knowledge—the things we expect other persons to know, sothat we rarely need to mention them. Considerwhat happens when you hear your telephonering. When you move to pick the receiver up, atfirst you see only one side of it, yet you feel thatyou’ve seen the entire thing; you anticipatehow it will fit in your grasp, and how it will feelwhen it contacts your ear; you know that youspeak into the bottom of the receiver and hearanswers from the top; and you know that ifyou dial a number, another phone will ringsomewhere else—and if anyone happens to an-

swer it, then you and that person can start toconverse. How does such scanty evidence makeit seem that the object you’re “looking at” hasbeen transported into your mind—where youcan move it and touch it and turn it around,even open it up and look inside? This chapterproposes some theories about how we acquire,represent, retrieve, and apply multiple kinds ofcommon-sense knowledge.

Chapter 7 develops more ideas about theCritic–Selector model of mind, and makes anattempt to classify our various ways to thinkabout things. How do we learn and organizeour new collections of resources? How do weknow when to quit or persist? And how do wechoose what next goal to pursue? I conjecturethat those strategies use mainly unconsciousprocesses, some of which may be cyclical incharacter.

Chapter 8 explores the sources of the humanbrain’s resourcefulness. My central conjecturehere is that our ancestors evolved several differ-ent ways to deal with every important threat tosurvival. Accordingly, we have evolved manyways to represent knowledge, each with differ-ent kinds of interconnections that give us a va-riety of methods for retrieving records thatmight be relevant. For example, we can de-scribe events in terms of lists of images, words,or story-like scripts, which in turn can be com-posed of pairs of “frames” that represent thesituations before and after an action is taken.These in turn can be represented by using thebranching web-like structures that researchersin artificial intelligence call “semantic net-works.” At lower levels, we also can use“connectionistic” neural or statistical net-works, which often are useful for rapid reac-tions, but tend to be extremely opaque to moresymbolic, abstract, or reflective methods. Fig-ure 38.11 illustrates the various structures wecan use to represent objects and ideas, whichenable us to see each situation from many dif-ferent points of view.4 Also, we seem to haveevolved special machinery that helps us toswitch rapidly among those perspectives; I callthose systems “panalogies” (for “parallel anal-ogies”) and would welcome future experimentsto see whether those structures really exist.

Chapter 9 proposes some theories aboutwhat we refer to as our “selves” or “identi-ties,” and discusses some possible reasons forwhy we find feelings so hard to describe, andalso why we like pleasure and dislike pain. Italso considers our sense of experiencing (and


why some philosophers regard this as verymysterious). The book concludes by arguingthat we don’t need to appeal to magical “gifts”to account for our human intelligence; the trickis that we each develop a massive collection ofdifferent ways to deal with each challenge thatfaces us.5

NOTES

1. Much of this chapter is condensed from parts of TheEmotion Machine (Minsky, 2006), which describesmore details about these ideas. Copyright 2006 byMarvin Minsky. Adapted by permission.

2. In earlier publications, I have used the term “agents”to refer to mental resources, but today that term is of-ten used for other purposes.

3. Some resources may simply turn on or off, while oth-ers may be aroused to various degrees. Also, somemay have serious conflicts with others—and this mayrequire additional Critics to recognize and deal withthese.

4. Figure 38.11 combines ideas about ways to representknowledge that range from the low-level k-lines andmicronemes of Chapters 8 and 20 of Minksy (1986)to the high-level semantic networks and narratives ofChapter 5 through 8 of Minksy (2006). The cognitive

levels of Figure 38.9 need not closely correspond tothe neurological levels of Figure 38.9 because similarstructures could be reused for different kinds of func-tions.

5. Several reviewers have complained about the lack ofneurological evidence for many of these hypotheses.However, it seems to me that most such objectionsmiss the mark, because the proposals herein arelargely based on computational concepts that havenot yet spread through the neuroscience community.So, naturally, we won’t find many experiments rele-vant to these kinds of conjectures until researchersbegin to think about them.

REFERENCES

Lewis, M. (1995). Self-conscious emotions. AmericanScientist, 83, 68–78.

Minsky, M. (1986). The society of mind. New York: Si-mon & Schuster.

Minsky, M. (2006). The emotion machine: Common-sense thinking, artificial intelligence, and the fu-ture of the human mind. New York: Simon &Schuster.

Newell, A., & Simon, H. A. (1963). GPS, a programthat simulates human thought. In E. A. Feigenbaum& J. Feldman (Eds.), Computers and thought. NewYork: McGraw-Hill.


FIGURE 38.11. The structures we can use at different levels to represent objects and ideas. Adapted fromMinsky (2006).

C H A P T E R 3 9

Appraisal TheoriesHow Cognition Shapes Affect into Emotion

GERALD L. CLORE and ANDREW ORTONY

At dusk recently, while walking in the woods, I(G. C.) was startled by an arresting sound. Itturned out to be a male deer announcing hispresence with a loud snort. The realization thatit was only a deer came quickly, and the wholesequence from startle to categorization tookonly an instant. But then I experienced a novel,surprising sensation as the hair on the back ofmy neck stood up. Despite the cognitive reas-surance that it was only a deer, the body wasapparently still preparing for a nameless threat.

What happened in this episode in thewoods? Were appraisals involved? Was fearelicited? If so, did fear cause the piloerection?To this last question, William James would cer-tainly have replied, “No,” claiming that we are“afraid because we tremble” rather than theother way around (James, 1890, p. 450). In thischapter, we explore such questions, and end upagreeing both with James and with appraisaltheorists.

APPRAISAL THEORYAND ITS CRITICS

Appraisal theory as we know it today is usuallyattributed to Magda Arnold (1960), who madean early and influential statement of the cogni-tive approach to emotion. She proposed thatpeople implicitly appraise or evaluate every-thing they encounter, and that such evaluationsoccur immediately and automatically. Amongothers who were important in defining cogni-tive approaches to emotion were Schachter andSinger (1962), Lazarus (1966), and Mandler(1975). But it was not until the 1980s that theseed planted by Arnold began to take root. Fol-lowing pioneering work by Roseman (1979,1984), a number of appraisal theories ap-peared, including an often-cited account bySmith and Ellsworth (1985), an influential the-ory by Frijda (1986), and accounts by Oatleyand Johnson-Laird (1987) and ourselves

628

(Ortony, Clore, & Collins, 1988). In addition,Scherer (1984) proposed a theory emphasizingthe temporal sequence of appraisals. These the-ories are compared in Clore, Schwarz, andConway (1994), and some are elaborated bytheir authors in a volume edited by Scherer,Schorr, and Johnstone (2001).

Despite this flowering of interest, appraisaltheories turned out to be controversial. Someinvestigators found them implausible, becausethey read accounts of appraisal structure asthough they were assertions about appraisalprocess (e.g., Prinz, 2005). Although appraisaltheorists themselves often helped blur the dis-tinction, it is an important one. The differencecan perhaps be appreciated from an analogousdifference between the formal rules of syntaxand the processes involved in speaking. Lin-guists might all agree that observing the rulesof syntax is important for communication, butnone assume that speakers consult such rulesbefore opening their mouths to speak. Simi-larly, theories of the structure of emotional ap-praisals do not claim that people consult suchrules before feeling anything. In addition, somecritics argue that emotions occur too quicklyfor cognitive appraisals to be possible. How-ever, the validity of structural accounts of emo-tion is no more contingent on the speed of emo-tion than the validity of structural accounts oflanguage is dependent on how fast people talk.

In this chapter we focus on issues relevantto theories of appraisal process rather thanappraisal structure. Also, instead of focusingon what has been said in the past, we empha-size ideas that have appeared recently. We askquestions about emotion and appraisal, includ-ing what emotions are and in what sense theyexist. One answer to the latter question isthat emotions exist only in the sense that theBig Dipper exists—namely, in our perceptions(Russell, 2003). We also ask whether emotionscause behavior. Some investigators argue thatunconscious affect causes behavior, but thatfull-blown emotions do not (Baumeister, Vohs,DeWall, & Zhang, 2007). We go on to askwhether appraisals are really necessary in theemotion process. The philosopher Prinz (2005)argues that appraisals are no more necessaryfor emotion than for pain. Finally, we ask whatprocess theorists have in mind when they referto “appraisals.” We review both dual-processmodels (Clore & Ortony, 2000; C. A. Smith,Griner, Kirby, & Scott, 1996; E. R. Smith &

Neumann, 2005) and alternatives to dual pro-cess models (e.g., Barrett, 2006a; Huron, 2006;Cunningham & Zelazo, 2007).

In our own view, “emotions” are cognitivelyelaborated affective states (e.g., Ortony et al.,1988; Ortony, Norman, & Revelle, 2005). Inthis characterization, we take “affective” tomean anything evaluative, and we propose that“states” exist when multiple components rep-resent or register the same internal or externalsituation in the same time frame. Thus one canthink of emotions as involving multiple repre-sentations of something as good or bad in someway. Appraisal theories address the nature ofsuch evaluations. The general view we take inthis chapter is that (as in the piloerection exam-ple with which we have begun this chapter)low-level bodily, hormonal, and affective reac-tions often get the emotional process started,and that cognitive appraisal processes act like asculptor, shaping general affective reactionsinto specific emotions. But before addressingthe process, we ask about the nature of theemotions that involve such appraisals.

THE NATURE OF EMOTIONS

Are Emotions Situated Constructions?

Theorists increasingly view psychological pro-cesses as distributed across multiple sources(e.g., Clark, 1997), and behavioral and affec-tive responses as situated or context-specific(e.g., Brown, Collins, & Duguid, 1989). Con-sistent with such approaches, Barrett (e.g.,2006a) has proposed that emotions are not dis-tinct states with clear boundaries, that they arenot hard-wired in the brain, and that they don’thave distinctive psychophysiological signaturesand facial expressions. Emotions do, of course,involve facial expressions, psychophysiology,and specialized brain areas, but the boundariesof these are not those of specific emotions suchas anger, fear, and shame. Barrett suggests thatemotions, rather than reflecting similar entitiesin the brain, involve combinations of processeswith much more variability than is implied bythe discrete-emotions or basic-emotions view.

The prevailing psychological picture hasbeen that emotions are distinct entities in thebody or evolved modules (e.g., Ekman, 1984).Once activated by emotional stimuli, theseemotions are evident in distinctive expressions,thoughts, feelings, neurochemistry, and behav-

39. Appraisal Theories 629

ior. But for Barrett positive and negative affec-tive reactions and arousal—the dimensions of“core affect” (Russell & Barrett, 1999)—maybe the only necessary givens in the body. Spe-cific emotions are situated instances of such af-fect. Thus “fear” is a label for negative affect insituations involving threat, whereas the samenegative affect in situations of loss may becalled “sadness,” and a reaction to blamewor-thy behavior may be called “anger” (for a re-lated view, see Sabini & Silver, 2005).

If Barrett (2006a) is right that the bound-aries between similarly valenced emotions arenot as distinct as is usually assumed, it wouldnot necessarily make specific emotions any lessimportant, powerful, or universal. It wouldsimply change the locus and nature of their dis-tinctiveness. In such a view what makes emo-tions universal is not their biological status, butthe situations to which they are responses. Ifthere is anything basic about fear, anger, joy,and disgust, it lies in the ubiquity of the life sit-uations to which these emotions are responses,rather than in distinctive biological signatures,which Barrett says research does not find.

In this alternative view the experience of spe-cific, distinct emotions arises partly from cul-tural knowledge about emotions. In a similarmanner, experiencing colors of similar wave-lengths as the same or as different dependspartly on the boundaries provided by a cul-ture’s language. Members of cultures with sep-arate concepts for blue and green may see twodifferent colors, where those without a sepa-rate category for green may see only variants ofblue. Barrett suggests that whether people ex-perience two points along the affect dimensionas two emotions, rather than as variants of one,may be similarly influenced. Thus conceptsmay augment sensations to create distinct emo-tions, as well as distinct colors.

Evidence that concepts do help create theboundaries of emotions comes from experi-ments inducing semantic satiation of an emo-tion word by multiple repetitions, which makesthe word temporarily lose meaning (Lindquist,Barrett, Bliss-Moreau, & Russell, 2006). Aftersemantic satiation of the word “anger,” for ex-ample, people were no longer able to swiftlyrecognize that two patently angry faces wereexpressing the same emotion, presumably be-cause the temporary breakdown of access tothe anger concept caused by the satiation ma-nipulation made it difficult to “see” anger inthe faces.

Barrett’s work appears likely to spark debatein much the same way as Mischel’s (1968) bookdid about personality. Mischel argued that theavailable evidence failed to show the high degreeof cross-situational consistency in behavior de-manded by traditional ideas about personality.He proposed a view of personality reflecting thesituated nature of individual differences inbehavior potential. After 40 years of debate andrevision, the field of personality study is argu-ably healthier, even though many personalitytheorists still disagree with Mischel’s position.Similarly, Barrett is saying that the evidence alsodoes not show that emotions are tightly orga-nized, evolved modules with distinctive expres-sive, experiential, and neural signatures. It re-mains to be seen whether her claims will have asimilar effect on the study of emotion.

Do Emotions Cause Expressions?1

Theorists have traditionally thought aboutemotion within a kind of latent-trait model.Thus research focuses on evidence of an unseenentity measured by indicators—expressions,physiology, behavior, and feelings that aresymptoms or consequences of an emotion. Analternative approach might treat emotion notas causing these disturbances, but as emergingfrom them. Such a model would not includearrows going outward from emotion to itsseveral indicators; instead, the arrows wouldgo the other way, pointing inward from cog-nitions, expressions, and feelings to the emo-tion they constitute. The factors considered asindicators of emotion in the traditional latent-trait model can be seen instead as constituentsof emotion in an emergence model. Bollen andLennox (1991) distinguish these as “effect indi-cator” and “causal indicator” models (see alsoBarrett, 2006b).

In this view, emotions are the conjunction ofexpressions, physiology, behavior inclination,experience, and so on (Barrett, Ochsner, &Gross, 2006; Clore & Centerbar, 2004; Clore& Ortony, 2000). It sees emotions as affectivestates, in which multiple components registerthe same emotional significance in differentways at the same time. For example, if threatwere registered simultaneously in facial expres-sion, posture, tone of voice, thought, motiva-tion, neurochemistry, autonomic activity, brainactivation, and so on, a person would be in astate of fear. Indeed, that is what we mean by“fear”—being in a state dedicated to threat.


In the conventional latent-trait or effect indi-cator model, expressions in the face, voice, andposture and actions, thoughts, and desires aredifferent indicators of an underlying emotion.In the emergence model, these things are con-stituents of emotions, so that an emotion ex-ists by virtue of their co-occurrence. If one’sthoughts turn elsewhere, one’s physiology re-verts to baseline, and one’s voice, face, posture,and motivation no longer represent threat, thenone is no longer afraid. Of course, in neithermodel is it necessary for each and every aspectto be evident. One can be afraid without gap-ing, opening one’s eyes widely, and developinga squeaky voice. But for one to be in a genuinestate of fear, threat must be multiply repre-sented.

In summary, in this section we have com-pared classical views that emotions areevolved, tightly organized modules with an al-ternative view of emotions as loosely orga-nized, psychologically constructed states, con-sisting of situated elicitations of core affect(Barrett, 2006a). We have considerable sympa-thy for the latter view (e.g., Ortony et al.,2005). Next, we have contrasted a model inwhich expressions and feelings are indicatorsof underlying emotion with an emergencemodel in which expressions and feelings areconstituents rather than consequences ofemotion. Our own view is compatible withan emergence variable model (e.g., Clore &Centerbar, 2004). We turn next to an examina-tion of the process of appraisal.

THE NATURE OF APPRAISALS

Are Appraisals Cognitive or Perceptual?

Critics of appraisal theory often object not toappraisals per se, but to the idea that they arecognitive in nature. Some consider themselvesappraisal theorists, but not cognitive appraisaltheorists (J. E. LeDoux, personal communica-tion, March 2006). In contrast to cognitive ac-counts of emotion elicitation, Parkinson (2007,p. 22) suggests that emotions are “direct ad-justments to relational dynamics, not articu-lated responses to propositional representa-tions of appraisal information.” He argues thatif one steps onto a bus and one’s carrier bagsplits apart, scattering groceries across thepavement, one feels frustrated directly in a waythat does not require verification of appraisal-

relevant propositions. Similarly, Parkinson(2007, p. 21) says that “the minimal precondi-tion for anger is simply resistance stopping usfrom getting through,” and that a person bran-dishing a gun can be directly perceived as scarywithout any need for (cognitive) appraisalsby the perceivers (for a similar view, seeBerkowitz, 1990).

Parkinson may be interpreting propositionalaccounts (e.g., Siemer & Reisenzein, 2007) ofgoals and goal blockage as requiring an explicitanalysis of situations before emotions canarise. But we doubt that anyone holds such aposition. A propositional analysis of the dis-tress at seeing one’s groceries scattered acrossthe pavement does not imply that people ex-plicitly entertain propositions. A propositionalanalysis is a formal description of the implicitmeaning of that perception. Similarly, a para-bolic equation might be a formal description ofthe trajectory of a ball thrown through the air.But neither the distress nor the flight of the ballrequires a propositional or mathematical anal-ysis for its occurrence. Seeing one’s groceriesspill out surely would produce distress directlyand without thought, but a propositional anal-ysis does not suggest otherwise. Whatever cog-nitive activity is involved is implicit. A numberof models (described below) have addressedsuch issues, arguing that any requisite cognitiveactivity can occur exceedingly rapidly andwithout conscious thought.

Nevertheless, we agree that many of the cri-teria for emotion elicitation are as much per-ceptual as cognitive. Gibson (1979) challengedorthodoxy by throwing out the Cartesian ideaof the mind as a separate entity that creates aperceptual model of the world from sensory in-puts and guides action. Perception, he said, re-flects the environment directly, and action re-flects its affordances. Many appraisals mayalso be perceptual rather than mental products:Is an outcome my own or another’s? Is it pastor future? Is it human or environmentallycaused? Thus some appraisals are already inthe topography of psychological situations, sothat explanations for emotion lie not in themind, but in what the mind is in (Buck, March4, 2007).

Are Appraisals Cognitive or Situational?

Thus far we have treated emotions as emer-gent, situated, and constructed from underly-ing affective dimensions. In this view, when the


elements in a situation match patterns of thekind specified in appraisal theories, the corre-sponding emotion can result. If so, then emo-tion might be thought of as radically situated,so that emotional variation directly reflects sit-uational variation.

If we think of emotions as reactions to differ-ent situational structures, then the only essencethat instances of emotion share is in the situa-tion they represent. If so, the fact that one in-stance of anger or fear or joy looks a lot likeanother is evidence not so much that a discreteemotion has been evoked, as that aspects ofwhatever situations elicit fear, anger, or joyhave constrained affective reactions in distinc-tive ways. Thus the distinctiveness of an emo-tion may lie in the nature of the situation it rep-resents, not in a stored pattern of latentemotional potential. In this view, emotions arenot entities, any more than cognitions exist asentities. Just as chameleons cope with varia-tions in their environment by changing color,so we cope by transforming ourselves emotion-ally into reflections of our environment.

Many different writers have tried to equateemotions and situations. For example, Polti(1921/1977) proposed that there are 36 basicplots in the history of drama correspondingto 36 emotions. They concern love, tragedy,hope, fear, betrayal, honor, sacrifice, passion,lust, sympathy, ambition, jealousy, shortsight-edness, courage, revelation, forgiveness, deliv-erance, rivalry, jealousy, and more. There isno reason to assume that these categories aretruly basic, or that another investigator mightnot find more or fewer. Moreover, each nar-rative situation would surely involve manyemotions. But the larger point is that thereare recurrent patterns of situations in humanaffairs, which have been the stuff of dramafrom the Greeks to the present. The actionsdepicted are energized and directed by univer-sal human motivations, and the dramaticturning points are marked by the emotions ofcharacters and audiences.

In summary, we have suggested that some as-pects of emotional appraisals may be percep-tual rather than cognitive, and that appraisalsoften directly reflect the structure of the situa-tions in which they arise, with little cognitiveelaboration. But cognition is involved in full-blown emotions, and a central question con-cerns how it is involved.

THE PROCESS OF APPRAISAL

Are Emotions Too Fast or Too Mindlessfor Cognitive Appraisals?

Automaticity

Reservations about cognitive appraisal theoriesoften rest on the claim that emotions are toofast or mindless to involve cognition. In fact,mindlessness is a feature of most cognitive re-actions as well. For example, many people aregood at determining whether the footsteps theyhear belong to a man or a woman (Huron,2006). It turns out that the sound is differentbecause the timing of first heel and then toe hit-ting the ground is shorter for small than forlarge shoes. People who are good at makingthis discrimination do it mindlessly, however.They are unaware of what cues they use, andcertainly never think of foot size or shoe move-ments in order to decide on the sex of the per-son. Instead, they directly hear a man or awoman. It is a learned association, but theyhave no explicit knowledge of what they havelearned.

In addition to being mindless, emotional re-actions are often assumed to be too fast forcognitive appraisals. However, Moors (2006)has recently reported research explicitly exam-ining appraisal time, and she concludes thatthere is no reason to assume that cognitive pro-cesses are too slow to produce emotion.Through clever experiments that build on priorresearch on automaticity and affective priming,she has shown that appraisal judgments canbe rapid and automatic. Also, Moors andDeHouwer (2001) have demonstrated thatstimulus valence and motivational relevancecan be determined rapidly and automatically,as assumed in many appraisal theories.

Other research has shown that people maybe faster to infer emotions (on the basis of briefvignettes) than to infer the appraisals on whichthese emotions are believed to depend (Siemer& Reisenzein, 2007). But, of course, this find-ing does not mean that emotions cannot bebased on cognitive appraisals. Nor do these au-thors suggest otherwise. Categorizing a figureas a bird may also be faster than judgingwhether the figure has wings, even though hav-ing wings may be one of the criteria for makingthat categorization in the first place. Since mul-tiple attributes contribute to birdness (or toemotion), a categorization may be made as


soon as the collective activation of some ofthem is adequate. This may occur before anysingle attribute is sufficiently activated to serveas a basis by itself, and long before respondentscan affirm the presence of that attribute in self-reports.

Siemer and Reisenzein (2007) adapt an exist-ing cognitive model (Anderson, 1983) to ex-plain why their finding does not mean thatemotions precede appraisal. They argue thatwith experience, appraisals become automated;appraisal programs get compiled so that notonly do the appraisals become rapid, but theintervening steps become inaccessible (Wegner& Vallacher, 1986). This dual-process ap-proach assumes that with practice, emotion in-ferences can become faster than appraisal infer-ences. As in playing the piano, routines becomeautomated as they are repeated over and over.One might think that this implies that whereasemotions might be quick to arise in adults, chil-dren might have to think before they feel. Butthis particular implication seems unlikely, sothat dual processes reflecting the automatic–controlled distinction, in which automaticity isachieved through practice, may be less relevantthan models in which one of the processes is,for example, heuristic or associative ratherthan simply well practiced.

Dual-Process Theories

Dual-process theories have become common inpsychology as a way of handling conflicting re-sults. Rather than having to choose whetherthinking is propositional or heuristic, dual-process theorists say that both are at work. Theclaim is that people engage in both heuristicand systematic (Chaiken, Liberman, & Eagly,1989) or both central and peripheral (Petty &Cacioppo, 1986) processing.

Dual-process models have similar power forhandling data about emotion. A useful dual-process model of appraisal has been based on adistinction between “associative processing”and “rule-based processing” (Sloman, 1996).This approach has been developed by CraigSmith and colleagues (e.g., Smith et al., 1996;Smith & Kirby, 2001), but others have alsofound the idea useful (Clore & Ortony, 2000;E. R. Smith & Neumann, 2005). Associativeprocessing is guided by subjective similarityand temporal contiguity, whereas rule-basedprocessing involves symbolic reasoning. In ev-

eryday categorization, people appear to useboth subjective similarity and rule-based rea-soning.

Both kinds of reasoning have also been pro-posed as playing a role in emotion elicitation.In new situations, emotions may involve con-siderable bottom-up processing. As perceptionsof a situation unfold and requirements for spe-cific emotions are satisfied, reactions may be-come correspondingly differentiated. Similarly,as children develop and become capable ofmaking relevant distinctions, their emotionsmay become more differentiated. Such rule-based reasoning is rarely conscious, explicit, ordeliberative, and preverbal infants are alreadysurprisingly adept at implicit rule-based pro-cessing (e.g., Kotovsky & Baillargeon, 1994;Needham & Baillargeon, 1993).

But emotions are presumably also elicited byassociative processes. One may become happy,angry, or anxious simply by being in situationssimilar to those in which one was previouslyhappy, angry, or anxious. And since cognitivesystems capitalize on prior experience ratherindiscriminately, such associative processingmay be the rule rather than the exception. Themeaning and significance of an event are per-haps always partly contingent on its resem-blance to other situations in one’s experience.LeDoux (1996) captured this idea by saying,“Emotion is memory” (p. 249). Rather thanconfronting each situation as a blank slate,people assume that the emotionally relevant as-pects of situations recur, so that prior apprais-als often get reinstated.

The kind of reinstatement mechanism wehave in mind was actually proposed by Freud.In his view, “the act of birth, moreover, is thefirst experience attended by anxiety, and isthus, the source and model of the affect of anx-iety” (Freud, 1911, p. 251). He also believedthat reactions to one’s father serve as a proto-type for later emotional orientations to author-ity figures, and that falling in love is a reinstate-ment of a child’s love for its parent. Some ofFreud’s examples seem bizarre and implausible,but the general mechanism of reinstatementhas great explanatory power.

Bowlby’s (1969) infant attachment theoryalso assumes that early emotional reactions oflove and attachment are the basis of later emo-tions. He saw the emotional protests of infantsseparated from caregivers as evidence of anevolved tendency for children to become


emotionally attached to caregivers. LaterAinsworth, Blehar, Waters, and Wall (1978)identified specific patterns of infantile attach-ment believed to be reinstated in later romanticattachments. The idea that early emotional pat-terns reappear when people fall in love and se-lect their mates continues to be fruitful in con-temporary attachment work (e.g., Morgan &Shaver, 1999; Shaver & Clark, 1994).

Although there may be more than one routeto emotion elicitation, the appraisal for a givenemotion remains the same. Regardless ofwhether fear or anger arises from computation,conditioning, imitation, or predisposition, fearis always a response to apparent threat, and an-ger to apparent infringement. Whereas the con-stituent thoughts, feelings, and physiology maydiffer, each instance of anger or fear involvessimilar perceptions. Lazarus (1994) referred tothese as co-relational themes. This idea, com-mon to all appraisal approaches, is that thereis an underlying constancy in situations thatmakes them sources of anger, fear, or joy.

In summary, in this section, we have exam-ined an example of a dual-process model(Clore & Ortony, 2000; see also Sloman, 1996;C. A. Smith et al., 1996). We assume that dif-ferent routes to emotional appraisal serve dif-ferent functions. The typically faster, associa-tive process of emotion reinstatement is usefulfor preparedness, whereas the typically slower,rule-based process of computation affords flex-ibility. Creatures with more restricted emo-tional repertoires are less capable of flexible re-sponding than creatures with more complexemotional repertoires. Presumably there is anadvantage to automatic preparation for re-sponding, as well as to flexibility in respond-ing. As Scherer (1984) pointed out, the evo-lutionary advantage of emotion was that astimulus could be registered and reacted towithout the organism’s being committed tobehavior.

These considerations suggest that emotionallows behavior to be contingent on a stimuluswithout being dictated by it. There are, there-fore, two fingers on the emotional trigger—onefrom early perceptual processes that identifythe emotional value of a stimulus to prepare foraction; and one from cognitive processes thatverify the nature of the stimulus, situate it, andappraise it. Before discussing cognitive involve-ment, however, let us examine further the roleof low-level, subcortical processes.

Is There a Low Route to Emotion?

Investigators of emotion have become increas-ingly interested in the role of subcortical pro-cesses in emotion. The work of LeDoux andcolleagues (e.g., LeDoux, Romanski, &Xagoraris, 1989) in finding a subcortical “lowroute” to the elicitation of fear-relevant re-sponses is well known. In studying fear condi-tioning in rats, LeDoux and his colleaguesfound a pathway from the sensory thalamus di-rectly to the amygdala, without first going torelevant areas of the cortex to be interpreted.To the extent that the amygdala is involved infear and other emotions, this low route ap-peared capable of generating emotional re-sponses before the organism could know whatit was responding to or have any experience offear. This finding was taken to imply that cog-nitive appraisal is not required for emotion.

It would be difficult to overstate the impactof this discovery on thinking about emotion. Itwas excellent science and has been importantin stimulating further research on subcorticalprocesses. However, many criticisms of the ac-cepted interpretation have appeared (for an ac-cessible review, see Storbeck, Robinson, &McCourt, 2006). The work was done on rats,and it now appears that the particular pathwayexamined may not be functional in primatesand humans. Moreover, by itself, this pathwaywould not be able to discriminate stimuli thathad acquired emotional significance from thosethat had none, unless reduced to light versusdark or similarly gross kinds of stimulation.Thus the popular narrative that this pathwayalone—that is, without cognitive involvement—could generate a genuine emotion of fear of asnake may be untenable. Similarly untenable isthe idea that human emotions, economic deci-sions, and political preferences take place with-out cortical involvement (see Clore, Storbeck,Robinson, & Centerbar, 2005; Davidson,2003). However, even if this particular path-way did not turn out to be as important as ini-tially believed, and despite the fact that anysubcortical process is insufficient to accountfor emotion, this work has stimulated impor-tant additional research and thinking about thecritical role of subcortical processes in emo-tion.

Increasingly, it is apparent that behavior-relevant processing in the brain is highly recur-sive (Storbeck & Clore, 2007). Incoming sen-sory information is progressively refined in an


iterative process (Cunningham & Zelazo,2007). The kinds of subcortical processes thatLeDoux and colleagues’ work has highlightedpresumably serve as early signals that some-thing should be processed further, so that itssignificance can be determined. The pervasive-ness and importance of these processes meanthat many existing appraisal theories are in-complete. We discuss in the next section twoaccounts that do include low-level, reflex-likeprocesses at the beginning of the emotion se-quence.

Is Appraisal Sequential?

Charles Osgood (e.g., Osgood, Suci, & Tanne-baum, 1957) proposed a theory of connotativemeaning that was in many ways an affectiveappraisal theory. He proposed (and producedrelevant evidence) that all words in all lan-guages convey connotative meaning alongthree dimensions—“evaluation,” “potency,”and “activity” (the dimensions of E, P, and A).Osgood (1969) was interested in the origins oflanguage and communication, and he specu-lated about what kinds of informationprelinguistic humans would have needed mostto communicate. He reasoned that the mostimportant information for survival would havebeen whether something was good or bad,whether it was strong or weak, and whether itsapproach was fast or slow. Once one could lo-cate something within this E-P-A space, copingbehavior would be appropriately constrained.His idea was that triangulations from theseconnotative dimensions would have allowedpeople to make behavior-relevant distinctionsbetween saber-toothed tigers and mosquitoes,and to communicate these distinctions to oth-ers.

Scherer (1984, 2001) echoes the importanceof E, P, and A, and also treats them as dimen-sions of emotional meaning. He suggests that Eis linked to appraisals of stimuli with regard togoals and needs, P to appraisals of coping po-tential, and A to appraisals of urgency. He hasproposed a series of “stimulus evaluationchecks” believed to underlie stimulus apprais-al. The type and intensity of any resulting emo-tion, then, reflect the profile of results of theappraisal process based on these stimulus eval-uation checks.

The stimulus evaluation checks proposed byScherer begin with a (1) “novelty check,” fol-lowed by (2) an “intrinsic pleasantness check,”

based on innate feature detectors and learnedassociations; (3) a “goal/need significancecheck,” evaluating whether an event is relevantto goals, conducive to goals, expected, and ur-gent; (4) a “coping potential check,” evaluatingcausation, coping potential, control over con-sequences, relative power, and options for in-ternal adjustment; and finally (5) a “norm–selfcompatibility check,” evaluating the compati-bility of actions or events with social norms,conventions, or expectations of others, as wellas with internalized norms or standards of self.Scherer assumes that the outcomes of thesechecks change various subsystems that serveemotion (physiology, expression, motivation,feelings), creating a telltale trace that is theemotion. In addition, he emphasizes that emo-tions are fluid, reflecting constant evaluativeactivity.

In a somewhat related way, the music cogni-tion theorist David Huron (2006) has analyzedthe emotion process and suggested six stages ofemotion elicitation in response to auditorystimulation. He proposes that the process maystart with various (1) “reflexive responses,” in-cluding the orienting response, the startle re-flex, defensive reflex, and reflex-like reactionsbased on various overlearned perceptualschemas. Thus an unexpected bang of a door ismarked physiologically by flexing of the shoul-der muscles and by the release of epinephrineand norepinephrine into the bloodstream,which cause increases in heart rate and respira-tion, sweating, pupil dilation, and so on. Thesereactions may facilitate perceptual intake andprotective action. At the same time, (2) “deno-tative responses” allow stimulus identification(e.g., “slamming door”) on the basis of passiveassociations; (3) “connotative responses” arealso passively learned processes using the phys-ical properties of timing, energy, proximity, andso on to determine what the sound is like (e.g.,“forceful,” “loud”). The reflexive, denotative,and connotative responses are all fast and auto-matic. (4) “Associative responses” are arbi-trary, learned associations that may activate anemotional response on the basis of memory(e.g., “That reminds me of my dad slammingdoors when he was upset”). (5) “Empatheticresponses” identify whether a sound was gen-erated by an animate agent and what state ofmind is signaled by the sound (e.g., anger).Finally, (6) “critical responses” are conscious,cognitive processes that evaluate the intentionsor sincerity of the agent; these may also involve


self-monitoring processes concerning the ap-propriateness of one’s own response.

These two accounts of the appraisal process(Scherer’s and Huron’s) have in common theidea that low-level and higher-level processesare both operative, rather than being alterna-tives. However, in these models reflexes, suchas the orienting or startle reflex, provide thelow-level impetus to emotion. Some other low-level affective or evaluative reactions are dis-cussed in the next section.

How Basic and Broadly DistributedIs Affect?

Although appraisal theories concentrate oncognitive distinctions, emotion does not startthere. The process often begins as a very low-level affective reaction—a reaction that is notyet an emotion (Barrett, 2006a; Baumeister etal., 2007; Berkowitz & Harmon-Jones, 2004;LeDoux, 1996; Ortony et al., 2005). Most cog-nitive appraisal theories have not included astage for such early affective reactions.

Although full-blown emotional states in-volve multiple components, including cogni-tion, what gets the emotional ball rolling issometimes quite low-level. For example, peopleapparently respond more positively to smooth,curved objects than to objects with sharp-angled edges (Bar & Neta, 2006). That prefer-ence also holds for roundish as opposed toangular faces (Zebrowitz, 1997). Some havehypothesized that sharp shapes may conveythreat, and round shapes warmth (Aronoff,Woike, & Hyman, 1992).

In a related vein, lightness and darkness ap-pear to have reliable affective values that mayhave both universal application and ancient or-igins (Meier, Robinson, & Clore, 2004). Highversus low physical location (Meier & Robin-son, 2004) and high versus low pitch in musicor speech (Huron, 2006) have a similarevaluative impact. Each of these stimulus char-acteristics may have an associated valence fordifferent reasons. It seems likely, for example,that young children’s preference for sweetnessand aversion to bitterness may be innate,whereas some other quite common preferencesmay be learned.

Beyond such specific preferences, familiarstimuli generally elicit more positive reactionsthan do novel stimuli—a preference that ap-pears to be a general design feature of verte-

brates. Moreover, this process is evident at thegroup level, and even at the cellular level.2 The“minimal-groups” effect is a highly reliablesocial-psychological effect (Tajfel, Billig,Bundy, & Flament, 1971). Any distinctionamong a collection of individuals, no matterhow arbitrary, leads to a surprising degreeof ingroup favoritism and outgroup disfavor.Such a principle is presumably also at work inxenophobia, racism, and religious intolerance.However, our point is that the criterion of “likeme” versus “not like me” may be a very basicprinciple of evaluation.

A more general statement about the lower-level nature of affective processes was recentlymade by Buck (March 4, 2007):

Brains, after all, are only about 0.6 billion yearsold, compared to the 3.5 billion year history of lifeon the earth; and . . . we have about 40% of ourgenes in common with microbes. . . . I think it isof significance that one can find in microbes genesthat encode for dopamine, serotonin, nor-epinephrine, ACTH [adrenocorticotropic hor-mone], many of the peptide neurohormones, etc.dating from long before the evolution of the brain.This suggests that prototypical motivational–emotional systems are design principles in themost elemental life forms.

What are the implications of such observa-tions? What does it mean that processes analo-gous to those of social groups are alreadypresent at the cellular level, and that the neuro-chemistry of evaluation may be widely distrib-uted among animate organisms down to thelevel of microbes? Since no one seriously cham-pions the idea of microbe emotions, a distinc-tion must be drawn between affective processesand emotional states.

Are Appraisals Sequential, Dual, Chaotic,or Recursive?

We have discussed one kind of dual-processmodel and two sequential models of emotionalappraisal. We turn next to dual-process mod-els, and alternatives to such models, that specif-ically distinguish affect from emotion.

Baumeister, Vohs, DeWall, and Zhang

A dual-process view proposed by Baumeister etal. (2007) not only specifies two processes, butsuggests how they interact. Drawing on litera-


ture reviewed by Schwarz and Clore (2007),they note that although there is abundant evi-dence for the influence of emotion on cogni-tion, there is scant evidence for a direct influ-ence of emotion on behavior. They proposethat behavior is controlled in a bottom-up wayby unconscious affect, in a manner similar tothat demonstrated by Winkielman, Berridge,and Wilbarger (2005) in their studies of howunconscious priming with happy faces stimu-lated thirsty people to drink more of a novelbeverage.

Baumeister et al. suggest that full-blown,conscious emotions are re-representations orconstructions of affectively significant situa-tions for the purpose of remembering the les-sons of those situations. They argue that hu-man social life is vastly more complex than thatof any other species, and that it requires a cor-responding richness and variety of emotionalrepresentation. In their view, emotion is anelaborated, conscious state that is memorableand hence useful for self-instruction. This theo-retical maneuver of drawing a sharp distinctionbetween affect and emotion strikes us as a use-ful one. In addition, specifying emotion as ahigh-level state that does not drive behav-ior, but that provides information to theexperiencer, may help resolve some of the in-herent conflicts between cognitive andnoncognitive approaches to affect and emo-tion. On the other hand, some theorists havealso suggested useful alternatives to dual-process accounts, as we see next.

Barrett, Ochsner, and Gross

In some accounts, both of the processes of adual-process model are handled in a single-network model (e.g., Barrett et al., 2006). Insuch a model, a psychological process may berepresented by activation distributed acrossmultiple nodes. Network models can either be“localist” (e.g., Thagard & Erb, 2002) or re-flect “parallel distributed processing” (Wager& Thagard, 2004). In the former, each nodemay correspond to a given emotion or emotioninstance; in the latter, the nodes may corre-spond to elements of emotion, with an emotionemerging from their joint action. Networkmodels operate through multiple-constraintsatisfaction. Each item of information in a net-work may constrain other items, such that theoverall state of the network at any given mo-

ment is emergent from these multiple con-straints. With respect to appraisal theory, thesolution of the multiple constraints would be aspecific emotion.

Rather than contending that such modelsrepresent a category of processing that is eitherautomatic (associative) or controlled (rule-governed), Barrett et al. (2006) envision a con-tinuum along which a given solution mightrepresent a given combination or partiallyautomated reaction. Processes combinecomponentially so that a given processingevent is conditionally automatic (Bargh, 1997),in that it falls on the continuum from auto-matic to controlled. This approach is compati-ble with the notion of emotion emergence dis-cussed earlier. If we think of the nodes as brainregions or perhaps as circuits corresponding toways of representing evaluation, then a partic-ular emotion would emerge as the best-fittingsolution to the constraints of the currently ac-tive goals, attitudes, perceptions, knowledge,and situational parameters. The funneling to-ward a solution presumably can take place veryrapidly and involves both top-down andbottom-up processing. As the bottom-up pro-cesses of constraint satisfaction take place, onemay feel visited by an emotion, whereas whenone engages in imaginative constructions ofemotional events, one may feel more like theauthor of one’s emotions. In this view, auto-matic and controlled processes have differentfunctions within the same system, rather thanbeing two different systems.

Ortony, Norman, and Revelle

A related view has been proposed by Ortony etal. (2005). They think of emotions similarly asinterpreted affect or affect with a cognitive–perceptual frame. Their view is that feeling isundifferentiated positive or negative affect, andthat specific emotions are transformations offeeling by appraisal. Appraisals, which aresometimes conscious but more often uncon-scious, situate and make sense of affect. Feelingis generated quite automatically, and its cogni-tive or perceptual framing may also generally beautomatic. People are necessarily aware of feel-ing, in that the idea of “unfelt feeling” involves acontradiction in terms, but such awareness neednot extend to the causes of feeling. Thus emo-tions in these views are cognitively elaboratedstates of affective feeling.


Cunningham and Zelazo

Finally, still another alternative to dual-processmodels is an approach in which levels of pro-cessing are iterative. Cunningham and Zelazo(2007) suggest that exposure to an object initi-ates an iterative sequence of evaluativeprocesses—“the evaluative cycle.” In this cycle,stimuli are interpreted and reinterpreted inlight of an increasingly rich set of contextuallymeaningful representations. Evaluations basedon few iterations of the cycle may be uncon-scious and automatic, whereas those based onadditional iterations become relatively reflec-tive. Thus implicit evaluations have fewer itera-tions and recruit fewer processes than explicitones.

Cunningham and Zelazo propose that afterinitial affective reactions fire, sensory informa-tion may be reprocessed. Then, after more de-tailed stimulus identification, the informationis again sent to the amygdala. One’s own vis-ceral reactions may also be reprocessed, so thatthe autonomic state also becomes corticallyrepresented. At each stage, the amygdala maybe used again, reacting to ever more detailedinformation at each iteration. In general, then,these authors view information processing as aseries of recursive feedback loops that involveadditional regions of the cortex as the processcontinues. With continual interaction of limbicand cortical areas, evaluations that start out asautomatic become situated and progressivelyrefined. In short, they become emotions.

Such an iterative model helps sharpen ournotion of implicit and explicit processes. Peo-ple tend to think of implicit emotional pro-cesses and attitudes as unconscious versions ofwhat they see in conscious, explicit versions.Thus when research (Phelps et al., 2000)showed that Implicit Association Test (IAT)measures of racial attitudes were related toamygdala activation of white participants in re-sponse to black faces, people might assumethat these reactions were unconscious versionsof explicit and fully formed attitudes or preju-dices. Indeed, in studies of racism, some writersseem to view conscious awareness solely as reg-ulating or suppressing fully formed attitudeslurking within.

However, an iterative model suggests a dif-ferent account. The amygdala is sensitive tonovelty and violations of expectation. For mostwhite research participants, black faces arenon-normative. The presented image is likely

to be processed and reprocessed, and eachtime, the amygdala receives a more and moredifferentiated form of the same information.Cortical processing, then, is not simply regula-tory, but also helps define the reaction. The ex-plicit, fully elaborated attitude is probably notthe same as whatever is reflected in responsetimes on the IAT, and the explicit, fully elabo-rated emotion is also not the same thing as ini-tial subcortical and neurochemical reactions ofaffect. They are the same thing only in the sensethat the block of marble that Leonardo daVinci selected for his statue of David was thesame thing as the statue that emerged from it.Both are made of the same material, but the lat-ter has a very different form as a result of beingprocessed and reprocessed many times. In asimilar way, affect and emotion are made of thesame stuff, but they have very different formsas a result of similarly iterative processing.

SUMMARY AND CONCLUSIONS

In this chapter we have focused on the processrather than the structure of emotion appraisal.We have argued that some kind of appraisal orevaluation is a necessity, since emotions are in-herently about various kinds of goodness orbadness. The issue of primary interest concernshow such evaluations are made. In the 1980s,Lazarus (1984) and Zajonc (1984) arguedabout whether affect and emotion requiredcognitive appraisals or not. It was an exchangethat was more heated than illuminating, be-cause critical terms were often used in differentways. In the end, both Lazarus and Zajoncwere correct, but they were talking about dif-ferent things. In general, cognitive theoristshave focused on full-blown emotional states in-volving subjective experience, whereas criticshave often focused on low-level, nonconscious,automatic processes. Both believe that they areexplaining emotion, but it might be more accu-rate to say that the latter are studying undiffer-entiated affect (Ortony et al., 2005), whereasthe former are studying emotion. If so, then animportant task is to ask how these fit into a sin-gle processing model. Leading up to that task,we have asked questions about the nature ofemotion as well as of appraisal and the apprais-al process.

Much research has been inspired by the as-sumption that emotional life issues from asmall number of basic emotions, which are de-


fined by distinctive physiology and neurology,and are marked by distinctive feelings, expres-sions, and actions. A failure to find the kind ofcoherence implied by that model might implyeither a limitation of method or of conceptual-ization. We have focused on alternative possi-ble conceptualizations both of emotions and ofthe appraisal processes that differentiate them.In one model, the emotions are treated as emer-gent states from partially redundant affectiverepresentations across multiple components.We have contrasted this emergence model, inwhich physiology, expression, and cognitionare constituents of emotion, with the tradi-tional latent-trait model, in which these are in-dicators of an underlying emotion. Taking theview of these as constituents, we end up inagreement with William James’s dictum thatwe are “angry because we strike” and “afraidbecause we tremble.”

In an examination of the idea of appraisal,we have suggested that the concept of appraisalshould probably be expanded to reflect the factthat appraisals are often as much perceptual ascognitive, and as much dictated by the topogra-phy of situations as by mental action. On thebasis of Sloman’s (1996) distinction betweenassociative and rule-based reasoning, we havesuggested that emotions can arise either fromrule-based processing or by reinstatement.

We have also reviewed the impact ofLeDoux’s (1996) proposal of a low route toemotion without cortical involvement, andhave cited critical reviews that cast doubt onthe relevance of this particular pathway for hu-man emotion. However, we have noted that thework in this area has done much to stimulatethe study of subcortical contributions to emo-tion.

We have then reviewed two sequence modelsof emotional appraisal (Huron, 2006; Scherer,2001), which propose that emotional processesare often initiated by reflexes such as the startleor orienting reflex, which activate appraisalprocesses. In the next section, we have re-viewed low-level affective reactions, includingpreferences for curved versus jagged lines, andthe evaluative implications of lightness anddarkness and of high versus low pitch. We havealso noted reactions to novelty versus identitythat are present even at the cellular level, andthat are perhaps continuous with behavioralreactions at the level of human groups, suggest-ing a surprising continuity of affective pro-cesses.

The observations of amazingly low-levelaffect-like processes make it clear that theoristsmust distinguish emotions from the affectivereactions that are their seeds. Hence we havereviewed several models that make this distinc-tion one way or another, including a dual-process model (Baumeister et al., 2007), aparallel-constraint model (Barrett et al., 2006),a sequence model (Ortony et al., 2005), and aniterative-process model (Cunningham &Zelazo, 2007).

In the end, the model of emotion and emo-tional appraisal that we have entertained is dif-ferent in many respects from the receivedmodel. It emphasizes emotions as emergentconstructions rather than as latent entities; itmakes a sharp distinction between affective re-actions and emotions; and it sees appraisal asan iterative process. Reflexes and low-level af-fective reactions often get the emotional ballrolling. These undifferentiated states are thenrefined, situated, further evaluated, and re-represented. The results are the rich andnuanced emotional states that mark the impor-tant occasions and turning points in people’slives, that embody people’s aspirations andfears, and that are capable of motivating theirbest and worst actions.

ACKNOWLEDGMENTS

Support is acknowledged from National Institute ofMental Health Grant No. MH 50074 and NationalScience Foundation Grant No. BCS 0518835.

NOTES

1. The writing of this section benefited greatly from dis-cussions with James Coan and Lisa Barrett.

2. Lydia Wraight (personal communication, November2006) has pointed out that a preference for the famil-iar is evident in the cells of vertebrates, which expressproteins that allow mutual recognition among, andpreference for the cells of, a given individual.

REFERENCES

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall,T. (1978). Patterns of attachment. Hillsdale, NJ:Erlbaum.

Anderson, J. R. (1983). The architecture of cognition.Cambridge, MA: Harvard University Press.

Arnold, M. B. (1960). Emotion and personality (2vols.). New York: Columbia University Press.


Aronoff, J., Woike, B. A., & Hyman, L. M. (1992).Which are the stimuli in facial displays of anger andhappiness?: Configurational bases of emotion recog-nition. Journal of Personality and Social Psychology,62, 1050–1066.

Bar, M., & Neta, M. (2006). Humans prefer curved vi-sual objects. Psychological Science, 17, 645–648.

Bargh, J. (1997). The automaticity of everyday life: Amanifesto. In R. S. Wyer (Ed.), Advances in socialcognition (pp. 1–61). Mahwah, NJ: Erlbaum.

Barrett, L. F. (2006a). Emotions as natural kinds? Per-spectives on Psychological Science, 1, 28–58.

Barrett, L. F. (2006b). Solving the emotion paradox:Categorization and the experience of emotion. Per-sonality and Social Psychology Review, 10, 20–46.

Barrett, L. F., Ochsner, K. N., & Gross, J. J. (2006). Onthe automaticity of emotion. In J. Bargh (Ed.), Socialpsychology and the unconscious: The automaticity ofhigher mental processes. New York: PsychologyPress.

Baumeister, R. F., Vohs, K. D., DeWall, C. N., & Zhang,L. (2007). How emotion shapes behavior: Feedback,anticipation, and reflection, rather than direct causa-tion. Personality and Social Psychology Review, 11,167–203.

Berkowitz, L. (1990). On the formation and regulationof anger and aggression: A cognitive–neoassocia-tionistic analysis. American Psychologist, 45, 494–503.

Berkowitz, L., & Harmon-Jones, E. (2004). Toward anunderstanding of the determinants of anger. Emo-tion, 4, 107–130.

Bollen, K. A., & Lennox, R. (1991). Conventional wis-dom on measurement: A structural equation perspec-tive. Psychological Bulletin, 110, 305–314.


Brown, J. S., Collins, A., & Duguid, S. (1989). Situatedcognition and the culture of learning. EducationalResearcher, 18(1), 32–42.

Buck, R. (2007, March 4). Message posted to listserv ofthe International Society for Research on Emotions.Postings may be requested from [email protected]

Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heu-ristic and systematic processing within and beyondthe persuasion context. In J. S. Uleman & J. A. Bargh(Eds.), Affect and social behavior (pp. 152–206).New York: Cambridge University Press.

Clark, A. (1997). Being there: Putting brain, body, andworld together again. Cambridge, MA: MIT Press.

Clore, G. L., & Centerbar, D. (2004). Analyzing anger:How to make people mad. Emotion, 4, 139–144.

Clore, G. L., & Ortony, A. (2000). Cognitive in emo-tion: Never, sometimes, or always? In L. Nadel & R.Lane (Eds.), The cognitive neuroscience of emotion(pp. 24–61). New York: Oxford University Press.

Clore, G. L., Schwarz, N., & Conway, M. (1994). Affec-tive causes and consequences of social information

processing. In R. S. Wyer & T. Srull (Eds.), Hand-book of social cognition (2nd ed., pp. 323–417).Hillsdale, NJ: Erlbaum.

Clore, G. L., Storbeck, J., Robinson, M. D., &Centerbar, D. (2005). Seven sins of research on un-conscious affect. In L. F. Barrett, P. Niedenthal, & P.Winkielman (Eds.), Emotion and consciousness(pp. 384–408). New York: Guilford Press.

Cunningham, W. A., & Zelazo, P. D. (2007). Attitudesand evaluations: A social cognitive neuroscience per-spective. Trends in Cognitive Sciences, 11, 97–104.

Davidson, R. J. (2003). Seven sins in the study of emo-tion: Correctives from affective neuroscience. Brainand Cognition, 52, 129–132.

Ekman, P. (1984). Expression and the nature of emo-tion. In K. Scherer & P. Ekman (Eds.), Approaches toemotion (pp. 319–343). Hillsdale, NJ: Erlbaum.

Freud, S. (1911). The interpretation of dreams (3rd ed.,Trans A. A. Brill).New York: Macmillan.

Frijda, N. H. (1986). The emotions. New York: Cam-bridge University Press.

Gibson, J. J. (1979). The ecological approach to visualperception. Boston: Houghton Mifflin.

Huron, D. (2006). Sweet anticipation: Music and thepsychology of expectation. Cambridge, MA: MITPress.

James, W. (1890). Principles of psychology. New York:Holt.

Kotovsky, L., & Baillargeon, R. (1994). Calibration-based reasoning about collision events in 11-month-old infants. Cognition, 51, 107–129.

Lazarus, R. S. (1966). Psychological stress and the cop-ing process. New York: McGraw-Hill.

Lazarus, R. S. (1984). On the primacy of cognition.American Psychologist, 39, 124–129.

Lazarus, R. S. (1994). Universal antecedents of the emo-tions. In P. Ekman & R. J. Davidson (Eds.), The na-ture of emotion: Fundamental questions (pp. 163–171). New York: Oxford University Press.


LeDoux, J. E., Romanski, L., & Xagoraris, A. (1989).Indelibility of subcortical emotional memories. Jour-nal of Cognitive Neuroscience, 1, 238–243.

Lindquist, K., Barrett, L. F., Bliss-Moreau, E., & Rus-sell, J. A. (2006). Language and the perception ofemotion. Emotion, 6, 125–138.


Meier, B. P., & Robinson, M. D. (2004). Why thesunny side is up: Associations between affect andvertical position. Psychological Science, 15, 243–247.

Meier, B. P., Robinson, M. D., & Clore, G. L. (2004).Why good guys wear white: Automatic inferencesabout stimulus valence based on color. PsychologicalScience, 15, 82–87.

Mischel, W. (1968). Personality and assessment. NewYork: Wiley.


Moors, A. (2006, April 18). Investigating the auto-maticity of constructive appraisals. Paper presentedat the 18th European Meeting on Cybernetics andSystems Research, Vienna.

Moors, A., & De Houwer, J. (2001). Automaticappraisal of motivational valence: Motivational af-fective priming and Simon effects. Cognition andEmotion, 15, 749–766.

Morgan, H. J., & Shaver, P. R. (1999). Attachment pro-cesses and commitment to romantic relationships. InJ. M. Adams & W. H. Jones (Eds.), Handbook of in-terpersonal commitment and relationship stability(pp. 109–124). New York: Plenum Press.

Needham, A., & Baillargeon, R. (1993). Intuitionsabout support in 4.5-month-old infants. Cognition,47, 121–148.

Oatley, K., & Johnson-Laird, P. N. (1987). Towards acognitive theory of the emotions. Cognition andEmotion, 1, 29–50.


Ortony, A., Norman, D. A., & Revelle, W. (2005). Af-fect and proto-affect in effective functioning. In J. M.Fellous & M. A. Arbib (Eds.), Who needs emotions?:The brain meets the machine (pp. 173–202). NewYork: Oxford University Press.

Osgood, C. E. (1969). On the whys and wherefores ofE, P, and A. Journal of Personality and Social Psy-chology, 12, 194–199.

Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957).The measurement of meaning. Urbana: University ofIllinois Press.

Parkinson, B. (2007). Getting from situations to emo-tions: Appraisal and other routes. Emotion, 7, 21–25.

Petty, R. E., & Cacioppo, J. T. (1986). The elabora-tion likelihood model of persuasion. In L.Berkowitz (Ed.), Advances in experimental socialpsychology (Vol. 19, pp. 123–205). New York: Ac-ademic Press.

Phelps, E. A., O’Connor, K. J., Cunningham, W. A.,Funayama, E. S., Gatenby, J. C., Gore, J. C., et al.(2000). Performance on indirect measures of raceevaluation predicts amygdala activation Journal ofCognitive Neuroscience, 12, 729–738.

Polti, G. (1977). The thirty-six dramatic situations.Boston: The Writer. (Original work published 1921)

Prinz, J. J. (2005). Emotions, embodiment, and aware-ness. In L. F. Barrett, P. M. Niedenthal, & P.Winkielman (Eds.), Emotion and consciousness(pp. 363–383). New York: Guilford Press.

Roseman, I. J. (1979, September). Cognitive aspects ofemotion and emotional behavior. Paper presented atthe 87th Annual Convention of the American Psy-chological Association, New York.

Roseman, I. J. (1984). Cognitive determinants of emo-tion: A structural theory. Review of Personality andSocial Psychology, 5, pp. 11–36.



Sabini, J., & Silver, M. (2005). Why emotion names andexperiences don’t neatly pair. Psychological Inquiry,16, 1–10.


Scherer, K. R. (1984). On the nature and function ofemotion: A component process approach. In K. R.Scherer & P. Ekman (Eds.), Approaches to emotion(pp. 293–317). Hillsdale, NJ: Erlbaum.

Scherer, K. R. (2001). Appraisal considered as a processof multilevel sequential processing. In K. R. Scherer,A. Schorr, & T. Johnstone (Eds.), Appraisal processesin emotion: Theory, methods, research (pp. 92–120).New York: Oxford University Press.

Scherer, K. R., Schorr, A., & Johnstone, T. (Eds.).(2001). Appraisal processes in emotion: Theory,methods, research. New York: Oxford UniversityPress.

Schwarz, N., & Clore, G. L. (2007). Feelings and phe-nomenal experiences. In A. K. Kruglanski & E. T.Higgins (Eds.), Social psychology: Handbook of ba-sic principles (2nd ed., pp. 385–407). New York:Guilford Press.

Shaver, P. R., & Clark, C. L. (1994). The psychodynam-ics of adult romantic attachment. In J. M. Masling &R. F. Bornstein (Eds.), Empirical perspectives on ob-ject relations theories (pp. 105–156). Washington,DC: American Psychological Association.

Siemer, M., & Reisenzein, R. (2007). The process ofemotion inference. Emotion, 7, 1–20.

Sloman, S. A. (1996). The empirical case for two sys-tems of reasoning. Psychological Bulletin, 119, 3–22.

Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cog-nitive appraisal. Journal of Personality and SocialPsychology, 48, 813–838.

Smith, C. A., Griner, L. A., Kirby, L. D., & Scott, H. S.(1996). Toward a process model of appraisal in emo-tion. In N. H. Frijda (Ed.), Proceedings of the NinthConference of the International Society for Researchon Emotion (pp. 101–105). Toronto: InternationalSociety for Research on Emotion.

Smith, C. A., & Kirby, L. D. (2001). Toward deliveringon the promise of appraisal theory. In K. R. Scherer,A. Schorr, & T. Johnstone (Eds.), Appraisal processesin emotion: Theory, methods, research (pp. 121–138). New York: Oxford University Press.

Smith, E. R., & Neumann, R. (2005). Emotion pro-cesses considered from the perspective of dual-process models. In L. F. Barrett, P. M. Niedenthal, &P. Winkielman (Eds.), Emotion and consciousness(pp. 287–311). New York: Guilford Press.


Storbeck, J., & Clore, G. L. (2007). On the interdepen-dence of cognition and emotion. Cognition and Emo-tion, 21, 1212–1237.

Storbeck, J., Robinson, M. D., & McCourt, M. E.(2006). Semantic processing precedes affect retrieval:The neurological case for cognitive primacy in visualprocessing. Review of General Psychology, 10, 41–55.

Tajfel, H., Billig, M. G., Bundy, R. P., & Flament, C.(1971). Social categorization and intergroup behav-iour. European Journal of Social Psychology, 1, 149–177.

Thagard, P., & Erb, J. (2002). Emotional gestalts:Appraisal, change, and the dynamics of affect. Per-sonality and Social Psychology Review, 6, 274–282.

Wager, B. M., & Thagard, P. (2004). Spiking PhineasGage: A neurocomputational theory of cognitive–

affective integration in decision making. Psychologi-cal Review, 111, 67–79.

Wegner, D. M., & Vallacher, R. R. (1986). Action iden-tification. In R. M. Sorrentino & E. T. Higgins (Eds.),Handbook of motivation and cognition: Vol. 1.Foundations of social behavior (pp. 550–582). NewYork: Guilford Press.

Winkielman, P., Berridge, K. C., & Wilbarger, J. (2005).Unconscious affective reactions to masked happyversus angry faces influence consumption behaviorand judgments of value. Personality and Social Psy-chology Bulletin, 31, 121–135.

Zajonc, R. B. (1984). On the primacy of affect. Ameri-can Psychologist, 39, 117–123.

Zebrowitz, L. A. (1997). Reading faces: Window to thesoul? Boulder, CO: Westview Press.


PA R T V I I

HEALTH AND EMOTIONS

C H A P T E R 4 0

Emotions and Health BehaviorA Self-Regulation Perspective

MICHAEL A. DIEFENBACH, SUZANNE M. MILLER,MATTHEW PORTER, ELLEN PETERS, MICHAEL STEFANEK,

and HOWARD LEVENTHAL

The study of emotions and health has had along history. To begin with, research was con-cerned with the influence of physiological reac-tions (later termed “stress”) to outside aversiveevents (Selye, 1951). Subsequent research ex-amined the influence of experienced negativeemotions on the body’s ability to fight infec-tions (Cohen & Wills, 1985). More recently,Leventhal and Patrick-Miller (2000) have ar-gued that emotions can be causes of healthstates as well as outcomes, and can even be in-dicators of health. Such relationships usuallyassume a direct pathway between health andemotions. However, it is also plausible to envi-sion indirect influences between emotions andhealth—for example, those in which emotionsinfluence health behaviors (e.g., screening forcancer), which in turn might influence healthstates.

Traditional health behavior theories, such asthe health belief model or the theory of plannedbehavior (Ajzen, 1985), have not been con-cerned with the influence of emotional stateson health behavior. Such theories stand incontrast to a self-regulation perspective, as

elaborated in the parallel-processing model(Leventhal, 1970) or the cognitive–socialhealth information processing (C-SHIP) frame-work (Miller, Shoda, & Hurley, 1996). In bothof these latter approaches, emotional states aregiven equal weight to the cognitive processingof a health threat. Still, the specific roles ofemotions in health cognitions and healthbehavior have not as yet been well described,compared to the body of literature that hasconcerned itself with the description and influ-ence of health-related beliefs and illness repre-sentations (Lau & Hartman, 1983; Meyer,Leventhal, & Gutmann, 1985; Miller,Diefenbach, Krantz, Baum, & Academy ofBehavioral Medicine Research, 1998).

This chapter explores the possible roles emo-tions may play in health cognitions and subse-quent health behaviors. In our effort to addressthis topic, we rely on a self-regulation perspec-tive and will also draw from other perspectives,including those from health, social, and cogni-tive psychology, as well as from the judgmentand decision literature. The chapter is struc-tured as follows. First, a historical perspective

645

relating results from the early fear studies (e.g.,Janis & Feshbach, 1953) to the emergenceof the cognitive–affective parallel–processingframework (Leventhal, 1970) is presented.Next, studies on negative emotions and breastcancer screening are described, to illustrate thecomplex relationship between emotions andhealth behavior. This description is followed bya brief discussion of cognitive models of deci-sion making, focusing primarily on a descrip-tion of heuristics, and their relationship to cog-nitive and affective threat representations ina self-regulation framework (Leventhal &Diefenbach, 1991; Miller et al., 1996, 1998).Following this, we return to the example ofbreast cancer screening, applying theory to re-solve divergent findings in that literature. Weconclude with a discussion of new areas of re-search emanating from our review; we focus inparticular on four functions of emotions de-rived from the decision-making literature andapplied to the self-regulation framework.

EARLY STUDIES ON EMOTIONSAND BEHAVIOR

It is worthwhile to briefly review the originalstudies on fear communications. Fear appeals(or communications) originated from the fear-drive reduction model advanced by Dollardand Miller (1950). The central hypothesis un-derlying this model was that fear acts as a moti-vating force prompting individuals to performrecommended behaviors, which in turn reducethe unpleasant fear state. This model generallyassumed that as the level of fear increases, sotoo does the likelihood of changes in behaviorand attitude. Interventions based on this modelconsisted of fear-arousing messages that pairedimages of the undesired outcome with a recom-mendation for new behaviors. For example,pictures of decaying teeth were presented withthe message to brush one’s teeth regularly.

One of the first experimental studies to testthis model was conducted by Janis andFeshbach (1953). Junior high school studentsreceived recommendations for oral hygiene(i.e., brushing one’s teeth three times a day)coupled with three different fear appeals. Thehigh-fear condition consisted of vivid picturesof decaying teeth and gums; the mild-fear con-dition consisted of less vivid pictures; and theno-fear condition had no pictures of decayingteeth. It was hypothesized that a positive linear

relationship between intensity of fear and aperson’s willingness for behavior change wouldbe found, such that higher levels of fear wouldbe associated with a greater likelihood ofbehavior change. Contrary to the researchers’expectations, students in the no-fear conditionwere more likely to change their dental hygienethan students in either the mild-fear or high-fear conditions. Janis and Feshbach attributedthese findings to defensive avoidance arousedamong students exposed to the fear messages(see also Janis, 1967).

A majority of subsequent studies, however,did not confirm these initial results or this in-terpretation. In studies that employed Janis andFeshbach’s basic experimental design, higherlevels of fear were consistently associated withstronger behavioral intentions or actual be-havior changes. One such study (Dabbs &Leventhal, 1966) exposed students to three lev-els of fear-arousing information about the con-sequences of failing to obtain a tetanus shot.Both intentions and actual inoculation be-havior correlated linearly with increased feararousal; that is, students with higher fear weremuch more likely to obtain a tetanus shot. Thisbasic pattern of results has been confirmed instudies using a number of paradigms, such assafe driving (Leventhal & Trembly, 1968;Rogers & Deckner, 1975), dental hygiene(Leventhal & Singer, 1966), and smoking ces-sation (Leventhal & Watts, 1966; Rogers &Thistlethwaite, 1970).

Another implication of these studies was therealization that the effects of fear appeals aretransient and short-lived (Leventhal & Niles,1965), and that intentions can best be trans-lated into behavior if an individual receives anaction plan. An “action plan” is a concrete setof instructions containing information abouthow, when, and where to execute a desired ac-tion in the context of the individual’s life (e.g.,getting a tetanus shot—Leventhal, Singer,& Jones, 1965; quitting smoking—Leventhal,Watts, & Pagano, 1967). Most importantly, itwas found that if an action plan was provided,the attitude change and desired behaviors weresustained over weeks or months (in the case ofsmoking cessation) beyond the immediate ef-fects of the fear appeal. The results suggestedthat neither fear (high or low) nor an actionplan alone was sufficient to result in behavioralchange; rather, the combination of fear and anaction plan resulted in changes to individuals’cognitive and emotional representations of a

646 VII. HEALTH AND EMOTIONS

threat, and these led in turn to changes inbehavior (Leventhal, 1970).

THE FIRST PARALLEL-PROCESSINGFORMULATION OF COGNITIONAND EMOTION

The realization that individuals’ threat repre-sentations are central to motivating behaviorhad two major implications. First, it led to re-search describing the content and nature ofthreat representations; second, it led to the de-velopment of a new theoretical framework, theparallel-processing model (Leventhal, 1970). Incontrast to the fear-drive reduction model,which assumed serial processing of cognitionand emotion (i.e., emotion follows cognition),this framework postulates the parallel process-ing of fear messages on both cognitive andemotional levels. Stimuli and potential re-sponses are evaluated and represented simulta-neously with both cognitions and emotions.The representations are situated within a pro-cessing network that is modified as new infor-mation enters.

The notion of “separate but equal” process-ing arms for cognition and emotion in a paral-lel framework raises the question of how thetwo arms interact to determine coping andappraisal behaviors. Much past research in thisvein has assumed the existence of a processingsystem heavily influenced by a cognitive ap-proach to information processing and decisionmaking (e.g., Simon, 1967; Kahneman &Tversky, 1982). Zajonc (1984) was a strongproponent of the importance of emotion. Henoted that although cognitions are tradition-ally viewed as the primary driver of decisions,emotions need not always be postcognitivephenomena. According to Zajonc, emotionscan be quite independent of cognitive opera-tions and can precede them in time. Indeed,some investigators have proposed that peopleuse an emotion heuristic for decision making;this means that individuals’ representations ofobjects or people are “tagged” with emotion(Finucane, Alhakami, Slovic, & Johnson,2000), and that individuals access this “affec-tive pool” in the process of making judgments.This is not unlike Damasio’s (1994) hypothesisthat thought is drawn largely from images con-sisting of symbolic representations. Individualsover time learn to “mark” these images as posi-tive or negative, based on personal experiences.

Hence, if a negative image is linked to an imageof a future outcome, it may serve as a barrier toa given action; if a positive image is associatedwith the outcome image, it may serve as an in-centive. This is important, as we shall see whendiscussing further developments of the parallel-processing framework, such as the common-sense model of illness behavior and the C-SHIPmodel (Miller et al., 1996).

THE EXAMPLE OF FEARAND BREAST CANCER SCREENING

Having reviewed the early literature on fearprocessing and the emergence of a psychologi-cal model to account for divergent experimen-tal findings, we turn to a more recent examplethat illustrates the complex relationship be-tween emotion and health behavior. We havechosen the example of how negative emotionalstates such as worry or fear affect breast cancerscreening, because it is one of the best-developed areas within the otherwise sparsebody of empirical studies on the relationshipbetween emotions and cancer screening.

The U.S. Department of Health and HumanServices (2000) report Healthy People 2010makes screening a cornerstone of cancer pre-vention and control efforts. It recognizes thatone of the most important weapons in the bat-tle against cancer is early detection throughscreening. Existing guidelines, endorsed by allmajor health institutes, recommend screeningfor breast, cervical, and colon cancer. Becausepatients’ uptake of these recommendations hasbeen variable, research on patients’ decisionsabout screening is of paramount importance,and this is an area where emotions play a keyrole.

We conducted broad searches of compre-hensive computerized databases (MEDLINE,PsycLIT, and CancerNet). This process identi-fied 21 relevant, quantitative, empirical studiesof breast cancer screening, published in Englishin peer-reviewed journals between 1990 and2006. All studies included quantitative mea-sures of worry or fear and actual screeningbehavior (not intentions to be screened).

“Cancer worry” or “cancer fear” is definedas a negative emotional state with specific ref-erence to cancer. It is often assessed with a four-item scale developed by Lerman and colleagues(Lerman, Trock, Rimer, Boyce, et al., 1991;Lerman, Trock, Rimer, Jepson, et al., 1991).

40. Emotions and Health Behavior 647

This scale not only measures the degree ofworry in response to cancer, but determines theimpact such worry has on daily functioning.An alternative approach is to use a single itemsuch as “When thinking about cancer, I amworried,” which is scored on a 4- or 5-pointLikert scale.

Results across the pool of 21 primary studieswere mixed and are summarized in Table 40.1.Thirteen studies reported a positive associationbetween anxiety and screening; that is, greateranxiety was associated with more screeningbehavior. Four studies reported a negative orinverse association; that is, greater anxiety wasassociated with less screening behavior. The re-maining four studies reported no relationshipbetween anxiety and screening behavior. Gen-erally, stronger relationships with screeningbehavior were demonstrated for cancer-specificanxiety than for trait anxiety (see Diefenbach,Miller, & Daly, 1999). In addition, severalstudies suggested that the relationship betweenanxiety and screening may be curvilinear, tak-ing the shape of an inverted-U curve—a rela-tionship where moderate anxiety is associatedwith greater screening, but very high and verylow levels of anxiety are associated with lessscreening (Bowen, Alfano, McGregor, &Andersen, 2004; Lerman et al., 1993;Schwartz, Taylor, & Willard, 2003). Substanti-ating such a pattern would require further re-search.

The diversity of results in part reflectsmethodological heterogeneity among thestudies. Primary studies differed in sampling(e.g., women with or without a family historyof breast cancer, women of different demo-graphic backgrounds), construct operational-ization (e.g., general vs. cancer-specific anxi-ety), and outcome (i.e., breast self-examination, clinical breast examination, ormammography attendance). Some of thestudies used a cross-sectional/retrospective de-sign, measuring both emotion and history ofscreening behavior at a single time point.Others employed a longitudinal/prospectivedesign, measuring emotion at baseline andsubsequent screening behavior at a later timepoint. Due to the quasi-experimental natureof all the investigations, establishing causalityis difficult. However, the longitudinal studiesat least establish that emotion precededscreening behavior, making a stronger argu-ment for the hypothesized causal relationshipbetween emotion and screening.

Based on the number and quality of studiesreviewed, we can cautiously support the notionthat cancer worry motivates breast cancerscreening behavior. Even so, the question ofhow to account for the different effects of can-cer worry on screening behavior remains. Toexplore why this may be so, we turn our atten-tion to the role of cognitive heuristics and bi-ases.

COGNITIVE HEURISTICS AND BIASESIN DECISION MAKING

Our knowledge of decision making and judg-ment under uncertainty has been heavily influ-enced by the study of cognitive heuristics andbiases (Kahneman & Tversky, 1982). The orig-inal work by Tversky and Kahneman (1981)described 12 cognitive biases related to subjec-tive judgment of probability, such as salienceand availability. The potential influence of themost common biases listed by Kahneman andTversky (1982) on cancer screening, preven-tion, or treatment decision making has notbeen systematically evaluated for their rele-vance; nor, obviously, have strategies beentested to ameliorate these biases. For example,according to the availability heuristic, knowl-edge that no one in one’s own family has hadany type of cancer associated with the humanpapillomavirus (HPV) could lead to a predispo-sition to judge HPV-related cancer as less likelyfor oneself (regardless of one’s own exposurehistory), thereby decreasing screening behavior.According to the salience heuristic, knowledgeof a friend or relative who endured a difficulttreatment regimen for breast cancer could leadto a predisposition to judge a diagnosis ofbreast cancer as more likely for oneself, duesimply to the greater cognitive salience of thatparticular information (which presumably re-sults from its emotional impact). In both exam-ples, screening decisions are based on cognitivebiases rather than objective reality. Yet theoperation of these cognitive biases in makingdecisions about cancer screening has not yetbeen systematically investigated; nor have in-terventions to correct them been developed ortested. The need to clarify the role of cognitivebiases in screening and treatment decisionmaking is heightened, because health care pro-viders often provide relatively complex inci-dence and prevalence statistics to patients whomay not be equipped to interpret them.



TABLE 40.1. Summary of Investigations of the Relationship between Anxietyand Breast Cancer Screening

Author DesignScreeningmodality Population Affect measure

Studies reporting a positive relationship

Diefenbach et al.(1999)

Prospective Mammogram FDRs of women withcancer (N = 213)

Breast cancer worry, 1item

McCaul et al.(1998)

Prospective BSE Women with (n = 65) andwithout (n = 70) familyhistory of breast cancer

Breast cancer worry, 4items

McCaul et al.(1996)

Prospective BSE,mammogram, CBE

Women from thecommunity (N = 353)

Breast cancer concerns,3 items

Tinley et al. (2004) Prospective BSE, CBE,mammogram

Women from BRCA1/2families (N = 112)

IES Intrusion Scale

Young & Severson(2005)

Prospective Mammogram African American womeneligible for federallyfunded cancer screeningprogram (N = 405)

Cancer fear, 11 items

Andersen et al.(2003)

Retrospective Mammogram Population sample (N =6,512), including somewomen with family history(n = 948)

Cancer worry, 5 items

Bowen et al.(2004)

Retrospective BSE,mammogram

Population-based (N =1,366) sample of women

Breast cancer worry, 4items; general anxietyand depression (BriefSymptom Inventory)

Burnett et al.(1999)

Retrospective BSE, CBE,mammogram

Women with family history(N = 139)

Breast cancer worry, 3items

Consedine et al.(2004)


Cluster sampling of olderwomen from six ethnicgroups (N = 1,364)

Cancer worry (CancerAttitude Inventory, 4items)

Epstein et al.(1997)

Retrospective BSE FDRs of women withcancer (N = 1,053);focused on 85 excessiveBSE performers

Cancer worry, 3 items

Gram & Slenker(1992)

Retrospective Mammogram Women invited to a freemammography screening(N = 1,349) in two cities

Breast cancer anxiety,recalled 1 year later

Lauver et al.(1999)


Patients at an urbanhospital who had not hada mammogram in the last13 months (N = 119)

Profile of Mood States

Stefanek & Wilcox(1991)


FDRs of breast cancerpatients (N = 125)

Breast cancer worry,no specific informationgiven

(continued)



Author DesignScreeningmodality Population Affect measure

Studies reporting a negative relationship

Schwartz et al.(2003)

Prospective Mammogram FDRs of women withbreast cancer (N = 159)

Breast cancer worry, 2items; general distressmental healthinventory

Lerman et al.(1993)

Retrospective Mammogram Women with family history(N = 140)

Breast cancer worry, 1item; IES IntrusionScale

Kash et al. (1992) Retrospective BSE, CBE,mammogram

Women enrolled in abreast protection program(N = 217)

Taylor ManifestAnxiety Scale; cancer-related anxiety andhelplessness scale

Lerman et al.(1994)a


Women < 50 years oldwith family history ofbreast cancer (N = 783)from three study sites

Breast cancer worry, 2items; psychologicaldistress; IES IntrusionScale

Studies reporting a null relationship

Aro et al. (1999) Prospective Mammogram Attenders and nonattendersof a mammographyscreening program (n‘s =946 and 641, respectively)

General anxiety anddepression

Sutton et al. (1995) Prospective Mammogram Community sample ofwomen 50–64 years old (N= 1,500)

State–Trait AnxietyInventory; GeneralHealth Questionnaire;7-item anxiety measure

Lindberg &Wellisch (2001)


Attenders and nonattendersof a mammographyscreening program (n‘s =191 and 174, respectively)

Breast cancer worry

Drossaert et al.(1996)

Retrospective Mammogram Women with and withoutfamily history of breastcancer (n‘s = 389 and 295,respectively)

Breast cancer anxiety;investigator-developedanxiety scale

Note. FDRs, first-degree relatives; BSE, breast self-examination; CBE, clinical breast examination; BRCA1/2, breast cancersusceptibility genes 1 and 2; IES, Impact of Events Scale.aThis study provides inconclusive results, as they vary by study site.

Although “virtually all current theories ofchoice under risk or uncertainty are cognitiveand consequentalist” (Loewenstein, Weber,Hsee, & Welch, 2001, p. 267), relatively few ofthese theories attempt to link the roles of cogni-tion and emotion. Schwarz (1990), for exam-ple, has pointed out that the anticipation ofemotion as a consequence of a decision is capa-ble of influencing the decision, as well as theformation of attitudes. Mellers and McGraw(2001) have found that anticipated pleasure isclosely connected to choice. They have pro-posed, as part of their theory of subjective ex-pected pleasure, that individuals predict thepleasurable emotion that could result fromeach potential outcome of a particular behav-ioral option, weight these by the perceivedchances of their occurrence, and combine themto form an average anticipated pleasure indexfor each behavioral option. This is done for allbehavioral options under consideration, andthen the option with the greatest average antic-ipated pleasure index is selected. A corollary ofthis theory is that individuals will avoid select-ing any option that could make them feelworse. The theory has not yet been applied tomedical decision situations, and thus it is un-known how accurate individuals are in predict-ing emotional consequences of future events(Loewenstein & Schkade, 1999; Ditto &Hawkins, 2005; Ubel, Loewenstein, Schwarz,& Smith, 2005).

A COMMON-SENSE MODEL OFCOGNITIVE–AFFECTIVE PROCESSING

As the discussion to this point has suggested,biased cognitive processing exists within alarger cognitive–affective framework. This ideais supported by research that attempted tocharacterize the cognitive attributes of a threatstimulus and led to the formulation of thecommon-sense model of illness representations(Meyer et al., 1985). At the beginning of thethreat identification and processing sequence isthe individual’s attempt to identify the threat,and thus to assign it a “label” (e.g., “I can feelsomething in my breast; is it cancer?”). The useof a label often determines subsequent attrib-utes, such as “timeline or duration,” “conse-quences,” “cause,” and “controllability” (Lau& Hartman, 1983). Each attribute answers adifferent question in the attempt to define the

unknown stimulus. Timeline or duration ad-dresses the perceived temporal progressionof the stimulus. It answers the question ofwhether the threat represents an acute condi-tion (e.g. a benign tumor), a chronic condition(e.g., cancer), or a condition that might disap-pear and reappear (e.g., an allergy). The conse-quence attribute defines the stimulus in termsof its potential impact on the individual’s over-all health—whether it is minor, major, or life-threatening. The cause attribute categorizes thestimulus in terms of potential factors thatmight have led to the stimulus. Controllabilityanswers the question of whether anything canbe done about it. Attributes of illness represen-tations are highly individualized and are notnecessarily consistent with medical knowledge.Although there is a lack of research on the in-fluence of cognitive biases and heuristics on therepresentational level, it is most likely that cog-nitive biases work at this initial encoding level.

The next step in the processing of the stim-ulus involves the selection of a coping re-sponse. Coping responses can be triggered asa result of a cognitive evaluation of thethreat, as a way to deal with the emotionalimpact of the threat, or as the result of aninteraction of the two. For example, copingwith a cold will probably prompt a person totake some over-the-counter cold medication—an act that involves only minimal emotionalprocessing. In contrast, the occurrence of apersistent stuffy nose with symptoms of achi-ness may lead the person to consider a moresevere diagnosis, triggering feelings of worryand anxiety. In that case, the coping responsemay be triggered as much to allay worryabout an unknown condition as it is to re-lieve the discomfort.

After selecting and executing a coping ac-tion, the individual next appraises the chosencoping response for its effectiveness. To con-tinue the prior example, if the cold symptomssubside after the person ingests the medication,the coping procedure has been successful. Ifhowever, the cold symptoms persist, the indi-vidual will have to change his or her initial di-agnosis to consider more severe illnesses, suchas a type of flu. This leads to a revision of therelevant illness representation attributes (i.e.,label, time line, consequences, cause, and con-trollability), triggering new negative feelingsand prompting the selection of other copingprocedures.


An often overlooked feature of the common-sense model is that coping requires not just therepresentation of the threat, but also the repre-sentation of coping procedures—the actionplan. Thus this approach integrates the earlierfinding from the fear communication literaturethat an action plan is an important addition tothe cognitive and emotional threat represen-tations. This approach leads to a coherent,action-oriented model, in which the threat isplaced in an action framework. Of course,the individual must appraise the available re-sponses as effective—and, more importantly,actions must seem necessary and feasible as away to limit the threat. If the responses are notadequate, an individual faces a health threatwithout an action plan, with a likely increase innegative emotional states.

COGNITIVE–SOCIAL HEALTHINFORMATION PROCESSING

The cognitive–affective approach has been ex-panded and applied to coping with cancerthreat and disease (Miller et al., 1996, 1998).The C-SHIP framework frees itself from aparallel-processing mechanism by postulatingthe existence of five main cognitive–affectivemediating units that process cancer-relevant in-formation. These units are interconnected andinclude many of the primary theoretical con-structs from existing health behavior theoriesdeveloped in the cognitive, social, and healthsciences. Specifically, the units consist of (1) theindividual’s encoding/self-construals of the can-cer threat (Champion, 1985, 1987, 1988; Laz-arus & Folkman, 1984; Leventhal, 1970,1983, 1990); (2) the individual’s cancer-relatedself-efficacy beliefs (Bandura, 1977) and out-come expectations (Bandura, 1986; Scheier &Carver, 1985); (3) the individual’s cancer-related values and goals (Lau & Hartman,1983); (4) the individual’s emotional responsesto cancer threats (Horowitz, 1991; Leventhal,1970); and (5) the individual’s repertoire ofself-regulatory/coping skills and strategies(Carver, Scheier, & Weintraub, 1989). Figure40.1 displays the interconnected network ofemotion and cognition at work in the exampleof making a decision about breast cancerscreening. Specifically, a woman may react toinformation about breast cancer risk, the effec-tiveness of mammograms, and screening rec-ommendations with increased risk perception

(Path 1), a belief that mammograms will beable to detect cancer in an early stage (Path 2),or the intention to adhere to a regular screeningregimen (Path 3). Beliefs of increased vulnera-bility to breast cancer may be followed by a re-lated belief that a lump may be found soon(Path 4), which will be cured by treatment(Path 8). However, to take advantage of earlytreatment, screening is necessary (Path 10), andthis realization may lead to the intention tohave regular mammograms. A woman mayalso react to information about breast cancerwith increased helplessness, anxiety, and de-pression (Paths 4 and 7). A belief that “a lumpwill be detected sooner or later” does not nec-essarily lead to action, as can be seen in Path11, which points to a predominantly negativeaffective reaction. It is possible, however, thatsome women use action in the form of screen-ing as a way to control their negative emotions.This possibility is accounted for in Paths 12and 13.

Thus, depending on the initial encoding ofan external or internal stimulus, different medi-ating units are activated that trigger individualbeliefs, values, and goals; outcome expecta-tions; and an emotional response to the stimu-lus. Ultimately, this processing will lead to theexecution or inhibition of behaviors related tocoping and adjustment to cancer risk and dis-ease.

The conceptualization of information pro-cessing within C-SHIP as a model of mediatingunits can be compared to the excitation ofnodes within the actual physiological nervoussystem. The mediating units, or nodes, are ex-cited or inhibited based on the information thatis processed through the connected nodes,eventually resulting in a behavioral action. Im-plicit in the C-SHIP framework is the devel-opment of information-processing signatures.These signatures are the result of repeated pro-cessing of similar information, resulting in sim-ilar responses. Thus the behavioral signaturesreflect a system’s tendency to react with similarbehavioral patterns, given that similar nodesare activated. In other words, these signaturescan be thought of as a sequence of “if–then”statements resulting in more or less consistentbehavioral responses. The identification ofsuch sequences or signatures will facilitate theaccurate prediction of individual responses tothreat messages and different health behaviorsby incorporating an individual’s habitual re-sponse style.


COGNITIVE AND EMOTIONALHEURISTICS WITHINA SELF-REGULATION FRAMEWORK

The self-regulation approach sees the patient asan active protagonist who is constantly evalu-ating his or her internal and external environ-ment to control health behavior. These evalua-tion behaviors can be executed on both anautomatic and a deliberate level. Research hasidentified a number of such mental rules orshortcuts in somatic processing and has con-nected them to the basic illness representationattributes. For example, the “symmetry” rule isused to link the somatic experience to the iden-

tity attribute or label. A patient experiencingsymptoms searches for a meaning-giving iden-tity to the symptomatic process. However, inthe case of an illness label, a patient is also in-clined to search for symptoms in support of hisor her own illness experience: symptoms arethere for a reason and have a name, and ill-nesses are characterized by a set of symptoms(Meyer et al., 1985). Other heuristics that helppatients to process somatic and experiential in-formation include the “stress–illness” heuristicand the “age–illness” heuristic. The stress–illness heuristic categorizes general somaticevents (e.g., fatigue, muscle aches, headaches)as either stress responses or, in the absence of


FIGURE 40.1. Example of hypothesized activation of cognition and affect in the C-SHIP model leadingto mammography screening. Thick arrows are activated paths; thin arrows are possible but nonactivatedpaths.

stressful events, as symptoms of an illness(Cameron, Leventhal, & Leventhal, 1995). Theage–illness heuristic similarly attributes somati-c events that fit with a perceived aging model toaging and not to illness (Prochaska, Keller,Leventhal, & Leventhal, 1987; Stoller, 1984).Illness attributions are made if the somaticevent significantly departs from the agingattribution—for example, if there is a sharppain that is distinguishable from regular ar-thritic pain. Other heuristics are concernedwith the “duration” of a somatic event (a moreenduring event is more serious and threatening)and its “prevalence.” Finally, the “affect” heu-ristic connects negative emotions, such as anx-ious and depressed moods, to increased diseasevulnerability (Salovey & Birnbaum, 1989).

Thus the self-regulation framework not onlyallows for the processing of the threat stimuluson the cognitive and emotional level, but alsopermits the formulation of some rules that gov-ern the emotional and cognitive processing ofhealth-threatening events. In addition, as inany complex regulatory system that strives forequilibrium, it provides for a feedback loopwith the potential to revise the underlyingcognitive and emotional representations, heur-istics, and subsequent selection of coping pro-cedures (see Figure 40.2). Such a feedbackmechanism increases the complexity of themodel—a cost that is offset by improved ex-planatory power for the processes hypothe-

sized to underlie health behaviors. Thus weargue that comprehensive models of healthbehavior need to include constructs of emo-tions as well as cognitive representations.Models such as the theory of planned behavior(Ajzen, 1985) and the health belief model,which lack such components, consequentlyhave limited explanatory utility, even thoughthey address cognitive factors as well as exter-nal societal factors. Similarly, single-variablemodels, such as the focus on self-efficacy(Bandura, 1977), have limited utility in predict-ing health behavior.

FUNCTIONS OF AFFECT1 ANDTHEIR ROLE IN DECISION MAKING

Although the role of emotion in decision mak-ing has received renewed attention over thepast several years, a formal description of thespecific roles of emotionally based heuristics islacking. A recent paper by Peters, Lipkus, andDiefenbach (2006) has begun to address thispoint by focusing on the functions of affect, aconcept closely related to emotion. Peters et al.argue that affect has four separable roles im-portant to health decision-making processes.First, affect can act as “information.” Thisfunction is best exemplified by the affect heu-ristic (Slovic, Peters, Finucane, & MacGregor,2005). That is, affect may serve as a cue for


FIGURE 40.2. Self-regulation or parllel processing model.

many important judgments, including proba-bility judgments (Slovic, Finucane, Peters, &MacGregor, 2002). Zajonc (1980) proposedthat emotional reactions to stimuli are oftenthe earliest reactions. Under the affect heuristicmodel, affect can be experienced first uponconsideration of a familiar technology, or it canbe the result of further information processing(LeDoux, 1996; Peters, Västfjäll, Gärling, &Slovic, 2006). In either event, affect then acts asinformation, guiding decisions and judgmentssuch as risk and benefit perceptions. The affectheuristic is substantially similar to models of“risk as feelings” and “mood as information”(Loewenstein et al., 2001; Schwarz & Clore,2003), and it has much in common with dualprocess theories discussed earlier (see, e.g.,Cameron & Leventhal, 2003; Leventhal, 1970;Leventhal, Diefenbach, & Leventhal, 1992;Epstein, 1994).

Whereas some theories focus exclusively onthe use of mood states in judgments (e.g.,Forgas, 1995; Schwarz & Clore, 2003), use ofthe affect heuristic is characterized by relianceon feelings attributed to an option or stimulusand experienced while considering it in judg-ments and decisions. Alhakami and Slovic(1994) proposed that the strength of positive ornegative affect associated with an activity (andexperienced while considering that activity)guides perceptions of its risks and benefits.Thus judgments about a technology such as anew medical treatment may be based not onlyon what people think about the treatment, butalso on how they feel about it. If feelings re-lated to a technology are more positive, deci-sion makers tend to judge its risks as low andits benefits as high; if their feelings are morenegative, they tend to judge the opposite—risksas high and benefits as low. For example, vir-tual colonoscopy is currently under much scru-tiny for the detection of colon cancer. Individ-uals with positive affect toward this technology(e.g., because they have learned that it is not in-vasive or that it is less embarrassing than actualcolonoscopy) may interpret new informationabout risks and benefits in ways that are con-sistent with their affect (i.e., they may perceiveit as having low risk and high benefit).

Considerably less work has been done on theother three proposed functions of affect. Affectalso can act as a “spotlight,” focusing people’sattention on different information (e.g., numer-ical cues, risks vs. benefits), depending on theextent of their affect. First, the extent or type of

affective feelings (e.g., weak vs. strong affect,or anger vs. fear) focuses the decision maker onnew information. Second, the new information(rather than the initial feelings themselves) isused to guide the judgment or decision. Affectcan also serve as a “motivator” for action orthe processing of information. For example, itmay provide an action tendency toward gettinga mammogram. Alternatively, it could changethe extent of deliberative effort the decisionmaker puts forth. Finally, affect may serve as a“common currency” in judgments and deci-sions allowing people to compare more effec-tively the values of very different decision op-tions or information—in effect, to compareapples to oranges (Cabanac, 1992). Montagueand Berns (2002) link this notion to “neural re-sponses in the orbitofrontal–striatal circuitwhich may support the conversion of disparatetypes of future rewards into a kind of internalcurrency, that is, a common scale used to com-pare the valuation of future behavioral acts orstimuli” (p. 265). By translating more complexthoughts into simpler emotional evaluations,decision makers can compare and integrategood and bad feelings, rather than attempt tomake sense out of a multitude of conflictinglogical reasons. This function is thus an exten-sion of the affect-as-information function intomore complex decisions that require integra-tion of information. It predicts that emotionalinformation can often be more easily and effec-tively integrated into complex judgments thancognitive information can be.

INTEGRATING THE FUNCTIONSOF AFFECT INTOA SELF-REGULATION FRAMEWORK

We argue that the functions of affect should notbe examined as constructs by themselves, butshould be integrated into the larger body ofwork that supports the self-regulation frame-work. There is considerable overlap betweenthe roles assigned to emotion in the parallel-processing model and C-SHIP on the one hand,and the functions of affect on the other. Thus,from a theoretical point of view, a combinationof the two approaches seems readily achiev-able. Theoretical models need experimentalconfirmation, and a greater effort must bemade to examine the role and functions ofemotion in both laboratory and field studies.Examining the functions of emotion in medical


decision making would be a particularly fruit-ful approach. Medical decision making underuncertainty, such as the decision of whether toundergo mammography or genetic testing forsusceptibility to breast cancer, is fraught withemotions. Yet few investigations have exam-ined the role of emotions in such situations.

THE RELATIONSHIPBETWEEN FEAR AND BREASTCANCER SCREENING REVISITED

In our review of the literature on the role ofemotions in breast cancer screening behavior,we have found provisional support for a posi-tive relationship between negative emotionsand such behavior. Some of the diverging re-sults might be explained by differences in pop-ulations and measures of emotion, but none ofthese factors appears to consistently influencethe direction of the effect. Some researchers(Bowen et al., 2004; Lerman et al., 1993;Schwartz et al., 2003) suggest that the effectof worry may best be characterized by aninverted-U curve: Very low and very high levelsof worry may be associated with lower levels ofscreening behavior—the former because lowlevels emotion lack motivational force, and thelatter because levels that are too high act as abarrier. Although intuitively attractive, this in-terpretation of the data is less plausible in thecontext of the literature on fear studies and itsresulting theoretical developments.

The driving force behind a screening deci-sion is likely to be a combination of factors.As suggested by the self-regulation approach,these factors consist of individual threat, vul-nerability, and coping representations; emo-tional reactions; and societal and cultural de-terminants. In addition, research on heuristicsand biases point to processing pitfalls thatcould potentially influence the information-processing sequence. Research on heuristicsand biases within the health or cancer contextis sparse; most of it has been conducted in thelaboratory with non-health-related or non-cancer-related scenarios, and thus the influenceof these variables in health-related decision sit-uations is as yet unknown.

Many of the same points made about re-search focusing on cognitive variables and deci-sion making can also be made with regard toresearch on emotion variables and decisionmaking. Much of it lacks a health focus and is

therefore not readily transferable to decisionmaking in the health or cancer context. Fur-thermore, researchers focusing on emotionvariables often do so to the exclusion of cogni-tive variables, making it impossible to comparethe individual contributions of each group ofvariables to the decision process.

As we have demonstrated, negative emotionscan motivate, or in some cases can interferewith, cancer screening behavior. However, wedo not currently understand the potential rolethat positive emotions—or the absence of nega-tive emotions—may play. For example, arethere some circumstances when a feeling oftranquility coupled with an objective assess-ment of costs and benefits may increase screen-ing behavior? Or to what extent may screeningbehavior be prompted by the anticipation of re-lief from a current unpleasant anxiety stateupon receiving good news? Anticipating one’saffect in a particular situation is akin to roleplaying or the pre-living of a screening result.The role-playing/pre-living technique has re-cently been applied to facilitate decision mak-ing for genetic testing: At-risk individuals areasked to pre-live the emotional and cognitiveconsequences of both positive and negative ge-netic testing results (Diefenbach & Hamrick,2003; Miller et al., 2001). However, itis necessary to examine the information-processing flow within that context more care-fully, particularly the use and influence ofheuristics and biases during the pre-living se-quences. It is also critical to continue examin-ing the degree to which individuals using therole-playing/pre-living technique can accu-rately predict their response to situations thatmay arise in real life.

Researchers interested in the diverse path-ways between emotions and health shouldstrive to incorporate the complex findings re-garding emotions and health behavior thathave been derived from diverse research areasinto more comprehensive and integrative theo-retical frameworks. With the notable exceptionof the C-SHIP framework and the parallel-processing model, efforts in model buildinghave been sparse. However, model buildingshould not consist of including an increasingnumber of variables in ever more complexmodels. In contrast, we suggest that researchersrigorously evaluate variables and their hypoth-esized relationships to other factors in a givenmodel. These tests should be conducted bothexperimentally in laboratory settings that sim-


ulate the appropriate health contexts, as well asnaturalistically in the field, employing bothquantitative and qualitative methods. Onlythen can we be confident that our theoreticalmodels are valid for predicting health-relatedbehavior.

The role of emotions in cancer screening re-search needs further exploration. We hope thatthis chapter has sensitized researchers as wellas clinicians to the necessity of examining andincorporating emotion variables in their re-search programs. Furthermore, we hope thatcognitive researchers from areas that tradition-ally do not have a health focus will be per-suaded to adopt such an outlook in their futureresearch.

ACKNOWLEDGMENTS

This chapter was supported in part by grants fromthe National Institutes of Health (No. R01CA104979, and No. P30 CA06927 to the Fox ChaseCancer Center Behavioral Research Core Facility),the Department of Defense (Nos. DAMD 17-01-01-1-0238 and DAMD 17-02-1-0382), and the Ameri-can Cancer Society (No. TURSG 02-227-01). We areindebted to Joanne Buzaglo and Linda Fleisher fortheir valuable feedback, and to John Scarpato andMary Anne Ryan for their technical assistance.

NOTE

1. Though the distinction between “affect” and “emo-tion” is clouded by the imprecision of language, af-fect is here understood to be a “basic invariant build-ing block of emotional life derived from the humanmind’s capacity to engage in processes of valuation”(i.e., judging whether something is helpful or harm-ful; see also Russell & Barrett, 1999). It follows thataffect is tantamount to a felt experience of valence—the specific quality of goodness or badness that bothis experienced as a feeling state (with or without con-scious awareness) and demarcates a positive or nega-tive quality of a stimulus. Such feelings can be used asinformation for judgment and decision making(Schwarz & Clore, 1988), and this usage can betermed “the affect heuristic” (Slovic, Finucane, Pe-ters, & MacGregor, 2002). By contrast, emotions,such as cancer-related worry, are in this context takento be akin to what William James (1884) called “themore complicated cases in which a wave of bodilydisturbance of some kind accompanies the percep-tion of the interesting sights or sounds, or the passageof the exciting train of ideas” (p. 188). Both strong,visceral emotions such as fear and anger, and muchsubtler feelings—the “faint whisper[s] of emotion”

we call affects—can play a role in risk perceptionsand behavior (e.g., Lerner & Keltner, 2000; Slovic &Peters, 2006).

REFERENCES

Ajzen, I. (1985). From intentions to actions: A theory ofplanned behavior. In J. Kuhl & J. Beckman (Eds.),Action control: From cognition to behavior (pp. 11–39). Berlin: Springer-Verlag.

Alhakami, A. S., & Slovic, P. (1994). A psychologicalstudy of the inverse relationship between perceivedrisk and perceived benefit. Risk Analysis, 14, 1085–1096.

Andersen, M. R., Smith, R., Meischke, H., Bowen, D.,& Urban, N. (2003). Breast cancer worry and mam-mography use by women with and without a familyhistory in a population-based sample. Cancer Epide-miology, Biomarkers and Prevention, 12(4), 314–320.

Aro, A. R., de Koning, H. J., Absetz, P., & Schreck, M.(1999). Psychosocial predictors of first attendancefor organised mammography screening. Journal ofMedical Screening, 6(2), 82–88.

Bandura, A. (1977). Self-efficacy: The exercise of con-trol. New York: Freeman.

Bandura, A. (1986). Social foundations of thought andaction: A social cognitive theory. Englewood Cliffs,NJ: Prentice Hall.

Bowen, D. J., Alfano, C. M., McGregor, B. A., &Andersen, M. R. (2004). The relationship betweenperceived risk, affect, and health behaviors. CancerDetection and Prevention, 28(6), 409–417.

Burnett, C. B., Steakley, C. S., Slack, R., Roth, J., &Lerman, C. (1999). Patterns of breast cancer screen-ing among lesbians at increased risk for breast cancer.Women and Health, 29(4), 35–55.

Cabanac, M. (1992). Pleasure: The common currency.Journal of Theoretical Biology, 155, 173–200.

Cameron, L. D., Leventhal, E. A., & Leventhal, H.(1995). Seeking medical care in response to symp-toms and life stress. Psychosomatic Medicine, 57,37–47.

Cameron, L. D., & Leventhal, H. (2003). Self-regulation, health, and illness: An overview. In L. D.Cameron & H. Leventhal (Eds.), The self-regulationof health and illness behaviour (pp. 1–14). London:Routledge.

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989).Assessing coping strategies: A theoretically based ap-proach. Journal of Personality and Social Psychol-ogy, 56(2), 267–283.

Champion, V. (1985). Use of the health belief model indetermining frequency of breast self exam. Researchin Nursing and Health, 8, 373–379.

Champion, V. (1987). The relationship of breast self-examination to health belief model variables. Re-search in Nursing and Health, 10, 375–382.

Champion, V. (1988). Attitudinal variables related to in-


tention, frequency and proficiency of breast self-examination in women 35 and over. Research inNursing and Health, 11, 283–291.

Cohen, S., & Wills, T. A. (1985). Stress, social support,and the buffering hypothesis. Psychological Bulletin,98, 310–357.

Consedine, N. S., Magai, C., & Neugut, A. I. (2004).The contribution of emotional characteristics tobreast cancer screening among women from six eth-nic groups. Preventive Medicine, 38(1), 64–77.

Dabbs, J. M., Jr., & Leventhal, H. (1966). Effects ofvarying the recommendations in a fear-arousing com-munication. Journal of Personality and Social Psy-chology, 4, 525–531.

Damasio, A. R. (1994). Descartes’ error: Emotions, rea-son and the human mind. New York: Avon.

Diefenbach, M. A., & Hamrick, N. (2003). Self-regulation and genetic testing. In L. D. Cameron &H. Leventhal (Eds.), The self-regulation of health andillness behaviour (pp. 314–331). London: Routledge.

Diefenbach, M. A., Miller, S. M., & Daly, M. B. (1999).Specific worry about breast cancer predicts mam-mography use in women at risk for breast and ovar-ian cancer. Health Psychology, 18(5), 532–536.

Ditto, P. H., & Hawkins, N. A. (2005). Advance direc-tives and cancer decision making near the end of life.Health Psychology, 24(4, Suppl.), S63–S70.

Dollard, J., & Miller, N. (1950). Personality and psy-chotherapy: An analysis in terms of learning, think-ing and culture. New York: McGraw-Hill.

Drossaert, C. C., Boer, H., & Seydel, E. R. (1996). Per-ceived risk, anxiety, mammogram uptake, and breastself-examination of women with a family history ofbreast cancer: The role of knowing to be at increasedrisk. Cancer Detection and Prevention, 20(1), 76–85.

Epstein, S. (1994). Integration of the cognitive and thepsychodynamic unconscious. American Psychologist,49(8), 709–724.

Epstein, S. A., Lin, T. H., Audrain, J., Stefanek, M.,Rimer, B., & Lerman, C. (1997). Excessive breastself-examination among first-degree relatives ofnewly diagnosed breast cancer patients. Psycho-somatics, 38(3), 253–261.

Finucane, M. L., Alhakami, A., Slovic, P., & Johnson, S.M. (2000). The affect heuristic in judgments of risksand benefits. Journal of Behavioral Decision Making,13, 1–17.

Forgas, J. P. (1995). Mood and judgment: The affect in-fusion model (AIM). Psychological Bulletin, 11, 39–66.

Gram, I. T., & Slenker, S. E. (1992). Cancer anxiety andattitudes toward mammography among screening at-tenders, nonattenders, and women never invited.American Journal of Public Health, 82(2), 249–251.

Horowitz, M. J. (1991). Person schemas and maladap-tive interpersonal patterns. Chicago: University ofChicago Press.


Janis, I. L. (1967). Effects of fear arousal on attitude

change: Recent developments in theory and research.In L. Berkowitz (Ed.), Advances in experimental so-cial psychology (Vol. 3, pp. 166–224). New York:Academic Press.

Janis, I. L., & Feshbach, S. (1953). Effect of fear-arousing communications. Journal of Abnormal Psy-chology, 48(1), 78–92.

Kahneman, D., & Tversky, A. (1982). The simulationheuristic. In D. Kahneman, P. Slovic, & A. Tversky(Eds.), Judgement under uncertainty: Heuristics andbiases (pp. 201–208). New York: Cambridge Univer-sity Press.

Kash, K. M., Holland, J. C., Halper, M. S., & Miller, D.G. (1992). Psychological distress and surveillance be-haviors of women with a family history of breast can-cer. Journal of the National Cancer Institute, 84(1),24–30.

Lau, R. R., & Hartman, K. A. (1983). Common senserepresentations of common illnesses. Health Psychol-ogy, 2, 167–185.

Lauver, D. R., Kane, J., Bodden, J., McNeel, J., &Smith, L. (1999). Engagement in breast cancerscreening behaviors. Oncology Nursing Forum,26(3), 545–554.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal,and coping. New York: Springer.


Lerman, C., Daly, M., Sands, C., Balshem, A.,Lustbader, E., Heggan, T., et al. (1993). Mammogra-phy adherence and psychological distress amongwomen at risk for breast cancer. Journal of the Na-tional Cancer Institute, 85(13), 1074–1080.

Lerman, C., Kash, K., & Stefanek, M. (1994). Youngerwomen at increased risk for breast cancer: Perceivedrisk, psychological well-being, and surveillancebehavior. Journal of the National Cancer InstituteMonographs, 16(16), 171–176.

Lerman, C., Trock, B., Rimer, B. K., Boyce, A., Jepson,C., & Engstrom, P. F. (1991). Psychological andbehavioral implications of abnormal mammograms.Annals of Internal Medicine, 114(8), 657–661.

Lerman, C., Trock, B., Rimer, B. K., Jepson, C., Brody,D., & Boyce, A. (1991). Psychological side effects ofbreast cancer screening. Health Psychology, 10(4),259–267.

Lerner, J. S., & Keltner, D. (2000). Beyond valence: To-ward a model of emotion-specific influences on judg-ment and choice. Cognition and Emotion, 14(4),473–493.

Leventhal, H. (1970). Findings and theory in the studyof fear communications. In L. Berkowitz (Ed.), Ad-vances in experimental social psychology (Vol. 5,pp. 119–186). New York: Academic Press.

Leventhal, H. (1983). Behavioral medicine: Psychologyin health care. In D. Mechanic (Ed.), Handbook ofhealth, health care, and the health professions(pp. 709–743). New York: Free Press.

Leventhal, H. (1990). Emotional and behavioural pro-cesses. In M. Johnston & L. Wallace (Eds.), Stress


and medical procedures (pp. 3–35). Oxford: OxfordUniversity Press.

Leventhal, H., & Diefenbach, M. A. (1991). The ac-tive side of illness cognition. In J. A. Skelton & R.T. Croyle (Eds.), Mental representation in healthand illness (pp. 247–272). New York: Springer-Verlag.

Leventhal, H., Diefenbach, M. A., & Leventhal, E. A.(1992). Illness cognition: Using common sense to un-derstand treatment adherence and affect cognitioninteractions. Cognitive Therapy and Research, 16,143–163.

Leventhal, H., & Niles, P. (1965). Persistence of influ-ence for varying durations of exposure to threat stim-uli. Psychological Reports, 16, 223–233.

Leventhal, H., & Patrick-Miller, L. (2000). Emotionsand physical illness: Causes and indicators of vulner-ability. In M. Lewis & J. H. Haviland-Jones (Eds.),Handbook of emotions (2nd ed., pp. 523–537). NewYork: Guilford Press.

Leventhal, H., & Singer, R. P. (1966). Affect arousal andpositioning of recommendations in persuasive com-munications. Journal of Personality and Social Psy-chology, 4, 137–146.

Leventhal, H., Singer, R. P., & Jones, S. (1965). Effectsof fear and specificity of recommendations upon atti-tudes and behavior. Journal of Personality and SocialPsychology, 2, 20–29.

Leventhal, H., & Trembly, G. (1968). Negative emo-tions and persuasion. Journal of Personality, 36,154–168.

Leventhal, H., & Watts, J. C. (1966). Sources of resis-tance to fear-arousing communications on smokingand lung cancer. Journal of Personality, 34, 155–175.

Leventhal, H., Watts, J. C., & Pagano, F. (1967). Effectsof fear and instructions on how to cope with danger.Journal of Personality and Social Psychology, 6,313–321.

Lindberg, N. M., & Wellisch, D. (2001). Anxiety andcompliance among women at high risk for breastcancer. Annals of Behavioral Medicine, 23(4), 298–303.

Loewenstein, G. F., & Schkade, D. (1999). Wouldn’t itbe nice?: Predicting future feelings. In D. Kahneman,E. Diener, & N. Schwarz (Eds.), Well-being: Thefoundations of hedonic psychology (pp. 85–105).New York: Russell Sage Foundation.

Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch,N. (2001). Risk as feelings. Psychological Bulletin,127(2), 267–286.

McCaul, K. D., Branstetter, A. D., O’Donnell, S. M., Ja-cobson, K., & Quinlan, K. B. (1998). A descriptivestudy of breast cancer worry. Journal of BehavioralMedicine, 21(6), 565–579.

McCaul, K. D., Schroeder, D. M., & Reid, P. A. (1996).Breast cancer worry and screening: Some prospectivedata. Health Psychology, 15(6), 430–433.

Mellers, B. A., & McGraw, A. P. (2001). Anticipatedemotions as guides to choice. Current Directions inPsychological Science, 10, 210–214.

Meyer, D., Leventhal, H., & Gutmann, M. (1985).Common-sense models of illness: The example ofhypertension. Health Psychology, 4, 115–135.

Miller, S., Diefenbach, M., Krantz, D. S., Baum, A., &Academy of Behavioral Medicine Research. (1998).The cognitive–social health information-processing(C-SHIP) model: A theoretical framework for re-search in behavioral oncology. In D. S. Krantz & A.Baum (Eds.), Technology and methods in behavioralmedicine (pp. 219–244). Mahwah, NJ: Erlbaum.

Miller, S. M., Diefenbach, M. A., Kruus, L., Ohls, L.,Hanks, G., Bruner, D., et al. (2001). Psychologicaland screening profiles of first degree relatives of pros-tate cancer patients. Journal of Behavioral Medicine,24, 247–258.

Miller, S. M., Shoda, Y., & Hurley, K. (1996). Applyingcognitive–social theory to health-protective behavior:Breast self-examination in cancer screening. Psycho-logical Bulletin, 119(1), 70–94.

Montague, R. P., & Berns, G. S. (2002). Neural eco-nomics and the biological substrates of valuation.Neuron, 36, 265–284.

Peters, E., Lipkus, I., & Diefenbach, M. (2006). Thefunctions of affect in health communication and inthe construction of health preferences. Journal ofCommunication, 56, S140–S162.

Peters, E., Västfjäll, D., Gärling, T., & Slovic, P. (2006).Affect and decision making: A “hot” topic. Journalof Behavioral Decision Making, 19(2), 79–85.

Prochaska, T. R., Keller, M. L., Leventhal, E. A., &Leventhal, H. (1987). Impact of symptoms and agingattribution on emotions and coping. Health Psychol-ogy, 6, 495–514.

Rogers, R. W., & Deckner, W. C. (1975). Effects of fearappeals and physiological arousal upon emotion, at-titudes, and cigarette smoking. Journal of Personalityand Social Psychology, 32, 222–230.

Rogers, R. W. & Thistlethwaite, D. L. (1970). Effects offear arousal and reassurance on attitude change.Journal of Personality and Social Psychology, 15,227–233.

Russell, J. A., & Barett, L. F. (1999). Core affect, proto-typical emotional episodes, and other things calledemotion: Dissecting the elephant. Journal of Person-ality and Social Psychology, 76(5), 805–819.

Salovey, P., & Birnbaum, D. (1989). Influence of moodon health-relevant cognitions. Journal of Personalityand Social Psychology, 57, 539–551.

Scheier, M. F., & Carver, C. S. (1985). Optimism, cop-ing, and health: Assessment and implications of gen-eralized outcome expectancies. Health Psychology, 4,219–247.

Schwarz, N. (1990). Feelings as information: Informa-tional and motivational functions of affective states.In E. T. Higgins & R. M. Sorrentino (Eds.), Hand-book of motivation and cognition: Foundations ofsocial behavior (Vol. 2, pp. 527–561). New York:Guilford Press.

Schwarz, N., & Clore, G. L. (1988). How do I feelabout it?: The informative function of affective


states. In K. Fielder & J. Forgas (Eds.), Affect, cogni-tion and social behavior: New evidence and inte-grative attempts (pp. 44–62). Toronto, ON: C. J.Hogrefe.

Schwarz, N., & Clore, G. L. (2003). Mood as informa-tion: 20 years later. Psychological Inquiry, 14, 296–303.

Schwartz, M. D., Taylor, K. L., & Willard, K. S. (2003).Prospective association between distress and mam-mography utilization among women with a familyhistory of breast cancer. Journal of Behavioral Medi-cine, 26(2), 105–117.

Selye, H. (1951). The physiology and pathology of ex-posure to stress. Oxford: Acta.

Simon, H. (1967). Motivational and emotional con-trols of cognition. Psychological Review, 74(1), 29–39.

Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D.G. (2002). The affect heuristic. In T. Gilovich, D. Grif-fin, & D. Kahneman (Eds.), Heuristics and biases: Thepsychology of intuitive judgement (pp. 397–420).Cambridge, UK: Cambridge University Press.

Slovic, P., & Peters, E. (2006). Risk Perception and Af-fect. Current Directions in Psychological Science,15(6), 322–325.

Slovic, P., Peters, E., Finucane, M. L., & Macgregor, D.G. (2005). Affect, risk, and decision making. HealthPsychology, 24(4, Suppl.), S35–S40.

Stefanek, M. E., & Wilcox, P. (1991). First degree rela-tives of breast cancer patients: Screening practicesand provision of risk information. Cancer Detectionand Prevention, 15(5), 379–384.

Stoller, E. P. (1984). Self-assessments of health by the el-derly: The impact of informal assistance. Journal ofHealth and Social Behavior, 25, 260–270.

Sutton, S., Saidi, G., Bickler, G., & Hunter, J. (1995).Does routine screening for breast cancer raise anxi-ety?: Results from a three wave prospective study inEngland. Journal of Epidemiology and CommunityHealth, 49(4), 413–418.

Tinley, S. T., Houfek, J., Watson, P., Wenzel, L., Clark,M. B., Coughlin, S., et al. (2004). Screening adher-ence in BRCA1/2 families is associated with primaryphysicians’ behavior. American Journal of MedicalGenetics A, 125(1), 5–11.

Tversky, A., & Kahneman, D. (1981). The framing ofdecisions and the psychology of choice. Science,211(4481), 453–458.

Ubel, P. A., Loewenstein, G. F., Schwarz, N., & Smith,D. (2005). Misimagining the unimaginable: The dis-ability paradox and healthcare decision making.Health Psychology.

U.S. Department of Health and Human Services.(2000). Heathy people 2010. Washington, DC: U.S.Government Printing Office.

Young, R. F., & Severson, R. K. (2005). Breast cancerscreening barriers and mammography completion inolder minority women. Breast Cancer Research andTreatment, 89(2), 111–118.

Zajonc, R. B. (1980). Feeling and thinking. Preferencesneed no inferences. American Psychologist, 35, 151–175.

Zajonc, R. B. (1984). On the primacy of affect. Ameri-can Psychologist, 39(2), 117–123.


C H A P T E R 4 1

Emotions, the Neuroendocrineand Immune Systems,

and Health

MARGARET E. KEMENY and AVGUSTA SHESTYUK

Although emotions and other affective experi-ences have been linked to specific neural struc-tures and circuits, much less is known aboutthe relationships between emotions and theperipheral neuroendocrine and neuroimmunesystems that contribute to physical health anddisease. Most research on the linkages betweenpsychological processes and these peripheralsystems has focused on exposure to stress-ors and the experience of diffuse cognitive–affective states of mind, such as “stress” or“distress” (see Kemeny & Schedlowski, 2007).Alternatively, researchers have examined theperipheral physiological patterns associatedwith psychiatric disorders, including mood dis-orders (e.g., major depression). Despite the factthat emotional states are often considered cen-tral and most proximal mediators of the effectsof environmental perturbation on peripheralphysiology (Lazarus & Folkman, 1984), insuf-ficient attention has been paid to the relationsbetween specific emotions and these criticallyimportant regulatory systems.

An overarching assumption underlyingmuch of the research in this area is that the

physiological response to difficult, stressful, orthreatening environmental perturbations is ageneralized one. Hans Selye shaped the think-ing on this topic for generations of researcherswith his generality model, which proposes acore set of physiological responses that arecommon to all stressors—including activationof the hypothalamic–pituitary–adrenocortical(HPA) axis and impairment of immune func-tion (Selye, 1956). Modern versions of Selye’sgenerality model include many more psycho-logical constructs than he envisioned. How-ever, these modern adaptations remain essen-tially generality models, because “all roads”lead to a general state of distress. In otherwords, although individuals may respond to aparticular event with different forms of affec-tive, cognitive and behavioral response, it is theextent of resulting distress that is assumed todetermine the physiological response. And dis-tress is assumed to have a uniform relationshipto these major physiological systems (Kemeny,2003a).

The notion of an adaptive specificity has be-gun to be entertained in the fields of psycho-

661

neuroendocrinology and psychoneuroimmun-ology, however. For example, Herbert Weiner(1992) argued that “organisms meet . . . chal-lenges and dangers by integrated behavioral,physiological patterns of response that are ap-propriate to the task” (p. 33). Specific condi-tions or environmental signals, then, elicit anintegrated psychobiological response that in-cludes a patterned array of neural and periph-eral changes capable of preparing the organismto deal with the specific situation. For example,fleeing from a predator, grieving over a loss,and demonstrating one’s social status requiredifferent kinds of psychological, physiological,and behavioral changes to respond adaptivelyto these situations (Weiner, 1992). Thus happi-ness is associated with approach behaviors,whereas fear in response to a threat to survivalis associated with specific behavioral and phys-iological changes that facilitate flight, which inturn is different from physiological and be-havioral manifestations of depressive affect(e.g., disengagement; Ekman, 1999; Levenson,1994). This specificity is proposed to includethe response of the HPA axis as well as the im-mune system. In addition, the specific nature ofthe physiological response depends not only onthe characteristics of the environmental cir-cumstance, but also on the cognitive appraisalof that circumstance and the specific emotionalreaction (Kemeny, 2003). Increasing evidencesupports specificity in the relationship amongstressors, cognitive and affective responses, andperipheral physiology. For example, a recentmeta-analysis found that all acute laboratorystressors did not activate the HPA axis. Instead,specific environmental characteristics were re-quired to induce an increase in cortisol, as de-scribed in more detail below (Dickerson &Kemeny, 2004).

This chapter considers how specific neuro-endocrine and immune changes may supportadaptive behavioral responses to specificelicitors. The role of emotion in these integrat-ed psychobiological responses to environmen-tal perturbations is discussed. The chapteremphasizes the evidence linking affective expe-rience with the neuroendocrine system (specifi-cally, the HPA axis), as well as with criticalfunctions of the immune system. In addition,we attempt to provide an initial overview of therelations between affective experience and cen-tral circuitry as a way of considering how thesecircuits mediate the relationship between affectand peripheral responses. The brain plays a key

role as a transducer of the effects of the envi-ronment on peripheral physiological responses.Neural mechanisms interpret elements of socialenvironment; instantiate individuals’ beliefs,expectations, and appraisals; and regulate (i.e.,activate and inhibit) peripheral systems, includ-ing the HPA and immune systems.

In this chapter, we define “emotions” asshort-duration, intensely felt states associatedwith distinctive behavioral predispositions andfacial expressions that have evolved to coordi-nate responses to specific eliciting conditions(Ekman, 1994; Levenson, 1992). “Moods” arelonger-term affective states that do not haveunique facial signatures or eliciting conditions(Ekman, 1994) and often contain persistentcognitions (such as negative thoughts about theself with depressed mood). “Affective traits”are predispositions to experience specifics af-fects over time and across situations. “Mooddisorders” are pathological variants of moodstates that interfere with daily living and mayrequire intervention.

Below we review several adaptive affective–motivational response patterns that involve theHPA and immune systems: the fight–flight re-sponse; the recuperative response and behavior-al disengagement; and the positive approach re-sponse. The “fight–flight response” involvesrecognition of a threat that requires physical ac-tion and the mobilization of adequate physio-logical resources to fight or flee the threat. Theseresponses are associated with emotions of fear(and, if persistent, anxiety) for the flight mode,and anger (and hostility) for the fight mode. The“recuperative response” and “behavioral disen-gagement” reflect physiological resource con-servation processes that facilitate healing anddisengagement from a threat. These responsesare associated with negative affect and depres-sion, as well as with shame and submissiveness.Finally, “positive approach responses” promotemotivation and physiological homeostasis; de-pending on behavioral goals, they rely on re-source building, mobilization, and conserva-tion. These responses are associated withhappiness, contentment, and excitement.

PSYCHOBIOLOGY OFFIGHT–FLIGHT RESPONSES

There is a long-standing recognition that or-ganisms, even quite primitive ones, containphysiological systems adapted to respond effec-


tively to danger and threat. The fight–fight sys-tem becomes activated in response to a threatto the central goal of preserving the physicalself. This system mobilizes bodily processesthat are required for threat-relevant physicalactions (such as fighting and fleeing), anddown-regulates restorative systems that are notcritical to addressing a threat (such as repro-duction, repair, and growth) (Sapolsky, 1993).Sapolsky argues that organisms engage in “re-source building” at a physiological level untilconfronted with an emergency situation, atwhich time there is a shift to “resource utiliza-tion.” For example, fighting requires energy,and activation of the HPA axis results in the re-lease of cortisol, which mobilizes energy re-sources by stimulating the conversion of othersubstrates into glucose, a metabolic fuel.Cortisol also exerts a permissive effect on thesympathetic nervous system (SNS), allowingSNS products to activate the cardiovascularsystem (e.g., increasing heart rate) in prepara-tion for physical exertion. A number of studiesin animals and humans have demonstrated thatexposure to physical and psychological threatscan activate the HPA axis (Dickerson &Kemeny, 2004).

There is a growing literature documentingthe effects of exposure to stressors and threatson the immune system, resulting in decrementsin the functioning of various cell types. Some ofthe immune system changes observed during athreat, however, have been argued to be centralcomponents of the adaptive physiological re-sponse to the threat, rather than nonfunctionalside effects of a more general stress response. Inanimal studies, Dhabhar and colleagues havedemonstrated that acute exposure to a stressoror threat shunts white blood cells to the site oftissue damage or exposure to a pathogen, re-sulting in an enhancement of the local im-mune response (Dhabhar & McEwen, 1999;Viswanathan & Dhabhar, 2005). Dhabhar(2003) argues that these stress-induced immu-nological changes may be an adaptive prepara-tion for the challenge of wounding or infection,which may occur during circumstances that in-volve fighting or fleeing. Just as soldiers aboutto fight an enemy leave the barracks and as-sume positions at battle stations, environmen-tal dangers and threats cause immune cells totravel from the immune organs through thebloodstream to “battle stations” in the skin, inorder to fight against infectious agents thatmay result from fight- or flight-induced

wounding. These effects are probably mediatedvia the activation of the SNS, a key system un-derlying other peripheral responses to fight–flight circumstances, since the SNS neuro-transmitters influence trafficking of whiteblood cells and the expression of moleculesthat regulate immune cell trafficking (Ottaway& Husband, 1992).

At the same time, an increase in a subset ofwhite blood cells, called “natural killer” (NK)cells, would be particularly effective at rapidand nonspecific killing of pathogens that enterthe system in the context of fight- or flight-induced tissue injury (Kemeny & Gruenewald,2000). NK cells can kill virally infected cellsnonspecifically without prior exposure to theorganism. Increases in the number and activityof NK cells immediately following exposureto an acute threat have been demonstrated(Segerstrom & Miller, 2004), while other typesof white blood cells do not show this enhance-ment in most cases.

Fear and Anxiety

Real or perceived threat and its anticipationhave been associated with a host of emotionalresponses, including fear, worry, and anxiety.Imminent threat (such as facing a predator) islikely to evoke fear, whereas an anticipation ofa threatening or negative event (such as waitingto receive a punishment) is likely to evoke anxi-ety and apprehension. Physiological profiles(i.e., neural responses and peripheral activa-tions) associated with these affective responsescan be conceptualized in the framework of pre-paratory mobilization for active avoidance andescape from the threat (i.e., the “flight” re-sponses). Thus fear and anxiety share similarpatterns of central and peripheral activation,but can be differentiated based on whether thethreatening stimulus is certain and imminent(i.e., requires immediate action) or is uncertainand anticipated (i.e., requires preparation foraction) (Cox & Taylor, 1999).

Peripheral autonomic nervous system (ANS)responses associated with fear and anxiety,such as increased heart rate and respiration, arelinked to noradrenergic activity, originating inthe locus coeruleus (Charney, 2004). There isevidence that repeated noradrenergic activa-tion may be linked to a greater risk of cardio-vascular disease, especially in individuals withanxiety disorders (Kubzansky, Kawachi, Weiss,& Sparrow, 1998). Brain areas involved in or-

41. Emotions, Neuroendocrine/Immune Systems, Health 663

chestrating these ANS responses include theperiaqueductal gray (PAG) and the brainstemautonomic nuclei (Price, 2005); these areas inturn are triggered by amygdala activation,which is hypothesized to play an assessmentrole in determining the salience and signifi-cance of external stimuli (Phelps & LeDoux,2005). These norepinephrine-mediated re-sponses are modulated by glutamate-dependent neuronal activity within the pre-frontal regions and hippocampus (Cortese &Phan, 2005), which provide inhibitory reg-ulation of the amygdala, PAG, and brain-stem nuclei. Insufficient activity within the pre-frontal network has been linked to the etiologyand manifestation of anxiety disorders, whichare marked by increased central and peripheralactivation in response to an imagined, unrealis-tic, or unsubstantiated threat.

Physical and psychosocial threats known toelicit fear and anxiety can also activate theHPA axis and trigger cortisol release (Erickson,Drevets, & Schulkin, 2003; Sapolsky, 2005).HPA activation in these situations has beenlinked to enhanced memory for threateningevents and may play a role in the developmentand manifestation of symptoms of posttrau-matic stress disorder (PTSD; McNally, 1999).In fact, HPA dysregulation is one of the mostprominent physiological features of PTSD. In-dividuals with this disorder demonstrate ab-normally low baseline levels of cortisol secre-tion, but exaggerated reactivity of the HPA axisin response to external threats or internalglucocorticoid release (e.g., on the dexametha-sone suppression test; Yehuda, 2000; Yehuda,Boisoneau, Mason, & Giller, 1993). Otheranxiety disorders have a less clear pattern ofHPA engagement. For example, panic attacksby themselves have been found insufficient intriggering HPA activation, but anticipatoryanxiety seems to be more effective in elicitingcortisol release (Fraeff, Garcia-Leal, Del-Ben,& Guimaraes, 2005), and HPA activation hasbeen found to precipitate panic attacks (Strohle& Holsboer, 2003).

In contrast to the ANS and HPA responses,there is less known about the relationship be-tween immune system alterations with fear andanxiety. Naturalistic studies of self-relevant sit-uations that produce genuine feelings of anxi-ety are more successful in eliciting immune ef-fects than laboratory studies. For example,patients reporting higher levels of anxietyabout cancer at their first outpatient visit had

significantly lower levels of NK cell activity(which tests the ability of NK cells to kill tumorcells), suggesting that feelings of anxiety mayhave immunological correlates (Koga et al.,2001). In addition, intervention studies demon-strate an increase in immune functioning asso-ciated with a decrease in anxiety symptoms(Anderson et al., 2004). Furthermore, individu-als with trait worry demonstrate distinct im-mune changes (e.g., fewer NK cells) in responseto naturalistic stressors, such as the aftermathof an earthquake (Segerstrom, Solomon,Kemeny, & Fahey, 1998), or acute laboratory-induced stressors (Segerstrom, Glover, Craske,& Fahey, 1999).

Anxiety disorders have also been linkedto compromised immune functioning. Spe-cifically, changes in anxiety levels from pre- toposttreatment in patients with panic disorderhave been associated with changes in pro-liferative responses to an in vitro stimulus andinterleukin-2 (IL-2) cytokine production—twoimportant aspects of immune cell function(Koh & Lee, 2004). However, such findings ofabnormal immune functioning in anxiety dis-orders are not ubiquitous, as some studies re-port no association between anxiety disordersand proliferative capacity, NK cell activity, orthe numbers of lymphocyte subsets (Schleifer,Keller, & Bartlett, 2002).

Anger and Hostility

Anger-related affective experiences, includinghostility and aggression, have been linked toperipheral systems and health outcomes. “An-ger” is “an unpleasant emotion ranging in in-tensity from irritation or annoyance to fury orrage” (Smith, 1994, p. 25), which often resultsfrom exposure to situations that involve the re-lational theme of unfair interference or harm(Lazarus, 1991). Anger involves the action ten-dency of inflicting harm through “aggression,”which is defined as attacking, destructive, orhurtful behavior (Smith, 1994). “Hostility” hasbeen defined as both an affective trait and amood state; definitions usually include a cogni-tive propensity toward cynicism. Smith (1994)defines hostility as a negative attitude towardothers that involves “a devaluation of theworth and motives of others, an expectationthat others are likely sources of wrong-doing, arelational view of being in opposition towardothers, and a desire to inflict harm or see othersharmed” (p. 26).


Anger, hostility, and aggressiveness (collec-tively termed AHA) have been thought to con-tribute to cardiovascular disease since at leastthe 19th century. These anger-related responsesare one of the most widely studied psychosocialrisk factors for cardiovascular disease and mor-tality. Most studies support this link (e.g.,Rozanski, Blumenthal, & Kaplan, 1999; Smith& Ruiz, 2002), including several recent pro-spective studies of initially healthy samples thathave been controlled for confounding factors(Smith, Glazer, Ruiz, & Gallo, 2004). For ex-ample, hostility is strongly and independentlyassociated with the development and progres-sion of coronary artery disease (Boyle et al.,2004; Miller, Smith, Turner, Guijarro, &Hallet, 1996). Other cardiovascular outcomespredicted by hostility include high blood pres-sure, atherosclerosis, coronary artery calci-fication, and mortality (Suinn, 2001). Effectsizes can be as large as those associated withtraditional risk factors, such as smoking anddiet.

The psychophysiological reactivity modelsuggests that cardiovascular risk associatedwith hostility is due to exaggerated cardiovas-cular and neuroendocrine responses to stress-ors (including blood pressure, heart rate, epi-nephrine, and cortisol; Williams, Barefoot, &Shekelle, 1985). ANS and HPA concomitantsof anger and hostility are thought to be linkedto increased activity within the PAG and hypo-thalamus, and to inadequate inhibitory controlfrom areas of the prefrontal cortex due to in-sufficient availability of serotonin (Gregg &Siegel, 2001; Davidson, Putnam, & Larson,2000). Increasing evidence supports a relation-ship between cardiovascular and neuroendo-crine reactivity to stressors, and the initiationand the progression of coronary artery disease(Rozanski et al., 1999; Smith & Ruiz, 2002).Other important physiological processes arealso enhanced in stressful situations amonghostile individuals, including cholesterol levelsand protein mediators of inflammatory pro-cesses (see Smith et al., 2004). Interestingly, re-search supports the hypothesis that hostilitypredicts an exaggerated physiological responseto stressors, particularly when the stressors aresocial (e.g., Smith & Gallo, 2001). For ex-ample, hostile individuals demonstrated greatphysiological reactivity to one task only whenit included harassment. Conceptualizing thisrisk factor as a social process rather than as afunction of personality alone is likely to lead to

important and interesting new directions forresearch in this area.

More recently, AHA constructs have beenfound to be associated with increased evidenceof peripheral inflammation, which is known topredict increased mortality and morbidity—inparticular, atherosclerosis and myocardial in-farction (Ridker, Hennekens, Buring, & Rifai,2000). For example, trait hostility, more sothan self-reported anger and verbal aggression,has been associated with increased productionof a proinflammatory cytokine (PIC) called tu-mor necrosis factor-α (TNF-α) (Suarez, Lewis,& Kuhn, 2002). TNF-α and other PICs play acritical role in orchestrating the inflammatoryresponse. Overall, the association with inflam-matory processes appears to depend on longterm constructs such as hostility rather thanshort-term experiences of anger. In addition,AHA-related constructs have predicted de-crements in cellular immunity. For example,higher levels of hostility coded from verbal andnonverbal behavior during a marital conflicttask predicted decreases in immune functionover the following 24-hour period. Specifically,individuals who behaved in a hostile way to-ward their spouses showed decreased ability ofimmune cells to proliferate in response to astimulus, as well as decreased ability of NKcells to kill tumor targets (Kiecolt-Glaser et al.,1993).

Although the majority of studies in this arearely on self-report measures of AHA, more re-cent studies have included observer ratings ofhostile behavior in social situations. Hostilitydefined on this basis in these behavioral assess-ment strategies is associated with increasedfacial expressions of anger and disgust(Brummett et al., 1998), reduced expressions offriendly appeasement (Prkachin & Silverman,2002), and increased negative cardiovascularoutcomes.

Unfortunately, few studies have directlycompared other affective constructs, such asanxiety or depression, to anger and hostility aspredictors of cardiovascular disease outcomesin the same studies. However, there is clearlyless evidence for the role of these affects in thisliterature. Whether this is due to the popularityof hostility as a construct in this research area,or to a psychobiological specificity that sug-gests something unique and physiologically im-portant about the neurobiology of anger andhostility, is an important issue and deserves sci-entific attention.


PSYCHOBIOLOGYOF DISENGAGEMENT

Whereas fear and anger reflect adaptive physi-ological responses to an imminent or antici-pated threat and result in either fight or flightbehaviors, exposure to an uncontrollable phys-ical or psychosocial threat often activates emo-tional and physiological systems associatedwith behavioral disengagement and with-drawal. Behavioral disengagement is repre-sented in a number of emotional responses,ranging from negative affect and depression (asresponses to a personal loss or uncontrollablephysical and psychosocial threat) to shame andembarrassment (as responses to a loss of socialstatus or exposure to a social threat). Theseemotions and withdrawal behaviors are associ-ated with a distinct pattern of central and pe-ripheral activation, including diminished sero-tonergic and dopaminergic neuronal activity,abnormal prefrontal functioning, HPA dysreg-ulation, and compromised immune function-ing.

In particular, a number of physiologicalchanges, such as increases in HPA activation,opioid-mediated analgesia, and lowered cellu-lar immune function, are associated with goaldisengagement in animals (Fleshner, Landens-lager, Simons, & Maher, 1989; Miczek,Thompson, & Shuster, 1982; Stefanski &Ben-Eliyahu, 1996). Moreover, the role of theimmune system in withdrawal behavior canbe traced to a similar adaptive behavioral re-sponse called “behavioral recuperation.” Fol-lowing infection or injury, PICs are releasedby immunological cells to facilitate the im-mune response against the organism, for ex-ample. More recently, it has become clearthat PICs play another critical role in re-sponse to infectious agents. These proteinsact on the brain, causing what is known as“sickness behavior”—that is, inducing in-creases in sleep and decreases in social, sex-ual, aggressive, exploratory, and other behav-iors. Careful animal work has shown thatthese behavioral changes are not functions ofweakness or incapacitation, but represent amotivational shift (e.g., away from fight andflight) toward behavior that would supportrecuperation. This behavioral disengagementappears to be an adaptive response that al-lows the organism to conserve energy andthus to maximize recuperative physiological

processes, such as the maintenance of fever(Maier & Watkins, 1998).

PICs may play a significant role in otherforms of adaptive “behavioral disengage-ment”—in particular, in uncontrollable con-texts where disengagement may be the mostadaptive response. Evidence suggests that ani-mals are more likely to demonstrate an increasein the production of PICs in response to an un-controllable stressor. For example, subordinateanimals confronted with an attack from a dom-inant animal have been found to exhibit behav-ioral disengagement behaviors in conjunctionwith PIC release (Kinsey, Bailey, Sheridan,Padgett, & Avitsur, 2007). In such case, behav-ioral disengagement is more adaptive, as itminimizes the chance of an attack from thedominant animal (Kemeny, 2006a). Similarly,animals injected with PICs cease to display of-fensive behaviors in an aggressive encounter(Cirulli, De Acetis, & Alleva, 1998). Finally,uncontrollable acute social threats in humanshave also been shown to increase levels of PICs,as discussed below (Ackerman, Martino,Heyman, Moyna, & Rabin, 1998; Dickerson,et al., 2007).

Negative Affect and Depression

Negative affect and depression have beenlinked to insufficient availability of serotoninand dopamine and to abnormal activation, es-pecially within structures of the prefrontal cor-tex such as dorsolateral prefrontal cortex(DLPFC), ventromedial prefrontal cortex(VMPFC), and orbito-frontal cortex (OFC;Phan, Wager, Taylor, & Liberzon, 2004;Drevets, 1999). In particular, low concentra-tions of prefrontal dopamine and diminishedactivity in these regions have been linked to aloss of motivation, lack of goal-oriented behav-ior, and behavioral withdrawal in depression—symptoms commonly seen in individuals withorganic prefrontal brain damage or frontal de-mentia (Rogers, Bradshaw, Pantelis, & Phillips,1998). Moreover, abnormal patterns of pre-frontal alpha electroencephalographic (EEG)asymmetry (i.e., relative increase in right-hemisphere activity relative to a decrease inleft-hemisphere activity) have been observedduring negative affect, in individuals at risk fordepression, in individuals with current or re-mitted depression, or in those high in neuroti-cism (Davidson, 1992).


Abnormal activity in the prefrontal brain re-gions, and thus insufficient regulatory controlexhibited during negative affect or depression,may account for the peripheral changes anddysregulation that accompany these emotionsand may prompt behavioral withdrawal anddisengagement. In humans, there is some em-pirical evidence that negative affect may lead toHPA activation and cortisol release; however,such evidence is limited, and observed effectsizes associated with the increases in cortisolare small compared to those elicited by physi-cal or social threats (Dickerson, Gruenewald,& Kemeny, 2004). Findings of HPA dysregula-tion in clinical depression are more compelling,although the pattern of HPA dysregulationvaries depending on depression subtype. Spe-cifically, melancholic, but not atypical, depres-sion has been linked to increased baselinecortisol secretion, abnormal slope of diurnalcortisol release, and limited HPA reactivityin response to acute laboratory stressors(Howland & Thase, 1999). Abnormal HPA ac-tivity in depression and negative affect maycontribute to the deleterious effects on the im-mune system often observed during negativeaffect or depression, as described below. More-over, patterns of prefrontal EEG asymmetryconsistent with negative affect have been linkedwith lower antibody titers after vaccination forinfluenza (Davidson et al., 2003) and withlower NK cell activity (Davidson, Coe, Dolski,& Donzella, 1999; Kang et al., 1991).

The link between compromised immunefunctioning and depression has received muchattention in recent years. In particular, studieshave examined relations between major de-pressive episodes and immune status. Majordepression has been associated with a varietyof functional and phenotypic changes in theimmune system, similar to those often observedin response to major life events and stressors(Irwin, Daniels, Bloom, & Weiner., 1986;Schleifer et al., 1984). Depression has been reli-ably associated with increases in neutrophilsand decreases in B and T cells, as well as withdeficiencies in the proliferative response andNK cell activity (Herbert & Cohen, 1993).These immune effects may be specific to an ac-tive state of depression, as they have beenshown to be ameliorated with treatment(Irwin, Lacher, & Caldwell, 1992). Immuneimpairment is most likely to be observed inthose individuals with major depression who

are older, have melancholic depression, or havea sleep disorder—conditions commonly associ-ated with abnormal HPA functioning (Irwin,2001).

Furthermore, there is increasing evidencethat clinical depression is associated with ele-vated levels and/or production of PICs, evenwhen samples are controlled for medical condi-tions, medications, and health behaviors (Maeset al., 1994; Miller, Stetler, Carney, Freedland,& Banks, 2002). It has been proposed that thatPICs may not only be consequences of majordepression, but may also contribute to its etiol-ogy (Dantzer et al., 2001; Yirmiya, 1996;Yirmiya et al., 1999). In addition to the find-ings above, demonstrating in a number of stud-ies that patients with major depression showevidence of increases in PICs, patients with in-flammatory diseases (e.g., rheumatoid arthri-tis) often experience depression symptoms inexcess of what is observed in individuals withdiseases that are not directly linked to inflam-mation. Also, administration of cytokines suchas interferon to boost the immune system dur-ing cancer or AIDS treatment often inducessymptoms of clinical depression, such asdysphoria, anhedonia, fatigue, and apathy,within days of the first administration. Thesesymptoms tend to remit when the cytokinetreatment is discontinued. Finally, animals in-jected with PICs demonstrate the behavioralchanges called sickness behavior—includingmotor retardation; lack of appetite, weightloss, and sleep disturbances; decreased sexual,social, and aggressive behavior; and anhe-donia. These behavioral changes bear a strikingresemblance to the vegetative symptoms ofsome forms of depression. Thus increases inthese inflammatory molecules may play acausal role in the initiation or maintenance ofdepressive symptoms.

Shame and Submissiveness

In addition to the motive to protect the “physi-cal self,” humans and other animals are moti-vated to protect the “social self.” Maintaininga sense of belonging and social connection ap-pears to be a fundamental motive that is ex-pressed across cultures. Threats to social con-nection can have a variety of adverse effects(Baumeister & Leary, 1995; Cacioppo et al., inpress). In particular, contexts that include rejec-tion, stigmatization, or discrimination are ex-


perienced as aversive, because they reflect alack of social value or status and can result indrops in self-esteem (Baumeister & Leary,1995). Social self-preservation theory arguesthat maintaining social status and value (i.e.,social self-preservation) is a central motive inhumans and other animals, akin to physicalself-preservation (Dickerson, Gruenewald, &Kemeny, 2004). Threats to one’s status orvalue, as in negative social evaluation, rejec-tion, or stigmatization, can have a variety ofpsychological and physiological effects. If anindividual is repeatedly exposed to evaluativeand rejecting circumstances, such as those ex-perienced in the context of low social status,persistent activation of these physiological sys-tems may result in health risks.

Threats to the social self often involve condi-tions in which one may be negatively judged byothers (i.e., social evaluative threat, or SET;Dickerson, Gruenewald, & Kemeny, 2004).SET can occur in performance contexts whereimportant attributes are displayed in front ofothers, in a context where group membership isat risk or where an uncontrollable aspect ofone’s identity is made salient to others. Self-conscious emotions are frequent consequencesof such contexts—notably shame, since thisemotion is experienced when a core aspect ofthe self is judged to be defective or inferior byothers (Gilbert, 1997; Tangney, 1995). The so-cial nature of shame is fundamental to mosthistorical and modern accounts of the emotion(e.g., Cooley, 1902; Benedict, 1946). For exam-ple, Darwin (1871/1899) argued that shame“relates almost exclusively to the judgment ofothers” (p. 114).

The motivational context, displays, and be-haviors associated with shame suggest a linkbetween this emotion and submissive behaviorand withdrawal. For example, shame is elicitedunder social self-threat conditions and is asso-ciated with the motivation to disengage orwithdraw (e.g., wanting to “shrink,” “dis-appear,” or avoid interpersonal interaction;Tangney, 1995). Submissive behavior in ani-mals is elicited when social status is threatened;the behavior is consistent with the motive todisengage and withdraw. Characteristic dis-plays of shame in humans are quite similar todisplays of submission and appeasement inprimates (e.g., gaze avoidance, head down,slumped posture, attempts to appear small;Gilbert, 1997; Keltner & Buswell, 1996).

Social self-preservation theory proposes thatthreats to the social self—such as threats toself-esteem, status, or acceptance—elicit anintegrated psychobiological response that sup-ports adaptive behavior in that context. Spe-cifically, social self-threats provide a set of elic-iting conditions for specific psychobiologicalchanges that would support appeasement andbehavioral disengagement. Recent researchsuggests that three components of this coordi-nated response may be central nervous system(CNS)-coordinated activation of the HPA, in-duced insensitivity of certain immune cells tothe suppressive effects of cortisol, and in-creased production of PICs.

Although a variety of stressful circumstancescan activate the HPA system, acute socialthreats are reliable and powerful elicitors ofsuch activation. In a meta-analytic review of208 acute laboratory stress studies, exposure toperformance tasks in the presence of evaluativeothers, particularly in an uncontrollable con-text, was associated with substantially greatercortisol and adrenocorticotropic hormone re-sponses than was exposure to other stressfultasks without social evaluation (Dickerson &Kemeny, 2004). Exposure to uncontrollableSET was also associated with slower recoveryof cortisol to baseline levels.

An experimental test of the link betweenSET and cortisol increase demonstrated thatperformance tasks that had some degree of un-controllability as well as an SET component(performance in front of an evaluative audi-ence) induced a cortisol response, whereas thesame uncontrollable task without SET did not(Gruenewald, Kemeny, Aziz, & Fahey, 2004).Although both conditions elicited a variety ofemotions, the strongest effects were for shameand self-conscious emotions. The greater theshame reported, the greater the increase incortisol in response to the SET. Other studieshave also shown correlations between self-conscious emotions and cortisol (e.g., Lewis &Ramsay, 2002). Supporting the link betweensocial evaluative processes and cortisol, fear ofnegative evaluation is a significant predictor ofcortisol response to a laboratory-induced SET,with those reporting higher levels of such fearshowing a greater cortisol response to the SETcondition (Gruenewald et al., 2004).

At the same time, studies of primates demon-strate higher levels of cortisol associated withsubordinate rank and submissive displays, sug-


gesting that processes related to social statusand submissive behavior in other animals mayhave been maintained and elaborated inhumans as responses to social self-threats(Gilbert, 1997; Price, Sloman, Gardner,Gilbert, & Rohde, 1994). This HPA reactivitytranslates into greater basal cortisol levels, aslower response to challenge, and impairedsensitivity of the HPA axis to negative feedbackregulation (Sapolsky, 2005). These effects havebeen argued to reflect the persistent exposureto social stressors and limited coping resources(Abbott et al., 2003).

HPA activation in response to SET is con-trolled by a specific neural network, consistingof cortical and subcortical structures. Animalresearch has shown that cortisol release in re-sponse to an acute physical or psychosocialthreat is triggered by activity of the para-ventricular nucleus (PVN) of the hypothala-mus. In turn, activity of the PVN is regulatedby limbic and higher cortical areas directly im-plicated in social cognition: the amygdala, hip-pocampus, anterior cingulate cortex, OFC,VMPFC, and DLPFC (Herman, Ostrander,Mueller, & Figueiredo, 2005; Floyd, Price,Ferry, Keay, & Bandler, 2001). These brain ar-eas demonstrate greater activation in responseto social stimuli and have been implicated in in-dividuals’ emotional, cognitive, and behavioralresponses in the social context (Ochsner,2004). For example, the OFC plays a crucialrole in the appraisal of complex social situa-tions and self-appraisal of social emotions,such that lesions in this area in humans resultin failures to recognize or experience such emo-tions as shame and to monitor social behavioronline (Beer, Heerey, Keltner, Scabini, &Knight, 2003; Beers, John, Scabini, & Knight,2006). Given these patterns of brain activationin response to social cues and self-appraisal, itis likely that these brain areas feature promi-nently in the translation of social appraisalsinto peripheral physiological responses.

A second physiological response that wouldbe adaptive in the context of social self-threatwould be production of PICs. Social status dis-ruption, which induces social submission anddefeat, has been shown to increase productionof PICs and to decrease sensitivity of im-munological cells to down-regulation byglucocorticoids (i.e., glucocorticoid resistance,or GCR) in animals. The primary model of so-cial threat utilized in this literature is a mouse

model of social status disruption, in which anaggressive intruder is placed in the home cagesof other animals, resulting in social defeat andsubordination in the home-caged animals.Mice that are socially defeated show greaterPIC production and greater GCR than do con-trol animals that were not socially defeated(Avitsur, Stark, & Sheridan, 2001; Stark,Avitsur, Padgett, & Sheridan, 2001; Quan etal., 2001). The degree of submissive behavior isassociated with the level of PIC production. Byrendering immune cells insensitive to the sup-pressive effects of glucocorticoids, GCR causesincreases in the production of PICs. Thus spe-cific forms of social threat (i.e., social subordi-nation and defeat) may cause increases in in-flammatory processes in animals via GCR.

There is much less known about the impactof social status disruption on inflammatoryprocesses in humans. Some studies indicatethat acute social threat in humans increasesPIC production. Recent findings suggestgreater TNF-α production by peripheral bloodmononuclear cells in an experimental contextthat involved performing stressful tasks withan evaluative audience, compared to perform-ing these tasks without the evaluative audience(Dickerson et al., 2007). The increase in TNF-αproduction was correlated with the extent towhich subjects described themselves as feelingmore evaluated during the task, suggesting thatcognitive appraisals of social threat may be akey mediator of these effects. In addition, in-duction of shame has been associated with in-creases in PICs. In one study, participants wereeither asked to write about a situation in whichthey blamed themselves or a neutral topic onthree different days during one week. Those as-signed to the self-blame condition showed in-creased levels of TNF-α receptor in oralmucosal transudate on each study day. In-creases were correlated with self-reported in-creases in shame, but not with guilt or otheremotions (Dickerson, Kemeny, Aziz, Kim, &Fahey, 2004).

Increases in cortisol, GCR, and PICs can beconsidered adaptive responses to relatively un-controllable social status threats. As describedearlier, increased levels of PICs would supportthe behavioral disengagement that would beadaptive in subordinate-ranked animals andwould support submissive appeasement-relatedbehavior, thus reducing the likelihood that thesubordinate animal would provoke an attack.


GCR would prevent the increased cortisol lev-els induced by these social stressors from inhib-iting the production of PICs.

PSYCHOBIOLOGY OFPOSITIVE APPROACH

Whereas negative emotions, for the most part,are associated with withdrawal behaviors, pos-itive emotions have been linked to approachbehaviors and are thought to facilitate explora-tion, pleasure and comfort seeking, and affilia-tion. As such, positive emotions are likelyto provide a psychological and physiologicalfoundation for motivation and proactive be-haviors. In fact, it is likely that positive affec-tive states represent an affective baseline forhealthy individuals (Seligman, 2002). Researchsuggests that psychologically healthy partici-pants anticipate more positive than negativeevents, have a predominantly positive view ofthe self, and exhibit preferential processing ofand brain activation in response to positive rel-ative to negative or neutral stimuli (Taylor &Armor, 1996; Shestyuk, Deldin, Brand, &Deveney, 2005). Moreover, positive affectivestates have been linked, for the most part, toenhanced immune functioning and better over-all health outcomes (Barak, 2006). However, itappears that the peripheral effects of positiveemotions may be moderated by the level ofphysiological arousal—such that intense ex-citement would induce significant ANS, HPA,and immune changes, whereas calm content-ment would not.

Positive emotions, such as happiness andpleasure, are associated with greater frontalleft-hemisphere asymmetry (Davidson, 1992)and greater activation of the medial prefrontalregions and the basal ganglia (Parkinson, Car-dinal, & Everitt, 2000). These processes aresupported by monoamine neurotransmittersystems, with greater levels of serotonin as-sociated with low-arousal contentment andpleasure, and greater levels of dopamine andnorepinephrine associated with high-arousalexcitement (Charney, 2004). Arousal can alsomoderate HPA activation. Specifically, somestudies demonstrate a decrease in cortisol levelsfollowing positive mood induction or in corre-lation with trait positive affect (Berk et al.,1989; Buchanan, al’Absi, & Lovallo, 1999;Zachariae et al., 1991). In contrast, HPA acti-vation has also been shown during induction of

positive affect, which may result from experi-enced arousal (Pressman & Cohen, 2005),although there are few studies in this area. Sim-ilarly, induction of positive affect in studies thatrequire active participation from participantshas been shown to increase dopaminergicand noradrenergic neuronal activity, leading toelevations in heart rate and blood pressure(Ekman, Levenson, & Friesen, 1983;Futterman, Kemeny, Shapiro, & Fahey, 1994;Schwartz, Weinberger, & Singer, 1981).

Positive affect is also associated withchanges in the immune system (Pressman &Cohen, 2005). For example, positive affect ispositively correlated with secretory immuno-globulin A, an antibody found in the mucosalimmune system (e.g., in saliva; Hucklebridgeet al., 2000; Lambert & Lambert, 1995;McClelland & Cheriff, 1997). Moreover, dailypositive affect has been linked with elevatedlevels of this antibody in response to a specificantigen measured on a daily basis (Stone, Cox,Valdimarsdottir, Jandorf, & Neale, 1987;Stone et al., 1994; but not Evans et al., 1993).Similarly, positive affect has been associatedwith increases in the number of certain subsetsof white blood cells—a result similar to the oneobserved in response to acute stressors, sug-gesting a potential moderating role of arousal(Segerstrom & Miller, 2004).

Furthermore, laboratory induction of posi-tive affect has been shown to lead to an in-crease in lymphocyte proliferative responseto mitogenic stimulation (Futterman, Kemeny,Shapiro, Polonsky, & Fahey, 1992). These ef-fects were not mediated by changes in cortisolin response to the manipulation. Positive moodinduction may also result in increased levels ofcertain cytokines (IL-2, IL-3), while resulting indecreases in others (i.e., interferon-α and TNF-α; Mittwoch-Jaffe, Shalit, Srendi, & Yehuda,1995). Moreover, Type I hypersensitivity to anallergen has been associated with positive af-fect in some studies (Kimata, 2001; Laidlaw,Booth, & Large, 1996), but not in others(Zachariae, Jorgensen, Egekvist, & Bjerring,2001). Finally, induction of positive affect canlead to an increase in NK cell numbers, whichis similar to the changes in response to negativeaffect and threatening stimuli (Futterman et al.,1994; Segerstrom & Miller, 2004). This effectis probably mediated by SNS arousal. Thusperipheral responses associated with positiveemotions are both distinct from and similar tothose elicited by negative emotions, depending


on the level of arousal and behavioral engage-ment.

SUMMARY

There is increasing evidence that affective expe-riences are associated with important changesin peripheral regulatory systems intimately tiedto health and disease. Further research in thisarea will be greatly benefited by, and may actu-ally require, an integration of the frameworksand tools utilized in affective neuroscience withthose of health psychology and psychoneuro-immunology. Although there are data on theneural substrates of specific emotions, as wellas data on the link between these emotions andperipheral physiology, very few studies have at-tempted to define the pathway from the neuralsubstrates of these affective experiences to pe-ripheral physiological systems and then on tophysical health indices (Kemeny, 2003b). Nocomprehensive mechanistic map of these rela-tionships exists to form the basis for these stud-ies. In the investigation of the role of affectiveexperience in physical disease, it is critical thatthe CNS be investigated, since the brain is themost proximal physiological substrate throughwhich psychological factors act on peripheralneural systems that affect pathophysiologicalmechanisms. Unfortunately, since the brain israrely studied in this area of research, a criticalstep in the pathway remains unknown. An inte-gration of the approaches in affective neurosci-ence with those utilized in health psychologyand psychoneuroimmunology would form astrong basis for determining the complete path-way from emotional experience to peripheralresponse, including the neural circuitry in-volved in translating individual differences inaffective response into important changes inthe neuroendocrine and neuroimmune systems.These findings may then have tremendousvalue for understanding neural regulation ofbiological systems that play a significant role inhealth and disease.

REFERENCES

Abbott, D. H., Keverne, E. B., Bercovitch, F. B., Shively,C. A., Mendoza, S. P., Saltzman, C. T., et al. (2003).Are subordinates always stressed?: A comparativeanalysis of rank differences in cortisol levels amongprimates. Hormones and Behavior, 43, 67–82.

Anderson, N. B., Kaplan, R. M., Kumanyika, S.,

Salovey, P., Kawachi, I., Dunbar-Jacob, J., et al.(2004). Encyclopedia of health and behavior. Thou-sand Oaks, CA: Sage.

Avitsur, R., Stark, J. L., & Sheridan, J. F. (2001). Socialstress induces glucocorticoid resistance in subordi-nate animals. Hormones and Behavior, 39, 247–257.

Barak, Y. (2006). The immune system and happiness.Autoimmunity Reviews, 5, 523–527.

Baumeister, R. F., & Leary, M. R. (1995). The need tobelong: Desire for interpersonal attachment as a fun-damental human motivation. Psychological Bulletin,117, 497–529.

Beer, J. S., Heerey, E. A., Keltner, D., Scabini, D., &Knight, R. T. (2003). The regulatory function of self-conscious emotion: Insights from patients withorbitofrontal damage. Journal of Personality and So-cial Psychology, 85, 594–604.

Beer, J. S., John, O. P., Scabini, D., & Knight, R. T.(2006). Orbitofrontal cortex and social behavior: In-tegrating self-monitoring and emotion–cognition in-teractions. Journal of Cognitive Neuroscience, 18,871–879.

Benedict, R. (1946). The chrysanthemum and thesword. Boston: Houghton Mifflin.

Berk, L. S., Tan, S. A., Fry, W. F., Napier, B. J., Lee, J. W.,Hubbard, R. W., et al. (1989). Neuroendocrine andstress hormone changes during mirthful laughter.American Journal of the Medical Sciences, 298, 390–396.

Boyle, S. H., Williams, R. B., Mark, D. B., Brummett, B.H., Siegler, I. C., Helms, M. J., et al. (2004). Hostilityas a predictor of survival in patients with coronaryartery disease. Psychosomatic Medicine, 66, 629–632.

Brummett, B. H., Maynard, K. E., Babyak, M. A.,Haney, T. L., Siegler, I. C., Helms, M. J., et al.(1998). Measures of hostility as predictors of facialaffect during social interaction: Evidence for con-struct validity. Annals of Behavioral Medicine, 20,168–173.

Buchanan, T. W., al’Absi, M., & Lovallo, W. R. (1999).Cortisol fluctuates with increases and decreases innegative affect. Psychoneuroendocrinology, 24, 227–241.

Cacioppo, J. T., Hawkley, L. C., Ernst, J. M., Burleson,M., Berntson, G. G., Nouriani, B., et al. (2006).Loneliness within a nomological net: An evolution-ary perspective. Journal of Research in Personality,40, 1054–1085.

Charney, D. S. (2004). Psychobiological mechanisms ofresilience and vulnerability: Implications for success-ful adaptation to extreme stress. American Journal ofPsychiatry, 161, 195–216.

Cirulli, F., De Acetis, L., & Alleva, E. (1998). Behavioraleffects of peripheral interleukin-1 administration inadult CD-1 mice: Specific inhibition of the offensivecomponents of intermale agonistic behavior. BrainResearch, 791, 308–312.

Cooley, C. N. (1902). Human nature and the social or-der. New York: Scribner.


Cortese, B. M., & Phan, K. L. (2005). The role of gluta-mate in anxiety and related disorders. CNS Spec-trums, 10, 820–830.

Cox, B. J., & Taylor, S. (1999). Anxiety disorders: Panicand phobias. In T. Millon, P. H. Blaney, & R. D. Da-vis (Eds.), Oxford textbook of psychopathology(pp. 81–113). New York: Oxford University Press.

Dantzer, R., Bluthe, R., Castanon, N., Cauvet, N.,Capuron, L., Goodall, G., et al. (2001). Cytokine ef-fects on behavior. In R. Ader, D. L. Felten, & N. Co-hen (Eds.), Psychoneuroimmunology (3rd ed., Vol. 1,pp. 703–727). San Diego: Academic Press.

Darwin, C. (1899). The descent of man (2nd ed.). Lon-don: Murray. (Original work published 1871)

Davidson, R. J. (1992). Emotion and affective style:Hemispheric substrates. Psychological Science, 3,39–43.

Davidson, R. J., Coe, C. C., Dolski, I., & Donzella, B.(1999). Individual differences in prefrontal activationasymmetry predict natural killer cell activity at restand in response to challenge. Brain, Behavior, andImmunity, 13, 93–108.

Davidson, R. J., Kabat-Zinn, J., Schumacher, J.,Rosenkranz, M., Muller, D., Santorelli, S. F., et al.(2003). Alterations in brain and immune functionproduced by mindfulness meditation. PsychosomaticMedicine, 65, 564–570.

Davidson, R. J., Putnam, K. M., & Larson, C. L.(2000). Dysfunction in the neural circuitry of emo-tion regulation: A possible prelude to violence. Sci-ence, 290, 1093–1095.

Dhabhar, F. S. (2003). Stress, leukocyte trafficking, andthe augmentation of skin immune function. Annalsof the New York Academy of Sciences, 992, 205–217.

Dhabhar, F. S., & McEwen, B. S. (1999). Enhancing ver-sus suppressive effects of stress hormones on skin im-mune function. Proceedings of the National Acad-emy of Sciences of the USA, 96, 1059–1064.

Dickerson, S. S., Gruenewald, T. L., & Kemeny, M. E.(2004). When the social self is threatened: Shame,physiology, and health. Journal of Personality, 72,1191–1216.

Dickerson, S. S., & Kemeny, M. E. (2004). Acute stress-ors and cortisol responses: A theoretical integrationand synthesis of laboratory research. PsychologicalBulletin, 130, 355–391.

Dickerson, S. S., Kemeny, M. E., Aziz, N., Kim, K. H.,& Fahey, J. L. (2004). Immunological effects of in-duced shame and guilt. Psychosomatic Medicine, 66,124–131.

Dickerson, S. S., Gables, S., Kemeny, M. E., Aziz, N.,Irwin, M., & Fahey, J. L. (2007). Social evaluativethreat induces increased production of TNF-α.Manuscript in preparation.

Drevets, W. C. (1999). Prefrontal cortical–amygdalarmetabolism in major depression. Annals of the NewYork Academy of Sciences, 29, 614–637.

Ekman, P. (1994). Moods, emotions, and traits. In P.

Ekman & R. J. Davidson (Eds.), The nature of emo-tion: Fundamental questions (pp. 56–58). New York:Oxford University Press.

Ekman, P. (1999). Basic emotions. In T. Dalgleish & M.Powers (Eds.), Handbook of cognition and emotion.Chichester, UK: Wiley.

Ekman, P., Levenson, R. W., & Friesen, W. V. (1983).Autonomic nervous system activity distinguishesamong emotions. Science, 221, 1208–1210.

Erickson, K., Drevets, W., & Schulkin, J. (2003).Glucocorticoid regulation of diverse cognitive func-tions in normal and pathological emotional states.Neuroscience and Biobehavioral Reviews, 27, 233–246.

Evans, P., Bristow, M., Hucklebridge, F., Clow, A., &Walters, N. (1993). The relationship between secre-tory immunity, mood and life-events. British Journalof Psychology, 32, 227–236.

Fleshner, M., Laudenslager, M. L., Simons, L., & Maier,S. F. (1989). Reduced serum antibodies associatedwith social defeat in rats. Physiology and Behavior,45, 1183–1187.

Floyd, N. S., Price, J. L., Ferry, A. T., Keay, K. A., &Bandler, R. (2001). Orbitomedial prefrontal corticalprojections to hypothalamus in the rat. Journal ofComparative Neurology, 432, 307–328.

Fraeff, F. G., Garcia-Leal, C., Del-Ben, C. M., &Guimaraes, F. S. (2005). Does the panic attack acti-vate the hypothalamic–pituitary–adrenal axis? Anaisda Academia Brasileira de Ciencias, 77, 477–491.

Futterman, A. D., Kemeny, M. E., Shapiro, D., & Fahey,J. L. (1994). Immunological and physiologicalchanges associated with induced positive and nega-tive mood. Psychosomatic Medicine, 56, 499–511.

Futterman, A. D., Kemeny, M. E., Shapiro, D.,Polonsky, W., & Fahey, J. L. (1992). Immunologicalvariability associated with experimentally-inducedpositive and negative affective states. PsychologicalMedicine, 22, 231–238.

Gilbert, P. (1997). The evolution of social attractivenessand its role in shame, humiliation, guilt and therapy.British Journal of Medical Psychology, 70, 113–147.

Gregg, T. R., & Siegel, A. (2000). Brain structures andneurotransmitters regulating aggression in cats: Im-plications for human aggression. Science, 289, 591–594.

Gruenewald, T. L., Kemeny, M. E., Aziz, N., & Fahey, J.L. (2004). Acute threat to the social self: Shame, so-cial self-esteem and cortisol activity. PsychosomaticMedicine, 66, 915–924.

Herbert, T. B., & Cohen, S. (1993). Depression and im-munity: A meta-analytic review. Psychological Bulle-tin, 113, 472–486.

Herman, J. P., Ostrander, M. M., Mueller, N. K., &Figueiredo, H. (2005). Limbic system mechanisms ofstress regulation: Hypothalamo–pituitary–adrenocorticalaxis. Progress in Neuro-Psychopharmacology andBiological Psychiatry, 29, 1201–1213.

Howland, R. H., & Thase, M. E. (1999). Affective dis-


orders: Biological aspects. In T. Millon, P. H. Blaney,& R. D. Davis (Eds.), Oxford textbook of psycho-pathology (pp. 166–202). New York: Oxford Uni-versity Press.

Hucklebridge, F., Lambert, S., Clow, A., Warburton, D.M., Evans, P. D., & Sherwood, N. (2000). Modula-tion of secretory immunoglobulin a in saliva; re-sponse to manipulation of mood. Biological Psychol-ogy, 53, 25–35.

Irwin, M. (2001). Depression and immunity. In R. Ader,D. L. Felten, & N. Cohen (Eds.), Psychoneuro-immunology (3rd ed., Vol. 2, pp. 383–398). SanDiego: Academic Press.

Irwin, M., Daniels, M., Bloom, E. T., & Weiner, H.(1986). Life events, depression, and natural killer cellactivity. Psychopharmacology Bulletin, 22, 1093–1096.

Irwin, M., Lacher, U., & Caldwell, C. (1992). Depres-sion and reduced natural killer cytotoxicity: A longi-tudinal study of depressed patients and control sub-jects. Psychological Medicine, 22, 1045–1050.

Kang, D. H., Davidson, R. I., Coe, C. L., Wheeler, R.W., Tomarken, A. J., & Ershler, W. B. (1991). Frontalbrain asymmetry and immune function. BehavioralNeuroscience, 105, 860–869.

Keltner, D., & Buswell, B. N. (1996). Evidence for thedistinctness of embarrassment, shame, and guilt: Astudy of recalled antecedents and facial expressionsof emotion. Cognition and Emotion, 10, 155–171.

Kemeny, M. E. (2003a). The psychobiology of stress.Current Directions in Psychological Science, 12,124–128.

Kemeny, M. E. (2003b). An interdisciplinary researchmodel to investigate psychosocial cofactors indisease: Application to HIV-1 pathogenesis. BrainBehavior and Immunity, 17, 62–72.

Kemeny, M. E. (2006a). Psychoneuroimmunology. In H.Friedman & R. Silver (Eds.), Foundations of healthpsychology (pp. 92–116). New York: Oxford Univer-sity Press.

Kemeny, M. E., & Gruenewald, T. L. (2000). Affect,cognition, the immune system and health. Progress inBrain Research, 122, 291–308.

Kemeny, M. E., & Schedlowski, M. (2007). Under-standing the interaction between psychosocial stressand immune-related disease: A stepwise progression.Brain, Behavior and Immunity, 21, 1009–1018.

Kiecolt-Glaser, J. K., Malarkey, W. B., Chee, M., New-ton, T., Cacioppo, J. T., Mao, H. Y., et al. (1993).Negative behavior during marital conflict is associ-ated with immunological down-regulation. Psycho-somatic Medicine, 55, 410–412.

Kimata, H. (2001). Effect of humor on allergen-inducedwheal reactions. Journal of the American MedicalAssociation, 285, 738.

Kinsey, S. G., Bailey, M. T., Sheridan, J. F., Padgett, D.A., & Avitsur, R. (2007). Prepeated social defeatcauses increased anxiety-like behavior and alters

splenocyte function in C57BL/6 and CD-1 mice.Brain, Behavior, and Immunity, 21, 458–466.

Koga, C., Itoh, K., Aoki, M., Suefuji, Y., Yoshida, M.,Asosina, S., et al. (2001). Anxiety and pain suppressthe natural killer cell activity in oral surgery outpa-tients. Oral Surgery, Oral Medicine, Oral Pathology,Oral Radiology, and Endodontics, 91, 654–658.

Koh, K. B., & Lee, Y. (2004). Reduced anxiety level bytherapeutic interventions and cell-mediated immu-nity in panic disorder patients. Psychotherapy andPsychosomatics, 73, 286–292.

Kubzansky, L. D., Kawachi, I., Weiss, S. T., & Sparrow,D. (1998). Anxiety and coronary heart disease: Asynthesis of epidemiological, psychological, and ex-perimental evidence. Annals of Behavioral Medicine,20, 47–58.

Laidlaw, T. M., Booth, R. J., & Large, R. G. (1996). Re-duction in skin reactions to histamine after a hyp-notic procedure. Psychosomatic Medicine, 58, 242–248.

Lambert, R. B., & Lambert, N. K. (1995). The effects ofhumor on secretory immunoglobulin A levels inschool-aged children. Pediatric Nursing, 21, 16–19.

Lazarus, R. S. (1991). Emotion and adaptation. Lon-don: Oxford University Press.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal,and coping. New York: Springer.


Levenson, R. W. (1994). The search for autonomic spec-ificity. In P. Ekman & R. J. Davidson (Eds.), The na-ture of emotion: Fundamental questions (pp. 252–257). New York: Oxford University Press.

Lewis, M., & Ramsay, D. (2002). Cortisol response toembarrassment and shame. Child Development, 73,1034–1045.

Maes, M., Scharpe, S., Meltzer, H. Y., Okayli, G.,Bosmans, E., D’Hondt, P., et al. (1994). Increasedneopterin and interferon-gamma secretion and loweravailability of L-tryptophan in major depression:Further evidence for an immune response. PsychiatryResearch, 54, 143–160.

Maier, S. F., & Watkins, L. R. (1998). Cytokines for psy-chologists: Implications of bidirectional immune-to-brain communication for understanding behavior,mood, and cognition. Psychological Review, 105,83–107.

McClelland, D. C., & Cheriff, A. D. (1997). Theimmunoenhancing effects of humor on secretory IgAand resistance to respiratory infections. Psychologyand Health, 12, 329–344.

McNally, R. J. (1999). Posttraumatic stress disorder. InT. Millon, P. H. Blaney, & R. D. Davis (Eds.), Oxfordtextbook of psychopathology (pp. 144–165). NewYork: Oxford University Press.

Miczek, K. A., Thompson, M. L., & Shuster, L. (1982).Opioid-like analgesia in defeated mice. Science, 215,1520–1522.


Miller, G. E., Stetler, C. A., Carney, R. M., Freedland, K.E., & Banks, W. A. (2002). Clinical depression andinflammatory risk markers for coronary heart dis-ease. American Journal of Cardiology, 90, 1279–1283.

Miller, T. Q., Smith, T. W., Turner, C. W., Guijarro, M.L., & Hallet, A. J. (1996). A meta-analytic review ofresearch on hostility and physical health. Psychologi-cal Bulletin, 119, 322–348.

Mittwoch-Jaffe, T., Shalit, F., Srendi, B., & Yehuda, S.(1995). Modification of cytokine secretion followingmild emotional stimuli. NeuroReport, 6, 789–792.

Ochsner, K. N. (2004). Current directions in social cog-nitive neuroscience. Current Opinion in Neurobiolo-gy, 14, 254–258.

Ottaway, C. A., & Husband, A. J. (1992). Central ner-vous system influences on lymphocyte migration.Brain, Behavior, and Immunity, 6, 97–116.

Parkinson, J. A., Cardinal, R. N., & Everitt, B. J.(2000). Limbic cortical–ventral striatal systems un-derlying appetitive conditions. Progress in Brain Re-search, 126, 263–285.

Phan, K. L., Wager, T. D., Taylor, S. F., & Liberzon, I.(2004). Functional neuroimaging studies of humanemotions. CNS Spectrums, 9, 258–266.

Phelps, E. A., & LeDoux, J. E. (2005). Contributions ofthe amygdala to emotion processing: From animalmodels to human behavior. Neuron, 48, 175–187.

Pressman, S., & Cohen, S. (2005). Does positive affectinfluence health? Psychological Bulletin, 131, 925–971.

Price, J., Sloman, L., Gardner, R., Jr., Gilbert, P., &Rohde, P. (1994). The social competition hypothesisof depression. British Journal of Psychiatry, 164,309–315.

Price, J. L. (2005). Free will versus survival: Brain sys-tems that underlie intrinsic constraints on behavior.Journal of Comparative Neurology, 493, 132–139.

Prkachin, K. M., & Silverman, B. E. (2002). Hostilityand facial expression in young men and women: Issocial regulation more important than negative af-fect? Health Psychology, 21, 33–39.

Quan, N., Avitsur, R., Stark, J., He, L., Shah, M.,Caliguiri, M., et al. (2001). Social stress increasessusceptibility to endotoxic shock. Journal of Neuro-immunology, 115, 36–45.

Ridker, P. M., Hennekens, C. H., Buring, J. E., & Rifai,N. (2000). C-reactive protein and other markers ofinflammation in the prediction of cardiovascular dis-ease in women. New England Journal of Medicine,342, 836–843.

Rogers, M. A., Bradshaw, J. L., Pantelis, C., & Phillips,J. G. (1998). Frontostriatal deficits in unipolar majordepression. Brain Research Bulletin, 47, 297–310.

Rozanski, A., Blumenthal, J. A., & Kaplan, J. (1999).Impact of psychological factors on the pathogenesisof cardiovascular disease and implications for ther-apy. Circulation, 99, 2192–2217.

Sapolsky, R. M. (1993). Endocrinology alfresco:

Psychoendocrine studies of wild baboons. RecentProgress in Hormone Research, 48, 437–468.

Sapolsky, R. M. (2005). The influence of social hierar-chy on primate health. Science, 308, 648–652.

Schleifer, S. J., Keller, S. E., & Bartlett, J. A. (2002).Panic disorder and immunity: Few effects on circulat-ing lymphocytes, mitogen response, and NK cell ac-tivity. Brain, Behavior, and Immunity, 16, 698–705.

Schleifer, S. J., Keller, S. E., Meyerson, A. T., Raskin, M.J., Davis, K. L., & Stein, M. (1984). Lymphocytefunction in major depressive disorder. Archives ofGeneral Psychiatry, 41, 484–486.

Schwartz, G. E., Weinberger, D. A., & Singer, J. A.(1981). Cardiovascular differentiation of happiness,sadness, anger, and fear following imagery and exer-cise. Psychosomatic Medicine, 43, 343–364.

Segerstrom, S. C., Glover, D. A., Craske, M. G., &Fahey, J. L. (1999). Worry affects the immune re-sponse to phobic fear. Brain, Behavior, and Immu-nity, 13, 80–92.

Segerstrom, S. C., & Miller, G. E. (2004). Psychologicalstress and the human immune system: A meta-analytic study of 30 years of inquiry. PsychologicalBulletin, 130, 601–630.

Segerstrom, S. C., Solomon, G. F., Kemeny, M. E., &Fahey, J. L. (1998). Relationship of worry to immunesequelae of the Northridge earthquake. Journal ofBehavioral Medicine, 21, 433–450.

Seligman, M. E. P. (2002). Positive psychology, positiveprevention, and positive therapy. In C. R. Snyder &S. J. Lopez (Eds.), Handbook of positive psychology(pp. 3–9). New York: Oxford University Press.

Selye, H. (1956). The stress of life. New York: McGraw-Hill.

Shestyuk, A. Y., Deldin, P. J., Brand, J. E., & Deveney,C. M. (2005). Reduced sustained brain activityduring processing of positive emotional stimuli inmajor depression. Biological Psychiatry, 57, 1089–1096.

Smith, T. W. (1994). Concepts and methods in the studyof anger, hostility, and health. In A. W. Siegman &T. W. Smith (Eds.), Anger, hostility and the heart(pp. 23–42). Hillsdale, NJ: Erlbaum.

Smith, T. W., & Gallo, L. C. (2001). Personality traits asrisk factors for physical illness. In A. Baum, T.Revenson, & J. Singer (Eds.), Handbook of healthpsychology (pp. 139–172). Mahwah, NJ: Erlbaum.

Smith, T. W., Glazer, K., Ruiz, J. M., & Gallo, L. C.(2004). Hostility, anger, aggressiveness, and coronaryheart disease: An interpersonal perspective on per-sonality, emotion, and health. Journal of Personality,72, 1217–1270.

Smith, T. W., & Ruiz, J. M. (2002). Psychosocial influ-ences on the development and course of coronaryheart disease: Current status and implications for re-search and practice. Journal of Consulting and Clini-cal Psychology, 7, 548–568.

Stark, J., Avitsur, R., Padgett, D. A., & Sheridan, J. F.(2001). Social stress induces glucocorticoid resistancein macrophages. American Journal of Physiology:


Regulatory, Integrative and Comparative Physiology,280, 1799–1805.

Stefanski, V., & Ben-Eliyahu, S. (1996). Social confron-tation and tumor metastasis in rats: Defeat and beta-adrenergic mechanisms. Physiology and Behavior,60, 277–282.

Stone, A. A., Cox, D. S., Valdimarsdottir, H., Jandorf,L., & Neale, J. M. (1987). Evidence that secretoryIgA antibody is associated with daily mood. Journalof Personality and Social Psychology, 52, 988–993.

Stone, A. A., Neale, J. M., Cox, D. S., Napoli, A.,Valdimarsdottir, H., & Kennedy-Moore, E. (1994).Daily events are associated with a secretory immuneresponse to an oral antigen in men. Health Psychol-ogy, 13, 440–446.

Strohle, A., & Holsboer, F. (2003). Stress responsiveneurohormones in depression and anxiety. Phar-macopsychiatry, 36, S207–S214.

Suarez, E. C., Lewis, J. G., & Kuhn, C. (2002). Therelation of aggression, hostility, and anger tolipopolysaccharide-stimulated tumor necrosis factor(TNF)-alpha by blood monocytes from normal men.Brain, Behavior and Immunity, 16, 675–684.

Suinn, R. M. (2001). The terrible twos—anger and anx-iety: Hazardous to your health. American Psycholo-gist, 56, 27–36.

Tangney, J. (1995). Recent advances in the empiricalstudy of shame and guilt. American Behavioral Scien-tist, 38, 1132–1145.

Taylor, S. E., & Armor, D. A. (1996). Positive illusionsand coping with adversity. Journal of Personality, 64,873–898.

Viswanathan, K., & Dhabhar, F. S. (2005). Stress-induced enhancement of leukocyte trafficking intosites of surgery or immune activation. Proceedings of

the National Academy of Sciences of the USA, 102,5808–5813.

Weiner, H. (1992). Perturbing the organism: The biol-ogy of stressful experiences. Chicago: University ofChicago Press.

Williams, R. B., Jr., Barefoot, J. C., & Shekelle, R.(1985). The health consequences of hostility. In M.A. Chesney & R. H. Rosenman (Eds.), Anger andhostility in cardiovascular and behavioral disorders(pp. 173–185). New York: Hemisphere.

Yehuda, R. (2000). Biology of posttraumatic stress dis-order. Journal of Clinical Psychiatry, 61, S14–S21.

Yehuda, R., Boisoneau, D., Mason, J. W., & Giller, E. L.(1993). Glucocorticoid receptor number and cortisolexcretion in mood, anxiety, and psychotic disorders.Biological Psychiatry, 34, 18–25.

Yirmiya, R. (1996). Endotoxin produces a depressive-like episode in rats. Brain Research, 711, 163–174.

Yirmiya, R., Weidenfeld, J., Pollak, U., Morag, M.,Morag, A., Avitsur, R., et al. (1999). Cytokines,“depression due to a general medical condition,”and antidepressant drugs. In R. Dantzer, E.Wollman, & R. Yirmiya, (Eds.), Cytokines, stressand depression (pp. 283–316). New York: KluwerAcademic.

Zachariae, R., Bjerring, P., Zachariae, C., Arendt-Nielsen, L., Nielsen, T., Eldrup, E., et al. (1991).Monocyte chemotactic activity in sera after hypnoti-cally induced emotional states. Scandinavian Journalof Immunology, 34, 71–79.

Zachariae, R., Jorgensen, M. M., Egekvist, H., &Bjerring, P. (2001). Skin reactions to histamine ofhealthy subjects after hypnotically induced emotionsof sadness, anger, and happiness. Allergy, 56, 734–740.


C H A P T E R 4 2

Health-Promotingand Health-Damaging Effects

of EmotionsThe View from Developmental Functionalism

NATHAN S. CONSEDINE

Nothing vivifies, and nothing kills, like the emotions.—JOSEPH ROUX

It may have taken the social sciences the bestpart of 200 years to subject Roux’s observationto systematic scrutiny, but the field of emotionsand health has gained increasing coherence andmomentum across the last few decades. Mon-ographs (Zautra, 2003), book chapters(Leventhal & Patrick-Miller, 2000; Mayne,2001), papers in special journal issues(Andersen, 2002; Hamilton & Malcarne,2004; Stefanek & McDonald, 2003; Zautra,Davis, & Smith, 2004), and review articles(Consedine, Magai, Krivoshekova, Ryzewicz,& Neugut, 2004; Mayne, 1999; Smith, Glazer,Ruiz, & Gallo, 2004) are appearing with in-creasing regularity. Admittedly, the field (andthe current chapter) might currently be bettertermed the study of emotions and disease, withone review suggesting that studies of negativeemotions outnumber those on positive emo-tions by 11:1 (Mayne, 1999). Nonetheless, anddespite differences in emphasis, numerous psy-chological, physiological, immunological, epi-

demiological, prevention, and intervention re-searchers have begun to document the perva-siveness with which emotions relate to healthoutcomes.

REVIEW OF THE LITERATURE

There are several ways in which one could re-view and arrange this blossoming literature—by disease outcome, by emotion, or by mecha-nism/pathway. In the interests of space, how-ever, and with an eye toward demonstratingthe pervasiveness of emotion’s relations withhealth outcomes, I have chosen to arrangestudies and theory relating emotions to healthoutcomes within a framework emphasizing fivepoints along a disease progression spectrum(see Figure 42.1): primary causation and pre-vention, secondary causation and prevention,symptom sensitivity, detection/screening, andtreatment choice/adherence. The current chap-

676

ter emphasizes these five more “psychological”aspects of emotions and health, and readers arereferred to other sources for more comprehen-sive discussions of the physiological, neuro-hormonal, and immunological effects of emo-tions (see, e.g., Kemeny & Shestyuk, Chapter41, this volume). Following a brief overview ofhow emotions relate to health at each point, abrief critique of the field as a whole is given,and areas for future research are suggested.

Primary Causative/Preventive Aspects

Greater levels of negative affect (NA) havebeen implicated in a full panorama of diseases,ranging from heart disease (Donker, 2000) andcancer (Reiche, Morimoto, & Nunes, 2006) toarthritis (Kiecolt-Glaser, McGuire, Robles, &Glaser, 2002), diabetes (Carnethon, Kinder,Fair, Stafford, & Fortmann, 2003), AIDS(Moscowitz, 2003), and the common cold (Co-hen et al., 1998). Most extensively researchedin the context of cardiovascular disease, suchtypes of NA as anger, anxiety, and depressionhave been implicated in basic disease processes(Krantz & McCeney, 2002; Sirois & Burg,2003), immunological functioning (Baum &Pozluszny, 1999), incidence (Rugulies, 2002),and early mortality (Harburg, Julius, Kaciroti,Gleiberman, & Schork, 2003). In a large num-ber of cases, the health-deleterious impact ofgreater negative emotion is found longitudi-nally (Kubzansky, Cole, Kawachi, Vokonas, &Sparrow, 2006), and even when health behav-iors are controlled for (Consedine, Magai, Co-

hen, & Gillespie, 2002). At this level, excessiveNA appears to promote disease and ill health.

Although positive affect (PA) is proportion-ally underrepresented in the literature, more re-cent data indicate that forms of PA may alsobe associated with health outcomes, includingmortality (see reviews in Chesney et al., 2005;Pressman & Cohen, 2005; Richman et al.,2005). PA predicted superior health acrossboth 5 weeks (Pettit, Kline, Gencoz, Gencoz, &Joiner, 2001) and 2 years (Ostir, Markides,Black, & Goodwin, 2000), even when baselinehealth, NA, health behaviors, and (in onestudy) self-report biases were controlled (Pettitet al., 2001). Global measures of PA have beenshown to predict lower readmission amonghospitalized older adults (Middleton & Byrd,1996) and lower mortality among patientswith HIV (Moskowitz, 2003), while the relatedconstruct of optimism may protect against pul-monary decline (Kubzansky et al., 2002). Hopehas been linked to increased survival time inpatients with cancer (Gottschalk, 1985), whilecuriosity has been linked to lower hypertensionand diabetes (Richman et al., 2005) and tolower mortality (Silvia, 2001; Swan &Carmelli, 1996). Recent work suggests that PAmay have health-promoting benefits for physi-ology in general (Fredrickson & Levenson,1998) and for hormone production and im-mune parameters in particular (Cohen, Doyle,Turner, Alper, & Skoner, 2003), or that positiveemotions may facilitate health-promoting cog-nitive styles (Tugade, Fredrickson, & Barrett,2004).

42. The View from Developmental Functionalism 677

FIGURE 42.1. Organizational model describing five points at which emotions can relate to health.

Secondary Causative/Preventive Aspectsof Behavior

A second point at which emotions influencehealth and disease processes lies in their rela-tions with behaviors that are often “aimed” atthe regulation of emotional states, but that alsopromote either health or disease (Brigham,Henningfield, & Stitzer, 1991). Although therole of NA in the initiation of unhealthy behav-ior profiles is less well established than its rolesin maintenance and relapse (see Kassel, Stroud,& Paronis, 2003, on this issue), a voluminousliterature implicates greater NA in several keyhealth behaviors—notably the use of cigarettes(Kassel et al., 2003), the consumption of high-fat/sugar “comfort” foods (Dube, LeBel, & Lu,2005), overeating (Arnow, Kenardy, & Agras,1995), drug/alcohol use (Wills, Sandy, Shinar,& Yaeger, 1999; Witkiewitz & Marlatt, 2004),and reduced exercise (Kawachi, Sparrow,Spiro, Vokonas, & Weiss, 1996).

As in other areas of emotions–health re-search, more work has been done regarding therole of NA than on that of PA. There are, how-ever, a growing number of exceptions (seePressman & Cohen, 2005). In one recent studyof 1,093 women participating in a worksitehealth promotion intervention, PA and copingboth predicted self-reported health andexercise (Kelsey, DeVellis, Begum, Belton, &Hooten, 2006). Other work has linked greaterPA to greater levels of exercise and better nutri-tion (Griffin, Friend, Eitel, & Lobel, 1993), al-though the motivation to up-regulate PA hasbeen linked to several risky behaviors, includ-ing unprotected sex (Kalichman et al., 1994)and preferences for stimulant use (Adams et al.,2003).

Symptom Perception and Sensitivity

Third, aspects of emotions appear related to re-ports of health symptoms and to symptom sen-sitivity. One line of research suggests that atleast a part of the general relation betweenforms of NA and poorer health reflects differ-ences in the tendency to attend to (or complainabout) somatic symptoms—a methodologicalnuisance factor (see reviews in Van Diest et al.,2005; Watson & Pennebaker, 1989). Consis-tent with this view, several longitudinal studieshave shown state NA, to be a better predictorof self-reported health symptoms than trait

NA (Brown & Moskowitz, 1997; Charles &Almeida, 2006; Leventhal, Hansell, Diefen-bach, Leventhal, & Glass, 1996).

Others have, however, argued that trait NAis not necessarily most usefully viewed as a bi-asing factor in reports of illness or healthcare (Mora, Robitaille, Leventhal, Swigar, &Leventhal, 2002). According to these authors,older adults with high trait NA—particularlyanxiety—will be cognitively more attentive totheir somatic state and have greater expertise indifferentiating current symptoms from theirbaseline somatic background (Mora et al.,2002). Such a claim is consistent with otherwork suggesting that nondepressed peoplegenerally tend to underreport disease symp-tomatology, while depressed persons reportsymptoms more accurately (see Mayne, 2001).The effect of NA on reporting may also varywith the specificity of the symptoms being as-sessed; “vague,” systemic symptoms appear tobe more strongly related than local, specificsymptoms to NA (Van Diest et al., 2005). Posi-tive emotions have scarcely been researched atthis interface. Although PA is thought to pro-mote a number of advantageous perceptualand coping changes (see, e.g., Folkman &Moskowitz, 2004; Fredrickson & Losada,2005), how (if at all) such advantages might beevident in the context of symptom perceptionand sensitivity is presently unclear.

Detection and Screening Behaviors

One of the most active areas of behavioralemotions–health research focuses on howemotions—particularly fear/anxiety, depres-sion, and embarrassment—relate to the fre-quency of medical contacts, as well as to de-tection and screening behaviors. Clinicallysignificant levels of NA generally predict morefrequent contact with physicians, even whenhealth status is controlled for; this finding is bestdocumented in the case of depression and anxi-ety (Tweed, Goering, Lin, & Williams, 1998).Distress and/or depression have been linked tomore frequent physician visits (Callahan, Hui,Nienaber, Musick, & Tierney, 1994; Karlsson,Lehtinen, & Joukamaa, 1995), perhaps morestrongly among women (Larsen, Schroll, &Avlund, 2006). Discrete emotions have also be-gun to be studied in this context. A recent studyof 345 older adults found that anger, but notfear, predicted increases in the number of physi-


cians’ visits but not admissions; sadness inter-acted with gender such that it related to moreadmissions for men, but fewer for women(McKeen, Chipperfield, & Campbell, 2004).

A large literature has examined how fear-related constructs relate to cancer screening(Dale, Bilir, Han, & Meltzer, 2005; Hay,Buckley, & Ostroff, 2005). In general, greateranxiety or worry appears to facilitate screeningamong women, at least when operationalizedin terms of trait anxiety or fear of cancer(Consedine, Magai, Krivoshekova, et al.,2004), and trait anxiety appears to predictmore frequent prostate screening among men(Dale et al., 2005). It is worth noting, however,that research in both men’s (Consedine,Morgenstern, Kudadjie-Gyamfi, Magai, &Neugut, 2006) and women’s (Andersen, Smith,Meischke, Bowen, & Urban, 2003; Diefen-bach, Miller, & Daly, 1999) screening hasfound evidence of nonlinearities in the relationbetween anxiety and screening (Miller &Diefenbach, 1998). Such findings suggest thatthere may be “threshold” rather than “dose”effects for anxiety’s relation to screening.

Whereas anxiety generally appears to en-hance the frequency of health-promoting medi-cal contacts, other types of NA appear to havemore complex (and potentially less salutary) re-lations with health. Although depression ap-pears to predict more frequent medical contactin general, it has also been associated with lessfrequent mammography (Husaini et al., 2001),and a greater likelihood of late-stage breast can-cer diagnosis (Desai, Bruce, & Kasl, 1999), al-though at least one study found no relation withPap smear frequency (Pirraglia, Sanyal, Singer,& Ferris, 2004). Greater embarrassment hasbeen found to relate negatively to general medi-cal visits (Consedine, Krivoshekova, & Harris,2007), dental visit frequency (Moore, Brodsgaard,& Rosenberg, 2004), and help seeking for incon-tinence (Hagglund, Walker-Engstrom, Larsson,& Leppert, 2003), as well as to be associatedwith less frequent breast (Consedine, Magai, &Neugut, 2004), testicular (Gascoigne, Mason,& Roberts, 1999), cervical (Taylor et al., 2002),and colon (Farraye et al., 2004) cancer screen-ings.

Treatment Decision Making and Adherence

Persons diagnosed with a serious health threator disease are confronted with a bewildering

array of information, advice, demands, andtreatment options (Lam, Fielding, Chan, Chow,& Or, 2005). Having eventually embarked on atreatment course and experienced any possiblebenefits and side effects, such persons mustthen decide whether to persist with the adoptedtreatment/adhere to physician recommenda-tions. Both initial treatment decision making(TDM) and adherence are inevitably complexand highly demanding processes that naturallylend themselves to emotions-based analyses.

Despite the presence of multiple, unclear,and/or competing goals in TDM contexts(Broadstock & Michie, 2000), most studies ofTDM assume a high degree of rationality (Lamet al., 2005). The consequence of this orienta-tion is that despite the importance of antici-pated benefits and costs of treatment options(Lam et al., 2005), including emotional costsand benefits, there are few studies of emotionsin TDM. Moreover, such research has tendedto conceptualize affect as an outcome factorthat must be weighed against treatment effi-cacy.

In some of the few studies to date of TDMand emotions, women reporting greater intru-sive thinking and cancer worry are more likelyto choose tamoxifen treatment for breast can-cer risk (Bober, Hoke, Duda, Regan, & Tung,2004), and health-related quality of life hasbeen related to the decision to modifychemotherapy treatment (Detmar, Muller,Schnornagel, Wever, & Aaronson, 2002).Other work suggests that anticipated conse-quences of treatment decisions with a high af-fective content (e.g., losing hair, weight gain)are influential in determining adjuvant breastcancer treatment decisions (Lam et al., 2005),and it has been argued that anxiety and depres-sion interfere with TDM (Petersen et al., 2003).

One interesting question that emerges in thisresearch lies in examining exactly what com-prises a “good decision.” Research conductedwithin the “affective heuristic” (Slovic, Peters,Finucane, & MacGregor, 2005) or “risk-as-feelings” (Loewenstein, Weber, Hsee, & Welch,2001) models discriminates between fast, in-stinctive, and intuitive reactions to threat (riskas feelings) and more effortful, logical, and rea-soned efforts to manage risk (risk as analysis).Importantly for TDM researchers, factors suchas age, cognitive load, stress, and ill health mayincrease the reliance on less analytic and moreaffectively driven styles (Slovic et al., 2005). In


the context of cancer TDM, this might producesituations in which high levels of cancer-relatedanxiety and restrictive regulatory styles—evident in several minority groups—promoterapid (or even preemptive) TDM as the processbecomes more heavily influenced by affectiveconsiderations.

While treatment adherence is inherently abehavioral and motivational issue, there hasbeen comparatively little research on the role ofemotions in adherence. Depression has been re-lated to poorer adherence to medications forhypertension (Carney et al., 1998; Kaplan,Bhalodkar, Brown, Wflite, & Brown, 2004;Wang et al., 2002) and HIV (Van Servellen,Chang, Garcia, & Lombardi, 2002), as well asto dialysis (Brownbridge & Fielding, 1994),and oral hypoglycemic treatments for diabetes(Kalsekar et al., 2006); it has also been linkedto withdrawal from dialysis (McDade-Montez,Christensen, Cvengros, & Lawton, 2006).It remains unclear whether such effects arecaused by greater sensitivity to side effects orare a consequence of intense emotions’ dom-inating coping efforts at the expense ofproblem-focused efforts (see Cameron, Petrie,Ellis, Buick, & Weinman, 2005).

The relations between anxiety and ad-herence appear complex. Perhaps surprisingly,anxiety has been shown to predict lower adher-ence to HIV medication (Van Servellen et al.,2002) and dialysis (Brownbridge & Fielding,1994) regimens, although few studies control-ling for the large number of confounds in ad-herence have been conducted. A meta-analysisof these two affects (DiMatteo, Lepper, &Croghan, 2000) found that the association be-tween anxiety and nonadherence was variable,but reported a strong negative relation betweendepression and nonadherence (odds ratio [OR]= 3.03).

Again, PA remains understudied. One studyreported that positive states of mind predictedadherence to HIV treatment and partially me-diated the relation between social support andadherence (Gonzalez et al., 2004), althoughother work has suggested that the optimism–adherence link may be nonlinear for anti-retroviral therapy adherence (Milam, Richard-son, Marks, Kemper, & McCutchan, 2004).Another study suggested that lower quality oflife marginally predicted (OR = 1.3) tamoxifen(a breast cancer treatment) discontinuance overa 2-year period (Fink, Gurwitz, Rakowski,Guadagnoli, & Silliman, 2004).

A CRITIQUE ON THE STATEOF THE ART

Although emotions–health research has madeconsiderable progress in recent years, a numberof key issues remain unclear. Although some ofthese issues are empirical and are likely to beclarified by ongoing research, several pressingconcerns are theoretical and stem from thefield’s historical origins (Donker, 2000; Fried-man & Booth-Kewley, 1987; Smith, 1992) andongoing applied focus. It has been noted thatmany datasets used to examine emotion–healthrelations were “pressed into service” after thefact, with the consequences that basic psycho-metric issues are frequently ignored (Smith,1992) and observations are infrequently placedwithin broader theoretical contexts (Friedman& Booth-Kewley, 1987; Miller et al., 2004).Below, a selection of the global issues confront-ing emotions and health research are described,and some suggestions regarding the field’s de-velopment are given.

Function

The emotions are thought to represent adapta-tions that evolved over time to deal with recur-rent adaptive challenges or opportunities, andthus they are viewed as serving key social,physiological, cognitive, and motivationalfunctions (Consedine & Magai, 2006; Frijda,1994; Keltner & Gross, 1999). Whether healthmay represent one of these challenges is un-clear, as is the question of whether the emo-tions have functions that are specifically relatedto health or whether the relations betweenemotions and health arise epiphenomenallywith respect to non-health-related functions(e.g., Mayne, 2001). There is a marked dif-ference between an emotion’s having conse-quences for health and the emotion’s havingevolved because of these consequences(Consedine, Magai, & Bonanno, 2002).

Few of the discrete emotions thus far consid-ered by health researchers have any obvioushealth-related functions. Nonetheless, under-standing what emotions are designed or“meant” for is critical to the advancement ofthe field, because it provides guidance in thesearch for mechanisms. Because each emotionhas evolved to solve specific adaptive chal-lenges, each involves a coordinated responseamong physiological (Larsen, Berntson,Poehlmann, Ito, & Cacioppo, Chapter 11, this


volume), cognitive (Stein, Hernandez, &Trabasso, Chapter 35, this volume), and moti-vational (Consedine, Magai, Krivoshekova, etal., 2004), systems that, in total, is aimed to-wards addressing that challenge. Understand-ing the purpose of emotional response sets al-lows researchers to “fill” the explanatory gapbetween emotions and health outcomes in asystematic and theoretically driven way.

Mechanisms

Perhaps because emotions are so central to hu-man functioning, they appear to affect healthprocesses at a large number of points across thehealth–disease spectrum; the mechanisms link-ing them are multiple and complex. Below,three major classes of mechanisms are brieflyreviewed, with an eye toward considering howdescriptions of the functions of emotions in re-gard to three different components (physiologi-cal, experiential/motivational, and cognitive)may supplement existing accounts.

Historically, the mechanisms thought to ex-plain the link between emotions and healthhave been classed as either “direct” or “in-direct,” depending on whether a particularmodel suggests that emotions have a direct im-pact on health via their role in generating auto-nomic and immunological changes, or whetherthey are linked through mediated (indirect)pathways. One grouping of direct-pathwayexplanations—the arousal, accumulation, or“wear-and-tear” models—suggests that the au-tonomic and immunological changes associ-ated with emotions such as anxiety and angerare what create health risk; such a view (albeitin reverse) is also evident in recent modelsexamining positive emotions (Fredrickson &Losada, 2005; Moskowitz, 2003; Tugade et al.,2004). Although this approach may have moredifficulty in accounting for direct links betweendepressed (nonarousing) affects and health, thegeneral notion that the activation of the neg-ative emotions was historically adaptive(Levenson, 1994), but has become vestigial andnow damages and/or impairs autonomic andimmune systems when repeated or chronic, hasgarnered an impressive body of support(Mayne, 2001).

A second grouping of models conceptualizesthe emotion–health relation as indirect, arisingpredominantly as a result of emotion’s strongmotivational functions. Models examining theeffects of emotions on health via health behav-

iors, screening, and adherence have tended tobe motivationally hedonic (e.g., Chapman &Coups, 2006; Consedine, Magai, Krivoshe-kova, et al., 2004); health-relevant behaviorsare argued to occur in order to regulate eitherexperienced or anticipated affect. For example,many behaviors with implications for poorerhealth (e.g., poor dietary preferences; cigarette,alcohol, and drug use) are enacted because theyhelp people minimize the experience of nega-tive emotions and maximize the experience ofpositive emotions.

However, in the context of detection,screening, and adherence behaviors, experi-enced negative emotion appears to have di-verse effects on health behavior, depending onthe affect in question. It is at this point thatthe functionalist approach to discrete emo-tions becomes most useful. Moderate levels offear/anxiety, be they general or specific to acondition, generally appear to promotehealth-enhancing behavior profiles. Func-tionally, when people are anxious, their cog-nitive processes and motivational prioritiesshift. They begin to evaluate the environmentfor the source of the threat and to engage ininformation- and support-seeking behaviors,as well as in behaviors that consciously orunconsciously serve to ameliorate the un-pleasant experience of anxiety (Consedine,Magai, Krivoshekova, et al., 2004). Thus, de-spite the fact that fear did not evolve in orderto increase medical care seeking or screening,fear of a disease outcome appears to motivatebehaviors that reduce risk so long as efficacyrequirements are met (Mayne, 2001).

Conversely, emotional syndromes such asdepression have been generally linked tohealth-deleterious behaviors, poorer cancerscreening, and poorer adherence. Given thatdepressed affects such as sadness induce passiv-ity rather than overt activity (Brehm, 1999),and are thought to function by motivating anindividual to cease directing effort toward alost goal (Campos & Barrett, 1984)—thus con-serving systemic resources (Clark & Watson,1994)—these relationships are perhaps not sur-prising. Given that depressed affect naturallylends itself to behavioral passivity, future re-search might benefit from considering the na-ture of these “lost” goals, examining whetherthey relate to health states, and consideringhow best to activate competing goal and affec-tive states that facilitate less unhealthy behav-ioral profiles.


A final pathway that has received increasingattention in recent times has involved the con-sideration of how emotions compete with, re-late to, and are manifest in health-relevantcognitive processes. For health researchers in-terested in emotions, engaging at this interfacehas been a necessary step, as most psychologi-cal models of health behavior remain heavilycognitive in orientation (Jessop, Rutter,Sharma, & Albery, 2004; Ogden, 2003; Zautraet al., 2004)—stressing characteristics such asperceptions of threat and risk, illness represen-tations, beliefs, knowledge, self- and treatmentefficacy, and outcome expectancies. Some theo-ries view emotions as affecting health onlythrough cognitive pathways (e.g., Witte, 1998);others view cognitive factors as operating viaaffective channels (Cameron & Diefenbach,2001; Chapman & Coups, 2006; Loewensteinet al., 2001; Mellers & McGraw, 2001); andstill others incorporate both emotional andcognitive elements (Cameron & Leventhal,1995; Leventhal, Diefenbach, & Leventhal,1992; Miller & Diefenbach, 1998). Clearly,there exists a recurrent, if unilluminating, ten-sion between cognitive and affective ap-proaches to human behavior.

In many ways, however, such competition isunnecessary for emotion researchers. It is wellestablished within emotion theory that manyfunctions are enacted through cognitive chan-nels (Keltner, Ellsworth, & Edwards, 1993);thus emotions are likely to affect health behav-iors via cognitive pathways. Take an emotionsuch as sadness. In addition to motivationallypreventing continued investment in a “lostcause,” sadness produces “dwelling” on thelost object—a cognitive process that allows theorganism time for introspection, rearrange-ment of compromised goal commitments, andconstruction of contingency-based plans todeal with the loss (Consedine & Magai, 2006).

In regard to forms of NA, it has been notedabove that the relation between symptom re-porting and greater NA—perhaps particularlydepression—may arise because depressed affecteither promotes introspection and greater at-tention to symptoms (Mora et al., 2002) oreliminates positivistic biases (Mayne, 2001),perhaps leading to a greater likelihood of inter-preting symptoms as threatening and respond-ing with care-seeking behavior. Conversely,however, depressed affect is known to be asso-ciated with hopeless/helplessness, which mayequate with low self- and treatment efficacy.

This relation may help explain the links be-tween depression on the one hand and lowscreening (Husaini et al., 2001) and poor ad-herence (DiMatteo et al., 2000) on the other.

It has been suggested that in contrast to neg-ative emotions, which tend to narrow atten-tional focus, PA may have health benefits inbroadening the scope of attention (Fredrickson& Losada, 2005), producing thorough yet flex-ible information-processing strategies (Isen,1999), improving outcome expectancies (Erez& Isen, 2002), and improving medical diagnos-tic reasoning (Isen, Chapter 34, this volume).Although it has been suggested that cognitiveinterventions have proven comparatively inef-fective in areas such as adherence (Beswick etal., 2005), the role of cognitive processes inhealth behavior is a key one, and continuedwork examining how emotions influence or aremanifest in health-promoting or health-deleterious cognitive processing is clearlyneeded.

Specificity: I. Discrete Emotions

Examining studies of the emotions–health in-terface while recalling their evolutionary designtends to suggest that research on more spe-cific classes of affect—discrete emotions—willprove critical in disentangling the complexlinks between emotions and health. Althoughmost research has concentrated on broad affec-tive constructs—anxiety, depression, andanger/hostility (Dickerson, Gruenewald, &Kemeny, 2004)—theory regarding the func-tions of discrete emotions provides guidancefor understanding their likely physiological,motivational, and cognitive correlates, andthus their impact on health outcomes. Table42.1 summarizes the current literature regard-ing 10 discrete emotions taken from Izard(1991), as well as embarrassment and globalPA and NA.

As can be seen in this summary, global PAand NA as well as fear/anxiety and sadness/de-pression have been the most systematically ex-amined affects across the disease spectrum. Joy,interest, shame, and particularly anger/hostilityhave received some attention, although onlyin primary and secondary causative arenas.The research on other discrete emotions—contempt, disgust, embarrassment, guilt, andsurprise—either remains isolated in only con-sidering a specific intersection or does not existat this point. Although relations between con-


tempt and surprise appear unlikely from a the-oretical point of view, there are good reasons tosuspect that disgust, embarrassment, and guiltshould relate to health outcomes. These threeemotions are discussed briefly below.

Disgust

Despite the fact that disgust is the only discreteemotion with clear health-related functions, ithas received almost no empirical attention.Fundamentally, disgust functions to obviatethreats to physical health by encouraging spit-ting or the ejection of a noxious substance(Rozin & Fallon, 1987; Rozin, Lowery, Imada,& Haidt, 1997), and it is associated with a de-sire to avoid contamination (Davey & Bond,2006). Disgust may influence dietary choices(Hamilton, 2006), obsessional behavior(Rachman, 2004), and sexual symptom report-ing (Berman et al., 2003).

Guilt

A second emotion that theory implies shouldhave relevance for the cessation of unhealthybehaviors and the motivation of healthy behav-iors is guilt (Lindsey, 2005). Experiences ofguilt serve a key function in moving an individ-

ual toward positive reparation and change(Dearing, Stuewig, & Tangney, 2005). A fewstudies have linked guilt to exercise motivation(Eyler & Vest, 2002), to attempts to make sub-sequent dietary or exercise-related improve-ments (Sukhdial & Boush, 2004), and toreduced odds of substance abuse problems(Dearing et al., 2005).

Embarrassment

Finally, because embarrassment is thought tohave evolved to prevent social ostracism arisingfrom norm violations, awkward interactions,and negative evaluations, the experience of em-barrassment appears to motivate avoidance ofpotential eliciting situations (Consedine et al.,2007). Failures in privacy regulation constitutea key source of embarrassment (Keltner & An-derson, 2000); thus many medical contexts(particularly cancer screenings, which ofteninvolve intimate examinations) are avoided.However, because embarrassment also arisesdue to physical inadequacies (Keltner &Buswell, 1996), such as bad breath or beingobese, anticipated embarrassment may poten-tially lead individuals to engage in health-promoting behaviors that prevent embarrass-ment.


TABLE 42.1. Summary of What Current Data Suggest Regarding How Two Global Affects and 11Discrete Emotions Relate to Health at Five Points of Health Interface

Affective construct

Point of interface

Primarycausative

Secondarycausative

Symptomsensitivity

Detectionand screening

Decisionsand adherence

Global NA – – – – + + / – + / –

Global PA + + + / – ? ? +

Fear/anxiety – + / – + / – + / – + / –

Anger/hostility – – – – ? ? ?

Contempt ? ? ? ? ?

Disgust ? ? ? ? ?

Sadness/depression – – – – + – – – –

Shame – – – ? ? ?

Embarrassment ? ? ? + / – ?

Guilt ? + / – ? ? ?

Joy + + + ? ? ?

Interest + / – + / – ? ? ?

Surprise ? ? ? ? ?

Note. ++, clear suggestion of health-promoting effects; – –, clear suggestion of health-deleterious effects; +, preliminary datasuggesting some health-promoting impact; –, preliminary data suggesting some health-deleterious impact; +/–, mixed effects.

Specificity: II. Discrete Elicitors

Understanding the role of emotions in healthbehavior is predicated on a clear understandingof the emotions’ elicitors. Specifically, becauseof the likely motivational and cognitive path-ways linking several discrete emotions tobehavioral outcomes, knowing exactly what itis that (for example) frightens or embarrasses aperson is probably critical in understandingwhen emotions will and will not promotehealth (Consedine et al., 2007). In the contextof breast cancer screening, for example, it hasbeen argued that fears pertaining to the screen-ing context (e.g., fear of pain or of the medicalestablishment) will produce avoidance of thefear-inducing situation, and screening willtherefore be infrequent. Conversely, wherefears relate to the disease and expectancies arethat screening will reduce anxiety, greater feargenerally facilitates screening (Consedine,Magai, Krivoshekova, et al., 2004). Similarly,although experienced or anticipated embar-rassment is typically a barrier to medical con-tacts (see above), it has been argued that under-standing exactly what embarrasses people (andthus deters them from medical contacts) is anecessary precursor to developing interven-tions designed to reduce the impact of this bar-rier (Consedine et al., 2007). In differentiatingamong elicitors, this study also demonstratedthat some elements of embarrassment fa-cilitated certain classes of medical contact(Consedine et al., 2007). Future research thatfurther discriminates among the elicitors ofemotions as they relate to health is likely toprove of equal benefit.

Anticipated versus Actual Emotions

The functionalist view of emotions reminds usthat emotions can be motivational in terms ofboth experience and anticipation. Some infre-quently studied discrete emotions (such asshame, guilt, and embarrassment) are thoughtto motivate prosocial choices predominantly inanticipation—through prompting behaviorsthat prevent the occurrence of these emotions(Frijda, 1994), and by motivating the acquisi-tion of experiences, skills, or attributes thatforeclose on potentially unpleasant experiences(Izard, 1991). This perspective on emotion co-heres nicely with the growing field of “affectiveforecasting,” which suggests that despite the

ultimate inaccuracy of such “forecasts” (Con-nolly & Reb, 2005), people nonetheless payconsiderable attention to the anticipated affec-tive consequences when making choices (Chap-man & Coups, 2006), and to the inherenttradeoffs between short- and long-term bene-fits (as in the case of diet or exercise; Chapmanet al., 2001). In one recent study of worry, re-gret, and influenza vaccination decisions, forexample, it was found that persons who didnot obtain vaccinations in Year 1 were morelikely to get vaccinated in a subsequent year ifthey anticipated and experienced high levels ofworry and regret (Chapman & Coups, 2006).Differentiating the relative contribution of ex-perienced and anticipated affect to health be-haviors, screening, and TDM, as well as exam-ining how the predicted relations evolve withthe experience of estimation–outcome contin-gencies, will be exciting areas for future re-search.

Model Generalizability

Finally, it is worth noting that few studies ex-amining emotion–health links have specificallyconsidered whether the relations between emo-tions and health outcomes are the same acrossgroups (Consedine, Magai, & Bonanno, 2002).There are questions regarding model general-izability across genders (Van Diest et al., 2005)and ethnicities (Fredrickson et al., 2000). In myown opinion, sufficient data have been pro-vided to make it seem likely that cultural vari-ables moderate many vectors linking emotionsand health. In one study, while NA predictedsleep disturbance in all groups, the effectwas reduced among black English-speakingCaribbeans (Consedine, Magai, Cohen, et al.,2002). Other work found that although greateranger predicted reports of worse health in gen-eral, anger was related to better health in sev-eral groups of minority women (Consedine,Magai, & Horton, 2005). Conversely, whiletrait sadness was positively associated with so-matic symptoms among U.S.-born white menand Dominican men, they were negatively as-sociated with symptoms among black English-speaking Caribbean men, and the relation foranger also differed marginally across groups(Consedine, Magai, et al., 2006).

In addition, there may well be gender differ-ences in the relations between negative emo-tions and disease or symptom clusters. One


study of 1,017 university students found thatdifferent symptom characteristics—severity,and somatic versus psychic nature—were dif-ferentially related to NA in men and women(Van Diest et al., 2005). Specifically, high-NAwomen tended to report more severe symp-toms, while high-NA men tended to reportmore common, nonspecific symptoms. Otherresearch based in emotion theory suggests thatthe emotion/personality variables that best ex-plained variance in cardiovascular diseasescores differed between men and women; anxi-ety predicted disease for women, and hostilitypredicted disease for men (Consedine, Magai,& Chin, 2004).

In part because the mechanisms linking emo-tions and health outcomes are so complex,there is little theory to explain why global NAand some discrete emotions appear to pose dif-ferent risks to persons from different ethnicand gender groups. We have argued that recon-ciling these complex data is likely to requireconsidering whether the presence or absence ofan emotion is telling us the same thing aboutpersons from different ethnic groups(Consedine, Magai, et al., 2006). Cultural psy-chology suggests that the experiences of emo-tions such as anger in persons among differentcultures and contexts (gender, nationality, de-velopmental, religious, socioeconomic) may bequite different (Mesquita & Leu, 2007). Forexample, the discrepancy between ideals andactual emotion may be as informative in termsof health as the emotions themselves (see Tsai,Knutson, & Fung, 2006). The extent to whichemotions “fit” their sociocultural contexts maybe an important determinant of their impact onhealth—or, conversely, reports of emotions thatare infrequent within such contexts (e.g., sad-ness for men, anger for women) may be index-ing a more balanced emotional repertoire thatpredicts superior health outcomes (Consedine,Magai, et al., 2006). Although such competingexplanatory possibilities have yet to be exam-ined, the fact that models relating emotionsand health do not appear to generalize easilyacross groups illuminates inadequacies in exist-ing theory and suggests that our understandingof mechanisms is probably incomplete. Giventhe disparities that exist in most domains ofhealth, research that systematically tests mod-els across age, gender, ethnic, cultural, na-tional, and socioeconomic groups is clearlyneeded.

CONCLUDING REMARKS

Human lives are born, lived, and ended in thepresence of emotions. Emotions act as a majorinterface among the intrapersonal, sociocul-tural, and biological spheres of human func-tioning, providing a near-seamless bridge be-tween person and environment. As such, thefact that they relate to a large number of healthbehaviors and health outcomes is not surpris-ing. What have perhaps been less readily fore-seen in the relations between emotions andhealth are the sheer pervasiveness of the links,the manifold pathways, and the litany of likelymechanisms and moderators. Although nega-tive emotions are broadly thought to causeworse health (and positive emotions betterhealth), the review provided above suggeststhat matters are vastly more complex—withvariations arising depending on whether oneconsiders direct physiological pathways, healthbehaviors, symptom sensitivity, screening anddetection, or TDM and adherence phases ofhealth and disease processes.

Comprehending the mechanisms underlyingthese links is a process that emotion researchershave only just begun, and a number of areaswould benefit from further research. Some ofthese have been described above. Nonetheless,there exist considerable grounds for encourage-ment. Emotion–health researchers have accessto a strongly theoretical tradition that describesthe impact of emotions on physiological, moti-vational, cognitive, and behavioral systems,and that can thus inform their search for mech-anisms.

ACKNOWLEDGMENTS

I would like to thank Carol Magai, Ratja Mesquita,and Judith Moskowitz for their helpful comments onearlier versions of this chapter.

REFERENCES

Adams, J. B., Heath, A. J., Young, S. E., Hewitt, J. K.,Corley, R. P., & Stallings, M. C. (2003). Relation-ships between personality and preferred substanceand motivations for use among adolescent substanceabusers. American Journal of Drug and AlcoholAbuse, 29, 691–712.

Andersen, B. L. (2002). Biobehavioral outcomes follow-ing psychological interventions for cancer patients.


Journal of Consulting and Clinical Psychology, 70,590–610.

Andersen, M. R., Smith, R., Meischke, H., Bowen, D.,& Urban, N. (2003). Breast cancer worry and mam-mography use by women with and without a familyhistory in a population-based sample. Cancer Epide-miology, Biomarkers, and Prevention, 12, 314–320.

Arnow, B., Kenardy, J., & Agras, W. S. (1995). TheEmotional Eating Scale: The development of a mea-sure to assess coping with negative affect by eating.International Journal of Eating Disorders, 18, 79–90.

Baum, A., & Pozluszny, D. M. (1999). Health psychol-ogy: Mapping biobehavioral contributions to healthand illness. Annual Review of Psychology, 50, 137–163.

Berman, L., Berman, J., Felder, S., Pollets, D., Chhabra,S., Miles, M., et al. (2003). Seeking help for sexualfunction complaints: What gynecologists need toknow about the female patient’s experience. Fertilityand Sterility, 79, 572–576.

Beswick, A. D., Rees, K., West, R. R., Taylor, F. C.,Burke, M., Griebsch, I., et al. (2005). Improving up-take and adherence in cardiac rehabilitation: Litera-ture review. Journal of Advanced Nursing, 49, 538–555.

Bober, S. L., Hoke, L. A., Duda, R. B., Regan, M. M., &Tung, N. M. (2004). Decision-making about tamoxi-fen in women at high risk for breast cancer: Clinicaland psychological factors. Journal of Clinical Oncol-ogy, 22, 4951–4957.

Brehm, J. W. (1999). The intensity of emotion. Personal-ity and Social Psychology Review, 3, 2–22.

Brigham, J., Henningfield, J. E., & Stitzer, M. L. (1991).Smoking relapse: A review. International Journal ofthe Addictions, 25, 1239–1255.

Broadstock, M., & Michie, S. (2000). Processes of pa-tient decision making: Theoretical and methodologi-cal issues. Psychology and Health, 15, 191–204.

Brown, K. W., & Moskowitz, D. S. (1997). Does unhap-piness make you sick?: The role of affect and neuroti-cism in the experience of common physical symp-toms. Journal of Personality and Social Psychology,72, 907–917.

Brownbridge, G., & Fielding, D. M. (1994). Psycho-social adjustment and adherence to dialysis treatmentregimes. Pediatric Nephrology, 8, 744–749.

Callahan, C. M., Hui, S. L., Nienaber, N. A., Musick, B.S., & Tierney, W. M. (1994). Longitudinal study ofdepression and health-services use among elderlyprimary-care patients. Journal of the American Geri-atrics Society, 42, 833–838.

Cameron, L. D., & Diefenbach, M. A. (2001). Re-sponses to information about psychosocial conse-quences of genetic testing for breast cancer suscep-tibility: Influences of cancer worry and riskperceptions. Journal of Health Psychology, 6, 47–59.

Cameron, L. D., & Leventhal, H. (1995). Vulnerabilitybeliefs, symptom experiences and the processing ofhealth threat information: A self-regulatory perspec-

tive. Journal of Applied Social Psychology, 25, 1859–1883.

Cameron, L. D., Petrie, K. J., Ellis, C. J., Buick, D., &Weinman, J. A. (2005). Trait negative affectivity andresponses to a health education intervention for myo-cardial infarction patients. Psychology and Health,20, 1–18.

Campos, J. J., & Barrett, K. C. (1984). Toward a newunderstanding of emotions and their development. InC. E. Izard, J. Kagan & R. B. Zajonc (Eds.), Emo-tions, cognition, and behavior (pp. 229–263). Cam-bridge, UK: Cambridge University Press.

Carnethon, M. R., Kinder, L. S., Fair, J. M., Stafford, R.S., & Fortmann, S. P. (2003). Symptoms of depres-sion as a risk factor for incident diabetes: Findingsfrom the National Health and Nutrition Examina-tion Epidemiologic Follow-Up study, 1971–1992.American Journal of Epidemiology, 158, 416–423.

Carney, R. M., Freedland, K. E., Eisen, S. A., Rich, M.W., Skala, J. A., & Jaffe, A. S. (1998). Adherence to aprophylactic medication regimen in patients withsymptomatic versus asymptomatic ischemic heartdisease. Behavioral Medicine, 24, 35–39.

Chapman, G. B., Brewer, N. T., Coups, E. J., Brownlee,S., Leventhal, H., & Leventhal, E. A. (2001). Valuefor the future and preventive health behavior. Journalof Experimental Psychology: Applied, 7, 235–250.

Chapman, G. B., & Coups, E. J. (2006). Emotions andpreventive health behavior: Worry, regret, and influ-enza vaccination. Health Psychology, 25, 82–90.

Charles, S. T., & Almeida, D. M. (2006). Daily reportsof symptoms and negative affect: Not all symptomsare the same. Psychology and Health, 21, 1–17.

Chesney, M. A., Darbes, L. A., Hoerster, K., Taylor, J.M., Chambers, D. B., & Anderson, D. E. (2005).Positive emotions: Exploring the other hemisphere inbehavioral medicine. International Journal of Behav-ioral Medicine, 12, 50–58.

Clark, L. A., & Watson, D. (1994). Distinguishing func-tional from dysfunctional affective bursts. In P.Ekman & R. J. Davidson (Eds.), The nature of emo-tion: Fundamental questions (pp. 131–136). NewYork: Oxford University Press.

Cohen, S., Doyle, W. J., Turner, R. B., Alper, C. M., &Skoner, D. P. (2003). Emotional style and susceptibil-ity to the common cold. Psychosomatic Medicine,65, 652–657.

Cohen, S., Frank, E., Doyle, W. J., Skoner, D. P., Rabin,B. S., & Gwaltney, J. M. (1998). Types of stressorsthat increase susceptibility to the common cold inhealth adults. Health Psychology, 17, 214–223.

Connolly, T., & Reb, J. (2005). Regret in cancer-relateddecisions. Health Psychology, 24, S29–S34.

Consedine, N. S., Krivoshekova, Y. S., & Harris, C. R.(2007). Bodily embarrassment and judgment concernas separable factors in the measurement of medicalembarrassment: Psychometric development and linksto treatment-seeking outcomes. British Journal ofHealth Psychology, 12, 439–462.

Consedine, N. S., & Magai, C. (2006). Emotion devel-


opment in adulthood: A developmental functionalistreview and critique. In C. Hoare (Ed.), The Oxfordhandbook of adult development and learning(pp. 209–244). New York: Oxford University Press.

Consedine, N. S., Magai, C., & Bonanno, G. A. (2002).Moderators of the emotion inhibition–health rela-tionship: A review and research agenda. Review ofGeneral Psychology, 6, 204–228.

Consedine, N. S., Magai, C., & Chin, S. (2004). Hostil-ity and anxiety fear differentially predict coronaryheart disease in men and women. Sex Roles, 50, 63–75.

Consedine, N. S., Magai, C., Cohen, C. I., & Gillespie,M. (2002). Ethnic variation in the impact of negativeaffect and emotion inhibition on the health of olderadults. Journal of Gerontology: Series B: Psychologi-cal Sciences, 57B, 396–408.

Consedine, N. S., Magai, C., & Horton, D. (2005). Eth-nic variation in the impact of emotion and emotionregulation on health: A replication and extension.Journals of Gerontology: Psychological Sciences,60B, P165–P173.

Consedine, N. S., Magai, C., Krivoshekova, Y. S.,Ryzewicz, L., & Neugut, A. I. (2004). Fear, anxiety,worry, and breast cancer screening behavior: A criti-cal review. Cancer Epidemiology, Biomarkers, andPrevention, 13, 501–510.

Consedine, N. S., Magai, C., Kudadjie-Gyamfi, E. K.,Kaluk Longfellow, J., Ungar, T. M., & King, A. R.(2006). Stress versus discrete negative emotion in theprediction of physical complaints: Does predictiveutility vary across groups? Cultural Diversity andEthnic Minority Psychology, 12, 541–557.

Consedine, N. S., Magai, C., & Neugut, A. I. (2004).The contribution of emotional characteristics tobreast cancer screening among women from six eth-nic groups. Preventive Medicine, 38, 64–77.

Consedine, N. S., Morgenstern, A. H., Kudadjie-Gyamfi, E., Magai, C., & Neugut, A. I. (2006). Pros-tate cancer screening behavior in men from seven eth-nic groups: The fear factor. Cancer Epidemiology,Biomarkers, and Prevention, 15, 228–237.

Dale, W., Bilir, P., Han, M., & Meltzer, D. (2005). Therole of anxiety in prostate carcinoma. Cancer, 104,467–478.

Davey, G. C. L., & Bond, N. (2006). Using controlledcomparisons in disgust psychopathology research:The case of disgust, hypochondriasis and health anxi-ety. Journal of Behavior Therapy and ExperimentalPsychiatry, 37, 4–15.

Dearing, R. L., Stuewig, J., & Tangney, J. P. (2005). Onthe importance of distinguishing shame from guilt:Relations to problematic alcohol and drug use. Ad-dictive Behaviors, 30, 1392–1404.

Desai, M. M., Bruce, M. L., & Kasl, S. V. (1999). Theeffects of major depression and phobia on stage at di-agnosis of breast cancer. International Journal ofPsychiatry in Medicine, 29, 29–45.

Detmar, S. B., Muller, M. J., Schnornagel, J. H., Wever,L. D. V., & Aaronson, N. K. (2002). Role of health-

related quality of life in palliative chemotherapytreatment decisions. Journal of Clinical Oncology,20, 1056–1062.

Dickerson, S. S., Gruenewald, T. L., & Kemeny, M. E.(2004). When the social self is threatened: Shame,physiology, and health. Journal of Personality, 72,1191–1216.

Diefenbach, M. A., Miller, S. M., & Daly, M. B. (1999).Specific worry about breast cancer predicts mam-mography use in women at risk for breast and ovar-ian cancer. Health Psychology, 18, 532–536.

DiMatteo, M. R., Lepper, H. S., & Croghan, T. W.(2000). Depression is a risk factor for noncompliancewith medical treatment: Meta-analysis of the effectsof anxiety and depression on patient adherence. Ar-chives of Internal Medicine, 160, 2101–2107.

Donker, F. J. S. (2000). Cardiac rehabilitation: A reviewof current developments. Clinical Psychology Re-view, 20, 923–943.

Dube, L., LeBel, J. L., & Lu, J. (2005). Affect asymme-try and comfort food consumption. Physiology andBehavior, 86, 559–567.

Erez, A., & Isen, A. M. (2002). The influence of positiveaffect on the components of expectancy motivation.Journal of Applied Psychology, 87, 1055–1067.

Eyler, A. A., & Vest, J. R. (2002). Environmental andpolicy factors related to physical activity in ruralwhite women. Women and Health, 36, 111–121.

Farraye, F. A., Wong, M., Hurwitz, S., Puleo, E.,Emmons, K., Wallace, M. B., et al. (2004). Barriers toendscopic colorectal cancer screening: Are womendifferent from men? American Journal ofGastroenterology, 99, 341–349.

Fink, A. K., Gurwitz, J., Rakowski, W., Guadagnoli, E.,& Silliman, R. A. (2004). Patient beliefs and tam-oxifen discontinuance in older women with estrogenreceptor-positive breast cancer. Journal of ClinicalOncology, 22, 3309–3315.

Folkman, S., & Moskowitz, J. T. (2004). Coping: Pit-falls and promise. Annual Review of Psychology, 55,745–774.

Fredrickson, B. L., & Levenson, R. W. (1998). Positiveemotions speed recovery from the cardiovascularsequelae of negative emotions. Cognition and Emo-tion, 12, 191–220.

Fredrickson, B. L., & Losada, M. F. (2005). Positive af-fect and the complex dynamics of human flourishing.American Psychologist, 60, 678–686.

Fredrickson, B. L., Maynard, K. E., Helms, M. J.,Haney, T. L., Siegler, I. C., & Barefoot, J. C. (2000).Hostility predicts magnitude and duration of bloodpressure responses to anger. Journal of BehavioralMedicine, 23, 229–243.

Friedman, H. S., & Booth-Kewley, S. (1987). The “dis-ease-prone personality”: A meta-analytic view of theconstruct. American Psychologist, 42, 539–555.

Frijda, N. H. (1994). Emotions are functional, most ofthe time. In P. Ekman & R. J. Davidson (Eds.), Thenature of emotion: Fundamental questions (pp. 112–122). New York: Oxford University Press.


Gascoigne, P., Mason, M. D., & Roberts, E. (1999).Factors affecting presentation and delay in patientswith testicular cancer: Results of a qualitative study.Psycho-Oncology, 8, 144–154.

Gonzalez, J. S., Penedo, F. J., Antoni, M. H., Durdan, R.E., Fernandez, M. I., McPherson-Baker, S., et al.(2004). Social support, positive states of mind, andHIV treatment adherence in men and women livingwith HIV/AIDS. Health Psychology, 23, 413–418.

Gottschalk, L. A. (1985). Hope and other deterrents toillness. American Journal of Psychotherapy, 39, 515–524.

Griffin, K. W., Friend, R., Eitel, P., & Lobel, M. (1993).Effects of environmental demands, stress, and moodon health practices. Journal of Behavioral Medicine,16, 643–661.

Hagglund, D., Walker-Engstrom, M. L., Larsson, G., &Leppert, J. (2003). Reasons why women with long-term urinary incontinence do not seek professionalhelp: A cross-sectional population-based cohortstudy. International Urogynecology Journal and Pel-vic Floor Dysfunction, 14, 296–304.

Hamilton, M. (2006). Disgust reactions to meat amongethically and health motivated vegetarians. Ecologyof Food and Nutrition, 45, 125–158.

Hamilton, N. A., & Malcarne, V. L. (2004). Cognition,emotion, and chronic illness. Cognitive Therapy andResearch, 28, 555–557.

Harburg, E., Julius, M., Kaciroti, N., Gleiberman, L., &Schork, M. A. (2003). Expressive/suppressive anger-coping responses, gender, and types of mortality: A17-year follow-up (Tecumseh, Michigan, 1971–1988). Psychosomatic Medicine, 65, 588–597.

Hay, J. L., Buckley, T. R., & Ostroff, J. S. (2005). Therole of cancer worry in cancer screening: A theoreti-cal and empirical review of the literature. Psycho-Oncology, 14, 517–534.

Husaini, B. A., Sherkat, D. E., Bragg, R., Levine, R.,Emerson, J. S., Mentes, C. M., et al. (2001). Predic-tors of breast cancer screening in a panel study of Af-rican American women. Women and Health, 34, 35–51.

Isen, A. M. (1999). Positive affect. In T. Dalgleish & M.Power (Eds.), Handbook of cognition and emotion(pp. 521–539). New York: Wiley.


Jessop, D. C., Rutter, D. R., Sharma, D., & Albery, I. P.(2004). Emotion and adherence to treatment in peo-ple with asthma: An application of the emotionalStroop paradigm. British Journal of Psychology, 95,127–147.

Kalichman, S. C., Johnson, J. R., Adair, V., Rompa, D.,Multhauf, K., & Kelly, J. A. (1994). Sexual SensationSeeking: Scale development and predicting AIDS-riskbehavior among homosexually active men. Journalof Personality Assessment, 62, 385–397.

Kalsekar, I. D., Madhavan, S. S., Amonkar, M. M.,Makela, E. H., Scott, V. G., Douglas, S. M., et al.(2006). Depression in patients with Type 2 diabetes:

Impact on adherence to oral hypoglycemic agents.Annals of Pharmacotherapy, 40, 605–611.

Kaplan, R. C., Bhalodkar, N. C., Brown, E. J., Wflite, J.,& Brown, D. L. (2004). Race, ethnicity, and socio-cultural characteristics predict noncompliance withlipid-lowering medications. Preventive Medicine, 39,1249–1255.

Karlsson, H., Lehtinen, V., & Joukamaa, M. (1995).Are frequent attenders of primary health-care dis-tressed? Scandinavian Journal of Primary HealthCare, 13, 32–38.

Kassel, J. D., Stroud, L. R., & Paronis, C. A. (2003).Smoking, stress, and negative affect: Correlation,causation, and context across stages of smoking. Psy-chological Bulletin, 129, 270–304.

Kawachi, I., Sparrow, D., Spiro, A., Vokonas, P., &Weiss, S. T. (1996). A prospective study of anger andcoronary heart disease: The Normative Aging Study.Circulation, 94, 2090–2095.

Kelsey, K. S., DeVellis, B. M., Begum, M., Belton, L., &Hooten, E. G. (2006). Positive affect, exercise andself-reported health in blue-collar women. AmericanJournal of Health Behavior, 30, 199–207.

Keltner, D., & Anderson, C. (2000). Saving face forDarwin: The functions and uses of embarrassment.Current Directions in Psychological Science, 9, 187–192.

Keltner, D., & Buswell, B. N. (1996). Evidence for thedistinctiveness of embarrassment, shame, and guilt:A study of recalled antecedents and facial expressionsof emotion. Cognition and Emotion, 10, 155–171.

Keltner, D., Ellsworth, P. C. E., & Edwards, K. (1993).Beyond simple pessimism: Effects of sadness and an-ger on social perception. Journal of Personality andSocial Psychology, 64, 740–752.

Keltner, D., & Gross, J. J. (1999). Functional accountsof emotions. Cognition and Emotion, 13, 467–480.

Kiecolt-Glaser, J. K., McGuire, L., Robles, T. F., &Glaser, R. (2002). Emotions, morbidity, and mortal-ity: New perspectives from psychoneuroimmunology.Annual Review of Psychology, 53, 83–107.

Krantz, D. S., & McCeney, M. K. (2002). Effects of psy-chological and social factors on organic disease: Acritical assessment of research on coronary heart dis-ease. Annual Review of Psychology, 53, 341–369.

Kubzansky, L. D., Cole, S. R., Kawachi, I., Vokonas, P.,& Sparrow, D. (2006). Shared and unique contribu-tions of anger, anxiety, and depression to coronaryheart disease: A prospective study in the NormativeAging Study. Annals of Behavioral Medicine, 31, 21–29.

Kubzansky, L. D., Wright, R. J., Cohen, S., Weiss, S.,Rosner, B., & Sparrow, D. (2002). Breathing easy: Aprospective study of optimism and pulmonary func-tion in the normative aging study. Annals of Behav-ioral Medicine, 24, 345–353.

Lam, W. W. T., Fielding, R., Chan, M., Chow, L., & Or,A. (2005). Gambling with your life: The process ofbreast cancer treatment decision making in Chinesewomen. Psycho-Oncology, 14, 1–15.


Larsen, K., Schroll, M., & Avlund, K. (2006). Depres-sive symptomatology at age 75 and subsequent use ofhealth and social services. Archives of Gerontologyand Geriatrics, 42, 125–139.

Levenson, R. W. (1994). Human emotion: A functionalview. In P. Ekman & R. J. Davidson (Eds.), The na-ture of emotion: Fundamental questions (pp. 123–126). New York: Oxford University Press.

Leventhal, E. A., Hansell, S., Diefenbach, M.,Leventhal, H., & Glass, D. C. (1996). Negative affectand self-report of physical symptoms: Two longitudi-nal studies of older adults. Health Psychology, 15,193–199.

Leventhal, H., Diefenbach, M., & Leventhal, E. A.(1992). Illness cognition: Using common sense to un-derstand treatment adherence and affect cognitioninteractions. Cognitive Therapy and Research, 16,143–163.

Leventhal, H., & Patrick-Miller, L. (2000). Emotionsand physical illness: Causes and indicators of vulner-ability. In M. Lewis & J. M. Haviland-Jones (Eds.),Handbook of emotions (2nd ed., pp. 523–537). NewYork: Guilford Press.

Lindsey, L. L. M. (2005). Anticipated guilt as behavioralmotivation: An examination of appeals to help un-known others through bone marrow donation. Hu-man Communication Research, 31, 453–481.

Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch,N. (2001). Risk as feelings. Psychological Bulletin,127, 267–286.

Mayne, T. J. (1999). Negative affect and health: The im-portance of being earnest. Cognition and Emotion,13, 601–635.

Mayne, T. J. (2001). Emotions and health. In T. J.Mayne & G. A. Bonanno (Eds.), Emotions: Currentissues and future directions (pp. 361–397). NewYork: Guilford Press.

McDade-Montez, E. A., Christensen, A. J., Cvengros, J.A., & Lawton, W. J. (2006). The role of depressionsymptoms in dialysis withdrawal. Health Psychology,25, 198–204.

McKeen, N. A., Chipperfield, J. G., & Campbell, D. W.(2004). A longitudinal analysis of discrete negativeemotions and health-services use in elderly individu-als. Journal of Aging and Health, 16, 204–227.

Mellers, B. A., & McGraw, A. P. (2001). Anticipatedemotions as guides to choice. Current Directions inPsychological Science, 10, 210–214.

Mesquita, B., & Leu, J. (2007). The cultural psychologyof emotion. In S. Kitayama & D. Cohen (Eds.),Handbook of cultural psychology (pp. 734–759).New York: Guilford Press.

Middleton, R. A., & Byrd, E. K. (1996). Psychosocialfactors and hospital readmission status of older per-sons with cardiovascular disease. Journal of AppliedRehabilitation Counseling, 27, 3–10.

Milam, J. E., Richardson, J. L., Marks, G., Kemper, C.A., & McCutchan, A. J. (2004). The roles ofdispositional optimism and pessimism in HIV diseaseprogression. Psychology and Health, 19, 167–181.

Miller, S. M., Bowen, D. J., Campbell, M. K.,Diefenbach, M. A., Gritz, E. R., Jacobsen, P. B., et al.(2004). Current research promises and challenges inbehavioral oncology: Report from the American So-ciety of Preventive Oncology Annual Meeting, 2002.Cancer Epidemiology, Biomarkers, and Prevention,13, 171–180.

Miller, S. M., & Diefenbach, M. A. (1998). Thecognitive–social health information processing (C-SHIP) model: A theoretical framework for researchin behavioral oncology. In D. Krantz & A. Baum(Eds.), Technology and methods in behavioral medi-cine (pp. 219–243). Mahwah, NJ: Erlbaum.

Moore, R., Brodsgaard, I., & Rosenberg, N. (2004).The contributions of embarrassment to phobic dentalanxiety: A qualitative research study. BMC Psychia-try, 4(10). Available online at www.biomedcentral.com/content/pdf/1471-244X-4-10.pdf

Mora, P. A., Robitaille, C., Leventhal, H., Swigar, M., &Leventhal, E. A. (2002). Trait negative affect relatesto prior-week symptoms, but not to reports of illnessepisodes, illness symptoms, and care seeking amongolder persons. Psychosomatic Medicine, 64, 436–449.

Moskowitz, J. T. (2003). Positive affect predicts lowerrisk of AIDS mortality. Psychosomatic Medicine, 65,620–626.

Ogden, J. (2003). Some problems with social cognitionmodels: A pragmatic and conceptual analysis. HealthPsychology, 22, 424–428.

Ostir, G. V., Markides, K. S., Black, S. A., & Goodwin,J. S. (2000). Emotional well-being predicts subse-quent functional independence and survival. Journalof the American Geriatrics Society, 48, 473–478.

Petersen, S., Schwartz, R. C., Sherman-Slate, E., Frost,H., Straub, J. L., & Damjanov, N. (2003). Relation-ship of depression and anxiety to cancer patients’medical decision making. Psychological Reports, 93,323–334.

Pettit, J. W., Kline, J. P., Gencoz, T., Gencoz, F., &Joiner, T. E. (2001). Are happy people healthier?: Thespecific role of positive affect in predicting self-reported health symptoms. Journal of Research inPersonality, 35, 521–536.

Pirraglia, P. A., Sanyal, P., Singer, D. E., & Ferris, T. G.(2004). Depressive symptom burden as a barrier toscreening for breast and cervical cancers. Journal ofWomen’s Health, 13, 731–738.

Pressman, S. D., & Cohen, S. (2005). Does positive af-fect influence health? Psychological Bulletin, 131,925–971.

Rachman, S. (2004). Fear of contamination. BehaviourResearch and Therapy, 42, 1227–1255.

Reiche, E. M. V., Morimoto, H. K., & Nunes, S. O. V.(2006). Stress and depression-induced immune dys-function: Implications for the development and pro-gression of cancer. International Review of Psychia-try, 17, 515–527.

Richman, L. S., Kubzansky, L., Maselko, J., Kawachi, I.,Choo, P., & Bauer, M. (2005). Positive emotion and


health: Going beyond the negative. Health Psychol-ogy, 24, 422–429.

Rozin, P., & Fallon, A. E. (1987). A perspective on dis-gust. Psychological Review, 94, 23–41.

Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1997). TheCAD triad hypothesis: A mapping between threemoral emotions (contempt, anger, disgust) and threemoral codes (community, autonomy, divinity). Jour-nal of Personality and Social Psychology, 76, 574–586.

Rugulies, R. (2002). Depression as a predictor for coro-nary heart disease: A review and meta-analysis.American Journal of Preventive Medicine, 23, 51–61.

Silvia, P. (2001). Interest and interests: The psychologyof constructive capriciousness. Review of GeneralPsychology, 5, 279–290.

Sirois, B. C., & Burg, M. M. (2003). Negative emotionand coronary heart disease. Behavioral Modification,27, 83–102.

Slovic, P., Peters, E., Finucane, M. L., & MacGregor, D.G. (2005). Affect, risk, and decision making. HealthPsychology, 24, S35–S40.

Smith, T. W. (1992). Hostility and health: Current statusof a psychosomatic hypothesis. Health Psychology,11, 139–150.

Smith, T. W., Glazer, K., Ruiz, J. M., & Gallo, L. C.(2004). Hostility, anger, aggressiveness, and coronaryheart disease: An interpersonal perspective on per-sonality, emotion, and health. Journal of Personality,72, 1217–1270.

Stefanek, M., & McDonald, P. G. (2003). Editorial: Bio-logical mechanisms of psychosocial effects on dis-ease: Implications for cancer control. Introduction tothe special issue. Brain, Behavior, and Immunity, 17,S1–S4.

Sukhdial, A., & Boush, D. M. (2004). Eating guilt:Measurement and relevance to consumer behavior.Advances in Consumer Research, 31, 575–576.

Swan, G. E., & Carmelli, D. (1996). Curiosity and mor-tality in aging adults: A 5-year follow-up of the West-ern Collaborative Group Study. Psychology andAging, 11, 449–453.

Taylor, V. M., Jackson, J. C., Tu, S.-P., Yasui, Y.,Schwartz, S. M., Kuniyuki, A., et al. (2002). Cervicalcancer screening among Chinese Americans. CancerDetection and Prevention, 26, 139–145.


Tugade, M. M., Fredrickson, B. L., & Barrett, L. F.(2004). Psychological resilience and positive emo-tional granularity: Examining the benefits of positiveemotions on coping and health. Journal of Personal-ity, 72, 1161–1190.

Tweed, D. L., Goering, P., Lin, E., & Williams, J. I.(1998). Psychiatric morbidity and physician visits:Lessons from Ontario. Medical Care, 36, 573–585.

Van Diest, I., De Peuter, S., Eertmans, A., Bogaerts, K.,Victoir, A., & Van den Bergh, O. (2005). Negativeaffectivity and enhanced symptom reports: Differen-tiating between symptoms in men and women. SocialScience and Medicine, 61, 1835–1845.

Van Servellen, G., Chang, B., Garcia, L., & Lombardi,E. (2002). Individual and system level factors associ-ated with treatment nonadherence in human immu-nodeficiency virus-infected men and women. AIDSPatient Care and Studies, 16, 269–281.

Wang, P. S., Bohn, R. L., Knight, E., Glynn, R. J.,Mogun, H., & Avorn, J. (2002). Noncompliancewith antihypertensive medications: The impact of de-pressive symptoms and psychosocial factors. Journalof General Internal Medicine, 17, 504–511.

Watson, D., & Pennebaker, J. W. (1989). Health com-plaints, stress, and distress: Exploring the central roleof negative affectivity. Psychological Review, 96,234–254.

Wills, T. A., Sandy, J. M., Shinar, O., & Yaeger, A.(1999). Contributions of positive and negative affectto adolescent substance use: Test of a bidirectionalmodel in a longitudinal study. Psychology of Addic-tive Behaviors, 13, 327–338.

Witkiewitz, K., & Marlatt, G. A. (2004). Relapse pre-vention for alcohol and drug problems: That wasZen, this is Tao. American Psychologist, 59, 224–235.

Witte, K. (1998). Fear as motivator, fear as inhibitor:Using the extended parallel process model to explainfear appeal successes and failures. In P. A. Andersen& L. K. Guerrero (Eds.), Handbook of communica-tion and emotion: Research, theory, application, andcontexts (pp. 423–450). San Diego, CA: AcademicPress.

Zautra, A. J. (2003). Emotions, stress, and health. NewYork: Oxford University Press.

Zautra, A. J., Davis, M. C., & Smith, B. W. (2004).Emotions, personality, and health: Introduction tothe special issue. Journal of Personality, 72, 1097–1104.


C H A P T E R 4 3

Emotion Disturbancesas Transdiagnostic Processes

in Psychopathology

ANN M. KRING

Emotion disturbances are ubiquitous in psy-chopathology. Even a cursory glance at thecurrent version of the American diagnosticsystem—the Diagnostic and Statistical Manualof Mental Disorders, fourth edition, text revi-sion (DSM-IV-TR; American Psychiatric Asso-ciation, 2000)—reveals that nearly all the diag-nostic categories include symptoms that tapone type of emotion disturbance or another(see Table 43.1). These disturbances span bothpositive and negative emotions, and they in-clude excesses of emotion (as in the case of spe-cific and social phobias, with marked and per-sistent fear); deficits in emotion (as in the caseof narcissistic personality disorder, with a lackof empathy); social emotional problems (as inautistic disorder, with a lack of emotional reci-procity); and regulation problems (as in bor-derline personality disorder, with difficulties incontrolling anger). The pervasiveness of emo-tion disturbances in psychopathology suggeststhe potential for commonalities across disor-ders. Indeed, there may be emotional distur-

bances that are central to a number of differentdisorders; yet the manifestation of these distur-bances may differ from disorder to disorder,thus helping to account for the different symp-tom constellations across disorders. In thischapter, I consider the utility of adopting atransdiagnostic approach to understandingemotion disturbances in psychopathologyacross several levels, including descriptive phe-nomenology, etiology, and treatment.

CONSTRAINING EMOTIONAND AFFECT

Emotions have developed through the courseof human evolutionary history to prepare or-ganisms to act in response to a number of envi-ronmental stimuli and challenges. This accountsuggests that emotions, under most circum-stances, serve a number of important intra- andinterpersonal functions (e.g., Frijda, 1986;Keltner & Kring, 1998; Lang, Bradley, &

691


TABLE 43.1. Emotion-Based Symptoms in DSM-IV-TR Disorders

Disorders Symptoms

Schizophrenia and other psychotic disorders

SchizophreniaSchizoaffective disorderSchizophreniform disorder Affective flattening, anhedonia

Mood disorders

Major depressive episode Depressed mood, anhedoniaDysthymia Depressed moodManic or hypomanic episode Elevated, expansive, or irritable mood

Anxiety disorders

Panic disorder Intense fear or discomfortAgoraphobia AnxietySpecific and social phobias Marked and persistent fear, anxious anticipationObsessive–compulsive disorder Marked anxiety or distressPosttraumatic stress disorder Irritability, anger, distress, anhedonia, restricted range of affectAcute stress disorder Anxiety or increased arousalGeneralized anxiety disorder Excessive anxiety and worry, irritability

Somatoform disorders

Hypochondriasis Preoccupation with fears of having disease

Eating disorders

Anorexia nervosa Fear of gaining weight

Sleep disorders

Sleep terror disorder Intense fear and signs of autonomic arousalCircadian rhythm sleep disorder,

nightmare disorderClinically significant distress

Impulse control disorders

Pathological gambling Irritability, dysphoric moodTrichotillomania Tension; pleasure or relief after hair pullingIntermittent explosive disorder Rage, angerPyromania, kleptomania Tension or excited mood

Adjustment disorders Marked distress

Personality disorders

Paranoid personality disorder Quickness to react angrilySchizoid personality disorder Emotional coldness, detachment, flattened affectSchizotypal personality disorder Inappropriate or constricted affect, excessive social anxietyAntisocial personality disorder Lack of remorse, irritabilityBorderline personality disorder Affective instability due to marked reactivity of mood,

inappropriate intense anger, or difficulty controlling angerHistrionic personality disorder Rapidly shifting and shallow expressions of emotionNarcissistic personality disorder Lack of empathyAvoidant personality disorder Fear of criticism, disapproval, or rejectionDependent personality disorder Fear of being unable to care for self, being left alone

Substance-related disorders

Alcohol intoxication Mood labilityAlcohol withdrawal AnxietyAmphetamine and cocaine intoxication Euphoria or affective blunting; anxiety, tension, angerAmphetamine and cocaine withdrawal Dysphoric mood

(continued)

Cuthbert, 1990; Levenson, 1994). As I have ar-gued elsewhere, the functions of emotion inpersons with psychopathology are comparableto those for persons without psychopathology(Keltner & Kring, 1998; Kring &Bachorowski, 1999). However, emotion distur-bances in psychopathology interfere with theachievement of emotion-related functions. Forexample, the absence of facial expressions in apatient with schizophrenia may evoke negativeresponses from others (Krause, Steimer-Krause,& Hufnagel, 1992), thus negatively affectinghis or her social relationships and interactions(e.g., Hooley, Richters, Weintraub, & Neale,1987).

Emotions are typically considered to havemultiple components, including expression, ex-perience, and physiology. The extent to whichthese emotion components correspond withone another or cohere remains a topic of de-bate (e.g., Barrett, 2006a; Bradley & Lang,2000; Mauss, Levenson, McCarter, Wilhelm,& Gross, 2005). Functionalist accounts ofemotion suggest that coherence among compo-nents is adaptive (e.g., Levenson, 1994), butthe empirical data supporting coherence aremixed (Barrett, 2006a). There are a number ofreasons why particular emotion componentsmay not cohere in any given study—includingsample characteristics, emotion elicitationmethods, emotion component measurements,

and data-analytic techniques, as well aswhether even under ideal circumstances, emo-tion coherence is the exception rather thanthe norm. Nevertheless, the lack of coherenceacross multiple emotions, situations, and con-texts has been observed in different psychologi-cal disorders (e.g., schizophrenia, psychopathy)and has been considered an emotional distur-bance (e.g., Kring, 2001).

Although the terms “affect” and “emotion”are used interchangeably in the psychopatholo-gy literature, there are important conceptualand empirical distinctions between the terms.Generally speaking, “affect” is most often usedin reference to feeling states, whereas “emo-tion” is used in reference to multiple compo-nents (only one of which is a feeling state).Barrett and colleagues have distinguished “coreaffect” from the more generic term “affect.”Core affect reflects neurophysiological statesthat are an omnipresent indicator of a person(srelationship to his or her environment at anygiven time (Barrett, 2006a; Barrett, Mesquita,Ochsner, & Gross, 2007; Russell, 2003; Barrett& Russell, 1999). Core affect is experienced asfeelings of pleasure or displeasure, and to alesser extent arousal or activation (Barrett,2006a; Barrett et al., 2007). Although core af-fect is observed across cultures, there are never-theless important individual and cultural dif-ferences (Barrett, 2006b; Mesquita &

43. Emotion Disturbances as Transdiagnostic Processes 693


Disorders Symptoms

Substance-related disorders (cont.)

Caffeine intoxication Nervousness, excitementCannabis intoxication Euphoria, anxiety

Dementias

Dementia due to Pick’s disease Emotional bluntingDementia due to Huntington’s disease Depression, irritability, anxiety

Childhood disorders

Autistic disorder, Asperger’s disordernonverbal behaviors, such asfacial expression

Lack of emotional reciprocity; marked impairment in

Separation anxiety disorder Distress, worry, fearfulnessOppositional defiant disorder Quickness to lose temper, get angry, be annoyed by othersChildhood disintegrative disorder Impairment in nonverbal behaviors; lack of social or emotional

reciprocity

Other conditions

Bereavement Guilt

Karasawa, 2002) that may be important forunderstanding disturbances in core affect inpsychopathology.

THE TRANSDIAGNOSTIC APPROACH

In a recent and influential book, Harvey,Watkins, Mansell, and Shafran (2004) havereviewed the literature on cognitive and be-havioral maintaining processes in psychopath-ology. Rather than organizing their book bydisorder, they have instead adopted a trans-diagnostic perspective, reviewing the evidencefor common processes across different disor-ders. Their analysis points to an accumulatingbody of evidence of disruptions in a number ofprocesses, such as attention, memory, reason-ing, and avoidance, that are common acrossmore than one adult disorder and that serve tomaintain or exacerbate the symptoms of thesedisorders. Similar conclusions have beenreached regarding common cognitive maintain-ing processes across the eating disorders(Fairburn, Cooper, & Shafran, 2003), andtreatment approaches that target transdiag-nostic processes across disorders have recentlybeen developed for depression, anxiety, andeating disorders (e.g., Barlow, Allen, & Choate,2004; Fairburn et al., 2003; Norton, Hayes, &Hope, 2004).

Harvey et al. (2004) have argued that thereare a number of key advantages to adopting atransdiagnostic perspective in psychopatholo-gy. First, the transdiagnostic perspective mayhelp to account for the high rates of comorbidi-ty among the current DSM disorders. Indeed,comorbidity is the norm rather than the excep-tion. In the National Comorbidity Survey(NCS), Kessler et al. (1994) found that nearly80% of individuals with a lifetime diagnosis ofone disorder had received at least one otherlifetime diagnosis. In the NCS replicationstudy, 45% of people who met criteria for onediagnosis in the prior 12 months met criteriafor at least one more diagnosis (Kessler, Chiu,Demler, & Walters, 2005). Other studies havefound similarly high rates of comorbidity. Forexample, nearly two-thirds of individuals withdepression meet diagnostic criteria for an anxi-ety disorder (e.g., Mineka, Watson, & Clark,1998), and as many as 50% of individuals whomeet criteria for an anxiety disorder are de-pressed (e.g., Brown et al., 2001). The highlevel of comorbidity may well reflect problems

in the current diagnostic system (e.g., poor dis-criminant validity). However, the rampantcomorbidity also suggests that there may becommon symptoms or processes across disor-ders, including emotional processes.

Second, a transdiagnostic approach may beuseful for developing and evaluating treat-ments. For example, recent theory and empiri-cal data point to the notion that currentlyavailable treatments for anxiety and depressionare changing common aspects of these disor-ders, rather than disorder-specific aspects (e.g.,Barlow et al., 2004; Hayes, Strosahl, & Wil-son, 1999; Persons, Roberts, & Zalecki, 2003).In addition, evidence from treatment outcomestudies suggests that interventions for one dis-order (e.g., depression) are also effective intreating other disorders (e.g., generalized anxi-ety disorder) (e.g., Brown & Barlow, 1992;Tsao, Mystkowski, Zucker, & Craske, 2002).A transdiagnostic perspective may illuminatethe common mechanisms or processes acrossdisorders, which may then be more directly tar-geted in treatment.

A transdiagnostic approach to emotion dis-turbances in psychopathology has been sug-gested, sometimes implicitly, by other theorists,researchers, and clinicians (e.g., Barlow et al.,2004; Patrick & Bernat, 2006; Thayer &Brosschot, 2005). However, the reach of suchan approach has been fairly limited thus far, forat least two reasons. First, some disorders lendthemselves more clearly to such an approachthan others, given their comorbidities, compa-rable treatments, and treatment responses. Forexample, mood and anxiety disorders—twobroad categories that are collectively referredto as the “emotional disorders” (e.g., Barlow,2004; Watson, 2005)—have been discussed intransdiagnostic terms across many levels.Other disorders that may well benefit from atransdiagnostic approach (e.g., personality dis-orders) have not yet stimulated muchtransdiagnostic theory and research. Many ofthe examples of the transdiagnostic approachto emotion disturbances in psychopathology il-lustrated throughout the chapter will involvethe mood and anxiety disorders.

A second reason why the transdiagnostic ap-proach to emotion disturbances has been fairlylimited is probably the fact that the prevailingparadigm in psychopathology research over thepast several decades has been “disorder-centric.” That is, most investigations are de-signed to answer questions about the symp-


toms, causes, and treatments of individualdisorders. In the realm of emotion and psycho-pathology, studies are typically designed tostudy a particular emotion disturbance in aparticular disorder. For example, the literatureis packed with studies of particular emotiondisturbances (e.g., facial emotion perceptiondeficits) in putatively distinct disorders (e.g.,depression, borderline personality disorder,schizophrenia), with little consideration of thepossibility that the emotion disturbance maycut across these disorders. Such disorder-centric research is perpetuated by the field’s rel-atively greater emphasis on internal validity(e.g., tightly controlled study of one disorder)over external validity (e.g., naturalistic study ofcomorbid disorders), as reflected in editorialpractices at top-flight journals and funding pri-orities at granting agencies. As such, the entrypoint for a majority of psychopathology stud-ies is typically a single disorder rather thanmultiple disorders, with relatively less empha-sis on mechanisms or processes (such as emo-tion disturbances) that may cut across tradi-tional diagnostic boundaries. Unfortunately,this disorder-centric focus overlooks the factthat most disorders do not occur in “pure”form, and thus conclusions regarding specific-ity of emotion disturbances in X disorder maynot be particularly informative with respect tounderstanding X disorder as it more commonlyoccurs in combination with Y disorder.

Despite these challenges, the promise of atransdiagnostic approach to emotion distur-bances in psychopathology is evident at manylevels. First, a transdiagnostic approach can in-form the ways in which different disorders areclassified. In other words, examining emotion-related commonalities at the symptom levelmay help to account for the high levels ofcomorbidity in the current diagnostic system,and in turn may provide guidance on refiningdiagnostic categories. Second, a transdiag-nostic approach can be informative with re-spect to identifying common emotion-relatedcausal and/or maintaining processes across dis-orders. Third, a transdiagnostic approach canbe useful for treatment development, with anemphasis on changing emotion processes. Inthe remaining sections of this chapter, I con-sider the relative merits of adopting atransdiagnostic approach to emotion distur-bances in psychopathology across these threelevels: descriptive phenomenology, etiology,and treatment.

DESCRIPTIVE PSYCHOPATHOLOGYAND DIAGNOSIS

In an effort to spur the field toward greaterattention to emotion and psychopathology,Berenbaum, Raghavan, Le, Vernon, andGomez (2003) have proposed a separate taxon-omy of emotional disturbances as a companionto the current diagnostic system. Their taxon-omy parses specific emotion (e.g., shame, guilt,happiness, fear) disturbances into three broadareas of disruption: “valence,” “intensity/regu-lation,” and “disconnections.” Each of thebroad categories is further subdivided toachieve greater specificity of particular emo-tion disturbances. Emotional valence distur-bances can involve pleasant or unpleasant emo-tions, as well as too much or too little of theseemotions. For example, the limited experienceof pleasure (i.e., anhedonia) that characterizesdepression and schizophrenia, as well as the ex-cess of fear in panic disorder, would constitutevalence disturbances. Emotional intensity/regu-lation disturbances are defined as over- orunderregulation of both pleasant and unpleas-ant emotions. For example, mania, which ischaracterized by excesses in both pleasant (joy,euphoria) and unpleasant (irritability) emo-tions, would be construed as an emotional in-tensity/regulation disturbance. Disconnectiondisturbances reflect disconnections between theexpressive component of emotion and othercomponents, as in schizophrenia, where pa-tients experience strong feelings yet do not ex-press them outwardly (e.g., Berenbaum &Oltmanns, 1992; Kring & Neale, 1996; Kring& Earnst, 1999). Disconnection disturbancesalso reflect a lack of conscious awareness ofone’s own emotional responses. Berenbaum etal. (2003) conclude their paper with nine rec-ommendations for future research, but these donot explicitly include the possibility of examin-ing these disturbances across disorders. In-stead, the recommendations are geared towardamplifying our understanding of specific disor-ders; examining emotion disturbances indepen-dently from diagnosed disorders; modifyingtreatments for specific disorders by targetingemotion disturbances; and examining similari-ties and differences in the disturbances acrossgender, culture, and the lifespan. Althoughthese are certainly important goals for futureresearch, the transdiagnostic implications ofthis taxonomy are also ripe for further investi-gation.


In a commentary on Berenbaum et al.’s arti-cle, Watson (2003) has argued that includingthe taxonomy as a companion to the currentdiagnostic system would add unnecessary com-plexity and serve to further reify already prob-lematic diagnostic categories. Although he doesnot suggest a transdiagnostic approach, Wat-son has nonetheless argued for “scrappingthese heterogeneous diagnostic categories andfocusing instead on homogeneous symptomclusters” (p. 237). In this view, focusing on thecurrent DSM symptoms, many of which reflectemotion disturbances, may be a more fruitfulway of classifying psychopathology.

More recently, Watson (2005) has expandedupon this idea, arguing that the current config-uration of mood disorders and anxiety disor-ders ought to be replaced with a differentconfiguration. Specifically, he proposes a quan-titative hierarchical model to account for thehigh levels of comorbidity between mood andanxiety disorders, drawing upon structuralanalyses of the diagnostic symptoms at boththe phenotypic and genotypic levels. His modelincludes an overarching domain referred to as“emotional disorders.” This would consist ofthree subdomains, including what Watson calls“distress disorders” (generalized anxiety disor-der, posttraumatic stress disorder, major de-pressive disorder, dysthymia), “bipolar disor-ders” (bipolar I and II disorders, cyclothymia),and “fear disorders” (specific and social pho-bias, panic disorder, agoraphobia). Watson’sproposal is more than a simple reshuffling ofdiagnostic categories, and this approach is cer-tainly a more rational and empirically sup-ported approach to situating the mood andanxiety disorders in the diagnostic system thanis the current, purely phenomenological ap-proach. The model also points to the promiseof a transdiagnostic approach to the descriptivephenomenology of mental disorders.

Indeed, part of the impetus for Watson’s(2005) proposed model has come from the in-fluential theory and research on common anddistinct emotion disturbances in the mood andanxiety disorders. Clark and Watson (1991)originally proposed the tripartite model to ac-count for the relationship between anxiety anddepression (see also Watson, Clark, et al.,1995; Watson, Weber, et al., 1995; Watson,Weise, Vaidya, & Tellegen, 1999). In thismodel, a general distress factor characterizedby high levels of negative activation (NA) iscommon to both anxiety and depression; a sec-

ond factor, characterized by low levels of posi-tive activation (PA) or pleasurable engagementwith the environment, is specific to depression;and a third factor, variously referred to as“anxious arousal” (AA) or “somatic arousal,”is specific to anxiety. A revision to the modelwas later proposed, to account better for theheterogeneity among the anxiety disorders(Mineka et al., 1998). The revised model,termed the integrative hierarchical model, fol-lowed from additional data suggesting thathigh levels of AA are more characteristic ofpanic disorder in particular than of the anxietydisorders in general (Brown, Chorpita, &Barlow, 1998; Zinbarg & Barlow, 1996).

Watson (2005) has argued that data gener-ated from the integrative hierarchical modelthus far are not sufficient to account fully forthe comorbidity among the mood and anxietydisorders. Thus he has proposed the new quan-titative model to account better for the exten-sive comorbidities, while at the same time re-taining the current diagnostic categories (i.e., adisorder-centric approach). An alternative toWatson’s model is the dissection of mood andanxiety disorders into categories based onemotion-related difficulties. Thus, instead ofseparate categories for the different mood andanxiety disorders, there might be an overarch-ing category of general distress disorders withsubdomains of low-PA disorders and high-AAdisorders. Brown et al. (1998) proposed such astructural model, showing that many of theanxiety disorders and depression were betteraccounted for by the emotion-based factors ofNA, PA, and AA, though they did not explicitlycall for the scrapping of the specific mood andanxiety disorder categories. More recently,Barlow and colleagues have pushed this ideafurther, noting that “DSM-IV emotional disor-der categories do not qualify in any sense asreal entities . . . but do seem to be useful con-cepts or constructs that emerge as “blips” on ageneral background of NAS [negative affectsyndrome]” (Barlow et al., 2004, p. 212), andsuggesting that future revisions of the DSMmay do well to eliminate many of the currentcategories of disorders (Moses & Barlow,2006). This is a fairly radical proposal, but per-haps one whose time has (nearly) come.

It is true that the field’s disorder-centric ap-proach, which has been central at least sincethe development of DSM-III (Wilson, 1993),has proven to be advantageous in the diagno-sis, assessment, and treatment of various disor-


ders. Furthermore, reconfiguring the currentdiagnostic categories based only on emotion-related disturbances might leave out much clin-ically relevant information. However, atransdiagnostic approach would not necessar-ily have to supplant the current diagnostic cate-gories for it to inform our understanding of thesymptoms of various disorders. Indeed, mentaldisorders do not just consist of emotion-relatedsymptoms. Different symptoms reflect otherprocesses, including cognitive (e.g., inattention,thought disorder), behavioral (e.g., avoidance),and interpersonal (e.g., no close friends or con-fidants) processes. However, understanding theemotion-related symptoms that cut across dis-orders may help to refine the current diagnosticcategories without necessarily reconfiguringthem, as Watson (2005) has suggested. For ex-ample, knowing that individuals with depres-sion and panic disorder share the heightenedexperience of NA, even though panic is alsocharacterized by AA and depression is charac-terized by low PA, provides key informationabout similarities and differences betweenthese two disorders that is not readily acknowl-edged in the current diagnostic system.

CAUSAL ANDMAINTAINING PROCESSES

There have been several recent reviews of emo-tion disturbances in psychopathology (e.g.,Kring & Werner, 2004; Rottenberg & Johnson,2007). Instead of duplicating these efforts, I fo-cus here on candidate transdiagnostic emotiondisturbances that may reflect causal or main-taining processes. There are many points alongthe temporal course of disorders at which atransdiagnostic perspective may be informative(Barnett & Gotlib, 1988; Harvey et al., 2004).Specifically, transdiagnostic emotion distur-bances may be antecedents (i.e., predisposingor vulnerability factors), concomitants, or con-sequences (i.e., perpetuating or maintainingfactors). For an emotion disturbance to be con-sidered an antecedent transdiagnostic process,it must be shown to precede the onset of a dis-order. Disturbances that are observed duringactive episodes of disorders may be more accu-rately construed as concomitant transdiag-nostic processes, and disturbances that persistafter active episodes have abated might be con-sidered consequences or maintaining processes.Much of the evidence to date regarding specific

emotion disturbances in particular disorders ismost readily interpreted as evidence for main-taining processes. There has been less theoreti-cal, conceptual, and empirical work on pos-sible transdiagnostic causal or maintainingemotion processes than in the domains of phe-nomenology and treatment. Indeed, thedisorder-centric focus of research has been par-ticularly dominant in studies of possible etio-logical emotion disturbances, making integra-tion across disorders more of a challenge.Although space constraints preclude me fromconsidering all possible transdiagnostic emo-tion disturbances, I briefly consider four prom-ising candidates here: core affect, emotionawareness, emotion regulation, and emotiondisconnections.

Core Affect

As discussed earlier, Barrett and colleagues de-fine “core affect” as neurophysiological statesthat are experienced as pleasant or unpleasantand are ever-present indicators of a person’s re-lationship to his or her environment (Barrett,2006a; Barrett et al., 2007; Russell, 2003;Barrett & Russell, 1999). Findings from severalstudies that have measured reports of feelingstates have indicated that the experience of ex-cessive unpleasant affect (although it is not nec-essarily conceived of as “core affect”) is com-mon across many different disorders, includingdepression (for a review, see Mineka et al.,1998), the anxiety disorders (Mineka et al.,1998), eating disorders (e.g., Stice, 2001),schizophrenia (for a review, see Kring, 2001),substance-related disorders (e.g., Kassel,Stroud, & Paronis, 2003), and a number ofpersonality disorders (e.g., Berenbaum et al.,2006; Huperich, 2005; Putnam & Silk, 2005).Integrating findings across behavioral andbrain imaging studies, Barrett et al. (2007)have suggested that disturbances in core affectmay reflect an important emotion-relatedtransdiagnostic process in psychopathology. In-deed, the conceptual, theoretical, and empiricaladvances regarding core affect among healthyindividuals are ripe for translation into therealm of psychopathology.

Emotion Awareness

Barrett and Gross (2001) have argued thatknowledge and awareness of one’s emotionsare necessary prerequisites to effective emotion


regulation. However, simply having knowledgeabout emotion is not sufficient; rather, greateraccessibility of that emotion knowledge is be-lieved to promote effective emotion regulation.Individuals who describe their feelings in amore differentiated manner (e.g., “sad,” “con-fused,” “elated”) rather than more globally(e.g., “good,” “bad”) have greater accessibilityto and awareness of emotion knowledge anduse this knowledge when the regulation ofemotion may be necessary (Barrett, Gross,Christensen, & Benvenuto, 2001).

What is the evidence for emotion awarenessdifficulties in psychopathology? Although be-ing aware of how one feels is at the heart ofmost types of psychotherapy, surprisingly littleresearch has explicitly examined emotionawareness in psychopathology and how it maywax or wane with the exacerbation and remis-sion of symptoms. There is some evidence sug-gesting that patients with schizophrenia do notdiffer from individuals without schizophreniain terms of their emotion knowledge. However,patients with schizophrenia differentiate lessamong emotional states, and thus may be lesseffective at emotion regulation (Kring, Barrett,& Gard, 2003). Clinical conceptualizations ofborderline personality disorder include the no-tion that patients have difficulty distinguishingamong different emotional states (e.g., Westen,1991), but the empirical confirmation of thisnotion is needed. A recent investigation ofschizotypal personality disorder symptomsfound that such symptoms were associatedwith poor emotion clarity, but more attentionto emotions (Berenbaum et al., 2006).

Related to emotion awareness, the constructof “alexithymia” refers to difficulties in verbal-izing feelings. The most widely used measure ofthis construct, the Toronto Alexithymia Scale(Bagby, Parker, & Taylor, 1994), includes twosubscales that appear important to emotionawareness: difficulty in identifying feelings anddifficulty in describing feelings. This measurehas been widely used in correlational studies ofpsychiatric symptoms (e.g., Grabe, Spitzer, &Freyberger, 2004; Parker, Bagby, & Taylor,1991), anxiety disorders (e.g., Frewin, Pain,Dozois, & Lanius, 2006; Parker, Taylor, Bagby,& Acklin, 1993; Turk, Heimberg, Luterek,Mennin, & Fresco, 2005), eating disorders(e.g., Cochrane et al., 1993), personality dis-orders (e.g., Berenbaum, 1996), substance-related disorders (e.g., Speranza et al., 2004),and insomnia (e.g., Lundh & Broman, 2006).

However, the linkages between this constructand the various disorders are not always repli-cated across studies. Additional work is thusneeded to clarify alexithymia as a trans-diagnostic process, beyond correlations be-tween the one measure and symptoms withindisorders.

Emotion Regulation

Emotion regulation problems have been at theforefront of discussions about emotion distur-bances in psychopathology for at least the last10 years (e.g., Barlow et al., 2004; Gross &Muñoz, 1995; Kring & Werner, 2004; Linehan,1993). Broadly, emotion regulation refers toprocesses that serve to modify what we feel,when we feel it, and how we use that feeling toguide behavior (e.g., Gross, 1998). Many cur-rent diagnostic criteria explicitly refer to emo-tion regulation difficulties. For example, thecriteria “difficulty controlling anger” in bor-derline personality disorder; “efforts to avoidfeelings” in posttraumatic stress disorder(PTSD); “difficulty controlling worry” in gen-eralized anxiety disorder; and “rapidly shiftingexpressions of emotion” in histrionic personal-ity disorder all point to difficulties in regulatingemotions.

Despite the perceived importance of emotionregulation deficits in psychopathology, it is dif-ficult to integrate the literatures across disor-ders, because of the myriad approaches to con-straining the concept of emotion regulationacross studies. For example, some researchersdo not distinguish emotional responding fromregulation, following from theory that suggeststhe processes are indistinguishable (e.g., Cam-pos, Frankel, & Camras, 2004; Davidson,2000). Other studies examine emotion regula-tion within the individual, whereas still otherstudies examine emotion regulation from theoutside (e.g., having others provide soothingto down-regulate negative emotion). Disentan-gling emotion from emotion regulation re-mains a critical challenge for the field(Rottenberg & Gross, 2003). Greater concep-tual clarity will advance our understanding ofhow emotion regulation difficulties may belinked across different disorders.

A close cousin to emotion regulation is thetime course, or chronometry, of emotional re-sponses (e.g., Davidson, 1998). Emotional re-sponses are not wholly temporally constrainedby the presence of an eliciting stimulus, but in-


stead vary in their peak and duration in waysthat may hold important information about thenature of emotion disturbances in psychopath-ology. Two elements of the time course of anemotional response that have been studied inhealthy populations are (1) the time from theonset to the peak intensity of the response, and(2) the recovery time, or the time it takes forthe emotional response to resolve. This latterprocess, recovery time, is a probable trans-diagnostic emotion disturbance that remains atopic for future research. For example, the pro-longed experience of NA associated with de-pression, generalized anxiety disorder, and eat-ing disorders, or the prolonged experience ofPA associated with bipolar disorder, may re-flect a difficulty in the recovery time of emo-tional responding (e.g., Tomarken & Keener,1998).

Emotion Disconnections

A good deal of evidence indicates that patientswith schizophrenia report experiencing strongemotions in response to a variety of emotion-ally evocative stimuli (films, pictures, social in-teractions); yet they do not often display thesefeelings outwardly (for reviews, see Kring,2001; Kring & Werner, 2004). In other words,these patients’ outward displays of emotion arenot often accurate reflections of their experi-enced emotion, indicating a disconnection be-tween emotion response components. There issome evidence to suggest that this disconnec-tion may be present prior to the onset of the ill-ness. Walker, Grimes, Davis, and Smith (1993)analyzed home movies of adults with schizo-phrenia that were made before the adults devel-oped schizophrenia. They found that girls dis-played fewer joy expressions, and that bothboys and girls displayed more negative facialexpressions, compared to their healthy siblings.

Studies using the emotion-modulated startleparadigm (Lang et al., 1990) have observed adifferent disconnection among individuals withpsychopathy. Compared to healthy controls,these individuals showed comparable startle in-hibition during exposure to pleasant stimuli,but they did not show startle potentiation dur-ing exposure to aversive stimuli (e.g., Patrick,1994; Patrick, Bradley, & Lang, 1993). How-ever, the individuals with psychopathy did notdiffer from controls in their reported emotionalexperience to the aversive stimuli. Additionalevidence for this disconnection has been found

in imagery studies (Patrick, Cuthbert, & Lang,1994) and incidental memory paradigms(Christianson et al., 1996).

In both of the examples above, patients’ re-ports of emotional experience were indistin-guishable from healthy controls, but theirbehavioral or psychophysiological responsesdiffered. It may well be that such emotion dis-turbances are better construed as reflecting arelatively intact core affect system with corre-sponding behavioral system disturbances.

TREATMENT DEVELOPMENTAND EVALUATION

Although medication is a common form oftreatment for many different disorders, veryfew investigations have explicitly adopted atransdiagnostic approach to evaluating phar-macological treatment. Nevertheless, the evi-dence that particular medications may be effec-tive for multiple disorders is hiding in plainsight. For example, studies have found that an-tidepressant medications are effective at reduc-ing the symptoms of several other disorders, in-cluding specific and social phobias (e.g., Steinet al., 1998; Van Ameringen et al., 2001), panicdisorder (White & Barlow, 2002), generalizedanxiety disorder (Lydiard & Monnier, 2004),obsessive–compulsive disorder (Steketee &Barlow, 2002), posttraumatic stress disorder(Brady et al., 2000), some of the personalitydisorders (e.g., Rinne, van den Brink, Wouters,& van Dyck, 2002), and eating disorders (e.g.,Walsh et al., 2000). Following from such evi-dence, medications that were originally ap-proved by the U.S. Food and Drug Administra-tion (FDA) for the treatment of depressionhave since received approval (or an “indica-tion,” in FDA terminology) for the treatmentof other disorders. For example, paroxetine(Paxil) was later approved for the treatment ofobsessive–compulsive disorder, panic disorder,generalized anxiety disorder, and social anxi-ety; fluoxetine (Prozac) was later approved forthe treatment of obsessive–compulsive disorderand bulimia nervosa; sertraline (Zoloft) waslater approved for the treatment of obsessive–compulsive disorder, panic disorder, social anx-iety, and PTSD.

Although it may be the case that antidepres-sant medications are rather blunt instrumentsfor targeting the general distress that is com-mon across disorders, little research has di-


rectly examined the emotion-related mecha-nisms by which the medications might exerttheir transdiagnostic effects. However, weknow that the selective serotonin reuptake in-hibitors work on the neurotransmitter seroto-nin (as well as others, including dopamine; e.g.,Svenningsson et al., 2002), functionally leavingmore serotonin in the synapse, and that disrup-tions in serotonin have been implicated in de-pression (e.g., Thase, Jindal, & Howland,2002), anxiety disorders (e.g., Stein, 1998),and eating disorders (e.g., Carrasco, Dyaz-Marsa, Hollander, Cesar, & Saiz-Ruiz, 2000;Kaye et al., 1998). We also know a good dealabout how serotonin works throughout thebrain, and perhaps not surprisingly, this neuro-transmitter is heavily concentrated in areas ofthe brain linked with emotion (e.g., Barrett etal., 2007; Wrase et al., 2006). Finally, researchhas indicated that serotonin levels are associ-ated with PA among healthy individuals (e.g.,Duffy et al., 2006; Zald & Depue, 2001). Thebuilding blocks are thus available for con-structing a transdiagnostic approach to medi-cation treatment that explicitly links pharma-cology, neuroscience, and emotion. Much ofthis integrative work remains to be done, but itis certainly a fruitful avenue for future re-search.

Historically, different forms of psychothera-py were conceived of as treatments that couldbe applied across disorders or clinical problems(e.g., psychoanalysis). Furthermore, despite thedistinctly different theoretical traditions under-pinning various types of psychotherapy (e.g.,psychodynamic, interpersonal, gestalt, client-centered, behavioral), each of these traditionshas included some consideration of emotion(for reviews, see Greenberg, 2002b; Greenberg& Safran, 1987). One form of psychotherapythat is relevant to the focus of this chapter isemotion-focused therapy (EFT), which was de-veloped by Leslie Greenberg (e.g., Greenberg,2002a). Boiled down to its essence, EFT isbased on the idea that some emotions are adap-tive, whereas others are maladaptive. Mal-adaptive emotions are based on an underlyingloneliness, abandonment, worthlessness, anger,or inadequacy, and they can interfere with aperson’s relationships and overall functioning.The primary therapeutic goal is for a client tobecome more aware of these maladaptive emo-tions, to understand the source of these feel-ings, and to learn emotion regulation skills. Ac-cording to Greenberg (2002a), EFT is better

suited to particular types of clinical conditions(including depression and generalized anxietydisorder) and less well suited to others (such aspanic disorder). Unfortunately, data regardingthe efficacy of this treatment are limited. Norandomized controlled clinical trials have beenconducted, although smaller studies examiningthe process of change within EFT indicate thatthe treatment is effective for some clinicalproblems (Greenberg, 2002a). The theoreticalfoundations of this treatment continue to beenhanced by research in emotion, emotion andpsychopathology, and affective neuroscience(e.g., Greenberg, 2002b, 2004, and Chapter 6of this volume); a worthwhile endeavor for fu-ture research would be to examine whether thistreatment is effective in targeting transdiag-nostic emotion-related disturbances.

The shift to more disorder-specific psycho-therapeutic approaches perhaps began in thelate 1950s, with the pioneering work of JosephWolpe (1958), who developed systematic de-sensitization for the treatment of specific pho-bias. Additional disorder-specific psychothera-pies were developed in the 1960s, as cognitive-behavioral therapies became more prominent.A number of other influences in the fieldsince the 1970s have converged to solidify adisorder-specific approach to treatment devel-opment. These have included the greater de-mand to show that psychotherapy is effective;the development of the DSM-III, which washeavily influenced by the medical model(Wilson, 1993); the sophistication of researchmethods to evaluate treatment outcomes(e.g., Barlow, 2004; Westen, Novotny, &Thompson-Brenner, 2004); the emphasis onevidence-based practice (e.g., Barlow, 1996;Chambless & Hollon, 1998; Kendall, 1998);the efforts to position psychology as a healthcare profession and thus in the larger healthcare context (e.g., Johnson, 2001); and the ob-servation that psychological treatments tai-lored to specific disorders are as effective as, ormore effective than, other types of interven-tions. Indeed, as Barlow (2004) has noted,“few would argue that diversity in proceduresto address specific aspects of pathology is notnecessary” (p. 873).

The disorder-specific approach to treatmentdevelopment, particularly in the context of em-pirically supported treatments (Chambless &Hollon, 1998), has spawned a large number ofindividual treatment protocols and manuals.This proliferation of different treatment proto-


cols has undoubtedly benefited countless indi-viduals who have received these treatments, asthey have been shown to be effective (Nathan& Gorman, 2002). However, the sheer magni-tude of treatment protocols has become a bitoverwhelming to treatment professionals, withrespect to both learning the varied protocolsand disseminating them to a broader range oftreatment providers (Barlow et al., 2004; Per-sons, 2005). Furthermore, the reality in clinicalpractice is that providers often select bits andpieces from a number of different protocols, inorder to provide the best possible treatment fora given individual patient (Persons, 2005).Partly in reaction to this overwhelming numberof treatment protocols, there has recently beena call for more unified treatments across disor-ders. This conversation has been situated pri-marily within the mood and anxiety disorders,and it has been informed by research onemotion-related disturbances that are commonacross these disorders.

Following from research on the structuralconfiguration of descriptive phenomenologyand shared etiologies across mood and anxietydisorders, Barlow and colleagues have pro-posed a unified treatment for these disorders(Barlow et al., 2004; Moses & Barlow, 2006).The focus of this intervention is on putativeemotion-related mechanisms that may be driv-ing the emotion disturbances that cut acrossmood and anxiety disorders. The treatment hasthree main components: (1) altering cognitivereappraisals, a key component in emotion reg-ulation processes; (2) preventing emotionalavoidance; and (3) changing emotion actiontendencies, or replacing emotion behaviors as-sociated with fear and anxiety with behaviorsrelated to positive emotions. These key compo-nents of the intervention have their origin inbasic science in emotion, in both healthy anddisordered individuals. Time will tell whetherthe intervention is as effective as other avail-able treatments. Because the intervention is de-signed to target emotion-related mechanismsthat cut across disorders, it is transdiagnostic atheart, and it seems likely that it will pay off. In-deed, there has been a call for more treatmentsto target mechanisms rather than disorders perse. For example, Rosen and Davison (2003)have argued that we should be defining empiri-cally supported principles of change ratherthan empirically supported treatments.

Other recent treatments have been designedto target emotional disturbances, such as

Mennin and colleagues’ emotion regulationtreatment for generalized anxiety disorder(e.g., Mennin, 2004). Though this interventionwas designed around the emotion regulationproblems associated with generalized anxietydisorder, it seems probable that it would beuseful for a number of disorders with similardifficulties in emotion regulation, such as an in-ability to down-regulate intense negative emo-tions and a lack of awareness of negative emo-tions.


Advances in affective neuroscience and basicbehavioral research in emotion have greatlycontributed to our understanding of emotiondisturbances in psychopathology. Indeed,methods, theories, and measures developed inthese domains have allowed us to achievegreater clarity regarding the reach of emotiondisturbances across many different disorders.With this clarity has come the realization thatmany of the observed emotion disturbancesmay be common across disorders. Progress inunderstanding the reach of transdiagnosticemotion disturbances has begun to be achievedat the levels of descriptive phenomenology andtreatment. Although transdiagnostic treatmentapproaches targeting emotion disturbances aregrounded in theory regarding emotion-basedmechanisms, more work is needed to unpackthe nature of transdiagnostic etiological pro-cesses that are emotion-based.

REFERENCES

American Psychiatric Association. (2000). Diagnosticand statistical manual of mental disorders (4th ed.,text rev.). Washington, DC: Author.

Bagby, M. R., Parker, J. D. A., & Taylor, G. J. (1994).The twenty-item Toronto Alexithymia Scale: I. Itemselection and cross validation of the factor struc-ture. Journal of Psychosomatic Research, 38, 23–32.

Barlow, D. H. (1996). The effectiveness of psychothera-py: Science and policy. Clinical Psychology: Scienceand Practice, 1, 109–122.

Barlow, D. H. (2004). Psychological treatments. Ameri-can Psychologist, 59, 869–878.

Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). To-ward a unified treatment for emotional disorders.Behavior Therapy, 35, 205–230.

Barnett, P. A., & Gotlib, I. H. (1988). Psychosocial func-tioning and depression: Distinguishing among ante-


cedents, concomitants, and consequences. Psycho-logical Bulletin, 104, 97–126.

Barrett, L. F. (2006a). Solving the emotion paradox:Categorization and the experience of emotion. Per-sonality and Social Psychology Review, 10, 20–46.

Barrett, L. F. (2006b). Valence is a basic building blockof emotional life. Journal of Research in Personality,40, 35–55.

Barrett, L. F., & Gross, J. J. (2001). Emotional intelli-gence: A process model of emotion representationand regulation. In T. J. Mayne & G. A. Bonanno(Eds.), Emotions: Current issues and future direc-tions (pp. 286–310). New York: Guilford Press.

Barrett, L. F., Gross, J. J., Christensen, T. C., &Benvenuto, M. (2001). Knowing what you’re feelingand knowing what to do about it: Mapping the rela-tion between emotion differentiation and emotionregulation. Cognition and Emotion, 15, 713–724.

Barrett, L. F., Mesquita, B., Ochsner, K., & Gross, J. J.(2007). The experience of emotion. Annual Reviewof Psychology, 58, 373–403.

Barrett, L. F., & Russell, J. A. (1999). The structure ofcurrent affect: Controversies and emerging consen-sus. Current Directions in Psychological Science, 8,10–14.

Berenbaum, H. (1996). Childhood abuse, alexithymia,and personality disorder. Journal of PsychosomaticResearch, 41, 585–595.

Berenbaum, H., Boden, M. T., Baker, J. P., Dizen, M.,Thompson, R. J., & Abramowitz, A. (2006). Emo-tional correlates of the different dimensions ofschizotypal personality disorder. Journal of Abnor-mal Psychology, 115, 359–368.

Berenbaum, H., & Oltmanns, T. F. (1992). Emotionalexperience and expression in schizophrenia and de-pression. Journal of Abnormal Psychology, 101, 37–44.

Berenbaum, H., Raghavan, G., Le, H.-N., Vernon, L. L.,& Gomez, J. J. (2003). A taxonomy of emotional dis-turbances. Clinical Psychology: Science and Practice,10, 206–226.

Bradley, M. M., & Lang, P. J. (2000). Measuring emo-tion: Behavior, feeling, and physiology. In R. D. Lane& L. Nadel (Eds.), Cognitive neuroscience of emo-tion (pp. 242–276). New York: Oxford UniversityPress.

Brady, K., Pearlstein, T., Asnis, G. M., Baker, D.,Rothbaum, B., Sikes, R., & Farfel, G. U. (2000). Effi-cacy and safety of sertraline treatment of posttrau-matic stress disorder. Journal of the American Medi-cal Association, 283, 1837–1844.

Brown, T. A., & Barlow, D. H. (1992). Comorbidityamong anxiety disorders: Implications for treatmentand DSM-IV. Journal of Consulting and Clinical Psy-chology, 60, 835–44.

Brown, T. A., Campbell, L. A., Lehman, C. L., et al.(2001). Current and lifetime comorbidity of theDSM-IV anxiety and mood disorders in a large clini-cal sample. Journal of Abnormal Psychology, 110,585–599.

Brown, T. A., Chorpita, B., & Barlow, D. H. (1998).Structural relationships among dimensions of theDSM-IV anxiety and mood disorders and dimensionsof negative affect, positive affect, and autonomicarousal. Journal of Abnormal Psychology, 107, 179–192.

Campos, J. J., Frankel, C. B., & Camras, L. (2004). Onthe nature of emotion regulation. Child Develop-ment, 75, 377–394.

Carrasco, J. L., Dyaz-Marsa, M., Hollander, E., Cesar,J., & Saiz-Ruiz, J. (2000). Decreased monoamineoxidase activity in female bulimia. EuropeanNeuropsychopharmacology, 10, 113–117.

Chambless, D., & Hollon, S. (1998). Defining empiri-cally supported therapies. Journal of Consulting andClinical Psychology, 66, 7–18.

Christianson, S. A., Forth, A. E., Hare, R. D., Strachan,C., Lidberg, L., & Thorell, L. H. (1996). Remem-bering details of emotional events: A comparison be-tween psychopathic and nonpsychopathic offenders.Personality and Individual Differences, 20, 437–443.

Clark, L. A., & Watson, D. (1991). Tripartite model ofanxiety and depression: Psychometric evidence andpsychometric implications. Journal of Abnormal Psy-chology, 100, 316–336.

Cochrane, C. E., Brewerton, T. D., Wilson, D. B., &Hodges, E. L. (1993). Alexithymia in the eating dis-orders. International Journal of Eating Disorders,14, 219–228.

Davidson, R. J. (1998). Affective style and affective dis-orders: Perspectives from affective neuroscience.Cognition and Emotion, 12, 307–330.

Davidson, R. J. (2000). The functional neuroanatomyof affective style. In R. D. Lane & L. Nadel (Eds.),Cognitive neuroscience of emotion (pp. 371–388).New York: Oxford University Press.

Duffy, M. E., Stewart-Knox, B. J., McConville, C.,Bradbury, I., O’Connor, J., Helander, A., et al.(2006). The relationship between whole blood sero-tonin and subjective mood in apparently healthypostmenopausal women. Biological Psychiatry, 73,165–168.

Fairburn, C. G., Cooper, Z., & Shafran, R. (2003). Cog-nitive behavioural treatment for eating disorders: Atransdiagnostic theory and treatment. Behaviour Re-search and Therapy, 41, 509–528.

Frewin, P. A., Pain, C., Dozois, D. J. A., & Lanius, R. A.(2006). Alexithymia in PTSD: fMRI and psychomet-ric studies. Annals of the New York Academy of Sci-ences, 1071, 397–400.


Grabe, H. J., Spitzer, C., & Freyberger, H. J. (2004).Alexithymia and personality in relation to dimen-sions of psychopathology. American Journal of Psy-chiatry, 161, 1299–1301.

Greenberg, L. S. (2002a). Emotion-focused therapy:Coaching clients to work through their feelings.Washington, DC: American Psychological Associa-tion.


Greenberg, L. S. (2002b). Integrating and emotion-focused approach to treatment into psychotherapyintegration. Journal of Psychotherapy Integration,12, 154–189.

Greenberg, L. S. (2004). Emotion-focused therapy. Clin-ical Psychology and Psychotherapy, 11, 3–16.

Greenberg, L. S., & Safran, J. D. (1987). Emotion inpsychotherapy. New York: Guilford Press.

Gross, J. J. (1998). The emerging field of emotion regu-lation: An integrative review. Review of General Psy-chology, 2, 271–299.

Gross, J. J., & Muñoz, R. F. (1995). Emotion regulationand mental health. Clinical Psychology: Science andPractice, 2, 151–164.

Harvey, A., Watkins, E., Mansell, W., & Shafran, R.(2004). Cognitive behavioural processes across psy-chological disorders: A transdiagnostic approach toresearch and treatment. New York: Oxford Univer-sity Press.

Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999).Acceptance and commitment therapy: An experien-tial approach to behavior change. New York:Guilford Press.

Hooley, J. M., Richters, J. E., Weintraub, S., & Neale, J.M. (1987). Psychopathology and marital distress:The positive side of positive symptoms. Journal ofAbnormal Psychology, 96, 27–33.

Huperich, S. K. (2005). Differentiating avoidant and de-pressive personality disorders. Journal of PersonalityDisorders, 19, 659–673.

Johnson, N. G. (2001, April). President’s column: Psy-chology’s mission includes health: An opportunity.Monitor on Psychology, 32(4). Retrieved fromwww.apa.org/monitor/apr01/pc.html

Linehan, M. M. (1993). Cognitive-behavioral treatmentof borderline personality disorder. New York:Guilford Press.

Kassel, J. D., Stroud, L. R., & Paronis, C. A. (2003).Smoking, stress, and negative affect: Correlation,causation, and context across stages of smoking. Psy-chological Bulletin, 129, 270–304.

Kaye, W. H., Greeno, C. G., Moss, H., Fernstrom, J.,Lilenfeld, L. R., Wahlund, B., et al. (1998). Alter-ations in serotonin activity and platelet monoamineoxidase and psychiatric symptoms after recoveryfrom bulimia nervosa. Archives of General Psychia-try, 55, 927–935.

Keltner, D., & Kring, A. M. (1998). Emotion, socialfunction, and psychopathology. Review of GeneralPsychology, 2, 320–342.

Kendall, P. C. (1998). Empirically supported psycholog-ical therapies. Journal of Consulting and ClinicalPsychology, 66, 3–6.

Kessler, R. C., Chiu, W. T., Demler, O., & Walters, E.(2005). Prevalence, severity, and comorbidity of 12month DSM-IV disorders in the National Comorbid-ity Survey replication. Archives of General Psychia-try, 62, 617–627.

Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C.B., Hughes, M., Eshleman, S., et al. (1994). Lifetime

and 12-month prevalence of DSM-III-R psychiatricdisorders in the United States: Results from the Na-tional Comorbidity Survey. Archives of General Psy-chiatry, 51, 8–19.

Krause, R., Steimer-Krause, E., & Hufnagel, H. (1992).Expression and experience of affects in paranoidschizophrenia. Revue Européenne de PsychologieAppliquée, 42, 131–138.

Kring, A. M. (2001). Emotion and psychopathology. InT. J. Mayne & G. Bonanno (Eds.), Emotion: Currentissues and future directions (pp. 337–360). NewYork: Guilford Press.

Kring, A. M., & Bachorowski, J.-.A. (1999). Emotionand psychopathology. Cognition and Emotion, 13,575–599.

Kring, A. M., Barrett, L. F., & Gard, D. E. (2003). Onthe broad applicability of the circumplex: Represen-tations of affective knowledge in schizophrenia. Psy-chological Science, 14, 207–214.

Kring, A. M., & Earnst, K. S. (1999). Stability of emo-tional responding in schizophrenia. Behavior Ther-apy, 30, 373–388.

Kring, A. M., & Neale, J. M. (1996). Do schizophrenicpatients show a disjunctive relationship among ex-pressive, experiential, and psychophysiological com-ponents of emotion? Journal of Abnormal Psychol-ogy, 105, 249–257.

Kring, A. M., & Werner, K. H. (2004). Emotion regula-tion in psychopathology. In P. Philippot & R. S.Feldman (Eds.), The regulation of emotion (pp. 359–385). Mahwah, NJ: Erlbaum.


Levenson, R. W. (1994). Human emotion: A functionalview. In P. Ekman & R. J. Davidson (Eds.), The na-ture of emotion (pp. 123–126). New York: OxfordUniversity Press.

Lundh, L.-G., & Broman, E.-L. (2006). Alexithymiaand insomnia. Personality and Individual Differ-ences, 40, 1615–1624.

Lydiard, R. B., & Monnier, J. (2004). Pharmacologicaltreatment. In R. G. Heimberg, C. L. Turk, & D. S.Mennin (Eds.), Generalized anxiety disorder(pp. 351–381). New York: Guilford Press.

Mauss, I., Levenson, R. W., McCarter, L., Wilhelm, F.,& Gross, J. J. (2005). The tie that binds?: Coherenceamong emotion experience, behavior, and physiol-ogy. Emotion, 5, 175–190.

Mennin, D. (2004). Emotion regulation therapy for gen-eralized anxiety disorder. Clinical Psychology andPsychotherapy, 11, 17–29.

Mesquita, B., & Karasawa, M. (2002). Different emo-tional lives. Cognition and Emotion, 16, 127–141.

Mineka, S., Watson, D., & Clark, L. A. (1998). Comor-bidity of anxiety and unipolar mood disorders. An-nual Review of Psychology, 49, 377–412.

Moses, E. B., & Barlow, D. H. (2006). A new unifiedtreatment approach for emotional disorders based on


emotion science. Current Directions in PsychologicalScience, 15, 146–150.

Nathan, P. E., & Gorman, J. M. (Eds.). (2002). A guideto treatments that work. London: Oxford UniversityPress.

Norton, P. J., Hayes, S. A., & Hope, D. A. (2004). Ef-fects of a transdiagnostic group treatment for anxietydisorders on secondary depression. Depression andAnxiety, 20, 198–202.

Parker, J. D. A., Bagby, M., & Taylor, G. J. (1991).Alexithymia and depression: Distinct or overlappingconstructs? Comprehensive Psychiatry, 32, 387–394.

Parker, J. D. A., Taylor, G. J., Bagby, R. M., & Acklin,M. W. (1993). Alexithymia in panic disorder andsimple phobia: A comparative study. American Jour-nal of Psychiatry, 150, 1105–1107.

Patrick, C. J. (1994). Emotion and psychopathy: Star-tling new insights. Psychophysiology, 31, 319–330.

Patrick, C. J., & Bernat, E. M. (2006). The construct ofemotion as a bridge between personality and psycho-pathology. In R. F. Krueger & J. L. Tackett (Eds.),Personality and psychopathology (pp. 174–209).New York: Guilford Press.

Patrick, C. J., Bradley, M. M., & Lang, P. J. (1993).Emotion in the criminal psychopath: Startle reflexmodulation. Journal of Abnormal Psychology, 102,82–92.

Patrick, C. J., Cuthbert, B. N., & Lang, P. J. (1994).Emotion in the criminal psychopath: Fear image pro-cessing. Journal of Abnormal Psychology, 103, 523–534.

Persons, J. B. (2005). Empiricism, mechanism, and thepractice of cognitive-behavior therapy. BehaviorTherapy, 36, 107–118.

Persons, J. B., Roberts, N. A., & Zalecki, C. (2003).Anxiety and depression change together during treat-ment. Behavior Therapy, 34, 149–163.

Putnam, K. M., & Silk, K. R. (2005). Emotion dysregu-lation and the development of borderline personalitydisorder. Development and Psychopathology, 17,899–925.

Rinne, T., van den Brink, W., Wouters, L., & van Dyck,R. (2002). SSRI treatment of borderline personalitydisorder: A randomized, placebo-controlled clinicaltrial for female patients with borderline personalitydisorder. American Journal of Psychiatry, 159, 2048–2054.

Rosen, G. M., & Davison, G. C. (2003). Psychologyshould list empirically supported principles of change(ESPs) and not credential trademarked therapies orother treatment packages. Behavior Modification,27, 300–312.

Rottenberg, J., & Gross, J. J. (2003). When emotiongoes wrong: Realizing the promise of affective sci-ence. Clinical Psychology: Science and Practice, 10,227–232.

Rottenberg, J., & Johnson, S. L. (Eds.). (2007). Emo-tion and psychopathology: Bridging affective andclinical science. Washington, DC: American Psycho-logical Association.


Speranza, M., Corcos, M., Stéphan, P., Loas, G., Pérez-Diaz, F., Lang, F., et al. (2004). Alexithymia, depres-sive experiences, and dependency in addictive disor-ders. Substance Use and Misuse, 39, 567–595.

Stein, M. B. (1998). Neurobiological perspectives on so-cial phobia: From affiliation to zoology. BiologicalPsychiatry, 44, 1277–1285.

Stein, M. B., Liebowitz, M. R., Lydiard, R. B., Pitts, C.D., Bushnell, W., & Gergel, I. (1998). Paroxetinetreatment of generalized social phobia (social anxietydisorder): A randomized clinical trial. Journal of theAmerican Medical Association, 280, 708–713.

Steketee, G., & Barlow, D. H. (2002). Obsessive–compulsive disorder. In D. H. Barlow, Anxiety andits disorders: The nature and treatment of anxietyand panic (pp. 516–550). New York: GuilfordPress.

Stice, E. (2001). A prospective test of the dual-pathwaymodel of bulimic pathology: Mediating effects of di-eting and negative affect. Journal of Abnormal Psy-chology, 110, 124–135.

Svenningsson, P., Tzavara, E. T., Witkin, J. M.,Fienberg, A. A., Nomikos, G. G., & Greengard, P.(2002). Involvement of striatal and extrastriatalDARPP-32 in biochemical and behavioral effectsof fluoxetine (Prozac). Proceedings of the Na-tional Academy of Sciences of the USA, 99, 3182–3187.

Thase, M. E., Jindal, R., & Howland, R. H. (2002). Bio-logical aspects of depression. In C. L. Hammen & I.H. Gotlib (Eds.), Handbook of depression (pp. 192–218). New York: Guilford Press.

Thayer, J. F., & Brosschot, J. F. (2005). Psychomaticsand psychopathology: Looking up and down fromthe brain. Psychoneuroendocrinology, 30, 1050–1058.

Tomarken, A. J., & Keener, A. D. (1998). Frontal brainasymmetry and depression: A self-regulatory per-spective. Cognition and Emotion, 12. 387–420.

Tsao, J. C. I., Mystkowski, J. L., Zucker, B. G., &Craske, M. G. (2002). Effects of cognitive behavioraltherapy for panic disorder on comorbid conditions:Replication and extension. Behaviour Therapy, 33,493–509.

Turk, C. L., Heimberg, R. G., Luterek, J. A., Mennin, D.S., & Fresco, D. M. (2005). Emotion dysregulation ingeneralized anxiety disorder: A comparison with so-cial anxiety disorder. Cognitive Therapy and Re-search, 29, 89–106.

Van Ameringen, M. A., Lane, R. M., Walker, J. R.,Bowen, R. C., Chokka, P. R., Goldner, E. M., et al.(2001). Sertraline treatment of generalized socialphobia: A 20-week, double-blind, placebo-controlledstudy. American Journal of Psychiatry, 158, 275–281.

Walker, E. F., Grimes, K. E., Davis, D. M., & Smith, A.J. (1993). Childhood precursors of schizophrenia:


Facial expressions of emotion. American Journal ofPsychiatry, 150, 1654–1660.

Walsh, B. T., Agras, S. W., Devlin, M. J., Fairburn, C.G., Wilson, G. T., Kahn, C., & Chally, M. K. (2000).Fluoxetine for bulimia nervosa following poor re-sponse to psychotherapy. American Journal of Psy-chiatry, 157, 1332–1334.

Watson, D. (2003). Subtypes, specifiers, epicycles, andeccentrics: Toward a more parsimonious taxonomyof psychopathology. Clinical Psychology: Science andPractice, 10, 233–238.

Watson, D. (2005). Rethinking the mood and anxietydisorders: A quantitative hierarchical model for DSM-V. Journal of Abnormal Psychology, 114, 522–536.

Watson, D., Clark, L. A., Weber, K., Assenheimer, J. S.,Strauss, M. E., & McCormick, R. A. (1995). Testinga tripartite model II: Exploring the symptom struc-ture of anxiety and depression in student, adult, andpatient samples. Journal of Abnormal Psychology,104, 15–25.

Watson, D., Weber, K., Assenheimer, J. S., Clark, L. A.,Strauss, M. E., & McCormick, R. A. (1995). Testinga tripartite model I: Evaluating the convergent anddiscriminant validity of anxiety and depressionscales. Journal of Abnormal Psychology, 104, 3–14.

Watson, D., Weise, D., Vaidya, J., & Tellegen, A.(1999). The two general activation systems of affect:Structural findings, evolutionary considerations, andpsychobiological evidence. Journal of Personalityand Social Psychology, 76, 820–838.

Westen, D. (1991). Cognitive-behavioral interventionsin the psychoanalytic psychotherapy of borderlinepersonality disorders. Clinical Psychology Review,11, 211–230.

Westen, D., Novotny, C. M., & Thompson-Brenner, H.(2004). The empirical status of empirically supportedpsychotherapies: Assumptions, findings, and report-ing in controlled clinical trials. Psychological Bulle-tin, 130, 631–663.

White, K. S., & Barlow, D. H. (2002). Panic disorderand agoraphobia. In D. H. Barlow, Anxiety and itsdisorders: The nature and treatment of anxiety andpanic (pp. 328–379). New York: Guilford Press.

Wilson, M. (1993). DSM-III and the transformation ofAmerican psychiatry: A history. American Journal ofPsychiatry, 150, 399–410.

Wolpe, J. (1958). Psychotherapy by reciprocal inhibi-tion. Stanford, CA: Stanford University Press.

Wrase, J., Reimold, M., Puls, I., Kienast, T., & Heinz, A.(2006). Serotonergic dysfunction: Brain imaging andbehavioral correlates. Cognitive, Affective, andBehavioral Neuroscience, 6, 53–61.

Zald, D. H., & Depue, R. A. (2001). Serotonergic func-tioning correlates with positive and negative affect inpsychiatrically healthy males. Personality and Indi-vidual Differences, 30, 71–86.

Zinbarg, R. E., & Barlow, D. H. (1996). Structure ofanxiety and the anxiety disorders: A hierarchicalmodel. Journal of Abnormal Psychology, 105, 181–193.


PA R T V I I I

SELECT EMOTIONS

C H A P T E R 4 4

Fear and AnxietyOverlaps and Dissociations

ARNE ÖHMAN

Fear is a ubiquitous experience among human-kind that can be traced back to a distant mam-malian heritage. Recent world events, withterrorist attacks randomly striking innocentbystanders at many places, highlight the long-standing insight that fear is an inevitable partof human existence. Throughout human his-tory, fear and its close ally, anxiety, providerecurrent themes for people pondering their ex-istential predicament and have inspired fre-quent artistic representations. In a clinical con-text, the vicissitudes of fear and anxiety havebeen understood as keys to the dynamics ofpsychopathology—whether these are concep-tualized in terms of a learnable drive support-ing escape and avoidance, or as the targets ofpsychologically distorting defense mechanisms.The ubiquity and controversial status of fearand anxiety have made them central topics forresearch and reflection; they have generated avoluminous literature (see, e.g., Barlow, 2002;

Bourke, 2006; Tuma & Maser, 1985), only atiny fraction of which can be represented in thischapter.

The point of departure for this chapter isthat fear and anxiety are closely related emo-tional phenomena originating in evolved mam-malian defense systems. Nonetheless, in spiteof their overlap, research during the last decadehas started to unravel important differences be-tween them. The chapter starts with a briefconceptual overview and proceeds to considerthe situational contexts of fear in an evolution-ary perspective. A model of fast, unconsciousmechanisms in the generation of fear is pre-sented and discussed in relation to behavioraldata from backward-masking and attentionalparadigms. Then follows a section devoted tothe biology of fear and anxiety, reviewing neu-roanatomical, genetic, and psychophysiologi-cal data that collectively provide a distinctionbetween them. The chapter concludes with a

709

discussion of the implications of the knowledgepresented in prior sections for the clinical un-derstanding of fear and anxiety.

THE PHENOMENA OF FEARAND ANXIETY

Fear and anxiety are obviously overlapping,aversive, activated states centered on threat.They both involve intense negative feelingsand strong bodily manifestations. Subjectively,however, they take somewhat different forms.Fear denotes dread of impending disaster andan intense urge to defend oneself, primarily bygetting out of the situation. Clinical anxiety, onthe other hand, has been described as an ineffa-ble and unpleasant feeling of foreboding (Lader& Marks, 1973). Accordingly, the glossary ofthe Diagnostic and Statistical Manual of Men-tal Disorders, fourth edition, text revision(DSM-IV-TR; American Psychiatric Associa-tion [APA], 2000), uses the term “anxiety” todenote “apprehensive anticipation of futuredanger or misfortune accompanied by a feelingof dysphoria or somatic symptoms of tension”(p. 820). “Fear” is said to differ from anxietyprimarily in having an identifiable elicitingstimulus. In a sense, therefore, anxiety is often“prestimulus” (i.e., anticipatory to [more orless real] threatening stimuli), whereas fear is“poststimulus” (i.e., elicited by a defined fearstimulus).

As eloquently argued by Epstein (1972), ex-ternal stimuli are insufficient to distinguish fearand anxiety. He concluded that fear is relatedto coping behavior, particularly escape andavoidance. However, when coping attempts fail(e.g., because the situation is uncontrollable),fear is turned into anxiety. In Epstein’s view,then, “fear is an avoidance motive. If therewere no restraints, internal or external, fearwould support the action of flight. Anxiety canbe defined as unresolved fear, or, alternatively,as a state of undirected arousal following theperception of threat” (1972, p. 311). In fear,therefore, it is an obvious (albeit not necessar-ily clearly perceived) danger located in spaceand time that must be dealt with; in anxiety, onthe other hand, the nature and location of thethreat remain more obscure and thus are diffi-cult to cope with by active defensive maneu-vers.

Fear may be focused on external sources, asin phobias. Or anxiety may be situationally

unfocused, as in free-floating or generalizedanxiety. The latter may come in the form of epi-sodic panic attacks (i.e., as sudden emotionalsurges dominated by physical symptoms, some-times with and sometimes without clearprecipitants), which form the backbone ofpanic disorder (PD). Or it may be a relative-ly constant mental preoccupation with moreor less reasonable threats and dangers—associated with somatic symptoms—as in gen-eralized anxiety disorder (GAD).

As these descriptions imply, fear and anxietycan be regarded both as emotional states,evoked in a particular context and having alimited duration, and as personality traits,characterizing individuals across time and situ-ations (e.g., Spielberger, 1972; Rapee, 1991).The differences between clinical and normalfear/anxiety include that the former is more re-current and persistent; that its intensity isunreasonable, given the objective danger orthreat; that it tends to paralyze individuals,making them helpless and unable to cope; andthat it results in impeded psychosocial or physi-ological functioning (e.g., Lader & Marks,1973).

THE SITUATIONAL CONTEXTOF FEAR AND ANXIETY

Basically, fear is a functional emotion with adeep evolutionary origin, reflecting the factthat earth has always been a hazardous envi-ronment to inhabit. Staying alive is a prerequi-site for the basic goal of biological evolution—sending genes on to subsequent generations.Hence even the most primitive of organismshave developed defense responses to deal withlife threats in their environment, whether theseare unhealthy chemicals in the surroundings,circumstances suggesting a hunting predator,or aggressive conspecifics. Viewed from theevolutionary perspective, fear is central tomammalian evolution. As a product of naturalselection, it is shaped and constrained by evolu-tionary contingencies. Our evolutionary his-tory is obvious in the fears and phobias that wehumans exhibit and readily learn. We are morelikely to fear events and situations that pro-vided threats to the survival of our ancestors,such as potentially deadly predators, heights,and wide-open spaces, than to fear the mostfrequently encountered potentially deadly ob-jects in our contemporary environment, such as

710 VIII. SELECT EMOTIONS

handguns or motorcycles (Öhman & Mineka,2001; Seligman, 1971).

Traumatic Situations

Extreme danger jeopardizing one’s life (or thelives of close kin) elicits intense fright and mayhave long-lasting consequences in the form ofposttraumatic stress disorder (PTSD). Traumamay involve natural catastrophes, such asfloods or hurricanes destroying homes or com-munities; or it may involve seeing others beingseriously injured or killed as a result of an acci-dent or physical violence. If the trauma resultsin PTSD, the traumatic event is persistentlyreexperienced (e.g., in the form of “flash-backs”); stimuli or events associated with thetrauma are avoided; and the person feels gener-ally numbed with regard to emotions. Com-mon anxiety symptoms experienced by personssuffering from PTSD include sleep and concen-tration difficulties, irritability or anger out-bursts, hypervigilance, and exaggerated startle.Some events, such as torture, frequently resultin PTSD, whereas others, such as natural disas-ters or car accidents, only occasionally result inthe disorder (APA, 2000).

Commonly Fearedand Potentially Phobic Situations

Survival considerations, either contemporaryor in an evolutionary perspective, are relevantfor most situational dimensions of humanfears. Arrindell, Pickersgill, Merckelbach,Ardon, and Cornet (1991) reviewed studiesfactor-analyzing questionnaire data on self-reported fear, arguing that the results fromnearly 200 studies fit a simple structure involv-ing four factors. The first factor was “fearsabout interpersonal events or situations.” It in-cluded fears of criticism and social interaction,rejection, conflicts, and evaluation, but also in-terpersonal aggression and display of sexualand aggressive scenes. The second factor was“fears related to death, injuries, illness, blood,and surgical procedures.” This factor had aquite heterogeneous content, with fear ofbodily ailments, illness, and death as the com-mon theme. The third factor, “fear of animals,”included common domestic animals; othersmall, often harmless animals; and creepingand crawling animals such as insects andreptiles. Finally, “agoraphobic fears” was thefourth factor. It involved fear of entering public

places (such as stores or shopping malls) andcrowds, but also fear of closed spaces (such aselevators, theaters, or churches) or places lack-ing an escape route to a safe haven (such asbridges and tunnels, and traveling alone intrains, buses, or airplanes).

All these four factors represent situations ofrelevance for human evolution. Human historyis replete with examples of how social conflictsthat have escalated out of control can createpotentially deadly dangers, not to speak of thesocial threat in terms of the defeat and humilia-tion they may involve (Öhman, 1986). Thus itcomes as no surprise that social interactions aresometimes feared. For fear of death and illness,and associated bodily conditions, there is noneed to elaborate the potential survival threat.Although many animals are friendly andsought as companions, there is no question thatanimals as predators have posed recurrentthreats that have been an important force inshaping the evolution of humankind. Indeed, itis reasonable to give reptiles a privileged posi-tion as the prototypical predators (Isbell, 2006;Öhman, 1986; Öhman, Dimberg, & Öst,1985; Öhman & Mineka, 2001, 2003). Finally,agoraphobic fears center on the lack of securityinherent in separation from safe bases and kin,and the avoidance of places associated withpanic and feelings of discomfort.

These factors correspond to four prominenttypes of phobias (intense fright and avoidanceof specific situations or events): social phobia,blood phobia, animal phobia, and agorapho-bia. Furthermore, the second factor incorpo-rates fears often encountered in PD, such asfears or syncope, and in obsessive–compulsivedisorder (OCD), such as contamination. De-parting from a preparedness perspective onphobias (Seligman, 1971), my colleagues and I(Öhman et al., 1985; Öhman & Mineka, 2001)have argued that these fears reflect basic,evolutionarily shaped behavioral systems. Spe-cifically, we have suggested that social fearsoriginated in a dominance–submissiveness sys-tem, the adaptive function of which was to pro-mote social order by facilitating the establish-ment of dominance hierarchies. Animal fears,on the other hand, are attributed to a predatorydefense system, originating in the fear of theexclusive predators of early mammals—snakesand reptiles—and later in fear of venomoussnakes among monkeys and apes (Isbell, 2006).

These evolved systems involve stimuli forwhich organisms were adaptively shaped by

44. Fear and Anxiety 711

evolution to associate easily with fear and de-fense. By co-occurring with actual danger, theymay through Pavlovian fear conditioning becompromised into producing social and animalphobias when the fear responses they engenderbecome conditioned to stimuli that actually areharmless in the ecology of modern humans(Öhman & Mineka, 2001). The basic argu-ment is that evolution has equipped humanswith a propensity to associate fear with situa-tions that threatened the survival of their an-cestors (Seligman, 1971). Hence, although hu-mans in general are prepared to acquire somefears (e.g., snake fears) easily, some individualsare more prepared than others to develop fearsof particular situations. Thus the developmentof phobias is jointly determined by genetic pre-dispositions and specific environmental expo-sures. We have tested these ideas (see reviewsby Dimberg & Öhman, 1996; McNally, 1987;Öhman, 1993; Öhman & Mineka, 2001) inautonomic conditioning experiments primarilycomparing acquisition and resistance to extinc-tion of skin conductance conditioning to po-tentially phobic (e.g., snakes, spiders, angryfaces) and neutral (e.g., flowers, mushrooms,neutral faces, or friendly faces) stimuli. Thegeneral finding has been that responses condi-tioned to potentially phobic stimuli show en-hanced resistance to extinction, compared toresponses conditioned to neutral stimuli(Öhman & Mineka, 2001; Dimberg &Öhman, 1996). Examples of data on skin con-ductance responses from a single cue condi-tioning paradigm are given in Figure 44.1.

THE ROLE OF UNCONSCIOUSPROCESSES IN ANXIETY AND FEAR

This section of the chapter focuses on the psy-chological mechanisms underlying the adaptiveinput–output relationships for fear and anxietythat have been discussed in previous sections. Amodel is presented that emphasizes the role ofautomatic, unconscious processes in the gener-ation of fear and anxiety.

A Functional Perspectiveon Fear and Anxiety

As I have pointed out earlier (Öhman, 1993)defense responses are of little use unless theyare appropriately elicited. Thus they requireperceptual systems that can effectively locate

threat. Clearly, false negatives (i.e., failing toelicit defense to a potentially hazardous stimu-lus) are more evolutionarily costly than falsepositives (i.e., eliciting the response to a stimu-lus that in effect is harmless). In an evolution-ary perspective, therefore, it is likely that per-ceptual systems are biased toward discoveringthreat. Indeed, this provides an evolutionaryreason for why there are anxiety disorders. Toguarantee effective defense when life is at stake,the system is biased to “play it safe” by some-times activating defense to what on closer ex-amination turn out to be nondangerous con-texts. Responses of the latter type may seemunnecessary and unreasonable, and may be un-derstood as “irrational anxiety” by both ob-servers and the responder.

Effective defense must be quickly activated.Consequently, there is a premium for early de-tection of threat. Furthermore, threat stimulimust be detected wherever they occur in theperceptual field, independently of the currentdirection of attention. Coupled with the biastoward false positives, these factors imply thatdiscovery of threat is better based on a quick,superficial analysis of potentially dangerousstimuli (wherever they are) than on an


FIGURE 44.1. Extinction of skin conductance re-sponses (SCRs) in subjects who were conditionedto potentially phobic stimuli (pictures of snakes;left panel) or neutral stimuli (pictures of houses;right panel) by receiving them paired with electricshock (unconditioned stimulus). Control subjectshad the pictures and the shocks unpaired (“Sensi-tization”) or were only exposed to pictures (“CSalone”). It is obvious that potentially phobic con-ditioned stimuli resulted in much larger resistanceto extinction than did neutral conditioned stimuli.Data from Öhman, Eriksson, and Olofsson(1975).

effortful, detailed, and complete extractionof the meaning of one particular stimulus(LeDoux, 1996; Öhman, 1986). Therefore, thefunctional, evolutionary perspective suggeststhat the burden of threat discovery should beplaced on early, parallel-processing perceptualmechanisms, which define threat on the basisof relatively simple stimulus features.

Automatic Information-Processing Routinesto Discover Threat

My colleagues and I (Öhman, 1986, 1993;Öhman et al., 1985; Öhman, Flykt, &Lundqvist, 2000; Öhman & Wiens, 2004) havedeveloped a theoretical perspective on the gen-eration of emotion that is consistent with thefunctional scenario discussed above. Origi-nating in a model of orienting response activa-tion (Öhman, 1979), the perspective rests on adistinction between “automatic” and “con-trolled” or “strategic” information processing(e.g., Posner, 1978; Schneider, Dumais, &Shiffrin, 1984) to argue that many perceptualchannels can be automatically and simulta-neously monitored for potential threat. Whenstimulus events implying threat are located bythe automatic system, attention is drawn to thestimulus, as the control for its further analysisis transferred to the strategic level of informa-tion processing.

The switch of control from automatic tostrategic information processing is associatedwith activation of physiological responses—particularly the orienting response as indexedby, for example, skin conductance responsesand heart rate deceleration (Öhman, 1979;Öhman, Hamm, & Hugdahl, 2000). The auto-matic system is involuntary, in the sense that itis hard to suppress consciously once it is initi-ated; it does not interfere with focal attention;it is not easily distracted by attended activities;and it is typically not available for consciousintrospection (Schneider et al., 1984). Con-trolled or strategic information processing, onthe other hand, is governed by intentions. It isresource- or capacity-limited, in the sense thatinterference is marked between strategicallycontrolled tasks; it works sequentially ratherthan in parallel; it requires effort; and it is morereadily available to consciousness (Schneider etal., 1984).

This conceptualization suggests that the au-tomatic sensory monitoring processes have acapacity for sensory events exceeding that of

the controlled or strategic processes. Thus theycan keep track of many channels, only one ofwhich can be selected for strategic processing.Sensory messages have to compete for access tothe strategic processing channel for completesensory analysis. Given the survival contingen-cies implied by potential threats in the externaland internal environment, it is a natural as-sumption that events implying some degree ofthreat should have selection priority for strate-gic processing. This theoretical analysis sug-gests that anxiety and fear are activated as cor-relates to recruitment of defense responses afterquick, unconscious analyses of stimuli. As a re-sult of these analyses, threatening stimuli areselected for further conscious, controlled pro-cessing. Because they are located by automaticperceptual mechanisms, the person is not nec-essarily aware of the eliciting stimuli, whichmay result in episodes of anxiety. Indeed, whatappear from the inside to be “spontaneous” ep-isodes of anxiety may in fact be the results ofunconscious stimulation. These theoreticalprinciples are illustrated in the schematicmodel shown in Figure 44.2.

Experimental Test of the Model:Unconscious Activation of Phobias

Backward Masking

The central theoretical tenet of this model is thatresponses of anxiety or fear can be elicited afteronly a preliminary, unconscious analysis of thestimulus. Its empirical examination, therefore,requires a means of presenting fear stimuli out-side of the subject’s awareness. Such a means isprovided by backward masking, which has beenheld to be the potentially most fruitful avenue tounconscious perception (Holender, 1986;Öhman, 1999). Thus if backwardly masked fearstimuli were presented to fearful subjects, andthese stimuli still elicited physiological re-sponses suggesting activation of fear/anxietyeven though conscious recognition could beruled out, the theoretical notions advanced herewould receive experimental support.

Unconscious Phobic Responses

We (Öhman & Soares, 1994) selected researchparticipants who feared snakes but not spiders,or vice versa. Participants in the control groupfeared neither stimulus. In the experiment, par-ticipants were exposed to series of pictures of


snakes and spiders, with pictures of flowersand mushrooms serving as controls, while skinconductance responses were measured. In thefirst series, presentations were effectivelymasked by similar pictures that were grosslysimilar to the target stimuli in colors and tex-ture, but lacked any recognizable central ob-ject. A pilot experiment using a forced-choicerecognition procedure ascertained that bothfearful and nonfearful participants consistentlyfailed to identify the target with the maskingparameters used. The masks interrupted pre-sentation of the target stimuli after 30 millisec-onds of exposure and remained on for 100 mil-liseconds during the masked presentationseries. In the following series of presentations,the stimuli were presented unmasked. Afterthese series, the participants rated the stimulifor arousal, valence, and control/dominanceduring separately presented masked andnonmasked rating series. Figure 44.3 showsskin conductance responses to masked (a) andnonmasked (b) presentations of the stimuli. Itis evident that the fearful participants re-sponded specifically to their feared stimulus,but did not differ from controls for the otherstimulus categories, independently of masking.This enhanced responding to the feared stimu-lus cannot be attributed to conscious percep-tion. Nevertheless, parallel data were obtainedfor all three rating dimensions, which suggests

that some aspect of the masked stimulation be-came indirectly available to the conscioussystem (maybe through bodily feedback). Thusthe fearful participants rated themselves asmore disliking, more aroused, and less in con-trol when exposed to masked presentations oftheir feared stimulus.

Conditioning of Unconscious Effects

The data presented in Figure 44.3 support thenotion of fear as elicitable after only prelimi-nary, preattentive, automatic, and unconsciousanalyses of the stimulus. Consistent with a pre-pared Pavlovian conditioning perspective onphobias, we (Öhman & Soares, 1993; Soares& Öhman, 1993a, 1993b) demonstrated thatPavlovian conditioning to unmasked presenta-tions of fear-relevant stimuli (snakes and spi-ders) in nonfearful individuals resulted in con-ditioned skin conductance responses thatsurvived backward masking. Conditioning tofear-irrelevant stimuli (flowers and mush-rooms), on the other hand, resulted in moreelusive responses that were abolished by mask-ing.

Similar data were obtained for another classof fear-relevant stimuli, angry faces, whichwere masked by neutral faces (Esteves,Dimberg, & Öhman, 1994; Parra, Esteves,Flykt, & Öhman, 1997). Furthermore, we


FIGURE 44.2. A schematic version of the model of emotion activation presented in the text. Externalstimuli (1) make contact with “feature detectors,” detecting potentially significant stimuli on the basis ofsignal features. If such features are present, information is passed on (2) to the “significance evaluator”and (3) to the “arousal system,” which initiates a surge of autonomic activity to prepare action. Feed-back is then provided (4) to tune the significance evaluator for threat evaluation, which is also biased (5)by an “expectancy system,” which relies on the organization of emotion in memory. These inputs areevaluated (still unconsciously) by the significance evaluator, which then can call on (6) a system for “con-scious perception of threat” for an elaborated analysis of the stimulus in interaction with the expectancysystem (7) and the arousal system (8). See Öhman (2000) for a more detailed presentation of the model.

(Esteves, Dimberg, Parra, & Öhman, 1994) re-ported that skin conductance responses couldbe conditioned to masked fear-relevant, but notto masked fear-irrelevant, stimuli. That is, afterconditioning to masked angry or happy faces,subjects showed enhanced responding to subse-quent nonmasked presentations of angry butnot happy faces. Similarly, we (Öhman &Soares, 1998) demonstrated unconscious con-ditioning to masked snakes or spiders, but notto masked flowers or mushrooms. Thus notonly could emotional responses be elicited tomasked stimuli, but they could also be learnedto such stimuli, provided that they wereevolutionarily fear-relevant. We (Öhman &Mineka, 2001) concluded that these types ofpreattentive effects are best interpreted in termsof the ability of biologically fear-relevant stim-uli to access directly a basic, noncognitive levelof conditioning.

Fear-Relevant Stimuli as Guides to Attention

The model suggests that fear-relevant stimuli re-cruit attention by providing a switch from auto-matic to controlled processing of the stimulus.Thus automatic perceptual processes may lo-cate emotionally relevant stimuli at apreattentive level, and then call on controlledprocessing resources in order to deal with thesituation. Examining this hypothesis, we(Öhman, Flykt, & Esteves, 2001) demonstratedfast detection of evolutionarily fear-relevantstimuli (pictures of snakes and spiders) in a com-plex array of stimuli (pictures of flowers andmushrooms) by research participants selected tobe nonfearful or specifically fearful of snakes orspiders. The research used a visual search para-digm, in which the participants were instructedto detect a discrepant target stimulus among dif-ferent exemplars of a background category of


FIGURE 44.3. Upper panels (a and b) show skin conductance responses from snake-fearful, spider-fear-ful, and nonfearful control participants to effectively masked (a) and nonmasked (b) presentations of pic-tures of snakes, spiders, flowers, and mushrooms. Fearful subjects showed elevated responding to theirfeared stimulus even if it was prevented from entering conscious perception by backward masking (a).Lower panels (c and d) show spontaneous skin conductance responses (SCRs) in the intervals betweenstimulation. Whereas controls did not change from rest during stimulation, the fearful subjects showedenhanced spontaneous responding, suggesting that they became anxious after both masked (c) andnonmasked (d) presentations of feared stimuli. Data from Öhman and Soares (1994).

stimuli. For example, the target could be a snakepicture presented among distractor pictures offlowers. All participants were faster to locatesnakes and spiders among flowers and mush-rooms than vice versa; this advantage for fear-relevant targets, in contrast to the detection ofneutral targets, was independent of the numberof distractors in the display, which suggests thatthe detection of snakes and spiders had apreattentive origin (see, e.g., Eastwood et al.,2005). The basic finding that threatening ani-mals are more quickly detected than plants incomplex displays has been replicated in severallaboratories (even though the interpretations ofthese effects may differ; see Brosch & Sharma,2005; Flykt, 2004; Lipp, 2006; Lipp et al.,2004; Lipp, Derakshan, Waters, & Logies,2004). Similarly, visual search studies that haveexamined detection of schematic facial targetstimuli showing different emotions amongdistractor faces consistently report faster detec-tion of threatening (angry) than friendly(happy) targets (Eastwood, Smilek, & Merikle,2001; Fox et al., 2000; Mather & Knight, 2006;Öhman, Lundqvist, & Esteves, 2001; Tipples,Atkinson, & Young, 2002). This effect is obvi-ous with both neutral and emotionaldistractors, and it is specific for threateningrather than negative faces (Öhman, Lundqvist,& Esteves, 2001). Furthermore, it is closely re-lated to emotional ratings of the stimuli, and thethreat advantage is primarily conveyed by theupper part of the face (Lundqvist & Öhman,2005; Tipples et al., 2002).

Snake-fearful participants were as quick asthe controls in detecting spiders, but they werestill faster in detecting snakes, and vice versafor spider-fearful participants (Öhman, Flykt,& Esteves, 2001). Thus people in general ap-pear to be sensitive to evolutionarily relevantfear stimuli, and this sensitivity seems to be fur-ther enhanced in individuals for whom thestimuli actually elicit fear.

Rinck, Reinecke, Ellwart, Heuer, and Becker(2005) confirmed that spider-phobic partici-pants were quicker to detect spiders than neu-tral stimuli (beetles and butterflies) amongneutral distractors (dragonflies), whereas non-anxious controls were slower than the phobicindividuals for spiders but not for beetles andbutterflies. In addition, the phobic participantswere slower to detect dragonflies among spi-ders than among beetles and butterflies, whichsuggests that phobic stimuli in fact were moredifficult to ignore as distractors than were neu-

tral pictures when the primary task was to findneutral targets. Similarly, examining partici-pants that were high and low in social anxiety,we (Juth, Lundqvist, Karlsson, & Öhman,2005) found that those high in social anxietydetected schematic threatening targets more ac-curately than friendly targets, particularlywhen social anxiety was elevated by a criticalobserver in the room, and when the distractorswere emotional (i.e., happy for angry targets,and angry for happy targets).

Miltner, Kriechel, Hecht, Trippe, and Weiss(2004) instructed spider-fearful and nonfearfulparticipants to search for a neutral (mush-room) target among neutral distractors (flow-ers). On occasional trials, however, a spiderwas presented among the flower distractors.Even though the voluntarily controlled atten-tion was occupied in searching for the desig-nated mushroom targets, the occasional spideramong the distractor captured the attention ofthe spider-fearful but not the control partici-pants, which resulted in prolonged search timesfor the target mushroom. Taken together, thesestudies show that fear stimuli can be automati-cally detected and guide attention, both underconditions of top-down attentional control(Öhman, Flykt, & Esteves, 2001; Rinck et al.,2005), and in a stimulus-driven mode whenthey are unexpectedly presented outside theattentional focus (Miltner et al., 2004). In addi-tion, if stimuli are actively feared (e.g., spidersin spider phobia), they also interfere with thedetection of neutral targets when serving as aclass of distractors to be ignored in a top-downmode (Rinck et al., 2005). Whereas the formertwo effects suggest that feared stimuli are effec-tive in shifting and engaging attention, the lat-ter effect suggests, in addition, that it is moredifficult to disengage attention from an activelyfeared stimulus (see Posner & Peterson, 1990).

Processing Biases in Anxiety

There is a large literature (see reviews byMathews & MacLeod, 1994; Mineka, Rafaeli,& Yovel, 2003; Mogg & Bradley, 1998) on cog-nitive functions in anxiety, which documentsthat trait anxiety is associated with anattentional bias to focus on threatening infor-mation in the surroundings. As Mathews (1990)put it, “anxiety and worry are associated withan automatic processing bias, initiated prior toawareness, but serving to attract attention to en-vironmental threat cues, and thus facilitating


the acquisition of threatening information”(p. 462). This view is consistent with the func-tional perspective on fear and anxiety presentedabove. Accordingly, participants selected forhigh trait anxiety or for fulfilling diagnostic cri-teria for an anxiety disorder (particularly GAD)preferentially direct their attention towardthreatening stimuli in different experimentaltasks (for reviews, see Mathews & MacLeod,1994; Mogg & Bradley, 1998).

MacLeod, Mathews, and Tata (1986) devel-oped a dot probe detection procedure in whichparticipants were presented with two brieflydisplayed words, one shown above and theother below a fixation cross that forewarnedthe words. When the words disappeared, a dotappeared at the previous location of one of thewords. On some critical trials one of the wordswas threat-related and the other neutral.MacLeod et al. (1986) reported that patientswith GAD were faster to respond to dots ap-pearing at the location previously occupied bya threat word than at the location occupied bya neutral word, which implied that the pa-tients’ attention was preferentially directed to-ward the threatening words.

The dot probe paradigm can be modified foruse with pictorial threatening and neutral stim-uli. Following presentation of a central fixationcross, Bradley et al. (1997) briefly presentedtwo pictures of faces, one showing an emo-tional and the other a neutral expression, sideby side, one at each side of the cross. When thefaces disappeared, a dot probe appeared at oneof the locations of the previously presentedfaces, and the participants were instructed toindicate its position (left or right) on the screen.As assessed by response latencies to the probes,the results showed a bias away from angry(compared to happy and neutral) faces innonanxious participants, and a bias for lookingat the angry faces among participants selectedfor high trait anxiety (Bradley, Mogg, Falla, &Hamilton, 1998). Furthermore, recordings ofeye movements showed that individuals withGAD more frequently and rapidly looked firstat threatening rather than neutral faces (Mogg,Millar, & Bradley, 2000).

The dot probe paradigm (with words—e.g.,Mogg, Bradley, & Williams, 1995; or withthreatening and neutral faces—Mogg &Bradley, 1999) can be used with masked stim-uli. Using a dot probe procedure with pictures,Mogg and Bradley (2002) presented briefly (for17 milliseconds) pairs of pictures of the same

actor displaying an emotional (threatening orhappy) and a neutral facial expression, andthen this stimulus pair was immediatelymasked by another pair of pictures of scram-bled faces (for 68 milliseconds). The resultsshowed that anxious individuals (particularlythose with social anxiety) preferentially ori-ented their attention to the location of themasked threatening faces, thus supporting apreattentive origin for the threat bias.

Anxiety and Difficulties in Disengaging Attentionfrom Threat

Taken together, the data reviewed in this sec-tion are quite consistent in showing not onlythat fear-relevant stimuli often guide attentionamong nonselected participants, but also thatthis effect is potentiated by high anxiety. Fur-thermore, these biases for threat are likely tohave preattentive origins. However, there arealternative interpretations of the nature of thereported attentional bias in the dot probe para-digm. If we assume (for the sake of the argu-ment) that there is no bias of shifting attentionto the threatening stimulus in the dot probeparadigm, participants are initially as likely tofocus on the threatening as on the neutral stim-ulus. However, when they happen to focus onthe threatening stimulus, they may have diffi-culties in disengaging attention, which willslow their detection latency when they have toshift attention to a dot appearing at the loca-tion of the neutral stimulus. Because they donot get stuck in this way on neutral stimuli,there is no interference with moving the atten-tion to dots appearing behind the threateningstimuli. Thus one would generally expectslower response latencies to neutral than tothreatening stimuli—not because of fast detec-tion of the latter, but because of slow detectionof neutral stimuli due to slow disengagementfrom threatening stimuli.

These interpretational hazards led Fox,Russo, Bowles, and Dutton (2001) to developparadigms for stringent testing of the hypothesisthat the effect of anxiety on threat detection isprimarily due to slow disengagement of atten-tion from threatening stimuli. They used a para-digm adopted from Posner, Inhoff, Friedrich,and Cohen (1987). A trial was initiated by theappearance of a fixation cross in the center of ascreen between two laterally placed boxes. Withthe cross still on, a neutral or threatening cue (aword or a face, depending on the experiment)


was briefly (for 100 milliseconds) presented inone of two boxes. The cue disappeared and wasfollowed by a dot probe in the same box in themajority of (e.g., 75%) the trials (valid cueing)or a dot in the other box on the remaining trials(invalid cueing). Invalid cueing carries a cost interms of slowed reaction time, because attentionhas to be disengaged from the validly cued loca-tion before it can be moved to the other location(Posner et al., 1987). Across four experiments,Fox et al. (2001) reported consistently longer re-action times to invalidly than validly cued dots.This effect was enhanced if the cue was a threat-ening word or face (compared to positive wordsor happy faces)—but only for highly anxiousparticipants. These findings suggest that high-anxiety individuals were slow to shift the atten-tion to a new location once they had engagedtheir attention in a threatening stimulus. Thusthe attention of people with high anxiety may getstuck on threatening events around them, whichmay interfere with flexible deployments of at-tention according to current performance de-mands. Subsequent research from Fox’s grouphas confirmed this effect (Fox, Russo, & Dutton,2002) and demonstrated that it is specific tothreatening (angry, fearful) rather than to nega-tive (sad) faces (Georgiou et al., 2005).

For people preferentially attending to threat-ening events and objects in the environment,the world will appear a dangerous and riskyplace, which is likely to influence their moodstate negatively. An attentional bias for threat,therefore, may be an important factor main-taining anxiety over time in a vicious circle(where anxiety promotes a bias for threat, theresult of which is more anxiety promoting fur-ther threat bias, etc.). Furthermore, people whohave difficulties disengaging attention fromthreatening information will be preoccupied bythreat. If the threatening event that holds atten-tion in this way is mental (a memory, thought,or image), this is tantamount to getting stuckon ruminative worry. Thus the central symp-tom of GAD may result from problems withthe control of attention.

TOWARD A NEUROBIOLOGYOF FEAR AND ANXIETY

The Amygdala and Fear

LeDoux (1996) has described the neuro-architecture of a system that accounts for manyof the behavioral findings reviewed above. Heand his coworkers have demonstrated a direct

neural link from auditory nuclei (medialgeniculate body) in the thalamus to the “signif-icance evaluator” and “fear effector system” inthe lateral and central amygdala, respectively.This monosynaptic link provides immediate in-formation to the amygdala of gross features ofemotionally relevant auditory stimuli. It by-passes the traditionally emphasized thalamo-cortical sensory pathway, which gives fullmeaning to the stimulus, and the cortico-amygdala link, which is presumed to activateemotion. It is described as a “quick and dirty”transmission route: It “probably does not tellthe amygdala much about the stimulus, cer-tainly not much about Gestalt or object proper-ties of the stimulus, but it at least informs theamygdala that the sensory receptors of a givenmodality have been activated and that a signifi-cant stimulus may be present” (LeDoux, 1990,p. 172), so that the amygdala can start early ac-tivation of defense responses. This system is ex-plicitly postulated to be adaptively biased to-ward false positives rather than false negatives.This is because it is less costly to abort falselyinitialized defense responses than to fail toelicit defense when the threat is real.

Tests of the Amygdala Model in Humans

LeDoux’s model is primarily based on data re-sulting from experiments manipulating the ro-dent brain. However, in recent years, data onbrain imaging in humans supporting the modelhave started to appear.

Activation of the Human Amygdalaby Fear Stimuli

The most basic implication of LeDoux’s model,of course, is that fear stimuli of the type thatI have been discussing actually engage theamygdala. This assertion is supported by arange of studies (see, e.g., Phelps, 2006). Forexample, my colleagues and I (Carlsson et al.,2004) reported positron emission tomographydata from participants who, like those in anearlier study (Öhman & Soares, 1994; see Fig-ure 44.3), were selected to be afraid of spidersbut not snakes, and vice versa. They showed bi-lateral activation of the amygdala to phobicpictures (e.g., snakes for snake-phobic par-ticipants) as compared to fear-relevant butnonfeared pictures (e.g., spiders), as well as re-liable activations of the midbrain periaque-ductal gray, the insula, and the anteriorcingulate cortex. Very similar data were re-


ported by Straube, Mentzel, and Miltner(2006) and Sabatinelli, Bradley, Fitzsimmons,and Lang (2005), using functional magneticresonance imaging and spider-phobic partici-pants. Amygdala activations, however, wereconspicuously absent in earlier brain imagingstudies of individuals with specific phobiaschallenged by their phobic objects in videoclips or in vivo (e.g., Fredrikson et al., 1993;Fredrikson, Wik, Annas, Ericson, & Stone-Elander, 1995; Rauch et al., 1995). These fail-ures can probably be attributed to the use oflengthy stimulus presentations rather than re-peated short stimulations, because theamygdala is likely to respond best to the onsetof an effective fear stimulus (e.g., Carlsson etal., 2004). In addition, many studies have re-ported larger responses to emotional (e.g., fear-ful) than to neutral faces (e.g., Anderson,Christoff, Panitz, De Rosa, & Gabrieli, 2003).

Amygdala Activation by UnconsciouslyPresented Stimuli

Because of its emphasis on the subcortical in-put to the amygdala, the LeDoux model im-plies that it should be possible to activate theamygdala by stimuli that are not consciouslyperceived. This prediction was supported in theCarlsson et al. (2004) study, which reported re-liable amygdala activation (compared to neu-tral pictures) both for masked phobic pictures(e.g., snakes for snake phobics) and for maskedfear-relevant but nonphobic (e.g., spiders) pic-tures. Furthermore, we (Morris, Öhman, &Dolan, 1998) demonstrated specific activationof the amygdala by masked conditioned angryfaces in human subjects, and Whalen et al.(1998) demonstrated significant amygdala acti-vation to masked fearful faces.

Binocular rivalry is an alternative method tobackward masking for presenting stimuli out-side of awareness. If two different stimuli areseparately projected on corresponding retinallocations in the two eyes, rather than appearingas a mixture, they will compete to determinethe percept. Given that the two stimuli areroughly matched in salience, the percept willspontaneously shift between them. However, ifone stimulus is in some sense more salient thanthe other, the latter will be suppressed, whichimplies that it is presented outside of the partic-ipant’s awareness. Two studies have reportedreliable amygdala activation to suppressed pic-tures of fearful faces in a binocular rivalryparadigm (Pasley, Mayes, & Schultz, 2004;

Williams, Morris, McGlone, Abbott, &Mattingley, 2004), thus providing convergingevidence to the effect that the amygdala can beactivated by nonconsciously perceived threatstimuli.

Amygdala Activation by Nonattended Fear Stimuli

The notion of a direct subcortical link to theamygdala implies that it should be possible toactivate the amygdala by nonattended stimuli.Vuilleumier, Armony, Driver, and Dolan (2001)examined this notion by exposing their partici-pants to pairs of pictures of faces and housesthat were either horizontally or vertically ar-ranged around a fixation cross. Attention wasmanipulated by requiring the participants tojudge whether the pairs of faces or housesshowed identical pictures on different trials. Incontrast to the fusiform facial area, whose re-sponse to fearful faces was enhanced whenfaces were attended to, amygdala activationwas larger to fearful than to neutral faces re-gardless of whether faces or houses were at-tended to. Thus these data support the hypoth-esis that the amygdala can be activated bynonattended stimuli.

This conclusion was further supported byAnderson et al. (2003) in a similar paradigm, inwhich participants attended to either faces(showing fear, disgust, or a neutral expression)or houses in different colors that were superim-posed on each other. In support of the hypothe-sis that the amygdala activation is independentof attention to the eliciting stimulus, these au-thors reported that the amygdala response tofearful faces did not differ as a function ofwhether the faces were attended or not.

Finally, Straube et al. (2006) required theirspider-phobic participants either to identifywhether pictures showed a spider or a mush-room, or to perform a perceptual judgmenttask on a circle superimposed on spiders ormushrooms. Amygdala responses to spiders(relative to mushrooms) were larger in phobicthan in nonphobic participants. Furthermore,in contrast to the responses in other structures(e.g., insula, anterior cingulate, dorsomedialprefrontal cortex), in which the responses tospiders were smaller during the perceptualtask, the amygdala responses were as large dur-ing this task as when participants focused onspiders and mushroom pictures in the pictureidentification task. These results again supportindependence of the amygdala to attentionalconditions.


Amygdala Facilitation of Visual Perception

The amygdala is richly connected with the neo-cortex. However, these connections are biased,in the sense that the amygdala projects to manymore cortical areas than it receives input from(Emery & Amaral, 2000; LeDoux, 1996). Withregard to the visual system, it receives inputfrom the primary visual cortex (V1) and theinferotemporal cortex (IT); however, it has ef-ferent pathways to all the major way stations inthe ventral visual processing stream, as well asthe lateral and medial orbito-frontal cortices(Emery & Amaral, 2000). This anatomical or-ganization might suggest that the amygdalatunes perceptual processing in order to fa-cilitate perception of potentially threateningevents.

To test this idea, Phelps, Ling, and Carrasco(2006) performed an experiment in which theybriefly cued a perceptual task (judging line ori-entations as a function of contrast betweenlines and background) by fearful or neutralfaces. Confirming that emotion facilitates per-ception of low-level visual features, their datashowed improved contrast sensitivity (i.e.,more accurate judgment of line orientation atlower spatial contrasts between line and back-ground) after fearful as opposed to neutralfaces. This effect, however, was not obtainedwith inverted faces, which provides support forthe emotional (amygdala?) origin of the effect.Furthermore, in a second experiment, Phelps etal. (2006) showed that this enhanced contrastsensitivity was stronger for emotional than forneutral upright (but not inverted) faces in bothfocused- and divided-attention conditions, sug-gesting that it did not depend on attention.However, the emotion effect was larger withfocused than with divided attention, which im-plies that the effect of emotion was potentiatedby attention.

Vuilleumier, Richardson, Armony, Driver,and Dolan (2004), in a study of patients withfocal amygdala and hippocampal lesions, pro-vided data indicating that these types of per-ceptual effects may be mediated by theamygdala. They replicated previously demon-strated emotional enhancement of an early cor-tical area involved in vision, the fusiform cor-tex, to fearful faces (e.g., Vuilleumier et al.,2001) in normal controls and patients with fo-cal hippocampal lesions. However, such emo-tional enhancement was not observed in thepatients with focused amygdala lesions (withor without hippocampal lesions). Thus these

findings imply that the amygdala is rapidly ac-tivated in order to prime early visual processingof emotional stimuli, as suggested by LeDoux(1996) and as behaviorally demonstrated byPhelps et al. (2006).

Amygdala Activation in the Absence of RelevantSensory Cortex

The notion of a direct route from sensory or-gans to the amygdala implies that it should bepossible to activate the amygdala even if corti-cal processing of the fear stimulus is impossiblebecause of lesions in the relevant sensory cor-tices. To examine this prediction, Morris,deGelder, Weiskrantz, and Dolan (2001) exam-ined a patient with lesions in V1, resulting in“blindsight”—a condition in which a persondenies any visual experience from stimulationsin the blind spot, yet is able to locate and some-times to discriminate stimuli in the blind field.When the patient with blindsight examined byMorris et al. (2001) was exposed to faces in theblind field, he showed reliable activation of theright amygdala to fearful as compared to neu-tral faces. Furthermore, similar results were ob-tained in a patient whose extensive occipital le-sion was associated with total blindnesswithout any signs of residual vision (Pegna,Khateb, Lazeyras, & Seqhier, 2005).

Is There a Subcortical Pathway to the Amygdala?

To examine the route to nonconsciousamygdala activation, we (Morris, Öhman, &Dolan, 1999) used our earlier data (Morris etal., 1998) to examine the neural connectivitybetween the amygdala and other brain regionswhen the amygdala was activated by maskedstimuli. In support of the hypothesis (LeDoux,1996), we reported that activation of the rightamygdala by masked stimuli could be reliablypredicted from activation of subcortical waystations in the visual systems (such as the supe-rior colliculus and the right pulvinar nucleus ofthe thalamus), but not from any cortical re-gions.

Liddell et al. (2005) examined the effect ofmasked fearful versus masked neutral faces onanatomically defined regions of interest. Con-firming our connectivity data (Morris et al.,1999) they found reliable activation to maskedfearful faces in the left superior colliculi, theleft pulvinar, and bilateral amygdalae. In addi-tion, they found activation in the locuscoeruleus and the anterior cingulate.


In an ingenious elaboration of the low-roadconcept, Vuilleumier, Armony, Driver, andDolan (2003) suggested that the amygdala op-erates primarily on gross, low-frequency infor-mation, because it is served by large, rapidlyconducting neurons. Accordingly, they filteredthe spatial frequency of pictures of faces to pro-duce facial stimuli that retained only high- orlow-frequency spatial information. Their re-sults showed that amygdala responses werelarger for low-frequency faces, provided thatthey showed expressions of fear. Moreover,they demonstrated activation of the pulvinarand superior colliculus by low-frequency butnot high-frequency fearful faces. Thus these re-sults suggest that there is a distinct superiorcolliculus–pulvinar pathway to the amygdalathat operates primarily on low-frequency infor-mation.

Finally, in a powerful new hypothesis of theevolutionary origin of the visualization of theprimate brain, Isbell (2006) argued that thekoniocellular pathways that connect the supe-rior colliculi and the pulvinar with the lateralgeniculate and the visual cortex (V2 area) wasdriven by the evolutionary functionality ofthese pathways in detecting venomous, poten-tially deadly snakes hidden on the ground andin foliage.

The low-road hypothesis has been contro-versial, and it has influential challengers (e.g.,Pessoa, 2005; Pessoa, McKenna, Gutierrez, &Ungerleider, 2002; Rolls, 1999). A recurrenttheme has been that input from IT is necessaryfor the evidence suggesting discriminatorybehavior by the amygdala to facial emotion.However, Pasley et al. (2004), who reportedamygdala activation to fearful faces suppressedfrom awareness by binocular rivalry, demon-strated that areas of the IT that responded tofaces under normal viewing did not discrimi-nate faces from objects under conditions ofbinocular rivalry. If we assume that the binocu-lar rivalry suppressed the V1, these datastrongly suggest a subcortical origin of the vi-sual input that allowed the amygdala to dis-criminate suppressed fearful from suppressedneutral faces.

In summary, the data reviewed here providesubstantial support for LeDoux’s (1996) con-ception of dual input to the amygdala, thussupporting a neural underpinning for thebehavioral data from masking and attentionparadigms.

Toward a Psychobiological Distinctionbetween Fear and Anxiety

A distinction between fear and anxiety that isbroadly consistent with Epstein’s (1972) ideas,but couched in terms of neuroanatomical struc-ture, has recently been proposed.

The Neuroanatomy of Fear and Anxiety: Relatedbut Distinct

Davis and his colleagues (Davis, 1998; Davis &Shi, 1999; Walker, Toufexis, & Davis, 2003)propose that both fear and anxiety are relatedto the extended amygdala complex, whichreaches from the medial temporal lobe to theventral striatum. The central nucleus of theamygdala mediates acute fear responses to spe-cific stimuli, whereas the anatomically closelyrelated bed nucleus of the stria terminalis medi-ates anxiety (i.e., more long-lasting, tonic re-sponses to aversive contexts without a neces-sary, discrete eliciting stimulus). Both thesestructures receive input from the basolateralamygdala complex. Because they both have ef-ferent connections to the same set of anatomi-cal targets in the hypothalamus and brainstemthat control and coordinate behavioral signs offear and anxiety (autonomic responses, defen-sive reflexes, etc.), this distinction retains thelargely overlapping response output of fear andanxiety. Nonetheless, Davis’s group has shownthat the involved structures can also be dissoci-ated.

At the behavioral level, Davis and colleagues(see, e.g., Davis, 2000, for references) use thestartle reflex as the measure of fear. “Startle” isa defensive reflex elicited by any abrupt, unex-pected, and relatively intense stimulus that“makes us jump.” In its complete form it in-volves the whole body, with a defensive for-ward thrust of the head and trunk, and with aspreading wave of flexor movements along theneural axis. In rats it is measured as whole-body startle, and in humans it is commonlymeasured by one of its components, the eye-blinking response. Movie directors demon-strate their insight into startle reflex dynamicsby making the audience jump to an abrupt cut(often coinciding with a sudden noise) that cul-minates a fright-inducing sequence. This is anexample of “fear-potentiated startle.” In thepsychological laboratory, fear-potentiated star-tle denotes evidence of an enhanced startle re-flex to auditory probe stimuli presented againsta fear-inducing as compared to a neutral back-


ground stimulus (Lang, Bradley, & Cuthbert,1990). This effect is mediated by the centralnucleus of the amygdala (Davis, 2000; Lang,Davis, & Öhman, 2000).

The secretion of corticotropin-releasinghormone (CRH), which triggers the variousbodily changes (via the pituitary) that arecommonly used as indices of stress and anxi-ety, is activated by cells in the hypothalamus,but it also acts as a neurotransmitter in thebrain. Infusion of CRH in the brains of ratsresults in strong and relatively long-lastingstartle potentiation, which is mediated not bythe amygdala but by dense CRH receptors inthe bed nucleus of the stria terminalis (Davis& Shi, 1999; Walker et al., 2003). Thus thereis a double-dissociation effect: Lesion of thecentral nucleus of the amygdala blocks fear-potentiated startle by conditioned fear stim-uli, but has no effect on the CRH-enhancedstartle; lesion of the bed nucleus of the striaterminalis, on the other hand, blocks CRH-enhanced startle but has no effect on condi-tioned startle potentiation. Furthermore, theacute fear response (“freezing”) in rats to un-controllable foot shock is abolished byamygdala lesion, but the more long-term es-cape deficit to aversive stimuli that resultsfrom uncontrollable shock (“learned helpless-ness”) is dependent on an intact bed nucleusof the stria terminalis (see review by Walkeret al., 2003).

To conclude, therefore, even though fear andanxiety are overlapping responses, they can bedistinguished in terms of stimuli (presence vs.absence of a discrete eliciting stimulus), behav-ior (coping vs. noncoping), and neuroanatomy(central nucleus of the amygdala vs. bed nu-cleus of the stria terminalis).

The Behavioral Geneticsof Fear- and Anxiety-Related Disorders

A distinction between primarily fear-relateddisorders such as the specific phobias on theone hand, and primarily anxiety-related disor-ders (PD, GAD, PTSD) and depression on theother, is emerging in the behavioral genetics lit-erature. These investigations use data on thecomorbidity of DSM anxiety disorders in largesamples of mono- and dizygotic twins to modelvariance in these disorders that can be attrib-uted to genes versus shared and unique envi-ronmental influences. The results suggest that amodel incorporating two additive genetic fac-

tors (as well as shared and unique environmen-tal factors) provides a good fit to data on sixDSM anxiety disorders in large samples ofmale and female twins (Hettema, Prescott,Myers, Neale, & Kendler, 2005). One geneticfactor showed substantial loadings for thestudied specific phobias (animal, situational),but low loadings for social phobia, GAD, PD,and agoraphobia.

The other genetic factor had high loadingsfor GAD, PD, and agoraphobia; some loadingfor social phobia; and virtually zero loadingsfor the two diagnoses of specific phobias.Furthermore, a twin study on Vietnam vet-erans demonstrated considerable geneticoverlap among PTSD, GAD, and PD(Chantarujikapong et al., 2001). Indeed, thebehavioral genetics literature is consistent in at-tributing specific phobias on the one hand, andGAD, PD, and sometimes depression on theother, to distinctly different genetic factors(Kendler, Prescott, Myers, & Neale, 2003).Our analysis would then imply that this secondfactor is an anxiety factor, which potentially isrelated to the bed nucleus of the striaterminalis. These conclusions are schematicallyillustrated in Figure 44.4.

Psychophysiological Data from Patientswith DSM Anxiety Disorders

The distinction between specific phobias on theone hand, and PD, GAD, and PTSD on theother, is vindicated by psychophysiologicaldata. Individuals diagnosed with specific pho-bias, and particularly animal phobias, show adistinct psychophysiological response when ex-posed to their feared stimuli. For example, we(Globisch, Hamm, Esteves, & Öhman, 1999)exposed student volunteer participants whowere specifically fearful (if not phobic) ofsnakes or spiders to pictures of snakes or spi-ders, household objects (neutral), and eroticaand cute animals (positive). The fearful partici-pants rated snake/spider pictures as much moreunpleasant and arousing than neutral and posi-tive pictures, compared to controls. Further-more, they showed clearly larger skin conduc-tance responses to pictures of their fearedanimals than controls did, and their responsesto these animals were larger than their re-sponses to neutral or positive pictures. As ex-pected from previous data (e.g., Hamm,Cuthbert, Globisch, & Vaitl, 1997; Hare &Blevings, 1975), fearful participants showed a


strong heart rate acceleration to snake/spiderpictures, which was in marked contrast to thedeceleration shown by controls to these ani-mals, and to the small and indistinguishable re-sponse shown by both groups to neutral stim-uli. The startle data showed a rapid relativestartle potentiation to feared compared to neu-tral stimuli, which was obvious to probes pre-sented as early as 300 milliseconds after pictureonset, and then remained (and even increased)for at least 4 seconds of picture viewing. Con-trols, on the other hand, showed a rapid rela-tive startle inhibition to animal stimuli thatlasted for more than a second. In concert, thesedata show that enhanced (phobic) fear to ani-mal stimuli is associated with a distinctpsychophysiological response suggesting acti-vation of fight–flight behavior, which clearlydiffers from the pattern elicited by these stimuliin nonfearful participants. This distinctstimulus-elicited fear response in specific pho-bias contrasts with the psychophysiological re-sponse to stress seen in PD and GAD. In spiteof the clinical description (confirmed by self-report data) of panic as involving a surge ofsympathetic activation much like that in fearresponses (Öhman, 1993), research shows thatPD and GAD are associated primarily withhigh resting autonomic activity, and, if any-

thing, less acute reactivity to common stresstasks (Hoehn-Saric & MacLeod, 2000).

Cuthbert et al. (2003) examined psycho-physiological responses across the whole spec-trum of DSM anxiety disorders in patients en-gaged in imagery of different fear scenes (Lang,1979). When cued, the patients were instructedto imagine fear scenes based on previouslylearned sentences briefly describing personalfear experiences (referring to “the worst fearyou ever experienced”), standard danger scenes(“Taking a shower, alone in the house, I hearthe sound of someone forcing the door, and Ipanic”), social fear scenes (“My heart poundsin the suddenly silent room; everyone is watch-ing me . . . ”) or neutral control scenes (“Softmusic is playing on the stereo as I snooze lazilyon my favorite chair”).

In spite of small differences between patientgroups in ratings of fear and vividness of imag-ery, heart rate increases when imagining thefear scenes were larger in the participants withspecific or social phobia than in the partici-pants with PTSD, PD, or no disorder, who didnot differ among each other. The group withspecific phobia showed considerably morestartle potentiation than the other diagnosticgroups when imagining not only their own per-sonal fear scene, but also the other fear scenes.


FIGURE 44.4. The genetic structure of fear-related and anxiety-related disorders (based on behavioralgenetics data reported by Hettema et al., 2005, and Chantarujikapong et al., 2001). There is one set ofgenes (“Fear”) influencing specific phobias (animal and situational phobias), and another (“Anxiety”)influencing panic disorder and agoraphobia, posttraumatic stress disorder (PTSD), generalized anxietydisorder (GAD), and depression. Social phobia shows relations to both clusters. These clusters may be re-lated to different neural structures (central nucleus of the amygdala vs. bed nucleus of the striaterminalis), differences in psychophysiological reactivity (enhanced responses to fear stimuli vs. elevatedresting levels and blunted stress responses), and different attention biases (shift of attention to threat vs.failure to disengage from threat).

These results are consistent with previousdata showing larger heart rate responses duringfear imagery for individuals with specific pho-bias than for those with social phobia (Cook,Melamed, Cuthbert, McNeil, & Lang, 1988;McNeil, Vrana, Melamed, Cuthbert, & Lang,1993), who both showed larger responses thanthose with agoraphobia (Cook et al., 1988).Similarly, Lang, McTeague, and Cuthbert(2005) observed larger fear-potentiated startleto fear imagery in patients with specific or so-cial phobia than in patients with PD or GAD.

Keane et al. (1998) examined psychophysio-logical responses to a standard stress task(mental arithmetic), a standardized series ofcombat pictures and sounds, and personalizedmental imagery scripts of experienced traumas(using the Lang [1979] methodology) in morethan 1,200 Vietnam veterans (see also the re-view by Orr & Roth, 2000). Veterans with acurrent PTSD diagnosis showed larger heartrate, skin conductance, and systolic and dia-stolic blood pressure responses to both stan-dard combat and imagined personal traumascenes than veterans never diagnosed withPTSD. Veterans with a lifetime but not a cur-rent PTSD diagnosis in general were intermedi-ate to the other two groups. In addition, thegroup with a current diagnosis showed higherresting heart rate and skin conductance, butlower heart rate and diastolic blood pressureresponses to standard stress, than the other twogroups.

Like the patients with PTSD examined byKeane et al. (1998), the patients with PD exam-ined by Cuthbert et al. (2003) showed elevatedheart rate between imagery trials. These dataconform to those reported by Hoehn-Saric andMacLeod (2000) showing elevated base levelsof heart rate and skin conductance, but, if any-thing, less reactivity to common stress tasks, inpatients with PD and GAD.

There are also interesting differences be-tween patients with specific phobias on the onehand, and those with PD, GAD, or PTSD onthe other, in self-reported anxiety and comor-bidity for other DSM anxiety disorders and de-pression. Whereas those with specific phobiasdid not differ from controls in self-reportedanxiety, patients diagnosed with social phobia,PTSD, PD, or GAD showed elevated levels ofself-reported anxiety and depression (Cuthbertet al., 2003; Lang et al., 2005). Likewise, inagreement with other data (Brown, Campbell,Lehman, Grisham, & Mancill, 2001), a major-

ity of patients diagnosed with PTSD, PD, orGAD had comorbidity for an additional DSManxiety or mood disorder, whereas this was lessfrequent among patients with social phobiaand particularly infrequent among patients di-agnosed with specific phobias (Cuthbert et al.,2003; Lang et al., 2005).

Overall, the data reviewed in this sectionshow that specific phobia is associated with adistinct fear response, as indexed by startlepotentiation and strong heart rate accelerationwhen patients are exposed to their feared ob-jects. In addition, they show little evidence ofelevated general anxiety levels, and only mod-est comorbidity with other DSM anxiety andmood disorders. Patients with social phobialikewise show enhanced startle potentiationand accelerative heart rate responses whenimagining fear scenes, but they also show someelevation in anxiety scores (Cuthbert et al.,2003; Lang et al., 2005), compared to controlsand patients with specific phobias. Patientswith PTSD appear to show both patterns, withenhanced psychophysiological responses totrauma-related scenes and elevated restingheart rate and skin conductance levels (Orr &Roth, 2000); they also exhibit enhanced self-reported anxiety and depression, and consider-able comorbidity for DSM anxiety and depres-sive disorders. These data converge nicely withthe previously reviewed neuroanatomical andbehavioral genetics results, suggesting that sim-ilar distinctions between fear and anxiety areemerging in all three of these research fields.

CONCLUDING DISCUSSION

To recapitulate, I have started with the obviouscommunality between fear and anxiety, andthen launched a discussion of differences withrespect to coping with an implied threat (Ep-stein, 1972). In fear, organisms try to cope withthe danger, but in anxiety, the situation doesnot allow effective means of coping. The theo-retical perspective I have presented stresses therole of rapid early information-processingmechanisms in fear activation. Here, both fearand anxiety have a joint origin in an uncon-scious mobilization to an as yet poorly definedthreat. With more time to appraise the situa-tion, the emotion can be resolved into fearwhen reflexive (or highly preferred) coping op-tions are available, or can turn into anxietywhen they are not.


A cognitive-psychobiological model hasbeen presented, which is broadly consistentwith behavioral data on the effect of maskingon responses to fear stimuli and on the powerof fear or anxiety stimuli to guide attention.This model has an empirically well-foundedneural underpinning. However, in contrast tomy previous position (Öhman, 1993) converg-ing data now appear to support a model withfundamental differences between fear and anx-iety in terms of genetics, neuroanatomy, andpsychophysiological reactivity.

Behavioral genetic studies suggest that spe-cific phobias on the one hand, and GAD, PD,and PTSD on the other, are related to differentgenetic factors (Hettema et al., 2005). This in-formation converges with the neuroanatomicaldistinction between the central nucleus of theamygdala as related to fear and the bed nucleusof the stria terminalis as related to anxiety(Walker et al., 2003). Furthermore, there aredistinct psychophysiological responses for spe-cific phobias (Cuthbert et al., 2003). There is asuggestion in the data on attention, fear, pho-bias, and anxiety that fear stimuli promote re-flexive shifts in focal attention. The attentionalbias of anxious individuals (and particularlypatients diagnosed with GAD), on the otherhand, primarily concerns difficulties of disen-gaging attention from threatening stimuli oncethey have become the focus of attention (Fox etal., 2001). Social phobia appears intermediateto the two genetic factors (Hettema et al.,2005), as well as to the two patterns ofpsychophysiological reactivity (Lang et al.,2005). Finally, PTSD belongs genetically to theanxiety factor, but psychophysiologically it in-volves both enhanced reactivity to feared stim-uli and elevated basal levels of psychophysio-logical activity (Keane et al., 1998).

Thus it is reasonable to talk about specificphobias as exclusively fear-related disorders,PD and perhaps particularly GAD as primaryanxiety-related disorders, and social phobiaand PTSD as mixed fear- and anxiety-relateddisorders. This conclusion is further supportedby data on comorbidity, which show only mod-est comorbidity with other DSM anxiety disor-ders and depression for specific phobias; somesuch comorbidity for social phobia; and con-siderable such comorbidity for PTSD, PD, andGAD (Brown et al., 2001).

As a final note, it is important to rememberthat the modeled genetic factors are indepen-dent and additive. This means that principles of

fear may be operating in other DSM anxietydisorders besides phobias. For example, fearconditioning is an important factor producingmany of the symptoms of PTSD (e.g., Öhman& Mineka, 2001), but factors shared with PDand GAD are also important for understandingother PTSD symptoms, such as the elevatedphysiological levels suggesting a generalizedstress impact. Similarly, the neuroendocrine re-sponse patterns that may predispose for PD,GAD, and PTSD also seems to facilitate fearconditioning (LeDoux, 1996). What makesspecific phobias unique is that they tend tohave one central problem, which is best con-ceptualized in terms of fear. For the other DSManxiety disorders, fear may be one of the com-ponents; however, it is then embedded in amuch wider context of emotional agitation,which is best conceptualized in terms of anxi-ety.

REFERENCES

American Psychiatric Association (APA). (2000). Diag-nostic and statistical manual of mental disorders (4thed., text rev.). Washington, DC: Author.

Anderson, A. K., Christoff, K., Panitz, D., De Rosa, E.,& Gabrieli, J. D. E. (2003). Neural correlates of theautomatic processing of threat facial signals. Journalof Neuroscience, 23, 5627–5633.

Arrindell, W. A., Pickersgill, M. J., Merckelbach, H.,Ardon, M. A., & Cornet, F. C. (1991). Phobic dimen-sions: III. Factor analytic approaches to the study ofcommon phobic fears: An updated review of findingsobtained with adult subjects. Advances in BehaviourResearch and Therapy, 13, 73–130.

Barlow, D. H. (2002). Anxiety and its disorders: The na-ture and treatment of anxiety and panic (2nd ed.).New York: Guilford Press.

Bourke, J. (2006). Fear: A cultural history. Emeryville,CA: Shoemaker & Hoard.

Bradley, B. P., Mogg, K., Falla, S. J., & Hamilton, L. R.(1998). Attentional bias for threatening facial expres-sions in anxiety: Effect of stimulus duration. Cogni-tion and Emotion, 12, 737–753.

Bradley, B. P., Mogg, K., Millar, N., Bonham-Carter, C.,Fergusson, E., Jenkins, J., et al. (1997). Attentionalbiases for emotional faces. Cognition and Emotion,11, 25–42.

Brosch, T., & Sharma, D. (2005). The role of fear-relevant stimuli in visual search: A comparison ofphylogenetic and ontogenetic stimuli. Emotion, 5,360–364.

Brown, T. A., Campbell, L. A., Lehman, C. L., Grisham,J. R., & Mancill, R. B. (2001). Current and lifetimecomorbidity of the DSM-IV anxiety and mood disor-ders in a large clinical sample. Journal of AbnormalPsychology, 110, 585–599.


Carlsson, K., Petersson, K. M., Lundqvist, D., Karlsson,A., Ingvar, M., & Öhman, A. (2004). Fear and theamygdala: Manipulation of awareness generates dif-ferential cerebral responses to phobic and fear-relevant (but non-feared) stimuli. Emotion, 4, 340–353.

Chantarujikapong, S. I., Scherrer, J. F., Xian, H., Eisen,S. A., Lyons, M. J., Goldberg, J., et al. (2001). A twinstudy of generalized anxiety disorder symptoms,panic disorder symptoms, and post-traumatic stressdisorder in men. Psychiatry Research, 103, 133–145.

Cook, E. W., III, Melamed, B. G., Cuthbert, B. N.,McNeil, D. W., & Lang, P. J. (1988). Emotional im-agery and the differential diagnosis of anxiety. Jour-nal of Consulting and Clinical Psychology, 56, 734–740.

Cuthbert, B. N., Lang, P. J., Strauss, C., Drobes, D., Pat-rick, C. J., & Bradley, M. M. (2003). The psycho-physiology of anxiety disorder: Fear memory imag-ery. Psychophysiology, 40, 407–422.

Davis, M. (1998). Are different parts of the extendedamygdala involved in fear versus anxiety? BiologicalPsychiatry, 44, 1239–1247.

Davis, M. (2000). The role of the amygdala in condi-tioned and unconditioned fear and anxiety. In J. P.Aggleton (Ed.), The amygdala (2nd ed., pp. 213–288). New York: Oxford University Press.

Davis, M., & Shi, C. (1999). The extended amygdala:Are the central nucleus of the amygdala and the bednucleus of the stria terminalis differentially involvedin fear versus anxiety? Annals of the New York Acad-emy of Sciences, 877, 281–291.

Dimberg, U., & Öhman, A. (1996). Behold the wrath:Psychophysiological responses to facial stimuli. Mo-tivation and Emotion, 20, 149–182.

Eastwood, J. D., Smilek, D., & Merikle, P. M. (2001).Differential attentional guidance by unattended facesexpressing positive and negative emotion. Perceptionand Psychophysics, 63, 1004–1013.

Eastwood, J. D., Smilek, D., Oakman, J. M., Farvolden,P., van Ameringen, M., Mancini, C., et al. (2005). In-dividuals with social phobia are biased to becomeaware of negative faces. Visual Cognition, 12, 159–179.

Emery, N. J., & Amaral, D. G. (2000). The role of theamygdala in primate social cognition. In R. D. Lane& L. Nadel (Eds.), Cognitive neuroscience of emo-tion (pp. 156–191). New York Oxford UniversityPress.

Epstein, S. (1972). The nature of anxiety with emphasisupon its relationship to expectancy. In C. D.Spielberger (Ed.), Anxiety: Current trends in theoryand research (Vol. 2, pp. 291–337). New York: Aca-demic Press.

Esteves, F., Dimberg, U., & Öhman, A. (1994). Auto-matically elicited fear: Conditioned skin conductanceresponses to masked facial expressions. Cognitionand Emotion, 8, 393–413.

Esteves, F., Dimberg, U., Parra, C., & Öhman, A.(1994). Nonconscious associative learning: Pavlov-

ian conditioning of skin conductance responses tomasked fear-relevant facial stimuli. Psychophysi-ology, 31, 375–385.

Flykt, A. (2004). Visual search with biological threatstimuli: Accuracy, reaction times, and heart ratechanges. Emotion, 5, 349–353.

Foa, E. B., & McNally, R. J. (1986). Sensitivity to fearedstimuli in obsessive–compulsives: A dichotic listeninganalysis. Cognitive Therapy and Research, 10, 477–486.

Fox, E., Lester, V., Russo, R., Bowles, R. J., Pichler, A.,& Dutton, K. (2000). Facial expressions of emotion:Are angry faces detected more efficiently? Cognitionand Emotion, 14, 61–92.

Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Dothreatening stimuli draw or hold visual attention insubclinical anxiety? Journal of Experimental Psy-chology: General, 130, 681–700.

Fox, E., Russo, R., & Dutton, K. (2002). Attentionalbias for threat: Evidence for delayed disengagementform emotional faces. Cognition and Emotion, 16,355–379.

Fredrikson, M., Wik, G., Annas, P., Ericson, K., &Stone-Elander, S. (1995). Functional neuroanatomyof visually elicited simple phobic fear: Additionaldata and theoretical analysis. Psychophysiology, 32,43–48.

Fredrikson, M., Wik, G., Greitz, T., Eriksson, L., Stone-Elander, S., Ericson, K., & Sedvall, G. (1993). Re-gional cerebral blood flow during experimental fear.Psychophysiology, 30, 126–130.

Georgiou, G. A., Bleakley, C., Hayward, J., Russo, R.,Dutton, K., Eltiti, S., et al. (2005). Focusing on fear:Attentional disengagement from emotional faces. Vi-sual Cognition, 12, 145–158.

Globisch, J., Hamm, A. O., Esteves, F., & Öhman, A.(1999). Fear appears fast: Temporal course of startlereflex potentiation in animal fearful subjects.Psychophysiology, 36, 66–75.

Hamm, A. O., Cuthbert, B. N., Globisch, J., & Vaitl, D.(1997). Fear and the startle reflex: Blink modulationand autonomic response patterns in animal and muti-lation fearful subjects. Psychophysiology, 34, 97–107.

Hare, R. D., & Blevings, G. (1975). Defensive re-sponses to phobic stimuli. Biological Psychology, 3,1–13.

Hettema, J. M., Prescott, C. A., Myers, J. M., Neale, M.C., & Kendler, K. S. (2005). The structure of geneticand environmental risk factors for anxiety disordersin men and women. Archives of General Psychiatry,62, 182–189.

Hoehn-Saric, R., & McLeod, D. R. (2000). Anxiety andarousal: Physiological changes and their perception.Journal of Affective Disorders, 61, 217–224.

Holender, D. (1986). Semantic activation without con-scious identification in dichotic listening, parafovealvision, and visual masking: A survey and appraisal.Behavioral and Brain Sciences, 9, 1–66.

Isbell, L. A. (2006). Snakes as agents of evolutionary


change in primate brains. Journal of Human Evolu-tion, 51, 1–35.

Juth, P., Lundqvist, D., Karlsson, A., & Öhman, A.(2005). Looking for foes and friends: Perceptual andemotional factors when finding a face in the crowd.Emotion, 5, 379–395.

Keane, T. M., Kolb, L. C., Kaloupek, D. G., Orr, S. P.,Blanchard, E. B., Thomas, R. G., et al. (1998). Utilityof psychophysiological measurement in the diagnosisof posttraumatic stress disorder: Results from the De-partment of Veterans Affairs Cooperative Study.Journal of Consulting and Clinical Psychology, 66,914–923.

Kendler, K. S., Prescott, C. A., Myers, J., & Neale, M. C.(2003). The structure of genetic and environmentalrisk factors for common psychiatric and substanceuse disorders in men and women. Archives of Gen-eral Psychiatry, 60, 929–937.

Lader, M., & Marks, I. (1973). Clinical anxiety. Lon-don: Heinemann.

Lang, P. J. (1979). A bio-informational theory of emo-tional imagery. Psychophysiology, 16, 495–512


Lang, P. J., Davis, M., & Öhman, A. (2000). Fear andanxiety: Animal models and human cognitivepsychophysiology. Journal of Affective Disorders,61, 137–159.

Lang, P. J., McTeague, L. M., & Cuthbert, B. N. (2005).Fear startle, and the anxiety disorder spectrum. In B.Rothbaum (Ed.) Pathological anxiety: Emotionalprocessing in etiology and treatment (pp. 56–77).New York: Guilford Press.

LeDoux, J. E. (1990). Fear pathways in the brain: Impli-cations for a theory of the emotional brain. In P. F.Brain, S. Parmigiani, R. J. Blanchard, & D. Mainardi(Eds.), Fear and defence (pp. 163–177). London:Harwood.


Liddell, B. J., Brown, K. J., Kemp, A. H., Barton, M. J.,Das, P., Peduto, A., et al. (2005). A direct brainstem–amygdala–cortical “alarm” system for subliminalsignals of fear. NeuroImage, 24, 235–243.

Lipp, O. V. (2006). Of snakes and flowers: Does prefer-ential detection of pictures of fear-relevant animals invisual search reflect on fear-relevance? Emotion, 6,296–308.

Lipp, O. V., Derakshan, N., Waters, A., & Logies, S.(2004). Snakes and cats in the flower bed: Fast detec-tion is not specific to pictures of fear-relevant stimuli.Emotion, 4, 233–250.

Lundqvist, D., & Öhman, A. (2005). Emotion regulatesattention: The relation between facial configuration,facial emotion, and visual attention. Visual Cogni-tion, 12, 51–84.

MacLeod, C., Mathews, A., & Tata, P. (1986).Attentional bias in emotional disorders. Journal ofAbnormal Psychology, 95, 15–20.

MacLeod, C., & Rutherford, E. M. (1992). Anxiety andthe selective processing of emotional information:Mediating roles of awareness, trait and state vari-ables, and personal relevance of stimulus materials.Behaviour Research and Therapy, 30, 479–491.

Mather, M., & Knight, M. R. (2006). Angry faces getnoticed quickly: Threat detection is not impairedamong older adults. Journal of Gerontology: B. Psy-chological Science, 61, 54–57.

Mathews, A. (1990). Why worry?: The cognitive func-tion of anxiety. Behaviour Research and Therapy, 28,455–468.

Mathews, A., & MacLeod, C. (1994). Cognitive ap-proaches to emotion and emotional disorders. An-nual Review of Psychology, 45, 25–50.

McNally, R. J. (1987). Preparedness and phobias: A re-view. Psychological Bulletin, 101, 283–303.

McNeil, D. W., Vrana, S. R., Melamed, B. G., Cuthbert,B. N., & Lang, P. J. (1993). Emotional imagery insimple and social phobia: Fear versus anxiety. Jour-nal of Abnormal Psychology, 102, 212–225.

Miltner, W. H. R., Kriechel, S., Hecht, H., Trippe, R., &Weiss, T. (2004). Eye movements and behavioral re-sponses to threatening and nonthreatening stimuliduring visual search in phobic and nonphobic sub-jects. Emotion, 4, 323–339.

Mineka, S., Rafaeli, E., & Yovel, I. (2003). Cognitive bi-ases in emotional disorders: Information processingand social-cognitive perspectives. In R. J. Davidson,K. R. Scherer, & H. H. Goldsmith (Eds.). Handbookof affective sciences (pp. 976–1009). New York: Ox-ford University Press.

Mogg, K., & Bradley, B. P. (1998). A cognitive–motivational analysis of anxiety. Behaviour Researchand Therapy, 36, 809–848.

Mogg, K., & Bradley, B. P. (1999). Orienting attentionto threatening facial expressions presented underconditions of restricted awareness. Cognition andEmotion, 13, 713–740.

Mogg, K., & Bradley, B. P. (2002). Selective orienting ofattention to masked threat faces in social anxiety. Be-haviour Research and Therapy, 40, 1403–1414.

Mogg, K., Bradley, B. P., & Williams, R. (1995).Attentional bias in anxiety and depression: The roleof awareness. British Journal of Clinical Psychology,34, 17–36.

Mogg, K., Millar, N., & Bradley, B. P. (2000). Biasesin eye movement to threatening facial expression ingeneralized anxiety disorder and depressive disor-der. Journal of Abnormal Psychology, 109, 695–704.

Morris, J. S., deGelder, B., Weiskrantz, L., & Dolan, R.J. (2001). Differential extrageniculostriate andamygdala responses to presentation of emotionalfaces in a cortically blind field. Brain, 124, 1241–1252.

Morris, J. S., Öhman, A., & Dolan, R. J. (1998). Con-scious and unconscious emotional learning in the hu-man amygdala. Nature, 393, 467–470.

Morris, J. S., Öhman, A., & Dolan, R. J. (1999). A


subcortical pathway to the right amygdala mediating“unseen” fear. Proceedings of the National Academyof Sciences USA, 96, 1680–1685.

Öhman, A. (1979). The orienting response, attention,and learning: An information processing perspective.In H. D. Kimmel, E. H. van Olst, & J. F. Orlebeke(Eds.), The orienting reflex in humans (pp. 443–472).Hillsdale, NJ: Erlbaum.

Öhman, A. (1986). Face the beast and fear the face: Ani-mal and social fears as prototypes for evolutionaryanalyses of emotion. Psychophysiology, 23, 123–145.

Öhman, A. (1993). Fear and anxiety as emotional phe-nomena: Clinical phenomenology, evolutionary per-spectives, and information processing mechanisms.In M. Lewis & J. M. Haviland (Eds.) Handbook ofemotions (pp. 511–536). New York: Guilford Press.

Öhman, A. (1999). Distinguishing unconscious fromconscious emotional processes: Methodological con-siderations and theoretical implications. In T.Dalgleish & M. Power (Eds.), Handbook of cognitionand emotion (pp. 321–352). Chichester, UK: Wiley.

Öhman, A. (2000). Fear and anxiety: Evolutionary, cog-nitive, and clinical perspectives. In M. Lewis & J. M.Haviland-Jones (Eds.), Handbook of emotions (2nded., pp. 573–593). New York: Guilford Press.

Öhman, A., Dimberg, U., & Öst, L.-G. (1985). Animaland social phobias: Biological constraints on learnedfear responses. In S. Reiss & R. R. Bootzin (Eds.),Theoretical issues in behavior therapy (pp. 123–175).New York: Academic Press.

Öhman, A., Eriksson, A., & Olofsson, C. (1975). Onetrial learning and superior resistance to extinction ofautonomic responses conditioned to potentially pho-bic stimuli. Journal of Comparative and Physiologi-cal Psychology, 88, 619–627.

Öhman, A., Flykt, A., & Esteves, F. (2001). Emotiondrives attention: Detecting the snake in the grass.Journal of Experimental Psychology: General, 130,466–478.

Öhman, A., Flykt, A., & Lundqvist, D. (2000). Uncon-scious emotion: Evolutionary perspectives, psycho-physiological data, and neuropsychological mecha-nisms. In R. Lane & L. Nadel (Eds.), The cognitiveneuroscience of emotion (pp. 296–327). New York:Oxford University Press.

Öhman, A., Hamm, A. O., & Hugdahl, K. (2000). Cog-nition and the autonomic nervous system: Orienting,anticipation, and conditioning. In J. T. Cacioppo, L.G. Tassinary, & G. G. Berntson (Eds.), Handbook ofpsychophysiology (pp. 533–575). New York: Cam-bridge University Press.

Öhman, A., Lundqvist, D., & Esteves, F. (2001). Theface in the crowd revisited: An anger superiority ef-fect with schematic stimuli. Journal of Personalityand Social Psychology, 80, 381–396.


Öhman, A., & Mineka, S. (2003). The malicious ser-pent: Snakes as a prototypical stimulus for anevolved module of fear. Current Directions in Psy-chological Science, 12, 2–9.

Öhman, A., & Soares, J. J. F. (1993). On theautomaticity of phobic fear: Conditioned skin con-ductance responses to masked fear-relevant stimuli.Journal of Abnormal Psychology, 102, 121–132.

Öhman, A., & Soares, J. J. F. (1994). “Unconscious anx-iety”: Phobic responses to masked stimuli. Journal ofAbnormal Psychology, 103, 231–240.

Öhman, A., & Soares, J. J. F. (1998). Emotional condi-tioning to masked stimuli: Expectancies for aversiveoutcomes following nonrecognized fear-relevantstimuli. Journal of Experimental Psychology: Gen-eral, 127, 69–82.

Öhman, A., & Wiens, S. (2004). The concept of anevolved fear module and cognitive theories of anxi-ety. In A. S. R. Manstead, N. Frijda, & A. Fischer(Eds.), Feelings and emotions: The Amsterdam Sym-posium (pp. 58–80). Cambridge, UK: CambridgeUniversity Press.

Orr, S. P., & Roth, W. T. (2000). Psychophysiological as-sessment: Clinical applications for PTSD. Journal ofAffective Disorders, 61, 225–240.

Parra, C., Esteves, F., Flykt, A., & Öhman, A. (1997).Pavlovian conditioning to social stimuli: Backwardmasking and the dissociation of implicit and explicitcognitive processes. European Psychologist, 2, 106–117.

Pasley, B. N., Mayes, L. C., & Schultz, R. T. (2004).Subcortical discrimination of unperceived objectsduring binocular rivalry. Neuron, 42, 163–172.

Pegna, A., Khateb, A., Lazeyras, F., & Seqhier, M.(2005). Discriminating emotional faces without pri-mary visual cortices involves the right amygdala, Na-ture Neuroscience, 8, 24–25.

Pessoa, L. (2005). To what extent are emotional visualstimuli processed without attention and awareness?Current Opinion in Neurobiology, 15, 188–196.

Pessoa, L., McKenna, M., Gutierrez, E., & Ungerleider,L. G. (2002). Neural processing of emotional facesrequires attention. Proceedings of the National Acad-emy of Sciences of the USA, 99, 11458–11463.

Phelps, E. A. (2006). Emotion and cognition: Insightsfrom studies of the human amygdala. Annual Reviewof Psychology, 57, 27–53.

Phelps, E. A., Ling, S., & Carasco, M. (2006). Emotionfacilitates perception and potentiates the perceptualbenefits of attention. Psychological Science, 17, 292–299.

Posner, M. I. (1978). Chronometric explorations ofmind. Hillsdale, NJ: Erlbaum.

Posner, M. I., Inhoff, A. W., Friedrich, F. J., & Cohen, A.(1987). Isolating attentional systems: A cognitive-anatomical analysis. Psychobiology, 15, 107–121.

Posner, M. I., & Peterson, S. E. (1990). The attentionsystem of the human brain. Annual Review of Neu-roscience, 13, 25–42.


Rapee, R. M. (1991). Generalized anxiety disorder: Areview of clinical features and theoretical concepts.Clinical Psychology Review, 11, 419–440.

Rauch, S. L., Savage, C. R., Nathaniel, M. A., Miguel,E., Baer, L., Breiter, H. C., et al. (1995). A positronemission tomographic study of simple phobic symp-tom provocation. Archives of General Psychiatry, 52,20–28.

Rinck, M., Reinecke, A., Ellwart, T., Heuer, K., &Becker, E. S. (2005). Speeded detection and increaseddistraction in fear of spiders: Evidence from eyemovements. Journal of Abnormal Psychology, 114,235–248.

Rolls, E. T. (1999). The brain and emotion. Oxford,UK: Oxford University Press.

Sabatinelli, D., Bradley, M. M., Fitzsimmons, J. R., &Lang, P. J. (2005). Parallel amygdala and infero-temporal activation reflect emotional intensity andfear relevance. NeuroImage, 24, 1265–1270.

Schneider, W., Dumais, S. T., & Shiffrin, R. M.(1984). Automatic and control processing and at-tention. In R. Parasuraman & D. R. Davies (Eds.),Varieties of attention (pp. 1–27). Orlando, FL: Aca-demic Press.

Seligman, M. E. P. (1971). Phobias and preparedness.Behavior Therapy, 2, 307–320.

Soares, J. J. F., & Öhman, A. (1993a). Backward mask-ing and skin conductance responses after condition-ing to non-feared but fear-relevant stimuli in fearfulsubjects. Psychophysiology, 30, 460–466.

Soares, J. J. F., & Öhman, A. (1993b). Preattentive pro-cessing, preparedness, and phobias: Effects of in-struction on conditioned electrodermal responses tomasked and non-masked fear-relevant stimuli. Be-haviour Research and Therapy, 31, 87–95.

Spielberger, C. D. (1972). Anxiety as an emotional state.In C. D. Spielberger (Ed.), Anxiety: Current trends in

theory and research (Vol. 1, pp. 23–49). New York:Academic Press.

Straube, T., Mentzel, H.-J., & Miltner, W. H. R. (2006).Neural mechanisms of automatic and direct process-ing of phobigenic stimuli in specific phobia. Biologi-cal Psychiatry, 59, 162–170.

Tipples, J., Atkinson, A. P., & Young, A. W. (2002). Theeyebrow frown: A salient social signal. Emotion, 2,288–296.

Tuma, A. H., & Maser, J. D. (Eds.). (1985). Anxiety andthe anxiety disorders. Hillsdale, NJ: Erlbaum.

Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J.(2001). Effects of attention and emotion on facesprocessing in the human brain: An event-relatedfMRI study. Neuron, 30, 829–841.

Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J.(2003). Distinct spatial frequency sensitivities forprocessing faces and emotional expressions. NatureNeuroscience, 6, 624–631.

Vuilleumier, P., Richardson, M. P., Armony, J. L., Driver,J., & Dolan, R. (2004). Distant influences ofamygdala lesion on visual cortical activation duringemotional face processing. Nature Neuroscience, 7,1271–1278.

Walker, D. L., Toufexis, D. J., & Davis, M. (2003). Roleof the bed nucleus of the stria terminalis versus theamygdala in fear, stress, and anxiety. European Jour-nal of Pharmacology, 463, 199–216.


Wilson, E. J., MacLeod, C., Mathews, A., &Rutherford, E. M. (2006). The causal role of inter-pretive bias in anxiety reactivity. Journal of Abnor-mal Psychology, 115, 103–111.


C H A P T E R 4 5

The Development of Angerand Hostile Interactions

ELIZABETH A. LEMERISE and KENNETH A. DODGE

Since the publication of the first and secondeditions of this volume (Lewis & Haviland,1993; Lewis & Haviland-Jones, 2000), re-search on emotions has grown to the pointwhere emotion is a major focus in the sixth edi-tion of the Handbook of Child Psychology(Eisenberg, 2006). Here we trace the normativedevelopment of anger in infants and children,with the thesis that anger develops and comesto be regulated in the context of transactionswith the social environment (Parke, 1994).Also, individual differences in abilities/capaci-ties and developmental changes contribute tothis transaction and shed light on importantprocesses (Bugenthal & Grusec, 2006). The lit-erature reviewed here includes studies on in-fants’ and children’s expression of anger; theirreactions to and appraisal of anger; their regu-lation of and/or coping with anger in them-selves and others; and their understanding ofanger, including display rules. We also considersocialization processes relevant to anger, andboth adaptive and maladaptive aspects of an-ger and angry interactions.

THE FUNCTIONAL SIGNIFICANCEOF ANGER

Although emotion theorists do not agree onwhether anger is a primary human emotion(Izard, 1991) or is differentiated from a gener-alized distress state (Camras, 1992), they doagree that anger serves a variety of adaptivefunctions. Anger organizes and regulates physi-ological and psychological processes related toself-defense and mastery and regulates socialand interpersonal behaviors (Izard & Kobak,1991; Lewis, Sullivan, Ramsay, & Alessandri,1992; Saarni, Campos, Camras, & Withering-ton, 2006). From a functionalist perspective onemotion, when there is an obstacle to goal at-tainment, anger’s function is to overcome ob-stacles in order to achieve goals (Saarni et al.,2006). Despite its adaptive significance, angerposes difficulties for social organisms in thatit repels others (Marsh, Ambady, & Kleck,2005), incurring long-term costs. Thus the reg-ulation and appropriate expression of angerare key developmental tasks. Individuals must

730

learn their culture’s “display rules,” which con-cern when, to whom, and how to express emo-tions in culturally acceptable ways. Indirect so-cialization of display rules can be observedquite early in infancy (Malatesta & Haviland,1982). Anger regulates interpersonal behaviorand comes to be regulated in an interpersonalcontext via socialization by caregivers, peers,and the larger social context (Oatley &Jenkins, 1996). Problems in regulating andappropriately expressing anger appropriatelyhave implications for social functioning(Eisenberg, Fabes, Nyman, Bernzweig, &Pineulas, 1994; Murphy, Shepard, Eisenberg,& Fabes, 2004), as well as for the developmentof psychopathology and disease (Barefoot,Dodge, Peterson, Dahlstrom, & Williams,1989; Casey & Schlosser, 1994; Cole, Teti, &Zahn-Waxler, 2003; Suinn, 2001).

THE DEVELOPMENTAL COURSEOF ANGER AND ITS REGULATIONAnger in Infants and Toddlers

Validity of Infant Anger

The question of whether young infants displaydiscrete negative emotions (including anger),or a more generalized distress reaction, isthe topic of an ongoing debate (Bennett,Bendersky, & Lewis, 2002, 2004; Camras,2004; Camras, Sullivan, & Michel, 1993;Izard, 2004; Izard et al., 1995; Oster, Hegley,& Nagel, 1992). In part, this debate reflects theinherent difficulties in studying infants. Al-though objective methods for coding infants’facial expressions of emotion are available(MAX and AFFEX; Izard, 1983; Izard,Dougherty, & Hembree, 1983), some research-ers have questioned whether infant facial ex-pressions reflect internal states (Saarni et al.,2006). Another source of difficulty lies in find-ing situations that reliably elicit particularemotions (Bennett et al., 2002, 2004; Camras,2004; Izard, 2004). Similar problems arisewhen considering data that young infants re-spond differentially to naturally presented,multimodal emotion displays, especially intheir mothers (e.g., Kahana-Kalman & Walker-Andrews, 2001; Montague & Walker-Andrews, 2001), because it is not clear whatcues are driving the responses or what the re-sponses mean regarding infants’ understandingof expressions (see Saarni et al., 2006). Unfor-tunately, this debate has had an inhibiting ef-

fect on research on infant anger. In the remain-der of this section, we summarize studies thatexamine anger in the first year. As a group,these studies used MAX/AFFEX—codingschemes that yield discrete codes for negativeemotions.

Early Development of Anger

In previous editions of this chapter (Lemerise& Dodge, 1993, 2000), we reviewed work byStenberg and Campos (1990) and Izard,Hembree, and Hueber (1987) suggesting thatthe first clear expressions of anger emerge atabout 4 months of age, and that anger expres-sions are targeted to a social figure by 7months. These expressions have been observedin response to arm restraint (Stenberg & Cam-pos, 1990) and inoculations (Izard et al.,1987). Field and colleagues (e.g., Field et al.,2005; Pickens & Field, 1993; see alsoReissland, Shepherd, & Herrera, 2005) haveobserved anger expressions in infants as youngas 3 months during face-to-face interactionswith mothers, particularly in infants of de-pressed mothers. Anger expressions have beenobserved in infants as young as 2 months in thecontext of a contingency-learning paradigm(Lewis, Alessandri, & Sullivan, 1990). In thisparadigm, infants learn that an arm-pulling re-sponse produces an interesting display. Goalblockage is introduced by removing the contin-gency between arm pulling and the display,thus fulfilling the functionalist definition of ananger-eliciting situation wherein the infant’sgoal can be specified with greater certaintythan in arm restraint, for example. At extinc-tion, anger (but not sadness) is associated witharm pulling, and the highest levels of joy, inter-est, and arm pulling occur when the contin-gency is reinstated (Lewis et al., 1990, 1992;Sullivan & Lewis, 2003; Sullivan, Lewis, &Alessandri, 1992). Lewis and Ramsay (2005)replicated these results in 4- to 6-month-oldswith contingency learning and a still-face pro-cedure; sadness, but not anger, was found to re-late to cortisol responses, supporting the hy-pothesis that sadness and anger are distinct inyoung infants.

A study by Buss and Goldsmith (1998) alsounderscores the adaptive significance of anger.In separate sessions, 6-, 12-, and 18-month-oldinfants were exposed to two anger-elicitingstimuli (a barrier problem and arm restraint)and two fear-eliciting stimuli, which elicited the

45. The Development of Anger 731

predicted emotions. Anger and fear were dis-tinct and not significantly related to one an-other, as might be expected if they were really“distress.” Infants used different regulatorystrategies for anger and fear; these regulatorybehaviors reduced their anger, but not theirfear, expressions. Buss and Goldsmith con-cluded that the emotion regulation literature’salmost exclusive focus on negative emotion asa global construct has obscured importantbetween-emotion differences in infants’ regula-tory strategies.

Factors Affecting the Development of Anger

The body of work reviewed suggests that angeris associated with infants’ attempts to masterthe physical environment, and that anger elicitsbehavioral strategies from infants that serveregulatory functions and contribute to problemsolving. An infant’s anger can also be a socialsignal, mobilizing a reaction from the caregiverthat assists the infant in modulating distress. Atthe same time, however, caregivers’ socializa-tion goals are to encourage positive and neutralemotions and to minimize negative emotions.Differential responding to positive and nega-tive emotions by caregivers is related to the in-creasing frequency of positive and neutral ex-pressions over time; also, mothers are morelikely to reinforce sadness than anger(Malatesta, Culver, Tesman, & Shepard, 1989;Malatesta & Haviland, 1982). Children asyoung as 24 months appear to be responsive tothese differential responses to negative emo-tions. In laboratory situations designed to elicitfear and anger, children did exhibit the targetedemotions, except when they looked at theirmothers; during these looks, sad expressionswere more likely (Buss & Kiel, 2004).

Campos and colleagues have suggested thatinfants’ evolving abilities alter transactionswith their environments (Campos et al., 2000;Campos, Kermoian, & Zumbahlen, 1992). Forexample, Campos et al. (1992) reported thatthe onset of locomotion is associated with par-ent report of anger in both infants and parents,as well as with the perception that the infantsare more autonomous and responsible.

Campos et al.’s (1992) work also suggeststhat seeing an infant as autonomous is linkedto a parent’s using direct socialization strategiesand being less permissive. The child’s increas-ing autonomy makes it more likely that theparent and child will find themselves at cross-purposes. In Goodenough’s (1931) classic

monograph, as children moved from infancy totoddlerhood, parents became less permissive ofanger and used a variety of strategies, such ascoaxing, diversion of attention, ignoring, andphysical restraint, to discourage it. With in-creasing age, ignoring continued to be used forgirls, whereas boys received more attention foranger in the form of power-assertive measuressuch as threats, spankings, and deprivation ofprivileges (Goodenough, 1931; see also Radke-Yarrow & Kochanska, 1990).

Although direct socialization methods areused more with older children, caregivers donot abandon indirect socialization methods.Denham (1993) found that mothers and their2-year-olds responded contingently to eachother’s emotion displays, and that mothersseemed to have agendas similar to those re-ported by Malatesta and Haviland (1982; i.e.,maximizing positive and neutral emotions,minimizing negative emotions). Mothers whoresponded to toddlers’ anger with calm neu-trality or cheerful displays had toddlers whoshowed interest in the environment, positiveemotions, and positive responses to strangersin their mothers’ absence. On the other hand,as Crockenberg (1985) demonstrated, angrymaternal responding to toddlers’ difficult be-haviors is associated with toddlers’ persistingin angry, noncompliant behavior and a lowerlikelihood of empathic responding to others.Thus socialization practices that act to modu-late toddlers’ negative arousal may make possi-ble more competent and empathic behavior,whereas socialization practices that exacerbatesystem arousal may hinder social development.

Another mechanism of indirect socializationinvolves exposure to interadult anger. Infantsas young as 3 months respond differentially tomothers’ natural (multimodal) displays ofemotion, including anger (e.g., Haviland &Lelwica, 1987), and 5-month-old infants re-spond differentially to positively and negativelytoned infant-directed vocalizations in theirown and unfamiliar languages (Fernald, 1993).According to maternal diaries of naturally oc-curring emotions, children as young as 12months may stare or “freeze,” look concernedor frown, or show distress by whimperingwhen witnessing others’ angry interactions.Slightly older children engage in social refer-encing (Walden & Ogan, 1988) of their moth-ers (Radke-Yarrow & Kochanska, 1990), in or-der to learn how to interpret social stimuli.

Beginning at 16–18 months, toddlers usemore direct strategies, such as covering their


ears, leaving the room, intervening verballyand/or physically (Cummings, Zahn-Waxler, &Radke-Yarrow, 1981; Radke-Yarrow &Kochanska, 1990), or behaving aggressively.Data from a laboratory study suggest that ex-posure to emotionally arousing situations cre-ates expectancies about adults’ interactions(Jenkins, Franco, Dolins, & Sewell, 1995).Jenkins et al. argued that interadult anger influ-ences children’s models of relationships, whichare used to interpret future interactions (seealso Dodge, Bates, & Pettit, 1990; Schwartz,Dodge, Pettit, & Bates, 1997).

Complicating the picture of the normativedevelopment of anger, there are individual dif-ferences in infants’ “anger-proneness,” consid-ered to be a component of temperament.Calkins, Dedman, Gill, Lomax, and Johnson(2002) compared 6-month-olds on “frustrationreactivity, . . . a pattern of negative affect char-acterized by anger and low tolerance for frus-tration” (p. 176). Based on a composite mea-sure derived from three laboratory frustrationtasks and maternal reports of distress, infantswere classified as easily frustrated versus noteasily frustrated. The easily frustrated infantsdisplayed greater physiological reactivity(which they were less able to regulate), poorerattention, and higher activity levels. Calkinsand colleagues suggested that the implicationsof greater frustration reactivity may depend onthe extent to which infants receive support intheir environments for managing their affect.Support for this hypothesis is found in longitu-dinal work by Kochanska and colleagues.Kochanska, Aksan, and Carlson (2005) foundthat maternal responsiveness and security of at-tachment with fathers promoted later receptivecooperation with parents in anger-prone 7-month-olds. Kochanska (2001) also found thatwhile securely attached children became lessangry from 14 to 33 months, insecurely at-tached children increased in negative emotionalexpressivity; disorganized attachment was as-sociated with large increases in anger and de-creases in fear (see also Lyons-Ruth, Alpern, &Repacholi, 1993; Shaw, Owens, Vondra,Keenan, & Winslow, 1996).

In summary, the emergence of anger in in-fants remains a topic of theoretical and meth-odological debate. An answer to this questionawaits the resolution of the thorny issue of howto measure infants’ emotions reliably and val-idly. However, the perception of infants as au-tonomous and responsible, coincident with the

emergence of locomotion, is associated withparental reports of infant anger. It may be thatlocomoting infants’ goals are clearer to adultobservers, and that the infants’ response togoal blockage is thus easier to interpret as an-ger. Research from a variety of investigatorssuggests that anger is an important energizer ofinfants’ adaptive behavior. However, the onsetof locomotion often places infants’ goals atodds with those of caregivers, increasing the in-cidence of anger between them. Parents be-come less permissive of angry outbursts, andsocialization of anger becomes more direct.Toddlers’ social-emotional competence hasbeen linked to maternal socialization practicesthat modulate strong negative emotions andencourage positive and neutral ones.

Anger and Its Regulationin Preschool-Age Children

Preschoolers continue to learn to coordinatetheir goals with those of others and are thus ex-pected to show increasing control of anger.Peers emerge as important new socializers ofanger, and language affords children a new wayto express emotions and gives parents a power-ful tool in socializing emotion. However, pre-schoolers still have relatively poor control overtheir emotion displays. An important devel-opmental task involves learning to managearousal in service of goals for play and affilia-tion with peers (Parker & Gottman, 1989).Emotion socialization processes in the familycontribute to social competence with peers andat preschool (e.g., Denham, 1998; Dunn &Brown, 1994; Gottman, Katz, & Hooven,1997).

The Expression and Socialization of Angerin the Family Context

Parents can indirectly influence children’sexpressivity through contingent emotion dis-plays during dyadic interaction. Also, with theadvent of language, parents can teach or coachchildren about emotion. Finally, parents influ-ence children’s opportunities to learn aboutemotion by regulating their access to peers,stimulating games, and television programs,and by protecting versus exposing them toemotional displays and arguments (Eisenberg,Cumberland, & Spinrad, 1998; Parke, 1994).

Because preschoolers generally have poorercontrol of their emotion displays than do olderchildren, observational methods in the home


and laboratory are often used to study emo-tions at this age. Across home and laboratorysettings, evidence indicates that negative affect(especially anger) in the context of parent–childinteraction—particularly if it is intense, recip-rocated, and/or more frequent than for mostchildren and parents—is associated withpoorer child outcomes (Carson & Parke, 1996;Cole et al., 2003; Denham, Mitchell-Copeland,Strandberg, Auerbach, & Blair, 1997; Hayden,Klein, & Durbin, 2005; Rubin, Burgess,Dwyer, & Hastings, 2003; Snyder, Stoolmiller,Wilson, & Yamamoto, 2003). In contrast,“emotion coaching” (combining low angerwith verbal strategies) is positively correlatedwith social competence (Denham et al., 1997;Dunn & Brown, 1994; Garner & Spears, 2000;Gottman et al., 1997). Also, converging evi-dence demonstrates that parental sensitive re-sponding, particularly as manifested in secureattachment, is protective for children who areanger-prone (e.g., Denham, Blair, Schmidt, &DeMulder, 2002; Kochanska, 2001).

Parents socialize children’s anger by regulat-ing their exposure to angry situations, either in-tentionally or unintentionally. Children of allages find adults’ anger stressful; exposure tointeradult anger may sensitize children towardanger, making it more likely that they willbe aggressive (Cummings, 1994; Schudlich,Shamir, & Cummings, 2004; Schwartz et al.,1997). Children’s negative emotionality mayexacerbate these effects (Davies & Cummings,1995). Children older than 6 years show in-creasing sensitivity to whether conflicts are re-solved and are less upset by resolved conflicts,but even 4- to 5-year-olds are somewhatsensitive to this information (El-Sheikh &Cummings, 1995; Shifflett-Simpson &Cummings, 1996).

The Expression and Socialization of Angerin the Peer Context

Preschoolers’ emotion regulation skills equipthem, for better or worse, for interactions withpeers. Peer interactions provide a unique con-text for becoming more socially competent andlearning to manage one’s emotions, particularlywhen in conflict with peers (e.g., Putallaz &Sheppard, 1992). Children who are high innegative emotionality play less with peers overtime (Fabes, Hanish, Martin, & Eisenberg,2002), thus missing important socialization ex-periences. Poorly regulated anger is predictiveof conduct problems (Cole et al., 2003; Rydell,

Berlin, & Bohlin, 2003) and lower levels of em-pathy (Strayer & Roberts, 2004), whereaspreschoolers’ emotion knowledge and emotionregulation skills predict social competence withpeers both concurrently and longitudinally(Denham, et al., 2003; Denham, Caverly, et al.,2002; Eisenberg, et al., 1997).

Arsenio and colleagues have examined pre-school children’s affect both during and outsideconflicts. Arsenio and Killen (1996) videotapedand coded with AFFEX triads of preschoolersduring table play. Conflict initiators displayedmore positive affect during conflicts, whereasrecipients expressed more anger and sadness.Children who were “happy initiators” weremore likely to be angry and surprised whenthey were targets of conflict. Recipients of con-flict who exhibited higher levels of anger andsadness were more at risk for both initiatingand being the targets of conflicts. Similar find-ings were reported for conflict and ag-gression observed during free play (Arsenio,Cooperman, & Lover, 2000; Arsenio & Lover,1997). Also, anger expressed outside conflictsituations predicted both initiation of aggres-sion and anger during aggression.

Another approach to studying anger in pre-schoolers is to present hypothetical situationsand to elicit verbal or behavioral responses. Anadvantage of this method is that anger-elicitingsituations can be presented in a standardizedmanner. When researchers are studying natu-rally occurring conflicts/anger, not all childrenexhibit the behaviors of interest, and the cir-cumstances surrounding conflict events varywidely and nonrandomly across participants(Murphy & Eisenberg, 1997). Denham, Bouril,and Belouad (1994) presented preschoolerswith hypothetical provocation and peer entrysituations, and asked them how they would feeland what they would do. Children who saidthat they would be angry were more likely toselect aggression and less likely to pickprosocial behavior as responses. In the class-room, these children were less expressive;teachers were more likely to rate them as miser-able/fearful. Eisenberg, Fabes, Minore, et al.(1994) used a puppet paradigm in which pre-schoolers acted out responses to hypotheticalconflict situations. Children’s puppet enact-ments related well to observed anger reactionsat preschool and to reports from mothers andteachers about how the children handled con-flict. High levels of emotionality were associ-ated with high levels of aggression and lessfriendliness enacted with puppets. Murphy and


Eisenberg (1997) found that boys’ friendly re-sponses to hypothetical provocateurs wererelated to socializers’ reports of high self-regulation and low emotional intensity. Acrossstudies, negative emotionality, especially an-ger, is associated with aggression in preschool-ers.

Preschoolers’ Knowledge about Anger

Preschoolers exhibit increasing understandingof the situational determinants of emotion andrules guiding the display of emotions; thisknowledge predicts social competence bothconcurrently and longitudinally (Denham,1998; Denham et al., 2003). Usinghypothetical-situation methodology, Zeman,Penza, Shipman, and Young (1997) investi-gated the effects of social partner (mother, fa-ther, friend) on 4-year-olds’ use of display rulesand expectancies about managing emotion.Children reported that they would feel angrymost often and regulate emotion more withmothers and friends than with fathers.Children expected mothers and fathers to bemore receptive than friends to displays of emo-tion; they were more likely to report that theywould express negative emotions when theyexpected to receive support or assistance. Thus4-year-olds demonstrated an awareness of dis-play rules and a sensitivity to the role of con-text in their use.

In summary, during the preschool period de-mands are placed on children by family andpeers to regulate and express emotions in so-cially constructive ways. Children make greatstrides in controlling emotion displays and un-derstanding situational determinants of emo-tion and display rules. Some children, however,begin to show deficits in managing emotions,especially anger; these difficulties predictpoorer social functioning and problem behav-iors.

Anger and Its Regulationin School-Age Children

A number of factors converge to make the peercontext salient for school-age children. Withschool entry, children spend more time withpeers and are exposed to a greater number andvariety of peers. School-age children acquire in-formation about their own competencies bycomparing themselves to peers; across the ele-mentary school years, children increase in their

use of social comparison (Stipek & MacIver,1989). Children show increasing sensitivity totheir position relative to peers; being acceptedby peers and having friends assume great impor-tance (Parker & Gottman, 1989), making peergroup norms for emotion regulation especiallysignificant. Parker and Gottman contend thatchildren learn a great deal about display rulesfrom the reactions of peers to their own and oth-ers’ emotion displays. Moreover, the peer normduring middle childhood is to avoid emotional-ity and give the impression of “being ‘cool’—calm, unruffled, and always under emotionalcontrol” (Parker & Gottman, 1989, p. 116).

Research on children’s use and understand-ing of display rules supports Parker andGottman’s (1989) view. Children expect morenegative reactions from peers than parents forexpressing anger and sadness, and report usingdisplay rules more with friends than with par-ents (Zeman & Shipman, 1998). Children ex-pect little support for expressing anger, particu-larly via sulking or aggression (Zeman &Shipman, 1996; Shipman, Zeman, Nesin, &Fitzgerald, 2003), and report more masking ofanger than of other emotions (Underwood,1997). Also, they see anger as hard to control,reporting lower self-efficacy for controlling an-ger than sadness (Zeman & Shipman, 1997).Underwood (1997) found no peer status differ-ences in emotion display rule knowledge, indi-cating that these peer group norms are quite sa-lient for all children.

Given the peer group norms for controllingone’s emotions, it is not surprising that indicesof peer group acceptance during middle child-hood are related to children’s skills at readingand managing emotions, including anger(Bryant, 1992; Cassidy, Parke, Butkovsky, &Braungart, 1992; Hubbard & Coie, 1994;Vosk, Forehand, & Figueroa, 1983). In con-trast, poor regulation of negative emotions, es-pecially anger, is associated with peer rejection(Eisenberg, Fabes, Guthrie, & Reiser, 2000)and victimization (Hanish et al., 2004).

Research and theory on social informationprocessing (SIP) provide a mechanism for un-derstanding how anger influences social be-havior (Crick & Dodge, 1994; Lemerise &Arsenio, 2000). One important process in regu-lating anger responses is a child’s skill in inter-preting others’ intentions in social situations.Children who are relatively poor at interpret-ing peers’ intentions, especially prosocial andbenign intentions, are unlikely to be sociallypopular with peers (Dodge, Murphy, &


Buchsbaum, 1984) and are likely to display ag-gressive behavior toward peers (Dodge, Pettit,McClaskey, & Brown, 1986). Biases in inter-pretations of peers’ intentions are particularlyrelevant to anger and aggression responses.Children who are biased toward attributinghostile intent in peers are likely to display an-ger, behave aggressively, and grow in aggres-sion across development (Dodge, 1980; Dodge& Pettit, 2003). Converging evidence showsthat harsh parenting contributes to anger per-ception bias and to later aggression and con-duct problems (Fine, Trentacosta, Izard,Mostow, & Campbell, 2004; Pollack, 2003;Schultz, Izard, & Bear, 2004; Schultz & Shaw,2003). Also, aggressive individuals report ex-periencing more anger and being less in controlof their anger than do nonaggressive childrenand adolescents (Camodeca & Goossens,2005; DiLiberto, Katz, Beauchamp, &Howells, 2002; Graham, Hudley, & Williams,1992; Orobio de Castro, Merk, Koops,Veerman, & Bosch, 2005). Other research hasshown that aggressive children are more sus-ceptible to anger expressed by others. In video-taped hypothetical situations, provocateurs’angry emotion displays have been found to fa-cilitate hostile social goals and problem solv-ing, whereas provocateurs’ happy displays sup-press hostile goals and are associated withfriendlier problem-solving responses in aggres-sive children (Lemerise, Fredstrom, Kelley,Bowersox, & Waford, 2006; Lemerise, Greg-ory, & Fredstrom, 2005).

In other SIP research, arousal has been ex-perimentally manipulated via staged situations.For example, when a “threat” was staged,rejected–aggressive boys made more SIP errorsof presumed hostility and more hostile attribu-tions than did adjusted–nonaggressive boys(Dodge & Somberg, 1987). Similar resultswere reported in a study using a manipulationthat involved unjustly losing a computer game(Orobio de Castro, Slot, Bosch, Koops, &Veerman, 2003). These studies suggest that foraggressive children, feelings of anger may“short-circuit” evaluations of intent, leading toaggressive responding. This interpretation re-ceives support in a study that examined secondgraders’ nonverbal behavior and physiologicalreactivity in the context of a manipulationwhere children lost a board game to a confed-erate who obviously cheated. Reactive aggres-sion (Dodge & Coie, 1987) was positively re-lated to nonverbal expression of anger andphysiological reactivity, both of which in-

creased throughout the course of the game forthose high in reactive aggression. On the otherhand, children low in reactive aggression dis-played low and steady levels of nonverbal ex-pressions of anger and physiological reactivitythroughout the game (Hubbard et al., 2002;see also Hubbard, 2001; Underwood &Hurley, 1999; Waschbusch et al., 2002).

The research on school-age children suggestsindividual differences in the ability to deal withanger and provocative situations. Rejected and/or aggressive children have difficulty regulatingtheir anger in provocative situations, and thiscontributes to goals and responses that are lesssocially competent and constructive. Con-verging research suggests that individual differ-ences in negative emotionality and regulationcontribute to peer status differences inemotion-related skills and social competence.Children who are high in regulatory skills andlow in negative emotionality, general emotionalintensity, and nonconstructive coping showhigh-quality social functioning, both concur-rently and longitudinally (Eisenberg et al.,1997, 2005; Murphy & Eisenberg, 1996;Rydell et al., 2003). Moreover, interventionsfor children with conduct problems that in-clude training in the recognition and manage-ment of anger have shown some degree of suc-cess (e.g., Greenberg & Kusche, 1993;Lochman & Wells, 2004; Weiss, Harris,Catron, & Han, 2003).


Examination of the developmental course ofanger and its regulation reveals that the roleand meaning of anger in a child’s repertoirechange developmentally. For infants, the ex-pression of anger is often associated with beingeffective. When someone is restraining an in-fant’s movement, or a barrier is placed betweenthe infant and a desirable object, anger ener-gizes and organizes behaviors (e.g., struggling,pushing away the barrier) that “work” in thesense that they reduce anger and obtain goals.Infants’ angry signals are also effective in pro-ducing responses that reduce anger and in-crease the infants’ sense of efficacy. However,the onset of locomotion changes parents’ viewsof infants; they see infants as autonomous, re-sponsible, and angry, and they are more likelyto report getting angry with their infants. Thusthe “rules” appear to change at this point, and


parents begin to socialize the expression of an-ger more actively. Although children begin tolearn that anger doesn’t always work, they stillfind that it is effective in obtaining instrumen-tal goals (Patterson, Littman, & Bricker, 1967).

The process of socializing anger and its regu-lation is complex. The role of children’s temper-ament seems to be important, but we know lessabout how parents’ temperament contributes tothe process. Clearly, some children experienceemotions more intensely and have more diffi-culty regulating their emotions, and these char-acteristics have long-term implications for theirsocial functioning. In several studies, socializa-tion techniques that combine minimization ofnegative emotions with emotion coaching havebeen found to facilitate empathy and prosocialresponding, and thus to promote emotional andsocial competence. However, it is less clearwhether these techniques work equally well withchildren who vary in emotional intensity andself-regulation skills.

As children move from infancy to earlychildhood, their social circle widens, and peersbecome important elicitors and socializers ofanger. Children must learn to coordinate theirgoals with those of others who are much lessindulgent than parents. Affiliative goals play acrucial role in motivating children to controlarousal in order to sustain the exciting and en-joyable activity of play with peers and to main-tain friendship. Children’s temperamental pro-files and/or their socialization experiences inthe family may prepare them to function moreor less effectively with peers. Difficulties withpeers may further reduce children’s opportuni-ties and motivation to learn to manage anger inways that preserve relationships.

By elementary school, children with deficitsin anger regulation are at risk for problem be-haviors, and children who have been diagnosedwith externalizing disorders show deficits inthe expression, appraisal, and regulation ofemotions, particularly anger. These findingsunderscore the importance of continuing tostudy how anger and other emotions regulateand come to be regulated in the context of so-cial interactions.

In conclusion, progress has been made in thestudy of anger and its development. The resultsreviewed here have considerable applied signif-icance for those working with children whohave difficulties with peers, as well as more se-rious problems. Continued progress dependson resolving methodological debates and devel-oping new techniques to study anger.

REFERENCES

Arsenio, W., Cooperman, S., & Lover, A. (2000). Affec-tive predictors of preschool children’s aggression andpeer acceptance: Direct and indirect effects. Develop-mental Psychology, 36, 438–448.

Arsenio, W., & Killen, M. (1996). Conflict-related emo-tions during peer disputes. Early Education and De-velopment, 7, 43–57.

Arsenio, W., & Lover, A. (1997). Emotions, conflictsand aggression during preschoolers’ free play. BritishJournal of Developmental Psychology, 15, 531–542.

Barefoot, J. C., Dodge, K. A., Peterson, B. L., Dahlstrom,W. G., & Williams, X. B. (1989). The Cook–MedleyHostility Scale: Item content and ability to predict sur-vival. Psychosomatic Medicine, 51, 46–57.

Bennett, D. S., Bendersky, M., & Lewis, M. (2002). Fa-cial expressivity at 4 months: A context by expres-sion analysis. Infancy, 3(1), 97–113.

Bennett, D. S., Bendersky, M., & Lewis, M. (2004). Onspecifying specificity: Facial expressions at 4 months.Infancy, 6(3), 425–429.

Bryant, B. K. (1992). Conflict resolution strategies in re-lation to children’s peer relations. Journal of AppliedDevelopmental Psychology, 13, 35–50.

Bugenthal, D. B., & Grusec, J. E. (2006). Socializationprocesses. In W. Damon & R. M. Lerner (Series Eds.)& N. Eisenberg (Vol. Ed.), Handbook of child psy-chology: Vol. 3. Social, emotional, and personalitydevelopment (6th ed., pp. 366–428). Hoboken, NJ:Wiley.

Buss, K. A., & Goldsmith, H. H. (1998). Fear and angerregulation in infancy: Effects on the temporal dynam-ics of affective expression. Child Development, 69,359–374.

Buss, K. A., & Kiel, E. J. (2004). Comparison of sad-ness, anger, and fear facial expressions when toddlerslook at their mothers. Child Development, 75, 1761–1773.

Calkins, S. D., Dedmon, S. E., Gill, K. L., Lomax, L. E.,& Johnson, L. M. (2002). Frustration in infancy: Im-plications for emotion regulation, physiological pro-cesses, and temperament. Infancy, 3(2), 175–197.

Camodeca, M., & Goossens, F. A. (2005). Aggression,social cognitions, anger and sadness in bullies andvictims. Journal of Child Psychology and Psychiatry,46(2), 186–197.

Campos, J. J., Anderson, D., Barbu-Roth, M., Hubbard,E., Herenstein, M., & Witherington, D. (2000).Travel broadens the mind. Infancy, 1, 149–219.

Campos, J. J., Kermoian, R., & Zumbahlen, M. R.(1992). Socioemotional transformations in the familysystem following infant crawling onset. In N.Eisenberg & R. Fabes (Eds.), New directions forchild development: Vol. 55. Emotion and its regula-tion in early development (pp. 25–40). San Francisco:Jossey-Bass.

Camras, L. A. (1992). Expressive development and ba-sic emotions. Cognition and Emotion, 6, 269–283.

Camras, L. A. (2004). An event–emotion or event–expression hypothesis?: A comment on the commen-


taries on Bennett, Bendersky, and Lewis (2002). In-fancy, 6(3), 431–433.

Camras, L. A., Sullivan, J., & Michel, G. (1993). Do in-fants express discrete emotions?: Adult judgments offacial, vocal, and body actions. Journal of NonverbalBehavior, 17, 171–186.

Carson, J. L., & Parke, R. D. (1996). Reciprocal nega-tive affect in parent–child interactions and children’speer competence. Child Development, 67, 2217–2226.

Casey, R. J., & Schlosser, S. (1994). Emotional re-sponses to peer praise in children with and without adiagnosed externalizing disorder. Merrill–PalmerQuarterly, 40, 60–81.

Cassidy, J., Parke, R. D., Butkovsky, L., & Braungart, J.M. (1992). Family–peer connections: The roles ofemotional expressiveness within the family and chil-dren’s understanding of emotions. Child Develop-ment, 63, 603–618.

Cole, P. M., Teti, L. O., & Zahn-Waxler, C. (2003). Mu-tual emotion regulation and the stability of conductproblems between preschool and early school age.Development and Psychopathology, 15, 1–18.

Crick, N. R., & Dodge, K. A. (1994). A review and re-formulation of social-information-processing mecha-nisms of children’s social adjustment. PsychologicalBulletin, 115, 74–101.

Crockenberg, S. (1985). Toddlers’ reactions to maternalanger. Merrill–Palmer Quarterly, 31, 361–373.

Cummings, E. M. (1994). Marital conflict and chil-dren’s functioning. Social Development, 3, 16–36.

Cummings, E. M., Zahn-Waxler, C., & Radke-Yarrow,M. (1981). Young children’s responses to expressionsof anger and affection by others in the family. ChildDevelopment, 52, 1274–1282.

Davies, P. T., & Cummings, E. M. (1995). Children’semotions as organizers of their reactions to interadultanger: A functionalist perspective. DevelopmentalPsychology, 31, 677–684.

Denham, S. A. (1993). Maternal emotional responsive-ness and toddlers’ social-emotional competence.Journal of Child Psychology and Psychiatry, 34,715–728.

Denham, S. A. (1998). Emotional development inyoung children. New York: Guilford Press.

Denham, S. A., Blair, K. A., DeMulder, E., Levitas, J.,Sawyer, K., Auerbach-Major, S., et al. (2003). Pre-school emotional competence: Pathway to socialcompetence? Child Development, 74, 238–256.

Denham, S. A., Blair, K. A., Schmidt, M., & DeMulder,E. (2002). Compromised emotional competence:Seeds of violence sown early? American Journal ofOrthopsychiatry, 72, 70–82.

Denham, S. A., Bouril, B., & Belouad, F. (1994). Pre-schoolers’ affect and cognition about challengingpeer situations. Child Study Journal, 24, 1–21.

Denham, S. A., Caverly, S., Schmidt, M., Blair, K.DeMulder, E., Caal, S., et al. (2002). Preschool un-derstanding of emotions: Contributions to classroomanger and aggression. Journal of Child Psychologyand Psychiatry, 43(7), 1–16.

Denham, S. A., Mitchell-Copeland, J., Strandberg, K.,Auerbach, S., & Blair, K. (1997). Parental contribu-tions to preschoolers’ emotional competence: Directand indirect effects. Motivation and Emotion, 27,65–86.

DiLiberto, L., Katz, R. C., Beauchamp, K. L., &Howells, G. N. (2002). Using articulated thoughts insimulated situations to assess cognitive activity in ag-gressive and nonaggressive adolescents. Journal ofChild and Family Studies, 11(2), 179–189.

Dodge, K. A. (1980). Social cognition and children’s ag-gressive behavior. Child Development, 51, 162–170.

Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mech-anisms in the cycle of violence. Science, 250, 1678–1683.

Dodge, K. A., & Coie, J. D. (1987). Social-information-processing factors in reactive and proactive aggres-sion in children’s peer groups. Journal of Personalityand Social Psychology, 53, 1146–1158.

Dodge, K. A., Murphy, R. R., & Buchsbaum, K. (1984).The assessment of intention-cue detection skills inchildren: Implications for developmental psycho-pathology. Child Development, 55, 163–173.

Dodge, K. A., & Pettit, G. S. (2003). A biopsychosocialmodel of the development of chronic conduct prob-lems in adolescence. Developmental Psychology,39(2), 349–371.

Dodge, K. A., Pettit, G. S., McClaskey, C. L., & Brown,M. (1986). Social competence in children. Mono-graphs of the Society for Research in Child Develop-ment, 51(2, Serial No. 213).

Dodge, K. A., & Somberg, D. R. (1987). Hostileattributional biases among aggressive boys are exac-erbated under conditions of threat to the self. ChildDevelopment, 58, 213–224.

Dunn, J., & Brown, J. (1994). Affect expression in thefamily, children’s understanding of emotion, andtheir interaction with others. Merrill–Palmer Quar-terly, 40, 120–137.

Eisenberg, N. (2006). Introduction. In W. Damon & R.M. Lerner (Series Eds.) & N. Eisenberg (Vol. Ed.),Handbook of child psychology: Vol. 3. Social, emo-tional, and personality development (6th ed., pp. 1–23). Hoboken, NJ: Wiley.

Eisenberg, N., Cumberland, A., & Spinrad, T. L.(1998). Parental socialization of emotion. Psycholog-ical Inquiry, 9, 241–273.

Eisenberg, N., Fabes, R. A., Guthrie, I. K., & Reiser, M.(2000). Dispositional emotionality and regulation:Their role in predicting the quality of social function-ing. Journal of Personality and Social Psychology,78, 136–157.

Eisenberg, N., Fabes, R. A., Minore, D., Mathy, R.,Hanish, L., & Brown, T. (1994). Children’s enactedinterpersonal strategies: Their relations to socialbehavior and negative emotionality. Merrill–PalmerQuarterly, 40, 212–232.

Eisenberg, N., Fabes, R. A., Nyman, M., Bernzweig, J.,& Pinuelas, A. (1994). The relations of emotionalityand regulation to children’s anger-related reactions.Child Development, 65, 109–128.


Eisenberg, N., Fabes, R. A., Shepard, S. A., Murphy, B.C., Guthrie, I. K., Jones, S., et al. (1997). Contempo-raneous and longitudinal prediction of children’s so-cial functioning from regulation and emotionality.Child Development, 68, 642–664.

Eisenberg, N., Sadovsky, A., Spinrad, T. L., Fabes, R. A.,Losoya, S. H., Valiente, C., et al. (2005). The rela-tions of problem behavior status to children’s nega-tive emotionality, effortful control, and impulsivity:Concurrent relations and prediction of change. De-velopmental Psychology, 41, 193–211.

El-Sheikh, M., & Cummings, E. M. (1995). Children’sresponses to angry adult behavior as a function of ex-perimentally manipulated exposure to resolved andunresolved conflict. Social Development, 4, 75–91.

Fabes, R. A., Hanish, L. D., Martin, C. L., & Eisenberg,N. (2002). Young children’s negative emotionalityand social isolation: A latent growth curve analysis.Merrill–Palmer Quarterly, 48, 284–307.

Fernald, A. (1993). Approval and disapproval: Infantresponsiveness to vocal affect in familiar and unfa-miliar languages. Child Development, 64, 657–674.

Field, T., Hernandez-Reif, M., Vera, Y., Gil, K., Diego,M., Bendell, D., et al. (2005). Anxiety and anger ef-fects on depressed mother–infant spontaneous andimitative interactions. Infant Behavior and Develop-ment, 28(1), 1–9.

Fine, S. E., Trentacosta, C. J., Izard, C. E., Mostow, A.J., & Campbell, J. L. (2004). Anger perception, care-givers’ use of physical discipline, and aggression inchildren at risk. Social Development, 13, 213–228.

Garner, P. W., & Spears, F. M. (2000). Emotion regula-tion in low-income preschoolers. Social Develop-ment, 9, 226–245.

Goodenough, F. L. (1931). Anger in young children.Minneapolis: University of Minnesota Press.

Gottman, J. M., Katz, L. F., & Hooven, C. (1997).Meta-emotion: How families communicate emotion-ally. Mahwah, NJ: Erlbaum.

Graham, S., Hudley, C., & Williams, E. (1992). Attri-butional and emotional determinants of aggres-sion among African-American and Latino youngadolescents. Developmental Psychology, 28, 731–740.

Greenberg, M. T., & Kusche, C. A. (1993). Promotingsocial and emotional development in deaf children:The PATHS Project. Seattle: University of Washing-ton Press.

Hanish, L. D., Eisenberg, N., Fabes, R. A., Spinrad, T.L., Ryan, P., & Schmidt, S. (2004). The expressionand regulation of negative emotions: Risk factors foryoung children’s peer victimization. Developmentand Psychopathology, 16, 335–353.

Haviland, J. M., & Lelwica, M. (1987). The induced af-fect response: 10-week-old infants’ responses to threeemotion expressions. Developmental Psychology, 23,97–104.

Hayden, E. P., Klein, D. N., & Durbin, C. E. (2005).Parent reports and laboratory assessments of childtemperament: A comparison of their associationswith risk for depression and externalizing disorders.

Journal of Psychopathology and Behavioral Assess-ment, 27, 89–100.

Hubbard, J. A. (2001). Emotion expression processes inchildren’s peer interaction: The role of peer rejection,aggression, and gender. Child Development, 72,1426–1438.

Hubbard, J. A., & Coie, J. D. (1994). Emotional corre-lates of social competence in children’s peer relation-ships. Merrill–Palmer Quarterly, 40, 1–20.

Hubbard, J. A., Smithmyer, C. M., Ramsden, S. R.,Parker, E. H., Flanagan, K. D., Dearing, K. F., et al.(2002). Observational, physiological, and self reportmeasures of children’s anger: Relations to reactiveversus proactive aggression. Child Development, 73,1101–1118.

Izard, C. E. (1983). The Maximally Discriminative Fa-cial Movement Coding System (MAX) (rev. ed.).Newark: Instructional Resources Center, Universityof Delaware.


Izard, C. E. (2004). The generality–specificity issue ininfants’ emotion responses: A comment on Bennett,Bendersky, and Lewis (2002). Infancy, 6(3), 417–423.

Izard, C. E., Dougherty, L., & Hembree, E. A. (1983). ASystem for Identifying Affect Expressions by HolisticJudgments (AFFEX) (rev. ed.). Newark: ComputerNetwork Services and University Media Services,University of Delaware.

Izard, C. E., Fantauzzo, C. A., Castle, J. M., Haynes, O.M., Rayias, M. F., & Putnam, P. H. (1995). The on-togeny and significance of infants’ facial expressionsin the first nine months of life. Developmental Psy-chology, 31, 997–1013.

Izard, C. E., Hembree, E. A., & Huebner, R. R. (1987).Infants’ emotion expressions to acute pain: Develop-mental change and stability of individual differences.Developmental Psychology, 23, 105–113.

Izard, C. E., & Kobak, R. R. (1991). Emotions systemfunctioning and emotion regulation. In J. Garber &K. A. Dodge (Eds.), The development of emotion reg-ulation and dysregulation (pp. 303–321). New York:Cambridge University Press.

Jenkins, J. M., Franco, F., Dolins, F., & Sewell, A.(1995). Toddlers’ reactions to negative emotion dis-plays: Forming models of relationships. InfantBehavior and Development, 18, 273–281.


Kochanska, G. (2001). Emotional development in chil-dren with different attachment histories: The firstthree years. Child Development, 72, 474–490.

Kochanska, G., Aksan, N., & Carlson, J. J. (2005).Temperament, relationships, and young children’s re-ceptive cooperation with their parents. Developmen-tal Psychology, 41, 648–660.

Lemerise, E. A., & Arsenio, W. F. (2000). An integratedmodel of emotion processes and cognition in social


information processing. Child Development, 71,107–118.

Lemerise, E. A., & Dodge, K. A. (1993). The develop-ment of anger and hostile interactions. In M. Lewis& J. M. Haviland (Eds.), Handbook of emotions(pp. 537–546). New York: Guilford Press.

Lemerise, E. A., & Dodge, K. A. (2000). The develop-ment of anger and hostile interactions. In M. Lewis& J. M. Haviland-Jones (Eds.), Handbook of emo-tions (2nd ed., pp. 594–606). New York: GuilfordPress.

Lemerise, E. A., Fredstrom, B. K., Kelley, B. M.,Bowersox, A. L., & Waford, R. N. (2006). Do provo-cateurs’ emotion displays influence children’s socialgoals and problem solving? Journal of AbnormalChild Psychology, 34, 559–571.

Lemerise, E. A., Gregory, D. S., & Fredstrom, B. K.(2005). The influence of provocateurs’ emotion dis-plays on the social information processing of childrenvarying in social adjustment and age. Journal of Ex-perimental Child Psychology, 90, 344–366.

Lewis, M., Alessandri, S. M., & Sullivan, M. W. (1990).Violation of expectancy, loss of control, and angerexpressions in young infants. Developmental Psy-chology, 26, 745–751.

Lewis, M., & Haviland, J. M. (Eds.). (1993). Handbookof emotions. New York: Guilford Press.

Lewis, M., & Haviland-Jones, J. M. (Eds.). (2000).Handbook of emotions (2nd ed.). New York:Guilford Press.

Lewis, M., & Ramsay, D. (2005). Infant emotional andcortisol responses to goal blockage. Child Develop-ment, 76, 518–530.

Lewis, M., Sullivan, M. W., Ramsay, D., & Alessandri,S. M. (1992). Individual differences in anger and sadexpressions during extinction: Antecedents and con-sequences. Infant Behavior and Development, 15,443–452.

Lochman, J. E., & Wells, K. C. (2004). The coping powerprogram for preadolescent boys and their parents:Outcome effects at the 1-year follow-up. Journal ofConsulting and Clinical Psychology, 72, 571–578.

Lyons-Ruth, K., Alpern, L., & Repacholi, B. (1993).Disorganized infant attachment classification andmaternal psycho-social problems as predictors ofhostile–aggressive behavior in the preschool class-room. Child Development, 64, 572–585.

Malatesta, C. Z., Culver, C., Tesman, J. R., & Shepard,B. (1989). The development of emotion during thefirst two years of life. Monographs of the Society forResearch in Child Development, 54(1–2, Serial No.219).

Malatesta, C. Z., & Haviland, J. M. (1982). Learningdisplay rules: The socialization of emotion ex-pression in infancy. Child Development, 53, 991–1003.

Marsh, A. A., Ambady, N., & Kleck, R. E. (2005). Theeffects of fear and anger facial expressions onapproach- and avoidance-related behaviors. Emo-tion, 5, 119–124.

Montague, D. P. F., & Walker-Andrews, A. S. (2001).Peekaboo: A new look at infants’ perception of emo-tion expressions. Developmental Psychology, 37,826–838.

Murphy, B. C., & Eisenberg, N. (1996). Provoked by apeer: Children’s anger-related responses and their re-lations to social functioning. Merrill–Palmer Quar-terly, 42, 103–124.

Murphy, B. C., & Eisenberg, N. (1997). Young chil-dren’s emotionality: Regulation and social function-ing and their responses when they are targets of apeer’s anger. Social Development, 6, 18–36.

Murphy, B. C., Shepard, S. A., Eisenberg, N., & Fabes,R. A. (2004). Concurrent and across time predictionof young adolescents’ social functioning: The role ofemotionality and regulation. Social Development,13, 56–86.

Oatley, K., & Jenkins, J. M. (1996). Understandingemotions. Oxford: Blackwell.

Orobio de Castro, B., Merk, W., Koops, W., Veerman, J.W., & Bosch, J. D. (2005). Emotions in social infor-mation processing and their relations with reactiveand proactive aggression in referred aggressive boys.Journal of Clinical Child and Adolescent Psychology,34, 105–116.

Orobio de Castro, B., Slot, N. W., Bosch, J. D., Koops,W., & Veerman, J. W. (2003). Negative affect exacer-bates hostile attributions of intent in highly aggres-sive boys. Journal of Clinical Child and AdolescentPsychology, 32, 57–66.

Oster, H., Hegley, D., & Nagel, L. (1992). Adult judg-ments and fine-grained analysis of infant facial ex-pressions: Testing the validity of a priori coding for-mulas. Developmental Psychology, 28, 1115–1131.

Parke, R. D. (1994). Progress, paradigms, and unre-solved problems: A commentary on recent advancesin our understanding of children’s emotions. Merrill–Palmer Quarterly, 40, 157–169.

Parker, J. G., & Gottman, J. M. (1989). Social and emo-tional development in a relational context: Friend-ship interaction from early childhood to adolescence.In T. J. Berndt & G. W. Ladd (Eds.), Peer relation-ships in child development (pp. 95–131). New York:Wiley.

Patterson, G. R., Littman, R. A., & Bricker, W. (1967).Assertive behavior in children: A step toward a the-ory of aggression. Monographs of the Society for Re-search in Child Development, 32(5, Serial No. 113).

Pickens, J., & Field, T. (1993). Facial expressivity in in-fants of depressed mothers. Developmental Psychol-ogy, 29, 986–988.

Pollack, S. D. (2003). Selective attention to facial emo-tion in physically abused children. Journal of Abnor-mal Psychology, 112(3), 323–338.

Putallaz, M., & Sheppard, B. H. (1992). Conflict man-agement and social competence. In C. U. Shantz &W. W. Hartup (Eds.), Conflict in child and adolescentdevelopment (pp. 330–355). New York: CambridgeUniversity Press.

Radke-Yarrow, M., & Kochanska, G. (1990). Anger in


young children. In N. L. Stein, B. Leventhal, & T.Trabasso (Eds.), Psychological and biological ap-proaches to emotion (pp. 297–310). Hillsdale, NJ:Erlbaum.

Reissland, N., Shepherd, J., & Herrera, E. (2005).Teasing play in infancy: Comparing mothers withand without self-reported depressed mood duringplay with their babies. European Journal of Develop-mental Psychology, 2(3), 271–283.

Rubin, K. H., Burgess, K. B., Dwyer, K. D., & Hastings,P. (2003). Predicting preschoolers’ externalizing be-haviors from toddler temperament, conflict, and ma-ternal negativity. Developmental Psychology, 39,164–176.

Rydell, A., Berlin, L., & Bohlin, G. (2003). Emotional-ity, emotion regulation, and adaptation in 5- to 8-year old children. Emotion, 3, 30–47.

Saarni, C., Campos, J. J., Camras, L. A., &Witherington, D. (2006). Emotional development:Action, communication, and understanding. In W.Damon & R. M. Lerner (Series Eds.) & N. Eisenberg(Vol. Ed.), Handbook of child psychology: Vol. 3. So-cial, emotional, and personality development(pp. 226–299). Hoboken, NJ: Wiley.

Schudlich, T. D. D. R., Shamir, H., & Cummings, E. M.(2004). Marital conflict, children’s representations offamily relationships, and children’s dispositions to-wards peer conflict strategies. Social Development,13, 171–192.

Schultz, D., Izard, C. E., & Bear, G. (2004). Children’semotion processing: Relations to emotionality andaggression. Development and Psychopathology, 16,371–387.

Schultz, D., & Shaw, D. S. (2003). Boys’ maladaptivesocial information processing, family emotional cli-mate, and pathways to early conduct problems. So-cial Development, 12, 440–460.

Schwartz, D., Dodge, K. A., Pettit, G. S., & Bates, J. E.(1997). The early socialization of aggressive victimsof bullying. Child Development, 68, 665–675.

Shaw, D. S., Owens, E. B., Vondra, J. I., Keenan, K., &Winslow, E. B. (1996). Early risk factors and path-ways in the development of early disruptive behaviorproblems. Development and Psychopathology, 8,679–699.

Shifflett-Simpson, K., & Cummings, E. M. (1996).Mixed message resolution and children’s responses tointeradult conflict. Child Development, 67, 437–448.

Shipman, K. L., Zeman, J., Nesin, A. E., & Fitzgerald,M. (2003). Children’s strategies for displaying angerand sadness: What works with whom? Merrill–Palmer Quarterly, 49, 100–122.

Snyder, J., Stoolmiller, M., Wilson, M., & Yamamoto,M. (2003). Child anger regulation, parental re-sponses to children’s anger displays, and early childantisocial behavior. Social Development, 12, 335–360.

Stenberg, C. R., & Campos, J. J. (1990). The develop-ment of anger expressions in infancy. In N. L. Stein,

B. Leventhal, & T. Trabasso (Eds.), Psychologicaland biological approaches to emotion (pp. 297–310).Hillsdale: NJ: Erlbaum.

Stipek, D., & MacIver, D. (1989). Developmentalchanges in children’s assessment of intellectual com-petence. Developmental Psychology, 60, 521–538.

Strayer, J., & Roberts, W. (2004). Empathy and ob-served anger and aggression in five-year-olds. SocialDevelopment, 13, 1–13.

Suinn, R. M. (2001). The terrible twos—anger and anx-iety: Hazardous to your health. American Psycholo-gist, 56, 27–36.

Sullivan, M. W., & Lewis, M. (2003). Contextual deter-minants of anger and other negative expressions inyoung infants. Developmental Psychology, 39, 693–705.

Sullivan, M. W., Lewis, M., & Alessandri, S. M. (1992).Cross-age stability in emotional expressions duringlearning and extinction. Developmental Psychology,28, 58–63.

Underwood, M. K. (1997). Peer social status and chil-dren’s understanding of the expression and control ofpositive and negative emotions. Merrill–PalmerQuarterly, 43, 610–634.

Underwood, M. K., & Hurley, J. C. (1999). Emotionregulation in peer relationships during middle child-hood. In L. Balter & C. Tamis-LeMonda (Eds.),Child psychology: A handbook of contemporary is-sues (pp. 58–87). Philadelphia: Psychology Press.

Vosk, B. N., Forehand, R., & Figueroa, R. (1983). Per-ceptions of emotions by accepted and rejected chil-dren. Journal of Behavioral Assessment, 5, 151–160.

Walden, T. A., & Ogan, T. A. (1988). The developmentof social referencing. Child Development, 59, 1230–1240.

Waschbusch, D. A., Pelham, W. E., Jennings, J. R.,Greiner, A. R., Tarter, R. E., & Moss, H. B. (2002).Reactive aggression in boys with disruptive behaviordisorders: Behavior, physiology and affect. Journal ofAbnormal Child Psychology, 30, 641–656.

Weiss, B., Harris, V., Catron, T., & Han, S. H. (2003).Efficacy of the RECAP intervention program for chil-dren with concurrent internalizing and externalizingproblems. Journal of Consulting and Clinical Psy-chology, 71, 364–374.

Zeman, J., Penza, S., Shipman, K., & Young, G. (1997).Preschoolers as functionalists: The impact of socialcontext on emotion regulation. Child Study Journal,27, 41–67.

Zeman, J., & Shipman, K. (1996). Children’s expressionof negative affect: Reasons and methods. Develop-mental Psychology, 32, 842–849.

Zeman, J., & Shipman, K. (1997). Social-contextual in-fluences on expectancies for managing anger andsadness: The transition from middle childhood to ad-olescence. Developmental Psychology, 33, 917–924.

Zeman, J., & Shipman, K. (1998). Influence of socialcontext on children’s affect regulation: A functional-ist perspective. Journal of Nonverbal Behavior,22(3), 141–165.


C H A P T E R 4 6

Self-Conscious EmotionsEmbarrassment, Pride, Shame, and Guilt

MICHAEL LEWIS

In Chapter 18, I have suggested a model for theemergence of emotional life in the first 3 to 4years of life. Here, I focus on a unique set ofemotions that emerge late and that requirecertain cognitive abilities for their elicitation.Whereas the emotions that appear early, suchas joy, sadness, fear, and anger, have receivedconsiderable attention, these later-appearingemotions have received relatively little atten-tion. There are likely to be many reasons forthis. One reason is that these self-consciousemotions cannot be described solely by exam-ining a particular set of facial movement; theynecessitate the observation of bodily actionmore than facial cues (Darwin, 1872/1965).

A second reason for their neglect is the real-ization that there are no clear, specific elicitorsof these particular emotions, although some re-searchers have claimed otherwise (e.g., Izard,1979). Whereas happiness can be elicited byseeing a significant other, and fear can be elic-ited by the approach of a stranger, there are fewspecific situations that will elicit shame, pride,guilt, or embarrassment. These self-consciousemotions are likely to require classes of events

that can only be identified by the individualsthemselves. Consider pride. What kinds ofelicitors are necessary for pride to take place?Pride requires a large number of factors, allhaving to do with cognitions related to the self(Lewis, 1992; Tracy & Robins, 2004). Prideoccurs when one makes a comparison or evalu-ates one’s behavior vis-à-vis some standard,rule, or goal (SRG) and finds that one has suc-ceeded. Shame or guilt, on the other hand, oc-curs when such an evaluation leads to the con-clusion that one has failed.

The elicitation of self-conscious emotions in-volves elaborate cognitive processes that have,at their heart, the notion of self. Although sometheories—psychoanalysis, for example (seeFreud, 1936/1963, and Erikson, 1950)—haveargued for some universal elicitors of shame,such as failure at toilet training or exposureof the backside, the idea of an automatic,noncognitive elicitor of these emotions seemsunlikely. Cognitive processes must be theelicitors of these complex emotions. It is theway we think or what we think about that be-comes the elicitor of pride, shame, guilt, or em-

742

barrassment. There may be a one-to-one corre-spondence between thinking certain thoughtsand the occurrence of a particular emotion;however, in the case of this class of emotions,the elicitor is likely to be some cognitive event.This does not mean that the earlier emotions,those called “primary” or “basic,” are elicitedby noncognitive events. Cognitive factors mayplay a role in the elicitation of any emotion;however, the nature of the cognitive events aremuch less articulated and differentiated in theearlier ones (Plutchik, 1980).

In order to explore these self-conscious emo-tions, we need first to articulate the role of selfin their elicitation. Following this, an attemptat a working definition through a cognitive–attributional model is presented. The chapterfocuses on shame, pride, guilt, and embarrass-ment, although other self-conscious emotionscould be included—for example, jealousy, em-pathy, and envy.

THE ROLE OF SELF

Elsewhere, I have attempted to clarify thosespecific aspects of self that are involved in self-conscious emotions—in particular, the self-conscious evaluative emotions (Lewis, 2003).Self-conscious evaluative emotions first involvea set of standards, rules, or goals (SRGs). TheseSRGs are inventions of the culture that aretransmitted to children and involve their learn-ing of, and willingness to consider, these SRGsas their own. This process of incorporating theSRGs has been discussed by Stipek, Recchia,and McClintic (1992). What is apparent fromthe work of Stipek et al. is that the process ofincorporation starts quite early in life. SRGsimply self-evaluation, for it would make littlesense if we had SRGs but had no evaluation ofour action in regard to them.

Having self-evaluative capacity allows fortwo distinct outcomes: We can evaluate ourbehavior and hold ourselves responsible for theaction being evaluated, or we can hold our-selves not responsible. In the attribution litera-ture, this distinction has been called an “inter-nal” or an “external” attribution, respectively(Weiner, 1986).

If we conclude that we are not responsible,then evaluation of our behavior ceases. How-ever, if we evaluate ourselves as responsible,then we can evaluate our behavior as successfulor unsuccessful vis-à-vis the SRGs. The deter-

mination of success and failure resides withinan individual and is based on the nature of thestandard that is set. For example, if a studentbelieves that only receiving an A in an examconstitutes success, then receiving a B repre-sents a failure for that student. On the otherhand, a B may be considered a success by an-other.

Still another cognition related to the self hasto do with the evaluation of the self in terms ofspecific or global attributions. “Global self-attributions” refer to the whole self, while“specific self-attributions” refer to specific fea-tures or actions of the self (see Beck, 1979;Elliot & Dweck, 2005). In almost every one ofthese processes, a concept of the self needs tobe considered.

The need for cognitive elicitors having to dowith the self was known to Darwin (1872/1965). Darwin not only described the basic,primary, or early emotions, but also dealt withthe self-conscious emotions. Darwin saw theselater emotions as involving the self, althoughhe was not able to distinguish between the vari-ous types (see Tomkins, 1963, for a similarproblem). For example, Darwin believed thatblushing, which could be a sign of either shy-ness, embarrassment, shame, or guilt, wascaused by how we appear to others; as he putit, “the thinking about others, thinking ofus . . . excites a blush” (Darwin, 1872/1965,p. 325). His observation in regard to blushingindicates his concern with two issues: the issueof appearance and the issue of consciousness.He repeatedly made the point that these emo-tions depend on sensitivity to the opinion ofothers, whether good or bad. Thus the distinc-tion between emotions that require opinion orthought of others and emotions that do notsuggests that two different kinds of cognitiveprocesses are involved.

The distinction between self-conscious emo-tions and primary or basic emotions remains ofconcern. The idea that there is a basic set ofemotions grows out of the idea of humaninstincts or propensities. If they are basic,prewired, or genetically given, they are likely tobe limited in number. Although we recognizean enormous variety of emotions, the existenceof each one as a unique and discrete “wiring” istoo burdensome a characterization of the ner-vous system. Instead of positing this complexset of emotions, many have argued that thereare only a few select basic, primary, or pureemotions (see Oatley & Johnson-Laird, 1987,

46. Self-Conscious Emotions 743

and Ortony, Clore, & Collins, 1988, for con-trary views). In order to resolve this problem inregard to self-conscious emotions, we mightinstead make the distinction between emo-tions that involve few or simple cognitive pro-cesses and emotions that involve complex cog-nitive processes (Darwin, 1872/1965; Plutchik,1980).

TOWARD A WORKING DEFINITION

Trying to distinguish among the different typesof self-conscious emotions (e.g., embarrass-ment, shyness, shame, and guilt) is not easy. AsDarwin’s analysis makes clear, all of these emo-tions are likely to produce blushing. Since Dar-win viewed blushing as a human species-specific behavior, he also viewed these emo-tions as unique to humans. However, blushingoccurs with any one of these emotions (or, as inmost cases, not at all), so it is clear that blush-ing cannot be used to distinguish betweenthem.

One can turn to the psychoanalytic litera-ture; however, its focus on guilt rather thanshame (see Broucek, 1991, and Lansky & Mor-rison, 1997, for exceptions) makes this litera-ture suspect. For example, Freud (1905/1953)discussed the function of guilt but said littleabout shame. For Freud, the superego—themechanism by which the standards of the par-ents are incorporated into the self, specificallyvia the child’s fear that the parents will respondto transgression by withdrawal of love or evenby punishment—is the initial source of the feel-ing of guilt. Freud’s discussion of guilt in rela-tionship to the superego is similar to his discus-sion of guilt in relation to the instinctual drivesand their expression. For Freud, anxiety or fearis translatable directly into guilt. The twostages in the development of the sense of guiltrelated to the superego are (1) the fear of au-thority and (2) the fear of the superego itself,once the authority standards are incorporated.In the well-developed superego, the sense ofguilt arises not only when a violation is com-mitted, but even when a violation is being an-ticipated.

The guilt that Freud focuses on is not a guiltrelated to the whole self, but rather a guilt re-lated to one’s action. For Freud, guilt is a spe-cific and focused response to a transgressionthat can be rectified by abstinence and pen-ance. Freud’s focus on guilt, not shame, can

also be found in his discussion of psychopath-ology. It is to be found in the overdevelopedsense of guilt resulting from an overdevelopedego. Within normal functioning, the superegocondemns the ego; this condemnation in turngives rise to normal guilt. When Freud didmention shame, he usually did so in the contextof drives and impulses that require restriction.So, for example, in discussing the impulses hav-ing to do with the erogenous zones, he statedthat these impulses

would seem in themselves to be perverse—that is,to arise from erogenic zones, and to derive theiractivity from instincts which, in view of the direc-tion of the subjects’ development, can arouse onlyunpleasant feelings. They [the impulses] conse-quently evoke opposing mental forces [reactingimpulses] which, in order to suppress this displea-sure affectively, build up the mental dams of . . .disgust, shame and morality. (Freud, 1905/1953,p. 178)

Erikson had no more success in distinguish-ing between shame and guilt than the earlierpsychoanalysts had. Erikson turned more tothe Darwinian view when he suggested thatshame arises when “one is completely exposedand conscious of being looked at, in a word,self-conscious” (1950, pp. 223–224). Again,this self-consciousness is an undifferentiatedstate of being—that is, shame, shyness, embar-rassment, and guilt. Erikson tried to differenti-ate these terms, but was not completely suc-cessful. For example, he discussed “visualshame” versus “auditory guilt,” but did not de-velop this concept. I imagine that the referenceto visual shame is based on Darwin’s theorythat shame derives from being looked at, andthat in feeling shame, one wishes to hide one’sface and to disappear. Although Erikson heldto a more interactional view, one involving selfand self-consciousness, he also indicated thatthe conditions necessary for feeling shame in-clude being in an upright and exposed position.As he stated, “Clinical observation leads me tobelieve that shame has much to do with a con-sciousness of having a front and a back, espe-cially a ‘behind’ ” (Erikson, 1950, pp. 223–224). Erikson believed that shame is related tospecific body acts, particularly toilet functions.Erikson’s theory of ego challenges offers theclearest differentiation between shame andguilt, their place in human life, and eventslikely to elicit them. Erikson’s second challengeis autonomy versus shame and doubt. Auton-


omy is the attempt of the child to achieve, to dofor him- or herself—an attempt that is relatedto a developing sense of the self. Achievingmuscular control, including control of the elim-ination of body waste, is the socialization andthe developmental challenge at this life stage.Shame and doubt arise during this stage as thecounterpoints to autonomy, the successfulachievement. In other words, shame and doubtarise from the child’s inability to control bodilyfunctions fully. It is only after this basic egotask that the third ego task, initiative versusguilt, becomes significant. Here Erikson sug-gested that guilt has a reparative function.Erikson’s developmental sequence indicates arecognition that shame and guilt are differentemotions—that shame precedes guilt, and thatthey are associated in counterpoint with differ-ent ego tasks. It also leads to the commonlyheld psychoanalytic view that guilt is a moremature emotion than shame.

There is very little agreement as to the spe-cific elicitors of shame, guilt, or embarrass-ment. Many events are capable of eliciting anyone of them. No particular stimulus event hasbeen identified as the trigger for shame andguilt. It would be easier to understand theseself-conscious emotions if we could specify theclass of external events likely to elicit them. If itwere true that shame and guilt are similar toanxiety and that they reflect the subject’s fearof uncontrollable impulses, then we could con-sider the causes of shame to be sexual or ag-gressive impulses. Alternatively, if we couldprove that situations having to do with toilet orgenital functions are likely to elicit shame, or ifwe could prove that the way we appear physi-cally or how we behave in front of others mayautomatically elicit embarrassment, we couldthen specify situations that would help us todefine these self-conscious emotions and in-crease our understanding of what causesthem. There is no such clear cause-and-effectpattern—no event that can be used consistentlyas an elicitor of each of these self-consciousemotions.

Alternative theories having to do with selfpsychology are necessary. Success or failure vis-à-vis our SRGs is likely to produce a signalto the self that results in self-reflection (seeMandler, 1975, for a discussion of events likelyto cause self-reflection). This cognitive reflec-tive process gives rise to self-attribution and tothe specific emotions that accompany the dif-ferent types of self-attribution. The importance

of such a view resides in three important fac-tors. First, the model does not attempt tospecify what constitutes success or failure, orhow the person goes about evaluating successor failure. Second, the model does not specifyany particular SRG. In other words, it is notclear whether there are any specific stimuli thatuniquely contribute to any of the self-consciousemotions. Third, the model assumes that self-attributions leading to specific emotions are in-ternal events that reside in people themselves,although the SRGs are taught by others.

Although this model is based on a phenome-nological and cognitive–attributional model,self-conscious emotions are not just epiphen-omenological and therefore deserve “lower sta-tus” than the cognitive–attributional processesthemselves. These self-conscious emotions mayhave discrete and specific locations, as well asspecific processes that are themselves “bodily”in nature (see Lindquist & Barrett, Chapter 32,and Niedenthal, Chapter 36, this volume). Forexample, shame but not embarrassment leadsto increases in cortisol responses (Lewis &Ramsay, 2002). The cognitions associated withthese emotions may serve simply as elicitors ofspecific emotions in the same way as do otherstimuli, such as the social behavior of others,loud noises, or sudden and uncontrolledevents. Specific emotions can be elicitedthrough a variety of attributions. The idea thatcognitions can lead to emotions has beenpoorly received by some, who believe that thisidea implies that cognitions have real status,while emotions are epiphenomenological andcertainly not bodily (Schachter & Singer,1962). We need to give emotions the same sta-tus as cognitions. Just as cognitions can lead toemotions, emotions can lead to cognitions. Thetheory implies no status difference.

A COGNITIVE–ATTRIBUTIONALTHEORY

Figure 46.1 presents a structural model for de-fining various self-conscious emotions. In thefigure, A, B, and C represent cognitive pro-cesses that serve as stimuli for these emotions.

Standards, Rules, and Goals

The first feature of the model has to do withthe SRGs that govern our behavior. All of ushave beliefs about what is acceptable for others


and for ourselves with regard to standards hav-ing to do with actions, thoughts, and feelings.This set of beliefs, or SRGs, constitutes the in-formation one acquires through culturalizationin a particular society. SRGs differ across dif-ferent societies, across groups within societies,across different time epochs, and among indi-viduals of different ages. The standards of ourculture are varied and complex, yet each of usknows at least some of them. Moreover, eachof us has a unique set. To become a member ofany group requires that we learn them. I canthink of no group that does not have SRGs, orin which violations of SRGs does not lead tonegative sanctions. These SRGs are acquiredthrough a variety of processes. They are alwaysassociated with human behavior, includingthinking, action, and feeling. They are pre-scribed by the culture, including the culture atlarge, as well as by the influences of specificgroups, such as clan, peers, and family.

It is safe to claim that by the age of 1 year,children are beginning to learn the appropriateaction patterns reflecting the SRGs of the cul-ture. By the second year of life, children showsome understanding about appropriate and in-appropriate behavior (Heckhausen, 1984). Theacquisition of these SRGs continues across thelifespan; however, some emerge early.

Evaluation

The evaluation of one’s actions, thoughts, andfeelings in terms of SRGs is the secondcognitive–evaluative process that serves as astimulus for self-conscious emotions. Two ma-jor aspects of this process are considered; thefirst has to do with the internal and external as-pects of evaluation. For the model to work in

describing the process of eliciting emotions, in-ternal evaluation, as opposed to either noevaluation or external evaluation, is necessary.Individuals differ in their characteristic eval-uative responses. Moreover, situations differ inthe likelihood that they will cause a particularevaluative response. The second considerationhas to do with how individuals make a deter-mination about success or failure in regard toany specific standard. Research seems to indi-cate that by the beginning of the third year oflife, children already have SRGs and seem toshow distress when they violate them (Lewis,Alessandri, & Sullivan, 1992; Stipek, 1983).

Responsibility

Within the field of attributional studies, theproblem of responsibility has received consid-erable attention (Weiner, 1986). Often whenpeople violate SRGs, they do not attribute thefailure to themselves; they may explain theirfailure in terms of chance or the actions of oth-ers (Seligman et al., 1984). Determining re-sponsibility is a function both of situationalfactors and of individual characteristics. Thereare people who are likely to blame themselvesno matter what happens. Dweck and associ-ates, in studying causes of success and failurewithin academic fields, found that many chil-dren blamed their success or failure on externalforces and not themselves, although there wereas many who were likely to evaluate successand failure in terms of their own actions (seeDweck & Leggett, 1988, for one example). In-terestingly, sex differences emerged: In aca-demic achievement, boys were more apt tohold themselves responsible for their successand others for their failure, whereas girls were


A. STANDARDS AND RULES

B. EVALUATION

SUCCESS FAILURE C. ATTRIBUTION OF SELF

HUBRIS SHAME GLOBAL

PRIDEGUILT/

REGRETSPECIFIC

FIGURE 46.1. Structural model for the elicitation of self-conscious evaluative emotions. From Lewis(1992). Copyright 1992 by Michael Lewis. Reprinted by permission.

apt to blame others for their success and them-selves for their failure.

Success or Failure

Another feature of the self-evaluation processhas to do with the socialization of what consti-tutes success or failure. Once one has assumedresponsibility (internal evaluation), exactlyhow one comes to evaluate an action, thought,or feeling as a success or a failure is not wellunderstood. This aspect of self-evaluation isparticularly important because, as we can seefrom Figure 46.1, the same SRGs can result inradically different feelings, depending uponwhether success or failure is attributed to one-self.

Many factors are involved in producing in-accurate or unique evaluations of success orfailure. These include early failures in the self-system leading to narcissistic disorders (seeMorrison, 1989), harsh socialization experi-ences, and high levels of reward for successor punishment for failure (see Alessandri &Lewis, 1996). The evaluation of one’s behaviorin terms of success and failure is a very impor-tant aspect of the organization of plans and thedetermination of new goals and new plans.

Attribution about Self

As noted earlier, another attribution in regardto the self has to do with “global” or “specific”self-attribution (Beck, 1979). Global attribu-tion refers to an individual’s propensity to fo-cus on the total self. Thus, for any particularbehavior violation, some individuals, some ofthe time, are likely to focus on the totality ofthe self; they use such self-evaluative phrases as“Because I did this, I am bad [or good].”Janoff-Bulman’s (1979) distinction between“characterological” and “behavioral” self-blame is particularly relevant here.

On such occasions the focus is upon the self,both as object and as subject. The self becomesembroiled in the self. It becomes embroiled be-cause the evaluation of the self by the self is to-tal. There is no way out. The focus is not on anindividual’s behavior, but on the total self.There is little wonder that in using such globalattribution people can think of nothing else,and they become confused and speechless (H.B. Lewis, 1971). They tend to focus uponthemselves, not upon their action. Because ofthis, they are unable to act and are driven

from the field of action into hiding or disap-pearing.

“Specific” attribution refers to individuals’propensity in some situations, some of thetime, to focus on specific actions of the self.That is, their self-evaluation is not global, butspecific. It is not the total self that has donesomething wrong or good; instead, particularspecific behaviors are judged. At such times asthese, individuals will use such evaluativephrases as “What I did was wrong. I mustn’t doit again.” Notice that for such occurrences, anindividual’s focus is not on the totality of theself, but on the specific behavior of the self in aspecific situation. The focus here is on thebehavior of the self in interaction with objectsor persons. Here attention is on the actions ofthe self or the effect on other selves.

Global versus specific self-focus may be apersonality style. Global attributions for nega-tive events are generally uncorrelated withglobal attributions for positive events. It is onlywhen positive or negative events are taken intoaccount that relatively stable and consistentattributional patterns are observed. Some indi-viduals are likely to be stable in their globaland specific evaluations; under most conditionsof success or failure, these subjects are likely tomaintain a global or specific posture in regardto self-attribution. In the attribution literature,such dispositional factors have important con-sequences upon a variety of fixed “personalitypatterns.” So, for example, depressed individu-als are likely to make stable global attributions,whereas nondepressed individuals are lesslikely to be stable in their global attributions(Beck, 1979).

In addition to the dispositional factorsrelating to specific or global attributions(Kochanska, 1997), there are likely to be situa-tional constraints as well. Some have calledthese “prototypic situations.” That is, althoughthere are dispositional factors, not all people allthe time are involved in either global or specificattributions. Unfortunately, these situationalfactors have not been well studied. It seems rea-sonable that certain classes of situations shouldbe more likely than others to elicit a particularfocus, but exactly what classes of stimuli arelikely to elicit global or specific attributions re-main unknown. Another way to view this dis-tinction has been made by Dweck (e.g., Dweck,Hong, & Chiu, 1993). In her work, Dweck hasreferred to two styles that are quite similar tospecific and global. She has referred to them as


“task focus” and “self/performance focus.” Intask focus, attention is paid to a specific action;in self/performance focus, attention is on thewhole self.

MAKING SENSE OF THE MODEL

Given these three sets of activities—(1) the es-tablishment of one’s SRGs; (2) the evaluationof success or failure of one’s action in regard tothese; and (3) the attribution of the self, interms of both responsibility (external vs. inter-nal) and self-focus (global vs. specific)—it isnow possible to see how these factors bear onself-conscious emotional states. This model issymmetrical in relation to positive and negativeself-conscious emotions. It is the cognitive–evaluative process of the organism itself thatelicits these states. The model distinguishesamong four emotional states.

Shame

Shame is the product of a complex set of cogni-tive activities: individuals’ evaluation of theiractions in regard to their SRGs and their globalevaluation of the self. The phenomenologicalexperience of a person having shame is that ofa wish to hide, disappear, or die (H. B. Lewis,1971; Lewis, 1992). It is a highly negative andpainful state that also results in the disruptionof ongoing behavior, confusion in thought, andan inability to speak. The physical action ac-companying shame is a shrinking of the body,as though to disappear from the eye of the selfor the other. Because of the intensity of thisemotional state, the global attack on the self-system, all that individuals can do when pre-sented with such a state is to attempt to ridthemselves of it. However, since it is a globalattack on the self, people have great difficultyin dissipating this emotion. There are specificactions individuals employ when shamed in ef-forts to undo the shame state, such as reinter-pretation, self-splitting (multiple personalities),or forgetting (repression); see Lewis (1992) fora fuller discussion.

Shame is not produced by any specific situa-tion, but rather by the individual’s interpreta-tion of the event. Even more important is thatshame is not related necessarily to the event’sbeing public or private. Although many holdthat shame is a public failure, this need not beso. Failure, attributed to the whole self, can be

either public or private. Shame may be public,but it is as likely to be private. Each of us canthink of private events when we say to our-selves, “I’m ashamed of having done that.”Shame can center around moral action as well.Thus, when persons violate some moral SRG,they are ashamed.

Guilt

The emotional state of guilt or regret is pro-duced when individuals evaluate their behavioras failure but focus on the specific features oractions of the self that led to the failure. Unlikethe focus in shame on the global self, the focusin guilt is on the self’s actions and behaviorsthat are likely to repair the failure. From a phe-nomenological point of view, individuals arepained by their failure, but this pained feeling isdirected to the cause of the failure or the objectof harm. Because the cognitive–attributionalprocess focuses on the action of the self ratherthan on the totality of self, the feeling that isproduced—guilt—is not as intensely negativeas shame and does not lead to confusion and tothe loss of action. In fact, the emotion of guiltalways has associated with it a corrective ac-tion that an individual can take (but does notnecessarily take) to repair the failure. Rectifica-tion of the failure and preventing it from occur-ring again are the two possible correctivepaths. Whereas in shame we see the bodyhunched over itself in an attempt to hide anddisappear, in guilt we see individuals moving inspace as if trying to repair their action (seeBarrett & Zahn-Waxler, 1987). The markedpostural differences that accompany guilt andshame are helpful both in distinguishing theseemotions and in measuring individual differ-ences. We might point to blushing as a measurealso distinguishing guilt from shame; however,because of the variability in the likelihood ofindividuals to blush, the use of blushing is notan accurate index.

Because in guilt the focus is on the specific,individuals are capable of ridding themselves ofthis emotional state through action. The cor-rective action can be directed toward the self aswell as toward the other; thus, unlike shame,which is a melding of the self as subject and ob-ject, in guilt the self is differentiated from theobject. As such, the emotion is less intense andmore capable of dissipation. Moreover, itshould be unrelated to maladaptive behavior.The problem is that guilt may merge into


shame; thus two types of guilt might be consid-ered, guilt and maladaptive guilt. The expecta-tion would be that guilt would not be corre-lated with shame, while maladaptive guiltwould be (Tangney & Dearing, 2002). Thus,should the corrective action not beforthcoming—in either thought, feeling, ordeed—it is possible that a guilt experience canbe converted into one of shame (H. B. Lewis,1971). Here, then, is another difference be-tween shame and guilt. We can be ashamed ofour guilty action, but we cannot be guilty overbeing ashamed, suggesting a levels differenceand a directional difference in the experiencingof these emotions. The emotion of guilt lacksthe negative intensity of shame. It is not self-destroying, and as such can be viewed as amore useful emotion in motivating specific andcorrective action. However, because it is less in-tense, it may not convey the motivation neces-sary for change or correction.

Hubris

Hubris is defined as exaggerated pride or self-confidence, often resulting in retribution. It isan example of pridefulness—something dis-likable and to be avoided. Hubris is a conse-quence of an evaluation of success in regard toone’s SRGs where the focus is on the globalself. In this emotion, the individual focuses onthe total self as successful. Hubris is associatedwith such descriptions as “puffed up.” In ex-treme cases, it is associated with grandiosity orwith narcissism (Morrison, 1989). Mueller andDweck (1998) have shown that too muchpraise of children may result in negativeachievement performance; the assumed mecha-nism may be that such high levels of praise mayelicit a global self-focus rather than a specificfocus—or, to use Dweck’s terms, may lead to aself/performance focus rather than a task focus.The global self-focus can lead in time to shameover failure. Given this powerful negative state,the desire to avoid it can lead to not taking re-sponsibility for the failure on the one hand, andto assuming success where others see failure onthe other. These two features have been associ-ated with narcissistic disorders (Lansky &Morrison, 1997). Hubris occurs because of theglobal or self/performance focus. Because ofthe global nature of the attributions of thisemotion, it is likely to be transient, and in orderto be able to maintain this state, one must ei-ther alter standards or reevaluate what consti-

tutes success. Unlike shame, it is highly positiveand emotionally rewarding; that is, one feelsgood about oneself.

Hubris is, however, an emotion difficult tosustain, since there is no specific action thatprecipitates the feeling. Because such feelingsare addictive, people prone to hubris derive lit-tle satisfaction from the emotion. Conse-quently, they seek out and invent situationslikely to repeat this emotional state. They cando this either by altering their SRGs or by re-evaluating what constitutes success in their ac-tions, thoughts, or feelings.

From the outside, other people observe theindividuals having hubris with some disdain.Hubristic people have difficulty in their inter-personal relations, since their own hubris islikely to interfere with the wishes, needs, anddesires of others, in which case there is likely tobe an interpersonal conflict. Moreover, giventhe contemptuousness associated with hubris,the other persons are likely to be shamed by thenature of the actions of the persons having thisemotion. The three problems associated withhubris, therefore, are that (1) it is a transientbut addictive emotion; (2) it is not related to aspecific action, and thus requires altering pat-terns of goal setting or evaluation of whatconstitutes success; and (3) it interferes with in-terpersonal relationships because of its con-temptuous and insolent nature. In its extremeform, it is quite similar to narcissism.

Pride

Pride, on the other hand, is the consequence ofa successful evaluation of a specific action. Thephenomenological experience is joy over an ac-tion, thought, or feeling well done. Here, thefocus of pleasure is specific and related to aparticular behavior. In pride, the self and objectare separated, as in guilt (see Tracy & Robins,2004). Unlike shame and hubris, where subjectand object are fused, pride occurs when peoplefocus on their actions; the persons are en-grossed in the specific actions that give thempride. Stipek et al. (1992) have likened thisstate to achievement motivation. Because thispositive state is associated with a particular ac-tion, individuals have available to themselvesthe means by which they can reproduce thestate. Notice that, unlike hubris, pride’s specificfocus allows for action. Because of the generaluse of the term “pride” to refer to “hubris,”“efficacy,” and “satisfaction,” the study of


pride as hubris has until recently received littleattention. However, more work now is begin-ning to appear. Recently, Tracy, Robins, andLagattuta (2005) have shown that this recogni-tion of pride occurs relatively early (by 4 yearsof age) and has a set of specific behavioralmarkers.

Embarrassment and Shyness

From a behavioral as well as a phenomenologi-cal point of view, shame and guilt can be differ-entiated; however, embarrassment and shynessappear to be confused with these as well.

Shyness

Izard and Tyson (1986) consider shyness to besheepishness, bashfulness, a feeling of uneasi-ness or psychological discomfort in social situ-ations, and oscillation between fear and inter-est or between avoidance and approach. In thisdescription, shyness is related to fear and is anonevaluative emotion centered around the in-dividual’s discomfort response to others. Such adescription fits Buss’s (1980) notion of shynessas an emotional response that is elicited by ex-periences of novelty or conspicuousness. ForBuss, shyness and fear are closely related andrepresent a fearfulness toward others. A way ofdistinguishing shyness from shame is that it ap-pears much earlier than either shame or guilt.

Such an approach to shyness seems reason-able because it fits with other notions relatingthe self to others, or what we might call the“social self.” Eysenck (1954) has characterizedpeople as social or asocial by genetic disposi-tion, and Kagan (2003) has pointed out thephysiological responses of children he calls “in-hibited.” These inhibited or shy children arewithdrawn, are uncomfortable in social situa-tions, and appear fearful. Shyness may bea dispositional factor not related to self-evaluation (DiBiase & Lewis, 1997; Lewis &Ramsay, 1997). Rather, it may simply be thediscomfort of being in the company of othersocial objects—in other words, the opposite ofsociability.

Embarrassment

For Tomkins (1963), embarrassment is closelylinked to shame. The most notable differencebetween embarrassment and shame is the in-

tensity level. Whereas shame appears to be anintense and disruptive emotion, embarrassmentis clearly less intense and does not involve thedisruption of thought and language that shamedoes. Second, in terms of body posture, peoplewho are embarrassed do not assume the pos-ture of one wishing to hide, disappear, or die.In fact, their bodies reflect an ambivalentapproach-and-avoidance posture. Repeatedlooking and then looking away, accompaniedby smiling behavior, seem to index embarrass-ment (see Edelman, 1987; Geppert, 1986;Lewis, Stanger, & Sullivan, 1989). Rarely in ashame situation do we see gaze aversion ac-companied by smiling behavior. Thus, from abehavioral point of view, these two emotionsappear to be different. Finally, whereas shamecan occur in the physical absence of another,embarrassment almost always occurs in thecompany of others.

Phenomenologically, embarrassment is lessclearly differentiated from shame than fromguilt. People often report that embarrassmentis “a less intense experience of shame.” Similarsituations that invoke shame are found to in-voke embarrassment, although its intensity, du-ration, and disruptive quality are not the same.However, in order to understand embarrass-ment, I have suggested (Lewis, 1995) that thereare two types of embarrassed behavior: (1) em-barrassment as exposure, and (2) embarrass-ment as less intense shame.

EMBARRASSMENT AS EXPOSURE

In certain situations of exposure, people be-come embarrassed. It is not related to negativeevaluation, as is shame. Perhaps the best exam-ple is the case of being complimented. One phe-nomenological experience of those who appearbefore audiences is that of embarrassmentcaused by the positive comments of the intro-duction. Consider the moment when thespeaker is introduced: The person introducingthe speaker extols his or her virtues. Surpris-ingly, praise, rather than displeasure or nega-tive evaluation, elicits embarrassment!

Another example of this type of embarrass-ment can be seen in reactions to public display.When people observe someone looking atthem, they are apt to become self-conscious,look away, and touch or adjust their bodies.When an observed person is a woman, she willoften adjust or touch her hair; men are less


likely to touch their hair, but may adjust theirclothes or change their body posture. In fewcases do the observed people look sad. If any-thing, they appear pleased by the attention.This combination—gaze turned away briefly,no frown, and nervous touching—looks likethis first type of embarrassment.

A third example of embarrassment as expo-sure can be seen in this experiment: When Iwish to demonstrate that embarrassment canbe elicited just by exposure, I announce that Iam going to point randomly to a student. I re-peatedly mention that my pointing is randomand that it does not reflect a judgment aboutthe person. I close my eyes and point. Mypointing invariably elicits embarrassment inthe student pointed to.

In each of these examples, there is no nega-tive evaluation of the self in regard to SRGs. Inthese situations, it is difficult to imagine embar-rassment as a less intense form of shame. Sincepraise cannot readily lead to an evaluation offailure, it is likely that embarrassment resultingfrom compliments, from being looked at, andfrom being pointed to has more to do with theexposure of the self than with evaluation. Situ-ations other than praise come to mind, inwhich a negative evaluation can be inferred, al-though it may not be the case. Take, for exam-ple, walking into a room before the speaker hasstarted to talk. It is possible to arrive on timeonly to find people already seated. When youare walking into the room, eyes turn towardyou, and you may experience embarrassment.One could say that there is a negative self-evaluation: “I should have been earlier; Ishould not have made noise (I did not makenoise).” I believe, however, that the experienceof embarrassment in this case may not be elic-ited by negative self-evaluation, but simply bypublic exposure. We (Lewis & Ramsay, 2002)have shown that this type of embarrassmentproduces less stress (e.g., lower cortisol re-sponses) than does shame.

EMBARRASSMENT AS LESS INTENSE SHAME

The second class of embarrassment, which Icall “embarrassment as less intense shame,”seems to me to be related to a negative self-evaluation. The difference in intensity canprobably be attributed to the nature of thefailed SRG. Some SRGs are more or less associ-ated with the core of self; for me, failure atdriving a car is less important than is failure athelping a student. Failures associated with lessimportant and central SRGs result in embar-rassment rather than shame. If this analysis iscorrect, then it is possible that each of the fourself-conscious emotions has a less intense form.

It may well be that embarrassment is not thesame as shame. From a phenomenologicalstance, they appear very different. On the otherhand, there is the possibility that embarrass-ment and shame are in fact related and thatthey vary only in intensity. It is safe to say that,as a working definition, there appear to be atleast two different types of embarrassment.

SHAME AND PSYCHOPATHOLOGY

Recently my colleagues and I have extendedour interest in the development of shame tostudy its relation to psychopathology. Spe-cifically, we are concerned with how traumaticevents influence shame, and how shame theninfluences psychosocial adjustment.

Figure 46.2 presents our general model. Onthe far left side of the model is the traumaticevent. Trauma leads to shame through the me-diation of cognitive attributions about theabuse (a), and shame (b) in turn leads to pooradjustment (d). The model also allows for thetrauma to directly influence shame (c) and ad-justment (e), although our central hypothesis isthat the outcomes of trauma are mediated byhow the child thinks about the event(s). Thesupport for this model comes from the research


FIGURE 46.2. A model of trauma and adjustment as they relate to shame.

suggesting that individuals who are shame-prone are more likely to evidence poor self-esteem, depression, and dissociation (Lewis,1992; Tangney, Wagner, & Gramzow, 1990).Consequently, an internal, stable, global attri-bution style for negative events constitutes arisk factor for shame and subsequent poor ad-justment.

Sexual Abuse and Shame

The experience of shame as a consequence ofsexual abuse is a central mechanism related tosubsequent behavior problems (Lewis, 1992).The theory of the elicitation of shame is partic-ularly relevant to abuse, because attributions asthey are related to self-conscious emotions andattributions related to the cause of abuse areconnected. If the abuse is attributed to an inter-nal, global cause, the resulting emotion isshame; therefore, how the victim evaluates thesexual abuse event(s) is important. It is likely tomediate subsequent long-lasting effects of theabuse (Janoff-Bulman & Frieze, 1983; Wyatt& Mickey, 1988). Given that making internal,stable, global attributions for negative events isrelated to poor adjustment, low self-esteem,helplessness, psychological distress, and de-pression (Metalsky, Abramson, Seligman,Semmel, & Peterson, 1982; Peterson &Seligman, 1983; Tangney et al., 1990), shameas the intervening cause of symptoms followingsexual abuse is likely.

The work available on the relation betweenattribution style and sexual abuse in adults

who were victimized in childhood supports theidea that attribution style mediates the long-term outcome of sexual abuse. Gold (1986)compared the current functioning of womenwho had been sexually abused as children tothat of women who did not report beingabused, and found a strong relation between avictim’s attribution style and her adult func-tioning. Women who were sexually victimizedin childhood, and who reported psychologicaldistress and low self-esteem, were more likelyto have an attribution style characterized by in-ternal, stable, global attributions for negativeevents.

We have been able to show the relationsamong sexual abuse, shame, and adjustment(Feiring, Taska, & Lewis, 2002). In a longitudi-nal study of children ages 8–15 years who wereknown to have been sexually abused, we exam-ined the relations of the severity of sexualabuse, shame, and attribution to adjustment (inthis case, depression), using risk and protectivefactors as covariants.

Figure 46.3 presents these data. Our findingsindicate that within 6 months of the reportedabuse (left side of the figure), severity of abuseand amount of shame were both related to de-pression. However, by 1 year after the reportedabuse (right side of the figure), only amount ofshame was related to depression. Perhaps moreimportantly, children whose shame decreasedover the year showed decreases in depression,while children whose shame stayed the same orincreased actually showed increases in depres-sion.


FIGURE 46.3. Depression and shame in sexually abused children. ap < .05; bp < .01; cp < .001. Data fromFeiring, Taska, and Lewis (2002).

Maltreatment, Shame, and Attribution

The same model underlying psychopathologyin sexually abused children should be evidentfor maltreated children. Although the literatureon maltreatment in children (i.e., physicalabuse and neglect) is extensive, there has beenrelatively little work on the relation betweenmaltreatment and shame. We (Alessandri &Lewis, 1996), for example, have looked at thisrelation and found that shame and pride inregard to failure and success in learningachievement-like tasks are influenced byshame. Specifically, maltreatment should resultin more shame and less pride relative tononmaltreatment. The results of this study in-dicate that maltreated children show less pridewhen they succeed and more shame when theyfail, relative to children from the same back-ground who have not been maltreated (see Fig-ure 46.4).

Moreover, important sex differences appear.Maltreated girls show more shame when theyfail a task, and less pride when they succeed,than do nonmaltreated girls. Boys, on the otherhand, show a suppression of both shame andpride. These sex difference findings are particu-larly important, since for girls trauma mayresult in depression (little pride and muchshame), while for boys trauma may result in asuppression of emotion in general and there-

fore an increase in the likelihood of aggression,since they are not constrained by feelings ofshame, guilt, or regret. Observations of theseboys indicate higher amounts of such behaviorsas throwing the test materials away, verballyaggressive statements, and (although not com-mon) angry faces. In our more recent work(Sullivan & Lewis, 1999), we have measuredthese behaviors more carefully and found sig-nificant differences for maltreated and nonmal-treated boys’ aggressive laboratory behavior. Ifthese findings persist, the sex differences in re-sponse to traumas like maltreatment may ex-plain why girls and women show high likeli-hood of depression, while boys and men showhigh likelihood of aggressive behavior, as aconsequence of similar traumas. It is also inter-esting to note that children who are physicallyabused show less shame than those who are ne-glected (Lewis & Sullivan, 2005).

Abuse, especially in early childhood, is likelyto lead to serious forms of pathology throughthe mediation of shame and attributions. Re-cently, we (Gold, Sullivan, & Lewis, 2007)studied over 100 adolescents who were incar-cerated for antisocial behavior. Examined washow parental punishment, particularly harshpunishment, affected these adolescents’ delin-quent behavior. The effect of harsh parentalpunishment was mediated by shame and attri-butions in regard to responsibility for their ag-


FIGURE 46.4. Shame and pride in maltreated children. M, maltreated; NM, nonmaltreated. Data fromAlessandri and Lewis (1996).

gressive behavior. Using data on parental harshpunishment (beating and physical abuse), aswell as data from the Test of Self-ConsciousAffect–2 (TOSCA-2; Tangney, Ferguson, Wag-ner, Crowley, & Gramzow, 1996), we wereable to examine the adolescents’ violent behav-ior. As seen in Figure 46.5, harsh punishmentwas directly related to violent delinquent acts.It was also related to not taking responsibility(blaming others), which mediated the effects ofpunishment on violent delinquent behavior. In-terestingly, the more they blamed others (theless responsibility they took), the less shamethese adolescents showed. Here again, the con-nection between attribution and shame is ap-parent.

CONCLUSION

The study of self-conscious emotions continuesto be conducted. The structural model outlinedhere (see Figure 46.1) provides a model for theself-conscious emotions. Without a more accu-

rate or agreed-upon taxonomy, the harder itwill be to study these emotions. Given the in-terest in emotional life, and in the relationbetween emotions and health, it seems in-creasingly necessary to study these complexemotions, rather than only the more “primary”or “basic” ones. Moreover, as others havepointed out and as we are now discovering,these self-conscious emotions are intimatelyconnected with other emotions, such as angerand sadness, as well as with psychopathology.Finally, given the place of self-evaluation inadult life, it seems clear that the self-consciousevaluative emotions are likely to stand in thecenter of our emotional life.

REFERENCES

Alessandri, S. A., & Lewis, M. (1996). Differencesin pride and shame in maltreated and nonmal-treated preschoolers. Child Development, 67, 1857–1869.

Barrett, K. C., & Zahn-Waxler, C. (1987, April). Dotoddlers express guilt? Poster presented at the bien-


FIGURE 46.5. The effects of harsh parenting on adolescents’ emotions, attributions, and delinquentbehavior: Model of relations among maltreatment, emotions/attributions, and delinquency. + p < .10;*p < .05; **p < .01. Data from Gold, Sullivan, and Lewis (2007).

nial meeting of the Society for Research in Child De-velopment, Toronto.

Beck, A. T. (1979). Cognitive therapy and the emotionaldisorders. New York: Meridian.

Broucek, F. J. (1991). Shame and the self. New York:Guilford Press.

Buss, A. H. (1980). Self-consciousness and social anxi-ety. San Francisco: Freeman.


DiBiase, R., & Lewis, M. (1997). The relation betweentemperament and embarrassment. Cognition andEmotion, 11, 259–271.

Dweck, C. S., Hong, Y., & Chiu, C. (1993). Implicit the-ories: Individual differences in the likelihood andmeaning of dispositional inferences. Personality andSocial Psychology Bulletin, 19, 644–656.

Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation and personality.Psychological Review, 95, 256–273.

Edelman, R. J. (1987). The psychology of embarrass-ment. Chichester, UK: Wiley.

Elliot, A. J., & Dweck, C. S. (2005). Competence andmotivation: Competence as the core of achievementmotivation. In A. J. Elliot & C. S. Dweck (Eds.),Handbook of competence and motivation (pp. 3–12). New York: Guilford Press.

Erikson, E. H. (1950). Childhood and society. NewYork: Norton.

Eysenck, H. J. (1954). The psychology of politics. Lon-don: Routledge & Kegan Paul.

Feiring, C., Taska, L., & Lewis, M. (2002). Adjustmentfollowing sexual abuse discovery: The role of shameand attributional style. Developmental Psychology,38(1), 79–92.

Freud, S. (1953). Three essays on the theory of sexuality.In J. Strachey (Ed. & Trans.), The standard edition ofthe complete psychological works of Sigmund Freud(Vol. 7, pp. 123–231). London: Hogarth Press.(Original work published 1905)

Freud, S. (1963). The problem of anxiety (H. A. Bunker,Trans.). New York: Norton. (Original work pub-lished 1936)

Geppert, U. (1986). A coding-system for analyzingbehavioral expressions of self-evaluative emotions.Munich: Max-Planck-Institute for Psychological Re-search.

Gold, E. R. (1986). Long-term effects of sexual victim-ization in childhood: An attributional approach.Journal of Consulting and Clinical Psychology, 54,471–475.

Gold, J., Sullivan, M., & Lewis, M. (2007). Emotionsand attributions as mediators between harsh andabusive parenting and juvenile delinquency. Manu-script submitted for publication.

Heckhausen, H. (1984). Emergent achievement behav-ior: Some early developments. In J. Nicholls (Ed.),The development of achievement motivation (pp. 1–32). Greenwich, CT: JAI Press.

Izard, C. E. (1979). The Maximally Discriminative Fa-cial Movement Coding System (MAX). Newark: In-structional Resources Center, University of Dela-ware.

Izard, C. E., & Tyson, M. C. (1986). Shyness as a dis-crete emotion. In W. H. Jones, J. M. Cheek, & S.R. Briggs (Eds.), Shyness: Perspectives on researchand treatment (pp. 147–160). New York: PlenumPress.

Janoff-Bulman, R. (1979). Characterological versusbehavioral self-blame: Inquiries into depression andrape. Journal of Personality and Social Psychology,37, 1798–1809.

Janoff-Bulman, R., & Frieze, I. (1983). A theoreticalperspective for understanding reactions to victimiza-tion. Journal of Social Issues, 39, 1–17.

Kagan, J. (2003). Biology, context, and developmentalinquiry. Annual Review of Psychology, 54, 1–23.

Kochanska, G. (1997). Multiple pathways to consciencefor children with different temperaments: Fromtoddlerhood to age 5. Developmental Psychology,33, 597–615.

Lansky, M. R., & Morrison, A. P. (1997). The wideningscope of shame. Hillsdale, NJ: Analytic Press.

Lewis, H. B. (1971). Shame and guilt in neurosis. NewYork: International Universities Press.

Lewis, M. (1992). Shame: The exposed self. New York:Free Press.

Lewis, M. (1995). Embarrassment: The emotion of self-exposure and evaluation. In J. P. Tangney & K. W.Fischer (Eds.), Self-conscious emotions: The psychol-ogy of shame, guilt, embarrassment, and pride(pp. 198–218). New York: Guilford Press.

Lewis, M. (2003). The emergence of consciousness andits role in human development. Annals of the NewYork Academy of Sciences, 1001, 1–29.

Lewis, M., Alessandri, S. M., & Sullivan, M. W. (1992).Differences in shame and pride as a function of chil-dren’s gender and task difficulty. Child Development,63, 630–638.

Lewis, M., & Ramsay, D. (1997). Stress reactivity andself-recognition. Child Development, 68, 621–629.

Lewis, M., & Ramsay, D. (2002). Cortisol response toembarrassment and shame. Child Development,73(4), 1034–1045.

Lewis, M., Stanger, C., & Sullivan, M. W. (1989). De-ception in three-year-olds. Developmental Psychol-ogy, 25, 439–443.

Lewis, M., & Sullivan, M. W. (2005). The developmentof self-conscious emotions. In A. J. Elliott & C. S.Dweck (Eds.), Handbook of competence and motiva-tion (pp. 185–201). New York: Guilford Press.


Metalsky, G. J., Abramson, L. Y., Seligman, M. E. P.,Semmel, A., & Peterson, C. (1982). Attributionalstyles and life events in the classroom: Vulnerabilityand invulnerability to depressive mood reactions.Journal of Personality and Social Psychology, 43,612–617.


Morrison, A. P. (1989). Shame: The underside of narcis-sism. Hillsdale, NJ: Analytic Press.

Mueller, C. M., & Dweck, C. S. (1998). Praise for intel-ligence can undermine children’s motivation and per-formance. Journal of Personality and Social Psychol-ogy, 75, 33–52.

Oatley, K., & Johnson-Laird, P. N. (1987). Toward acognitive theory of emotions. Cognitions and Emo-tion, 1, 29–50.


Peterson, C., & Seligman, M. E. P. (1983). Learnedhelplessness and victimization. Journal of Social Is-sues, 39, 105–118.

Plutchik, R. (1980). A general psychoevolutionary the-ory of emotion. In R. Plutchik & H. Kellerman(Eds.), Emotion: Theory, research, and experience(Vol. 1, pp. 3–33). New York: Academic Press.


Seligman, M. E. P., Peterson, C., Kaslow, N., Tanen-baum, R., Alloy, L., & Abramson, L. (1984). Attribu-tional style and depressive symptoms among children.Journal of Abnormal Psychology, 39, 235–238.

Stipek, D. J. (1983). A developmental analysis of prideand shame. Human Development, 25, 42–54.

Stipek, D. J., Recchia, S., & McClintic, S. (1992). Self-evaluation in young children. Monographs of the So-ciety for Research in Child Development, 57(1, SerialNo. 226).

Sullivan, M. W., & Lewis, M. (1999, April). The emo-tions of maltreated children in response to successand failure. Paper presented at the biennial meetingof the Society for Research in Child Development,Albuquerque, NM.

Tangney, J. P., & Dearing, R. L. (2002). Shame andguilt. New York: Guilford Press.

Tangney, J. P., Ferguson, T. J., Wagner, P. E., Crowley,S. L., & Gramzow, R. (1996). The Test of Self-Conscious Affect–2 (TOSCA-2). Fairfax, VA: GeorgeMason University.

Tangney, J. P., Wagner, P., & Gramzow, R. (1990, June).Shame-proneness, but not guilt-proneness, is linkedto psychological maladjustment. Poster presented atthe meeting of the American Psychological Society,Dallas, TX.

Tomkins, S. S. (1963). Affect, imagery, and con-sciousness: Vol. 2. The negative affects. New York:Springer.

Tracy, J. L., & Robins, R. W. (2004). Show your pride:Evidence for a discrete emotion expression. Psycho-logical Science, 15(3), 194–197.

Tracy, J. L., Robins, R. W., & Lagattuta, K. H. (2005).Can children recognize pride? Emotion, 5(3), 251–257.

Weiner, B. (1986). An attributional theory of motivationand emotion. New York: Springer-Verlag.

Wyatt, G. E., & Mickey, M. R. (1988). The support ofparents and others as it mediates the effects of childsexual abuse: An exploratory study. In G. E. Wyatt &G. J. Powell (Eds.), Lasting effects of child sexualabuse (pp. 211–226). Newbury Park, CA: Sage.


C H A P T E R 4 7

Disgust

PAUL ROZIN, JONATHAN HAIDT, and CLARK R. McCAULEY

For North Americans, elicitors of disgust comefrom nine domains: food, body products, ani-mals, sexual behaviors, contact with death orcorpses, violations of the exterior envelope ofthe body (including gore and deformity), poorhygiene, interpersonal contamination (contactwith unsavory human beings), and certainmoral offenses (Haidt, McCauley, & Rozin,1994; Rozin, Haidt, & McCauley, 1993;Rozin, Haidt, McCauley, & Imada, 1997).What unites these disparate domains? Al-though all involve negative or unpleasantevents, there are many negative events, such aspain and loss, that are not disgusting. The pri-mary goal of this chapter is to make sense ofthis varied set of elicitors—that is, to describethe meaning of disgust within evolutionary, de-velopmental, and cultural contexts. We arguefor a path of development in individuals andcultures that extends from the presumed originof disgust as a rejection response that protectsthe body from “bad” foods, to a rejection sys-tem that protects the soul from the full range ofelicitors listed above.

DEFINING DISGUST

There are two classic papers describing disgust,published some 70 years apart. The first, achapter in Darwin’s The Expression of theEmotions in Man and Animals (1872/1965),defined disgust as referring to “something re-volting, primarily in relation to the sense oftaste, as actually perceived or vividly imagined;and secondarily to anything which causes asimilar feeling, through the sense of smell,touch and even of eyesight” (p. 253). Darwinrelated disgust not only to the experience of re-vulsion but to a characteristic facial expression.The second paper, by psychoanalyst AndrasAngyal (1941), held that disgust is revulsion atthe prospect of oral incorporation of an offen-sive object. He identified body waste productsas a focus of disgust, and related the strength ofdisgust to the degree of intimacy of contact.Our own description of disgust, or what wecall “core disgust,” follows on Angyal’s defini-tion above, adding this sentence: “The offen-sive objects are contaminants; that is, if they

757

even briefly contact an acceptable food, theytend to render that food unacceptable” (Rozin& Fallon, 1987, p. 23).

Most definitions focus on the mouth andreal or imagined ingestion. Tomkins (1963,1982) held that of all the emotions, disgust hasthe clearest linkage to a specific motivation(hunger), and functions to oppose this motive.Ekman and Friesen (1975) see disgust asan aversion that centers on oral rejection.Wierzbicka (1986) defines disgust as a bad feel-ing about another person’s action, “similar towhat one feels when one has something in one’smouth that tastes bad and when one wants tocause it to come to be out of one’s mouth”(p. 590).

Some have proposed systems other than in-gestion as the origin of disgust. Freud (1905/1953) predictably linked it to sex, and others(e.g., Renner, 1944; Plutchik, 1980) see its ori-gin as a defense against infection, with the skinplaying a central role. Curtis and her colleagues(Curtis & Biran, 2001; Curtis, Aunger, &Rabie, 2004) have presented evidence that thebest single account for what is currently dis-gusting is infection potential, and have sug-gested that protection from infection providesthe adaptive reason for the evolution of dis-gust. The fact that contamination sensitivity isa basic feature of disgust supports this claim.Kelly (2007) proposes that disgust has a dualorigin. His entanglement hypothesis posits aconvergence of an orally focused toxin avoid-ance system (present in nonhumans) and abroader parasite avoidance system that aroseduring human evolution.

It seems likely that threats of disease and in-fection shaped the disgust response as humansincreased their intake of foods of animal originand as group densities increased. Both of thesemajor changes increased parasite risk. Still, wefind the arguments for a food origin convincing(Rozin & Fallon, 1987). The English term “dis-gust” itself means “bad taste,” and the facialexpression of disgust can be seen as functionalin rejecting unwanted foods and odors. Themost distinct physiological concomitant ofdisgust—nausea—is a food-related sensationthat inhibits ingestion. Finally, the brain regionmost often activated in studies of disgust (in-cluding non-food-related stimuli, such as muti-lations and disgust faces) is the anterior insula(Husted, Shapira, & Goodman, 2006), which,among its other functions, is the gustatory cor-tex in primates (Rolls, 1994).

COMPONENTS OF DISGUST

Almost all of the literature on emotion prior to1990 focused on fear, anger, happiness, andsadness. However, disgust seems to have “ar-rived” in the 1990s. Between 1997 and 2006,PsycINFO listed 178 articles with “disgust” inthe title (see also Olatunji & Sawchuk, 2005).In the same period, two academically orientedbooks about disgust were published (W. I.Miller, 1997; S. B. Miller, 2004), and twoinfluential trade books about psychology de-voted considerable attention to disgust (Bloom,2004; Pinker, 1997). One stimulus to the inves-tigation of disgust has probably been the factthat it appears primitive and basic, while at thesame time the broad range of elicitors impli-cates disgust in many uniquely human con-cerns, including morality and divinity. For thisreason, we have described disgust in the title ofone publication as “the body and soul emo-tion” (Rozin, Haidt, & McCauley, 1999), andthis contrast has provoked considerable inter-est (e.g., Bloom, 2004). A related framing ofdisgust conceives it as the emotion that is theguardian of the borders of both the bodily selfand the social self (Fessler & Haley, 2006; S. B.Miller, 2004; Rozin, Nemeroff, Horowitz,Gordon, & Voet, 1995).

Paul Ekman (1992) has provided the clearestarticulation of the characteristics of an emo-tion, and disgust meets all nine of his criteria.We consider here four of Ekman’s criteria.

Behavioral Component

Disgust is manifested as a distancing fromsome object, event, or situation, and can becharacterized as a rejection.

Physiological Component

Disgust is associated with a specific physiologi-cal state—nausea—that is typically measuredby self-report. Another specific physiologicalaspect of disgust has been suggested by Angyal(1941), who pointed to increased salivation, it-self associated with nausea, as a concomitantof disgust. We know of no experimental studiesof the relation of disgust to nausea or saliva-tion. More conventional psychophysiologicalinvestigations of disgust suggest that it is asso-ciated with some degree of parasympatheticautonomic response, particularly lowered heartrate, whereas fear and anger are associated


with a predominantly sympathetic response(Levenson, Ekman, & Friesen, 1990; Leven-son, 1992).

Expressive Component

The expressive component of disgust has beenstudied almost entirely with reference to theface. The characteristics of the “disgust face”have received particular attention from Darwin(1872/1965), Izard (1971), Ekman (Ekman,1972; Ekman & Friesen, 1975), and Rozin,Lowery, and Ebert (1994). Scholars are not incomplete agreement about a prototypical dis-gust face, but the three main components seemto be the gape (Action Unit [AU] 25 or 26in theFacial Action Coding System [FACS; Ekman &Friesen, 1978]), retraction of the upper lip (AU10), and the nose wrinkle (AU 9). Activity cen-ters around the mouth and nose, and the move-ments tend either to discourage entry into thebody (e.g., nose wrinkle) or to encourage dis-charge (gape with or without tongue exten-sion). Facial electromyographic measurementsconfirm the observational data, involving someof the same facial muscles, and including someof the muscles around the eyes (Wolf et al.,2005). Laughter is a common response (as op-posed to the disgust face) in some disgust-eliciting situations (Hemenover & Schimmack,2007).

The N��ya��stra (Masson & Patwardhan,1970), an ancient Hindu treatise on drama (seeShweder, Haidt, Horton, & Joseph, Chapter25, this volume), treats disgust as one ofeight or nine basic emotions. As described byHejmadi (2000), the multiple portrayals of dis-gust designated in this document are dynamic(as opposed to the standard “frozen face” usedin almost all Western research), and involve ac-tions of the whole body, especially the hands.Americans as well as Indians are able to iden-tify these disgust expressions remarkably well(Hejmadi, Davidson, & Rozin, 2000).

Qualia

Qualia, the mental or feeling component ofemotion, may be at once the most centralcomponent of disgust and the most difficultto study. The qualia of disgust is often de-scribed as revulsion. In comparison to otheremotions, the experience of disgust appearsto be rather short in duration (Scherer &Wallbott, 1994).

CORE DISGUST

We believe that disgust was shaped by evolu-tionary forces that elaborated upon an olderfood rejection system based on distaste. In thisand subsequent sections, we describe how coredisgust differs from distaste, and then how dis-gust may have expanded to a much wider rangeof elicitors—including reminders of our animalnature, and certain types of interpersonal con-tact and moral violations.

Core disgust is one of four categories of foodrejection, the others being distaste (rejectionmotivated by bad sensory properties), danger(rejection motivated by fear of bodily harm),and inappropriateness (rejection of a foodculturally classified as not edible) (Fallon &Rozin, 1983; Rozin & Fallon, 1987). Like in-appropriateness, disgust is defined by idea-tional forces: beliefs about the nature or originof a potential food. Unlike inappropriate enti-ties, disgusting entities are presumed to be bothdistasteful and dangerous. The appraisal thatelicits core disgust requires (1) a sense of poten-tial oral incorporation (and hence a linkagewith food or eating), (2) a sense of offensive-ness, and (3) contamination potency (Angyal,1944; Rozin & Fallon, 1987).

Oral Incorporation

Rozin and Fallon (1987) noted that the mouthis the principal route by which material thingsenter the body, and hence can be thought of asthe gateway to the body. Aversion to an offen-sive entity in the mouth is usually stronger thanaversion to the same entity on the body surfacenear but not inside the mouth, or inside thestomach (Rozin et al., 1995).

The threat of oral incorporation is framedby a widespread belief that one takes on theproperties of the food one eats (“You arewhat you eat”). In The Golden Bough, Frazer(1890/1922) noted: “The savage commonlybelieves that by eating the flesh of an animalor man, he acquires not only the physical buteven the moral and intellectual qualitieswhich are characteristic of that animal orman” (p. 573). This belief is consistent withour general experience that when two thingscombine (in this case, a food and a person),the product resembles both. Nemeroff andRozin (1989) found, using the Aschimpression-listing technique, that Americancollege students attribute boar-like qualities

47. Disgust 759

to boar eaters, and turtle-like qualities to tur-tle eaters.

Offensive Entities:Animals and Their Products

Angyal (1941) held that the center of disgustis animal (including human) waste products,which he saw as debasing. Body products are afocus of disgust, and are central to the relatedanthropological concept of pollution (Douglas,1966; Meigs, 1978, 1984). There is widespreadhistorical and cultural evidence for aversionand disgust to virtually all body products,including feces, vomit, urine, and blood(especially menstrual blood). In accord withAngyal’s (1941) suggestion of an animal focusfor disgust, Rozin and Fallon (1987) proposedthat the elicitor category for core disgust con-sists of all animals and their products as poten-tial foods. Relatedly, Martins and Pliner (2006)report a dimension of livingness/animalnessemerging from multidimensional scaling of rat-ings of the disgustingness of a wide range ofnovel foods. Almost all cultures eat only asmall subset of potential animal foods. Angyal(1941) pointed out that in many cultures somecare is taken to disguise the animal origin ofanimal food by cutting, chopping, and otherculinary preparations, as well as by havingnames for animal foods (e.g., “pork” and“beef” in English) that are distinct from thecorresponding animal names.

Animals and their products seem cross-culturally to be both the most favored of foodsand the most tabooed. In short, animal foodsare emotionally charged (Tambiah, 1969) andtend to give rise to ambivalent responses. Manyanimal taboos involve disgust. Some animalsare disgusting because they bear some resem-blance to body products such as mucus (e.g.,slugs), or because they are commonly in con-tact with rotting animal flesh, feces, or otherhuman wastes (e.g., flies, cockroaches, rats,vultures, and other scavengers). Carnivorousland animals eat raw, often decaying animalflesh, and produce putrid feces; they are dis-gusting at both ends. Herbivores are much lesslikely to be prohibited cross-culturally. Eventhe hunter–gatherer !Kung bushmen, who eat amuch wider variety of species than most West-erners do, reject rodents, carnivores, and mostinsects (Howell, 1986).

Two other categories of animal food prohibi-tions deserve mention. Animals that are close

to humans, either in appearance (e.g., otherprimates) or by virtue of a relationship withhumans as pets, are rarely eaten. And finally,there is a group of anomalous animals thatseem to produce a mixture of fear (danger) anddisgust (e.g., spiders and snakes). These ani-mals are feared, although they are rarely harm-ful to humans. Davey and his colleagues(Davey, 1993; Matchett & Davey, 1991; Ware,Jain, Burgess, & Davey, 1993; Webb & Davey,1993) offer evidence that the aversion to theseanimals is based more on disgust than fear.

Contamination

The contamination response—rejection of apotential food if it even briefly contacted a dis-gusting entity—appears to be powerful anduniversal among adults. North American col-lege students reject liked beverages after thesehave briefly contacted a sterilized cockroach(Rozin, Millman, & Nemeroff, 1986), and vir-tually all North Americans reject foods thathave been handled or bitten by either unsavoryor disliked persons (Rozin, Nemeroff, Wane, &Sherrod, 1989). Although this aversion is typi-cally justified as an avoidance of disease, re-moval of this justification (e.g., by sterilizingthe offending dead cockroach) typically hasonly a small effect. Contamination may havebeen shaped as an adaptation for disease avoid-ance, but it operates largely independently ofconscious beliefs about disease.

Rozin and his colleagues have suggested thatcontamination effects may be instances of thesympathetic magical law of contagion (Tylor,1871/1974; Frazer, 1890/1922; Mauss, 1902/1972), which essentially holds that “once incontact, always in contact” (Rozin & Fallon,1987; Rozin & Nemeroff, 1990). The law ofcontagion as applied to disgust is potentiallycrippling; everything people might eat or touchis potentially contaminated. Humans deal withthis problem in a number of ways. First, con-tamination rules are developed in some cul-tures, such as the explicit rules establishing athreshold for contamination in the Hebrew di-etary system (Grunfeld, 1982). These rules pro-vide ritualistic relief but not necessarily psycho-logical relief of a sense of contamination(Nemeroff & Rozin, 1992).

Most often, framing is the strategy thatkeeps potential contamination out ofconsideration—as when we do not think of thepeople in the kitchen who prepare our food in a


restaurant, or the animal that was the source ofour meat, or the fact that our body contains ahost of disgusting substances. The framing so-lution fails when the source of contamination istoo salient.

A second law of sympathetic magic, the lawof similarity, accounts for some other aspects ofdisgust. The law of similarity, also dating fromTylor, Frazer, and Mauss (for reviews, seeRozin & Nemeroff, 1990; Nemeroff & Rozin,2000), holds in one form that if things are su-perficially similar, then they resemble eachother in a deep sense as well. In other words,appearance is reality. The law of similarity isevident when objects that look like somethingdisgusting are treated as disgusting. For exam-ple, many North Americans are reluctant toconsume imitation dog feces that they knoware made out of chocolate fudge (Rozin,Millman, & Nemeroff, 1986).

ANIMAL-NATURE DISGUST

Our discussion of disgust up to this point hasfocused on issues surrounding food and eating.We have presented core disgust as an oral de-fense against harm from potential foods, orthings that can easily contaminate foods suchas body products and some animals. However,when we asked North American and Japaneserespondents to list the things they thought weredisgusting, fewer than 25% of listed examplescame from the three core disgust domains offood, animals, and body products (Haidt,Rozin, McCauley, & Imada, 1997). Many ofthe other examples could be classified into fouradditional domains: inappropriate sexual acts,poor hygiene, death, and violations of the idealbody “envelope” or exterior form (e.g., gore,deformity, obesity). In the four additional do-mains, the focus of threat has spread from themouth to the body in general. This spread iscaptured in a psychoanalytic treatment of dis-gust: “In summary, any modality that repre-sents a means of entry into the self or body—the mouth, the nose, the skin, the eyes—seemsto play a part in the disgust experience” (S. B.Miller, 1986, p. 300). All four of these domainsinvolve potential sources of biological conta-gion and infection (e.g., venereal diseases fromsex, or skin-to-skin or hair-to-hair infectionfrom parasites on an unclean person); thus coredisgust was preadapted and easily expanded toapply contamination sensitivity to these addi-

tional classes of threats. However, we thinkthat something more symbolic was and is goingon as well.

Contact with death and corpses is a particu-larly potent elicitor of disgust. Two of the itemsin our 32-item Disgust Scale (discussed in moredetail later) that correlate most highly with thetotal score are about contact with dead bodies(Haidt et al., 1994). The prototypical odor ofdisgust is the odor of decay, which is the odorof death. The centrality of death in disgust sug-gests a more general construal of disgust withina modified psychoanalytic framework.

Becker (1973) has argued that the most im-portant threat to the psyche is not sexualityand aggression, but the certainty of death.Only human animals know they are to die, andonly humans need to repress this threat. In thisframework, Becker’s “denial of death” isserved by disgust, which helps to suppressthoughts or experiences that suggest humanmortality. Research on terror management the-ory has shown a strong connection betweendisgust and the fear of death: People who areasked to imagine their own death later show anincrease in disgust sensitivity, and an increasein liking for an essay that argues for humanuniqueness rather than human continuitywith other animals (Goldenberg et al., 2001).Conversely, exposure to disgusting stim-uli, under some conditions, increases im-plicit death-related ideation (Cox, Goldenberg,Pyszczynski, & Weise, 2007).

These speculations about death lead to anoverarching description of disgust elicitors:Anything that reminds us that we are animalselicits disgust (Rozin & Fallon, 1987). Humansmust eat, excrete, and have sex, just like otheranimals. Each culture prescribes the properway to perform these actions—for example, byplacing most animals off limits as potentialfoods, and all animals and most people off lim-its as potential sexual partners. People who ig-nore these prescriptions are reviled as disgust-ing and animal-like. Furthermore, humans arelike animals in having fragile body envelopesthat, when breached, reveal blood and soft vis-cera that display our commonalities with ani-mals. Human bodies, like animal bodies, die.Envelope violations and death are disgustingbecause they are uncomfortable reminders ofour animal vulnerability. Finally, hygienic rulesgovern the proper use and maintenance of thehuman body, and the failure to meet these cul-turally defined standards places a person below

47. Disgust 761

the level of humans. Animals are (often inap-propriately) seen as dirty and inattentive to hy-giene. Insofar as we humans behave like ani-mals, the distinction between humans andanimals is blurred, and we see ourselves as low-ered, debased, and mortal.

Elias (1939/1978), in The History of Man-ners, concludes that “people, in the course ofthe civilizing process, seek to suppress in them-selves every characteristic that they feel to be‘animal’ ” (p. 120). Tambiah (1969) empha-sizes the importance of this distinction for hu-mans, and points to the paradox of human fas-cination with and aversion to animals. Ortner(1973) notes that the one body product thatdoes not reliably elicit disgust is tears, andthese are seen as uniquely human. And Leach(1964) has pointed out that animal words areused as insults in many cultures. In general, theethnographic literature is filled with referencesto the fact that humans consider themselvesbetter than animals, and work to maintain aclear animal–human boundary. Violations ofthat boundary—for example, treating an ani-mal as a person in a pet relationship—arerather rare cross-culturally.

W. I. Miller’s (1997) broad, historically basedconception of disgust comes to a conclusionlike ours: “ultimately the basis for all disgust isus—that we live and die and that the process isa messy one emitting substances and odors thatmake us doubt ourselves and fear our neigh-bors” (p. xiv).

INTERPERSONAL DISGUST

The fact that contact with other people canelicit disgust was noted by Darwin (1872/1965). Furthermore, Angyal (1941) noted thatother persons, as containers of waste products,are potentially disgusting. There is widespreadevidence in the United States for aversion tocontact with possessions, utensils, clothing,cars, and rooms used by unknown or unde-sirable persons (Rozin et al., 1989; Rozin,Markwith, & McCauley, 1994). Interpersonalaversion can be analyzed into four separatelyidentifiable components: strangeness, disease,misfortune, and moral taint (Rozin, Markwith,& McCauley, 1994). Thus a sweater worn onceby a healthy stranger and then laundered is lessdesirable than an unworn sweater (aversion tostrangeness). This negativity is substantially en-hanced if the stranger has had a misfortune

(e.g., an amputated leg), a disease (e.g., tuber-culosis), or a moral taint (e.g., a conviction formurder). These types of contacts are both of-fensive and contaminating; thus they seem tobe instances of disgust.

Interpersonal disgust clearly discouragescontact with other human beings who are notintimates. This is probably adaptive by reduc-ing an infection risk, and can serve the purposeof maintaining social distinctiveness and socialhierarchies. In Hindu India, interpersonal con-tagion, mediated in part by contacts with food,is a major feature of society and a major ba-sis for the maintenance of the caste system(Appadurai, 1981; Marriott, 1968).

MORAL DISGUST

Studies that ask people to recall times theywere disgusted elicit stories that often focus onmoral violations, and that involve high levels ofanger as well (Haidt et al., 1997; Izard, 1977;Nabi, 2002; Scherer, 1997). Some of these sto-ries involve issues of sexuality, gore, or otherinstances of the misuse or abuse of human bod-ies, and are thus consistent with animal-naturedisgust. However, many of the stories peopletell about disgust do not involve the body at all;for North Americans, they often involve suchissues as betrayal, hypocrisy, and racism. Dothese disembodied moral violations really elicitdisgust?

Nabi (2002) has argued that they do not. Shefound that the phrase “grossed out” did an ex-cellent job of eliciting stories about core dis-gust, but the words “disgust” and “disgusted”elicited stories that were closer to those elicitedby “anger” than those elicited by “grossedout.” She suggests that the lay understandingof the word “disgust” is a mixture of disgustand anger, and that researchers should be care-ful about following ordinary language by as-suming that the blend is a single emotion. Inshort, disgust is really about bodily issues, andthe moral part of moral disgust is really anger(see also Royzman & Sabini, 2001).

We agree that the lay use of emotion wordscan be misleading, particularly in cross-culturalresearch (Haidt & Keltner, 1999). But there isevidence to suggest that cases of apparentmoral disgust are really disgust; they are notjust linguistic errors made by English speakers.First, if the broad expansion of the word “dis-gusting” into the sociomoral domain is a quirk


of the English language, it is also a quirk of al-most every language we have looked at. Frenchdégoût, German Ekel, Russian otvraschenie,Spanish asco, Hebrew go-al, Japanese ken-o,Chinese aw-shin, and Bengali ghenna all have asemantic domain covering concerns about thebody as well as concerns about other people’ssocial behavior (Haidt et al., 1997). People ofdiverse cultures and languages apparently feelsome similarity in their emotional reactions tofeces and to sleazy politicians.

Second, research in neuroscience is increas-ingly focusing on the anterior insula as acrucial site of “somatic marking” (Damasio,2003), the process by which interoceptive in-formation (gut feelings) meets up with higher-level social cognition to produce motives forsocial approach or avoidance. In a functionalmagnetic resonance imaging (fMRI) study ofpeople playing the “ultimatum game,” thosegiven a very low “take it or leave it” division ofa pool of money usually left it, and their deci-sion was well predicted by a surge of activity inthe anterior insula (Sanfey, Rilling, Aronson,Nystrom, & Cohen, 2003)—the brain areamost often linked to disgust in neuroimagingstudies.

Third, Sherman, Haidt, and Coan (2007)found that in response to a video about Ameri-can neo-Nazis that elicited very high ratings ofdisgust and moderately high ratings of anger,heart rates went down—the expected physio-logical response to disgust, and the opposite ofthe usual response to anger. Furthermore, theheart rate decrement was much larger in thesubset of subjects who reported tightness orclenching in their throats (a marker of core dis-gust). In other words, Nazis really are disgust-ing, at least to some people.

These moral offenses on the outer limits ofdisgust’s expansion show not just the propertyof offensiveness but also the property of con-tamination. Indirect contact with people whohave committed moral offenses (such as mur-ders) is highly aversive, to about the same ex-tent as similar contact with someone with a se-rious contagious illness (Rozin, Markwith, &McCauley, 1994).

Shweder, Much, Mahapatra, and Park(1997) offer a theory of moral judgment thatmay help clarify the moral significance of dis-gust, contempt, and anger (the three other-condemning moral emotions; Izard’s [1977]hostility triad). The theory proposes that threecodes of ethics underlie the morality of most

cultures. One code, called the “ethics of com-munity,” focuses on issues of duty, hierarchy,and the proper fulfillment of one’s social roles.Violations of this code seem to elicit the emo-tion of contempt. A second code, the “ethics ofautonomy,” encompasses issues of rights andjustice. This is the most fully elaborated code inWestern societies, where violations of this codeare usually associated with anger. A third code,the “ethics of divinity,” focuses on the self as aspiritual entity and seeks to protect that entityfrom degrading or polluting acts. We see arough match between Shweder et al.’s threemoral codes and the three other-condemningmoral emotions, with disgust as the emotionelicited by violations of the ethics of divinity,the guardian of the sanctity of the soul as wellas purity of the body. We call this the “CADtriad” hypothesis (community/contempt, au-tonomy/anger, divinity/disgust), and provideevidence supporting it from the correspon-dence between the three moral codes and thethree emotions in Japanese and Americans(Rozin, Lowery, Imada, & Haidt, 1999).

Disgust plays a special role in the moral do-main as a means of socialization. Insofar as en-tities viewed as immoral are also disgusting,there is no temptation to have traffic withthem. For example, as cigarette smoking hasmoved from being a preference to a negativemoral value in modern North America, there isan accompanying increase in disgust responsesto cigarettes, cigarette smoke, cigarette residues(e.g., ashes), and cigarette smokers (Rozin &Singh, 1999). This process of conversion of anentity from a preference into a value has beencalled “moralization” (Rozin, 1997). It is oftenassociated with the recruitment of a disgust re-sponse to the entity or activity in question.

PREADAPTATIONAND THE CULTURAL EVOLUTIONOF DISGUST

We believe that the output side of disgust(physiology, behavior, expression) has re-mained relatively constant over human history,and that it still bears noticeable similarities toits animal precursors. However, the input side(elicitors and meanings) has been transformedand greatly expanded.

We have suggested a course of biological andcultural evolution of disgust, summarized inTable 47.1 (Rozin et al., 1993, 1997). The pro-

47. Disgust 763

posed origin is the rejection response to bad-tasting foods, even though taste in the mouthultimately has little to do with the emotion ofdisgust. However, oral rejection remains an or-ganizing principle of disgust reactions, in whatwe have called “core disgust.” Core disgust canbe thought of as a guardian of the mouth, andtherefore as a guardian of the physical body.Food and its potential contaminants (bodyproducts and some animals) are the elicitors forcore disgust.

Disgust then expanded further to become aguardian of the temple of the body, respondingto direct threats of contagion or infection toparts other than the mouth, and also to any evi-dence that our bodies are really no differentfrom animal bodies (i.e., animal-nature disgustin the domains of sexuality, body envelope vio-lations, death, and hygiene). Driving a desire todistinguish ourselves from animals may be ourfear of animal mortality.

Interpersonal disgust and moral disgust arenot easily accounted for as reminders of ouranimal nature. They may both be linked to theprior forms of disgust, because they are exten-sions of a disease avoidance mechanism to be-come a broader social avoidance mechanism.This model suggests what might be called anopportunistic accretion of new domains ofelicitors to a rejection system that is already inplace. A parallel to this model in evolutionarybiology is the concept of “preadaptation”(Mayr, 1960). Mayr suggests that the majorsource of evolutionary “novelties” is the coopt-ing of an existing system for a new function.Preadaptation can operate either to replace anoriginal function, or to accrete new functionsto an existing system. A particularly appropri-ate example is the human mouth, whose teethand tongue clearly evolved for food handling.However, by a process of preadaptation, they

have come to be shared by the language expres-sion system. Teeth and tongue are critical inpronunciation, but they did not evolve for thatpurpose. We suggest that in both cultural evo-lution and individual development, a processlike preadaptation occurs; in development, itcan be described as the accessing of previouslyinaccessible systems for a wider range of activi-ties, functions, or elicitors (Rozin, 1976).

We have described the cultural evolution ofdisgust as a sequence of stages that takes dis-gust further and further away from its mouth-and-food origins, through a process ofpreadaptation. But it has not really expandedthat far beyond food, because, by a parallelprocess of preadaptation, food itself has cometo serve many functions—aesthetic, social, andmoral—besides its original nutritive function(Kass, 1994). In parallel, the food vocabularyhas taken on other, metaphorical functions,again by a process of preadaptation. Thus thevery words “taste” and “distaste” have cometo indicate general aesthetic judgments. InHindu India, food and eating are quintessen-tially social and moral activities (Appadurai,1981).

The latter part of the 19th century includedtwo events with important impact on the evo-lution of disgust in the Western world: Dar-win’s theory of evolution (which blurred thehuman–animal distinction) and the rise of germtheory following on the work of Pasteur andothers. Detailed analyses of lay and elite think-ing about disgust during the 19th century havebeen provided in two cultural histories ofFrance (Corbin, 1986; Barnes, 2006). Prior tothe development of germ theory, the Frenchhad already adapted to serious contagion risksby showing disgust to and avoidance of odorsof decay and contact with ill individuals. Quar-antine was in practice before germ theory, and


TABLE 47.1. Proposed Pathway of Expansion of Disgust and Disgust Elicitors

Disgust stage

0. Distaste 1. Core 2. Animal nature 3. Interpersonal 4. Moral

Function Protect bodyfrom poison

Protect body fromdisease/infection

Protect body andsoul; denymortality

Protect body,soul, and socialorder

Protect socialorder

Elicitors Bad tastes Food/eating, bodyproducts, animals

Sex, death,hygiene, envelopeviolations

Direct andindirect contactwith strangers orundesirables

Certain moraloffenses

odor was believed to be part and parcel of thesource of illness, rather than a correlate of it.Germ theory thus provided scientific justifica-tion for disgust sensibilities that were alreadypresent, and germ theory and disgust were bothadvanced in what Barnes (2006) has describedas the marriage of lay contagion beliefs andgerm theory—a “sanitary–bacteriological syn-thesis.”

THE DEVELOPMENT OF DISGUST

For adults, feces seems to be a universal disgustsubstance (Angyal, 1941; Rozin & Fallon,1987), with the odor of decay as perhaps themost potent sensory attribute associated withdisgust. It is also conceivable that vomit isa primary substance for disgust. Since feces,vomit, and decay are associated with diseasevectors, it would be reasonable to suppose thatthere would be an innate rejection of suchthings. However, neither one seems to be reli-ably rejected by nonhuman animals or youngchildren (Rozin, Hammer, Oster, Horowitz, &Marmara, 1986). Rather, it appears that in-fants may be attracted to feces, and that disgustis a powerful cultural force that turns thisattraction into aversion (Freud, 1910/1957;Jones, 1912/1948). The preponderance of evi-dence suggests that there are no innately nega-tive nonirritant odors, and that a rejection ofdecay odors (without a referent object present)appears somewhere between 3 and 7 years ofage (Petó, 1936; Schmidt & Beauchamp, 1988;Stein, Ottenberg, & Roulet, 1958; but seeSteiner, 1979).

As far as we know, there is no sense of of-fensiveness or rejection outside of the sen-sory realm in either infants or nonhumans,and hence no gape elicitors other thannegative tastes. Disgust seems to requireenculturation—a supposition confirmed byMalson’s (1964/1972) review of some 50 feralhumans, none of whom showed any sign of dis-gust.

Toilet training is probably the initial disgust-generating experience. For 3-year-olds, fecesare rejected, but not contaminating and possi-bly not offensive. In the period following toilettraining, children develop an aversion for sub-stances resembling feces (e.g., mud, dirt, andmushy substances) and sometimes a markedconcern for cleanliness (Senn & Solnit, 1968;Ferenczi, 1914/1952).

Further extension of aversion to core disgustelicitors is likely to occur either by generaliza-tion from existing disgusting entities (e.g., fromfeces to mud) or by evaluative conditioning (al-though attempts to capture this phenomenonin the laboratory have yielded mixed outcomes;see Schienle, Stark, & Vaitl, 2001; Rozin,Wrzesniewski, & Byrnes, 1998). There may bea predisposition or expectancy to associate cer-tain entities, such as certain types of ani-mals, with already disgusting entities (Davey,Cavanagh, & Lamb, 2003). Disgust may be ac-quired by witnessing facial displays of emo-tions that elicit the experience of those emo-tions (Tomkins, 1963), perhaps engagingprocesses that involve mirror neurons (Gallese,Keysers, & Rizzolatti, 2004).

Locating the onset of true disgust in develop-ment depends on subtle measures of “offen-siveness” or “ideational rejection” and theappearance of contamination sensitivity. Con-tamination sensitivity is not present in childrenunder 3–5 years of age (Fallon, Rozin, &Pliner, 1984; Rozin, Fallon, & Augstoni-Ziskind, 1985; Siegal, 1988; Siegal & Share,1990; Hejmadi, Rozin, & Siegal, 2004). Con-tamination sensitivity is a sophisticated ability,requiring a separation of appearance and real-ity. There is no sensory residue of past contami-nation in a contaminated entity; it is the historyof contact that is critical (Rozin & Nemeroff,1990; Nemeroff & Rozin, 2000). Furthermore,contamination implies some conception of in-visible entities (e.g., traces of cockroach) thatare the vehicle of contamination. The notionof invisible entities and the notion that ap-pearance is distinct from reality are cognitiveachievements of considerable abstraction, andboth seem to be absent in young children(Piaget & Inhelder, 1941/1974; Flavell, 1986;Rosen & Rozin, 1993; but see Siegal & Share,1990). This cognitive limitation may be theprincipal barrier to a full childhood acquisitionof disgust.

Adult contamination sensitivity is a mixtureof at least two types of conceptions. One in-volves transfer of invisible material throughcontact, and hence is often sensitive to manipu-lations like washing (material essence). A sec-ond is more indelible and involves the passingof some type of “spiritual” force that is notsubject to removal by chemical and physicaltreatments (“spiritual essence”; Nemeroff &Rozin, 1994). There is evidence that at its firstappearance in children, the essence producing

47. Disgust 765

contamination sensitivity is more like the in-delible, “spiritual” than the material form(Hejmadi et al., 2004), and it appears some-what earlier and in greater intensity in HinduIndian children than in American children.

A measure of contamination sensitivity thatfocused on disgust contaminants showed a sub-stantial correlation between young adults andtheir parents: r = .52 for Americans (Rozin,Fallon, & Mandell, 1984); r = .33 in Britain(Davey, Forster, & Mayhew, 1991). A study us-ing the broader Disgust Scale (see below) founda more modest correlation (r = .21) across threegenerations of Americans (Rozin & Wolf,2007).

CULTURAL DIFFERENCESIN DISGUST

Almost the entire literature on disgust comesfrom the approximately 6% of the world inwhich English is the native language. We be-lieve that the cultural evolution of disgust hasmade few changes on the output side, as notedabove, but that it has created substantial cul-tural variation on the input side. The simplestvariations can be seen when cultures differ inthe particular elicitors of disgust within one ofthe domains we have described. For example,most cultures value some kind of decayed/fer-mented food that is disgusting in most othercultures, but such food varies quite a bit (e.g.,cheese for Europeans, decayed meat for Inuit,fermented fish sauce for Southeast Asians).Similarly, cultures differ about whether dogsare best friends or dirty scavengers, aboutwhether or not corpses should be touched dur-ing mourning, or about whether mouth-to-mouth kissing is erotic or disgusting.

It is primarily in the last two steps of the ex-pansion of disgust—interpersonal and moraldisgust—that cultural differences seem to begreatest. Interpersonal and moral disgust ap-pear to be particularly elaborated in Hindu In-dia, compared to Western nations in whichpeople rarely worry about the caste or back-ground of the people cooking their food. Purityis a moral virtue to be protected in India, and inthis respect, food is a “biomoral” substance(Appadurai, 1981). Moral disgust in the UnitedStates seems to focus on acts that strip others oftheir basic humanity (of their souls, Bloom[2004] would say; e.g., acts of brutality, cruelty,and racism), as well as to more mundane acts

of sleaziness and insincerity (such as hypocrisyand fawning). In Japan, participants appliedthe word ken-o more to situations in whichthere had been a failure to achieve a good fit insocial relationships, such as when somebodyelse ignored them or criticized them unfairly.American moral disgust may be guardingagainst threats to an individualist, rights-basedsocial order, whereas Japanese ken-o may beguarding against threats to a more collectivist,interdependent social order (Haidt et al.,1997).

An additional cultural difference is found inthe moral significance attached to the activitiesthat disgust regulates. Haidt, Koller, and Dias(1993) asked North Americans and Braziliansof higher and lower socioeconomic statusabout a number of actions that were disgustingyet harmless, including incestuous kissing, eat-ing one’s dead pets, and eating a chicken onehas just had sex with. They found that NorthAmericans of high socioeconomic status sep-arated their emotional reactions from theirmoral judgments, while other groups weremore likely to condemn disgusting actions,even when they were harmless. W. I. Miller(1997) suggests that many Westerners may beuncomfortable using disgust as a moral emo-tion, because it is often at odds with our egali-tarian ethos: Disgust puts people down, and itis easily used to condemn people who areobese, are deformed, or have sexual prefer-ences at odds with the majority. It is partly forthis reason that Nussbaum (1999) argues thatdisgust should play little or no role in the legalsystem or the legislature. She disagrees stronglywith Kass’s (1997) claim that disgust some-times embodies “wisdom” about being humanthat is difficult for us to articulate rationally.

INDIVIDUAL DIFFERENCESIN DISGUST SENSITIVITY

Based on our theorizing about the expansion ofdisgust, we created a paper-and-pencil measureof individual differences in disgust sensitivitytoward seven kinds of elicitors (Haidt et al.,1994). Three were core disgust elicitors (food,animals, and body products), and four wereanimal-nature disgust elicitors (inappropriatesexuality, envelope violations, death, and poorhygiene). We also included an eighth categoryof magical thinking across the various kinds ofelicitors. The 32-item Disgust Scale (DS) in-


cluded two true–false and two disgust-ratingitems for each of these eight categories. The DShas an overall alpha of .84, and it has beenshown to predict hands-on disgust-relevantbehavior among Americans (Rozin, Haidt,McCauley, Dunlop, & Ashmore, 1999), andSwedes (Björklund & Hursti, 2004), and topredict the behavior of people with obsessive–compulsive symptoms even after other nega-tive affects were controlled for (Olatunji, Lohr,Sawchuck, & Tolin, 2007). It also predicts thedegree to which brain regions associated withdisgust are activated when subjects look at dis-gusting pictures in an fMRI scanner (Caseras etal., 2007).

Studies using the DS have consistently foundthat women score higher than men (Haidt etal., 1994), that disgust sensitivity declines withage in adulthood (Quigley, Sherman, &Sherman, 1996; Fessler & Navarette, 2005),and that it may decline faster for women thanfor men (Doctoroff & McCauley, 1996). Dis-gust sensitivity is inversely related to educationand socioeconomic status (Doctoroff &McCauley, 1996). Personality correlates of theDS are consistent with disgust’s role as an in-hibitor of approach and consumption. It corre-lates positively with scales related to anxiety(e.g., Big Five neuroticism), particularly anxi-eties related to mortality and bodily concerns(e.g., blood–injection phobia, contaminationfears), and it correlates negatively with scalesrelated to sensation seeking and openness toexperience. (See Haidt et al., 1994, for a list ofarticles reporting findings using the DS; seepeople.virginia.edu/~jdh6n/disgustscale.html)for information about the scale itself.

Our goal in creating the DS was not toachieve a high alpha by selecting items thatwere similar to each other. Rather, it was to cre-ate a broad instrument that would allow re-searchers to look for relationships betweensubtypes of disgust and various other behaviorsand clinical conditions. For example, the sexand death subscales are good predictors of reli-gious obsessions even after fearfulness is con-trolled (Olatunji, Tolin, Huppert, & Lohr,2005), as terror management theorists wouldexpect.

It is now clear, however, that our originalsubscales are not reliably discriminated in psy-chometric analyses. Rather, the 32 items in-clude just three psychometrically stable factors(Olatunji et al., 2007): core disgust (most of thefood, animal, and body product items); animal-

reminder disgust (most of the death and en-velope violation items), and interpersonal-contamination disgust (items in which the in-terpersonal nature of the contact is salient).These three subscales have much higher inter-nal consistency (alphas above .70) than thefour-item scales of the original DS.

In a series of studies, Olatunji and his col-laborators have found that the core andinterpersonal-contamination subscales predictclinical conditions, such as the contamina-tion aspects of obsessive–compulsive disorder(OCD) (Olatunji, Williams, Lohr, & Saw-chuck, 2005). The animal-reminder scale pre-dicts clinical conditions such as blood–injection–injury fears (Olatunji, Sawchuck, deJong, & Lohr, 2006).

In recent years, two additional measures ofdisgust sensitivity have been published. TheDisgust Propensity and Sensitivity Scale—Revised (Van Overveld, de Jong, Peters,Cavanagh, & Davey, 2006) offers one subscalethat assesses frequency of disgust experiences,and another that assesses the degree to whichdisgust experiences are upsetting. The DisgustEmotion Scale (Kleinknecht, Kleinknecht, &Thorndike, 1997) offers five subscales to assessdisgust toward rotten foods, small animals, in-jections and blood draws, mutilation anddeath, and bad smells. The development ofthese two scales was motivated in part by thelow internal reliabilities of the eight DS sub-scales. A similar concern led to Olatunji et al.’s(2007) modification of the DS to create the 25-item, three-factor DS-R.

DISGUST AND THE BRAIN

Since the early 1990s, there has been a greatdeal of research on the neural correlates of dis-gust. A main concern of this research is to iden-tify areas of the brain that are activated by ormediate disgust. An early study (confirmed bysome later work) demonstrated that disgust ex-periences are associated with increased activityin the right frontal cortex, a broad region asso-ciated with negative affect (Davidson, 1992).More specific is the finding that people withHuntington’s disease, caused by late-onset de-generation of the basal ganglia, show a remark-ably specific deficit in identifying disgust facialexpressions (Sprengelmeyer et al., 1996;Sprengelmeyer, Rausch, Eysel, & Przuntek,1998) that may extend to other modalities

47. Disgust 767

(e.g., Mitchell, Heims, Neville, & Rickards,2005). Disgust recognition deficits are alsoseen in people who have the Huntington geno-type, but are still too young to show any of theclassical symptoms (Gray, Young, Barker,Curtis, & Gibson, 1997).

A good part of the neuroimaging literaturehas aimed to establish a link among OCD, dis-gust, and the basal ganglia or insula (Shapira etal., 2003; Phillips et al., 2000; for a review, seeHusted et al., 2006). Dozens of experiments us-ing imaging techniques (principally fMRI), be-ginning with a study by Phillips et al. (1997; seealso Wicker et al., 2003; Wright, He, Shapira,Goodman, & Liu, 2004), have suggested thatthree interconnected brain areas are character-istically activated when there is exposure todisgust faces, disgust-eliciting images, disgust-related odors, or thoughts about disgusting en-tities. The areas are the anterior insula, thebasal ganglia, and parts of the prefrontal cor-tex (for a review, see Husted et al., 2006). Theevidence for a linkage between the anteriorinsula and basal ganglia and disgust is con-firmed by a demonstration that a patient withdamage to both areas showed a selective im-pairment in both the recognition and experi-ence of disgust (Calder, Keane, Manes, Antoun,& Young, 2000). However, one line of researchquestions whether insula and basal ganglia ac-tivation distinguishes disgust from fear (e.g.,Stark et al., 2003; Schienle, Schafer, Stark, Wal-ter, & Vaitl, 2005).

Research on the neural basis of disgust canalso provide evidence related to psychologicalissues. First, the relation between the anteriorinsula and disgust supports the idea that dis-gust has particularly strong links to food andeating, since the insula is part of the gustatorycortex and is activated by unpleasant tastes andsmells (Rolls, 1994). Second, with regard to thecontroversy about whether moral disgust is re-ally disgust, Moll et al. (2005) have reportedoverlap in brain areas activated by core oranimal-nature disgust elicitors and by moraldisgust elicitors.

DISGUST AND PSYCHOPATHOLOGY

Interest in disgust in relation to psychopatholo-gy has increased dramatically since 1990.Davey and his colleagues (Matchett & Davey,1991; Davey, 1993) provided evidence for alink between disgust and some types of animal

phobias, and Power and Dalgleish (1997) pro-posed links between disgust and manypsychopathologies, including depression. Acentral role for disgust in anxiety disorders(specifically, phobias and OCD; see below) hasbeen described in special issues of the Journalof Anxiety Disorders (McKay, 2002) and theJournal of Behavior Therapy and ExperimentalPsychiatry (Olatunji & McKay, 2006), aswell as in three review articles (Woody &Teachman, 2000; Berle & Phillips, 2006;Olatunji & Sawchuk, 2005). A plausible link-age between disgust and eating disorders hasalso been proposed (Quigley et al., 1996;Davey, Buckland, Tantow, & Dallos, 1998;Troop, Treasure, & Serpell, 2002). Further-more, relations between increased disgust sen-sitivity and schizophrenia (Schienle et al.,2003) and hypochondriasis (Davey & Bond,2006) have been reported.

The relations between disgust and two majorcategories of anxiety disorders—phobias (par-ticularly spider and blood–injury phobia) andOCD—have received most attention. Sevenkinds of evidence have been advanced.

First is identification of disgust in the symp-toms of the disorder. OCD has an obvious linkwith disgust, since a common form of OCD in-volves excessive cleaning and washing based onfears of contamination. Hypersensitivity to dis-gust (in all or some domains) is one possible ac-count of OCD, as is a potentially separable hy-persensitivity to contamination (for reviews,see Berle & Phillips, 2006; Husted et al., 2006;and Olatunji & Sawchuk, 2005). Similarly,phobias often overtly include feelings of disgust(Davey, 1993; Matchett & Davey, 1991;Woody, McLean, & Klassen, 2005; Woody &Teachman, 2000).

Second is correlation of disgust sensitivitywith the disorder. For OCD, both the clinicallydefined disorder and individual differences inOCD tendency have been found to be associ-ated with disgust sensitivity (e.g., Quigley etal., 1996; Rozin, Taylor, Ross, Bennett, &Hejmadi, 2005; Tolin, Woods, & Abramowitz,2006). Findings are quite consistent, withcleaning OCD more closely related to disgustsensitivity than other forms of OCD (e.g.,Mancini, Gragnani, & D’Olimpio, 2001).There are indications that sensitivity to inter-personal contamination and concern with hy-giene may be specifically related to cleaningOCD (Olatunji, Williams, Lohr, & Sawchuk,2005; Tolin et al., 2006). Disgust sensitivity is


also generally higher in individuals with small-animal or blood–injury phobias (Tolin, Lohr,Sawchuk, & Lee, 1997; Koch, O’Neill,Sawchuk, & Connolly, 2002), but there is stillsome question as to whether this is general dis-gust sensitivity or sensitivity in particular do-mains (for a review, see Woody & Teachman,2000). There are also important questionsabout distinguishing between state and traitdisgust (e.g., Woody & Tolin, 2002; Woody etal., 2005) as the principal correlates of pho-bias.

Third is demonstration that recognition ofdisgust faces or disgust situations is enhancedor compromised in the disorder. Whereas thereare abundant data on disgust face recognitionin Huntington’s disease (Sprengelmeyer et al.,1996), there are no parallel data for phobias,and there are mixed reports on disgust face rec-ognition in OCD.

Fourth is demonstration of deficits or en-hancements in disgust expressions (facial orotherwise). There are no data on this point forOCD, but there is evidence indicating en-hanced disgust facial expressions for personswith animal phobias in the presence of phobicelicitors (Lumley & Melamed, 1992; Schienle,Schafer, Walter, Stark, & Vaitl, 2005).

Fifth is demonstration of a parallelpsychophysiology in disgust and the disorder inquestion. The signature physiological sign ofdisgust, nausea, has not been studied in thiscontext for either phobias or OCD. Fainting,associated with parasympathetic activation, isa frequent feature of blood–injury phobia, anda link between disgust and fainting in blood–injury phobia has been reported (Page, 1994,2003).

Sixth is demonstration that disgust stimuliactivate (or inhibit) brain areas associated withthe disorder. Supportive data exist for OCDand disgust: There are signs of an overlap in ac-tivated brain areas, with the anterior insulaprominently engaged in both (e.g., Shapira etal., 2003; Husted et al., 2006).

Seventh is a more refined analysis thatprobes whether disgust is a direct cause, anoncausal correlate, an indirect cause, or aconsequence of the disorder. In the case of pho-bias, the issue has largely been whether disgustincreases symptoms directly, or increases themindirectly by increasing anxiety. Davey and hiscolleagues have addressed this issue with nor-mal individuals, showing that while inducedanxiety increases disgust, induced disgust does

not increase anxiety (Marzillier & Davey,2005). The same group has shown that experi-enced disgust causes a negative interpretationalbias that may enhance threat perceptions, andthat this effect is not fully accounted for by adisgust–anxiety linkage (Davey, Bickerstaffe, &MacDonald, 2006; but see Sawchuk, Meunier,Lohr, & Westendorf, 2002). Davey and Bond(2006) have dissected independent roles fortrait anxiety and trait disgust in hypochon-driasis and health anxiety. For acquisition ofdisgust-related phobias, a likely mechanism isevaluative conditioning (Schienle et al., 2001).

The psychopathology literature has under-standably focused on above-normal activationof disgust (and other emotions). However, weclose this section by noting that very low acti-vation of disgust may generate a highly antiso-cial person, since disgust is in many respectsthe emotion of civilization. Scores of non-disordered subjects on the DS correlate nega-tively with scores on the psychopathy subscaleof the Eysenck Personality Questionnaire(Eysenck & Eysenck, 1975); however, it is notknown whether persons with true psychopathyscore lower than nondisordered individuals onmeasures of disgust sensitivity (J. Blair, per-sonal communication, February 18, 2007).

THE DELICATE BOUNDARYBETWEEN DISGUST AND PLEASURE

Given that the human body is a repository ofdisgusting entities, and that we live in a con-taminated environment, humans are frequentlypoised on the edge of potential disgust. Whilewe manage most of the time, by habituation orframing (e.g., Rozin, in press), to ignore thedisgust elicitors all around us, there are somesituations in which we seek out and enjoy dis-gust. The powerful negativity of disgust seemsperversely to encourage its involvement and en-joyment in at least two domains: humor andromantic attachments.

Disgust plays a significant role in humor, viajokes, cartoons, and casual word play. Disguststimuli often elicit amusement. Generally, itseems that disgust can be amusing when it isnot personally threatening; when a person informal wear (other than the self) steps in dogfeces, it is amusing. Disgust plays a central rolein the humor of boys and adolescent males,who use it to tease, to question or confrontadult norms, and to establish status within

47. Disgust 769

their peer groups (Fine, 1988). Bloom (2004)emphasizes the dignity-destroying aspect ofdisgust, and sees disgust humor as taking ad-vantage of the fact that the human body is dis-gusting, at the same time that it abhors disgustand sees itself as at a higher plane of existence.The shift in perspective, from soul to body, is afertile base for humor.

A number of authors, including S. B. Miller(2004), Fessler and Haley (2006), and our-selves (e.g., Rozin et al., 1995) have empha-sized the self-boundary or gateway-guardingfunction of disgust. A distinct self generally re-sists mixing of the self’s substance with the sub-stance of another. Thus treating the other asself, and enjoying what are usually disgustinginteractions such as mouth-to-mouth kissingbetween lovers, is a way of affirming love andintimacy (W. I. Miller, 1997; Bloom, 2004). AsBloom (2004, p. 180) suggests, “In love, yousee the person not as a body, but as a soul.”

DISGUST ININTERGROUP RELATIONS

Recent research in social psychology has givennew attention to intergroup emotions in rela-tion to intergroup conflict and intergroup vio-lence, including feelings of disgust toward anenemy or minority group.

Dehumanization is often cited as part of theexplanation of particularly horrific forms of in-tergroup violence, including genocide (Chirot& McCauley, 2006). In a theoretical analysis ofdehumanization, Haslam (2006) distinguishesbetween “animalistic dehumanization,” whichmakes others less human by making them morelike animals, and “mechanistic dehumaniza-tion,” which makes enemies less human by de-nying them uniquely human emotions andtraits and making them more like machines.Haslam suggests that disgust is the emotionalreaction associated with animalistic dehuman-ization, whereas indifference is the reaction tomechanistic dehumanization.

In a related analysis of perception of out-groups, Fiske, Cuddy, Glick, and Xu (2002)followed Brown (1965) in suggesting that so-cial perceptions can be understood in rela-tion to two dimensions: status and solidarity.Groups seen as low in status and dissimilar toone’s own group (as welfare mothers were seenby participants in the study by Fiske et al.) tendto be viewed with disgust and contempt. Dis-

gust in relation to low-status and dissimilaroutgroups is also implicated in results showingthat disgust sensitivity is positively correlatedwith negative attitudes to foreigners, out-groups, immigrants, and deviant individuals(Faulkner, Schaller, Park, & Duncan, 2004;Hodson & Costello, 2007; Navarette &Fessler, 2006). To some degree, this effect ismediated by fear of infection or contamination(Faulkner et al., 2004; Navarette & Fessler,2006).

Cottrell and Neuberg (2005) have offereda more complex categorization of perceivedthreats and associated emotional reactions tothese threats. Groups representing physical ormoral contaminants (gay men for the partici-pants in this study) elicit disgust, whereasgroups representing barriers to desired goals(fundamentalist Christians, for these partici-pants) elicit anger. These studies have in com-mon that group perceptions are analyzed be-yond a simple dimension of positive or negativeaffect to distinguish the different characteristicsand different threats associated with differentgroups. Associated with these different ap-praisals are different emotional reactions, in-cluding feelings of disgust for groups seen asanimal-like, low-status, and dissimilar. Disgustmay signal a particularly potent threat, thethreat of contamination.

CONCLUSION

Darwin and Angyal offered prescient analysesof the emotion of disgust. Despite this early at-tention to an all-too-common emotional expe-rience, and the accessibility of experimentalmanipulations of this emotion, empirical inves-tigation of disgust has taken off only since1990. As a result, there are many unansweredquestions. We know little about the evolution-ary and ancient history of disgust. It is absentin nonhuman primates, yet extremely frequentand probably universal among contemporaryhumans. We do not know much about the se-quence of events that introduced and expandeddisgust over historical time (but see W. I. Miller,1997, for the most thorough analysis of this ex-pansion for Western cultures). We do not knowwhether the acceptance of the theory of evolu-tion, and hence of human continuity with ani-mals, played a role in the development or ex-pression of animal nature disgust. We do notknow how disgust originates in development;


nor what the principal causes of differences indisgust sensitivity are; nor why it is a focus ofhumor, especially in children. Many of thefundamental questions posed by Darwin andAngyal remain unanswered.

Our analysis suggests a cultural evolution ofdisgust that brings it to the heart of what itmeans to be human. We have suggested thatdisgust originated as a rejection response tobad tastes, and then evolved into a much moreabstract and ideational emotion. In this evolu-tion, the function of disgust shifted: A mecha-nism for avoiding harm to the body became amechanism for avoiding harm to the soul. Theelicitors of disgust may have expanded to thepoint where the only thing they have in com-mon is that decent people want nothing to dowith them. At this level, disgust becomes amoral emotion and a powerful form of nega-tive socialization. We have presented a skeletonof evidence in support of this analysis, butthere are many alternatives and points of diffi-culty. In our view, because cultures have capi-talized on disgust as a way of internalizingsome of their particular negative attitudes,there is no overarching abstract definition ofthe class of disgust elicitors. In the view ofRoyzman and Sabini (2001), the lack of asingle abstract description of disgust elicitorscompromises the status of disgust as an emo-tion. The elicitors of anger and fear, they sug-gest, are more coherent (i.e., insult and threatappraisals, respectively). Our view is that theappraisal that elicits disgust is indeed morecognitively complex than other emotions.There are multiple appraisals, as we haveindicated—some quite abstract, such as re-minders of our animal nature. The complexityof disgust reflects the complexity of a speciesthat is both animal and human.

ACKNOWLEDGMENTS

We thank the Whitehall Foundation for supportingsome of the research reported in this chapter and thepreparation of the original (1993) version of thischapter, and the Edmund J. and Louise W. KahnChair for Faculty Excellence Fund for supporting thepreparation of this revision.

REFERENCES

Angyal, A. (1941). Disgust and related aversions. Jour-nal of Abnormal and Social Psychology, 36, 393–412.

Appadurai, A. (1981). Gastro-politics in Hindu SouthAsia. American Ethnologist, 8, 494–511.

Barnes, D. S. (2006). The great stink of Paris and themarriage of filth and germs. Baltimore: JohnsHopkins University Press.

Becker, E. (1973). The denial of death. New York: FreePress.

Berle, D., & Phillips, E. S. (2006). Disgust andobsessive–compulsive disorder: An update. Psychia-try, 69, 228–238.

Björklund, F., & Hursti, T. (2004). A Swedish transla-tion and validation of the Disgust Scale: A measure ofdisgust sensitivity. Scandinavian Journal of Psychol-ogy, 45, 279–284.

Bloom, P. (2004). Descartes’ baby: How the science ofchild development explains what makes us human.New York: Basic Books

Brown, R. (1965). Social psychology. New York: FreePress.

Calder, A. J., Keane, J., Manes, F., Antoun, N., &Young, A. W. (2000). Impaired recognition and expe-rience of disgust following brain injury. Nature Neu-roscience, 3, 1077–1088.

Caseras, X., Mataix-Cols, D., An, S. K., Lawrence, N.S., Speckens, A., Giampietro, V., et al. (2007). Sexdifferences in neural responses to disgusting visualstimuli: Implications for disgust-related psychiatricdisorders. Biological Psychiatry, 62, 464–471.

Chirot, D., & McCauley, C. (2006). Why not kill all ofthem?: The logic and prevention of mass politicalmurder. Princeton, NJ: Princeton University Press.

Corbin, A. (1986). The foul and the fragrant: Odor andthe French social imagination. Cambridge, MA: Har-vard University Press.

Cottrell, C. A., & Neuberg, S. L. (2005). Different emo-tional reactions to different groups: A sociofunc-tional threat-based approach to “prejudice.” Jour-nal of Personality and Social Psychology, 88, 770–789.

Cox, C. R., Goldenberg, J. L., Pyszczynski, T., & Weise,D. (2007). Disgust, creatureliness and the accessibil-ity of death-related thoughts. European Journal ofSocial Psychology, 37, 494–507.

Curtis, V., Aunger, R., & Rabie, T. (2004). Evidence thatdisgust evolved to protect from risk of disease. Pro-ceedings of the Royal Society of London, Series B,271(Suppl.), S131–S133.

Curtis, V., & Biran, A. (2001). Dirt, disgust, and dis-ease: Is hygiene in our genes? Perspectives in Biologyand Medicine, 44, 17–31.

Damasio, A. (2003). Looking for Spinoza. Orlando, FL:Harcourt.


Davey, G. C. L. (1993). Factors influencing self-ratedfear to a novel animal. Cognition and Emotion, 7,461–471.

Davey, G. C. L., Bickerstaffe, S., & MacDonald, B. A.(2006). Experienced disgust causes a negative inter-

47. Disgust 771

pretation bias: A causal role for disgust in anxiouspsychopathology. Behaviour Research and Therapy,44, 1375–1384.

Davey, G. C. L., & Bond, N. (2006). Using controlledcomparisons in disgust psychopathology research:The case of disgust, hypochondriasis and health anxi-ety. Journal of Behavior Therapy and ExperimentalPsychiatry. 37, 4–15.

Davey, G. C. L., Buckland, G., Tantow, B., & Dallos, R.(1998). Disgust and eating disorders. EuropeanEating Disorders Review, 6, 201–211.

Davey, G. C. L., Cavanagh, K., & Lamb, A. (2003). Dif-ferential aversive outcome expectancies for high- andlow-predation fear-relevant animals. Journal ofBehavior Therapy and Experimental Psychiatry, 34,117–128.

Davey, G. C. L., Forster, L., & Mayhew, G. (1991). Fa-milial resemblance in disgust sensitivity and animalphobias. Behaviour Research and Therapy, 31, 41–50.

Davidson, R. J. (1992). Emotion and affective style:Hemispheric substrates. Psychological Science, 3,39–43.

Doctoroff, G., & McCauley, C. (1996, March). Demo-graphic differences in sensitivity to disgust. Posterpresented at the annual meeting of the Eastern Psy-chological Association, Philadelphia.

Douglas, M. (1966). Purity and danger. London: Rout-ledge & Kegan Paul.

Ekman, P. (1972). Universals and cultural differences infacial expressions of emotion. In J. K. Cole (Ed.), Ne-braska Symposium on Motivation (Vol. 19, pp. 207–283). Lincoln: University of Nebraska Press.


Ekman, P., & Friesen, W. V. (1975). Unmasking theface. Englewood Cliffs, NJ: Prentice Hall.

Ekman, P., & Friesen, W. V. (1978). Facial ActionCoding System: A technique for the measurement offacial movement. Palo Alto, CA: Consulting Psychol-ogists Press.

Elias, N. (1978). The history of manners: Vol. 1. Thecivilizing process (E. Jephcott, Trans.). New York:Pantheon Books. (Original work published 1939)

Eysenck, H. J., & Eysenck, S. B. G. (1975). The EysenckPersonality Questionnaire. London: Hodder &Stoughton.

Fallon, A. E., & Rozin, P. (1983). The psychologicalbases of food rejections by humans. Ecology of Foodand Nutrition, 13, 15–26.

Fallon, A. E., Rozin, P., & Pliner, P. (1984). The child’sconception of food: The development of food rejec-tions with special reference to disgust and contami-nation sensitivity. Child Development, 55, 566–575.

Faulkner, J., Schaller, M., Park, J. H., & Duncan, L. A.(2004). Evolved disease-avoidance mechanisms andcontemporary xenophobic attitudes. Group Pro-cesses and Intergroup Relations, 7, 333–353.

Ferenczi, S. (1952). The ontogenesis of the interest inmoney. In S. Ferenczi (Ed.) & E. Jones (Trans.), First

contributions to psychoanalysis (pp. 319–331). Lon-don: Hogarth Press. (Original work published 1914)

Fessler, D. M. T., & Haley, K. J. (2006). Guarding theperimeter: The outside–inside dichotomy in disgustand bodily experience. Cognition and Emotion, 20,3–19.

Fessler, D. M. T., & Navarrete, C. D. (2005). The effectof age on death disgust: A critique of terror manage-ment perspectives. Journal of Evolutionary Psychol-ogy, 3, 279–296.

Fine, G. A. (1988). Good children and dirty play. Playand Culture, 1, 43–56.

Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002).A model of (often mixed) stereotype content: Compe-tence and warmth. Journal of Personality and SocialPsychology, 82, 878–902.

Flavell, J. (1986). The development of children’s knowl-edge of the appearance–reality distinction. AmericanPsychologist, 41, 418–425.

Frazer, J. G. (1922). The golden bough: A study inmagic and religion (abridged ed., T. H. Gaster, Ed.).New York: Macmillan. (Original work published1890)

Freud, S. (1953). Three essays on the theory of sexuality.In J. Strachey (Ed. & Trans.), The standard edition ofthe complete psychological works of Sigmund Freud(Vol. 7, pp. 123–231). London: Hogarth Press.(Original work published 1905)

Freud, S. (1957). Five lectures on psychoanalysis. In J.Strachey (Ed. & Trans.), The standard edition of thecomplete psychological works of Sigmund Freud(Vol. 11, pp. 3–56). London: Hogarth Press. (Origi-nal work published 1910)

Gallese, V., Keysers, C., & Rizzolatti, G. (2004). A uni-fying view of the basis of social cognition. Trends inCognitive Science, 8, 398–403.

Goldenberg, J. L., Pyszczynski, T., Greenberg, J., Solo-mon, S., Kluck, B., & Cornwell, R. (2001). I am notan animal: Mortality salience, disgust, and the denialof human creatureliness. Journal of ExperimentalPsychology: General, 130, 427–435.

Gray, J. M., Young, A. W., Barker, W. A., Curtis, A., &Gibson, D. (1997). Impaired recognition of disgust inHuntington’s disease gene carriers. Brain, 120,2029–2038.

Grunfeld, D. I. (1982). The Jewish dietary laws: Vol. 1.Dietary laws regarding forbidden and permittedfoods, with particular reference to meat and meatproducts (3rd ed.). London: Soncino Press.

Haidt, J., & Keltner, D. (1999). Culture and facial ex-pression: Open-ended methods find more faces and agradient of recognition. Cognition and Emotion, 13,225–266.

Haidt, J., Koller, S., & Dias, M. (1993). Affect, culture,and morality, or is it wrong to eat your dog? Journalof Personality and Social Psychology, 65, 613–628.

Haidt, J., McCauley, C. R., & Rozin, P. (1994). A scaleto measure disgust sensitivity. Personality and Indi-vidual Differences, 16, 701–713.

Haidt, J., Rozin, P., McCauley, C. R., & Imada, S.


(1997). Body, psyche, and culture: The relationshipbetween disgust and morality. Psychology and De-veloping Societies, 9, 107–131.

Haslam, N. (2006). Dehumanization: An integrative re-view. Personality and Social Psychology Review, 10,252–264.

Hejmadi, A. (2000). Rasa or aesthetic emotion: An an-cient Hindu perspective. Unpublished manuscript.

Hejmadi, A., Davidson, R., & Rozin, P. (2000). Ex-ploring Hindu Indian emotion expressions: Evidencefor accurate recognition by Americans and Indians.Psychological Science, 11, 183–187.

Hejmadi, A., Rozin, P., & Siegal, M. (2004). Once incontact, always in contact: Contagious essence andconceptions of purification in American and HinduIndian children. Developmental Psychology, 40,467–476.

Hemenover, S. H., & Schimmack, U. (2007). That’s dis-gusting! . . . But very amusing: Mixed feelings ofamusement and disgust. Cognition and Emotion, 21,1102–1113.

Hodson, G., & Costello, K. (2007). Interpersonal dis-gust, ideological orientations and dehumanization aspredictors of intergroup attitudes. Psychological Sci-ence, 18, 691–698.

Howell, N. (1986). Feedbacks and buffers in relation toscarcity and abundance: Studies of hunter–gathererpopulations. In R. Scofield (Ed.), Beyond Malthus(pp. 156–187). Cambridge, UK: Cambridge Univer-sity Press.

Husted, D. S., Shapira, N. A., & Goodman, W. K.(2006). The neurocircuitry of obsessive–compulsivedisorder and disgust. Progress in Neuro-Psycho-pharmacology and Biological Psychiatry, 30, 389–399.



Jones, E. (1948). Anal–erotic character traits. In E.Jones, Papers on psychoanalysis (pp. 413–437).Boston: Beacon Press. (Original work published1912)

Kass, L. R. (1994). The hungry soul. New York: FreePress.

Kass, L. R. (1997, June 2). The wisdom of repugnance.The New Republic, pp. 17–26.

Kelly, D. R. (2007). Towards a cognitive theory of dis-gust. Manuscript in preparation.

Kleinknecht, R. A., Kleinknecht, E. A., & Thorndike, R.M. (1997). The role of disgust and fear in blood andinjection-related fainting symptoms: A structuralequation model. Behaviour Research and Therapy,35, 1075–1087.

Koch, M. D., O’Neill, H. K., Sawchuk, C. N., & Con-nolly, K. (2002). Domain-specific and generalizeddisgust sensitivity in blood–injection–injury phobia:The application of behavioral approach/avoidancetasks, 16, 511–527.

Leach, E. (1964). Anthropological aspects of language:

Animal categories and verbal abuse. In E. Lenneberg(Ed.), New directions in the study of language(pp. 23–64). Cambridge, MA: MIT Press.



Lumley, M. A., & Melamed, B. G. (1992). Bloodphobics and nonphobics: psychological differencesand affect during exposure. Behaviour Research andTherapy, 30, 425–434.

Malson, L. (1972). Wolf children (E. Fawcett, P. Ayrton,& J. White, Trans.). New York: Monthly ReviewPress. (Original work published 1964)

Mancini, F., Gragnani, A., & D’Olimpio, F. (2001). Theconnection between disgust and obsessions and com-pulsions in a non-clinical sample. Personality and In-dividual Differences, 31, 1173–1180.

Marriott, M. (1968). Caste ranking and food transac-tions: A matrix analysis. In M. Singer & B. S. Cohn(Eds.), Structure and change in Indian society (pp.133–171). Chicago: Aldine.

Martins, Y., & Pliner, P. (2006). “Ugh! That’s disgust-ing!”: Identification of the characteristics of foodsunderlying rejections based on disgust. Appetite, 46,75–85.

Marzillier, S. L., & Davey, G. C. L. (2005). Anxiety anddisgust: Evidence for a unidirectional relationship.Cognition and Emotion, 19, 729–750.

Masson, J. L., & Patwardhan, M. V. (1970). Aestheticrapture: the R�sadhy�ya of the N��ya��stra. Poona,India: Deccan College.

Matchett, G., & Davey, G. C. L. (1991). A test of adisease-avoidance model of animal phobias. Behav-iour Research and Therapy, 29, 91–94.

Mauss, M. (1972). A general theory of magic (R. Brain,Trans.). New York: Norton. (Original work pub-lished 1902)

Mayr, E. (1960). The emergence of evolutionary novel-ties. In S. Tax (Ed.), Evolution after Darwin: Vol. 1.The evolution of life (pp. 349–380). Chicago: Univer-sity of Chicago Press.

McKay, D. (2002). Introduction to the special issue: Therole of disgust in anxiety disorders. Journal of Anxi-ety Disorders, 16, 475–476.

Meigs, A. S. (1978). A Papuan perspective on pollution.Man, 13, 304–318.

Meigs, A. S. (1984). Food, sex, and pollution: A NewGuinea religion. New Brunswick, NJ: Rutgers Uni-versity Press.

Miller, S. B. (1986). Disgust: Conceptualization, devel-opment, and dynamics. International Review of Psy-choanalysis, 13, 295–307.

Miller, S. B. (2004). Disgust: The gatekeeper emotion.Hillsdale, NJ: Analytic Press.

Miller, W. I. (1997). The anatomy of disgust. Cam-bridge, MA: Harvard University Press.

47. Disgust 773

Mitchell, I. J., Heims, H., Neville, E. A., & Rickards, H.(2005). Huntington’s disease patients show impairedperception of disgust in the gustatory and olfactorymodalities. Journal of Neuropsychiatry and ClinicalNeurosciences, 17, 119–121.

Moll, J., deOliveira-Souza, R., Moll, F. T., Ignacio, F. A.,Bramati, I. E., Caparelli, D. E. M., et al. (2005). Themoral affiliations of disgust: A functional MRI study.Cognitive and Behavioral Neurology, 8, 68–78.

Nabi, R. L. (2002). The theoretical versus the lay mean-ing of disgust: Implications for emotion research.Cognition and Emotion, 16, 695–703.

Navarette, C. D., & Fessler, D. M. T. (2006). Diseaseavoidance and ethnocentrism: The effects of diseasevulnerability and disgust sensitivity on intergroup at-titudes. Evolution and Human Behavior, 27, 270–282.

Nemeroff, C., & Rozin, P. (1989). “You are what youeat”: Applying the demand-free “impressions”tech-nique to an unacknowledged belief. Ethos: The Jour-nal of Psychological Anthropology, 17, 50–69.

Nemeroff, C., & Rozin, P. (1992). Sympathetic magicalbeliefs and kosher dietary practice: The interaction ofrules and feelings. Ethos: The Journal of Psychologi-cal Anthropology, 20, 96–115.

Nemeroff, C., & Rozin, P. (1994). The contagion con-cept in adult thinking in the United States: Transmis-sion of germs and interpersonal influence. Ethos: TheJournal of Psychological Anthropology, 22, 158–186.

Nemeroff, C., & Rozin, P. (2000). The makings of themagical mind. In K. Rosengren, C. Johnson, & P.Harris (Eds.), Imagining the impossible: Magical, sci-entific, and religious thinking in children (pp. 1–34).New York: Cambridge University Press.

Nussbaum, M. C. (1999). “Secret sewers of vice”: Dis-gust, bodies, and the law. In S. A. Bandes (Ed.), Thepassions of law (pp. 19–62). New York: New YorkUniversity Press.

Olatunji, B. O., Lohr, J.-M., Sawchuk, C.-N., & Tolin,D.-F. (2007). Multimodal assessment of disgust incontamination-related obsessive–compulsive disor-der. Behaviour Research and Therapy, 45, 263–276.

Olatunji, B. O., & McKay, D. (2006). Introduction tothe special series: Disgust sensitivity in anxiety disor-ders. Journal of Behavior Therapy and ExperimentalPsychiatry, 37, 1–3.

Olatunji, B. O., & Sawchuk, C. N. (2005). Disgust:Characteristic features, social manifestations, andclinical implications. Journal of Social and ClinicalPsychology, 24, 932–962.

Olatunji, B. O., Sawchuk, C. N., de Jong, P. J., & Lohr,J. M. (2006). The structural relation between disgustsensitivity and blood–injection–injury fears: A cross-cultural comparison of US and Dutch data. Journalof Behavior Therapy and Experimental Psychiatry,37, 16–29.

Olatunji, B. O., Tolin, D. F., Huppert, J. D., & Lohr, J.M. (2005). The relation between fearfulness, disgustsensitivity and religious obsessions in a non-clinical

sample. Personality and Individual Differences, 38,713–722.

Olatunji, B. O., Williams, N. L., Lohr, J. M., &Sawchuk, C. N. (2005). The structure of disgust: Do-main specificity in relation to contamination ideationand excessive washing. Behaviour Research andTherapy, 43, 1069–1086.

Olatunji, B. O., Williams, N. L., Tolin, D. F., Sawchuck,C. N., Abramowitz, J. S., Lohr, J. M., et al. (2007).The Disgust Scale: Item analysis, factor structure,and suggestions for refinement. Psychological Assess-ment, 19, 281–297.

Ortner, S. B. (1973). Sherpa purity. American Anthro-pologist, 75, 49–63.

Page, A. C. (1994). Blood–injury phobia. Clinical Psy-chology Review, 14, 443–461.

Page, A. C. (2003). The role of disgust in faintness elic-ited by blood and injection stimuli. Journal of Anxi-ety Disorders, 17, 45–58.

Petó, E. (1936). Contribution to the development ofsmell feeling. British Journal of Medical Psychology,15, 314–320.

Phillips, M. L., Marks, I. M., Senior, C., Lythgoe, D.,O’Dwyer, A. M., Meehan, O., et al. (2000). A differ-ential neural response in obsessive–compulsive disor-der patients with washing compared with checkingsymptoms to disgust. Psychological Medicine, 30,1037–1050.

Phillips, M. L., Young, A. W., Senior, C., Brammer, M.,Andrew, C., Williams, S. C. R., et al. (1997). A spe-cific neural substrate for perceiving facial expressionsof disgust. Nature, 389, 495–498.

Piaget, J., & Inhelder, B. (1974). From conservation toatomism. In J. Piaget & B. Inhelder, The child’s con-struction of quantities (pp. 67–116). London: Rout-ledge & Kegan Paul. (Original work published 1941)



Power, M., & Dalgleish, T. (1997). Cognition and emo-tion: From order to disorder. Hove, UK: PsychologyPress.

Quigley, J. F., Sherman, M., & Sherman, N. (1996,March). Personality disorder symptoms, gender, andage as predictors of adolescent disgust sensitivity.Poster presented at the annual meeting of the EasternPsychological Association, Philadelphia.

Renner, H. D. (1944). The origin of food habits. Lon-don: Faber & Faber.

Rolls, E. T. (1994). Central taste anatomy and physiol-ogy. In R. L. Doty (Ed.), Handbook of clinicalolfaction and gustation (pp. 549–573). New York:Marcel Dekker.

Rosen, A., & Rozin, P. (1993). Now you see it. . . . Nowyou don’t: The preschool child’s conception of invisi-ble particles in the context of dissolving. Develop-mental Psychology, 29, 300–311.

Royzman, E. B., & Sabini, J. (2001). Something it takesto be an emotion: The interesting case of disgust.


Journal for the Theory of Social Behaviour, 31, 29–59.

Rozin, P. (1976). The evolution of intelligence and ac-cess to the cognitive unconscious. In J. A. Sprague &A. N. Epstein (Eds.), Progress in psychobiology andphysiological psychology (Vol. 6, pp. 245–280). NewYork: Academic Press.

Rozin, P. (1997). Moralization. In A. Brandt & P. Rozin(Eds.), Morality and health (pp. 379–401). NewYork: Routledge.

Rozin, P. (in press). Hedonic “adaptation”: Specific ad-aptation to disgust/death elcicitors as a result of dis-secting a cadaver. Judgment and Decision Making.

Rozin, P., & Fallon, A. E. (1987). A perspective on dis-gust. Psychological Review, 94, 23–41.

Rozin, P., Fallon, A. E., & Augustoni-Ziskind, M.(1985). The child’s conception of food: The develop-ment of contamination sensitivity to “disgusting”substances. Developmental Psychology, 21, 1075–1079.

Rozin, P., Fallon, A. E., & Mandell, R. (1984). Familyresemblance in attitudes to food. Developmental Psy-chology, 20, 309–314.

Rozin, P., Haidt, J., & McCauley, C. R. (1993). Disgust.In M. Lewis & J. M. Haviland (Eds.), Handbook ofemotions (pp. 575–594). New York: Guilford Press.

Rozin, P., Haidt, J., & McCauley, C. R. (1999). Disgust:The body and soul emotion. In T. Dalgleish & M.Power (Eds.), Handbook of cognition and emotion(Pp. 429–445). Chichester, UK: Wiley.

Rozin, P., Haidt, J., & McCauley, C. R. (2000). Disgust.In M. Lewis & J. M. Haviland-Jones (Eds.), Hand-book of emotions (2nd ed., pp. 637–653). New York:Guilford Press.

Rozin, P., Haidt, J., McCauley, C. R., Dunlop, L., &Ashmore, M. (1999). Individual differences in dis-gust sensitivity: Comparisons and evaluations ofpaper-and-pencil versus behavioral measures. Jour-nal of Research in Personality, 33, 330–351.

Rozin, P., Haidt, J., McCauley, C. R., & Imada, S.(1997). The cultural evolution of disgust. In H. M.Macbeth (Ed.), Food preferences and taste: Continu-ity and change (pp. 65–82). Oxford: Berghahn.

Rozin, P., Hammer, L., Oster, H., Horowitz, T., &Marmara, V. (1986). The child’s conception offood: Differentiation of categories of rejected sub-stances in the 1.4 to 5 year age range. Appetite, 7,141–151.

Rozin, P., Lowery, L., & Ebert, R. (1994). Varieties ofdisgust faces and the structure of disgust. Journal ofPersonality and Social Psychology, 66, 870–881.

Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). TheCAD triad hypothesis: A mapping between threemoral emotions (contempt, anger, disgust) and threemoral codes (community, autonomy, divinity). Jour-nal of Personality and Social Psychology, 76, 574–586.

Rozin, P., Markwith, M., & McCauley, C. R. (1994).The nature of aversion to indirect contact with otherpersons: AIDS aversion as a composite of aversion to

strangers, infection, moral taint and misfortune.Journal of Abnormal Psychology, 103, 495–504.

Rozin, P., Millman, L., & Nemeroff, C. (1986). Opera-tion of the laws of sympathetic magic in disgust andother domains. Journal of Personality and Social Psy-chology, 50, 703–712.

Rozin, P., & Nemeroff, C. J. (1990). The laws of sympa-thetic magic: A psychological analysis of similarityand contagion. In J. Stigler, G. Herdt, & R. A.Shweder (Eds.), Cultural psychology: Essays on com-parative human development (pp. 205–232). Cam-bridge, UK: Cambridge University Press.

Rozin, P., Nemeroff, C., Horowitz, M., Gordon, B., &Voet, W. (1995). The borders of the self: Contamina-tion sensitivity and potency of the mouth, other aper-tures and body parts. Journal of Research in Person-ality, 29, 318–340.

Rozin, P., Nemeroff, C., Wane, M., & Sherrod, A.(1989). Operation of the sympathetic magical law ofcontagion in interpersonal attitudes among Ameri-cans. Bulletin of the Psychonomic Society, 27, 367–370.

Rozin, P., & Singh, L. (1999). The moralization of ciga-rette smoking in America. Journal of ConsumerBehavior, 8, 321–337.

Rozin, P., Taylor, C., Ross, L., Bennett, G., & Hejmadi,A. (2005). General and specific abilities to recognisenegative emotions, especially disgust, as portrayed inthe face and the body. Cognition and Emotion, 19,397–412.

Rozin, P., & Wolf, S. (2007). Family resemblance inpreferences and values across three generations ofAmericans. Manuscript in preparation.

Rozin, P., Wrzesniewski, A., & Byrnes, D. (1998). Theelusiveness of evaluative conditioning. Learning andMotivation, 29, 397–415.

Sanfey, A. G., Rilling, J. K., Aronson, J. A., Nystrom, L.E., & Cohen, J. D. (2003). The neural basis of eco-nomic decision-making in the ultimatum game. Sci-ence, 300, 1755–1758.

Sawchuk, C. N., Meunier, S. A., Lohr, J. M., &Westendorf, D. H. (2002). Fear, disgust, and infor-mation processing in specific phobia: The applicationof signal detection theory. Journal of Anxiety Disor-ders, 16, 495–510.

Scherer, K. R. (1997). The role of culture in emotion–antecedent appraisal. Journal of Personality and So-cial Psychology, 73, 902–922.

Scherer, K. R., & Wallbott, H. G. (1994). Evidence foruniversality and cultural variation of differentialemotion response patterning. Journal of Personalityand Social Psychology, 66, 310–328.

Schienle, A., Schafer, A., Stark, R., Walter, B., Franz, M.,& Vaitl, D. (2003). Disgust sensitivity in psychiatricdisorders: A questionnaire study. Journal of Nervousand Mental Diseases, 191, 831–834.

Schienle, A., Schafer, A., Stark, R., Walter, B., & Vaitl,D. (2005). Relationship between disgust sensitivity,trait anxiety and brain activity during disgust induc-tion. Neuropsychobiology, 51, 86–92.

47. Disgust 775

Schienle, A., Schafer, A., Walter, B., Stark, R., & Vaitl,D. (2005). Elevated disgust sensitivity in blood pho-bia. Cognition and Emotion, 19, 1229–1241.

Schienle, A., Stark, R., & Vaitl, D. (2001). Evaluativeconditioning: A possible explanation for the acquisi-tion of disgust responses? Learning and Motivation,32, 65–83.

Schmidt, H., & Beauchamp, G. (1988). Adult-like odorpreferences and aversions in three-year-old children.Child Development, 59, 1136–1143.

Senn, M. J. E., & Solnit, A. J. (1968). Problems in childbehavior and development. Philadelphia: Lea &Febiger.

Shapira, N. A., Liu, Y., He, A. G., Bradley, M. M.,Lessig, M. C., James, G. A., et al. (2003). Brain acti-vation by disgust-inducing pictures in obsessive–compulsive disorder. Biological Psychiatry, 54, 751–756.

Sherman, G., Haidt, J., & Coan, J. (2007). Nazis reallyare disgusting. Unpublished manuscript, Universityof Virginia.

Shweder, R. A., Much, N. C., Mahapatra, M., & Park,L. (1997). The “big three” of morality (autonomy,community, divinity), and the “big three” explana-tions of suffering. In A. Brandt & P. Rozin (Eds.),Morality and health (pp. 119–169). New York:Routledge.

Siegal, M. (1988). Children’s knowledge of contagionand contamination as causes of illness. Child Devel-opment, 59, 1353–1359.

Siegal, M., & Share, D. L. (1990). Contamination sensi-tivity in young children. Developmental Psychology,26, 455–458.

Sprengelmeyer, R., Rausch, M., Eysel, U. T., &Przuntek, H. (1998). Neural structures associatedwith recognition of facial expressions of basic emo-tions. Proceedings of the Royal Society of London,Series B, 265, 1927–1931.

Sprengelmeyer, R., Young, A. W., Calder, A. J., Karnat,A., Lange, H., Homberg, V., et al. (1996). Loss of dis-gust: perception of faces and emotions in Hunting-ton’s disease. Brain, 119, 1647–1665.

Stark, R., Schienle, A., Walter, B., Kirsch, P., Sammer, G.,Ott, U., et al. (2003). Hemodynamic responses to fearand disgust-inducing pictures: An fMRI study. Inter-national Journal of Psychophysiology, 50, 225–234.

Stein, M., Ottenberg, P., & Roulet, N. (1958). A studyof the development of olfactory preferences. Archivesof Neurology and Psychiatry, 80, 264–266.

Steiner, J. E. (1979). Human facial expressions in re-sponse to taste and smell stimulation. In H. W. Reese& L. P. Lipsitt (Eds.), Advances in child developmentand behavior (Vol. 13, pp. 257–295). New York: Ac-ademic Press.

Tambiah, S. J. (1969). Animals are good to think andgood to prohibit. Ethnology, 8, 423–459.

Tolin, D. F., Lohr, J. M., Sawchuk, C. M., & Lee, T. C.(1997). Disgust and disgust sensitivity in blood–

injection–injury and spider phobia. Behaviour Re-search and Therapy, 35, 949–953.

Tolin, D. F., Woods, C. M., & Abramowitz, J. S. (2006).Disgust sensitivity and obsessive–compulsive symp-toms in a non-clinical sample. Journal of BehaviorTherapy and Experimental Psychiatry, 37, 30–40.

Tomkins, S. S. (1963). Affect, imagery, consciousness:Vol. 2. The negative affects. New York: Springer.

Tomkins, S. S. (1982). Affect theory. In P. Ekman (Ed.),Emotion in the human face (2nd ed., pp. 353–395).Cambridge, UK: Cambridge University Press.

Troop, N. A., Treasure, J. L., & Serpell, L. (2002). Afurther exploration of disgust in eating disorders. Eu-ropean Eating Disorders Review, 10, 218–226.

Tylor, E. B. (1974). Primitive culture: Researches intothe development of mythology, philosophy, religion,art and custom. New York: Gordon Press. (Originalwork published 1871)

Van Overveld, W. J. M., de Jong, P. J., Peters, M. L.,Cavanagh, K., & Davey, G. C. L. (2006). Disgustpropensity and disgust sensitivity: Separate con-structs that are differentially related to specific fears.Personality and Individual Differences, 41, 1241–1252.

Ware, J., Jain, K., Burgess, L., & Davey, G. C. L. (1994).Disease-avoidance model: Factor analysis of com-mon animal fears. Behaviour Research and Therapy,32, 57–63.

Webb, K., & Davey, G. C. L. (1993). Disgust sensitivityand fear of animals: effect of exposure to violent orrepulsive material. Anxiety, Coping, and Stress, 5,329–335.

Wicker, B., Keysers, C., Plailly, J., Royet, J. P., Gallese,V., & Rizzolatti, G. (2003). Both of us disgusted inmy insula: The common neural basis of seeing andfeeling disgust. Neuron, 40, 655–664.

Wierzbicka, A. (1986). Emotions: Universal or culture-specific? American Anthropologist, 88, 584–594.

Wolf, K., Mass, R., Ingenbleek, T., Kiefer, F., Naber, D.,& Wiedemann, K. (2005). The facial pattern of dis-gust, appetence, excited joy and relaxed joy: An im-proved facial EMG study. Scandinavian Journal ofPsychology, 46, 403–409.

Woody, S. R., McLean, C., & Klassen, T. (2005). Dis-gust as a motivator of avoidance of spiders. Journalof Anxiety Disorders, 19, 461–475.

Woody, S. R., & Teachman, B. A. (2000). Intersection ofdisgust and fear: Normative and pathological views.Clinical Psychology: Science and Practice, 7, 291–311.

Woody, S. R., & Tolin, D. F. (2002). The relationshipbetween disgust sensitivity and avoidant behavior:Studies of clinical and nonclinical samples. Journal ofAnxiety Disorders, 16, 543–559.

Wright, P., He, G., Shapira, N. A., Goodman, W. K., &Liu, Y. (2004). Disgust and the insula: fMRI re-sponses to pictures of mutilation and contamination.NeuroReport, 15, 2347–2351.


C H A P T E R 4 8

Positive Emotions

BARBARA L. FREDRICKSON and MICHAEL A. COHN

Positive emotions have attracted increased sci-entific attention in the past decade. They havelong been studied as markers of people’s overallwell-being or happiness (Diener & Seligman,2004; Kahneman, Kreuger, & Schkade, 2004),but looking at positive emotions as outcomes isjust the beginning. In large, well-controlledstudies, positive emotions and experiences havebeen found to predict or contribute to manydifferent life outcomes (Lyubomirsky, King, &Diener, 2005), as well as increased longevity(Danner, Snowdon, & Friesen, 2001; Levy,Slade, & Kunkel, 2002; Moskowitz, 2003;Ostir, Markides, & Black, 2000); improvedimmune function (Cohen, Doyle, & Turner,2003); and less pain, impairment, and mortal-ity in people with chronic disease (Gil et al.,2004; Cohen & Pressman, 2006). The“broaden-and-build” theory of positive emo-tions (Fredrickson, 1998, 2001) encompassesthis great variety of empirical results, and thischapter uses this theory as a framework for or-ganizing and interrelating past findings andcurrent questions about positive emotions.Through this synthesis, we hope to explain the

central paradox of positive emotions: How is itthat fleeting experiences of joy, interest, orlove—which can be so easily squelched ordismissed—produce lasting gains in strengthsand well-being?

OUTLINE OF THE CHAPTER

We begin the chapter with a general definitionof positive emotions and discuss some of the is-sues involved in studying them. We then brieflylay out our broaden-and-build theory, whichwe use as a framework for organizing researchon positive emotions throughout the chapter.The broaden-and-build theory begins with theimmediate effects of positive emotions, whichserve to broaden attention, cognition, and be-havioral repertoires. These lead to the long-term effects of frequent positive emotions,which serve to build resources that make last-ing contributions to survival, health, and hap-piness. Finally, we take stock of what is knownabout positive emotions, what questions re-main, and how the study of positive emotions

777

can contribute to the field of positive psychol-ogy more generally.

DEFINING POSITIVE EMOTIONS

The theories of emotions that dominated psy-chology for most of its history proved fruitfulfor studying negative emotions, but were oftenunsuitable for the study of positive emotions(Fredrickson, 1998). In the past 10 years, posi-tive emotions have come into their own. Therenaissance in positive emotions research stemsfrom two sources: a growing interest in the psy-chology of the “good life” (Ryff & Singer,1998; Fredrickson, 1998; Csikszentmihalyi &Csikszentmihalyi, 2006); and several researchprograms that have sought to build an empiri-cal, bottom-up model of positive emotions,rather than shoehorning them into older mod-els that were constructed primarily for thenegative emotions. Below we review thesefindings—refining what makes an emotion orother state “positive,” determining what differ-entiates a positive emotion from other pleasantaffective states, and investigating some of thechallenges unique to the study of positive emo-tions.

Positive Emotionsversus Other Positive Affective States

The distinctions between positive emotions andother closely related affective states, such assensory pleasure and positive mood, have oftenbeen blurry. Although working definitions ofemotions vary somewhat across researchers, aconsensus is emerging that emotions (bothpositive and negative) are best conceptual-ized as multicomponent response tendencies—incorporating muscle tension, hormone release,cardiovascular changes, facial expression, at-tention, and cognition, among other changes—that unfold over a relatively short time span.Typically, emotions begin with an individual’sassessment of the personal meaning of someantecedent event—what Lazarus (1991) calledthe “person–environment relationship” or “ad-aptational encounter.” Either conscious or un-conscious, this appraisal process triggers acascade of responses incorporating mental,physical, and subjective changes.

Sensory pleasure includes such experiencesas sexual gratification, satiation of hunger orthirst, and the remedying of unpleasant states

(e.g., cold, pain, or excessive noise). Cabanac(1971) suggested that sensory pleasure ariseswhenever a stimulus “corrects an internal trou-ble” (e.g., cooling down when overheated, eat-ing when hungry). Sensory pleasure shares withpositive emotions a pleasant subjective feel andmay include physiological changes, but anemotion also requires an appraisal of somestimulus or an assessment of its meaning. Emo-tion and sensation often co-occur: A good mealsatisfies hunger, and can also lead to feelings ofcontentment; sex provides pleasant sensations,and may also lead to gratitude or love towardone’s partner. Positive emotions can also occurwithout a physical stimulus (e.g., joy at receiv-ing good news, or interest in a new idea).Berridge (Berridge & Robinson, 2003, Peciña,Smith, & Berridge, 2006) finds evidence forperhaps a similar distinction at the neurologi-cal level: Positive affect includes a passive “lik-ing” component, mediated by opioid receptors,and a motivational “wanting” component, me-diated by dopamine.

Positive emotions also resemble positivemoods. Yet emotions differ from moods, inthat emotions are about some personally mean-ingful circumstance (i.e., they have an object),are typically short-lived, and occupy the fore-ground of consciousness. In contrast, moodsare typically free-floating or objectless, aremore long-lasting, and occupy the backgroundof consciousness (Oatley & Jenkins, 1996;Rosenberg, 1998). These distinctions betweenemotions and moods, however, are made moreoften at theoretical than at empirical levels. Inresearch practice, virtually identical techniquesare used for inducing positive moods and posi-tive emotions (e.g., giving gifts, viewing come-dies). Many experimental techniques involvepresenting a positive stimulus in order to lead aparticipant to respond to an unrelated taskwith a generally positive mindset. In this case,multiple forms of positive mood or emotionmay lead to the same results. However, re-search on the tendencies linked to specific emo-tions (e.g., gratitude, pride, awe), or on positivestates with more specific meaning (e.g., opti-mism, confidence, enjoyment), requires a morecareful distinction between mood and emotion.

Links to Urges to Approach or Continue

Most commonly, the function common to allpositive emotions has been conceptualized asfacilitating approach behavior (Cacioppo,


Priester, & Berntson, 1993; Davidson, 1993;Frijda, 1994) or continued action (Carver &Scheier, 1990; Clore, 1994). From this perspec-tive, experiences of positive emotions promptindividuals to engage with their environmentsand take part in activities, many of which areevolutionarily adaptive for the individuals,their species, or both. This link between posi-tive emotions and activity engagement providesan explanation for the often-documented“positivity offset,” or the tendency for individ-uals to experience mild positive affect fre-quently, even in neutral contexts (Diener &Diener, 1996; Ito & Cacioppo, 1999). Withoutsuch an offset, individuals most often would beunmotivated to engage with their environ-ments. Yet with such an offset, individuals ex-hibit the adaptive bias to approach and explorenovel objects, people, or situations.

However, other positive affective states sharethese effects. Sensory pleasure, for instance,motivates people to approach and continueconsuming whatever stimulus is (or appears)biologically useful at the moment (Cabanac,1971). Free-floating positive moods motivatepeople to continue along any line of thinking oraction that they have initiated (Clore, 1994).Thus the approach model can be seen as a low-est common denominator underlying subjec-tively pleasant states, but it is not sufficient todefine or capture the effects of the class of posi-tive emotions. Furthermore, positive emotionssuch as relief can be conceptualized as ap-proaching a desired state by avoiding an unde-sired one, but “approach” is by itself a poordescription of the associated elicitors or be-havioral tendencies. Clearly, positive emotionscan arise from a variety of approach- andavoidance-related situations.

Core Appraisal Dimensions

Theorists differ as to whether emotions are bestmodeled as points on a two-dimensional plane(Russell, Weiss, & Mendelson, 1989), pointsin a higher-dimensional space (Smith &Ellsworth, 1985), or separately evolved mod-ules (Tooby & Cosmides, Chapter 8, this vol-ume). However, there is general agreement thata primary characteristic of every emotion is va-lence on a bipolar continuum from highly un-pleasant to highly pleasant (reviewed in Smith& Ellsworth, 1985). Indeed, this pleasantnessrating may be one of the earliest determina-tions we make when processing sensory input

from our environment (Chen & Bargh, 1999).An appraisal of pleasantness can arise whena stimulus fulfills a biological need (e.g.,Cabanac, 1971), when it contributes to a per-sonally relevant goal, or when it remedies anoxious or goal-inconsistent state. Others haveargued that an appraisal of pleasantness isbased on a favorable comparison between ouractual rate of goal attainment and our expectedrate (Carver, 2003). This distinction may relateto the distinction between sensations and emo-tions: Anything that we recognize as progresstoward a goal should elicit an immediatelypleasant response, but when appraised in thecontext of our expectations or other consider-ations, it may or may not give rise to a pleasant(positive) emotion.

The pleasantness dimension of emotions isseparate from and orthogonal to other aspectsof emotional experience. Consider, for exam-ple, the emotions of joy (high arousal) versuscontentment (low arousal), or gratitude (lowpersonal control) versus pride (high control).Past emotion measures have often conflatedpleasantness with either high arousal or highpersonal control, even though pleasant emo-tions can span the range of these dimensions.For example, one prominent scale, the Positiveand Negative Activation Schedule1 (PANAS;Watson, Wiese, Vaidya, & Tellegen, 1999),deliberately focuses on high-arousal positiveemotions for psychometric reasons. ThisPANAS has been effective in the past, perhapsbecause different positive emotions tend to cor-relate, or because Americans tend to valuehigh-arousal positive emotions more highlythan low-arousal ones (Tsai, Knutson, & Fung,2006). There are two reasons to favor mea-sures with greater distinction between positiveemotions. First, the fact that different positiveemotions co-occur does not mean that do nothave distinct effects. Low-arousal positiveemotions are likely to have different thought–action tendencies from high-arousal ones, andignoring these emotions impedes our ability tomake specific predictions about emotions andbehavior. Second, looking at only a subset ofpositive emotions can tell us whether a personis generally having a positive experience, but itis less helpful in quantitatively measuring theperson’s level of positive emotion. Increasingevidence suggests that a person’s ratio of posi-tive to negative emotions is an important pre-dictor of psychological and social outcomes(Fredrickson & Losada, 2005; Gottman, 1994).

48. Positive Emotions 779

These quantitative measures are best assembledwhen the participant rates a more inclusive setof positive emotions. Therefore, positive emo-tion measures perhaps work best when theycast the widest possible net.

Additional Appraisal Dimensions

Appraisal-based emotion theories have gen-erally found that positive emotions are lesscognitively distinct than negative ones (Smith& Ellsworth, 1985; Ellsworth & Smith, 1988;Fredrickson & Branigan, 2001). This dovetailswith experience-sampling research showingthat experiences of various positive emotionscovary more strongly than experiences of nega-tive emotions (i.e., people are more likely tofeel multiple positive emotions at one timethan multiple negative ones; Barrett, Gross,Christensen, & Benvenuto, 2001). Becausenegative emotions are adapted for specific,survival-critical situations (see the descriptionof the broaden-and-build theory later in thischapter), it is plausible that negative situationswould evoke only one emotion, or a closelyrelated cluster. Positive emotions initiate abroader range of thoughts and actions, so it ismore likely that—for example—an experienceof contentment could lead to thoughts aboutnew challenges to take on, leading quickly toexperiences of pride or excitement.

However, there is also evidence that our pastresearch methods have been restrictive. Tong(2007) proposed new appraisal dimensionsbased on research on individual positive emo-tions, and found evidence of several higher-order dimensions—each associated with sev-eral discrete positive emotions—that made thepositive emotions appear just as distinct fromone another as the negative emotions. Whereasthe negative emotions seem to be differentiatedfrom each other by appraisals of threat, per-sonal responsibility, and self-efficacy, the posi-tive emotions require additional dimensions,such as interpersonal relationship, mastery, andspiritual experience. Similarly, Shiota, Keltner,and John (2006) found that the Big Five per-sonality traits predict dispositions toward dif-ferent positive emotions, beyond the well-documented association between extraversionand positive affect more generally. Thus recentresults suggest that the antecedents, effects, andsubjective dimensions of such emotions as con-tentment, compassion, and amusement are justas different as those of any negative emotions.Finally, individuals show different levels of

ability in noting fine distinctions between emo-tions (lexithymia), with higher levelspredicting lower correlations among the expe-riences of different positive emotions (Tugade,Fredrickson, & Barrett, 2004).

ISSUES IN THE STUDYOF POSITIVE EMOTIONS

Qualitatively Distinctfrom Negative Emotions

Historically, emotion research has focused onnegative emotions. The most general reason isthat psychology as a whole tends to focus onunderstanding and ameliorating psychologicalproblems (Seligman & Csikszentmihalyi,2000). Negative emotions—when extreme,prolonged, or contextually inappropriate—areimplicated in many grave problems, includinganxiety disorders, aggression and violence, eat-ing disorders and self-injury, and depressionand suicide. Medicine and health psychologyhave also focused on negative emotions, whichcontribute to problems ranging from sexualdysfunction to life-threatening immune dis-orders (Kiecolt-Glaser, McGuire, Robles, &Glaser, 2002). Although positive emotions cancontribute to problems (e.g., mania, drug ad-diction), negative emotions are more prom-inent in psychopathology and thus havecaptured the majority of research attention.Studies of positive functioning and strengthshave only recently begun to catch up, concom-itantly raising interest in the contributions ofpositive emotions. We also argue later in thechapter that even the study of pathology hasbeen hindered by overlooking positive emo-tions, which play a critical role in recoveringfrom adversity and developing compensatorystrengths.

The study of positive emotions has also suf-fered from the long-running effort to create asingle, general theory of emotion. Such modelsare typically built with the more attention-grabbing negative emotions (e.g., fear andanger) as prototypes, with positive emotionssqueezed in later. For instance, many theoriesof emotions associate each emotion with a spe-cific action tendency (Frijda, 1986; Frijda,Kuipers, & Schure, 1989; Lazarus, 1991;Levenson, 1994; Oatley & Jenkins, 1996;Tooby & Cosmides, 1990). Fear, for example,is linked with the urge to escape, anger with theurge to attack, disgust with the urge to expel,and so on. No theorist argues that these are ir-


resistible compulsions, but rather that people’sideas about possible courses of action narrowin on a specific set of behavioral options. Oftenthe specific action tendency is linked to theemotion’s evolutionary adaptive value. Its ac-tion tendencies are presumed to incline peopletoward behaviors that helped get early humansout of life-or-death situations. Another keyidea is that specific action tendencies and phys-iological changes go hand in hand: For exam-ple, when you feel fear your thoughts will tendtoward escape, and your autonomic nervoussystem will change in ways that would help yourun or climb.

Although specific action tendencies were in-voked to describe the form and function of pos-itive emotions as well, these tendencies werenotably vague and underspecified (Fredrickson& Levenson, 1998). Joy, for instance, waslinked with aimless activation, interest withattending, and contentment with inactivity(Frijda, 1986). These tendencies seem far toogeneral to be called specific; nor do they pres-ent the same obvious adaptive value as negativeemotions’ action tendencies (Fredrickson,1998). Although a few theorists noted that fit-ting positive emotions into emotion-generalmodels posed problems (Ekman, 1992; Laza-rus, 1991), this acknowledgment was not ac-companied by any new or revised models tobetter accommodate the positive emotions. In-stead, the difficulty of shoehorning the positiveemotions into emotion-general models merelytended to marginalize them further. The pre-vailing models were successful for studying thenegative emotions, so researchers pursuedthat work and often neglected the positiveemotions—either because it was more difficultto make progress, or because the positive emo-tions challenged the validity of the model.

Challenging to Measure and Evoke

Another reason why positive emotions mayhave been featured in less past research is sim-ply that they are harder to study. Negativeemotions, as we have discussed, lead to a focuson a single, clear action tendency. In the lab,this translates into clear, reliable results. Theresults of positive emotions are more diffuseand generally less urgent, making them moredifficult to operationalize or observe.

Empirical evidence from many areas of psy-chology (Baumeister, Bratslavsky, Finkenauer,& Vohs, 2001; Rozin & Royzman, 2001;Cacioppo, Gardner, & Berntson, 1999) and

economics (Tversky & Kahneman, 1981) sug-gests that negative emotions command moreattention than positive ones, and that negativeevents evoke a stronger response than positiveevents of the same magnitude. The observedhealth effects of prolonged negative emotionsalso contribute to the impression that negativeemotions are more significant. In addition, psy-chologists have a wide range of establishedprotocols for evoking fear, disgust, and anger,but fewer for evoking joy, contentment, orlove.

Thus scientists who wish to study positiveemotions face procedural difficulties, less obvi-ous dependent variables, and smaller effectsizes. However, this should not be taken tomean that positive emotions are less important!Negative emotions appear strongly in the mo-ment because they were sculpted by evolutionto mobilize immediate action; positive emo-tions exert their power over the long term, andare critical to building a healthy and fruitful life(Fredrickson & Losada, 2005; Lyubomirsky etal., 2005). We will return to this asymmetry re-peatedly: Negative emotions help us respond toa single, immediate threat; positive emotionshelp us take advantage of life’s numerous op-portunities.

Widely, but Not Universally, Appropriate

Some critics have suggested that researcherswho study the benefits of positive emotionspathologize negative emotions, ignore situa-tions where feeling good is inappropriate, ortacitly endorse an unsophisticated hedonism ora Pollyanna-like disregard for life’s difficulties(Lazarus, 2003; see also Fineman, 2006). Werecognize the critical role of negative emotionsand their associated physiological responses indealing with threatening situations, as well asin appropriately marking the emotional impor-tance of serious losses. We also acknowledgethat extremes of positive emotion can be dis-ruptive, especially in high-performance states,which are often marked by an overall lack ofself-focus (Csikszentmihalyi, 1990).

The emerging picture is not that peopleshould experience high levels of positive emo-tion at all times and in all situations, or thatpositive emotions are a panacea for all of life’schallenges. Rather, it appears beneficial forpeople to cultivate positive emotions as a gen-eral backdrop to their emotional lives, whilestill responding positively or negatively to emo-tionally meaningful events as they occur. It may


also be natural: Recall the evidence for a per-vasive and cross-cultural “positivity offset”(Diener & Diener, 1996; Ito & Cacioppo,1999), suggesting that people usually feel a lit-tle better than neutral. This is appropriate ifthey are usually in a position to amass re-sources, rather than responding to an imme-diate threat. Negative emotions help peopleachieve survival and short-term material ends,while positive emotions help them build to-ward a more widely construed “good life”(Ryff & Singer, 1998; Keyes & Haidt, 2003).

THE BROADEN-AND-BUILD THEORYOF POSITIVE EMOTIONS

The broaden-and-build theory (Fredrickson,1998) arose from a desire to move beyondnegative-emotion-based models of emotion,and capture the unique effects of positive emo-tions. Fredrickson’s broaden-and-build theoryof positive emotions holds that positive emo-tions broaden people’s momentary thought–action repertoires and lead to actions that buildenduring personal resources (Fredrickson,1998, 2001).

The specific action tendencies described bytraditional models are appropriate descriptionsof the function of negative emotions: They arethe outcomes of thought–action repertoiresthat narrow individuals’ urges and perceivedaffordances so that they are likely to act in aspecific way (e.g., escape, attack, expel). In alife-threatening situation, a narrowed thought–action repertoire promotes quick and decisiveaction that carries direct and immediate bene-fit.2 The various negative emotions available tohumans comprise the thought–action reper-toires that worked best to save their ancestors’lives and limbs (or genes) in similar situations.

Positive emotions, in contrast, seldom occurin response to life-threatening situations. Thusthere is less need for them to evoke specific, fo-cused response tendencies. Instead, positiveemotions lead to broadened and more flexibleresponse tendencies, widening the array of thethoughts and actions that come to mind(Fredrickson, 1998). Joy, for instance, createsthe urge to play, push the limits, and becreative—urges evident not only in social andphysical behavior, but also in intellectual andartistic behavior. Interest, a phenomenological-ly distinct positive emotion, creates the urge toexplore, take in new information and experi-

ences, and expand the self in the process.Contentment, a third distinct positive emotion,creates the urge to sit back and savor currentlife circumstances, and to integrate these cir-cumstances into new views of self and of theworld. And love—which we view as an amal-gam of distinct positive emotions (e.g., joy, in-terest, and contentment) experienced withincontexts of safe, close relationships—createsrecurring cycles of urges to play with, explore,and savor loved ones. The appraisal dimen-sions that are unique to positive emotions, suchas interpersonal relationship, mastery, and spir-ituality (Tong, 2007), can also be seen as differ-ent domains in which various thought–actiontendencies (to play, to explore, or to savor andintegrate) broaden habitual modes of thinkingor acting.

In contrast to negative emotions, which wereshaped by evolution to provide direct and im-mediate adaptive benefits when survival isthreatened, the broadened thought–action rep-ertoires triggered by positive emotions evolvedbecause of their indirect and long-term adap-tive benefits. Broadening builds enduring per-sonal resources.

Take play as an example. Specific forms ofchasing play evident in juveniles of a species—such as running into a flexible sapling orbranch and catapulting oneself in an unex-pected direction—are reenacted in adults ofthat species exclusively during predator avoid-ance (Dolhinow, 1987). Such correspondencesbetween juvenile play maneuvers and adultsurvival maneuvers suggest that juvenile playbuilds enduring physical resources (Boulton &Smith, 1992; Caro, 1988). Social play alsobuilds enduring social resources. Laughter ap-pears to function as a social signal of opennessto new, friendly interactions (broadening),which can lead to lasting social bonds and at-tachments (building; Gervais & Wilson, 2005).Shared amusement and smiles have manyof the same effects (Lee, 1983; Simons,McCluskey-Fawcett, & Papini, 1986; Keltner& Bonanno, 1997). Childhood play also buildsenduring intellectual resources, by increasinglevels of creativity (Sherrod & Singer, 1989)and fueling brain development (Panksepp,1998). Similarly, the exploration prompted bythe positive emotion of interest creates knowl-edge and intellectual complexity, and the savor-ing prompted by contentment produces self-insight and alters world views. So these phe-nomenologically distinct positive emotions all


share the feature of augmenting individuals’personal resources, ranging from physical andsocial resources to intellectual and psycho-logical ones (for more detailed reviews, seeFredrickson, 1998, 2001; Fredrickson &Branigan, 2001).

Importantly, the personal resources accruedduring states of positive emotions are durable.They outlast the transient emotional states thatled to their acquisition. These resources can bedrawn on in subsequent moments and in dif-ferent emotional states. So through experi-ences of positive emotions, people transformthemselves—becoming more creative, know-ledgeable, resilient, socially integrated, andhealthy individuals. Figure 48.1 representsthese three sequential effects of positive emo-tions (broadening, building, transforming). Italso suggests that initial experiences of positiveemotions produce upward spirals toward fur-ther experiences of positive emotions—a pointwe have begun to investigate empirically(Fredrickson & Joiner, 2002; see the followingsection for more details).

In short, the broaden-and-build theory de-scribes the form of positive emotions in termsof broadened thought–action repertoires, anddescribes their function in terms of buildingenduring personal resources. The theory ex-plains the evolved adaptive significance of

positive emotions. Those of our ancestorswho succumbed to the urges sparked by posi-tive emotions—to play, explore, and so on—would have by consequence accrued morepersonal resources. When these same ances-tors later faced inevitable threats to life andlimb, their greater personal resources wouldhave translated into greater odds of survival,and in turn greater odds of living longenough to reproduce. To the extent that thecapacity to experience positive emotions is in-nate and heritable, natural selection wouldhave driven it to become a basic trait sharedby our entire species (for early evidence thatthe broaden effect of positive emotions is in-deed cross-cultural, see Waugh, Hejmadi,Otake, & Fredrickson, 2006).

SHORT-TERM AND LONG-TERMEFFECTS OF POSITIVE EMOTIONS

Our empirical investigation of the broaden-and-build theory has rested on two hypotheses:the “broaden hypothesis,” which targets theways people change while experiencing a posi-tive emotion, and the “build hypothesis,”which targets the lasting changes that followrepeated positive emotional experiences overtime.


FIGURE 48.1. The broaden-and-build theory of positive emotions.

The Broaden Hypothesis

The first central claim of the broaden-and-build theory is that experiences of positiveemotions broaden a person’s momentarythought–action repertoire. We call this the“broaden hypothesis,” and have found evi-dence for it across a range of domains, fromvisual attention to self-construal and socialgroup perception.

Visual Attention

The most cognitively basic form of broadeningwe have examined appears in global–local vi-sual processing tasks. Participants are asked tomake a choice about a figure that can be judgedon the basis of either its global, overall shape orits local detail elements (see Figure 48.2a for anexample). Positive emotions, with their broad-ened focus, produce a preference for the globallevel, whereas negative emotions often producea preference for the details. This pattern holdsboth for emotionally relevant traits like op-timism and anxiety (Basso, Schefft, Ris, &Dember, 1996), and for emotional states in-duced through a variety of means (Brandt,Derryberry, & Reed, 1992, cited in Derryberry& Tucker, 1994; Fredrickson & Branigan,2005; Johnson & Fredrickson, 2005). Wad-linger and Isacowitz (2006) tracked partici-pants’ eye movements, and found that inducedpositive emotion broadened visual search pat-terns, leading to increased attention to periph-eral stimuli.

Our lab’s work with the global–local visualprocessing paradigm was the first to introducea neutral control condition, and to demon-strate that positive emotions broaden attentionrelative to a nonemotional baseline (Fredrick-son & Branigan, 2005). We also tested distinctemotions in each valence type (e.g., content-ment and joy; anxiety and anger), to build thecase that the broaden effect was linked to posi-tive emotions in general, and not to arousallevel or another specific property of the emo-tions tested.

Cognition and Behavior

Emotions affect both the focus and the processof cognition, and many long-standing findingson the effects of positive affect on cognitionand behavior are consistent with the broadenhypothesis. For instance, Isen and colleaguestested the effects of positive states on a wide

range of cognitive outcomes, ranging from cre-ativity puzzles to simulations of complex, life-or-death work situations (Estrada, Isen, &Young, 1997). Their work demonstrates thatpositive emotions produce patterns of thoughtthat are notably unusual (Isen, Johnson, Mertz,& Robinson, 1985), flexible and inclusive (Isen& Daubman, 1984), creative (Isen, Daubman,& Nowicki, 1987), and receptive to new infor-mation (Estrada et al., 1997). Confirming aninterpretation of these results in terms of thebroaden effect, Rowe, Hirsch, and Anderson(2007) replicated Isen et al.’s (1987) findings ofimproved performance on the verbal RemoteAssociates Test (Figure 48.2b), and found thatthis improvement was correlated with de-creased performance on a visual task that re-quired participants to ignore peripheral cues(i.e., a task that required attentional narrow-ing).

In the domain of more personally relevantbehavior, we (Fredrickson & Branigan, 2005)induced positive, negative, or no emotions, andasked participants to step away from the spe-cifics of the induction and list all the thingsthey felt like doing. Participants induced to feelpositive emotions listed more and more variedpotential actions, relative to the neutral group;participants induced to feel negative emotionslisted fewer potential actions than the neutralgroup. Similar research has shown that positiveemotions produce more creative (Isen et al.,1987) and variable (Kahn & Isen, 1993) ac-tions. (See also Isen, Chapter 34, this volume.)

Another perspective on positive affectand cognition comes from the mood-as-information theorists. Their view resembles thebroaden hypothesis in suggesting that positiveemotions lead to creative, unusual, or integra-tive thinking, but it differs in predicting a con-comitant reduction in attention to detail andnegative feedback, sometimes leading to anoverreliance on heuristics or stereotypes. Thereis substantial empirical support for this view(for a review, see the volume by Martin &Clore, 2001). However, other work suggeststhat people in positive emotional states aremore likely to incorporate challenging evidence(Trope & Pomerantz, 1998) and carefully con-sider difficult problems (for reviews, seeAspinwall, 1998, and Abele, 1992; see also thework by Isen et al. cited above).

It is unlikely that either of these bodies of evi-dence is entirely spurious, or that either themood-as-information or the mood-as-resourcemodel will be entirely disproven by future evi-


dence. The broaden-and-build model as it cur-rently stands is not specific enough to offer a pri-ori guidance: The general term “broadening”could apply to either vague, heuristic thinking orthorough, nondefensive exploration, and littlehas been done to reconcile these opposing inter-pretations. What work there is suggests thatflexibility and openness are important attributesof positive emotions’ cognitive effects(Dreisbach & Goschke, 2004; Bless et al., 1996).Properties of the particular situation may welldetermine whether these attributes are beneficialor harmful. We are continuing this work, consid-ering variables such as method of emotion in-duction, level of intrinsic motivation, social andpersonal relevance, and amount of mental setswitching involved, to determine when thethought–action tendencies associated with posi-tive emotions are beneficial to problem solvingand when they are detrimental.

Social Cognition

Broadening in the social domain takes the formof enhanced attention to others and reduceddistinctions between self and other, or betweendifferent groups. Participants experiencing pos-

itive emotions report more overlap betweentheir concept of themselves and their conceptof their best friend (Waugh & Fredrickson,2006; Waugh et al., 2006), and they be-come more imaginative and attentive regardingthings they could do for friends, relativeto things friends could do for them (Otake,Waugh, & Fredrickson, 2007). When a closerelationship does not yet exist, induced posi-tive emotions can increase trust (Dunn &Schweitzer, 2005), and may underlie the cre-ation of a wide variety of bonds and interde-pendence opportunities (Cohn & Fredrickson,2006; Gable, Reis, Impett, & Asher, 2004).

Positive emotions also broaden social groupconcepts and break down an essentialized senseof “us versus them” (Dovidio, Gaertner, Isen,Rust, & Guerra, 1995). We have discovered thesame result in a racial context: When we inducepositive emotions in participants, people be-come better at remembering the faces of indi-viduals of other races,3 and simultaneouslyworse at perceiving physical differences be-tween races (Figure 48.2c) (Johnson, 2005;Johnson & Fredrickson, 2005).

The studies we have discussed demonstratevariety in the broaden effect, but more impor-


FIGURE 48.2. Three forms of broadened attention. (a) The participant is instructed to find the letter T asquickly as possible. It is present in both figures, but finding the first is facilitated by a broadened visualfocus, while finding the second is facilitated by a narrowed (detail-oriented) focus (Johnson, 2005;Fredrickson & Branigan, 2005). (b) In this item from the Remote Associates Test, the participant is askedto find a word that ties the three stimulus words together. Participants are more likely to find the answer(“jack”) when experiencing a positive emotion (Isen, Daubman, & Nowicki, 1987). (c) White individu-als are typically poor at distinguishing one black face from another (solid box) and good at determiningwhere a morphed series crosses from “more white” to “more black.” Positive emotions improve face rec-ognition (Johnson & Fredrickson, 2005) and impair racial categorization (Johnson, 2005).

tantly, the outcomes they involve can make asubstantive difference in how people act. Thesechanges can then lead to differences in theircircumstances, abilities, resources, and rela-tionships in the future. In other words, thesebroadened mindsets can lead people to buildenduring resources.

The Build Hypothesis

The second central claim of the broaden-and-build theory is that temporary and transient ex-periences of positive emotions, by encouraginga broadened range of actions, over time buildenduring personal resources. It is now es-tablished that positive emotions function ascauses, results, and concomitants of success inlife. A large meta-analysis by Lyubomirsky etal. (2005) reviews the links between positive af-fect and outcomes ranging from satisfaction atwork and in relationships, to physical healthand effective problem solving. We refer inter-ested readers to their excellent review for spe-cifics. In this section we review a selection ofresearch that helps fill in the missing pieces nec-essary to test our overarching model: that posi-tive emotions lead to broader thought–actionrepertoires, and that these broadened mindsetsenable people to build resources over time.

Correlational and experimental studies ofhumans and animals help us link positive traits,positive states, and behaviors linked with posi-tive states (e.g., play) to increases in physical,intellectual, and social resources. As previouslymentioned, ethologists who have observednonhuman mammals have associated juvenileplay with the development of specific survivalmaneuvers evident in both predator avoidanceand aggressive fighting (Boulton & Smith,1992; Caro, 1988), suggesting that play buildsenduring physical resources. In laboratory ex-periments, rats deprived of juvenile social playwere slower to learn a complex motor taskthan nondeprived controls were (Einon, Mor-gan, & Kibbler, 1978).

Evidence suggesting that positive emotionsbuild intellectual resources can be drawn fromstudies on individual differences in attachmentstyles. Securely attached children—those whoexperience the most consistent caregiver love—are more persistent, flexible, and resourcefulproblem solvers than their peers (Arend, Gove,& Sroufe, 1979; Matas, Arend, & Sroufe,1978). They also engage in more independentexploration of novel places, and, as a conse-

quence, develop superior cognitive maps ofthose spaces (Hazen & Durrett, 1982). The in-tellectual resources associated with secure at-tachment also appear to last into adulthood.Securely attached adults are more curious andopen to information than their insecurely at-tached peers (Mikulincer, 1997). Experimentswith children ranging from preschool to highschool age reinforce the claim that positiveemotions build intellectual resources by show-ing that induced positive states—in comparisonto neutral and negative states—produce fasterlearning and improved intellectual perfor-mance (Bryan & Bryan, 1991; Bryan, Mathur,& Sullivan, 1996; Masters, Barden, & Ford,1979). Finally, correlational studies with bothhumans and nonhuman mammals suggest thatsocial play builds enduring social relation-ships (Boulton & Smith, 1992; Lee, 1983;Martineau, 1972). Research from our lab hasshown that positive emotions, experiencedover the course of a month, predict whether anew acquaintance becomes a friend (Waugh &Fredrickson, 2006). Mutually supportive socialrelationships in turn predict longevity (Brown,Nesse, & Vinokur, 2003), and in times of needthey can directly influence survival.

In a direct test of the build hypothesis, werandomly assigned working adults to an inter-vention to increase daily experiences of positiveemotions over the course of 8 weeks. Partici-pants in the experimental group were trained inloving-kindness meditation—a practice that issimilar to mindfulness meditation (Davidson etal., 2003; Kabat-Zinn, 2005), but that focuseson deliberately generating broadened mindsetsand the positive emotions of compassion andlove. They were compared to those assigned toa wait list for the same meditation workshop.All participants in this study reported daily ontheir experience of several discrete positive andnegative emotions, and filled out pre- andpostintervention batteries to assess psychologi-cal, social, mental, and physical resources, plustheir life satisfaction.

After 3 weeks of practice, participants in themeditation group began experiencing higherdaily levels of various positive emotions thanthose in the wait-list control group. After 8weeks, these participants also showed increasesin a number of personal resources, includingphysical wellness, agency for achieving impor-tant goals, ability to savor positive experiences,and quality of close relationships. These gainswere mediated by increased positive emotion.


Finally, meditators showed an increase relativeto the control group in life satisfaction, and thisgain was mediated by the increase in resources(see Figure 48.3). These results provide strongand specific evidence for the build hypothesis:Positive emotions led people to build a varietyof important resources, and these resourcesproved valuable in increasing their life satisfac-tion and functioning in general (Fredrickson,Cohn, Coffey, Pek, & Finkel, 2008).

Positive Emotions and Stress

Researchers have also examined the ways inwhich positive emotions affect coping withchronic stressors. Prolonged negative situationssuch as bereavement or joblessness evoke nega-tive emotions, but cannot be solved by the kindof immediate, narrowly defined action thatnegative emotions encourage. Consistent withthis view, studies have shown that people whoexperienced some level of positive emotionsduring bereavement (alongside their negativeemotions) showed greater psychological well-being a year or more later, and that this oc-curred partly because positive emotions wereassociated with the ability to take a longer viewand develop plans and goals for the future(Stein, Folkman, Trabasso, & Richards, 1997;Moskowitz, Folkman, & Acree, 2003). Simi-larly, we performed a longitudinal assessmentof college students’ emotions and mental healthbefore and after the terrorists attacks of Sep-tember 11, 2001 (Fredrickson, Tugade, Waugh,& Larkin, 2003). We found that precrisis trait

resilience predicted psychological growth andreduced risk of depression, but that this differ-ence was fully mediated by experiences of posi-tive emotion in the wake of the attacks. Resil-ient participants fared better, and this wasbecause they were more likely than nonresilientparticipants to experience positive emotions.Resilient participants were not devoid of nega-tive emotions—they felt fear and grief, much astheir less resilient peers did—but finding occa-sional opportunities to feel positive emotionsseems to have alleviated some of the negativeeffects of a prolonged narrowed mindset. Itmay be difficult to find positive emotions thatare fully appropriate while in a prolonged neg-ative situation; perhaps this is why people whocan feel a wider variety of finely differentiatedpositive emotions show greater psychologicalresilience (Tugade, Fredrickson, & Barrett,2004).

These results contradict common-sense criti-cisms that positive emotions are unhelpful orinappropriate for people in negative circum-stances: Even adults dealing with suicidalthoughts (Joiner, Pettit, Perez, & Burns, 2001)or disclosure of childhood sexual abuse(Bonanno et al., 2002) showed better copingwhen some degree of positivity accompaniedtheir painful feelings. In a longitudinal study ofcollege students coping with ordinary life prob-lems (Fredrickson & Joiner, 2002), we foundthat state positive emotions correlated with theuse of creative and broad-minded coping strat-egies, and that use of these strategies in turnpredicted increased positive emotions 5 weeks


FIGURE 48.3. Results from a positive emotions intervention study. The intervention increased daily pos-itive emotions, which led to building physical resources (top box), psychological resources (middle box),and social resources (bottom box). Resource building, in turn, led to increased life satisfaction.

later (above and beyond initial level of positiveemotion).

The literature on depression has long docu-mented a downward spiral in which depressedmood and the narrowed, pessimistic thinking itengenders feed into one another, leading toever-worsening moods and even clinical levelsof depression (Peterson & Seligman, 1984).The findings we have reviewed suggest thatpositive emotions can disrupt the spiral, and insome cases even initiate a comparable upwardspiral: Positive emotions make people feel goodin the present moment, which broadens theirmindsets, which allows them to build resourcesthat increase their likelihood of feeling good inthe future. The upward spiral effect is repre-sented by the feedback loop in Figure 48.1.

Positive Emotions and Health

One of the newest frontiers in positive emo-tions research is evidence linking positive emo-tions to physiology. People who experiencehigh levels of positive emotions tend to experi-ence less pain and disability related to chronichealth conditions (Gil et al., 2004), to fight offillness and disease more successfully (Cohen &Pressman, 2006; Ong & Allaire, 2005), andeven to live longer (Danner et al., 2001; Levyet al., 2002; Moskowitz, 2003; Ostir et al.,2000). We believe that these findings may beexplained by the ability of positive emotions tolift people out of stressed, narrowed states. Itis already established that the physiologicalchanges accompanying negative emotions arebeneficial for decisive, short-term action, butdetrimental to long-term health. If biochem-icals such as cortisol and epinephrine remainelevated for too long, they can lead to physicaldeterioration and immune dysregulation(Sapolsky, 1999). In line with this, individualswhose biological stress markers take longer toreturn to baseline following a stressor showpoorer health than those who return to base-line quickly (McEwen & Seeman, 1999). Thuswe have explored whether positive emotionsplay a role in physiological as well as psycho-logical well-being.

The basic observation that positive and neg-ative emotions (or key components of them)are to some extent incompatible has a long lin-eage in work on anxiety disorders (e.g., system-atic desensitization—Wolpe, 1958), motivation(e.g., opponent-process theory; Solomon &Corbit, 1974), and aggression (e.g., the princi-

ple of incompatible responses; Baron, 1976).4

Even so, the mechanism ultimately responsiblefor this incompatibility has not been ade-quately identified. The broaden effect may turnout to be the mechanism. Negative emotionsare linked to preparation for a specific action,and if a positive emotion dissipates that focus,then the negative emotion loses its hold.

Because negative emotions were shaped byevolution to deal with pressing threats, wewould not expect them to be easily neutralized;this may be the basis for the “bad-is-stronger-than-good” effect referenced earlier (Bau-meister et al., 2001; Rozin & Royzman, 2001;Tversky & Kahneman, 1981). Instead, we havehypothesized that the “undo effect” may func-tion when the threat is no longer apparent, butthe narrowed focus and biochemical responseto threat nonetheless linger.

Our laboratory has tested this by first induc-ing a high-arousal negative emotion in all par-ticipants, removing the negative stimulus, andthen immediately inducing a randomly as-signed emotion by showing a short, emotion-ally evocative film clip. Participants in the twopositive emotion conditions (mild joy and con-tentment) exhibited faster cardiovascular re-covery than those in the neutral control condi-tion, who recovered more quickly than those inthe sadness condition (Fredrickson, Mancuso,Branigan, & Tugade, 2000, Study 1; see alsoFredrickson & Levenson, 1998). When partici-pants viewed the positive films without a pre-ceding stressor, they had no cardiovascular ef-fects of any kind (Fredrickson et al., 2000,Study 2). In other words, the positive and neu-tral films did not differ in what they did to thecardiovascular system, but they differed inwhat they could undo within this system.

In subsequent work, we discovered that self-reported levels of psychological resilience(Block & Kremen, 1996) correlate with speedof cardiovascular recovery, and that this rela-tionship is mediated by self-generated levels ofpositive emotions during recovery (Tugade &Fredrickson, 2004). Resilient individuals seemto be experts at harnessing the undo effect ofpositive emotions on their own.

We take these laboratory experiments as amicrocosm for the influence of emotions oncoping, and of coping on health. Imagine thatsome individuals typically seek positive emo-tions to help them bounce back quickly fromlife’s stressors, while others spend more time re-maining physiologically activated and prepared


to react, even after the threat is gone. Overtime, the latter group will accumulate morephysiological wear and tear, and will be morevulnerable to a wide range of stress-related ill-nesses (McEwen & Seeman, 1999; Kiecolt-Glaser et al., 2002). Whether the undo effect ofpositive emotions factors into long-term healthin this way is a challenging but deeply impor-tant question for future research.

DIRECTIONS FORFUTURE RESEARCH

There is growing evidence that the broaden-and-build theory describes the most basic formand function of positive emotions, but there isalso much yet to learn. In this section, we de-scribe some critical directions for future re-search.

Physiological andNeurological Connections

We are eager to see empirical findings on posi-tive emotions embedded in a broader physio-logical context. We know that positive emo-tions such as liking and intimacy can lead tocloser relationships (Waugh & Fredrickson,2006; Aron, Normal, Aron, McKenna, &Heyman, 2000). Recent work suggests thatthese emotions are linked to heightened levelsof the hormones oxytocin (Zak, Kurzban, &Matzner, 2005; Gonzaga, Turner, Keltner,Campos, & Altemus, 2006) and progesterone(Schultheiss, Wirth, & Stanton, 2004), and thatexogenous oxytocin can induce positive socialfeelings (Kosfelt, Heinrichs, & Zak, 2005).Further research on this relationship will allowus to bring the human research on relationshipformation and the predominantly nonhumanresearch on hormone release and behavior tobear on each other, and to expand our under-standing of the antecedents and consequencesof relationship-relevant emotions.

The broaden effect is also amenable to physi-ological exploration. A review by Ashby, Isen,and Turken (1999) suggests that the broadeneffect may be associated with release of meso-limbic dopamine, which enhances cognitiveflexibility, set switching, and proactive curios-ity. Notably, this is the same neurological sys-tem Berridge and Robinson (2003) associatewith the motivational component of positiveaffect. It is also the mesolimbic dopamine sys-

tem that is impaired by older antipsychoticdrugs, which lead to notable cognitive narrow-ing and rigidity (Berger et al., 1989). There isno doubt that the neurological substrate of thebroaden effect will turn out to be more com-plex than a single neurotransmitter or neuronalsystem, but Ashby et al.’s observations helppave the way for future investigation.

Scattered results linking neurology and posi-tive emotional effects are emerging in otherareas. Haidt (2005) has early but sugges-tive evidence linking elevation and othermoral emotions to changes in vagal tone. Re-sults from neuroimaging studies demonstrateheightened left-hemispheric activation both inpersons experiencing short-term positive emo-tions, and tonically in individuals with highertrait positive emotionality (Davidson, 2004).Our work on the undo effect (Fredrickson &Levenson, 1998; Fredrickson et al., 2000) dem-onstrates that positive emotions can reduce theduration of cardiovascular response evoked bya stressor. We are currently undertaking workto examine relationships between psychophysi-ological measures and meditation outcomes,and among positive emotions, affiliation-related hormones, and relationship formation.

Interventions

Our loving-kindness meditation interventionwas the first to investigate the effects of posi-tive emotions in light of the broaden-and-buildtheory, but other investigators have also foundways to create relatively lasting increases inpositive emotions. Emmons and McCullough(2003) designed an intervention based oncounting blessings; Sheldon and Lyubomirsky(2006) combined counting blessings with visu-alizing one’s best possible self; and Seligman,Steen, and Park (2005) had participants countblessings, express gratitude, and practice usingtheir signature strengths. Practitioners ofcognitive-behavioral therapy also have a stableof techniques for generating positive emotions(Beck, 1995), although these are usually testedonly among depressed individuals.

These early efforts might best be seen asproof of the concepts that positive emotionscan be reliably evoked and that they have sig-nificant effects beyond momentary hedonic sat-isfaction. Next, positive emotion interventionsneed to become more mature: We can deter-mine how best to deliver them, which interven-tions function for different populations, and


how to maximize their effectiveness while min-imizing their cost and time commitment. Also,larger studies could test long-term health out-comes, relationships with others, and school orwork outcomes, in addition to the personaloutcomes already investigated. Unlike inter-ventions that focus on specific problems andtherefore address only a specific group of peo-ple (e.g., those with depression, drug depend-ence, or extreme aggression), interventions thatteach how to generate positive emotions couldbe of value to nearly anyone, in any situation.Therefore, they pose a wider variety of theoret-ical challenges, and possibly offer greater po-tential benefits as well.

Properties of Specific Emotions

So far, the empirical evidence suggests that thebroaden effect is common to many positiveemotions, and may describe their most generalshared effect on cognition and attention. How-ever, different positive emotions should alsohave distinct thought–action repertoires, sub-jective components, and physiological effects.For example, Tiedens and Linton (2001) com-pared the cognitive effects of contentment andpleasant surprise along a certainty–uncertaintycontinuum, and Gonzaga et al. (2006) studiedcognitive, behavioral, and biological distinc-tions among romantic love, friendship, andsexual desire. Tong (2007) found that distin-guishing different positive emotions may re-quire attention to dimensions of experiencethat we have not previously thought of as in-herent to emotion, such as social connectionand spiritual experience.

However, most work on discrete positiveemotions remains inconclusive (Barrett, 2006).We have suggested, for example, that joy is de-signed to encourage physical play and explora-tion, and so has a physiological arousal compo-nent (Fredrickson, 1998). Gratitude seems toencourage reflection and reconceptualization ofone’s experiences, targeted toward encouragingreciprocal behavior in the future (McCullough,Kilpatrick, Emmons, & Larson, 2001; Bartlett& DeSteno, 2006), and may differ from the re-lated negative emotion of indebtedness by in-volving more insight into the perspective and de-sires of the benefactor (F. Flynn, personalcommunication, April 2006; Fredrickson,2004). Keltner and Haidt (2003) suggest thatawe functions to facilitate a Piagetian process of“accommodation,” in which our mental struc-

tures enlarge to incorporate an important andwholly new experience; Haidt (2003) has sug-gested that there is a phenomenologically dis-tinct and generally recognized emotion called“elevation,” which causes people to attend toand emulate other people who show skillful ormorally excellent behavior (i.e., people broadentheir self-concept to incorporate the others’good example). Empirical tests of these propos-als are not difficult to imagine, and at present thefield is wide open for exploration.

Models

The broaden-and-build theory provides a de-scription of the short-term effects of positiveemotions and a plausible mechanism for long-term growth, but at a very general level. Thereis much more to learn about how the broadeneffect works and what it does in specific situa-tions. How does a broad mindset affect per-ceived familiarity with and interest in a new re-lationship partner? In a learning situation, is itlikely to increase interest in the topic at hand,or increase the tendency to switch between top-ics? In what situations does it lead to use ofheuristics versus careful processing?

Similarly, the build effect is difficult to ob-serve directly. We have demonstrated that posi-tive emotions lead people to build personal re-sources, but what specific actions lead to thisgrowth? What situational or personal charac-teristics beyond positive emotions encouragebehavior that leads to new resources? Becausepositive emotions encourage novel, divergentbehavior that leads to long-term benefits, wemay find that traditional methods based on asingle manipulation and a single, clear outcomeare inadequate to capture all their effects. In-stead, we may need analytical methods thattake into account individuals’ specific re-sources, their adaptation to their situation, andthe opportunities they can detect in theirenvironment. Our lab has collaborated withMarcial Losada, who has used nonlinear dy-namic models (similar to those used to analyzehighly complex systems, such as weatherand stock market fluctuations) to representmoment-by-moment interactions within high-and low-performing business teams. People ex-periencing positive emotions and teams achiev-ing high performance show patterns character-istic of complex, multiply determined systems,whereas people experiencing fewer positiveemotions and teams with lower performance


are characterized by less complex, more pre-dictable patterns (Schuldberg & Gottlieb,2002; Losada & Heaphy, 2004). Our analysissuggests that the shift to novel behavior and re-source building may begin precipitously whenthe ratio between positive and negative emo-tions exceeds a critical threshold of about 3:1(Fredrickson & Losada, 2005), and early evi-dence supports this view (Cohn et al., 2007;Waugh & Fredrickson, 2006). Using thesemethods, we may be able to look more closelyat the dynamics of resource building within asingle observation session, or to recognize thecourse that the build effect takes over time,even when we are limited to a few discrete as-sessments within a longitudinal study.

Limitations of Positive Emotions

Most of the research we have surveyed dis-cusses benefits of positive emotions, but a fullscientific understanding should include pitfallsand boundary conditions of these benefits. Forexample, although more securely attachedinfants show faster cognitive development, in-secure attachment may be an appropriate re-sponse to threatening or unreliable family cir-cumstances. Positive emotions help undo thelingering cardiovascular effects of stressors, butpeople in particularly dangerous circumstancesmay have good reason to remain ready to act,even when a threat appears to be gone. Addi-tionally, there is some suggestion that ex-tremely high levels of positive emotions, un-tempered by sufficient negative emotions, candegrade performance (Diener, 2004; Fredrick-son & Losada, 2005). We would like to de-velop a nuanced understanding of when posi-tive emotions can help resolve a negativesituation (as in Stein et al., 1997; Moskowitz etal., 2003; Fredrickson et al., 2003), and whenthey might be dangerous, excessive, or unac-ceptably costly.

Regardless of these caveats, we believe thatthere is good reason to celebrate and encouragepositive emotions. Research on critical ratios ofpositivity to negativity (Fredrickson & Losada,2005; Gottman, 1994) suggests that nearly allindividuals and groups have low ratios. Thereare important questions about when and howto experience positive emotions, and whichemotions are appropriate in differentsituations—but few of us are fortunate enoughto have the problem of simply experiencing toomuch joy, interest, contentment, and love.

CONCLUSION

At the most general level, the broaden-and-build theory offers three new avenues for ex-ploration:

1. For researchers on positive emotions, itprovides an empirical framework for investi-gating the dynamics of such emotions andmore clearly specifying their effects. Whetherthe theory is corroborated “as is” or whether itundergoes substantial changes, it is currentlythe best starting point for learning more.

2. For emotion theorists, the broaden-and-build theory suggests that the action tendenciesmodel be expanded to include not just the nar-rowed repertoires associated with negativeemotions, but also the broadened, less predict-able repertoires associated with positive emo-tions. Psychology’s understanding of emotionwill be hampered if we deal exclusively withresource-consuming emotions that operate in ashort time frame, and not with resource-building emotions that show their adaptive ef-fects over longer periods.

3. Finally, the broaden-and-build theorysuggests that positive emotions are fruitful tar-gets for basic and applied research in any fieldthat deals with personal growth, change, learn-ing, social coordination, or physical health. Aswe have seen, positive emotions are more thanjust desirable endpoints; they also affect a widevariety of cognitive, social, and biological pro-cesses, and can help people to build a variety ofpersonal resources. Right now we have a basicscheme for learning more about how positiveemotions work and how they affect other areasof life; we look forward to broadening andbuilding on it as research continues.

ACKNOWLEDGMENTS

Barbara L. Fredrickson’s research on positive emo-tions is supported by the National Institute of Men-tal Health (Grant No. MH59615).

NOTES

1. The PANAS was formerly the Positive and NegativeAffect Schedule. The authors renamed it to indicatethat it actually measures valence mostly for high-activation states (Watson et al., 1999).

2. Note that a life-saving response need not involve ac-tion per se. For example, evolutionary research sug-


gests that sadness may be adaptive because it pre-vents action, thus conserving resources at times whenprospects are so poor that no available action wouldbe worth the risk and/or calories it requires (Keller &Nesse, 2005).

3. There is strong evidence that face recognition typi-cally relies on a holistic (broad) representation of theface, and that impaired other-race face recognitionresults partly from a focus on narrowly construed ra-cial features (Johnson & Fredrickson, 2005).

4. We previously reviewed studies in which frightenedor grieving individuals reported both positive andnegative emotions. The incompatibility of positiveand negative emotions relates to feelings in a singlemoment, whereas the participants in bereavementstudies were surveyed about their emotions over along period of time, which can encompass many indi-vidual episodes of different emotions.

REFERENCES

Abele, A. (1992). Positive versus negative mood influ-ences on problem solving. Polish Psychological Bul-letin, 23, 187–202.

Arend, R., Gove, F. L., & Sroufe, L. A. (1979). Continu-ity of individual adaptation from infancy to kinder-garten: A predictive study of ego-resiliency and curi-osity in preschoolers. Child Development, 50, 950–959.

Aron, A., Norman, C. C., Aron, E. N., McKenna, C., &Heyman, R. (2000). Couples shared participation innovel and arousing activities and experienced rela-tionship quality. Journal of Personality and SocialPsychology, 78, 273–283.

Ashby, F. G., Isen, A. M., & Turken, U. (1999). A neuro-psychological theory of positive affect and its influ-ence on cognition. Psychological Review, 106, 529–550.

Aspinwall, L. (1998). Rethinking the role of positive af-fect in self-regulation. Motivation and Emotion,22(1), 1–32.

Baron, R. A. (1976). The reduction of human aggres-sion: A field study of the influence of incompatiblereactions. Journal of Applied Social Psychology, 6,260–274.

Barrett, L. F. (2006). Emotions as natural kinds? Per-spectives on Psychological Science, 1, 28–58.

Barrett, L. F., Gross, J., Christensen, T. C., &Benvenuto, M. (2001). Knowing what you’re feelingand knowing what to do about it: Mapping the rela-tion between emotion differentiation and emotionregulation. Cognition and Emotion, 15, 713–724.

Bartlett, M. Y., & Desteno, D. (2006). Gratitude andprosocial behavior: Helping when it costs you. Psy-chological Science, 17, 319–325.

Basso, M. R., Schefft, B. K., Ris, M. D., & Dember, W.N. (1996). Mood and global–local visual processing.Journal of the International Neuropsychological So-ciety, 2, 249–255.

Baumeister, R. F., Bratslavsky, E., Finkenauer, C., &Vohs, K. D. (2001). Bad is stronger than good. Re-view of General Psychology, 5, 323–370.

Beck, J. S. (1995). Cognitive therapy: Basics and be-yond. New York: Guilford Press.

Berger, H. J. C., van Hoof J. J. M., van Spaendonck,K. P. M., Horstink, M. W. I., van den Bercken, J.H. L., Jaspers, R., et al. (1989). Haloperidoland cognitive shifting. Neuropsychologia, 27, 629–639.

Berridge, K. C., & Robinson, T. E. (2003). Parsing re-ward. Trends in Neuroscience, 26, 507–513.

Bless, H., Clore, G., Schwarz, N., Golisano, V., Rabe,C., & Wölk, M. (1996). Mood and the use of scripts:Does a happy mood really lead to mindlessness?Journal of Personality and Social Psychology, 71(4),665–679.

Block, J., & Kremen, A. M. (1996). IQ and ego-resilience: Conceptual and empirical connections andseparateness. Journal of Personality and Social Psy-chology, 70, 349–361.

Bonanno, G. A., Keltner, D., Noll, J. G., Putnam, F. W.,Trickett, P. K., LeJune, J., et al. (2002). When the facereveals what words do not: Facial expressions ofemotion, smiling, and the willingness to disclosechildhood sexual abuse. Journal of Personality andSocial Psychology, 83, 94–110.

Boulton, M. J., & Smith, P. K. (1992). The social natureof play fighting and play chasing: Mechanisms andstrategies underlying cooperation and compromise.In J. H. Barkow, L. Cosmides, & J. Tooby (Eds.), Theadapted mind: Evolutionary psychology and the gen-eration of culture (pp. 429–444). New York: OxfordUniversity Press.

Brown, S. L., Nesse, R. M., Vinokur, A. D., & Smith, D.M. (2003). Providing social support may be morebeneficial than receiving it: Results from a prospec-tive study of mortality. Psychological Science, 14,320–327.

Bryan, T., & Bryan, J. (1991). Positive mood and mathperformance. Journal of Learning Disabilities, 24,490–494.

Bryan, T., Mathur, S., & Sullivan, K. (1996). The impactof positive mood on learning. Learning DisabilitiesQuarterly, 19, 153–162.

Cabanac, M. (1971). Physiological role of pleasure. Sci-ence, 173, 1103–1107.

Cacioppo, J. T., Gardner, W. L., & Berntson, G. G.(1999). The affect system has parallel and integrativeprocessing components: Form follows function. Jour-nal of Personality and Social Psychology, 76, 839–855.

Cacioppo, J. T., Priester, J. R., & Berntson, G. G.(1993). Rudimentary determinants of attitudes: II.Arm flexion and extension have differential effectson attitudes. Journal of Personality and Social Psy-chology, 65, 5–17.

Caro, T. M. (1988). Adaptive significance of play: Arewe getting closer? Tree, 3, 50–54.

Carver, C. S. (2003). Pleasure as a sign you can attend to


something else: Placing positive feelings within a gen-eral model of affect. Cognition and Emotion, 17,241–261.


Chen, M., & Bargh, J. (1999). Consequences of auto-matic evaluation: Immediate behavioral disposi-tions to approach or avoid the stimulus. Person-ality and Social Psychology Bulletin, 25(2) 215–224.

Clore, G. L. (1994). Why emotions are felt. In P. Ekman& R. Davidson (Eds.), The nature of emotion: Fun-damental questions (pp. 103–111). New York: Ox-ford University Press.

Cohen, S., Doyle, W. J., & Turner, R. B. (2003). Emo-tional style and susceptibility to the common cold.Psychosomatic Medicine, 65, 652–657.

Cohen, S., & Pressman, S. D. (2006). Positive affect andhealth. Current Directions in Psychological Science,15, 122–125.

Cohn, M. A., & Fredrickson, B. L. (2006). Beyond themoment, beyond the self: Shared ground between se-lective investment theory and the broaden-and-buildtheory of positive emotion. Psychological Inquiry,17, 39–44.

Cohn, M. A., Fredrickson, B. L., Brown, S. L., Mikels, J.A., & Conway, A. (2007). Happiness unpacked: Pos-itive emotions increase life satisfaction by buildingresilience. Manuscript in preparation.

Csikszentmihalyi, M. (1990). Flow: The psychology ofoptimal experience. New York: Harper & Row.

Csikszentmihalyi, M., & Csikszentmihalyi, I. (2006). Alife worth living: Contributions to positive psychol-ogy. New York: Oxford University Press.

Danner, D., Snowdon, D. A., & Friesen, W. V. (2001).Positive emotions in early life and longevity: Findingsfrom the nun study. Journal of Personality and SocialPsychology, 80, 804–813.

Davidson, R. J. (1993). The neuropsychology of emo-tion and affective style. In M. Lewis & J. M.Haviland (Eds.), Handbook of emotions (pp. 143–154). New York: Guilford Press.

Davidson, R. J. (2004). What does the prefrontal cortex‘do’ in affect?: Perspectives on frontal EEG asymme-try research. Biological Psychiatry, 67, 219–233.

Davidson, R. J., Kabat-Zinn, J., Schumacher, J.,Rosenkrantz, M., Muller, D., Santorelli, S. F., et al.(2003). Alterations in brain and immune functionproduced by mindfulness meditation. PsychosomaticMedicine, 65, 564–570.

Derryberry, D., & Tucker, D. M. (1994). Motivating thefocus of attention. In P. M. Niedenthal & S.Kitayama (Eds.), The heart’s eye: Emotional influ-ences in perception and attention (pp. 167–196). SanDiego, CA: Academic Press.

Diener, E. (2004, February). “Was Pollyanna right? De-termining when positive affect will be beneficial.”Talk presented at the meeting of the Society for Per-sonality and Social Psychology, Austin TX.

Diener, E., & Diener, C. (1996). Most people are happy.Psychological Science, 7, 181–185.

Diener, E., & Seligman, M. E. P. (2004). Beyond money:Toward an economy of well-being. Psychological Sci-ence in the Public Interest, 5, 1–31.

Dolhinow, P. J. (1987). At play in the fields. In H.Topoff (Ed.), The natural history reader in animalbehavior (pp. 229–237). New York: Columbia Uni-versity Press.

Dovidio, J., Gaertner, S., Isen, A., Rust, M., & Guerra,P. (1995). Positive affect and the reduction of inter-group bias. In C. Sedikides, J. Schopler, & C. A.Insko (Eds.), Intergroup cognition and intergroupbehavior (pp. 337–366). Mahwah, NJ: Erlbaum.

Dreisbach, G., & Goschke, T. (2004). How positiveaffect modulates cognitive control: Reduced per-severation at the cost of increased distractibility.Journal of Experimental Psychology: Memory andCognition, 30, 343–353.

Dunn, J. R., & Schweitzer, M. E. (2005). Feeling andBelieving: The influence of emotion on trust. Journalof Personality and Social Psychology, 88, 736–748.

Einon, D. F., Morgan, M. J., & Kibbler, C. C. (1978).Brief periods of socialization and later behavior inthe rat. Developmental Psychobiology, 11, 213–225.


Ellsworth, P. C., & Smith, C. A. (1988). Shades of joy:Patterns of appraisal differentiating pleasant emo-tions. Cognition and Emotion, 2, 301–331.

Emmons, R. A., & McCullough, M. E. (2003).Counting blessings vs. burdens: An experimental in-vestigation of gratitude and subjective well-being indaily life. Journal of Personality and Social Psychol-ogy, 84, 377–389.

Estrada, C. A., Isen, A. M., & Young, M. J. (1997). Pos-itive affect facilitates integration of information anddecreases anchoring in reasoning among physicians.Organizational Behavior and Human Decision Pro-cesses, 72, 117–135.

Fineman, S. (2006). On being positive: Concerns andcounterpoints. Academy of Management Review, 31,270–91.

Fredrickson, B. L. (1998). What good are positive emo-tions? Review of General Psychology, 2, 300–319.

Fredrickson, B. L. (2000). Cultivating positive emotionsto optimize health and well-being. Prevention andTreatment, 3. Article 0001a. Retrieved from con-tent.apa.org/journals/pre/3/1/1a.html


Fredrickson, B. L. (2004). Gratitude, like other positiveemotions, broadens and builds. In R. A. Emmons &M. E. McCullough (Eds.), The psychology of grati-tude (pp. 145–166). New York: Oxford UniversityPress.

Fredrickson, B. L., & Branigan, C. (2001) Positive emo-tions. In T. J. Mayne & G. A. Bonnano (Eds.), Emo-


tions: Current issues and future directions. NewYork: Guilford Press.

Fredrickson, B. L., & Branigan, C. (2005). Positiveemotions broaden thought–action repertoires: Evi-dence for the broaden-and-build model. Cogntionand Emotion, 19, 313–332.

Fredrickson, B. L., Cohn, M. A., Coffey, K. A., Pek, J.,& Finkel, S. M. (2008). Open hearts build lives: Posi-tive emotions, induced through loving-kindness med-itation, meditation, build consequential personal re-sources. Manuscript under review.

Fredrickson, B. L., & Joiner, T. (2000). Positive emo-tions trigger upward spirals toward emotional well-being. Psychological Science, 13, 172–175.

Fredrickson, B. L., & Levenson, R. W. (1998). Positiveemotions speed recovery from the cardiovascularsequelae of negative emotions. Cognition and Emo-tion, 12, 191–220.

Fredrickson, B. L., & Losada, M. F. (2005). Positive af-fect and the complex dynamics of human flourishing.American Psychologist, 60, 678–686.

Fredrickson, B. L., Mancuso, R. A., Branigan, C., &Tugade, M. (2000). The undoing effect of positiveemotions. Motivation and Emotion, 24(4), 237–258.

Fredrickson, B. L., Tugade, M. M., Waugh, C. E., &Larkin, G. (2003). What good are positive emotionsin crises? A prospective study of resilience and emo-tions following the terrorist attacks on the UnitedStates on September 11, 2001. Journal of Personalityand Social Psychology, 84, 365–376.


Frijda, N. H. (1994). Emotions are functional, most ofthe time. In P. Ekman & R. Davidson (Eds.), The na-ture of emotion: Fundamental questions (pp. 112–122). New York: Oxford University Press.

Frijda, N. H., Kuipers, P., & Schure, E. (1989). Rela-tions among emotion, appraisal, and emotional ac-tion readiness. Journal of Personality and Social Psy-chology, 57, 212–228.

Gable, S. L., Reis, H. T., Impett, E. A., & Asher, E. R.(2004). What do you do when things go right?: Theintrapersonal and interpersonal benefits of sharingpositive events. Journal of Personality and SocialPsychology, 87, 228–245.

Gervais, M., & Wilson, D. S. (2005). The evolution andfunctions of laughter and humor: A synthetic ap-proach. Quarterly Review of Biology, 80, 395–451.

Gil, K. M., Carson, J. W., Porter, L. S., Scipio, C.,Bediako, S. M., & Orringer, E. (2004). Daily moodand stress predict pain, health care use, and work ac-tivity in African American adults with sickle-cell dis-ease. Health Psychology, 23, 267–274.

Gonzaga, G. C., Turner, R. A., Keltner, D., Campos, B.,& Altemus, M. (2006). Romantic love and sexual de-sire in close relationships. Emotion, 6, 163–179.

Gottman, J. M. (1994). What predicts divorce?: The re-lationship between marital processes and maritaloutcomes. Hillsdale, NJ: Erlbaum.

Haidt, J. (2003). Elevation and the positive psychologyof morality. In C. L. M. Keyes & J. Haidt (Eds.),Flourishing: Positive psychology and the life well-lived (pp. 275–289). Washington, DC: American Psy-chological Association.

Haidt, J. (2005). The happiness hypothesis: Finding themodern truth in ancient wisdom. New York: BasicBooks.

Hazen, N. L., & Durrett, M. E. (1982). Relationship ofsecurity of attachment and cognitive mapping abili-ties in 2-year-olds. Developmental Psychology, 18,751–759.

Isen, A. M., & Daubman, K. A. (1984). The influence ofaffect on categorization. Journal of Personality andSocial Psychology, 47, 1206–1217.


Isen, A. M., Johnson, M. M. S., Mertz, E., & Robinson,G. F. (1985). The influence of positive affect on theunusualness of word associations. Journal of Person-ality and Social Psychology, 48, 1413–1426.

Ito, T. A., & Cacioppo, J. T. (1999). The psycho-physiology of utility appraisals. In D. Kahneman, E.Diener, & N. Schwartz (Eds.), Well-being: Founda-tions of hedonic psychology (pp. 470–488). NewYork: Russell Sage Foundation.

Johnson, K. J. (2005). We all look the same to me: Posi-tive emotions eliminate the own-race bias in face rec-ognition. Unpublished doctoral dissertation, Univer-sity of Michigan.

Johnson, K. J., & Fredrickson, B. L. (2005). “We alllook the same to me:” Positive emotions eliminatethe own-race bias in face recognition. PsychologicalScience, 16, 875–881.

Joiner, T. E., Pettit, J. W., Perez, M., & Burns, A. B.(2001). Can positive emotion influence problem-solving among suicidal adults? Professional Psychol-ogy: Research and Practice, 32, 507–512.

Kabat-Zinn, J. (2005). Full catastrophe living: Using thewisdom of your body and mind to face stress, pain,and illness (15th anniversary ed.). New York: Ban-tam.

Kahn, B. E., & Isen, A. M. (1993). The influence of pos-itive affect on variety seeking among safe, enjoyableproducts. Journal of Consumer Research, 20, 257–270.

Kahneman, D., Kreuger, A. B., & Schkade, D. A.(2004). A survey method for characterizing daily lifeexperience: The day reconstruction method. Science,306, 1776–1780.

Keller, M. C., & Nesse, R. M. (2005). Is low mood anadaptation?: Evidence for subtypes with symptomsthat match precipitants. Journal of Affective Disor-ders, 86, 27–35.

Keltner, D., & Bonanno, G. A. (1997). A study of laugh-ter and dissociation: Distinct correlates of laughter


and smiling during bereavement. Journal of Person-ality and Social Psychology, 73, 687–702.

Keltner, D., & Haidt, J. (2003). Approaching awe, amoral, spiritual, and aesthetic emotion. Cognitionand Emotion, 17, 297–314.

Keyes, C. L. M., & Haidt, J. (Eds.). (2003). Flourishing:Positive psychology and the life well-lived. Washing-ton, DC: American Psychological Association.

Kiecolt-Glaser, J. K., McGuire, L., Robles, T. F., &Glaser, R. (2002). Emotions, morbidity, and mortal-ity. Annual Review of Psychology, 53, 83–107.

Kosfelt, M., Heinrichs, M., & Zak, P. J. (2005). Oxy-tocin increases trust in humans. Nature, 435, 673–676.


Lazarus, R. S. (2003). Does the positive psychologymovement have legs? Psychological Inquiry, 14, 93–109.

Lee, P. C. (1983). Play as a means for developing rela-tionships. In R. A. Hinde (Ed.), Primate social rela-tionships (pp. 82–89). Oxford, UK: Blackwell.

Levenson, R. W. (1994). Human emotions: A functionalview. In P. Ekman & R. Davidson (Eds.), The natureof emotion: Fundamental questions (pp. 123–126).New York: Oxford University Press.

Levy, B. R., Slade, M. D., & Kunkel, S. R. (2002). Lon-gevity increased by positive self-perceptions of aging.Journal of Personality and Social Psychology, 83,261–270.

Losada, M., & Heaphy, E. (2004). The role of positivityand connectivity in the performance of businessteams: A nonlinear dynamics model. AmericanBehavioral Scientist, 47(6), 740–765.

Lyubomirsky, S. L., King, L., & Diener, E. (2005). Thebenefits of frequent positive affect: Does happinesslead to success? Psychological Bulletin, 131, 803–855.

Martin, L. L., & Clore, G. (Eds.). (2001). Theories ofmood and cognition: A user’s guidebook. Mahwah,NJ: Erlbaum.

Martineau, W. H. (1972). A model of the social func-tions of humor. In J. H. Goldstein & P. E. McGee(Eds.), The psychology of humor: Theoretical per-spectives and empirical issues (pp. 101–128). NewYork: Academic Press.

Masters, J. C., Barden, R. C., & Ford, M. E. (1979). Af-fective states, expressive behavior, and learning inchildren. Journal of Personality and Social Psychol-ogy, 37, 380–390.

Matas, L., Arend, R. A., & Sroufe, L. A. (1978). Conti-nuity of adaptation in the second year: The relation-ship between quality of attachment and later compe-tence. Child Development, 49, 547–556.

McCullough, M. E., Kirkpatrick, S. D., Emmons, R. A.,& Larson, D. B. (2001). Is gratitude a moral affect?Psychological Bulletin, 127, 249–266.

McEwen, B. S., & Seeman, T. (1999). Protective anddamaging effects of mediators of stress: Concepts of

allostasis and allostatic load. Annals of the New YorkAcademy of Sciences, 896, 30–47.

Mikulincer, M. (1997). Adult attachment style and in-formation processing: Individual differences in curi-osity and cognitive closure. Journal of Personalityand Social Psychology, 72, 1217–1230.

Moskowitz, J. T. (2003). Positive affect predicts lowerrisk of AIDS mortality. Psychosomatic Medicine, 65,620–626.

Moskowitz, J. T., Folkman, S., & Acree, M. (2003). Dopositive psychological states shed light on recoveryfrom bereavement?: Findings from a 3-year longitu-dinal study. Death Studies, 27, 471–500.

Oatley, K., & Jenkins, J. M. (1996). Understandingemotions. Cambridge, MA: Blackwell.

Ong, A. D., & Allaire, J. C. (2005). Cardiovascularintraindividual variability in later life: The influenceof social connectedness and positive emotions. Psy-chology and Aging, 20, 476–485.

Ostir, G. V., Markides, K. S., & Black, S. A. (2000).Emotional well-being predicts subsequent functionalindependence and survival. Journal of the AmericanGeriatrics Society, 48, 473–78.

Otake, K., Waugh, C. E., & Fredrickson, B. L. (2007).Positive emotions unlock other-focused thinking.Manuscript in preparation.

Panksepp, J. (1998). Attention deficit hyperactivity dis-orders, psychostimulants, and intolerance of child-hood playfulness: A tragedy in the making? CurrentDirections in Psychological Science, 7, 91–98.

Peciña, S., Smith, K. S., & Berridge, K. C. (2006).Hedonic hot spots in the brain. Neuroscientist, 12,500–511.

Peterson, C., & Seligman, M. E. P. (1984). Causal expla-nations as a risk factor for depression: Theory andevidence. Psychological Review, 91, 347–374.

Rosenberg, E. L. (1998). Levels of analysis and the orga-nization of affect. Review of General Psychology, 2,247–270.

Rowe, G., Hirsch, J., & Anderson, A. K. (2007). Posi-tive affect increases the “breadth” of attentional se-lection. Proceedings of the National Academy of Sci-ences of the USA, 104, 338–383.

Rozin, P., & Royzman, E. B. (2001). Negativity bias,negativity dominance, and contagion. Personalityand Social Psychology Review, 5, 296–320.

Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989).Affect Grid: A single-item scale of pleasure andarousal. Journal of Personality and Social Psychol-ogy, 57, 493–502.

Ryff, C. D., & Singer, B. (1998). The contours of posi-tive human health. Psychological Inquiry, 9, 1–28.

Sapolsky, R. M. (1999). The physiology and pathophys-iology of unhappiness. In D. Kahneman, E. Diener, &N. Schwarz (Eds.), Well-being: The foundations ofhedonic psychology (pp. 453–469). New York: Rus-sell Sage Foundation.

Schuldberg, D., & Gottlieb, J. (2002). Dynamics andcorrelates of microscopic changes in affect. Nonlin-


ear Dynamics, Psychology, and Life Sciences, 6, 231–257.

Schultheiss, O. C., Wirth, M. M., & Stanton, S. J.(2004). Effects of affiliation and power motivationarousal on salivary progesterone and testosterone.Hormones and Behavior, 46, 592–599.

Seligman, M. E. P., & Csikszentmihalyi, M. (2000). Pos-itive psychology: An introduction. American Psy-chologist, 55, 5–14.

Seligman, M. E. P., Steen, T. A., & Park, N. (2005). Pos-itive psychology progress: Empirical validation of in-terventions. American Psychologist, 60, 410–421.

Sheldon, K. M., & Lyubomirsky, S. (2006). How to in-crease and sustain positive emotion: The effects ofexpressing gratitude and visualizing best possibleselves. Journal of Positive Psychology, 1, 73–82.

Sherrod, L. R., & Singer, J. L. (1989). The developmentof make-believe play. In J. Goldstein (Ed.), Sports,games and play (pp. 1–38). Hillsdale, NJ: Erlbaum.

Shiota, M. N., Keltner, D., & John, O. P. (2006). Posi-tive emotion dispositions differentially associatedwith Big Five personality and attachment style. Jour-nal of Positive Psychology, 1, 61–71.

Simons, C. J. R., McCluskey-Fawcett, K. A., & Papini,D. R. (1986). Theoretical and functional perspectiveon the development of humor during infancy, child-hood, and adolescence. In L. Nahemow, K. A.McCluskey-Fawcett, & P. E. McGhee (Eds.), Humorand aging (pp. 53–77). Orlando, FL: Academic Press.

Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cog-nitive appraisal in emotion. Journal of Personalityand Social Psychology, 48, 813–838.

Solomon, R. L., & Corbit, J. D. (1974). An opponent-process theory of motivation: I. Temporal dynamicsof affect. Psychological Review, 81, 119–145.

Stein, N. L., Folkman, S., Trabasso, T., & Richards, T.A. (1997). Appraisal and goal processes as predictorsof psychological well-being in bereaved caregivers.Journal of Personality and Social Psychology, 72,872–884.

Tiedens, L. Z., & Linton, S. (2001). Judgment underemotional certainty and uncertainty: The effects ofspecific emotions on information processing. Journalof Personality and Social Psychology, 81, 973–988.

Tong, E. M. W. (2007). Appraisal processes in emo-tional experiences. Dissertation Abstracts Interna-

tional: Section B: The Sciences and Engineering, 67,4161.

Tooby, J., & Cosmides, L. (1990). The past explains thepresent: Emotional adaptations and the structure ofancestral environments. Ethology and Sociobiology,11, 375–424.

Trope, Y., & Pomerantz, E. (1998). Resolving conflictsamong self-evaluative motives: Positive experiencesas a resource for overcoming defensiveness. Motiva-tion and Emotion, 22(1), 53–72.


Tugade, M. M., & Fredrickson, B. L. (2004). Resilientindividuals use positive emotions to bounce backfrom negative emotional experiences. Journal of Per-sonality and Social Psychology, 86, 320–333.

Tugade, M. M., Fredrickson, B. L., & Barrett, L. F.(2004). Psychological resilience and positive emo-tional granularity: Examining the benefits of positiveemotions on coping and health. Journal of Personal-ity, 72, 1161–1190.

Tversky, A., & Kahneman, D. (1981). The framing ofdecisions and the psychology of choice. Science, 211,453–458.

Wadlinger, H. A., & Isaacowitz, D. M. (2006). Positivemood broadens visual attention to positive stimuli.Motivation and Emotion.


Waugh, C. E., & Fredrickson, B. L. (2006). Nice to knowyou: Positive emotions, self–other overlap, and com-plex understanding in the formation of a new relation-ship. Journal of Positive Psychology, 1, 93–106.

Waugh, C. E., Hejmadi, A., Otake, K., & Fredrickson,B. L. (2006). Cross-cultural evidence that positiveemotions broaden views of self to include close oth-ers. Unpublished raw data.

Wolpe, J. (1958). Psychotherapy by reciprocal inhibi-tion. Stanford, CA: Stanford University Press.

Zak, P. J., Kurzban, R., & Matzner, W. T. (2005).Oxytocin is associated with human trustworthiness.Hormones and Behavior, 48, 522–527.


C H A P T E R 4 9

Sadness and Grief

GEORGE A. BONANNO, LAURA GOORIN, and KARIN G. COIFMAN

Sadness and grief are universal and inevitableaspects of the human experience. They are con-ceptually and phenomenally similar—so simi-lar, in fact, that sometimes the terms are usedinterchangeably. But it is important to under-score that sadness and grief are not the same,and that their distinction helps illuminateunique and important aspects of human behav-ior. We begin this chapter by detailing four cru-cial ways in which sadness and grief differ. Wenext review the functional role played by sad-ness in normal human experience, and we de-scribe how, when those functions go awry, sad-ness can become depression. We then considerhow similar patterns may characterize the nor-mative experience of sadness in response to thedeath of a loved one, and how bereavement-related sadness can sometimes deteriorate intocomplicated grief reactions. To this end, wespeculate about the possible role played by sad-ness and grief in our ancestral past, and also inthe social life of nonhuman primates. Finally,we consider the role of positive emotional ex-periences after loss, as well as the growing

body of evidence regarding the oscillation ofpositive and negative affect and its role in nor-mal self-regulatory processes.

FOUR CRUCIAL DIFFERENCESBETWEEN SADNESS AND GRIEF

We begin our chapter with the question of howsadness and grief differ from each other. We ar-gue that sadness is a basic emotion, but thatgrief represents a broader and more elaborateconstruct similar to (though not identical with)depression (Bonanno, 2001; Lazarus, 1991).Grief and emotion each involve complexbehavioral responses whose respective opera-tional definitions have been subject to consid-erable debate. (For reviews of some of the keydefinitional issues regarding emotion, seeBarrett, 2006a, 2006b; Ekman, 1992; Ekman& Davidson, 1994; Izard, 1994; Oatley &Jenkins, 1996; and Russell, 1994. Regardinggrief, see Bonanno, 2001; Bonanno & Kalt-man, 1999; Hansson, Carpenter, & Fairchild,

797

1993; M. S. Stroebe, Hansson, Stroebe, &Schut, 2001; W. Stroebe, Schut, & Stroebe,2005.)

Grief is typically a highly emotionally dis-tressing experience, and at a superficial level itappears to share features with specific emo-tions, most notably sadness (Lazarus, 1991).Perhaps for this reason, some investigatorshave preferred to blur the distinction betweenthe concepts, and have viewed grief as a formof emotion (Averill & Nunley, 1993; Stearns &Knapp, 1996; Panksepp, 2005). However, acareful analysis reveals that grief is most appro-priately conceptualized as a complex and en-during molar experience that generates variousmolecular components, including a range ofspecific emotions (Bonanno, 2001; Lazarus,1991). Most prominent among the emotionsexperienced during grief is sadness.

There are at least four ways in which theemotion of sadness differs from the molar ex-perience of grief (Bonanno, 2001). First, sad-ness and grief each encompass dramaticallydifferent temporal intervals. Emotions, suchas sadness, are commonly defined as ephem-eral phenomena, generally lasting a few sec-onds but sometimes up to several hours(Ekman, 1984; Ekman & Davidson, 1994;Izard, 1993; Chow, Ram, Boker, Fujita, &Clore, 2005). In contrast, grief is an enduringstate that for most bereaved individuals per-sists for several weeks and up to several years(Bonanno, 2004; Bonanno & Kaltman,2001). In some cases, aspects of grief havebeen found to endure 7 to 8 years or longerafter the loss (Lehman, Wortman, & Wil-liams, 1987; Lundin, 1984).

Second, myriad different emotions typicallyoccur within the course of a single period ofgrief. Although the death of a loved one is mostcommonly associated with sadness, grieving isfar from a one-dimensional emotional phe-nomenon. In addition to sadness, grief has beenassociated with a wide range of negative emo-tions, such as anger, contempt, hostility, fear,and guilt (Abraham, 1924; Belitsky & Jacobs,1986; Bonanno & Keltner, 1997; Bonanno,Mihalecz, & LeJeune, 1998; Bowlby, 1980;Cerney & Buskirk, 1991; Kavanagh, 1990;Lazare, 1989; Osterweis, Solomon, & Green,1984; Raphael, 1983)—and, as we discuss ingreater detail below, genuinely positive emo-tional experiences related to amusement, affec-tion, happiness, and pride (see Bonanno &Kaltman, 1999, 2001).

Third, grief and emotions are associatedwith different types of underlying meaningstructures. Emotions are typically linked to rel-atively simple, proximal appraisals related tothe immediate situational context. Emotion-related appraisals often encompass, for exam-ple, issues of personal danger or benefit, copingpotential, or their interaction with moti-vational states (Frijda, 1993; Lazarus, 1991;Roseman, Antoniou, & Jose, 1996). In manycases, emotional responses occur without thebenefit of even these simple cognitive apprais-als. For instance, the chemical and physical re-sponses associated with basic emotions, such asfear, can be triggered solely on the basis ofrapid, automated, subcortical processing ofcrude perceptual information (LeDoux, 1989,1996; Phelps, 2006). Moreover, there is emerg-ing evidence that the amygdala also respondsto sadness-related stimuli (Wang, McCarthy,Song, & LaBar, 2005).

The emotion of sadness in particular is gen-erally associated with the appraisal of perma-nent loss. Grief is, of course, also associatedwith the cognitive understanding of loss. How-ever, in contrast to the relatively simple waythis appraisal manifests itself in sadness, thesense of loss that informs grief is typically farmore profound and all-encompassing. Whenthe loss involves the death of someone of im-portance in a person’s life, the impact on mean-ing structures includes a dramatic impact onthat person’s identity and cognitive under-standing of the world (Schwartzberg & Janoff-Bulman, 1991) and the future (Horowitz et al.,1997; Lehman et al., 1987; Shuchter &Zisook, 1993). Indeed, bereaved people com-monly report that they feel as if “a piece of meis missing” (Kastenbaum, 1995; Shuchter &Zisook, 1993). These longer-term appraisalstypically encompass a bereaved person’s evalu-ation and understanding of the entire course ofbereavement, as well as major portions of hisor her own life (Bonanno & Kaltman, 1999,2001).

Fourth, grief and emotions evoke differenttypes of coping responses. Emotions are proxi-mal, and thus are generally implicated in proxi-mal, short-term coping responses aimed at ei-ther changing or maintaining the immediatepsychological or physical state (Gross & John,2003; Mauss, Levenson, McCarter, Wilhelm,& Gross, 2005). Indeed, emotion and proximalcoping are so intimately related that Folkmanand Lazarus (1988, 1990) described coping as


a mediator of emotion. Grieving, on the otherhand, typically evokes longer-term coping ef-forts aimed at ameliorating the enduring emo-tional upsets as well as myriad concrete disrup-tions wrought by the loss, such as changes insocial roles, economic situation, or familialconfiguration (Bonanno & Keltner, 1997; Laz-arus, 1991; Neimeyer, 2006; Neimeyer,Prigerson, & Davies, 2002; Shuchter &Zisook, 1993; W. Stroebe & Stroebe, 1987).

THE FUNCTIONS OF SADNESS

In this section, we consider the functional roleplayed by the everyday experience and expres-sion of sadness. A key adaptive function of sad-ness is to promote personal reflection followingthe irrevocable loss of a person or object of im-portance to the self (Lazarus, 1991). The expe-rience of sadness turns our attention inward,promoting resignation and acceptance (Izard,1977, 1993; Lazarus, 1991; Stearns, 1993).Physiological arousal is decreased, allowing fora “time out” to update cognitive structures andto accommodate lost objects (Welling, 2003).The reflective function of sadness, therefore,opportunely affords us a pause, allowing us totake stock and to revise our goals and plans(Bonanno & Keltner, 1997; Oatley & Johnson-Laird, 1996).

An extensive body of experimental data hasassociated sadness with more detail-orientedinformation processing, more accurate perfor-mance appraisals, and less overall reliance onheuristics and stereotyping for decision making(see Bodenhausen, Gabriel, & Lineberger,2000; Schwarz, 1998). Overskeid (2000) hasargued that the decreased arousal associatedwith sadness facilitates problem solving byallowing for the deployment of more time-consuming analytic strategies. Alternatively,Schwarz (1990) has suggested that sadnesstends to be accompanied by a decrease in peo-ple’s confidence in their first impressions. In anattempt to compensate for this insecurity, anindividual experiencing sadness may engage ina more extensive deliberation during decisionmaking. Based on their research showing thatinduction of a sad emotional state decreases thelikelihood of false-memory bias, Storbeck andClore (2005) similarly concluded that “withsadness comes accuracy” (p. 785).

In addition to the reflective function associ-ated with the experience of sadness, the non-

verbal expression of sadness is thought to servean important interpersonal function. From asocial-functional perspective, expressions ofemotion in mammals are evolutionary adapta-tions to social environments related to the cre-ation and maintenance of social relationshipsand the organization of interindividual interac-tions (Darwin, 1872/1998; Keltner & Kring,1998). Facial displays of emotion evoke andshape the responses of others by inducingspecific emotional responses and reinforcingor discouraging social behaviors (Keltner &Kring, 1998). The facial expression of sadnessis thought to support group social behavior byevoking sympathy and helping responses inothers (Keltner & Kring, 1998; Izard, 1977,1993; Lazarus, 1991; Stearns, 1993).

Research on distress eliciting sympathysuggests that sadness functions in a reciprocalmanner (Batson & Shaw, 1991; Keltner &Kring, 1998). Sad images evoke both sad af-fect (Gross & Levenson, 1995) and increasedamygdala activation in observers (Wang etal., 2005). Moreover, experimental researchreveals that newborn infants are capable ofdistinguishing between an audiotape record-ing of their own cry and another neonate’scry, and that the sound of another newborn’scry provokes distress in the infants, evidencedthrough consistent facial grimaces, turningred-faced, and showing visible signs of agita-tion (Dondi, Simion, & Caltran, 1999). Phys-iological responses accompanying sympathyin adults, including concerned gaze and re-duced heart rate, are predictive of altruisticor helping behaviors. Such reciprocal re-sponses increase the probability that individu-als expressing sadness will receive needed at-tention and/or assistance from others (Keltner& Kring, 1998).

It is important to note, however, that thefunctional benefits of sadness are not entirelyfree of cost. For example, Gray (2001) showedthat experimentally induced sadness enhancedsome aspects of working memory while reduc-ing others. In particular, sadness induction en-hanced spatial memory while reducing verbalmemory. Notably, the opposite pattern was evi-dent in a condition that involved a happinessinduction (Gray, 2001). Similarly, despitethe fact that sadness is generally linked todecreased susceptibility to judgmental bias,Bodenhausen et al. (2000) found that a sadnessinduction resulted in greater vulnerability to ananchoring bias wherein final judgments were

49. Sadness and Grief 799

altered toward a provided starting point, evenif the particular starting point was arbitrary.

There are also social consequences associ-ated with sadness. Ambady and Gray (2002)found that a sadness induction led to reducedaccuracy in participants’ social judgments ofbrief video clips. In one of their studies, for ex-ample, sadness reduced participants’ ability toaccurately gauge teacher effectiveness. In an-other study, sadness reduced participants’ abil-ity to correctly categorize the type of relation-ship enacted in brief video clips of dyadicinteractions.

WHEN SADNESSBECOMES DEPRESSION

Sadness can sometimes deteriorate into morechronic dysphoric mood states, or, in extremecases, depressogenic states. In contrast to thecognitive and social benefits associated withbrief sadness episodes, more prolonged dys-phoric states have been associated with with-drawal and despair, as well as with the elicita-tion of rejection from others (Bonanno &Keltner, 1997; Lazarus, 1991; Smith &Ellsworth, 1985).

One of the mechanisms that probably medi-ates the transition from the brief, episodic ex-perience of sadness to a more elaborate anddysfunctional depressive state is rumination.Considerable research indicates that transientdysphoric mood is most likely to develop into amore prolonged depressive state when peo-ple engage in ruminative responses, such as“repetitively and passively focusing on symp-toms of distress and the possible causes andconsequences of these symptoms” (Nolen-Hoeksema, Wisco & Lyubomirsky, in press).Moreover, the tendency to ruminate appears tobe a relatively stable response style—one linkedto increased vulnerability not only for depres-sion, but also for negative thinking, decreasedproblem solving abilities, disruption in the exe-cution of instrumental behaviors, and the dis-solution of social relationships (see Nolen-Hoeksema, 1991; Nolen-Hoeksema et al., inpress). And rumination has been shown to me-diate other risk factors for depression, suchas negative cognitive styles, self-criticism,neediness, and a history of depression(Spasojevic & Alloy, 2001).

The potentially dysfunctional influence ofdepressive rumination is perhaps most pro-

nounced in situations where psychologicalthreat is acute, as in the case of a serious inter-personal loss. For example, Nolen-Hoeksema,Morrow, and Frederickson (1993) found thatin a longitudinal study of bereaved men, initialrumination levels were uniquely predictive ofdepression 12 months later, even after initialdepressive symptomatology was controlled for.Recent evidence indicates that rumination pre-dicts the onset of depression, and is moderatedby negative cognitive styles in predicting theduration of depression (Nolen-Hoeksema etal., in press).

There also appear to be genetic vulnerabili-ties to depression that emerge when people areexposed to loss or other extremely aversiveevents, suggesting a gene × environment inter-action (Moffitt, Caspi, & Rutter, 2006). Con-vincing prospective evidence has shown, for ex-ample, that a functional polymorphism in thepromoter region of the serotonin transportergene (5-HTT) moderates the impact of stressfullife events on depression (Caspi et al., 2003).Such findings suggest the compelling possibilitythat ruminative responses to sad states may beat least partially informed by a person’s geneticmakeup.

But is the transition from sadness to depres-sion always maladaptive? This pivotal questionbecomes especially important in the context ofloss, where bereavement scholars have tradi-tionally assigned functional significance togrief reactions (e.g., Bowlby, 1980). Nesse(2000) has argued, for example, that the cog-nitive and behavioral manifestations ofdepression—specifically, pessimism and lack ofmotivation—help to inhibit potentially danger-ous actions that might lead to further loss, thusconserving necessary resources for survival. Intheir situation–symptom congruence hypothe-sis, Keller and Nesse (2006) theorize that be-cause of the distinct adaptive functions of vari-ous depressive symptoms, different contexts(including failure and social loss) are likely toevoke discrete depressive symptoms, which willin turn enhance survival capacities in these di-versely challenging situations.

However, other fundamental social and cog-nitive survival skills may be lost at the expenseof these inhibited behaviors. Whereas transientsadness may boost some forms of problemsolving, when sadness becomes depression, theconcomitant lack of motivation and pessimismwill tend to interrupt problem-solving efforts(Overskeid, 2000). Similarly, whereas brief dis-


plays of sadness evoke sympathy and helpingresponses from others, more prolonged dys-phoric expressions tend to extract a serious tollon personal relationships and to threaten over-all social adjustment. Several studies have sug-gested that intense and prolonged expressionsof negative emotions, such as sadness, tend todrive away people who might otherwise offersupport (Coyne, 1976; Gottlieb, 1991; Harber& Pennebaker, 1992; Pennebaker, 1993). In hisinterpersonal theory of depression, for in-stance, Coyne (1976) argued that depression istransmitted between people through excessivereassurance seeking. However, reassuranceseeking in turn tends to lead to social rejection(Joiner, 1999; Joiner, Alfano, & Metalsky,1992).

The type of negative social exchange fueledby excessive dysphoric expressions creates adownward spiral of social rejection. For exam-ple, Strack and Coyne (1983) found that par-ticipants who had engaged in 15 minutes ofconversation with depressed partners endorsedgreater anxiety, depression, and hostile moodfollowing the interaction, compared to thosewho conversed with nondepressed partners.Participants who conversed with a depressedpartner also endorsed less willingness to inter-act with the partner in the future and a greaterwillingness to share negative social perceptionswith the depressed partner. The depressed part-ners, in turn, had accurately anticipated theirpartners’ rejection, and acted reciprocally byrejecting their nondepressed counterparts.

WHEN SADNESS BECOMES GRIEF

The phenomenology of sadness as it moves intomore depressed states is mirrored, to some ex-tent, by the ways in which sadness appears togive way to more complicated grief reactionsduring bereavement. From a comparative andhistorical perspective, it is easy to envision howthe functional relevance of sadness would facil-itate human survival of loss. However, it ismore difficult to fathom that there would bemuch usefulness to prolonged dysphoric states.In this section, we briefly consider what the ex-perience of loss might have been like in our an-cestral past, and how it may still operateamong our nonhuman primate relatives.

One salient feature of virtually our entire an-cestral past is that humans and protohumanswere constantly faced with the death of genetic

relatives or others within the broader socialgroup. It was only recently, for example, withthe advent of modern medicine, that humansbegan to exceed a normal life expectancy of20–35 years (Galor & Moav, 2005). For all buta relatively brief period of the time humanshave lived on earth, daily life has been besetwith disease, poor nutrition, predation, andother threats to survival. Moreover, becausehumans and protohumans were until only re-cently nomadic, any psychological mechanismsthat might have evolved to deal with these con-tinual losses would have by necessity been amechanism commensurate to group life on themove.

In this context, grief-related sadness in re-sponse to the loss would clearly be most adap-tive as an ephemeral reaction that ran its coursein relatively short order. Each of the intra-personal functional characteristics we have dis-cussed above—the turning of attention inward,the temporary decrease in arousal, the promo-tion of resignation and acceptance, the facilita-tion of problem solving, and the updatingof cognitive structures—offer obvious mecha-nisms that would help a person adjust to thesocial and personal changes wrought by thedeath of an important relative or group mem-ber. Likewise, the expressive functions ofsadness—the evocation of sympathy and help-ing responses in others—would also facilitatebereavement in the context of the broader so-cial group.

By the same token, it seems obvious thatmore prolonged sadness reactions, or moreelaborate grief or depressive states, would beincompatible with nomadic life and would puta bereaved nomad in this context at consider-able risk. The withdrawal and despair associ-ated with longer-term sadness and depressivestates would make it difficult to keep up withthe moving tribe or group, and would be likelyto increase susceptibility to disease or preda-tion. Similarly, the association of prolongedsadness and depression with rejection fromothers would mean that the bereaved nomadwould receive less support or perhaps evenelicit treachery from other members of thegroup or tribe.

We can no longer directly observe the press-ing demands of nomadic group living as thesewould have manifested themselves in our ownancestral past, when easy-communication,mass produced goods or foods, and effectivemedical care were nonexistent. However, it is


still possible to observe nomadic group livingin our closest primate relatives. Here JaneGoodall’s famously patient observations ofwild chimpanzees in Gombe National Park,Tanzania, provide compelling examples.

In her field observations, Goodall (1986) re-peatedly described apparent instances of sad-ness, depression, and grief in young chimpan-zees following the death of their mothers.Chimps younger than 3 years of age are still de-pendent on mothers’ milk, and thus cannot sur-vive unless adopted by another female. Yet allof the orphaned chimps Goodall observed inthis age group eventually died following theirmothers’ deaths, regardless of whether or notthey were adopted by another female. Most ofthe older orphans survived, even though someexhibited ostensible developmental deficits(e.g., unusual levels of violence, poor coordina-tion) that may have been occasioned by theloss. For younger orphans, however, prolongedsadness or depression-like behaviors were al-most always lethal. In fact, each of the chimpsthat failed to survive following its mother’sdeath, regardless of age, had been observed aslistless and lethargic. By stark contrast, noneof the surviving orphans exhibited lethargy;moreover, the orphans that were oldest (i.e., 7–9 years) at the time of their mothers’ deathsshowed no noticeable adverse reactions.

It is simply not possible to ascribe emotionalstates to animals with full confidence. How-ever, laboratory studies have shown that inter-ruption of the mother–child bond in non-human primates tends to produce significantbehavioral disturbances, depression-like reac-tions (e.g., lethargy, slouched posture), andwhat appear to be sad facial expressions (e.g.,Reite, Short, Seiler, & Pauley, 1981). Nonhu-man primates also appear to recognize (or atleast to attend to) sad facial expressions, evenin humans. For example, Japanese monkeyshave been found to engage in longer searchtime when presented with sad than with neu-tral human faces (Kanazawa, 1996). Becausemonkeys do not have the same capacity foreyebrow movement as humans, they tend todevote greater attention in differentiating sad-ness to the cheek muscles (Kanazawa, 1998).

Goodall (1986) consistently observed proto-emotional expressions in the orphaned chimpsthat were similar to human sadness expres-sions. We could speculate that these expres-sions would play some role in restructuring so-cial relations following the loss. A key element

of chimpanzee survival with particular rele-vance to bereavement is social bonding withsiblings and other adults. All of the orphanedchimps in Goodall’s sample, even those that didnot survive, had either been adopted by an-other adult or sibling, or traveled with anotherchimp when the group was on the move. Sur-vival would no doubt depend, at least in part,on the success of these bonds (Trivers, 1972).The orphans that did not survive appeared tohave somewhat troubled and ambivalent rela-tions with their newly adopted partners, and itseems that these difficulties might be attributedat least in part to the social friction caused bytheir lethargic expressions and neediness. Forexample, one 4- to 5-year-old orphan, given thename Merlin, was adopted by an older femalesibling. Merlin was described as lethargic, withsunken eyes; as engaging in minimal play; andas socially unresponsive. The sibling allowedMerlin to sleep with her, but rejected him whenhe attempted to ride her dorsally, as youngchimps often do with their mothers. He died 18months later (Goodall, 1986, p. 102).

Together, these data suggest that the experi-ence and expressions of sadness are function-ally useful following an important loss, butalso that more prolonged and extreme sadnesswill be less functionally relevant; as sadnessgives way to depressive expressions and leth-argy, it begins to compromise functioning andto undermine potentially crucial avenues for in-terpersonal bonding and support.

POSITIVE EMOTION ANDOSCILLATION IN GRIEF REACTIONS

As noted earlier, although sadness is the proto-typical emotion of bereavement, it is not theonly emotion evoked by the death of a lovedone. Somewhat counterintuitively, loss experi-ences also commonly involve positive memo-ries and reflections, as well as positive emo-tional expressions (Bonanno & Kaltman,1999; Bonanno, Moskowitz, Papa, &Folkman, 2005; Moskowitz, Folkman, &Acree, 2003; Shuchter & Zisook, 1993; Steinet al., 1997). Facial coding from videotaped in-terviews conducted within months after thedeath of a spouse showed, for example, thatmost bereaved people exhibited at least onegenuine laugh or smile, even as they discussedtheir recent loss (Bonanno & Keltner, 1997).Moreover, consistent with the adaptive value of


positive emotions observed in other contexts(Fredrickson, 1998, 2001; Fredrickson &Joiner, 2002), bereaved people who expressedpositive emotions early in bereavement hadbetter long-term adjustment (Bonanno &Keltner, 1997) and also evoked more favorablereactions from untrained observers whoviewed videotapes of the interviews (Keltner &Bonanno, 1997).

How can grief involve sadness, longing,withdrawal, and disorientation on the onehand, and laughter and positive emotional ex-periences on the other? The answer seems to bethat bereavement, like other stress reactions,has a kind of wave-like periodicity character-ized by oscillation between a focus inward (onthe stressor event and its implications andmeanings) and a focus outward (toward the ex-ternal world and other people). Grief reactionsare not uniform or static. Rather, as many be-reavement investigators have noted, grief seemsto occur in waves. The wave-like nature of griefhas been observed, for example, in studies of19th-century American diaries (e.g., Rosen-blatt, 1983). Robert Kastenbaum (1995), oneof the first social scientists to consider how hu-mans adapt to death and loss, noted that “dis-tress does not end with the first wave of shockand grief. After the realization that a loved oneis dead often comes the realization that life issupposed to go on” (p. 316).

The oscillatory nature of these reactions hasbeen documented in many theories about se-vere or pathological reactions to loss and trau-ma. The key point is that a moderate level ofoscillation is an adaptive feature of normal,short-lived stress reactions. The inability toself-regulate sadness and other emotions dur-ing bereavement, however, tends to be associ-ated with more extreme and unregulated formsof oscillation, with more enduring and extremenegative affective reactions, and consequentlywith more chronic or complicated grief reac-tions.

The oscillating quality of grief reactions isnot unlike similar wave-like or oscillatory reac-tions observed in response to other stressorevents. Litz (quoted in Greer, 2005) has noted,for example, “We tend to think of [posttrau-matic stress disorder] as this discrete disorderthat is a steady state. But research suggests it’smuch more dynamic . . . the symptoms canwax and wane over time” (p. 39). Similarly, ina widely cited book about trauma, Herman(1992) describes the “dialectic of trauma . . . In

the aftermath of an experience of overwhelm-ing danger, the two contradictory responses ofintrusion and constriction establish an oscillat-ing rhythm” (p. 47).

In an attempt to move beyond the traditionalemphasis in bereavement theory on grief work,M. S. Stroebe and Schut (1999) have incorpo-rated the wave-like nature of grief into abroader model of bereavement coping behav-iors: the dual-process model of bereavement.The dual-process model specifies that adaptivecoping with bereavement requires two types ofprocesses or coping orientations: a “loss orien-tation” and a “restoration orientation.” A lossorientation is described as dealing with “pro-cessing of some aspect of the loss experience it-self, most particularly, with respect to the de-ceased person” (M. S. Stroebe & Schut, 1999,p. 212). By contrast, a restoration orientationinvolves processes that deal with the secondarysources of stress associated with the loss—theindirect consequences of a loss that must bedealt with in order to move beyond the loss.

The crucial theoretical advance of the dual-process model is that it emphasizes the alter-nating quality or oscillation of the two orienta-tions. Not only are both loss-oriented andrestoration-oriented processes deemed neces-sary for successful recovery from loss; accord-ing to the model, “optimal adjustment” (M. S.Stroebe & Schut, 1999, p. 216) over time alsorequires a relatively constant switching backand forth, or oscillation, between the pro-cesses. As we discuss below, however, becausethe Stroebe and Schut dual-process model fo-cuses primarily on coping rather than emotionduring bereavement, it may overestimate theduration and extent of oscillation required forhealthy adjustment.

Oscillation in Emotion Regulation

Under normal conditions, we can conceptual-ize self-regulation as the relatively simple andstraightforward task of maintaining an optimalpsychological and emotional equilibrium(Bonanno, 2001; Carver, 1998; Carver &Scheier, 1982, 1990; Westphal & Bonanno,2004). One key issue to consider when wethink of oscillation is how positive and nega-tive states may function in relation to eachother. There has been quite a bit of debate inpsychology about this issue. Some theoristshave argued that positive and negative statesrepresent two ends of a bipolar continuum


(e.g., Barrett & Russell, 1998). One of the pri-mary sources of evidence for this position co-mes from factor-analytic studies of the waypeople describe their affective states; it seemsthat when people say they are feeling good,they don’t say they are feeling bad, and viceversa. Other theorists have argued, however,that positive and negative states are relativelyindependent, and thus represent distinct anduncorrelated dimensions of experience. Evi-dence for the independence of emotional statescomes from studies of biological markers ofemotion, such as brain functions (Cacioppo,Gardner, & Berntson, 1997, 1999) and facialexpressions (Bonanno & Keltner, 1997).

One way to think about this controversyis that under typical conditions, when self-regulation demands are considerably mild, pos-itive and negative states will appear to functionrelatively independently. By contrast, however,when there is a serious disturbance or aversivethreat to the system, we should see extremes ofthese behaviors as individuals tend to experi-ence more bipolar affective reactions. Stressfulsituations tend naturally to evoke negativeemotions, and these emotions should servetheir functional purpose. For example, gettingangry helps people to mobilize their resourcesand communicate their readiness to defendthemselves. However, people also tend to expe-rience positive emotions and even moments ofcalm in the aftermath of extreme adversity. Thepresence of positive emotion in the context ofadversity is supported by evidence suggestingthat positive emotions may play a crucial rolein undoing or regulating the physiologicaleffects of negative emotions (Bonanno &Keltner, 1997; Fredrickson & Levenson, 1998;Fredrickson, 1998, 2001; Tugade &Fredrickson, 2004). Because it is rather diffi-cult, if not impossible, to experience negativeand positive emotions simultaneously, and be-cause the experience of positive emotions mayminimize or undo negative emotions, it is rea-sonable to expect that such experiences willtend to alternate.

This is precisely what is predicted by Zautraand colleagues’ dynamic model of affect(Reich, Zautra, & Davis, 2003; Zautra, Reich,Davis, Potter, & Nicolson, 2000; Zautra,Berkhof, & Nicolson, 2002). They argue thatwhen there is low stress, it is relatively easy forpeople to engage in complex, differentiated,and multidimensional processing of the sur-rounding environment and of their own affec-

tive reactions in response to that environment.During an ordinary social gathering withfriends, for instance, one may have differentperceptions and attributions about one’s ownbehavior or the behavior of the various peoplein the room, and these perceptions and attribu-tions will probably result in a range of affectiveexperiences as the gathering progresses. In sucha casual situation, there is relatively little de-mand on attention, and the mind is ostensiblyfree to roam among, monitor, and cataloguethese affective states. However, in more de-manding and stressful situations—say, for ex-ample, a heated interpersonal conflict or a rushto catch a plane—one’s attention becomesmore clearly concentrated on the most immedi-ate and necessary behaviors and information inthe environment. Thus perceptions and attribu-tions become more narrowly focused, and theexperience of positive or negative states seemsto “collapse into a single bipolar dimensionwith highly inversely coupled affect” (Reich etal., 2003, p. 70). It is important to note thatstress does not necessarily change the functionsof positive and negative emotions, and indeedthe biological and behavioral data suggest thatpositive and negative emotions are functionallyindependent (Bonanno & Keltner 1997;Cacioppo et al., 1999). What does change instressful situations is the subjective experienceof affect; one’s perception is shifted toward ex-periencing affect more as a bipolar, or one-or-the-other, phenomenon.

The inverse relationship between positiveand negative affect in times of stress suggests adirect relationship among the severity of thestress, the strength of the bipolarity, and thestrength of the regulatory oscillation (Bisconti,Bergeman, & Boker, 2004). For example, inlongitudinal studies of people suffering fromchronic pain syndromes, Zautra and Smith(2001) found that across a number of weeks,positive and negative affect tended to show amild correlation—and, not surprisingly, thatpain tended to increase negative affect and de-crease positive affect. More interestingly, how-ever, they also found that when pain was morepronounced, the presence of positive affect waspredictive of a weaker relationship betweenpain and negative affect. In other words, “aspain escalates, positive affect appears to playan increasing role in the regulation of negativeaffect” (Zautra & Smith, 2001, p. 690).

Recently, Bisconti, Bergeman, Boker, andOng (Bisconti et al., 2004; Bisconti, Bergeman,


& Boker, 2006; Ong, Bergeman, & Bisconti,2004) replicated these findings, using daily rat-ings of positive and negative affect across sev-eral different samples, including a recentlybereaved sample. They obtained reports ofemotional well-being and depression from asmall sample of widows each day during thefirst 1–4 months of bereavement. The explicitgoal of this study was to test the idea, central tothis chapter, that “a stressful life event, such asthe death of a spouse, perturbs the emotionalwell-being state of the individual away fromequilibrium, contributing to emotional shiftsthat vacillate between negative and positive af-fect” (Bisconti et al., 2004, p. 164). Thus theypredicted that the widows’ daily well-being rat-ings would conform to a linear oscillator model(Bisconti et al., 2004; Chow et al., 2005) thatlooks something like a “pendulum with fric-tion” (Bisconti et al., 2004, p. 159). The mostobvious features of this type of model are itsgradual decrease in oscillation and the slope ofchange over time. For example, well-being ismore variable and swings to its lowest pointearly in bereavement, and then gradually be-comes more stable and increases on averageacross time. Essentially, these types of dynam-ics indicate that change in reaction to stress isnot static, but rather a product of different self-regulation processes and behaviors over time.Thus the change is best characterized as an os-cillation.

It goes without saying that immediately fol-lowing the death of someone of personal im-portance, people are likely to feel deep andpainful despair. However, even in the early daysof bereavement, people can also temporarilyforget about the loss and laugh, or experiencemoments of normal or even heightened plea-sure. Because of the severity of the disturbanceto normal equilibrium, and the volatility ofthe system, the oscillations tend to be morefrequent and extreme—to have a greateramplitude—and then gradually to lessen or“dampen” across time.

Oscillation and Resilience after Loss

Bisconti et al.’s (2004) data on the first 4months of bereavement conform nicely to thelinear oscillator model. What’s more, althoughthese findings generally seem to support M. S.Stroebe and Schut’s (1999) dual-process model,Bisconti et al. (2004) conclude that the regula-tory oscillations “occur more rapidly” (p. 165)

and earlier in bereavement, and also seem todampen more quickly than the dual-processmodel would imply.

Stroebe and Schut’s dual-process model alsopredicts that because oscillation has an adap-tive regulatory function, the strength of the os-cillation can be taken as a marker of adjust-ment to bereavement. This idea leads to thecompatible assumption that “people who showlittle or no oscillation will adapt less well toloss” (M. S. Stroebe, Schut, & Stroebe, 2005,p. 52). However, both the dynamic model of af-fect and the work of Bisconti and colleagues(e.g., Bisconti et al., 2004, 2006) suggest theopposite hypothesis: that people who showhealthy adaptation or resilience after loss willbe less dysregulated by the loss, and thus willevidence less and not more oscillation.

Support for this idea was further evidencedin a recent study of recently bereaved adultswho participated in a series of brief laboratoryinterviews (Coifman, Bonanno, & Rafaeli,2007). The most resilient bereaved individualsin this study (“resilience” was defined as hav-ing relatively minimal disruption in functioningacross the first 18 months of bereavement; seeBonanno et al., 2005) showed less oscillation inaffect across interview topics. More specifi-cally, when current levels of distress were statis-tically controlled for, resilient bereaved individ-uals had weaker correlations between self-reported negative and positive affect, suggest-ing relatively greater independence in affect,less bipolarity, and therefore less oscillation. Incontrast, for those individuals who had chroni-cally elevated symptom levels similar to com-plicated grief reactions, positive and negativeaffect were more strongly inversely correlated,suggesting less independence between affects,greater bipolarity, and therefore more oscilla-tion.

In another recent study, Bisconti et al. (2006)also suggested a relationship between resilienceafter loss and individual differences in the oscil-latory pattern during bereavement. They con-cluded that bereaved individuals whose out-come trajectory was similar to that observedamong resilient individuals in previous be-reavement studies (e.g., Bonanno et al., 2002)showed “a temporary influx of lability in emo-tional well-being, followed by a high level ofpositive and stable functioning” (Bisconti et al.,2006, p. 596). They further contrast this resil-ient pattern of oscillation with another trajec-tory more clearly suggestive of the chronic


or complicated grief reactions, in which thewidow was “exhibiting a continued amount ofoscillation across the duration of the study”(p. 596).

CONCLUSION

The results of these investigations not only sup-port our understanding of oscillatory processesduring bereavement; they also illuminate thecrucial distinction between the emotion of sad-ness and the more enduring and complex expe-riences captured under the broader construct ofgrief. As we have described earlier, sadness ex-ists within the experience of grief but is not ex-clusive to grief. By the same token, the experi-ence of grief is dominated by, but not limitedto, sadness; the process of grieving encom-passes a variety of other negative emotions(such as anger or guilt), as well as positive emo-tions (such as happiness, amusement, and af-fection). Moreover, the periodic experience ofpositive emotion while grieving appears toserve the valuable function of helping to regu-late sadness and other negative emotional ex-periences over the course of bereavement. Al-though this model is somewhat at odds withmodern conceptualizations of grief as an exclu-sively painful and somber experience, the oscil-latory nature of positive and negative affectduring bereavement makes sense from the per-spective of our nomadic ancestral heritage. Be-cause long-term grief would almost certainlyhave led to isolation, abandonment, and preda-tion, the utility of experiencing episodes of sad-ness punctuated with positive emotional expe-riences would have been favored over moreruminative and depressogenic responses, andwould have evolved as an adaptive means toachieve both necessary respite and resourcesfrom within the broader social group.

REFERENCES

Abraham, K. (1953). A short study of the developmentof the libido, viewed in the light of mental disorders.In D. Bryan & J. A. Strachey (Eds.), Selected papersof Karl Abraham (pp. 418–501). New York: BasicBooks.

Ambady, N., & Gray, H. M. (2002). On being sad andmistaken: Mood effects on the accuracy of thin-slicejudgments. Journal of Personality and Social Psy-chology, 83, 947–961.

Averill, J. R., & Nunley, E. P. (1993). Grief as an emo-tion and as a disease: A social-constructionist per-spective. In M. S. Stroebe, W. Stroebe, & R. O.Hansson (Eds.), Handbook of bereavement: Theory,research, and intervention (pp. 77–90). New York:Cambridge University Press.

Barrett, L. F. (2006a). Solving the emotion paradox:Categorization and the experience of emotion. Per-sonality and Social Psychology Review, 10, 20–46.

Barrett, L. F. (2006b). Are emotions natural kinds? Cur-rent Directions in Psychological Science, 1, 28–58.

Barrett, L. F., & Russell, J. A. (1998). Independence andbipolarity in the structure of current affect. Journalof Personality and Social Psychology, 74(4), 967–984.

Batson, C. D., & Shaw, L. L. (1991). Evidence for altru-ism: Toward a pluralism of prosocial motives. Psy-chological Inquiry, 2, 107–122.

Belitsky, R., & Jacobs, S. (1986). Bereavement, attach-ment theory, and mental disorders. Psychiatric An-nals, 16(5), 276–280.

Bisconti, T. L., Bergeman, C. S., & Boker, S. M. (2004).Emotional well-being in recently bereaved widows: Adynamic systems approach. Journals of Gerontology:Series B. Psychological Sciences and Social Sciences,59B, 158–168.

Bisconti, T. L., Bergeman, C. S., & Boker, S. M. (2006).Social support as a predictor of variability: An exam-ination of recent widows’ adjustment trajectories.Psychology and Aging, 21, 590–599.

Bodenhausen, G. V., Gabriel, S., & Lineberger, M.(2000). Sadness and susceptibility to judgmentalbias: The case of anchoring. Psychological Science,11(4), 320–323.

Bonanno, G. A. (2001). Grief and emotion: A socio-functional perspective. In M. Stroebe, R. O.Hansson, W. Stroebe, & H. Schut (Eds.), Handbookof bereavement research: Consequences, coping, andcare (pp. 493–516). Washington, DC: American Psy-chological Association.

Bonanno, G. A. (2004). Loss, trauma, and human resil-ience: Have we underestimated the human capacityto thrive after extremely aversive events? AmericanPsychologist, 59, 20–28.

Bonanno, G. A., & Kaltman, S. (1999). Toward an inte-grative perspective on bereavement. PsychologicalBulletin, 125(6), 760–776.

Bonanno, G. A., & Kaltman, S. (2001). The varieties ofgrief experience. Clinical Psychology Review, 21(5),705–734.

Bonanno, G. A., & Keltner, D. (1997). Facial expres-sions of emotion and the course of conjugal bereave-ment. Journal of Abnormal Psychology, 106(1), 126–137.

Bonanno, G. A., & Keltner, D. (2004). The coherence ofemotion systems: Comparing “on-line” measures ofappraisal and facial expression, and self-report. Cog-nition and Emotion, 18, 431–444.

Bonanno, G. A., Mihalecz, M. C., & LeJeune, J. T.


(1999). The core emotion themes of conjugal be-reavement. Motivation and Emotion, 23, 175–201.

Bonanno, G. A., Moskowitz, J. T., Papa, A., &Folkman, S. (2005). Resilience to loss in bereavedspouses, bereaved parents, and bereaved gay men.Journal of Personality and Social Psychology, 88,827–843.

Bonanno, G. A., Wortman, C. B., Lehman, D. R.,Tweed, R. G., Haring, M., Sonnega, J., et al. (2002).Resilience to loss and chronic grief: A prospectivestudy from pre-loss to 18 months post-loss. Journalof Personality and Social Psychology, 83, 1150–1164.

Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss:Sadness and depression. New York: Basic Books.

Cacioppo, J. T., Gardner, W. L., & Berntson, G. G.(1997). Beyond bipolar conceptualizations and mea-sures: The case of attitudes and evaluative space. Per-sonality and Social Psychology Review, 1(1), 3–25.

Cacioppo, J. T., Gardner, W. L., & Berntson, G. G.(1999). The affect system has parallel and integrativeprocessing components: Form follows function. Jour-nal of Personality and Social Psychology, 76(5), 839–855.

Carver, C. S. (1998). Resilience and thriving: Issues,models, and linkages. Journal of Social Issues, 54,245–268.

Carver, C. S., & Scheier, M. F. (1982). Control theory: Auseful conceptual framework for personality–social,clinical and health psychology. Psychological Bulle-tin, 92, 111–135.


Caspi, A., Sugden, K., Moffitt, T., Taylor, T., Craig, I.W., Harrington, H., et al. (2003). Influence of lifestress on depression: Moderation by a polymorphismin the 5-HTT gene. Science, 301, 386–389.

Cerney, M. S., & Buskirk, J. R. (1991). Anger: The hid-den part of grief. Bulletin of the Menninger Clinic,55(2), 228–237.

Chow, S.-Y., Ram, N., Boker, S. M., Fujita, F., & Clore,G. (2005). Emotion as a thermostat: Representingemotion regulation using a damped oscillator model.Emotion, 5, 208–225.

Coifman, K. G., Bonanno, G. A., & Rafaeli, E. (2007).Affective dynamics, bereavement, and resilience toloss. Journal of Happiness Studies, 8, 371–392.

Coyne, J. C. (1976). Depression and the response ofothers. Journal of Abnormal Psychology, 85, 186–193.


Dondi, M., Simion, F., & Caltran, G. (1999). Can new-borns discriminate between their own cry and the cryof another newborn infant? Developmental Psychol-ogy, 35(2), 418–426.

Ekman, P. (1984). Expression and the nature of emo-

tion. In K. Scherer & P. Ekman (Eds.), Approaches toemotion (pp. 319–344). Hillsdale, NJ: Erlbaum.

Ekman, P. (1992). Are there basic emotions? Psycholog-ical Review, 99(3), 550–553.

Ekman, P., & Davidson, R. J. (Eds.). (1994). The natureof emotion: Fundamental questions. New York: Ox-ford University Press.

Folkman, S., & Lazarus, R. S. (1988). Coping as a medi-ator of emotion. Journal of Personality and SocialPsychology, 54(3), 466–475.

Folkman, S., & Lazarus, R. S. (1990). Coping and emo-tion. From N. L. Stein, B. Leventhal, & T. Trabasso(Eds.), Psychological and biological approaches toemotion (pp. 313–332). Hillsdale, NJ: Erlbaum.

Fredrickson, B. L. (1998). What good are positive emo-tions? Review of General Psychology, 2, 300–319.


Fredrickson, B. L., & Joiner, T. (2002). Positive emo-tions trigger upward spirals toward emotional well-being . Psychological Science, 13, 172–175.

Fredrickson, B. L., & Levenson, R. W. (1998). Positiveemotions speed recovery from the cardiovascularsequelae of negative emotions. Cognition and Emo-tion, 12(2), 191–220.

Frijda, N. H. (1993). Moods, emotion episodes, andemotions. In M. Lewis & J. M. Haviland (Eds.),Handbook of emotions (pp. 381–403). New York:Guilford Press.

Galor, O., & Moav, O. (2005). Natural selection andthe evolution of life expectancy (Minerva Center forEconomic Growth, Paper No. 02-05). Retrievedfrom ssrn.com/abstract=563741

Goodall, J. (1986). The chimpanzees of Gombe. Cam-bridge, MA: Belknap Press of Harvard UniversityPress.

Gottlieb, B. H. (1991). Advances in research on socialsupport and issues applying to adolescents at risk. InG. Albrecht, H.-W. Otto, S. Karstedt-Henke, K.Bollert, & F. Deutsche (Eds.), Social prevention andthe social sciences: Theoretical controversies, re-search problems, and evaluation strategies (pp. 199–216). Berlin: de Gruyter.

Gray, J. R. (2001). Emotional modulation of cognitivecontrol: Approach–withdrawal states double-disso-ciate spatial from verbal two-back task performance.Journal of Experimental Psychology: General,130(3), 436–452.

Greer, M. (2005). A new kind of war. American Psycho-logical Association Monitor on Psychology, 36, 40.

Gross, J. J., & John, O. P. (2003). Individual differencesin two emotion regulation processes: Implications foraffect, relationships, and well-being. Journal of Per-sonality and Social Psychology, 85(2), 348–362.

Gross, J. J., & Levenson, R. W. (1995). Emotion elicita-tion using films. Cognition and Emotion, 9, 87–108.

Hansson, R. O., Carpenter, B. N., & Fairchild, S. K.


(1993). Measurement issues in bereavement. In M. S.Stroebe, W. Stroebe, & R. O. Hansson (Eds.), Hand-book of bereavement: Theory, research, and interven-tion (pp. 62–74). New York: Cambridge UniversityPress.

Harber, K. D., & Pennebaker, J. W. (1992). Overcomingtraumatic memories. In S.-A. Christianson (Ed.), Thehandbook of emotion and memory: Research andtheory (pp. 359–387). Hillsdale, NJ: Erlbaum.

Herman, J. L. (1992). Trauma and recovery. New York:Basic Books.

Horowitz, M. J., Siegel, B., Holen, A., Bonanno, G. A.,Milbrath, C., & Stinson, C. (1997). Diagnostic crite-ria for complicated grief disorder. American Journalof Psychiatry, 154(7), 904–910.

Izard, C. (1977). Human emotions. New York: PlenumPress.

Izard, C. E. (1993). Four systems for emotion activa-tion: Cognitive and noncognitive processes. Psycho-logical Review, 100(1), 68–90.

Izard, C. E. (1994). Innate and universal facial expres-sions: Evidence from developmental and cross-cultural research. Psychological Bulletin, 115(2),288–299.

Jenkins, J. M., & Oatley, K. (1996). Emotional episodesand emotionality through the life span. In C. Magai& S. H. McFadden (Eds.), Handbook of emotion,adult development, and aging (pp. 421–441). SanDiego, CA: Academic Press.

Joiner, T. E. (1999). A test of interpersonal theory of de-pression in youth psychiatric inpatients. Journal ofAbnormal Child Psychology, 27, 77–85.

Joiner, T. E., Alfano, M. S., & Metalsky, G. I. (1992).When depression breeds contempt: Reassuranceseeking, self-esteem, and rejection of depressed col-lege students by their roommates. Journal of Abnor-mal Psychology, 101(1), 165–173.

Kanazawa, S. (1996). Recognition of facial expressionsin a Japanese monkey (Macaca fuscata) and humans(Homo sapiens). Primates, 37, 25–38.

Kanazawa, S. (1998). What facial part is important forJapanese monkeys (Macaca fuscata) in recognition ofsmiling and sad faces of humans (Homo sapiens)?Journal of Comparative Psychology, 112, 363–370.

Kastenbaum, R. J. (1995). Death, society, and humanexperience (5th ed.). Boston: Allyn & Bacon.

Kavanagh, D. J. (1990). Towards a cognitive-behavioural intervention for adult grief reactions.British Journal of Psychiatry, 157, 373–383.

Keller, M. C., & Nesse, R. M. (2006). The evolutionarysignificance of depressive symptoms: Different ad-verse situations lead to different depressive symptompatterns. Journal of Personality and Social Psychol-ogy, 91(2), 316–330.

Keltner, D., & Bonanno, G. A. (1997). A study of laugh-ter and dissociation: Distinct correlates of laughterand smiling during bereavement. Journal of Person-ality and Social Psychology, 73, 687–702.

Keltner, D., & Kring, A. (1998). Emotion, social func-

tion, and psychopathology. Review of General Psy-chology, 2(3), 320–342.

Lazare, A. (1989). Bereavement and unresolved grief. InA. Lazare (Ed.), Outpatient psychiatry: Diagnosisand treatment (2nd ed., pp. 381–397). Baltimore:Williams & Wilkins.


LeDoux, J. E. (1989). Central pathways of emotionalplasticity. In H. Weiner, I. Florin, R. Murison, & D.Hellhammer (Eds.), Frontiers of stress research(pp. 122–136). Lewiston, NY: Huber.

LeDoux, J. E. (1996). The emotional brain: The myste-rious underpinnings of emotional life. New York: Si-mon & Schuster.

Lehman, D. R., Wortman, C. B., & Williams, A. F.(1987). Long-term effects of losing a spouse or childin a motor vehicle crash. Journal of Personality andSocial Psychology, 52(1), 218–231.

Litz, B. T. (2005). Has resilience to severe trauma beenunderestimated? American Psychologist, 60(3), 262.

Lundin, T. (1984). Long-term outcome of bereavement.British Journal of Psychiatry, 145, 424–428.

Mauss, I. B., Levenson, R. W., McCarter, L., Wilhelm, F.H., & Gross, J. J. (2005). The tie that binds?: Coher-ence among emotion experience, behavior, and physi-ology. Emotion, 5(2), 175–190.

Moffitt, T. E., Caspi, A., & Rutter, M. (2006). Measuresof gene–environment interactions in psychopatholo-gy: Concepts, research strategies, and implicationsfor research, intervention, and public understandingof genetics. Perspectives on Psychological Science, 1,5–27.

Moskowitz, J., Folkman, S., & Acree, M. (2003). Dopositive psychological states shed light on recoveryfrom bereavement?: Findings from a 3-year longitu-dinal study. Death Studies, 27(6), 471–500.

Neimeyer, R. A. (2006). Widowhood, grief, and thequest for meaning: A narrative perspective on resil-ience. In D. Carr, R. M. Nesse, & C. B. Wortman(Eds.), Late life widowhood: New directions in re-search, theory, and practice (pp. 227–246). NewYork: Springer.

Neimeyer, R. A., Prigerson, H. G., & Davies, B. (2002).Mourning and meaning. American Behavioral Scien-tist, 46(2), 235–251.

Nesse, R. (2001). Is depression an adaptation? Archivesof General Psychiatry, 57, 14–20.

Nolen-Hoeksema, S. (1991). Responses to depressionand their effects on the duration of depressive epi-sodes. Journal of Abnormal Psychology, 100(4),569–582.

Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L.(1993). Response styles and the duration of episodesof depressed mood. Journal of Abnormal Psychology,102(1), 20–28.

Nolen-Hoeksema, S., Wisco, B., & Lyubomirsky, S. (inpress). Rethinking rumination. Perspectives in Psy-chological Science.


Oatley, K., & Jenkins, J. M. (1996). Understandingemotions. Cambridge, MA: Blackwell.

Oatley, K., & Johnson-Laird, P. N. (1996). The commu-nicative theory of emotions: Empirical tests, mentalmodels, and implications for social interaction. In L.L. Martin & A. Tesser (Eds.), Goals and affect(pp. 363–393). Hillsdale, NJ: Erlbaum.

Ong, A. D., Bergeman, C. S., & Bisconti, T. L. (2004).The role of daily positive emotions during conjugalbereavement. Journals of Gerontology: Series B. Psy-chological Sciences and Social Sciences, 59B(4),P168–P176.

Osterweis, M., Solomon, F., & Green, F. (Eds.). (1984).Bereavement: Reactions, consequences, and care.Washington, DC: National Academy Press.

Overskeid, G. (2000). The slave of the passions: Experi-encing problems and selecting solutions. Review ofGeneral Psychology, 4(3), 284–309.

Panksepp, J. (2005). Why does separation distresshurt?: Comment on MacDonald and Leary (2005).Psychological Bulletin, 131, 224–230.

Pennebaker, J. W. (1993). Social mechanisms of con-straint. In D. M. Wegner & J. W. Pennebaker (Eds.),Handbook of mental control (pp. 200–219).Englewood Cliffs, NJ: Prentice-Hall.

Phelps, E. A. (2006). Emotion and cognition: Insightsfrom the study of the human amygdala. Annual Re-view of Psychology, 57, 27–53.

Raphael, B. (1983). The anatomy of bereavement. NewYork: Basic Books.

Reich, J. W., Zautra, A. J., & Davis, M. (2003). Dimen-sions of affect relationships: Models and their inte-grative implications. Review of General Psychology,7(1) 66–83.

Reite, M., Short, R., Seiler, C., & Pauley, J. D. (1981).Attachment, loss, and depression. Journal of ChildPsychology and Psychiatry, 22, 141–169.

Roseman, I. J., Antoniou, A. A., & Jose, P. E. (1996).Appraisal determinants of emotions: Constructing amore accurate and comprehensive theory. Cognitionand Emotion, 10(3), 241–277.

Rosenblatt, P. C. (1993). Grief: The social context ofprivate feelings. In. M. S. Stroebe, W. Stroebe, & R.O. Hansson (Eds.), Handbook of bereavement: The-ory, research, and intervention (pp. 102–111). NewYork: Cambridge University Press.

Russell, J. A. (1994). Is there universal recognition ofemotion from facial expression?: A review of cross-cultural studies. Psychological Bulletin, 115, 102–141.

Schwartzberg, S. S., & Janoff-Bulman, R. (1991). Griefand the search for meaning: Exploring the assump-tive worlds of bereaved college students. Journal ofSocial and Clinical Psychology, 10(3), 270–288.

Schwarz, N. (1990). Feelings as information: Informa-tional and motivational functions of affective states.In E. T. Higgins & R. M. Sorrentino (Eds.), Hand-book of motivation and cognition: Foundations of

social behavior (Vol. 2, pp. 527–561). New York:Guilford Press.

Schwarz, N. (1998). Warmer and more social: Recentdevelopments in cognitive social psychology. AnnualReview of Sociology, 24, 239–264.

Shuchter, S. R., & Zisook, S. (1993). The course of nor-mal grief. In M. S. Stroebe, W. Stroebe, & R. O.Hansson (Eds.), Handbook of bereavement: Theory,research, and intervention (pp. 23–43). New York:Cambridge University Press.

Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cog-nitive appraisal in emotion. Journal of Personalityand Social Psychology, 48(4), 813–838.

Spasojevic, J., & Alloy, L. B. (2001). Rumination as acommon mechanism relating depressive risk factorsto depression. Emotion, 1(1), 25–37.

Stearns, C. Z. (1993). Sadness. In M. Lewis & J. M.Haviland (Eds.), Handbook of emotions (pp. 547–561). New York: Guilford Press.

Stearns, P. N., & Knapp, M. (1996). Historical perspec-tives on grief. In R. Harr & W. G. Parrott (Eds.), Theemotions: Social, cultural, and biological dimensions(pp. 132–150). Thousand Oaks, CA: Sage.

Stein, N., Folkman, S., Trabasso, T., & Richards, T.(1997). Appraisal and goal processes as predictors ofpsychological well-being in bereaved caregivers.Journal of Personality and Social Psychology, 72(4),872–884.

Strack, S., & Coyne, J. C. (1983). Social confirmation ofdysphoria: Shared and private reactions to depres-sion. Journal of Personality and Social Psychology,44, 798–806.

Storbeck, J., & Clore, G. (2005). With sadness comesaccuracy; With happiness, false memory: Mood andthe false memory effect. Psychological Science,16(10), 785–791.

Stroebe, M., Schut, H., & Stroebe, W. (2005). Attach-ment in coping with bereavement: A theoretical inte-gration. Review of General Psychology, 9, 48–66.

Stroebe, M. S., Hansson, R. O., Stroebe, W., & Schut,H. (2001). Concepts and issues in contemporary re-search on bereavement. In M. S. Stroebe, R. O.Hansson, W. Stroebe, & H. Schut (Eds.), Handbookof bereavement research: Consequences, coping, andcare (pp. 3–22). Washington, DC: American Psycho-logical Association.

Stroebe, M. S., & Schut, H. (1999). The dual processmodel of coping with bereavement: Rationale anddescription. Death Studies, 23(3), 197–224.

Stroebe, W., Schut, H., & Stroebe, M. S. (2005). Griefwork, disclosure and counseling: Do they help the be-reaved? Clinical Psychology Review, 25(4), 395–414.

Stroebe, W., & Stroebe, M. S. (1987). Bereavement andhealth: The psychological and physical consequencesof partner loss. New York: Cambridge UniversityPress.

Trivers, R. L. (1972). Parental investment and sexual se-lection. In B. Campbell (Ed.), Sexual selection and


the descent of man (pp. 136–179). Chicago: Aldine-Atherton.

Tugade, M. M., & Fredrickson, B. L. (2004). Resilientindividuals use positive emotions to bounce backfrom negative emotional experiences. Journal of Per-sonality and Social Psychology, 86, 320–333.

Wang, L., McCarthy, G., Song, A. W., & LaBar, K. S.(2005). Amygdala activation to sad pictures duringhigh-field (4 tesla) functional magnetic resonance im-aging. Emotion, 5, 12–22.

Welling, H. (2003). An evolutionary function of the de-pressive reaction: The cognitive map hypothesis.New Ideas in Psychology, 21(2), 147–156.

Westphal, M., & Bonanno, G. A. (2004). Emotion self-regulation. In M. Beauregard (Ed.), Consciousness,

emotional self-regulation and the brain (pp. 1–33).Amsterdam: Benjamins.

Zautra, A. J., Berkhof, J., & Nicolson, N. A. (2002).Changes in affect interrelations as a function ofstressful events. Cognition and Emotion, 16, 309–318.

Zautra, A. J., Reich, J. W., Davis, M. C., Potter, P., &Nicolson, N. A. (2000). The role of stressful events inthe relationship between positive and negative af-fects: Evidence from field and experimental studies.Journal of Personality, 68, 927–951.

Zautra, A. J., & Smith, B. W. (2001). Depression and re-activity to stress in older women with rheumatoid ar-thritis and osteoarthritis. Psychosomatic Medicine,63, 687–696.


Author Index

Aaronson, N. K., 679Aarts, H., 239Abbott, D. H., 669Abe, J. A. A., 221, 292, 293Abele, A., 594, 784Abelson, J. L., 503Abelson, R. P., 542, 593Abraham, K., 798Abramowitz, J. S., 768Abramson, L. Y., 752Abrisqueta-Gomez, J., 603Achee, J. W., 562Ackerl, K., 242Ackerman, B. P., 291, 292, 377Ackermann, H., 283, 285Acklin, M. W., 522, 698Acree, M., 787, 802Adams, C. M., 147Adams, J. B., 678Adams, K., 358Adams, R. M., 205Adams, S., 524Adkins-Regan, E., 238, 241Adler, D., 144, 145Adolphs, R., 169, 170, 184, 224, 257, 281,

602, 603, 610, 611Aftanas, L. I., 521, 522Aggleton, J. P., 160, 239Agras, W. S., 190, 678Aguirre, G. K., 253Aharonov, G., 164, 252Ahern, G. L., 187, 521Ahmetzanov, M. V., 609Ainslie, G., 146Ainsworth, M. D. S., 634Ajzen, I., 645, 654Akert, R., 228Akil, H., 57Aksan, N., 733al’Absi, M., 670Albert, D., 505Albery, I. P., 682Albro, E. R., 575, 576, 582Albus, J. S., 258Alcaine, O., 89Alcaro, A., 52Alea, N., 383Aleman, A., 237, 603

Alessandri, S. A., 747, 753Alessandri, S. M., 226, 300, 578, 730, 731,

746Alexander, A. L., 252Alfano, C. M., 648Alfano, M. S., 801Alhakami, A. S., 647, 655Alkon, A., 402Allaire, J. C., 788Allen, G. D., 476Allen, J. J. B., 188, 189, 284Allen, J. R., 228Allen, L. B., 694Allen, M., 401Allen, R., 386Alleva, E., 666Allison, K., 358Allman, J. M., 280, 282, 283Alloy, L. B., 503, 800Almeida, D. M., 344, 386, 678Alonso-Arbiol, I., 517Alper, C. M., 190, 677Alpern, L., 733Alpers, G. W., 187Alpert, M., 199, 201, 202, 385, 386Als, H., 309Altemus, M., 789Alvarado, N., 215, 218Alvarez-Buylla, A., 237Alvarez-Conrad, J., 91Alçalar, N.Amabile, T. M., 549Amaral, D. G., 164, 166, 169, 170, 172,

250, 252, 280, 603, 720Ambady, N., 223, 224, 730, 800Amorapanth, P., 165, 252An, X., 250, 260, 261, 263Anagnostaras, S. G., 164Ancoli, S., 188, 217Andayani, S., 400Anders, S., 603Anders, T. F., 240Andersen, M. R., 648, 649, 679Anderson, A., 603, 611Anderson, A. K., 170, 259, 603, 719, 784Anderson, C., 458, 683Anderson, J. A., 505Anderson, J. R., 593, 633

Anderson, S. W., 253Andrade, E. B., 148, 149Andrew, D., 276, 277Andrew, R. J., 188Andrews, D. W., 493Andrews, G., 323, 349, 350, 351, 359Anglada Figueroa, D., 252Angrilli, A., 253Angus, L., 89, 93Angyal, A., 757, 758, 759, 760, 762, 765,

770, 771Annas, P., 719Anson, C., 477Antoniou, A. A., 798Antos, S. J., 602Antoun, N., 224, 768Aoki, Y., 398Apfel, R. J., 536Appadurai, A., 762, 764, 766Apple, W., 199Arak, A., 243Archer, D., 227, 228Archer, J., 127, 459Arcus, D., 490Ardon, M. A., 711Arend, R. A., 786Argyle, M., 310Ariely, D., 142, 148, 149Aristotle, 3, 4, 5, 11, 12, 14, 15, 104, 422,

534Arluke, A., 35Armony, J. L., 162, 163, 166, 170, 603, 604,

719, 720, 721Armor, D. A., 670Arn, I., 285Arnell, K. M., 170Arnold, M. B., 68, 70, 72, 73, 81, 82, 577,

628Arnow, B., 678Aro, A. R., 650Aronoff, J., 636Aronson, J. A., 147, 148, 763Aronson, K. R., 272Arrindell, W. A., 711Arsenio, W. F., 325, 342, 734, 735Ashbasker, M., 397Ashby, F. G., 149, 550, 552, 558, 568, 789Asher, E. R., 785

811

Asher, S., 337, 343Ashmore, M., 767Ashton-James, C., 540Aspinwall, L. G., 549, 552, 557, 784Atkinson, A. P., 716Ato-García, M., 522Atzmueller, M., 242Au, W. T., 456, 459Auerbach, J. G., 487Auerbach, S., 337, 734Aunger, R., 758Austen, J., 103Austin, J. L., 9Auszra, L., 92Averill, J. R., 12, 18, 798Avero, P., 402Avitsur, R., 666, 669Avlund, K., 678Axel, R., 236Axelrod, B. N., 236Aycock, J., 386Aziz, N., 668, 669

Baars, B. J., 50, 171Baas, D., 237, 603Babinsky, R., 171, 603Bachevalier, J., 165Bachorowski, J. A., 103, 197, 198, 199,

200, 203, 204, 205, 206, 223, 351, 693Baddeley, A., 171Badinter, E., 21Badre, D., 257Bagby, M. R., 698Bagby, R. M., 521, 522, 523, 698Bagozzi, R. P., 515, 516Bahrick, L. E., 367, 368, 369, 371, 373Baier, B., 284Bailey, M. T., 666Baillargeon, R., 633Bainum, C. K., 337Baird, B. M., 478Bajgar, J., 520, 521Baldaro, B., 222Baldwin, D. A., 351Ballard, M., 341Balogh, R. D., 240Baltes, M. M., 379, 380Baltes, P. B., 379, 380, 382, 445Banaji, M. R., 202, 257Bandes, S., 451, 452Bandler, R., 250, 669Bandura, A., 443, 652, 654Bandy, D., 259, 279Banerjee, M., 321, 353Banks, W. A., 667Banse, R., 187, 197, 198, 199, 200, 201,

202, 217Bar, M., 636Barak, Y., 670Barbalet, J. M., 37, 44Bard, K. A., 217Bard, P., 159Barden, R. C., 322, 786Barefoot, J. C., 665, 731Barger, S. D., 329Bargh, J. A., 71, 83, 205, 207, 458, 507,

608, 637, 779Barker, W. A., 224, 768Barkow, J., 116Barksdale, C. M., 55Barlow, D. H., 459, 508, 694, 696, 698,

699, 700, 701, 709Barna, J., 352Barndollar, K., 507Barnes, D. S., 764, 765Barnett, P. A., 697Barnhart, K. T., 241Bar-On, R., 538, 539Baron, R. A., 788Baron-Cohen, S., 119, 323Barone, M. J., 549Barrett, H. C., 115Barrett, K. C., 296, 335, 341, 370, 456, 681,

745, 748

Barrett, L. F., 52, 62, 76, 181, 183, 201,249, 253, 259, 262, 263, 272, 274, 282,285, 292, 398, 401, 431, 437, 458, 459,498, 499, 509, 513, 514, 515, 516, 517,518, 519, 520, 522, 523, 525, 536, 588,589, 590, 597, 629, 630, 631, 636, 637,639, 677, 693, 697, 698, 700, 780, 787,790, 797, 804

Barron, J. L., 164Barry, W. J., 200Barsade, S. G., 462, 549Barsalou, L. W., 82, 518, 591, 594, 595,

596, 597, 598Barsky, A. J., 522Bartel, C. A., 462Barter, J., 472Bartlett, J. A., 664Bartlett, M. Y., 152, 436, 463, 790Bartsch, K., 324Bartussek, D., 188Basso, M. R., 784Bates, J. E., 342, 485, 486, 487, 489, 490,

492, 493, 733Batson, C. D., 442, 447, 449, 799Batson, J. G., 449Bauer, A. M., 584Baum, A., 645, 677Bauman, C. W., 435Baumann, N., 553, 554, 555Baumeister, R. F., 134, 456, 459, 629, 636,

637, 639, 667, 668, 781, 788Baumgartner, T., 285Bavelas, J. B., 458Baxter, L. C., 278Baxter, M. G., 252Bean, N. J., 54, 56Bear, G., 736Bear, M. F., 252Beardsall, L., 321, 327, 337Beattie, J., 141Beatty, W. W., 56Beauchamp, G., 765Beauchamp, K. L., 736Beaulieu, D., 342, 344Beaupré, M. G., 459Beauregard, M., 169Bechara, A., 53, 142, 143, 151, 167, 189,

281, 284, 610Becht, M. C., 459Beck, A. T., 508, 743, 747Beck, D. M., 171Beck, J. S., 789Becker, E. S., 716Becker, E., 761Bedell, B. T., 536, 541Bedford, E., 9, 11Beebe, B., 364, 441Beeghly, M., 592Beer, J. S., 103, 253, 669Beers, M., 458Begum, M., 678Behar, E., 89Behbehani, M. M., 250Behniea, H., 170Bekerian, D. A., 593Belitsky, R., 798Bellgowan, P. S., 162Belouad, F., 734Belova, M. A., 252Belton, L., 678Ben-Eliyahu, S., 666Bendersky, M., 300, 306, 731Benedetti, F., 258Benedict, R., 668Benetti-McQuoid, J., 398Benhabib, J., 150Bennett, D. S., 300, 306, 731Bennett, G., 768Bennett, S. L., 364Bennett, S. M., 147Bensafi, M., 237Bentley, M., 68, 71Benton, D., 241Benvenuto, M., 516, 698, 780

Berenbaum, H., 222, 387, 521, 695, 696,697, 698

Bergeman, C. S., 384, 516, 525, 804, 805Berger, H. J. C., 789Berger, J., 35Bergmann, G., 200Berk, L. S., 670Berkhof, J., 804Berkow, R., 130Berkowitz, L., 432, 631, 636Berle, D., 768Berlin, H. A., 253Berlin, L. J., 518, 734Berliner, D., 241Berman, L., 683Bermpohl, F., 50Bernat, E. M., 694Bernhard, R., 225Bernheim, B. D., 150Bernoulli, D., 140Berns, G. S., 144, 655Berntson, G. G., 180, 181, 183, 185, 186,

189, 190, 779, 781, 804Bernzweig, J., 731Berridge, K. C., 51, 52, 61, 71, 72, 83, 181,

184, 224, 253, 260, 637, 778, 789Berry, D. C., 611Berry, E. M., 386Berthier, M., 284Berthoz, S., 521Bertolucci, P. H., 603Bestmann, S. J., 240Beswick, A. D., 682Bettencourt, B. A., 401Bettens, F., 240Beutler, L. E., 91, 92Beyers, J. M., 492Bhalodkar, N. C., 680Bibi, U., 605Bickerstaffe, S., 769Biederman, I., 517Biehl, M., 401Bigelow, A. E., 364Bikle, P. C., 611Bilir, P., 679Billig, M. G., 636Billings, R. S., 551Bimson, W. E., 281Bindra, D., 77Biran, A., 758Birditt, K., 382, 386Birdsall, T. G., 124Birkeli, K., 477Birnbaum, D., 654Bischof-Köhler, D., 316, 445Bisconti, T. L., 804, 805Bishop, P., 54, 55, 56Bisin, A., 150Biswas-Diener, R., 476, 515Bixler, M. A., 160Bjerring, P., 670Bjork, R. A., 257Björklund, F., 767Black, A., 458Black, S. A., 677, 777Blainey, K., 541Blair, K., 337, 338, 734Blair, R. J. R., 223, 359Blairy, S., 187Blakemore, C., 171Blakemore, S. J., 257Blanchard, C. D., 160Blanchard, R. J., 160Blanchard-Fields, F., 386Blascovich, J., 119Blass, E. M., 55, 444Blehar, M. C., 634Bless, H., 549, 553, 555, 558, 565, 566,

606, 785Blevings, G., 722Bliss-Moreau, E., 183, 272, 515, 520,

630Block, J., 788Bloom, E. T., 667

812 Author Index

Bloom, F. E., 307Bloom, K., 372Bloom, P., 758, 766, 770Blount, S., 148Bluck, S., 383, 607Blum, L. A., 445Blumberg, S. J., 502Blume, J. H., 452Blumenthal, J. A., 665Blumstein, S. E., 198Bober, S. L., 679Boca, S., 401, 458Boddaert, N., 284Bodenhausen, G. V., 549, 550, 552, 558,

565, 799Bodkin, J. A., 54Bohannon, J. N., 605, 606Bohart, A. C., 92Bohlin, G., 734Bohner, G., 549Boisoneau, D., 664Boker, S. M., 798, 804, 805Bollen, K. A., 524, 630Bonanno, G. A., 95, 206, 215, 217, 218,

219, 221, 480, 498, 680, 684, 782, 787,797, 798, 799, 800, 802, 803, 804, 805

Bond, N., 683, 768, 769Bonett, D. G., 241Bongar, B., 91Boomsma, D. I., 475Booth, D. A., 241Booth, R. J., 670Booth-Kewley, S., 680Boring, E. G., 186Borke, H., 355Borkovec, T. D., 89Bormann-Kischkel, C., 358Bornstein, M. H., 489Bornstein, R. F., 611Borod, J. C., 222, 260, 385, 386Borton, R. W., 367Bosch, J. D., 736Bosquet, H., 307Bouchard, B., 107Boulifard, D., 297Boulton, M. J., 782, 786Bourgouin, P., 169, 256Bouril, B., 734Bourke, J., 709Bourne, E., 417Boush, D. M., 683Bouton, M. E., 78, 164Bowen, D. J., 648, 649, 679Bower, G. H., 593, 594Bowersox, A. L., 736Bowlby, J., 54, 308, 497, 798, 800Bowles, R., 717Boyce, W. T., 342, 402, 575, 578, 579, 582,

583, 584Boyer, P., 128Boyes-Braem, P., 517Boyle, P. A., 524Boyle, S. H., 665Boyle, S., 203Brackett, M. A., 537, 539, 540Bradburn, N. M., 472Bradley, B. P., 717Bradley, M. M., 73, 77, 78, 81, 181, 184,

187, 203, 602, 605, 691, 693, 699, 719,722

Bradmetz, J., 324, 359Bradshaw, D., 298Bradshaw, J. L., 666Brady, K., 699Braendgaard, H., 263Braesicke, K., 252Brand, G., 200, 237, 238Brand, J. E., 670Branigan, C., 95, 555, 780, 783, 784, 788Branscombe, N. R., 431, 435, 463Brant, L. J., 382Bratslavsky, E., 781Braun, J. J., 241Braungart, J. M., 327, 337, 735

Braungart-Rieker, J. M., 488, 489Braver, T. S., 278Braverman, D. L., 536Brazelton, T. B., 309Breder, C. D., 253Brehm, J. W., 681Breitenstein, C., 202Breiter, H. C., 223Brejnak, C., 522Bremmer, J., 24Bremner, J. D., 165Brentano, F., 9Bresin, R., 199Bretherton, I., 321, 337, 353, 592Brett, M., 256Brewer, J., 170Brewer, M. B., 437Bricker, W., 737Brickman, P., 144, 151, 480Bridges, K. M. B., 293, 294, 309, 315Brierley, B., 603Brigham, J., 678Brissette, I., 458Briton, N. J., 396Britton, J. C., 184, 237Broaders, S., 575Broadstock, M., 679Brocas, I., 150Brodal, A., 159, 160Brodsgaard, I., 679Brody, L. R., 357, 395, 396, 397, 398, 400,

401, 402, 403, 404, 405Broks, P., 224Broman, E. L., 698Bronfenbrenner, U., 332, 343Brooks, J. C., 278, 280, 281, 315Brooks-Gunn, J., 313, 445Brooksbank, B. W. L., 241Brosch, T., 716Brosgole, L., 386Brosschot, J. F., 694Brothers, L., 441Broucek, F. J., 744Brown, A., 219Brown, C. E., 205Brown, D., 131Brown, D. L., 680Brown, E. J., 680Brown, J. R., 328Brown, J., 327, 328, 337, 733, 734Brown, J. S., 629Brown, J. W., 278Brown, K. W., 678Brown, M., 736Brown, R., 605, 606, 607, 770Brown, R. P., 285Brown, R. W., 517Brown, S. L., 187, 188, 786Brown, T. A., 694, 696, 724, 725Brownbridge, G., 472, 680Browne, M. J., 151Bruce, M. L., 679Bruder, S., 525Brummett, B. H., 449, 665Bruner, J., 109Bruno, J. P., 252Brunstein, J. C., 478Brushi, C., 341Bryan, J., 786Bryan, T., 786Bryant, B. K., 735Buchanan, T. W., 602, 603, 610, 611, 670Buchel, C., 167Buck, L., 236Buck, R. W., 219, 228Buck, R., 73, 76, 77, 79, 402, 536, 631, 636Buckland, G., 768Buckley, T. R., 679Buckner, R. L., 611Bucy, P. C., 160Budd, M., 102, 106Budson, A. E., 605, 608Bueno, O. F., 603Buffalo, E. A., 607

Bugental, D. B., 342, 344Bugental, D. E., 399Bugenthal, D. B., 730Buhler, C., 371Buick, D., 680Bulka, D., 380Bullock, M., 349, 354, 355, 356, 358, 593Bundy, R. P., 636Bunge, S. A., 169, 258, 504, 507Bunt, L., 106, 107Buntaine, R. L., 397Burg, M. M., 677Burgdorf, J., 56, 57, 58Burgess, K. B., 492, 734Burgess, L., 760Burghardt, G. M., 56Buring, J. E., 665Burke, A., 583Burke, P. J., 21, 38, 39Burke, R. J., 399Burke, S., 57Burnett, C. B., 649Burney, R., 93Burns, A. B., 787Burns, G. A., 171Burns, J. W., 402Burns, L. R., 503Burrows, L., 507Bursik, K., 398Burton, L. A., 611Bush, G., 257Bush, L. K., 187Bush, S. I., 605, 606Bushara, K. O., 282Bushman, B. J., 503Bushnell, M. C., 278Buskirk, J. R., 798Busool, A. N., 425Buss, A. H., 750Buss, D. M., 79, 116, 120, 124, 133Buss, K. A., 731, 732Buswell, B. N., 226, 459, 668, 683Butkovsky, L., 327, 337, 735Butler, E. A., 504Butler, L. T., 611Butterfield, P., 212, 351Bynum, C., 23Byrd, E. K., 677Byrne, N., 522Byrne, P. A., 54Byrnes, D., 765

Cabanac, M., 59, 274, 655, 778, 779Cabeza, R., 604Cacioppo, J. T., 180, 181, 183, 184, 185,

186, 187, 188, 189, 190, 200, 205, 219,458, 472, 516, 633, 778, 779, 781, 782,804

Cador, M., 252Cahill, L., 165, 166, 170, 171, 602, 603,

604, 605Cahill, S. E., 35Cain, W. S., 235Cajdric, A., 436, 463Calder, A. J., 224, 768Caldwell, C., 667Calhoun, C., 12Calkins, S. D., 488, 489, 733Callahan, C. M., 678Callan, V., 200Callanan, M. A., 337, 404Caltran, G., 799Calvert, G. A., 368Calvo, M. G., 402Camerer, C., 148Cameron, L. D., 654, 655, 680, 682Camodeca, M., 736Campbell, D. T., 480Campbell, D. W., 679Campbell, J. L., 736Campbell, L. A., 724Campbell, N., 207Campbell, S. B., 493Campbell, W. C., 602

Author Index 813

Campeau, S., 162Campos, B., 226, 789Campos, J. J., 131, 212, 225, 296, 298, 299,

306, 315, 334, 335, 351, 352, 370, 371,459, 487, 578, 681, 698, 730, 731, 732

Campos, R., 334Camras, L. A., 77, 215, 222, 295, 298, 299,

311, 333, 334, 336, 350, 358, 487, 698,730, 731

Cancian, F. M., 22, 24Canli, T., 170, 260, 268, 478Cannon, W. B., 10, 119, 159, 182, 183, 185,

190Cantagallo, A., 253Canteras, N. S., 164Cantor, J. J., 525Cantor, N., 479, 533, 551, 587Capatides, J., 582Caplin, A., 151Caporael, L. R., 437Cappella, J., 336Capps, L. M., 222Cardinal, R. N., 670Carleton, A., 237Carlo, G., 338, 443Carlson, J. J., 733Carlsson, K., 718Carmelli, D., 236, 677Carmichael, S. T., 164, 250, 603Carnethon, M. R., 677Carnevale, P. J. D., 549, 553, 555Carney, R. M., 667, 680Caro, R. A., 450Caro, T. M., 782, 786Caron, A. J., 313, 368, 369, 371Caron, R. F., 313, 368Carpenter, B. N., 797Carrasco, J. L., 700Carrasco, M., 170Carrier, B., 278Carrillo, J. D., 150Carrive, P., 163Carroll, J. M., 180, 553Carroll, S. A., 399Carson, J. L., 734Carstensen, L. L., 219, 379, 380, 382, 384,

385, 386, 399, 501, 515, 516, 525, 602Carstensen, L. O., 379Carter, C. S., 55, 61Carter, J. D., 404Carter, R. M., 171Caruso, D. R., 225, 399, 535, 539Carver, C. S., 472, 542, 652, 779, 803Cary, J., 447Caseras, X., 767Casey, R. J., 731Caspi, A., 487, 800Cassidy, J., 327, 337, 338, 518, 735Cassirer, E., 74Castonguay, L. G., 94Catron, T., 736Cauldwell, R. T., 202Cavanagh, K., 765, 767Cavanagh, P., 284Cayrou, G., 68Cechetto, D. F., 253Ceci, S. J., 332, 343, 556Centerbar, D., 81, 630, 631, 634Cerney, M. S., 798Cernoch, J. M., 240Cervantes, C. A., 337, 404Cesar, J., 700Cestari, V., 164Chagnon, N., 128Chaiken, S., 181, 608Chalelle, G., 272, 285Chalfonte, B. L., 610Chalip, L., 410Chalmers, D., 83Chambless, D., 700Champagne, F. A., 54Champion, V., 652Chan, M., 679Chang, B., 680Changeux, J. P., 171

Chantarujikapong, S. I., 722Chao, L. L., 257Chaplin, T. M., 397, 399, 404Chapman, G. B., 681, 682, 684Charland, L. C., 577Charles, S. T., 379, 382, 383, 384, 386, 387,

501, 525, 602, 678Charlesworth, W. R., 366Charmaz, K., 34Charney, D. S., 663, 670Chartrand, T. L., 185, 205, 207, 458Chawla, P., 399Chen, D., 241, 242Chen, K., 259, 279Chen, M., 507, 779Chen, Y., 399Cheng, D. T., 168Cheriff, A. D., 670Cherry, E. C., 170Chesney, M. A., 215, 222, 677Chess, S., 486Cheung, N., 459Cheung, V. H. W., 104Chevalier-Skolnikoff, S., 214Chiba, A. A., 252Chida, K., 160Chikama, M., 280Chin, S., 685Chiodo, L. M., 380, 384Chipperfield, J. G., 679Chirot, D., 770Chiu, C. Y., 506, 747Chiu, W. T., 694Choate, M. L., 694Chomsky, N., 103Chorpita, B., 696Chovil, N., 399Chow, L., 679Chow, S. Y., 798, 805Christensen, A. J., 680Christensen, E. E., 517Christensen, P. N., 404Christensen, T. C., 516, 517, 698, 780Christianson, S. A., 602, 605, 606, 699Christoff, K., 719Chu, E., 225Chua, P., 604Chung, J., 109Churchland, P. S., 11Chwalisz, K., 183Cialdini, R. B., 461Ciarrochi, J., 400, 520Cicchetti, D., 306Cicchetti, P., 161, 162, 163Cicero, M. T., 196Ciompi, L., 48Cirulli, F., 666Cisamolo, D., 588Claparède, E., 80Clark, A., 629Clark, A. E., 474Clark, C., 34Clark, C. L., 634Clark, L. A., 431, 472, 681, 694, 696Clark, M., 152, 552Clark, M. S., 458, 460Clark, R. D., 441, 444Clarkin, J. F., 91Classen, C., 236Clements, F. E., 332Cliff, N., 202Clore, G. L., 71, 81, 82, 143, 180, 181, 307,

308, 312, 355, 396, 398, 400, 474, 516,518, 523, 524, 535, 537, 553, 554, 555,556, 566, 567, 588, 593, 594, 606, 629,630, 631, 633, 634, 636, 637, 655, 657,744, 779, 784, 798, 799

Coan, J. A., 188, 189, 763Coates, D., 144Coats, E. J., 399, 403, 457Cobos, P., 183Cochrane, C. E., 698Codispoti, M., 73, 222Coe, C. C., 667Coffey, K. A., 787

Coffield, C., 368Cohen, A., 717Cohen, C., 385Cohen, C. E., 587Cohen, C. I., 677Cohen, D., 410Cohen, J. C., 163Cohen, J. D., 146, 147, 148, 152, 162, 171,

763Cohen, N. J., 160, 164, 166Cohen, S., 190, 645, 667, 670, 677, 678,

777, 788Cohn, J., 293Cohn, M. A., 785, 787, 791Coie, J. D., 735, 736Coifman, K. G., 805Colantuoni, C., 54Cole, A., 338Cole, J. O., 54Cole, P. M., 341, 397, 487, 731, 734Cole, S. R., 677Coleman, D., 89Coleman, J. R., 161Coleridge, S. T., 108Coles, M. G. H., 181, 182Collie, A., 610Collingwood, R. G., 111Collins, A., 71, 307, 355, 537, 593, 594,

629, 744Collins, M. A., 611Collins, P. F., 62Collins, R., 36, 40Collins, W. A., 489Comblain, C., 602, 606Conforti, N., 252Conlin, M., 145Conner, R., 54Conner, T., 515Conner, W. E., 244Connolly, K., 769Connolly, T., 684Connor, T., 517, 520Conrad, C. D., 165, 166Consedine, N. S., 377, 381, 382, 383, 649,

676, 677, 679, 680, 681, 682, 683, 684,685

Constantinidis, C., 171Contreras, J., 338Contreras-Grau, J., 337Conway, M. A., 520, 602, 605, 629Cook, C. L., 507Cook, E. W., III, 724Cook, M., 119, 128, 212Cooke, A., 611Cooke, D., 106Cooke, T., 323, 359Cooke-Carney, C., 221Cooley, C. H., 32Cooley, C. N., 668Coombs, M. M., 89Cooper, Z., 694Cooperman, S., 342, 734Copland, A., 107Corbin, A., 24, 764Corbit, J. D., 788Coren, S., 356Corkin, S., 602, 603, 604, 605, 606, 608, 610Corneille, O., 523Cornet, F. C., 711Corodimas, K. P., 162, 165, 166Cortese, B. M., 664Cosmides, L., 115, 116, 117, 118, 119, 120,

121, 124, 127, 128, 129, 131, 133, 134,224, 779, 780

Costa, P. T., Jr., 382, 387, 478Costello, J., 200Costello, K., 770Costenbader, V. K., 397Coté, S., 458Cottrell, C. A., 436, 437, 770Couchoud, E. A., 327, 354, 355, 356, 371Coull, J. T., 282Coulmas, F., 102Coulombe, D., 160Coups, E. J., 681, 682, 684

814 Author Index

Cousins, M. S., 253Cowan, R., 360Cowan, W. M., 251Cowie, R., 197, 207Cowley, J. J., 241Cox, A., 540Cox, A. W., 252Cox, B. J., 663Cox, C. R., 761Cox, D. L., 92Cox, D. S., 670Cox, J. F., 50, 57Cox, V. C., 251, 261Coy, D. H., 54Coyle, N., 384, 387Coyne, J. C., 801Craig, A. D., 52, 253, 259, 272, 273, 275,

276, 277, 278, 279, 281, 284, 285Craig, F. W., Jr., 281Craig, K. B., 222, 536Cramer, P., 400Cranford, J. L., 163Craske, M. G., 664, 694Crawford, C., 116Crepeau, L., 55, 57Crick, N. R., 493, 735Critchley, H. D., 171, 183, 190, 282Crittenden, N., 400Crockenberg, S. C., 492, 732Croghan, T. W., 680Crosby, F. J., 463Cross, D., 324Cross, S. E., 395Crowe, E., 328Crowley, S. L., 754Crown, C. L., 364, 441Cryder, C., 147Csikszentmihalyi, I., 778Csikszentmihalyi, M., 44, 778, 780, 781Cuddy, A. J. C., 770Culver, C., 377, 385, 732Culver, L. C., 377, 378Cumberland, A., 733Cumming, E., 379Cummings, E. M., 341, 493, 733, 734Cunningham, A. R., 236Cunningham, J. G., 106Cunningham, M. R., 558, 561Cunningham, W. A., 171, 629, 635, 638,

639Cupchik, G. C., 108, 239, 240, 243, 244Curtis, A., 224, 768Curtis, C. E., 171Curtis, J. T., 94Curtis, V., 758Cuthbert, B. N., 73, 699, 722, 723, 724, 725Cvengros, J. A., 680

Dabbs, J. M., Jr., 646Dahl, R. E., 190, 220Dahlstrom, W. G., 731Daily, L. Z., 525Dale, W., 679Dalgleish, T., 143, 768Dallos, R., 768Daly, M. B., 648, 679Daly, M., 116, 120, 128Damasio, A. R., 52, 53, 54, 61, 104, 142,

151, 170, 182, 189, 224, 253, 272, 274,281, 610

Damasio, A., 12, 73, 77, 82, 83, 160, 172,181, 763

Damasio, H., 53, 142, 151, 224, 253, 284,610

D’Amato, F. R., 55Dana, J., 152Dandoy, A. C., 503D’Andrade, R. G., 410Daneman, M., 525Daniels, M., 667Danner, D., 777, 788Dantzer, R., 667D’Argembeau, A., 602, 606, 608Darlington, R. B., 152Darwin, C. E., 377

Darwin, C. R., 306, 307, 308, 315, 317,742, 743, 744, 757, 759, 762, 770, 771

Darwin, C., 63, 95, 104, 186, 197, 205, 211,212, 214, 219, 222, 225, 227, 249, 273,291, 534, 668, 799

Dasgupta, N., 436, 463Daubman, K. A., 536, 549, 551, 552, 553,

567, 784Daum, I., 202Davachi, L., 610Davey, A., 382, 387Davey, G. C. L., 683, 760, 765, 766, 767,

768, 769David, A. S., 603Davidson, R. J., 96, 188, 189, 190, 219,

223, 224, 252, 260, 284, 307, 411, 503,603, 634, 665, 666, 667, 670, 698, 767,779, 786, 789, 797, 798

Davidson, R., 759Davies, B., 799Davies, P. T., 493, 734Davies, P., 205Davies, S., 107Davis, D. M., 699Davis, M. C., 516, 525, 676, 804Davis, M., 160, 162, 163, 164, 167, 252,

721, 722, 804Davis, P. J., 398Davison, G. C., 701Dawes, R. M., 149Dawkins, R., 203, 212Dawson, M. E., 241Day, A. L., 399De Rosa, E., 719de Rosnay, M., 324, 326, 328, 349, 359,

518, 521de Sousa, R., 10, 13, 14, 82de Swaan, A., 22De Vries, N. K., 141De Waal, F. B. M., 81, 217, 283, 285De Acetis, L., 666Deacon, T. W., 205Dean, J. L., 285Dean, J., 379Deane, F. P., 520Dearing, R. L., 39, 683, 749Deary, I. J., 282Deci, E. L., 479, 560Decker, L. R., 252Deckner, W. C., 646Dedmon, S. E., 489De Dreu, C. K. W., 460DeEskinazi, F. G., 56, 60, 61Degel, J., 239de Gelder, B., 170, 223, 521, 720De Geus, E. J. C., 475de Greck, M., 50De Groot, D. M. G., 263de Gucht, V., 522DeGroat, D. A., 242Dehaene, S., 171De Houwer, J., 83Deichert, N. T., 281de Jong, P. J., 459, 767Del-Ben, C. M., 664Deldin, P. J., 670Delgado, M. R., 167, 169, 605Della Sala, S., 386Del Mar, D., 28Delumeau, J., 20Demaree, H. A., 386, 504Demaret, C., 523DeMatteo, M. R., 227Dember, W. N., 784Demler, O., 694Demos, J., 20, 23Dempesy, C. W., 258DeMulder, E., 338, 734den Dulk, P., 163Denham, S. A., 327, 334, 337, 338, 349, 354,

355, 356, 371, 489, 732, 733, 734, 735Dennett, D. C., 13, 69Denney, C., 200Dennis, T. A., 487Denton, D., 53, 60, 62

D’Entremont, B., 350Depue, R. A., 62, 700Derakshan, N., 716Derryberry, D., 96, 126, 478, 486, 549, 553,

554, 555, 784Desai, M. M., 679Descartes, R., 4, 6, 7, 10, 68, 83Desmond, J. E., 257, 260D’Esposito, M., 171, 253DeSteno, D., 152, 436, 463, 594Detmar, S. B., 679Detweiler, J. B., 536, 541Deutsch, R., 507DeVellis, B. M., 678Deveney, C. M., 670Devereux, P. G., 206Devine, P. G., 396Devinsky, O., 47, 253DeVivo, M. J., 474DeVoe, M., 380, 384, 386Devos, T., 431DeWall, C. N., 629, 636Dewey, J., 9, 10, 249, 335Dewhurst, S. A., 605, 606DeWinter, A. F., 492Dhabhar, F. S., 663Dhar, R., 146Diamond, D. M., 166Dias, M., 409, 766DiBiase, R., 750Dickerson, J., 184Dickerson, S. S., 662, 663, 666, 667, 668,

669, 682Dickson, K., 298Diefenbach, M. A., 646, 648, 649, 654, 655,

656, 678, 679, 682Diehl, M., 380, 384, 387Diener, C., 480, 779, 782Diener, E., 183, 387, 397, 398, 472, 473,

474, 476, 477, 478, 479, 480, 481, 515,589, 777, 779, 782, 791

Diener, M., 476, 479Dietrichs, E., 258DiGirolamo, G. J., 596Dijker, A. J., 429Dijkers, M., 474DiLiberto, L., 736Dillon, D. G., 503DiMatteo, M. R., 228, 680, 682Dimberg, U., 187, 212, 224, 243, 397, 399,

458, 460, 711, 712, 714, 715Dimock, E. C., 411Dindia, K., 401DiPaolo, M., 536Dishion, T. J., 491, 493Ditto, B., 522Ditto, P. H., 151, 651Djamasbi, J., 561Dobbins, I. G., 611Doctoroff, G., 767Dodd, G. H., 235Dodge, K. A., 343, 489, 490, 492, 493, 731,

733, 735, 736Doerksen, S., 608, 610Doherty, R. W., 397Dolan, R. J., 52, 167, 170, 171, 183, 223,

603, 604, 605, 609, 719, 720, 721Dolcos, F., 604, 605, 608Dolgin, K., 366Dolhinow, P. J., 782D’Olimpio, F., 768Dolins, F., 733Doll, E., 383Dollard, J., 646Dolski, I., 667Donchin, E., 181Dondi, M., 799Donker, F. J. S., 677, 680Donnellan, M. B., 473Donzella, B., 667Doosje, B., 431, 434, 435, 463Doosje, E. J., 463Dore, J., 364, 441Dorries, K. M., 238, 241Dorval, B., 385

Author Index 815

Doty, R. L., 236Dougherty, L. M., 293, 731Douglas, L., 298Douglas, M., 760Douglas-Cowie, E., 207Dovidio, J. F., 202, 205, 429, 785Downey, J. E., 104, 107Downhill, J. E., 188Doyal, G., 521Doyle, W. J., 190, 677, 777Dozois, D. J. A., 698Dreifus, C., 388Dreisbach, G., 555, 785Drevets, W. C., 165, 664, 666Driver, J., 170, 596, 603, 719, 720, 721Driver-Linn, E., 145Dror, O. E., 28Drossaert, C. C., 650Du Bois, M. A., 520Dube, L., 678DuBrin, J. R., 92Duchenne de Boulogne, G. B., 212Duda, R. B., 679Duffy, E., 71Duffy, M. E., 700Dufwenberg, M., 148Duguid, S., 629Duke, M. P., 228Dulin, J., 563Dumais, S. T., 713Dumas, G., 71, 75, 81Dumont, M., 433, 463Dunbar, R. I. M., 205Duncan, E., 74Duncan, G. H., 278Duncan, L. A., 770Duncan, S. W., 322Duncker, K., 552Dunlop, L., 767Dunn, B. D., 143Dunn, E. W., 540Dunn, J., 321, 327, 328, 337, 353, 733, 734, 785Dura, J. R., 190Durbin, C. E., 734Durkheim, E., 32, 40Durrett, M. E., 786Dutton, K., 717, 718Duval, S., 313Duvernoy, H. M., 256Dweck, C. S., 506, 743, 746, 747, 749Dwyer, K. D., 734Dwyer, K. M., 492Dyaz-Marsa, M., 700

Eagly, A. H., 181Early, S., 442Earnst, K. S., 695Easterbrook, J. A., 553Eastwood, J. D., 716Ebert, R., 227, 759Ebstein, R., 487Edelman, G. M., 249, 518Edelman, R. J., 750Edmans, A., 143Edmondson, D., 371Edmonson, M. S., 205Edwards, K., 219, 682Edwards, R., 327Egekvist, H., 670Eggleston, R., 35Egloff, B., 503, 505Ehlers, A., 608, 611Ehlert, U., 285Ehrlichman, H., 594Eiben, L. A., 240Eibl-Ebesfeldt, I., 133, 205, 206, 212, 214Eich, E., 594Eichenbaum, H., 160, 164, 166Eid, M., 474Einon, D. F., 786Eisenberg, E., 443Eisenberg, N., 147, 212, 336, 338, 343, 397,

443, 448, 458, 487, 488, 489, 501, 730,731, 733, 734, 735, 736

Eitel, P., 678Ekman, P., 75, 76, 79, 103, 119, 126, 127,

131, 181, 183, 184, 185, 187, 188, 190,211, 212, 213, 214, 215, 216, 217, 218,219, 221, 222, 223, 224, 225, 226, 227,291, 292, 295, 307, 310, 333, 385, 411,513, 534, 592, 629, 662, 670, 758, 759,781, 797, 798

Elfenbein, H. A., 223Elias, N., 21, 22, 23, 762Eliot, T. S., 109Ellestad, J., 38Ellgring, H., 215, 222Elliot, A. J., 472, 743Elliott, D., 182Elliott, R., 88Ellis, B. J., 342, 344Ellis, C. J., 680Ellis, G. F. R., 48, 53Ellison, C. G., 476Ellsworth, P. C., 78, 151, 185, 212, 219,

414, 499, 628, 682, 779, 780, 800Ellwart, T., 716Ellyson, S. L., 205Elmehed, K., 187, 243Elmore, B. C., 228Else-Quest, N. M., 403El-Sheikh, M., 341, 734Elster, J., 70, 71, 72, 151, 307Elvin, M., 23Ely, T. D., 170, 603, 604Emde, N. E., 441Emde, R. N., 131, 200, 212, 298, 299, 315,

337, 351, 370, 371, 459, 578Emerson, R. M., 42Emery, G., 508Emery, N. J., 720Emmons, R. A., 472, 478, 515, 789, 790Eng, A., 111Engelberg, E., 606Engle, R. W., 519, 525Ennis, M., 250Enquist, M., 243Epel, E. S., 538Epstein, E. B., 151Epstein, S., 387, 534, 649, 655, 710, 724Erb, J., 637Erb, M., 603Erber, M. W., 559Erber, R., 537, 541, 559Erez, A., 551, 558, 560, 566, 682Erickson, K., 664Ericson, K. M., 152, 719Erikson, E. H., 19, 742, 744, 745Ernst, J., 119Essex, M. J., 342Esteves, F., 212, 224, 714, 715, 716, 722Estrada, C. A., 549, 564, 784Etzel, J. A., 184Eustace, N., 22, 26Evans, A. C., 255, 278Evans, J. St. B. T., 149Evans, M. K., 225Evans, P., 670Evans, S. M., 263Evans, T. W., 536Everitt, B. J., 160, 164, 165, 252, 670Evers, C., 457, 458, 507Ewart, C. K., 190Eyler, A. A., 683Eyre, H. L., 397Eysel, U. T., 767Eysenck, H. J., 750, 769Eysenck, S. B. G., 769Eyssell, K. M., 398, 458, 516

Fabes, R. A., 335, 336, 338, 343, 443, 448,731, 734, 735

Fabre-Thorpe, M., 171Fahey, J. L., 664, 668, 669, 670Fainsilber, L., 180Fair, J. M., 677Fair, P. L., 187Fairburn, C. G., 694

Fairchild, S. K., 797Falla, S. J., 717Fallon, A. E., 683, 758, 759, 760, 761, 765,

766Fanelli, M. M., 522Fanselow, M. S., 160, 162, 163, 164, 252Fant, G., 198Farah, M. J., 253Farb, C., 161Farchione, T., 521Faroy, M., 487Farraye, F. A., 679Faulkner, J., 770Febvre, L., 19Fedorikhin, A., 152Feeney, J. A., 336Feeny, N. C., 91Fehr, B., 356, 517, 518, 592, 593Fehr, E., 56Feinman, S., 306, 351Feiring, C., 752Feldman, L. A., 516, 518, 525Feldman, R. S., 399, 457, 465Feldman, S., 252Feldstein, S., 364, 441Fendt, M., 163Fera, F., 223Ferenczi, S., 765Ferguson, T. J., 397, 398, 754Fernald, A., 352, 370, 732Fernandez de Molina, A., 160Fernández-Dols, J. M., 103, 197, 223, 333Ferris, T. G., 679Ferry, A. T., 256, 669Feshbach, S., 646Fessler, D. M. T., 151, 758, 767, 770Fiddick, L., 128Fiedler, K., 594Field, T., 731Fielding, D. M., 680Fielding, R., 679Figley, C. R., 443Figueiredo, H., 669Figueroa, R., 735Filipkowski, R. K., 164Filsinger, E. E., 241Finch, D. M., 253Fine, G. A., 770Fine, S. E., 377, 736Fineman, S., 781Fingerman, K. L., 382, 386Fink, A. K., 680Finkel, S. M., 787Finkenauer, C., 781Finucane, M. L., 561, 647, 654, 655, 657, 679Fiore, M. J., 385Fischbacher, U., 56Fischer, A. H., 396, 397, 400, 401, 402, 457,

458, 459, 460, 461, 463Fischer, K. R., 206Fischhoff, B., 151Fischler, B., 522Fisek, M. H., 35Fish, S., 410Fiske, S. T., 770Fitch, W. T., 103Fite, J. E., 493Fitness, J., 458Fitzgerald, M., 735Fitzpatrick, M., 94Fitzsimmons, J. R., 719Fivush, R., 337, 518, 520, 524Flack, W. F., Jr., 281Flament, C., 636Flamson, T. J., 151Flandrin, J. L., 20, 24Flavell, J., 765Fleeson, W., 382Fleming, A. S., 240Fleshner, M., 666Fletcher, C., 537Fletcher, P., 604Flicker, L., 459Flom, R., 367, 368, 369, 371

816 Author Index

Florer, F., 605Floyd, N. S., 669Flykt, A., 713, 714, 715, 716Foa, E. B., 91Fodor, J., 594Fogarty, S. J., 548Fogel, A., 205, 297, 298, 299, 300, 301,

334, 372Foisy, M. L., 222Foley, M. A., 607Folkman, S., 502, 575, 652, 661, 678, 787,

798, 802Fong, G. W., 147Fontana, C. J., 258Foran, J. M., 187Ford, C. E., 462Ford, M. E., 786Forehand, R., 735Forest, D., 552, 558Forgas, J. P., 553, 594, 655Forgays, D. K., 339Forster, J., 553, 555Forster, L., 766Fortmann, S. P., 677Fosha, D., 88Foss, M. A., 180, 518, 588Fox, E., 716, 717, 718, 725Fox, K. E., 539Fox, N. A., 188, 307, 342, 358, 488Fox, P. T., 255Frackowiak, R. S. J., 255Fraeff, F. G., 664Fraiberg, S., 333Fraley, C., 517Franchi, C., 240Francis, D. D., 54Franco, F., 733Frank, J., 94Frank, M. G., 215, 217, 223, 225Frank, R. H., 80, 82, 212, 128Frankel, C. B., 487, 698Frankland, P. W., 164Franks, D., 43Frawley, W., 521Frazer, J. G., 759, 760, 761Frederick, S., 146, 147, 480Frederickson, B., 95Fredrickson, B. L., 126, 480, 503, 525, 549,

553, 555, 565, 677, 678, 681, 682, 684, 777,778, 779, 780, 781, 782, 783, 784, 785, 786,787, 788, 789, 790, 791, 803, 804

Fredrikson, M., 719Fredstrom, B. K., 736Freedland, K. E., 667Freedle, R., 337Fresco, D. M., 698Fretter, P. B., 94Freud, A., 507Freud, S., 8, 9, 10, 11, 12, 88, 109, 312,

497, 534, 633, 742, 744, 758, 765Frewin, P. A., 698Freyberg, R., 239Freyberger, H. J., 698Frick, R. W., 197Fridel, Z., 162Fridlund, A. J., 127, 187, 212, 291, 292,

456, 457, 460Friedlander, L., 521Friedlmeier, W., 297, 300Friedman, H. S., 228, 680Friedman, R. S., 553, 555Friedrich, F. J., 717Friend, R., 678Friesen, W. V., 126, 183, 187, 188, 212, 213,

214, 217, 219, 222, 223, 225, 226, 295,310, 385, 670, 758, 759, 777

Frieze, I., 752Frijda, N. H., 71, 72, 73, 74, 76, 77, 78, 80,

83, 110, 181, 414, 430, 431, 434, 456,498, 499, 535, 574, 577, 580, 628, 680,684, 691, 779, 780, 781, 798

Frisch, D., 147Friston, K. J., 167, 255Frith, C. D., 171, 223

Frith, U., 257Fritz, J., 337, 353Fudenberg, D., 150Fujita, F., 387, 473, 479, 480, 798Funayama, E. S., 167Fung, H. H., 379, 410, 505, 685, 779Furnham, A., 539Fuster, J. M., 171Futterman, A. D., 670

Gable, S. L., 785Gabriel, C., 107Gabriel, S., 549, 799Gabrieli, J. D. E., 169, 170, 257, 258, 260,

504, 719Gabrielsson, A., 104Gaddy, G., 328Gaertner, S. L., 429, 785Gaffan, D., 160, 171Gaffan, E. A., 160Gagnon, L., 107, 180Gaines, S. O., 540Galanter, E., 145, 577Gall, M., 540Gallagher, D., 183Gallagher, M., 252Gallese, V., 74, 82, 441, 592, 594, 595,

765Gallistel, C. R., 583Gallo, D. A., 608Gallo, L. C., 665, 676Gallois, C., 200, 336Gallup, G. G., 135Galor, O., 801Galotti, K. M., 199Ganchrow, J. R., 444Gandevia, S. C., 278Gangemi, A., 59, 111Ganz, P. A., 477Garavan, H., 603Garcia, C., 183García, D., 143Garcia, J., 128Garcia, L., 680Garcia, R., 164Garcia-Leal, C., 664Gard, D. E., 698Gardiner, I. M., 604Gardner, H., 533Gardner, R., Jr., 669Gardner, W. L., 180, 781, 804Garfield, D. A. S., 514, 521Gärling, T., 655Garner, P. W., 328, 734Garnham, A., 109Garoff-Eaton, R. J., 606, 608Gartner, J., 476Gary, Y., 221Gaschke, Y. N., 536, 537, 538, 589Gascoigne, P., 679Gasper, K., 553, 554, 555, 566, 567Gatenby, J. C., 167Gates, G. S., 357Gati, I., 551Gatz, M., 382Gay, D. A., 476Gay, P., 19Gazzaniga, M. S., 181Geen, T., 214Geer, J. H., 605, 606Geertz, C., 410, 420Geiselmann, B., 382Gencoz, F., 677Gencoz, T., 677Gendron, M., 292Gentzler, A., 337, 338George, J. M., 462George, L. K., 477Georgellis, Y., 474Georgiadis, J. R., 250Georgiou, G. A., 718Geppert, U., 750Gerard, G., 170Gerbarg, P. L., 285

Gergen, K. J., 22Gerhards, J., 25Gerrig, R. J., 102Gervais, M., 782Geva, N., 552, 557, 563Gewirtz, J. C., 162, 164Gianetto, R. M., 399Gibbard, A., 13, 14Gibson, D., 224, 768Gibson, E. J., 333, 367Gibson, J. J., 306, 367, 373, 631Gigerenzer, G., 124Gil, K. M., 788Gilbert, A. N., 239Gilbert, D. T., 142, 145, 151, 502Gilbert, P., 668, 669Gill, K. L., 489, 733Giller, E. L., 664Gillespie, M., 677Gilligan, S. G., 594Gillis, J. R., 19, 22, 26Gilman, B. I., 104Gilmer, B., 534Ginsburg, G. P., 206Giovanello, K. S., 610Girbaumer, N., 603Girdler, S. S., 398Givens, B., 252Glaser, D., 61Glaser, R., 190, 677, 780, 789Glass, D. C., 678Glass, T. A., 476Glassner, B., 29Glazer, H. I., 81Glazer, K., 665, 676Gleiberman, L., 677Glenn, N. D., 476Glenn, P. J., 205Glick, P., 770Globisch, J., 722Gloor, P., 160Glover, D. A., 664Glover, G., 260Gnepp, J. C., 446Gnoli, R., 411Goddard, G., 160Godinez, A., 278Goering, P., 678Goffman, E., 34Gogate, L. J., 373Goguen, L. A., 380Gohm, C. L., 400, 523, 524Gold, E. R., 752Gold, J., 753Gold, P. W., 165Goldbeck, T., 199Goldberg, S., 307Goldenberg, J. L., 761Goldfried, M. R., 88, 94Goldie, P., 11Goldin, P. R., 504, 505Goldman, R. N., 94Goldman, R., 94Goldman, S. L., 538Goldman, S., 370, 472Goldman-Rakic, P. S., 169, 171Goldsmith, H. H., 403, 731, 732Goldstein, A. G., 503Goldstein, D. C., 152Goldstein, D. G., 152Goldyn, D. T., 370Goleman, D., 538, 543Gollwitzer, P. M., 507Golman, S. L., 90Gomberg, D., 385Gomberg-Kaufman, S., 541Gomez, J. J., 695Gonsalves, B., 609Gonzaga, G. C., 103, 212, 226, 789, 790Gonzalez, J. S., 680Gonzalez, R. M., 190, 220Gonzalez, R., 541Gonzalez-Lima, F., 164Goodall, J., 802

Author Index 817

Goode, W. J., 27Goodenough, F. L., 732Goodlin-Jones, B. L., 240Goodman, G. S., 583Goodman, W. K., 758, 768Goodnight, J. A., 493Goodnow, J., 341, 410Goodwin, J. S., 677Goodwin, R. D., 403Goosens, K. A., 252Goossens, F. A., 736Gopnik, A., 352, 360Gordijn, E., 433, 463Gordon, A. H., 396, 402Gordon, B., 758Gordon, L. B., 89Gordon, N. S., 57Gordon, R. M., 12Gordon, S. L., 18, 22, 34Gore, J. C., 167Gorgoglione, J. M., 567Gorman, J. M., 165, 701Goschke, T., 555, 785Goss, C. M., 214Gosselin, P., 215, 218, 356Gotlib, I. H., 403, 508, 697Gottlieb, B. H., 801Gottlieb, J., 791Gottman, J. M., 98, 224, 226, 335, 336,

379, 399, 458, 459, 537, 584, 733, 734,735, 779, 791

Gottschalk, L. A., 677Gould, S. J., 81, 120Gove, F. L., 786Gove, W. R., 476Govender, R., 608Grabe, H. J., 698Graff, J., 379Grafman, J., 282Grafton, S. T., 170, 187, 604Gragnani, A., 768Graham, B. G., 244Graham, S., 447, 736Gralinski, J. H., 226Gram, I. T., 649Grammer, K., 205, 206, 242Gramzow, R., 537, 752, 754Grandjean, D., 574Granic, I., 491, 492, 494Gratton, A., 253Gravenstein, S., 190Graves, A. B., 236Gray, H. M., 214, 225, 800Gray, J. A., 52, 78, 165, 478, 487Gray, J. M., 224, 768Gray, J. R., 553, 799Gray, W. D., 517Grayson, C., 400Graziano, W. G., 397Green, D. F., 472Green, D. P., 180Green, F., 798Green, R. S., 202Greenberg, L. S., 88, 89, 90, 91, 92, 93, 94,

95, 96, 98, 700Greenberg, M. T., 736Greenberg, P. S., 187Greene, D., 149Greene, T. R., 549Greenhouse, L., 451Greenlee, M. W., 171Greenspan, J. D., 279Greenspan, P., 14Greenspan, R. J., 50Greenwald, M. K., 187, 602, 605Greenwood, A., 310Greer, M., 803Gregg, T. R., 250, 665Gregory, D. S., 736Greve, D., 170Greven, P. J., Jr., 21Grewal, D., 535, 540Griffin, K. W., 678Griffith, R. M., 28Griffiths, P., 14, 456

Griffiths, T. D., 285Grill-Spector, K., 611Grillon, C., 167Grimes, K. E., 699Griner, L. A., 629Grisham, J. R., 724Griswold, R. L., 22Grob, A., 476Grodd, W., 603Gross, J. J., 34, 91, 169, 181, 216, 217, 219,

249, 258, 382, 386, 387, 397, 400, 437,457, 497, 498, 499, 500, 501, 503, 504,505, 506, 507, 508, 515, 516, 630, 680,693, 697, 698, 780, 798, 799

Gross, J. N., 325Grosser, B., 241Grossman, M., 396Growdon, J. H., 602, 603Gruendel, J., 322Gruenewald, T. L., 663, 667, 668, 682Grühn, D., 383Grunfeld, D. I., 760Grusec, J. E., 730Guadagnoli, E., 680Guerra, P., 785Guijarro, M. L., 665Guimaraes, F. S., 664Guinn, J. S., 39Gundersen, H. J. G., 263Gunnar, M. R., 307Guo, L., 169Gurwitz, J., 680Gusella, J. L., 370Gustavson, A. R., 241Gustavson, B. J., 521Guth, W., 148Guthrie, I. K., 735Gutierrez, E., 721Gutmann, D., 403Gutmann, M., 645

Haan, N., 221Haber, S. N., 280Hadjikhani, N., 170Hadwin, J., 324Haft, W., 364, 441Hagemann, D., 188, 190Hager, J. C., 188Hagglund, D., 679Haidt, J., 68, 219, 224, 409, 410, 412, 421,

456, 462, 593, 683, 757, 758, 759, 761,762, 763, 766, 767, 782, 789, 790

Haines, D. E., 256, 258Hakeem, A. Y., 280Hakin-Larson, J., 380Halberstadt, A. G., 337, 399Halberstadt, J. B., 594Haley, K. J., 758, 770Halgren, E., 160Halisch, C., 226Halisch, F., 226Hall, I. E., 91Hall, J. A., 227, 396, 398, 399, 403, 404, 405Hallet, A. J., 665Hallett, M., 282Halligan, S. L., 611Halpern, B. P., 238, 241Halpern, J. N., 594Halpern, M., 238Halverson, C. F., Jr., 382Hamann, S. B., 170, 603, 604Hambrick, D. Z., 525Hamilton, D. L., 434Hamilton, L. R., 717Hamilton, M., 683Hamilton, N. A., 676Hamilton, W. D., 133Hamm, A. O., 187, 605, 713, 722Hammer, L., 765Hamrick, N., 656Hamsher, J. H., 228Han, M., 679Han, S. H., 736Han, Z. S., 276Hanish, L. D., 343, 734, 735

Hankin, B. L., 404Hansell, S., 678Hansen, S., 222Hanslick, E., 104, 108Hanson, E., 536Hansson, R. O., 797, 798Harber, K. D., 542, 801Harburg, E., 677Hardin, C. D., 202Hardman, C. E., 322, 357Hare, R. D., 722Hareli, S., 440Haring, M. J., 477Hariri, A. R., 190, 223Haririr, A. R., 220Harker, L. A., 221Harlow, T. F., 542Harmon-Jones, E., 189, 636Harré, R., 22, 73, 76Harris, C. R., 215, 217, 218, 504, 679Harris, J. R., 475Harris, P. L., 321, 322, 323, 324, 325, 326,

328, 329, 349, 353, 357, 358, 359, 361,518, 521

Harris, P., 305Harris, V., 736Harrison, A. A., 611Harrison, D. W., 386, 399Harrison, R. H., 357Harrison, S., 160Harro, J., 52, 61, 62Harrod, M. M., 39Harsch, N., 605, 607Hartman, K. A., 645, 651, 652Harver, A., 186Harvey, A., 694, 697Haselton, M. G., 124Hasher, L., 396Hashtroudi, S., 608Haslam, N., 770Haslam, S. A., 429Hasselmo, M. E., 252Hastings, P. D., 492, 734Hatfield, E., 200, 202, 203, 205, 458, 461Hauser, M., 103, 212, 214Haviland, J. M., 350, 364, 365, 385, 521,

730, 731, 732Haviland-Jones, J., 239, 241, 242, 243, 244,

297, 377, 378, 384, 387, 486, 730Hawilo, M. E., 61Hawkins, N. A., 651Hay, D. F., 444Hay, J. L., 679Hayden, E. P., 734Hayes, A. M., 94Hayes, C. W., 399Hayes, S. C., 694Haynal, V., 216, 222Haynes, O. M., 387Hazen, N. L., 786He, G., 768Headey, B., 474, 478Heaphy, E., 791Heath, R. G., 49, 53, 62, 252, 258Heatherton, T. F., 134, 459Hebb, D. O., 71, 80, 81Heberlein, A. S., 170Hebl, M. R., 404Hecht, H., 716Hecht, M. A., 399Heckhausen, H., 746Heerebout, B. T., 163Heerey, E. A., 222, 253, 669Heesacker, M., 399Hegel, G. W. F., 8Hegley, D., 293, 731Hegtvedt, K. A., 42Heidegger, M., 9Heider, K. G., 219, 226Heilman, K. M., 284Heim, C., 54Heiman, R., 410Heimberg, R. G., 698Heims, H., 768Heine, S. J., 515

818 Author Index

Heinrichs, M., 56, 285, 789Heinze, H., 184, 383, 402Heinzel, A., 50Heise, D. R., 32, 38, 39Heiser, W., 522Hejmadi, A., 411, 759, 765, 766, 768, 783Hektner, J. M., 44Hellawell, D. J., 224Heller, M., 216, 222Hellstrom, P., 222Helmstetter, F. J., 162, 168Heltman, K., 205Hembree, E. A., 91, 293, 731Hemenover, S. H., 759Henderson, L. A., 278, 281Henderson, L. N., 451, 452Hendriks, M., 239Hennekens, C. H., 665Henningfield, J. E., 678Henriques, J. B., 188, 189, 260Henry, S., 429Henry, W., 379Henson, R. N., 604Henson, R., 611Herbert, T. B., 190, 662, 667Herkenham, M., 161Hermalin, B. E., 562Herman, B. H., 54, 55, 56Herman, J. L., 803Herman, J. P., 669Hernandez, M. W., 575, 576, 577, 580, 581,

582Herrera, C., 370Herrera, E., 731Herrmann, I., 92Herrnstein, R. J., 146Hershey, J., 143Hertenstein, M. J., 212, 225, 351, 361Herz, R. S., 239Hess, M., 187Hess, U., 187, 197, 199, 204, 205, 207, 217,

336, 396, 459Hess, W. R., 49, 61, 62Heth, C. D., 164Hetherington, E. M., 489Hettema, J. M., 722, 725Heuer, F., 583, 611Heuer, K., 716Heyman, R., 789Hiatt, S., 299Hicken, B. L., 474Hickling, A. K., 447Highberger, L., 449Hildebrand-Pascher, S., 358Hill, A. L., 489Hilton, S. M., 160Hiltunen, T., 197Hinsz, V. B., 404Hird, K., 199Hirsch, J., 784Hirshleifer, D., 143, 561Hirshleifer, J., 128Hirt, E. R., 551Hitchcock, D. F. A., 306Hjort, M., 102Ho, K., 141Hobson, P. R., 222Hoch, S. J., 139Hochschild, A. R., 25, 32, 34Hodges, S. D., 399Hodson, G., 770Hoehn-Saric, R., 723, 724Hoeksma, J. B., 488Hof, P. R., 253Hofer, L., 364, 441Hoffman, C., 379Hoffman, J. M., 603Hoffman, M. L., 441, 442, 443, 444, 445,

446, 448, 449, 451Hoffman, R. R., 180Hogan, P. C., 110Hogg, M. A., 428, 434, 461Hoke, L. A., 679Holen, A., 498Holender, D., 713

Holland, R. W., 239Hollander, E., 700Hollbach, S., 433, 434, 437, 462Hollon, S., 700Holmse, D. A., 54Holodynski, M., 297, 300Holowaty, K. A. M., 91Holsboer, F., 664Holstege, G., 54, 250Holt, R., 327Hommer, D., 147Hong, M., 402Hong, Y., 747Hong, Y. Y., 506Hoogstra, L., 410Hooker, K., 477Hooley, J. M., 693Hooten, E. G., 678Hooven, C., 335, 733Hopcroft, J. E., 107Hope, D. A., 694Horgan, T. G., 404Horn, N., 152Hornak, J., 253Hornung, K., 222, 306Horowitz, F. D., 365Horowitz, M. J., 498, 652, 758Horowitz, T., 765Horstmann, G., 223Horton, D., 684Horton, R., 414, 415, 418Houlbrooke, R., 24House, P., 149Houser, J. A., 35, 36Howard, C. V., 263Howard, D., 285Howell, N., 760Howells, G. N., 736Howes, D., 236Howland, R. H., 667, 700Hoyt, R. E., 151Hsee, C. K., 138, 143, 151, 200, 542, 651,

679Hsu, H., 298Hsu, H. C., 205Hua, L. H., 278, 279Hubbard, J. A., 337, 735, 736Huber, R., 49, 52Hucklebridge, F., 670Hudley, C., 736Hudson, J. A., 239Huebner, R. R., 293, 298Huff, N. C., 164Hufnagel, H., 693Hugdahl, K., 167, 713Hui, S. L., 678Hume, D., 3, 7, 8, 309, 442, 534Hunsley, J., 89Hunsperger, R. W., 160Hunt, M. G., 92Hunt, W. A., 71Hunziker, J., 378Huperich, S. K., 697Hupka, R. B., 76Huppert, J. D., 767Hurley, J. C., 337, 736Hurley, K., 645Huron, D., 102, 629, 632, 635, 636, 639Hursti, T., 767Husaini, B. A., 679Husband, A. J., 663Husserl, E., 9Husted, D. S., 758, 768, 769Huttenlocher, J., 353Hutton, D., 323, 359Hyde, J. S., 396, 403Hyde, S., 222Hyman, A., 6Hyman, L. M., 636Hymel, S., 343Hynes, K., 400Hyyppa, M. T., 522

Iacoboni, M., 441Igarashi, M., 163

Iglesias, J., 366Ikemoto, S., 52, 56, 59Imada, S., 219, 683, 757, 761, 763Impett, E. A., 785Infrasca, R. R., 522Ingalls, D. H., 110Ingram, R. E., 593Ingvar, M., 282Inhelder, B., 765Inhoff, A. W., 717Insel, T. R., 55, 61, 278Insler, R. Z., 257Intriligator, J., 284Iran-Nejad, A., 472Irons, D., 249Irwin, M., 667Irwin, W., 603Isaac, R., 21Isaacowitz, D. M., 382Isbell, L. A., 711, 721Isen, A. M., 96, 149, 152, 536, 548, 549,

550, 551, 552, 553, 554, 555, 556, 557,558, 559, 560, 561, 562, 563, 564, 566,567, 568, 606, 682, 784, 785, 789

Ishai, A., 611Islam, A., 205Ito, T. A., 183, 185, 779, 782Ives, C., 107Iwakabe, S., 91Iwata, J., 160, 163Iyengar, S. S., 561Iyer, A., 435, 463Izard, C. E., 75, 76, 187, 211, 221, 223,

224, 291, 292, 293, 298, 307, 309, 311,315, 349, 356, 358, 365, 376, 377, 378,385, 387, 513, 535, 540, 682, 684, 730,731, 736, 742, 750, 759, 762, 797, 798,799

Jackendoff, R., 105Jackson, B., 400Jackson, D. C., 188, 503Jackson, J. J., 342Jackson, O., 610Jacob, S., 241Jacobs, S., 798Jacobs, W. J., 164Jacobson, J. A., 151Jacobson, N., 584Jacobson, N. L., 244Jaffe, J., 364, 441Jain, E., 184, 383, 402Jain, K., 760Jalowiec, J., 56, 57James, T. T., 521James, W., 3, 9, 10, 58, 60, 70, 76, 159, 182,

185, 249, 272, 281, 335, 499, 534, 540,628, 639, 657

Jandorf, L., 670Janis, I. L., 646Janoff-Bulman, R., 144, 752, 798Jarrard, L. E., 164Jasnow, M. D., 364, 441Jasso, G., 42Java, R. I., 604Jaycox, L. H., 91Jellis, J. B., 91Jenkins, J. M., 103, 457, 460, 731, 733, 778,

780, 797Jennings-White, C., 241Jensen, L. A., 409, 421Jessop, D. C., 682Jewett, J., 574, 580, 582Jindal, R., 700Joergen Grabe, H., 521Johanson, C., 337John, O. P., 34, 397, 400, 457, 458, 500,

501, 504, 505, 506, 669, 780, 798Johnsen, E. L., 184Johnson, C., 36Johnson, C. N., 323, 359Johnson, D. B., 445, 450Johnson, D. M., 517Johnson, E., 563Johnson, E. J., 143, 151, 536

Author Index 819

Johnson, J. T., 396, 398Johnson, K., 198Johnson, K. J., 565, 784, 785, 791Johnson, L. M., 489, 733Johnson, L. R., 322, 355, 576Johnson, M., 224Johnson, M. K., 171, 607, 608, 609, 610Johnson, M. M. S., 551, 784Johnson, N. G., 700Johnson, S., 89Johnson, S. C., 278Johnson, S. L., 508, 697Johnson, S. M., 89, 647Johnson, T. E., 388Johnson, W., 475Johnson, W. F., 200, 201Johnson-Frey, S. H., 187Johnson-Laird, P. N., 59, 103, 104, 105,

107, 108, 109, 111, 171, 518, 574, 577,580, 591, 592, 593, 594, 628, 743, 799

Johnsrude, I. S., 256Johnstone, T., 181, 197, 199, 200, 202, 205,

252, 503, 629Joiner, T. E., 555, 677, 787, 801, 803Jonas, R., 377Jones, B., 160Jones, C., 17Jones, D. C., 328Jones, D. E., 593Jones, E., 765Jones, E. E., 89Jones, S. K., 147Jones, S., 646Jonides, J., 169, 171, 253, 257Jorgensen, M. M., 670Jose, P. E., 798Joseph, C., 409, 421, 424Josephs, I., 335Joukamaa, M., 678Judd, C. M., 490Juergen Freyberger, H., 521Julius, M., 677Junghanns, K., 523Juslin, P. N., 104, 107, 197, 199, 200, 201,

202, 203Juth, P., 716

Kaas, J., 280Kabat-Zinn, J., 93, 786Kaciroti, N., 677Kadis, J., 540Kagan, J., 104, 342, 487, 750Kahana, M., 487Kahana-Kalman, R., 350, 351, 369, 370, 731Kahlbaugh, P. E., 385Kahn, B. E., 149, 551, 560, 784Kahn, R. S., 237, 603Kahneman, D., 140, 143, 145, 152, 185,

245, 472, 473, 481, 501, 550, 563, 647,648, 777, 781, 788

Kakolewski, J. W., 251, 261Kalichman, S. C., 678Kalin, N. H., 55, 252Kalogerakos, F., 89Kalsekar, I. D., 680Kalso, E., 282Kaltenhäuser, M., 279Kaltman, S., 797, 798, 802Kalyanaraman, S., 146Kamisar, Y., 447Kan, I. P., 596Kanazawa, S., 802Kandel, E. R., 170Kane, F., 160Kanfer, R., 558, 559Kang, D. H., 667Kang, S., 523, 525Kano, M., 284, 523Kant, I., 8, 309Kaplan, E. A., 449Kaplan, J., 665Kaplan, R. C., 680Kapp, B. S., 160, 166Kappas, A., 187, 197, 199, 204, 217, 459Karbon, M., 443

Kardes, F. R., 553Karlson, P., 238, 240Karlsson, A., 716Karlsson, H., 678Karnath, H. O., 284Karp, L., 548Karrass, J., 340Kasch, K. L., 508Kash, K. M., 650Kashima, Y., 401Kashub, C., 515Kasl, S. V., 679Kass, L. R., 764Kassel, J. D., 678, 697Kasser, T., 479Kasson, J. F., 24Kastenbaum, R. J., 798, 803Kastner, S., 605Katdare, A., 242Katkin, E. S., 186, 402Katz, L. D., 568Katz, L. F., 335, 336, 343, 733Katz, R. C., 736Kauffman, M., 334Kavanagh, D. J., 798Kawachi, I., 663, 677, 678Kawakami, K., 202, 311Kaye, W. H., 700Kazen, M., 549, 555, 568Keane, J., 224, 768Keane, M. M., 610Keane, T. M., 724Keating, C. F., 205Keay, K. A., 669Keenan, K., 733Keener, A. D., 699Kehoe, P., 55Keiley, M. K., 489Keinan, A., 151Keith, A. B., 411Keller, M., 325Keller, M. C., 791, 800Keller, M. L., 654Keller, S. E., 664Kelley, B. M., 736Kellogg, S., 19Kelly, D. R., 758Kelly, J., 28Kelly, J. L., 170Kelly, J. R., 462Kelly, T., 28Kelsey, K. S., 678Kelso, J., 297Keltner, D., 80, 95, 103, 151, 206, 212, 215,

216, 217, 218, 219, 221, 222, 224, 226,253, 396, 412, 456, 458, 459, 462, 498,499, 593, 594, 657, 668, 669, 680, 682,683, 691, 693, 762, 780, 782, 789, 790,798, 799, 800, 802, 803, 804

Kemeny, M. E., 661, 662, 663, 664, 666,667, 668, 669, 670, 671, 677, 682

Kemper, C. A., 680Kemper, T. D., 32, 36, 43Kenardy, J., 678Kendall, P. C., 700Kendler, K. S., 722Keniston, K., 447Kenny, A., 13, 73Kensinger, E. A., 602, 603, 604, 605, 606,

607, 608, 609, 610, 611Kermer, D. A., 145, 152Kermoian, R., 298, 334, 732Kerns, K., 337, 338Kerr, D. C. R., 489Kerszberg, M., 171Kessel, F., 410Kessler, R. C., 694Kessler, T., 433, 434, 437, 462Ketelaar, T., 459, 478Keubli, J., 524Keverne, E. B., 54, 55Keyes, C. L. M., 782Keysers, C., 765Khan, S., 503Khantzian, E. J., 54

Khateb, A., 720Kibbler, C. C., 786Kibler, J., 119Kiecolt-Glaser, J. K., 190, 402, 665, 677,

780, 789Kieffer, B. L., 55Kiel, E. J., 732Kiesler, D. J., 94Kihlstrom, J. F., 171, 533Killcross, S., 165Killen, M., 734Killian, C., 42Killinger, W. A., Jr., 605Kilpatrick, L., 604Kilts, C. D., 170, 603, 604Kim, E., 147Kim, G., 351Kim, H., 252Kim, H. S., 187Kim, J. J., 162, 164, 166, 611Kim, K. H., 669Kim, M., 336Kimata, H., 670Kimchi, R., 554Kinder, L. S., 677Kinderman, W., 107King, L., 480, 777King, R., 445Kinsey, S. G., 666Kirby, K. N., 146Kirby, L. D., 629, 633Kirchsteiger, G., 148Kirk-Smith, M. D., 241Kirkevold, B., 205Kirkpatrick, B., 55Kirouac, G., 204, 205, 207, 336Kirsch, P., 223Kirschbaum, C., 285Kirsh, D., 410Kirsh, S., 338Kirsner, K., 199Kirson, D., 517, 581, 588Kischka, U., 253Kitayama, S., 410, 460, 479, 515, 516, 608Kivetz, R., 151, 152Kivley, L. S., 520Klassen, T., 768Kleban, M. H., 379Kleck, R. E., 224, 730Klein, D. J., 183, 185Klein, D. N., 734Klein, K. J. K., 399Klein, M. H., 94Klein, T. R., 449Kleinginna, A. M., 69, 71Kleinginna, P. R., 69, 71Kleinknecht, E. A., 767Kleinknecht, R. A., 767Kleinman, S., 35Klepac, L., 338Klerman, G. L., 187Klerman, G. R., 187Kline, J. P., 284, 677Kling, K. C., 534Klinnert, M. D., 200, 212, 225, 298, 351,

371, 459Kluger, R., 448Kluver, H., 160Knapp, M., 798Knetsch, J. L., 145Knight, D. C., 168Knight, M. R., 716Knight, R. T., 171, 253, 669Knutson, B., 56, 57, 147, 148, 505, 685, 779Kobak, R. R., 730Koch, C., 171Koch, H. E., 386Koch, M. D., 769Kochanska, G., 325, 341, 489, 490, 732,

733, 734, 747Koehler, D. J., 151Koelsch, S., 285Koff, E., 222Koga, C., 664Kogut, T., 148

820 Author Index

Koh, K. B., 664Kolb, B., 54, 62Koller, S., 409, 766Kollock, P., 42Kong, J., 258Konijn, E., 110Koopmann, B., 220Koopmans, T. C., 144Koops, W., 736Korman, L. M., 94Kosfeld, M., 56Kosfelt, M., 789Koslowski, B., 309Kosman, L. A., 425Kosson, D. S., 493Koster, B. A., 22, 25Koster, E. P., 239Köszegi, B., 141Kotchemidova, C., 18Kotovsky, L., 633Kotter, R., 160Kouznetsova, N., 400Kovacs, G. L., 56Kozin, M., 602, 605Kraemer, D. T., 538Kraemer, H. C., 190Kraiger, K., 551Kramer, G. P., 549Kramer, R., 325Krantz, D. S., 645, 677Krause, R., 222, 693Krauss, R. M., 199, 399Krauth-Gruber, S., 82, 588, 595, 597Krebs, D., 116Krebs, J. R., 203, 212Kreiman, G., 171Kremen, A. M., 788Krendl, A. C., 605Kreuger, A. B., 777Kreutzer, M. A., 366Kriechel, S., 716Kring, A. M., 212, 396, 397, 402, 459, 498,

691, 693, 695, 697, 698, 699, 799Kringelbach, M. L., 259, 268Krivoshekova, Y. S., 383, 676, 679, 681, 684Kroes, R. A., 57, 58Krogh-Jespersen, S., 368Krout, K., 276Krubitzer, L., 280Krueger, R. F., 475Kruger, A. C., 336Kruger, T., 608Krumhansl, C. L., 106, 184Krumhuber, E., 459Krupnick, J., 341Krupp, D., 400Kuang, J. X., 152Kubany, E., 134Kuboki, T., 522Kubzansky, L. D., 663, 677Kuczaj, S. A., 353Kudadjie-Gyamfi, E., 679Kuhl, J., 549, 553, 554, 555, 568Kuhl, P. K., 372Kuhn, C., 665Kuiken, D., 109, 110Kuipers, P., 430, 780Kulik, J., 605, 606, 607Kumano, H., 522Kunkel, S. R., 777Kunreuther, H., 143, 151Kunzmann, U., 382, 383, 387Kuppens, P., 460Kupperbusch, C., 217, 387Kurokawa, M., 515Kurzban, R., 789Kusche, C. A., 736Kutas, M., 608

Laakso, M. L., 197LaBar, K. S., 167, 168, 171, 503, 604, 606,

607, 611, 798LaBarbera, J. D., 365Labouvie-Vief, G., 184, 380, 383, 384, 386,

387, 402, 521

Lacher, U., 667Ladd, D. R., 200, 202Lader, M., 710LaFrance, M., 217, 399Lagattuta, K. H., 321, 322, 325, 750Laibson, D. I., 146, 152Laidlaw, T. M., 670Laird, A. R., 255Laird, J., 336Lakatos, K., 487Lake, N., 329Lakoff, G., 592, 594, 595Lam, W. W. T., 679Lamb, A., 765Lambert, N. K., 670Lambert, R. B., 670Lamberts, K., 588Lambie, J. A., 51, 60, 75, 83Lambrecht, L., 299, 311Lamoreaux, R. R., 165Lancaster, J. L., 255Landerman, R., 477Landis, C., 71Lane, R. D., 90, 228, 257, 269, 514, 516,

517, 518, 519, 520, 521, 522, 604Lane, S., 329Lang, E., 279Lang, F. R., 380Lang, P. J., 73, 181, 184, 185, 187, 203,

593, 602, 605, 691, 693, 699, 719, 722,723, 724, 725

Langston, C. A., 500Lanius, R. A., 698Lansford, R., 237Lansky, M. R., 744, 749Large, R. G., 670Larkin, G., 787Larsen, J. T., 180, 183, 185, 187, 189, 472,

516Larsen, K., 678, 680Larsen, R. J., 387, 397, 472, 478, 589Larsen, R. W., 334, 344Larsen, S. F., 110Larson, C. L., 503, 665Larson, D. B., 476, 790Larson, J., 400, 541Larson, R., 477Larson, R. W., 401Larsson, G., 679Lassegard, M., 462Lau, R. R., 645, 651, 652Laub, D., 450Laukka, P., 107, 197, 199, 200, 201, 202,

203Laurenceau, J. P., 459Lauver, D. R., 649Lavenu, I., 386Laver, S., 102Lavie, N., 171Lavond, D. G., 611Lawler, E. J., 42Lawrence, A. D., 143, 250, 272Lawrence, K., 223Lawson, J., 521Lawson, M. E., 447Lawton, M. P., 378, 379, 382, 386Lawton, W. J., 680Lazare, A., 798Lazarus, B., 124Lazarus, R. S., 71, 78, 79, 82, 124, 219,

304, 305, 307, 414, 421, 437, 456, 497,498, 502, 508, 537, 574, 577, 602, 628,634, 638, 652, 661, 664, 778, 780, 781,797, 798, 799, 800

Lazeyras, F., 720Le, H. N., 695Leach, C. W., 435, 457, 463Leach, E., 762Leahy, J., 151Leahy, R. L., 90Leary, M. R., 456, 667, 668Leavitt, L. A., 366LeBel, J. L., 678Lebert, F., 386Lebo, K., 515

Leboyer, M., 56Lebron, K., 164Lederhendler, I., 55LeDoux, J. E., 53, 71, 79, 119, 124, 129,

160, 161, 162, 163, 164, 165, 166, 167,168, 171, 181, 252, 576, 578, 602, 611,631, 633, 634, 635, 636, 655, 664, 713,718, 719, 720, 725, 798

Lee, C. S., 105Lee, J., 163Lee, J. J., 477Lee, M., 226, 227Lee, P. C., 782, 786Lee, R. R., 279Lee, T. C., 769Lee, Y., 515, 664Lee, Y. L., 252, 268Lee-Chai, A., 507Leeper, R. W., 186, 535Leerkes, E. M., 492Legerstee, M., 352Leggett, E. L., 746Lehman, C. L., 724Lehman, D. R., 515, 798Lehtinen, V., 522, 678Leifer, D., 279Leighton, E. A., 386Leiguarda, R., 284Leinonen, L., 197, 199, 200, 201Leites, E., 20, 22LeJeune, J. T., 798Lelwica, M., 350, 364, 365, 732Lemerise, E. A., 731, 735, 736Lemery, C. R., 458Lemu, B. A., 425Lennon, E. M., 367, 368Lennon, R., 397Lennox, R., 524, 630Lensing, P., 56Lenton, A. P., 76Lenz, F. A., 279Lenzi, G. L., 441Lepage, M., 604Lepper, H. S., 680Lepper, M. R., 561Leppert, J., 679Lerdahl, F., 105Lerman, C., 647, 648, 650, 656Lerner, J. S., 147, 151, 190, 216, 220, 561,

562, 563, 594, 657Lerner, N., 539Lerner, R. M., 334LeRoux, J. A., 225Leslie, K. R., 187Leu, J., 516, 685Levenson, R. W., 77, 96, 126, 181, 183, 185,

216, 217, 219, 220, 224, 226, 310, 379,382, 385, 387, 399, 402, 456, 458, 459,498, 500, 504, 505, 507, 536, 662, 670,677, 681, 693, 759, 780, 781, 788, 789,798, 799, 804

Leventhal, E. A., 654, 655, 678, 682Leventhal, H., 645, 646, 647, 652, 654, 655,

676, 678, 682Levesque, J., 169, 504Levin, P. F., 152, 558Levine, D. K., 150Levine, J., 487Levine, L., 574, 575, 576, 580, 582, 583Levine, L. J., 180, 355, 383, 414, 576, 607LeVine, R. A., 410Levine, S., 311Levy, B. R., 777, 788Levy, R. I., 422Levy Paluck, E., 399Lewicka, M., 356Lewis, H. B., 39, 747, 748, 749Lewis, J., 24Lewis, J. G., 665Lewis, M., 70, 75, 77, 171, 226, 298, 300,

304, 305, 306, 307, 308, 309, 310, 311,312, 313, 314, 315, 316, 317, 335, 336,337, 351, 414, 445, 486, 487, 578, 625,668, 730, 731, 742, 743, 745, 746, 747,748, 750, 751, 752, 753

Author Index 821

Lewkowicz, D. J., 367, 371Lewontin, R. C., 81Leyden, J. J., 241Li, X. F., 162, 163Libero, D. Z., 387Liberzon, I., 250, 252, 255, 259, 269, 284,

503, 666Lichtenstein, S., 151Lichtman, R. R., 479Lickliter, R., 367, 369, 371Liddell, B. J., 720Liebal, K., 217Lieberman, D., 119, 121, 130, 134Lieberman, M. D., 147, 282Lieberman, P., 198, 200Liénard, P., 128Lietaer, G., 88Light, K. C., 398Lin, E., 678Lin, M. H., 325Lindberg, C., 222Lindberg, N. M., 650Lindenberg, S., 492Linder, D. E., 241Lindholm, T., 522Lindquist, K., 292, 630Lindsay, D. S., 608Lindsey, L. L. M., 683Lindsley, D. B., 188Lindström, E., 104Lineberger, M., 799Linehan, M. M., 91, 92, 93, 698Ling, S., 170, 720Linnankoski, I., 197Linton, M., 602Linton, S., 147, 431, 790Liotti, M., 54Lipkus, I., 654Lipp, O. V., 716Lipps, T., 110, 441Lipworth, L., 93Little, T. D., 382Littman, R. A., 737Littrell, J., 92Litz, B. T., 803Liu, Y., 768Liwag, M. D., 574, 575, 582Lledo, P., 237Lobel, M., 678Lochman, J. E., 736Lodge, D., 110Loeches, A., 366Loewenstein, G., 138, 139, 142, 144, 145,

146, 147, 148, 150, 151, 152, 480, 559,561, 563, 651, 655, 679, 682

Loewy, A. D., 251Lofland, L., 22Lofthouse, N., 489Loftus, E. F., 593Logan, G. D., 488Logies, S., 716Lohr, J. M., 767, 768, 769Lomax, L. E., 489, 733Lombardi, E., 680Londahl, E. A., 212, 226Long, L., 370Longuet-Higgins, H. C., 105Loomes, G., 141Lopes, P. N., 458, 539, 540Lopez, N. L., 489Lorch, M. P., 222Losada, M. F., 678, 681, 682, 779, 781, 791LoSasso, G. L., 236Losch, M. E., 187Lotze, M., 603Lounsbury, K., 337Lourenço, O., 325Lovaglia, M. J., 35Lovallo, W. R., 670Love, L. R., 399Lover, A., 342, 734Lovett, M. C., 525Lovick, T. A., 250Lowery, L., 219, 227, 683, 759, 763Lu, J., 678

Lucas, N., 242, 244Lucas, R. E., 396, 473, 474, 475, 476, 477,

478, 480Ludemann, P., 366Lukas, S. E., 54Lumer, E., 171Luminet, O., 521, 522, 523, 596Lumley, M. A., 184, 383, 402, 521, 522,

536, 769Lundh, L. G., 698Lundh, L., 522Lundin, T., 798Lundquist, L., 397, 399Lundqvist, D., 713, 716Lunquist, L. O., 458, 460Lupien, S. J., 166Luscher, M., 238, 240Lustick, L. S., 252Luterek, J. A., 698Lutz, C., 76, 323, 414, 420Luu, P., 257Lydiard, R. B., 699Lykken, D., 475Lyons, D., 10Lyons-Ruth, K., 733Lystra, K., 24Lyubomirsky, S. L., 541, 777, 781, 789, 800

Macaulay, D., 594Maccoby, E. E., 489MacDonald, B. A., 769MacDonald, C., 327Macdonald, N. H., 199MacDonald, T. K., 151Macefield, V. G., 278MacGregor, D. G., 561, 679MacIver, D., 735MacKay, D. G., 608, 609, 610Macker, C. E., 152Mackie, D. M., 430, 431, 433, 434, 435,

436, 437, 549, 565MacLean, D. J., 368MacLean, P. D., 50, 53, 159MacLeod, C., 716, 717, 723, 724MacNamara, A., 549, 552Macphail, E. M., 283MacPherson, S. E., 386Macrae, C. N., 257Madden-Derdich, D., 336Maddison, S., 164Maddock, R. J., 258, 268Madon, S., 399Madson, L., 395Maes, M., 667Magai, C., 297, 377, 378, 381, 382, 383,

384, 385, 387, 536, 676, 677, 679, 680,681, 682, 684, 685

Magnus, P., 475Magnusson, D., 285Mahapatra, M., 409, 421, 763Mahoney, M., 94Mahr, R. N., 236Maia, T. V., 143Maier, S. F., 666Maietti, A., 253Maihöfner, C., 279Main, M., 309Maitner, A. T., 434, 435, 437Majidishad, P., 164, 252Malamut, B., 165Malarkey, W. B., 190Malatesta, C., 292, 298, 372, 377, 378, 385,

731, 732Malatesta-Magai, C., 377, 383, 385Malcarne, V. L., 676Malmstadt, J. R., 503Malti, T., 325Mancill, R. B., 724Mancini, F., 59, 111, 768Mancuso, R. A., 95, 788Mandel, M. R., 187Mandell, R., 766Mandler, G., 76, 182, 185, 307, 514, 576,

628, 745Manes, F., 224, 284, 768

Mano, H., 553Mansell, W., 694Manstead, A. S. R., 327, 396, 397, 400, 401,

402, 431, 457, 458, 459, 460, 463Mantani, T., 258, 521, 522Mantel, S. P., 566Manuck, S. B., 402Mar, A., 54Mar, R., 111Maratos, E. J., 604Marcel, A. J., 51, 60Marcel, A., 75, 83Marcus, R., 62Maren, S., 160, 162, 164, 252, 611Margarinos, A. M., 165, 166Maril, A., 257Mariske, M., 380Markham, R., 358Markides, K. S., 677, 777Markman, E. M., 349Markowitsch, H. J., 171, 603Markowitz, H. M., 140, 141Marks, G., 680Marks, I., 119, 120, 121, 135, 710Markus, H. R., 410, 460, 479, 515Markwith, M., 762, 763Marlatt, G. A., 678Marler, P., 367, 368Marmara, V., 765Marriott, M., 762Marsh, A. A., 224, 730Martel, F. L., 54Martin, A., 257, 611Martin, C. L., 343, 446, 734Martin, G. B., 441, 444Martin, J. H., 256Martin, L. L., 186, 397, 474, 505, 542, 784Martin, L. M., 562Martin, S. E., 487Martineau, W. H., 786Martínez-Sánchez, F., 522Martins, Y., 760Martzke, J. S., 219Marucha, P. T., 190Maruff, P., 610Marx, K., 32Marzillier, S. L., 769Masataka, N., 104Mascagni, F., 161, 162, 169Maser, J. D., 709Maslow, A. H., 478Mason, J. W., 664Mason, M. D., 679Mason, M. F., 282Masson, J. L., 411, 412, 759Mastekaasa, A., 476Masters, J. C., 322, 786Matas, L., 786Matchett, G., 760, 768Mather, M., 386, 602, 610, 716Mathews, A., 716, 717Mathias, C. J., 171Mathieu-Coughlan, P., 94Mathur, S., 786Matias, R., 293Matsumoto, D., 212, 214, 216, 217, 219,

220, 221, 222, 223, 225, 226, 227, 400,401

Matt, G. E., 602Matt, S. J., 25Mattay, V. S., 223Matthews, G., 399, 543Matthews, S. C., 273Matus-Amat, P., 164Matzner, W. T., 789Maurer, M., 537Mauss, I. B., 181, 190, 216, 217, 218, 219,

498, 507, 508, 509, 693, 798Mauss, M., 78, 760, 761May, C. P., 386Mayberg, H. S., 258, 269Mayer, J. D., 90, 225, 399, 464, 523, 533,

534, 535, 536, 537, 538, 539, 541, 543,589, 594

Mayer, K. U., 382

822 Author Index

Mayes, A. R., 610Mayes, L. C., 719Mayhew, E., 368Mayhew, G., 766Maynard Smith, J., 128Mayne, T. J., 676, 678, 680, 681, 682Mayr, E., 764Mayr, U., 382, 515Mazziotta, J. C., 255McBride, A., 541McCall, G. J., 38McCanne, T. R., 505McCarter, L., 181, 217, 498, 693, 798McCarthy, G., 798McCarthy, K., 520McCaul, K. D., 649McCauley, C. R., 757, 758, 761, 762, 763,

767, 770McCeney, M. K., 677McClain, T. M., 400McClary, S., 107McClaskey, C. L., 736McClellan, W. M., 477McClelland, D. C., 670McClelland, G. H., 490McClelland, J. L., 143McClintic, S., 310, 743McClintock, M. K., 240, 241McClure, E. B., 399McClure, S. M., 146, 147, 152McCluskey-Fawcett, K. A., 782McComas, H. C., 205McCormick, L. J., 594McCourt, M. E., 634McCrae, R. R., 382, 387, 478McCullough, M. E., 789, 790McCutchan, A. J., 680McDade-Montez, E. A., 680McDaniel, P. A., 22McDermott, K. B., 556, 608McDonald, A. J., 161, 166, 169, 250McDonald, P. G., 676McDonald, P. W., 521, 522McDonald, R. J., 164McDougall, W., 72, 77McDowell, C. L., 386McDowell, D. J., 336, 339McEwen, B. S., 165, 166, 663, 788, 789McFadden, S. H., 377, 378McFadyen-Ketchum, S., 490McFarland, N. R., 280McGarty, C., 463McGaugh, J. L., 160, 165, 166, 170, 171,

602, 603, 604, 605, 606McGhee, P. E., 219McGinnes, G. C., 293McGonigal, K. M., 505McGraw, A. P., 180, 472, 516, 651, 682McGraw, K., 134McGregor, B. A., 648McGue, M., 475McGuire, L., 677, 780McGuire, T., 242McIntosh, D. N., 183, 184, 185McIntosh, W. D., 542McKay, D., 768McKeen, N. A., 679McKenna, C., 789McKenna, M., 721McKinley, M., 327McKnew, D., 341McLean, C., 768McLeod, J. D., 492McNally, R. J., 602, 664, 712McNaughton, S., 336McNeil, D. W., 724McPherson, R., 383McRae, K., 257, 504, 509McTeague, L. M., 724Mead, G. H., 32Meaney, M. J., 54Medford, N., 602, 603Medler, M., 380Mednick, E. V., 552Mednick, M. T., 552

Mednick, S. A., 552Meerum Terwogt, M., 322, 357, 360, 521Mehrabian, A., 589Meier, B. P., 636Meigs, A. S., 760Meischke, H., 679Melamed, B. G., 724, 769Melia, K. R., 162Mellers, B. A., 141, 516, 562, 651, 682Melton, R. J., 549, 557, 566Meltzer, D., 679Meltzoff, A. N., 82, 367, 372Mendelson, T., 240Mennin, D. S., 698, 701Mennuti-Washburn, J., 518Menon, U., 413, 419, 420, 421Mentzel, H. J., 719Mercado, A. M., 190Merckelbach, H., 711Meredith, M. A., 238, 368Mergenthaler, E., 94Merikle, P. M., 716Merk, W., 736Merker, B., 60Merleau-Ponty, M., 13Merritt, M., 225Mertz, E., 551, 784Mervis, C. B., 517Mesholam, R. I., 236Mesias, W., 378Mesquita, B., 74, 76, 80, 212, 249, 401,

410, 414, 437, 458, 498, 505, 515, 522,685, 693

Messina, J., 385Messinger, D., 298Mesulam, M. M., 253Meszaros, J., 143Metalsky, G. I., 801Metalsky, G. J., 752Metzinger, T., 70, 84Meulders, M., 460Meunier, S. A., 769Meyer, D., 645, 651, 653Meyer, L. B., 106Meyer, N., 160Miall, D. S., 109, 110Michael, T., 611Michalos, A. C., 479Michalson, L., 171, 305, 309, 314, 315,

316, 337, 414Michel, G. F., 299, 311, 731Michie, S., 679Mickey, M. R., 752Miczek, K. A., 666Middleton, F. A., 258, 259Middleton, R. A., 677Mihalecz, M. C., 798Mikels, J. A., 184, 386Mikulincer, M., 786Milad, M. R., 164, 168, 253Milam, J. E., 680Miliaressis, E., 160Mill, J. S., 447Millar, N., 717Miller, D. A., 435, 436Miller, E. K., 171Miller, G., 104Miller, G. A., 398, 577Miller, G. E., 663, 667, 670Miller, J. G., 417, 421Miller, K., 410Miller, N., 398, 401, 646Miller, N. E., 185Miller, P., 337, 410Miller, P. A., 338, 443Miller, P. J., 341, 518Miller, R. S., 459Miller, S., 130Miller, S. B., 758, 761, 770Miller, S. M., 645, 647, 648, 656, 676, 679,

680, 682Miller, T. Q., 665Miller, W. I., 758, 762, 766, 770Milligan, M. J., 34Millman, L., 760, 761

Millot, J. L., 200Mills, J., 552Miltner, W. H. R., 716, 719Mineka, S., 70, 119, 128, 167, 212, 244,

694, 696, 697, 711, 712, 715, 716, 725Miniard, P. W., 549Minor, J. K., 596Minoshima, S., 258Minsky, M., 619, 621, 627Mintz, J., 410Mintz, S., 19Mischel, W., 146, 497, 503, 533, 587, 630Miserendino, M. J., 162Mishkin, M., 160, 165Miskawayh, A. M., 424Missirlian, T., 95Mistlin, A. J., 368Mitchell, I. J., 768Mitchell, J. P., 257Mitchell-Copeland, J., 337, 734Mithen, S., 102Mittwoch-Jaffe, T., 670Mitzman, A., 21Miyake, K., 79, 298Moav, O., 801Moberg, P. J., 236Moffitt, T. E., 221, 800Mogg, K., 716, 717Mohiuddin, S., 611Molden, D. C., 506Moles, A., 55Moll, J., 768Mondillon, L., 596Monnier, J., 699Montague, D. P. F., 365, 370, 731Montague, R. P., 655Monte, W. C., 241Montebarocci, O., 222Monteiro, K. P., 594Monti-Bloch, L., 241Moore, B. F., 28Moore, M. K., 372Moore, R., 679Moors, A., 83, 632Mora, P. A., 678, 682Morelli, G., 79Moreno, A. J., 370Moreno, C., 385, 386Moreno, K. N., 549Morewedge, C. K., 142Morgan, H. J., 634Morgan, M. A., 164Morgan, M. J., 786Morgenstern, A. H., 679Mori, E., 603Moriarty, A., 333Morimoto, H. K., 677Morral, A. R., 91Morrell, J. M. B., 328, 359Morrell, M. J., 47, 253Morris, A. S., 501Morris, J. S., 167, 170, 223, 604, 719, 720Morrison, A. P., 744, 747, 749Morrison, J. H., 165Morrison, S. E., 252Morrow, J., 503, 541, 542, 800Morse, P. A., 366Moscovitch, M., 606, 608, 609Moses, E. B., 508, 696, 701Moses, L. J., 351, 352Moskal, J. R., 57, 58Moskowitz, D. S., 678Moskowitz, J. T., 677, 678, 681, 777, 787,

788, 791, 802Mosley, J., 337Moss, M. S., 379Mossey, J. M., 477Mostow, A. J., 736Mowat, R. R., 122Mower, G., 274Mowrer, O. H., 78, 165Moyer, D., 503Mroczek, D. K., 382, 477Much, N. C., 409, 763Muchembled, R., 20

Author Index 823

Muderrisoglu, S., 400Mueller, C. M., 749Mueller, J. S., 549Mueller, N. K., 669Mufson, E. J., 253Muir, D. W., 350, 370Mulford, M., 149Mullen, E., 435Muller, J., 162Muller, M. J., 679Muller, N. G., 171Mullett, J., 458Mumme, D. L., 334, 352, 370Munn, P., 321, 337, 353Muñoz, R. F., 501, 508, 698Murdaya, U., 517Murphy, B., 343Murphy, B. C., 731, 734, 736Murphy, F. C., 250, 260, 272, 284Murphy, G. L., 517Murphy, L., 333Murray, D., 124Murray, E. A., 171, 252Murray, H. A., 478Murray, K. T., 489Murray, N., 551Musick, B. S., 678Mussweiler, T., 549Myers, C., 610Myers, D. G., 471Myers, J. M., 722Myers, R. S., 313, 366Myers, W. A., 258Myhrer, T., 477Mystkowski, J. L., 694

Nabi, R. L., 762Naccache, L., 171Nadel, L., 164Nader, K., 165, 252Nagel, L., 293, 731Nagele, T., 284Nagin, D., 142Nakao, M., 522Nakash-Eisikovits, O., 400Naqvi, N. H., 53, 189, 284Nash, A., 444Nath, L. E., 397, 398Nathan, P. E., 701Naumann, E., 188Navarette, C. D., 767, 770Navon, D., 554Neale, J. M., 670, 693, 695Neale, M. C., 475, 722Nearing, K. I., 167Needham, A., 633Neimeyer, R. A., 94, 799Neisser, U., 367, 605, 607Nelson, B., 27Nelson, C. A., 307, 366, 371Nelson, E., 55, 56, 477Nelson, K., 322, 337Nelson, N. L., 353, 354, 356, 358, 361Nelson-Goens, G., 298Nemeroff, C., 758, 759, 760, 761, 765Nemeroff, C. B., 54, 285Nemeroff, C. J., 760, 761, 765Nes, R. B., 475Nesin, A. E., 735Nesse, R. M., 116, 117, 129, 508, 786, 791,

800Nesselroade, J. R., 382, 515Nester, M. A., 566Neta, M., 636Neu, J., 12Neuberg, S. L., 436, 437, 444, 770Neugut, A. I., 676, 679Neumann, R., 187, 203, 434, 629, 633Neumann, S. A., 402Neundörfer, B., 279Neville, E. A., 768Nevison, C. M., 54Newell, A., 622Newman, J. D., 55Newman, J. P., 493

Newman, J. R., 400Nezlek, J. B., 220, 516Nicholas, M., 222Nicholls, M. E., 596Nichols, K. E., 325Nichols, M. P., 92Nichols, T. E., 257Nicolich, M., 377Nicolson, N. A., 804Niedenthal, P. M., 82, 551, 588, 593, 594,

595, 596, 597Niedzwienska, A., 398Nielson, K. A., 521, 522Nienaber, N. A., 678Niesink, R. J., 56Nietzsche, F., 8, 9Nieuwenhuis, J. A., 89Nigg, J. T., 487Niit, T., 356Niles, P., 646Nimchinsky, E. A., 253Nimmo-Smith, I., 250, 272Nir, A., 337Nisbett, R. E., 410, 515Nishijo, H., 160Nocjar, C., 57Noice, H., 549Nolen-Hoeksema, S., 400, 503, 541, 542,

800Noller, P., 336Norcross, J. C., 98Norli, Ø., 143Norman, D. A., 629, 637Norman, R. Z., 35Normansell, L. A., 50, 55, 56, 57Norris, C. J., 187Northoff, G., 50, 62Norton, P. J., 694Notarius, C. I., 536Novemsky, N., 146Novotny, C. M., 700Nowicki, G. P., 536, 549, 552, 553, 784Nowicki, S. J., 228Nowlis, S., 146Nundy, S., 110Nunes, S. O. V., 677Nunley, E. P., 798Nusbaum, B., 377, 378Nusbaum, H., 410Nussbaum, M. C., 5, 12, 69, 71, 102, 104,

111, 766Nwokah, E. E., 205Nyengaard, J. R., 263Nygren, T. E., 149, 552, 563, 564Nyman, M., 731Nystrom, L. E., 147, 148, 763

Oakes, P. J., 428, 429, 461Oatley, K., 71, 74, 75, 77, 78, 103, 104,

105, 108, 109, 110, 111, 457, 460, 465,518, 574, 577, 580, 591, 592, 593, 594,628, 731, 743, 778, 780, 797, 799

Ober, C., 240O’Boyle, C. G., 503O’Brien, J., 606Ochsner, K. N., 169, 249, 257, 258, 400,

403, 498, 499, 503, 504, 515, 605, 606,608, 630, 637, 669, 693

O’Connor, C., 517, 581, 588O’Connor, F., 109O’Doherty, J. P., 170, 171O’Donoghue, T., 145, 150Ogan, T. A., 732Ogawa, S., 263Ogden, J., 682Öhman, A., 70, 73, 76, 78, 167, 171, 183,

212, 224, 244, 245, 282, 608, 711, 712,713, 714, 715, 716, 718, 719, 720, 722,723, 725

Oishi, S., 473, 474, 476, 479, 515, 516Okada, G., 258, 521Okamoto, Y., 258O’Keefe, J., 164Okun, M. A., 477Olatunji, B. O., 758, 767, 768

Oldehinkel, A. J., 492Olds, J., 160Oliner, P. M., 447Oliner, S. P., 447Oliveira, M. G., 603Olson, C. R., 253Olson, S. L., 489Olsson, A., 168Olthof, T., 322, 357Oltmanns, T. F., 222, 695O’Neil, R., 336O’Neill, H. K., 769Ong, A. D., 384, 516, 525, 788, 804, 805Ongur, D., 250, 256, 278, 282Ono, T., 160Oosterveld, P., 521Oppenheim, D., 337Or, A., 679Ormel, J., 492Orobio de Castro, B., 736Orr, S. P., 119, 724Ortiz-Soria, B., 522Ortner, S. B., 762Ortony, A., 71, 76, 77, 180, 307, 308, 312,

355, 518, 537, 580, 588, 593, 594, 602,629, 630, 631, 633, 634, 636, 637, 638,639, 744

Orwoll, L., 380Osgood, C. E., 180, 589, 635Oshinsky, J. S., 202Öst, L. G., 711Oster, H., 227, 293, 295, 298, 299, 301,

311, 731, 765Osterweis, M., 798Ostir, G. V., 677, 777, 788Ostrander, M. M., 669Ostroff, J. S., 679O’Sullivan, M., 212, 217, 225Otake, K., 783, 785Ottaway, C. A., 663Ottenberg, P., 765Ottersen, O. P., 164Otto, N. J., 57Overskeid, G., 799, 800Owen, A. M., 256Owen, D., 351Owens, E. B., 733Owren, M. J., 197, 198, 199, 200, 203, 204,

205, 206, 207, 351Ozonoff, S., 222, 323

Packard, M. G., 165Padgett, D. A., 666, 669Paepke, A. J., 240Paeschke, A., 200Pagano, F., 646Page, A. C., 769Paglia, C., 109Pain, C., 698Paivio, S. C., 89, 91Pakosz, M., 197, 201, 202, 203Palermo, D. S., 180Palfai, T. P., 90, 536, 538Paller, K. A., 609Palmer, S. E., 554Palomba, D., 253Panitz, D., 719Panksepp, J., 48, 49, 50, 51, 52, 53, 54, 55,

56, 57, 58, 59, 60, 61, 62, 63, 104, 250,513, 782, 798

Pantelis, C., 666Papa, A., 802Papageorgis, D., 536Papez, J. W., 159, 160Papini, D. R., 782Papousek, H., 372Papousek, M., 372Paradis, C. M., 605Paradiso, S., 284Pardo, J. V., 237Pare, D., 163Pare-Blagoev, E. J., 257Parent, M. B., 165Paris, F. A., 356, 357, 358Parish, S., 413, 419

824 Author Index

Parisi, S. A., 365Park, J. H., 770Park, L., 409, 763Park, N., 789Parke, R. D., 327, 336, 337, 338, 339, 730,

733, 734, 735Parker, J. D. A., 521, 522, 523, 698Parker, J. G., 733, 735Parkhurst, J., 343Parkinson, B., 457, 461, 465, 631Parkinson, J. A., 670Paronis, C. A., 678, 697Parr, L. A., 217, 223Parra, C., 714, 715Parrott, W. G., 73, 76, 81, 457, 465, 498Parry, L. A., 605, 606Parry, M. H., 310Parsons, L. M., 255Parten, S., 367, 368Partridge, R. T., 521Pascoe, J. P., 160Pascual-Leone, A., 88Pascual-Leone, J., 94, 96Pasley, B. N., 719, 721Pasqualini, M. S., 217Pasquier, F., 386Pasternak, T., 171Pasupathi, M., 382, 445, 515Paternoster, R., 142Paton, J. J., 252, 260, 261Patrick, C. J., 222, 694, 699Patrick, R., 563Patrick-Miller, L., 645, 676Patterson, C. M., 493Patterson, G. R., 491, 493, 494, 737Patwardhan, M. V., 110, 411, 412, 759Pauley, J. D., 802Pauls, C. A., 184Paulus, M. P., 273, 283, 285Pavlicevic, M., 106, 107Pavlov, I. P., 164Pavlov, S. V., 521Pavot, W., 472, 478Pearlman, L. A., 443Pecher, D., 596Peciña, S., 51, 61, 778Pedersen, J., 444Peeke, H., 240Pegna, A., 720Pek, J., 787Pell, M. D., 199, 200Pelli, D. G., 164Pellis, S. M., 56Pellis, V. C., 56Pendergrass, J. C., 603Pendleton, B. F., 476Peng, K., 515Pennebaker, J. W., 93, 329, 383, 477, 541,

542, 678, 801Pennington, B. F., 222, 323Penza, S., 735Pereira, C., 200, 202Perepeluk, D., 92Peretz, I., 104, 107, 180Perez, M., 787Perls, F. S., 88Perner, J., 324Perrett, D. I., 223, 224, 368Perry, P., 537Perry, R., 171Perry-Jenkins, M., 401Persad, S. M., 222Persons, J. B., 694, 701Pessoa, L., 605, 611, 721Peters, D., 583Peters, E., 561, 654, 655, 657, 679Peters, M. L., 767Petersen, S., 679Peterson, B. L., 731Peterson, C., 336, 752, 788Peterson, G., 34Peterson, J. M., 447Peterson, S. E., 716Petersson, K. M., 282Petit, H., 386

Petó, E., 765Petreanu, L. T., 237Petrides, K. V., 539Petrie, K. J., 680Petrovic, P., 282Petry, M. C., 602Pettigrew, T. F., 436Pettit, G. S., 486, 489, 490, 492, 493, 733,

736Pettit, J. W., 677, 787Petty, R. E., 181, 187, 219, 552, 559, 594,

633Petzold, P., 594Pezdek, K., 605Pfennig, J., 387Pfister, J., 23Phaf, R. H., 163Phan, K. L., 250, 252, 255, 260, 284, 610,

664, 666Phelps, E. A., 167, 168, 169, 170, 171, 259,

498, 597, 603, 605, 606, 638, 664, 718,720, 798

Philippot, P., 187, 272, 285, 401, 457, 458,588, 605, 606

Phillips, A. T., 352Phillips, E. S., 768Phillips, J. G., 666Phillips, K. L., 28Phillips, K., 240, 244Phillips, L. H., 386Phillips, M. L., 223, 225, 281, 768Phillips, R. G., 164Piaget, J., 448, 583, 765Pick, A. D., 351, 369Pickens, J., 367, 731Pickersgill, M. J., 711Pierce, R. A., 92Pietromonaco, P. R., 398, 458, 459, 516Pietrowsky, R., 237Pika, S., 217Pike, K. M., 399Pillsworth, E. G., 151Pillutla, M. M., 148Pincus, D. I., 60Pinel, E. C., 502Pinker, S., 104, 116, 243, 758Pinquart, M., 477Pirraglia, P. A., 679Pitkänen, A., 160, 163, 164, 250, 603Pitman, R., 119Pittam, J., 198, 200Pizarro, D. A., 151Pizzagalli, D. A., 260Plant, E. A., 396Plato, 3, 4Pliner, P., 760, 765Plomin, R., 441Plotsky, P. M., 278Plutchik, R., 73, 77, 308, 421, 743, 744, 758Poehlmann, K. M., 183Poherecky, L. A., 81Poldrack, R. A., 257, 263Polefrone, J. M., 402Polivy, J., 222Pollack, S. D., 736Pollak, L. H., 34Pollak, S. D., 222, 306, 371Pollermann, B. Z., 514, 518, 519Pollio, H., 337Pollner, M., 476Pollock, L. A., 21Poloma, M. M., 476Polonsky, W., 670Polti, G., 632Pomerantz, E. M., 549, 552, 557, 784Pons, F., 326, 328, 349, 359, 360, 518, 521Porges, S. W., 181Porter, R. H., 235, 240Pos, A. E., 94Posada, G. E., 370Posner, M. I., 171, 257, 486, 487, 596, 713,

716, 717, 718Posner, R. A., 449Posthuma, D., 475Pott, M., 79

Potter, P. T., 516, 804Potts, R., 410Poulin, R., 336Pourtois, G., 521Povel, D. J., 105Power, M., 768Pozluszny, D. M., 677Pratto, F., 608Prelec, D., 144, 147Prescott, C. A., 722Pressman, S. D., 670, 677, 678, 777, 788Preti, G., 241Preuss, T. M., 169Price, D. D., 278Price, J. L., 164, 250, 256, 278, 282, 603,

664, 669Prigerson, H. G., 799Prince, L. M., 228Prins, B., 171Prinz, J. J., 12, 13, 14, 629Prizmic-Larsen, Z., 334Prkachin, K. M., 222, 521, 522, 536, 665Prochaska, T. R., 654Procyk, E., 171Protopapas, A., 200Proust, M., 238, 239, 241Provine, R. R., 57, 206, 458Przuntek, H., 767Putallaz, M., 734Putnam, K. M., 665, 697Putnam, P., 387Putzke, J. D., 474Puukka, P., 522Pyszczynski, T., 761

Quan, N., 669Quas, J. A., 402, 584Quierdo, J. G., 364Quigley, J. F., 767, 768Quigley, K. S., 181, 183, 189, 272Quinlan, D. M., 519Quirk, G. J., 162, 164, 166, 168, 252, 253

Rabie, T., 758Rabin, M., 141, 145, 148Rachlin, H., 146Rachman, S. J., 583, 683Radke-Yarrow, M., 444, 445, 732, 733Radley, J. J., 165Rafaeli, E., 716, 805Raghavan, G., 695Rahhat, T., 386Raikes, H. A., 338Raineri, A., 146Rainville, P., 189, 190, 278, 281Rajagopal, D., 379Rakowski, W., 680Ram, N., 798Ramel, W., 504Ramsay, D., 300, 306, 311, 668, 730, 731,

745, 750, 751Ranganath, C., 171Rangel, A., 150Rapee, R. M., 710Raphael, B., 798Rapport, L. J., 236Rapson, R. L., 458Ratner, H. H., 336Rauch, S. L., 169, 719Raue, P. J., 94Rausch, M., 767Ray, R. D., 503Ray, W. J., 181Raye, C. L., 171, 607, 609Raymond, E. S., 123Raymond, J. E., 170Read, D., 146Reading, S., 253, 255, 257Ready, R. E., 525Reb, J., 684Rebucal, K. A., 386Recchia, S., 310, 743Redelmeier, D. A., 501Reder, L. M., 525Redican, W. K., 214

Author Index 825

Reed, A., 222, 306Reed, M. A., 96, 478Reed, S. D., 186Rees, G., 171Reeve, J., 551, 557, 558, 560, 566Regan, M. M., 679Reich, J. W., 516, 804Reichborn-Kjennerud, T., 475Reiche, E. M. V., 677Reicher, S. D., 428, 461, 462Reid, J. B., 491Reidel, R., 228Reiman, E. M., 259, 269, 278, 279Reinecke, A., 716Reis, D. J., 161, 163Reis, H. T., 459, 785Reisberg, D., 583, 611Reise, S. P., 521Reisenzein, R., 182, 631, 632, 633Reiser, M., 735Reissland, N., 334, 731Reite, M., 802Reith, G., 536Rendall, D., 197, 203, 205, 207, 351Renner, H. D., 758Rentfrow, P. J., 39Repacholi, B. M., 352, 733Rescorla, R. A., 164, 165Resnick, G., 338Retzinger, S. M., 40, 44Reuda, M. R., 486Reva, N. V., 521Revelle, W., 629Reynolds, C. A., 382Reynolds, S. M., 260Ric, F., 82, 588, 595, 597Rich, W. J., 451Richards, J. M., 34, 500, 504, 505Richards, J. S., 474Richards, M. H., 401Richards, T. A., 787Richardson, J. L., 680Richardson, M. P., 170, 603, 604, 720Richman, L. S., 677Richters, J. E., 693Rick, S., 147Rickards, H., 768Ricoeur, P., 9Ridge, B., 490Ridgeway, C. L., 36, 37Ridgeway, D., 337, 353Ridker, P. M., 665Riecker, A., 283, 285Riedel, R., 521Rieffe, C., 360, 521, 523, 525Riemann, B., 602Rifai, N., 665Riggio, R. E., 228Riis, J., 473Rikke, B. A., 388Riksen-Walraven, J. M., 342Rilling, J. K., 147, 148, 763Rimé, B., 75, 401, 458, 521, 588Rinck, M., 716Rinne, T., 699Ris, M. D., 784Risen, J. L., 142Risinger, R. C., 603Risser, D., 293Ritov, I., 141, 148Rivers, S. E., 539Rivest, J., 284Rizvi, T. A., 250Rizzolatti, G., 765Roach, P., 207Robbins, T. W., 160, 164, 165, 252Roberts, E., 679Roberts, K. C., 282Roberts, N. A., 253, 694Roberts, R., 12Roberts, R. C., 425Roberts, R. D., 399, 543Roberts, W., 734Roberts, W. L., 337Robin, L., 398, 458, 516

Robins, R. W., 226, 742, 749, 750Robinson, D. T., 39Robinson, G. F., 551, 784Robinson, J., 102Robinson, J. L., 441Robinson, K. J., 556Robinson, M. D., 81, 396, 398, 516, 525,

634, 636Robinson, R. G., 188, 284Robinson, T. E., 52, 253, 778, 789Robitaille, C., 678Robles, T. F., 677, 780Rochat, P., 364Rodin, J., 537Rodrigues, S. M., 162Rodriguez, M. L., 497Rodriguez Mosquera, P. M., 397, 457Roedema, T. M., 521Roediger, H. L., III, 556, 608Roehm, M., 549Rogan, K., 91Rogan, M. T., 162Rogers, C. R., 88Rogers, M. A., 666Rogers, P. L., 227Rogers, R. W., 646Rogers, S. J., 222, 323Rohde, P., 669Rollock, D., 402Rolls, E. T., 52, 78, 160, 164, 171, 237, 253,

269, 272, 274, 721, 758, 768Romanski, L. M., 161, 162, 164, 634Romeo, J. B., 549Roodenburg, H., 24Roozendaal, B., 606Rorty, A., 13Rosaldo, M. Z., 420Rosario, H. H., 242Rosch, E. H., 517, 592Rose, A. J., 337, 404Roseman, I. J., 431, 457, 460, 461, 513,

574, 577, 628, 798Rosen, A., 765Rosen, G. M., 701Rosen, J. B., 252Rosenberg, E. L., 212, 216, 217, 218, 219,

222, 223, 778Rosenberg, M., 34Rosenberg, N., 679Rosenblatt, P. C., 22, 803Rosenblum, L., 305Rosenfeld, H. M., 365Rosenhan, D. L., 152Rosenkranz, H. C., 236Rosenthal, R., 227Rosenthal, T. L., 443Rosenwein, B. H., 21, 23, 28Rosenzweig, A. S., 549Rosicky, J. G., 351Ross, H., 575Ross, L., 149, 768Ross, M., 516, 575Ross, T. J., 603Rossi, J., III, 55Rossi, N., 222Roter, D. L., 398Roth, T. L., 54Roth, W. T., 724Rothbart, M. K., 342, 485, 486, 487, 489,

490, 493, 503Rothbaum, F., 79, 492, 502Rothgerber, H., 404Rotshtein, P., 171, 183Rottenberg, J., 508, 509, 697, 698Rottenstreich, Y., 143, 151Rotteveel, M., 458Rotton, J., 236Rotundo, A., 28Rouhi, A. M., 236Roulet, N., 765Rowe, G., 784Royet, J.-P., 603Roysamb, E., 475Royzman, E. B., 762, 771, 781, 788Rozanski, A., 665

Rozin, P., 81, 219, 227, 411, 421, 683, 757,758, 759, 760, 761, 762, 763, 764, 765,766, 767, 768, 769, 770, 781, 788

Rubin, D. C., 602, 605, 607Rubin, K. H., 492, 734Ruble, D. N., 446Ruch, W., 216, 218, 219Ruckdeschel, K., 379Rucker, D. D., 594Rudolph, K. D., 404Rudrauf, D., 53, 284Rudy, J. W., 164Ruffman, T., 328Rugg, M. D., 604, 609Ruggiero, D. A., 161Rugulies, R., 677Ruiz, J. M., 665, 676Rukavina, I., 525Rumsfeld, J. S., 477Runciman, W. G., 430Rush, A. J., 508Russell, B., 9Russell, J. A., 52, 60, 62, 76, 103, 111, 180,

181, 187, 197, 223, 274, 292, 333, 349,353, 354, 355, 356, 357, 358, 359, 414,431, 432, 472, 514, 517, 518, 521, 522,524, 553, 589, 590, 591, 592, 593, 605,629, 630, 657, 693, 697, 779, 797, 804

Russell, M., 240Russell, T. A., 225Russo, G. K., 164Russo, R., 717, 718Rust, M., 785Rusting, C. L., 400, 478Rutter, D. R., 682Rutter, M., 800Ryan, L., 594Ryan, R. M., 478, 479, 560Rydell, A., 734, 736Ryff, C. D., 477, 778, 782Ryle, G., 10, 11Ryzewicz, L., 676

Saakvitne, K. W., 443Saalbach, H., 325Saarni, C., 310, 312, 334, 335, 336, 337,

339, 340, 349, 361, 534, 730, 731Saavedra, R., 462Sabatinelli, D., 719Sabean, D., 23Sabini, J., 630, 762, 771Sacks, D. S., 57Sackur, J., 171Saddoris, M. P., 252Sadovsky, A., 489Safran, J. D., 700Sagi, A., 441, 444, 448Sahley, T. L., 54, 55Saiz-Ruiz, J., 700Sakaguchi, A., 161Salamone, J. D., 253Sales, J. M., 518Salovey, P., 90, 180, 225, 399, 458, 464,

472, 534, 535, 536, 537, 538, 539, 540,541, 543, 654

Salt, P., 187Salvarani, G., 442Salzman, C. D., 252Sameroff, A. J., 489Samoilov, A., 88Samuelson, P., 144Sananes, C. B., 162Sanbonmatsu, D. M., 553Sánchez, M., 183, 278Sander, L. W., 309Sanderson, C. A., 479Sandler, D., 574, 580Sandvik, E., 397, 472, 473, 476Sandy, J. M., 678Sanfey, A. G., 147, 148, 763Sani, F., 462Santos, A., 518Sanyal, P., 679Saper, C. B., 251, 253Saphier, D., 252

826 Author Index

Sapolsky, R. M., 663, 664, 669, 788Sarnyai, Z., 56Saron, C. D., 188Sarter, M., 185, 252Sartre, J. P., 9, 10, 14Sato, A., 276Sattely-Miller, E. A., 244Savander, V., 160Savic, I., 237, 239Sawchuk, C. N., 758, 768, 769Scabini, D., 253, 669Scannell, J. W., 171Scarantino, A., 456Schaal, B., 235Schachar, R. J., 488Schachter, S., 10, 182, 185, 186, 307, 308,

628, 745Schacter, D. L., 171, 257, 603, 604, 606,

607, 608, 609, 610, 611Schaefer, A., 605, 606Schaefer, S. M., 504Schafe, G. E., 162Schafer, A., 402, 768, 769Schaffer, H. R., 310, 313, 315Schaffer, L. F., 534Schaller, M., 338, 443, 770Schechner, R., 411Schedlowski, M., 661Scheff, T. J., 39, 40, 44, 110Schefft, B. K., 784Scheier, M. F., 472, 542, 652, 779, 803Scheler, M., 9Schelling, T. C., 148Scherer, K. R., 36, 71, 73, 74, 76, 77, 78, 80,

107, 108, 181, 197, 198, 199, 200, 201,202, 203, 205, 207, 212, 397, 401, 431,499, 503, 574, 577, 580, 629, 634, 635,639, 759, 762

Schieman, S., 382, 397Schienle, A., 402, 765, 768, 769Schiffman, S. S., 236, 244Schilling, E. M., 489Schimmack, U., 472, 473, 474, 759Schipper, E. M., 488Schkade, D., 152, 481, 651, 777Schleifer, S. J., 664, 667Schlosberg, H., 237Schlosser, S., 731Schmahmann, J. D., 258, 279Schmid Mast, M., 399Schmidt, H., 765Schmidt, J. A., 44Schmidt, L. A., 189Schmidt, M., 338, 734Schmidt, R. F., 276Schmidt, S. R., 607, 609Schmitt, C. S., 34Schmittberger, R., 148Schmolck, H., 607Schmukle, S. C., 503Schneider, R., 324, 359Schneider, W., 507, 713Schneiderman, E., 540Schneirla, T. C., 310Schnirman, G., 610Schnog, N., 23Schnornagel, J. H., 679Schnyer, D. M., 610, 611Schoeberlein, S., 380Schoen, M., 534Schoenbaum, G., 252Schoff, K. M., 377Schore, A., 364Schork, M. A., 677Schorr, A., 181, 503, 629Schredl, M., 383Schröder, M., 199Schroeder, T., 84Schroll, M., 678Schudlich, T. D. D. R., 734Schul, R., 602Schuldberg, D., 791Schulkin, J., 165, 664Schulman, A., 206Schultheiss, O. C., 789

Schultz, D., 736Schultz, R. T., 719Schultz, W., 52, 59Schulz, S., 187Schure, E., 780Schut, H., 798, 803, 805Schutte, N. S., 539Schwartz, A., 141Schwartz, D., 733, 734Schwartz, G. E., 90, 187, 188, 219, 305,

516, 519, 520, 521, 670Schwartz, J., 517, 581, 588Schwartz, J. C., 76Schwartz, M. D., 648, 650, 656Schwartz, M. F., 152Schwartzberg, S. S., 798Schwarz, N., 143, 474, 535, 536, 549, 553,

555, 558, 559, 565, 566, 567, 606, 629,637, 651, 655, 657, 799

Schwarze, B., 148Schweinhardt, P., 281Schweitzer, M. E., 785Scollon, C. N., 480Scollon, N. C., 515, 516, 525Scolton, K., 338Scott, E., 57Scott, H. H., 77Scott, H. S., 629Scott, J. P., 54Searle, J., 13Sechrest, L., 228, 516, 520, 521Sedikides, C., 437Seebeck, T., 240Seeley, W. W., 282, 284Seeman, T., 788, 789Segal, J., 538Seger, C., 430, 437Segerstrom, S. C., 663, 664, 670Seidlitz, L., 398, 473, 476Seidman, S., 25Seiler, C., 802Seilman, U., 110Selden, N. R., 164Seligman, M. E. P., 481, 670, 711, 712, 746,

752, 777, 780, 788, 789Sell, A., 119, 127, 131, 132Sellers, W., 93Seltzer, M. H., 89Selye, H., 645, 661Sem-Jacobsen, C. W., 254, 259, 261Semegon, A. B., 383Semin, G. R., 459Semmel, A., 752Sendlmeier, W. F., 200Seneca, 5, 7Senn, M. J. E., 765Sennett, R., 37Senulis, J. A., 188Seqhier, M., 720Sergent, C., 171Sergerie, K., 604Serpell, L., 768Serrano, J. M., 366Servan-Schreiber, D., 162, 163Setlow, B., 252Setterlund, M. B., 593, 594Settersten, R. A., 382Severson, R. K., 649Sewards, M. A., 250Sewards, T. V., 250Sewell, A., 733Seymour, B., 171, 283Shackman, J. E., 371Shafran, R., 694Shakespeare, W., 110Shalit, F., 670Shalker, T., 548Shallice, T., 171Shamasundara, P., 520Shamir, H., 734Shanahan, M. J., 492Shand, A. F., 73Shanks, D., 588Shao, L., 476Shapira, N. A., 758, 768, 769

Shapiro, A., 298Shapiro, D., 670Shapiro, D. E., 520Shapiro, E., 477Shapiro, F., 89Shapiro, K. L., 170Share, D. L., 765Sharma, D., 682, 716Sharot, T., 605, 608Shaver, P. R., 76, 517, 518, 523, 525, 581,

588, 592, 593, 597, 634Shaw, B. F., 508Shaw, D. S., 733, 736Shaw, L. L., 449, 799Shaw, P., 603Sheard, D. E., 605Sheehan, G., 336Shefrin, H. M., 150Shekelle, R., 665Sheldon, K. M., 789Sheldrick, R., 575Shelford, V. E., 332Shelton, S. E., 55, 252Shepard, B., 377, 732Shepard, R. N., 128Shepard, S. A., 343, 731Shepherd, J., 334, 731Shepherd, M., 122Sheppard, B. H., 734Sheridan, J. F., 190, 666, 669Sherif, M. A., 423, 425Sherman, G., 763Sherman, J. C., 258Sherman, M., 767Sherman, N., 767Sherrington, C. S., 273Sherrod, A., 760Sherrod, L. R., 782Sherwood, A., 398Sherwood, R. J., 260Shestyuk, A. Y., 670Shewmon, D. A., 54, 62Shi, C., 252, 721, 722Shields, S. A., 22, 25, 398Shifflett-Simpson, K., 734Shiffman, S., 539Shiffrar, M. M., 517Shiffrin, R. M., 507, 713Shimamura, A., 608, 610Shimmack, U., 515, 516Shin, H., 476Shin, L. M., 169Shinar, O., 678Shiota, M. N., 224, 226, 780Shipley, M. T., 250Shipman, K., 400, 735Shirao, N., 258, 268, 521Shiv, B., 151, 152Shmidt, E., 553, 555, 566, 568Shoda, Y., 497, 645Shoqeirat, M. A., 610Shors, T. J., 165Short, R., 802Shorter, E., 20Showers, C. J., 534Shrestha, S., 341Shuchter, S. R., 798, 799, 802Shumway, D. R., 25Shumway, T., 143, 561Shuster, L., 666Shweder, R. A., 68, 409, 410, 413, 414, 417,

419, 420, 421, 759, 763Siau, J., 477Siegal, M., 765Siegel, A., 250, 665Siegler, I. C., 477Siegman, A. W., 203Siemer, M., 631, 632, 633Sifneos, P. E., 521, 536Sigman, M., 222Silberschatz, G., 94Silk, K. R., 697Silliman, R. A., 680Silva, A. J., 164Silver, L. A., 431

Author Index 827

Silver, M., 630Silverman, B. E., 665Silverman, K. E. A., 200Silvia, P., 677Simard, J., 356Simion, F., 799Simmonds, S. F., 152, 552, 558Simmons, A., 273Simmons, J. C., 539Simmons, J. L., 38Simmons, S., 539Simmons, W. K., 518Simner, M. L., 441, 444, 448Simon, B., 434Simon, H. A., 622, 647Simon, R. W., 397, 398Simons, C. J. R., 782Simons, L., 666Simons, R. F., 521Simonson, I., 152, 550Simonsson-Sarnecki, M. M., 522Simonton, D. K., 108Sinclair, A., 321, 353Singer, B., 778, 782Singer, D. E., 679Singer, J., 10Singer, J. A., 670Singer, J. E., 182, 185, 186, 307, 308, 514,

628, 745Singer, J. L., 782Singer, R. P., 646Singer, W., 252Singh, K. D., 281Singh, L., 763Sinha, P., 222Sirois, B. C., 677Sisman, E., 108Sitarenios, G., 225, 539Siviy, S. M., 56Skinner, B. F., 51Skitka, L. J., 435Skoner, D. P., 190, 677Slade, L., 328Slade, M. D., 777Slenker, S. E., 649Sloboda, J. A., 104, 107Sloman, A., 69Sloman, L., 669Sloman, S. A., 633, 634Slomkowski, C., 328Slot, N. W., 736Slovic, P., 151, 561, 562, 647, 654, 655,

657, 679Small, D. A., 148, 561Small, D. M., 603Smetana, J. G., 325Smilek, D., 716Smiley, P., 353Smith, A., 8, 150, 441, 442Smith, A. C., 35Smith, A. J., 699Smith, A. P., 604, 609, 611Smith, B. W., 516, 676, 804Smith, C. A., 77, 151, 437, 533, 629, 633,

779, 780, 800Smith, C. L., 487, 489Smith, C. N., 168Smith, D., 651Smith, E. E., 169, 171, 250, 257Smith, E. R., 429, 430, 431, 433, 434, 435,

436, 437, 461, 629, 633Smith, H., 473Smith, J., 382Smith, K., 51Smith, K. S., 778Smith, L. B., 372, 373Smith, M. C., 217, 605Smith, N. K., 185Smith, P. K., 327, 782, 786Smith, R., 679Smith, R. N., 521Smith, S. E., 552Smith, T. W., 664, 665, 676, 680Smith, Y., 163Smith LeBeau, L., 403

Smith-Lovin, L., 39Smoski, M. J., 205, 206Snowdon, C. T., 214Snowdon, D. A., 777Snyder, B. J., 253Snyder, C. R., 462Snyder, J., 170, 734Snyder, S. S., 502Soares, J. J. F., 713, 714, 715, 718Sobin, C., 199, 201, 202Sodian, B., 325Soken, N. H., 351, 369Solnit, A. J., 765Solomon, F., 798Solomon, G. E. A., 517Solomon, G. F., 664Solomon, K. O., 596, 598Solomon, L. Z., 605Solomon, R., 12, 13Solomon, R. C., 73, 82Solomon, R. H., 23Solomon, R. L., 165, 788Somberg, D. R., 736Somerville, L. H., 252, 270Sondheimer, S. J., 241Song, A. W., 798Sonnby-Borgström, M., 458Sonnemans, J., 74Sontag, S., 450Sorabji, R., 5Sorce, J. F., 212, 298, 351, 352, 371, 459Sorenson, E. R., 223, 292Soto, J. A., 216Sparrow, D., 663, 677, 678Spasojevic, J., 503, 800Spassova, G., 561Spears, F. M., 734Spears, R., 431, 463Speicher, C. E., 190Spelke, E. S., 352Spence, C., 596Spencer, D. D., 167Spencer, W. B., 382Speranza, M., 698Sperry, L. L., 337, 518Spiegler, B. J., 160Spielberger, C. D., 339, 710Spinoza, B., 7, 12, 71, 72, 83, 95Spinrad, T. L., 336, 487, 488, 489, 733Spiro, A., 382, 477, 678Spitz, R. A., 309Spitzer, C., 521, 698Spoor, J. R., 462Spracklen, K. M., 493Sprengelmeyer, R., 223, 767, 769Squire, L. R., 160, 166, 170, 252, 607, 610Srendi, B., 670Srivastava, S., 505, 506Sroufe, A., 578Sroufe, L. A., 93, 294, 295, 298, 309, 786Srull, T. K., 507Stack, D. M., 370Stafford, R. S., 677Stalikas, A., 91, 94Stanger, C., 313, 750Stansbury, K., 343Stanton, A., 91Stanton, S. J., 789Stark, C. E., 610Stark, J. L., 669Stark, R., 402, 765, 768, 769Starkstein, S., 284Staubli, U. V., 162Staudinger, U. M., 380, 445Staw, B. M., 549Stearns, C. Z., 19, 22, 24, 25, 27, 799Stearns, P. N., 19, 22, 24, 25, 26, 27, 29,

798Steckley, P., 89Steele, J. J., 236Steen, T. A., 789Stefanacci, L., 169Stefanek, M. E., 649, 676Stefanski, V., 666Stegagno, L., 253

Stegbauer, G., 358Steidl, S., 611Steimer-Krause, E., 222, 693Stein, B. E., 368Stein, E. A., 168, 603Stein, M., 765Stein, M. B., 273, 283, 699, 700Stein, N., 71, 78, 414, 421, 802Stein, N. L., 180, 322, 355, 574, 575, 576,

577, 578, 579, 580, 581, 582, 583, 584,787, 791

Stein, S., 533Steinberg, L., 489Steinberg, S., 336Steiner, C., 537Steiner, J. E., 61, 444, 765Steketee, G., 699Stemmler, G., 77, 184, 190, 503Stenberg, C. R., 131, 306, 315, 578, 731Stennett, J., 223Stephan, C. W., 429Stephan, K. E., 609Stephan, W. G., 429Stepper, S., 186, 397, 505Sterling, R. S., 106Stermac, L., 89Stern, D. N., 364, 371, 372, 441Stern, K., 240, 241Stern, P., 447Stern, R. M., 181Sternberg, R. J., 273, 533Sternthal, B., 549Sterpenich, V., 604Stetler, C. A., 667Stets, J. E., 33, 35, 38, 39, 43Stettner, L., 336, 522, 536Stevens, A. A., 523, 538Stevens, K. N., 198, 199Stice, E., 697Stifter, C. A., 336, 358, 488, 489, 503Stigler, J., 410Stiles, T. C., 89Stillwell, A. M., 134, 459Stipek, D. J., 226, 310, 317, 735, 743, 746,

749Stitzer, M. L., 678Stock, W. A., 477Stocker, M., 10Stockhorst, U., 237Stoddart, D. M., 236Stohler, C. S., 258Stoller, E. P., 654Stone, A. A., 481, 670Stone, L., 20Stone, L. A., 521, 522Stone, L. D., 383Stone-Elander, S., 719Stoner, S. B., 382Stoolmiller, M., 734Storbeck, J., 81, 556, 566, 567, 606, 634,

799Stote, D. L., 164Stoumbos, J., 367Stouthamer-Loeber, M., 221Stowe, H. B., 446, 447, 450Strack, F., 186, 203, 397, 404, 434, 474,

505, 507, 549Strack, S., 801Stradler, M. A., 608Strandberg, K., 337, 734Strange, B. A., 170, 604Straube, T., 719Straus, R., 540Stravinsky, I., 104Strayer, J., 337, 443, 734Streeter, L. A., 199, 202Striano, T., 364, 370, 373Strick, P. L., 258, 259Strigo, I. A., 273, 278Stroebe, M. S., 538, 798, 799, 803, 805Stroebe, W., 538, 798, 799, 805Stroehm, W., 594Strohle, A., 664Strong, S. E., 594Strosahl, K. D., 694

828 Author Index

Stroud, L. R., 538, 678, 697Stryker, S., 38Stubbe, J. H., 475Stuewig, J., 683Su, A., 163Suarez, E. C., 665Suarez, S. D., 135Subic-Wrana, C., 525Suci, G. J., 180, 589, 635Suengas, A. G., 607Sugden, R., 141, 150Suggs, M. S., 244Suh, E. M., 473, 476, 477, 479Suinn, R. M., 665, 731Sujan, H., 551Sujan, M., 551Sukhdial, A., 683Sullivan, J., 299, 731Sullivan, K., 786Sullivan, M. W., 212, 217, 225, 226, 300,

306, 313, 316, 414, 730, 731, 746, 750,753

Sullivan, R. M., 54, 253Summerfield, A. B., 36, 397Summers, R., 182Summers-Effler, E., 40, 41Sunderland, M., 56Suomi, S., 311Supple, W. F., 160Surcinelli, P., 222Suslow, T., 523Susser, K., 549Sutherland, R. J., 164Sutton, J., 327Sutton, R. I., 500Sutton, S. K., 188, 472, 650Svartberg, M., 89Svejda, M., 371Svenningsson, P., 700Swan, G. E., 236, 677Swanberg, K., 606Swann, W. B., Jr., 39Swanson, L. W., 159, 160, 164, 251Swartz, T. S., 457Swartzentruber, D., 164Swets, J. A., 124Swettenham, J., 327Swigar, M., 678Swinkels, A., 523Switzer, G., 443Symons, D., 121, 133Synnott, A., 236Syversen, J. L., 477Szabo, G., 56Sznycer, D., 119Szporn, A., 446

Tajfel, H., 428, 636Takatsuji, K., 61Takeuchi, S., 400Talairach, J., 256Talarico, J. M., 607Talmi, D., 606, 608, 609Tamang, B., 341, 342Tambiah, S. J., 760, 762Tambs, K., 475Tamir, M., 500, 505, 506, 509Tan, E. S., 110Tan, J., 323Tang, Y., 263Tangney, J. P., 39, 339, 459, 537, 668, 683,

749, 752, 754Tannenbaum, P. H., 180Tanner, W. D., 124Tannock, R., 488Tantow, B., 768Tardif, T., 324Taska, L., 752Tassinary, L. G., 181, 187, 200, 219, 397Tata, P., 717Taylor, C., 768Taylor, C. B., 190Taylor, G. J., 521, 522, 523, 698Taylor, K. L., 648Taylor, P. J., 563

Taylor, S., 663Taylor, S. E., 185, 190, 220, 479, 549, 552,

557, 670Taylor, S. F., 250, 252, 255, 269, 284, 604,

666Taylor, V. M., 679Tcherkassof, A., 80Teachman, B. A., 768, 769Teasdale, J. D., 93, 548, 593Tees, C., 54, 62Teich, A. H., 164Tellegen, A., 181, 472, 475, 478, 553, 589,

696, 779ter Schure, E., 430Terman, L. M., 534Terracciano, A., 225, 382Tesla, C., 328Tesman, J. R., 732Tesser, A., 542Tessitore, A., 223Teti, L. O., 731Tett, R. P., 539Thagard, P., 637Thaler, R. H., 145, 150Thamm, R. A., 43Thanosoulis, L. C., 166Thase, M. E., 667, 700Thayer, J. F., 188, 190, 694Thayer, R. E., 400Thein, S., 351Thelen, E., 79, 297, 298Thell, S., 243Theodorou, P., 221Thistlethwaite, D. L., 646Thoits, P. A., 34, 507Thomas, A., 486Thomas, W., 525Thompson, M. L., 666Thompson, R., 338Thompson, R. A., 338, 499, 500, 501Thompson, R. F., 611Thompson, T., 605Thompson-Brenner, H., 700Thompson-Schill, S. L., 253, 257, 596Thorndike, E. L., 533Thorndike, R. L., 533Thorndike, R. M., 767Thorpe, S. J., 164, 171Thrash, T. M., 472Thunberg, M., 187, 243Thye, S. R., 42Tickle, A., 200Tidball, G., 351Tiebout, J., 329Tiedens, L. Z., 147, 212, 431, 457, 460, 790Tierney, W. M., 678Timmers, M., 396, 400, 401, 460Tinley, S. T., 649Tipples, J., 716Tobin, R. M., 397, 400, 404Todarello, O. O., 522Todd, R. M., 449Todt, D., 205, 206Tohyama, M., 61Tolin, D. F., 767, 768, 769Tolkmitt, F. J., 200Tomaka, J., 119Tomarken, A. J., 188, 205, 699Tomasello, M., 217, 336Tomaz, C., 165Tomhave, J. A., 404Tomich, P., 338Tomkins, S., 291, 292Tomkins, S. D., 307, 315Tomkins, S. S., 72, 75, 76, 186, 214, 378,

513, 535, 540, 743, 750, 758, 765Tomonaga, M., 214Tong, E. M. W., 780, 782, 790Tononi, G., 249Tooby, J., 115, 116, 117, 118, 119, 120,

121, 124, 125, 127, 128, 129, 131, 133,134, 224, 779, 780

Toronchuk, J. A., 48, 53Totterdell, P., 461, 462Toufexis, D. J., 721

Toukmanian, S., 95Trabasso, T., 71, 78, 109, 322, 355, 574,

575, 576, 577, 580, 581, 787Tracey, I., 278Tracy, J. L., 226, 742, 749, 750Trainor, L. J., 104Tran, N., 91Tranel, D., 142, 170, 181, 182, 184, 224,

253, 603, 610Trask, O. J., 190Treasure, J. L., 768Trehub, S. E., 104Trembly, G., 646Trentacosta, C. J., 736Trevarthen, C., 335, 364Treynor, W., 541Triandis, H., 473Trimboli, A., 228Trippe, R., 716Trivedi, N., 206Trivers, R. L., 133, 802Troisi, A., 521Tronick, E. Z., 300, 364, 370, 372Troop, N. A., 768Trope, Y., 549, 552, 557, 784Trötschel, R., 507Trouvain, J., 200Troyer, D., 443Truax, K. M., 180Trumbach, R., 20Tsai, J. L., 219, 382, 383, 384, 505, 515,

685, 779Tsang, C. D., 104Tsao, J. C. I., 694Tse, P. U., 284Tsoudis, O., 39Tsushima, T., 35Tucker, D. M., 126, 784Tugade, M. M., 95, 516, 519, 525, 677,

681, 780, 787, 788, 804Tuma, A. H., 709Tung, N. M., 679Turiel, E., 421Turk, C. L., 698Turken, A. U., 550Turken, U., 789Turkewitz, G., 367Turner, B., 161Turner, C. W., 665Turner, J. C., 428, 429, 434, 461Turner, J. H., 33, 38, 39, 43Turner, J. R., 398Turner, R. A., 789Turner, R. B., 190, 677Turner, T., 76, 77Turner, T. J., 580, 602Turvey, C., 90, 538Tuttle, E., 477Tversky, A., 140, 143, 151, 185, 245, 536,

550, 551, 563, 647, 648, 781, 788Tweed, D. L., 678Tylor, E. B., 760, 761Tyson, M. C., 750

Ubel, P. A., 651Ueno, A., 214Ueno, Y., 214Ullman, J. D., 107Underwood, M. K., 337, 735, 736Ungerleider, L. G., 605, 611, 721Urban, N., 679Urry, H. L., 169, 188Uvnas-Moberg, K., 285

Vaidya, J., 181, 589, 696, 779Vaish, A., 370, 373Vaitl, D., 402, 722, 765, 768, 769Valdimarsdottir, H., 670Valenstein, E. S., 251, 261, 262Valiente, C., 443Vallacher, R. R., 633Vallance, D. D., 357Van Ameringen, M. A., 699Van de Wal, C., 224van den Brink, W., 699

Author Index 829

Van der Horst, V. G., 250van der Kolk, B. A., 93Van der Linden, M., 386van der Linden, M., 602, 606, 608, 609van der Plight, J., 141Van der Pligt, J., 459Vanderschuren, L. J., 56, 57Van Diest, I., 678van Dyck, R., 699Van Goozen, S., 74Van Gundy, K., 397Van Hoof, J. A. R. A., 214van Hooff, J. A. R. A., 188, 205Van Hulle, C. A., 403Van Kleef, G. A., 460Van Lancker, D., 202Van Mechelen, I., 460, 516Vanman, E. J., 397Van Overveld, W. J. M., 767Van Ree, J. M., 56van Reekum, C. M., 199Van Servellen, G., 680Vansteelandt, K., 516Van Toller, S., 235Van Velsor, P., 92van Vianen, A. E. M., 397van Vianen, E. A. M., 459Van Zomeren, M., 463Varlamov, A. A., 521, 522Vassend, D., 477Västfjäll, D., 104, 655Veenhoven, R., 474Veenstra, R., 492Veerman, J. W., 736Vera, M. N., 183Verfaellie, M., 610Vermeulen, N., 523, 596Vernon, L. L., 695Vest, J. R., 683Vettin, J., 205, 206Vianna, E. P. M., 182Vick, S. J., 217Vietze, P., 365Vila, I., 183Vilberg, T., 54Villablanca, J., 62Vingerhoets, A. J. J. M., 459Vinokur, A. D., 786Visscher, K. M., 282Viswanathan, K., 663Vitterso, J., 475Voet, W., 758Vogel, D. L., 399Vogelsang, T. J., 145Vogt, B. A., 47, 253Vogt, L. J., 253Vohs, K. D., 629, 636, 781Vokonas, P., 677, 678Vondra, J. I., 733von Helversen, O., 240, 244von Neumann, J., 140von Salisch, M., 335, 338Vosk, B. N., 735Vrana, S. R., 188, 402, 724Vroomen, J., 521Vuilleumier, P., 170, 171, 603, 605, 719,

720, 721

Wachs, T. D., 486, 489Wadlinger, H. A., 784Waford, R. N., 736Wager, B. M., 637Wager, T. D., 253, 255, 257, 258, 259, 284,

403, 666Waggoner, J. E., 180Wagner, A. D., 257, 610Wagner, G., 222Wagner, H., 183Wagner, P. E., 537, 754Wagner, P., 752Wagner, R. K., 533Walden, T. A., 340, 351, 732Waldstein, S. R., 190, 402Walker, A. S., 367, 369Walker, D. L., 252, 721, 722, 725

Walker, E. F., 699Walker, M., 228Walker, P. A., 519Walker-Andrews, A. S., 349, 350, 351, 365,

367, 368, 369, 370, 371, 373, 731Walker-Engstrom, M. L., 679Wall, T., 629, 634, 636Wallace, K. J., 252Wallace, M. T., 368Wallbott, H. G., 36, 199, 200, 397, 759Waller, B. M., 217, 227Wallon, H., 72Walsh, B. T., 699Walsh, J., 6Walter, B., 768, 769Walters, E., 694Walthall, A., 26Walton, K., 13Wane, M., 760Wang, A., 539Wang, J., 146Wang, L., 798, 799Wang, P. S., 680Wang, Z., 55, 61Waraczynski, M. A., 252Ward, D. W., 562Ward, J. H., 447Ware, J., 760Warner, R. M., 539Warr, P., 472Warren, J. D., 285Warren, S., 337Warwar, N., 92, 95Warwar, S., 95Waschbusch, D. A., 736Waskett, L., 239Wasserman, G., 336Wastell, V., 241Waters, A., 716Waters, E., 353, 634Watkins, E., 694Watkins, L. R., 666Watson, D., 181, 431, 472, 477, 553, 589,

678, 681, 694, 696, 697, 779, 791Watson, J., 57, 88, 324Watson, J. B., 51, 308Watson, J. C., 89, 90, 94, 96Watson, J. M., 608Watson, K. K., 280Watson, S., 134Watt, D. F., 60Watten, R. G., 477Watts, J. C., 646Waugh, C. E., 783, 785, 786, 787, 789, 791Wearing, A., 474, 478Webb, K., 760Weber, E. U., 138, 651, 679Weber, H., 336Weber, M., 171Weber, R. A., 152Webster, M., Jr., 35Wedekind, C., 240Weeks, J. L., 447Wegener, D. T., 552, 559, 594Weghorst, S. J., 120Wegner, D. M., 59, 541, 633Wehmer, F., 522, 536Weimer, B., 337, 338Weinberg, K., 300Weinberg, M. K., 372Weinberger, D. A., 305, 670Weinberger, D. R., 223Weinberger, M., 525Weinberger, N. M., 162, 166, 252Weindruch, R., 388Weiner, B., 440, 447, 549, 743, 746Weiner, H., 662Weinman, J. A., 680Weins, S., 282Weinstock, J., 182Weintraub, J. K., 652Weintraub, S., 693Weir, B., 39Weir, T., 399Weise, D., 696, 761

Weisfeld, G. E., 205Weiskrantz, L., 83, 160, 521, 720Weisman, M. S., 200Weiss, A., 779Weiss, B., 736Weiss, J. M., 81Weiss, M., 313Weiss, S. T., 663, 678Weiss, T., 716Weissman, D. H., 282Weisz, J. R., 79, 492, 502Welch, N., 138, 651, 679Welkowitz, J., 385, 386Weller, S. J., 244Welling, H., 799Wellisch, D., 650Wellman, H. M., 321, 322, 324, 352, 353,

355, 356, 360, 447, 489Wells, K. C., 736Werner, H., 349Werner, K. H., 697, 698, 699Westen, D., 698, 700Westendorf, D. H., 769Wester, S. R., 399Westphal, M., 803Wetherell, M. S., 428, 461Wever, L. D. V., 679Wflite, J., 680Whalen, P. J., 160, 163, 169, 223, 252, 268,

270, 719Wheatley, T. P., 502Wheeler, R. E., 188Whelton, W. J., 88, 92Whike, K., 56White, G., 420White, G. M., 333White, J. M., 476White, K., 25White, K. S., 699White, R. T., 602Whitfield, I. C., 163Whitman, R. D., 236Whittlesea, B. W. A., 608Wicke, J. D., 188Wicker, B., 147, 368, 768Wicklund, R. A., 313Widen, S. C., 349, 353, 354, 355, 356, 357,

358, 359Wiener, S. G., 311Wiens, S., 171, 183, 272, 713Wierzbicka, A., 68, 69, 333, 336, 412, 420,

591, 758Wiese, D., 181, 589, 779Wiest, C., 457Wigboldus, D., 433, 463Wiggs, C. L., 611Wik, G., 719Wilbarger, J. L., 181, 224, 637Wilcox, P., 649Wilensky, A. E., 162Wilhelm, F., 693Wilhelm, F. H., 181, 498, 503, 507, 508,

798Wilhelm, F. L., 217, 219Wilhelm, F. W., 508Willard, K. S., 648Willatts, P., 578Williams, A. F., 798Williams, C., 443Williams, C. E., 199Williams, C. L., 165Williams, C. M., 236Williams, E., 736Williams, G., 343Williams, G. C., 119, 120Williams, J. I., 678Williams, K. D., 147, 456Williams, L., 283Williams, L. D., 608Williams, L. E., 507Williams, N. L., 767, 768Williams, R., 717, 719Williams, R. B., Jr., 665Williams, T., 109Williams, X. B., 731

830 Author Index

Willingham, B., 216, 219Wills, T. A., 645, 678Wilson, A., 377, 378Wilson, C. D., 518Wilson, D. S., 782Wilson, K. G., 694Wilson, M., 116, 120, 128, 696, 700, 734Wilson, P. J., 239, 242, 244Wilson, T. D., 142, 145, 410, 502Wilson, W., 476, 477Wilson-Cohn, C., 223Wimmer, G. E., 147Winblad, S., 222Windmann, S., 608Winkielman, P., 181, 224, 596, 597, 637Winkler, L., 240Winograd, E., 605Winslow, E. B., 733Winston, A., 89Winston, J., 257Winston, J. S., 170Winter, L., 379Wintre, M. G., 357Wirth, M. M., 789Wisco, B., 800Wisecup, A. K., 39Wiser, S., 94Witherington, D. C., 334, 351, 730Witkiewitz, K., 678Witte, K., 682Witter, R. A., 477Wittgenstein, L., 12, 321Woike, B. A., 636Woldorff, M. G., 282Wolf, K., 759Wolf, S., 759, 766Wolff, P., 298Wolitzky, D., 228Wolpe, J., 700, 788Wong, N., 515Wood, J. V., 479Wood, S., 610Wood, W., 396, 404Woodin, E. M., 224, 336Woods, C. M., 768Woodward, A., 578, 583Woodworth, R. S., 237, 534Woody, S. R., 768, 769Woodzicka, J., 217

Woolery, A., 538Woolley, J. D., 355Worsley, K. J., 255Worth, L. T., 549, 565Wortman, C. B., 798Wouters, C., 22, 25, 29Wouters, L., 699Wrase, J., 700Wright, P., 768Wrzesniewski, A., 765Wu, S., 76Wurm, L. H., 386Wyatt, G. E., 752Wyer, R. A., 562Wyer, R. S., 507Wylie, T., 415Wysocki, C. J., 239, 241

Xagoraris, A. E., 161, 164, 634Xu, J., 770

Yaeger, A., 678Yamamoto, M., 734Yamawaki, S., 258, 521Yang, H., 556Yasui, Y., 253Yates, G., 56Yatiswarananda, S., 422Yechiam, E., 494Yehuda, R., 664Yehuda, S., 670Yi, Y., 515Yirmiya, N., 222Yirmiya, R., 667Yonelinas, A. P., 604Yoo, S. H., 225Yoon, J., 42Young, A. W., 224, 716, 768Young, E. A., 503Young, G., 735Young, H., 202Young, L. J., 55, 57, 61, 278Young, M. J., 549, 784Young, M. P., 171Young, P. T., 534Young, R. F., 649Young-Browne, G., 365Youngblade, L., 328Yovel, I., 716

Yrizarry, N., 223Yücel, B., 522Yuill, N., 323Yurak, T. J., 147Yzerbyt, V. Y., 433, 463

Zaalberg, R., 457Zachariae, R., 670Zacks, R. T., 396Zahn-Waxler, C., 337, 341, 353, 397, 441,

444, 445, 731, 733, 748Zajonc, R. B., 71, 82, 83, 183, 184, 185,

205, 207, 244, 304, 305, 574, 611, 638,647, 655

Zak, P. J., 56, 789Zald, D. H., 237, 610, 700Zalecki, C., 694Zambreanu, L., 278Zanakos, S., 541Zautra, A. J., 285, 516, 676, 682, 804Zbrozyna, A. W., 160Zebrowitz, L. A., 636Zeelenberg, M., 141, 459Zeelenberg, R., 596Zeidner, M., 399, 543Zelazo, P. D., 629, 635, 638, 639Zelditch, M., Jr., 35Zelizer, V., 24Zeman, J., 400, 735Zentner, M. R., 104Zevon, M. A., 472Zhang, E. T., 276Zhang, L., 629Zhao, Z., 170, 260Ziaie, H., 503Zillmann, D., 109Zimmermann, L. K., 343Zinbarg, R. E., 696Zipursky, B., 451Zisook, S., 798, 799, 802Zoellner, L. A., 91Zola, S. M., 160, 166Zola-Morgan, S., 252Zoller, D., 371Zonderman, A. B., 225, 382Zornberg, C. L., 54Zubieta, J. K., 55, 258Zucker, B. G., 694Zumbahlen, M. R., 298, 732

Author Index 831

Subject Index

Page numbers followed by an n, f, or t indicate notes, figures, or tables.

Abandonment, 89Aboutness, 13Abstinence, 423tAbuse, 89, 306, 752–754AC. see Anterior cingulate cortexAcademic competence, 341ACC. see Anterior cingulate cortexAcceptance, 69, 71Accommodation, 790Aching, 273–274, 278Acoustics. see Vocal acousticsAcquisition of standards/rules, 316fAction, 68, 72–76, 134, 165–166, 181, 332,

414, 557, 591t, 780–781Activation coordinates, 257fActivation states

descriptive models of, 590fdimensions approach and, 589disgust and, 769evolution and, 125fear and, 713–718, 714fgender and, 402–403positive emotions and, 779psychology and, 72, 74, 76semantic-network approach and, 594

Activism, 41Activity, 589, 635Actors, 41–42Adaptation. see Evolution; Flexibility;

ResilienceAdaptational encounters, 778Adaptive emotions, 89–90Addiction, 54Adjustment, 480, 692–693t, 751fAdolescents, 754fAdrenal glands, 166Adrenaline. see EpinephrineAdrenocortical response, 311, 661–670Adults, 292–293, 300–301, 447, 767Advice, 35Affect

in animals, 49–50, 50fANS and, 190appraisal and, 636behavior and, 629behavior/cognition and, 52–54cognition and, 568cognitive flexibility and, 550–557

cognitive-affective theory and, 380–381complexity and, 517, 521conclusions, 62–63, 568–570consciousness and, 61–62control theory, 39core. see Core affectculture and, 414–415, 417decision making and, 548–550, 561–568defining, 373n, 662descriptive models of, 590fdevelopment and, 376–382differential-emotions theory and, 378dimensions approach and, 589diminishment of, 50–52disengagement and, 666–667disgust and, 767disturbances and, 691–692dual-aspect monism and, 60–61dynamic model of, 804emotion and, 10, 262nemotional intelligence and, 537vs. emotions, 578–580, 657n, 778evolution and, 59–60expressive development and, 295facial expressions and. see Facial

expressionforecasting errors and, 144–145, 151ngender and, 401, 404–405health behavior and, 647, 651–652, 654–

655, 678, 681–682, 683timaging and, 259infants and, 366–367, 369, 371–372interoception and, 284–285joy/laughter and, 56–58memory and, 611motivation and, 557–561music and, 106nature of emotion and, 181neurology and, 49fneuroscience ontology and, 58overview of, 47–48positive approach and, 670positive emotions and, 789prediction of, 152nprogram, 11psychology and, 71, 73rewards and, 143sadness/panic and, 54–56

self-reporting and, 397social context and, 339socioemotional selectivity theory and, 380subjective well-being and, 473, 477–478vocalization and, 197, 203–207

Affect Communication Test, 228nAffect heuristic model, 655, 657nAffect Intensity Measure, 397–398Affection, 21Affective forecasting, 684Affective information, 373nAffective neuroscience, 52–54Affective valence, 71Affectus, 68Afferents, 275, 285AFFEX facial coding system, 293, 295, 301,

731, 734Affiliation function of emotion, 457–460Affordances, 333, 367Age considerations, 376–387, 516, 606Age-illness heuristic, 653–654Agency, 282–284, 313, 627n, 787fAggravation, 517Aggression, 115–116, 325, 492, 665, 734,

736, 780, 788, 790Agonistic behavior, 217Agoraphobia, 692–693t, 711, 723fAgranular insula, 253AIM. see Affect Intensity MeasureaINS. see Anterior insulaAir hunger, 273–274, 278Akhlaq, 423, 425nAlcohol use. see Substance useAlexithymia, 521–523, 698Algorithms, 123–124, 134Alienation, 32Altruism, 130, 424, 558Alzheimer’s disease, 236, 386Amazement, 411Ambivalent emotions, 326Amiability, 423tAmino acid transmission, 162Amusement, 411, 413Amusia, 284Amygdala

affect and, 53appraisal and, 638central, 162f

832

circuit theory and, 159cognition and, 169–170disengagement and, 669as emotional computer, 160facial expressions and, 223–224fear acquisition and, 167–168fear and, 718–721fear conditioning and, 160–167, 163ffear response and, 168–169grief and, 798imaging and, 250, 252, 259interoception and, 280flateral, 162f, 163–164limbic system and, 160memory and, 602–606, 608–610olfaction and, 237perception and, 169–170psychology and, 79, 84sadness and, 799

Analogy, 621Analysis, 535fAnalytic philosophy, 13AND. see AndrostadienoneAndrostadienone, 241Androstenol, 241Androstenone, 241Anger

affect and, 657nANS and, 183–185appraisal and, 632, 634baring teeth and, 291as basic-level emotion, 350built-in critics and, 620fchildren and, 350, 353, 355–356, 357f,

358children’s verbalization of, 321complexity and, 516–518conclusions, 736–737coping strategies for, 35core affect and, 630culture and, 341–342, 402, 414–418, 420–

421, 581defining, 12development and, 315, 316f, 385, 731–

735dimensions approach and, 589disgust and, 763disturbances and, 692–693t, 698elicitation and, 770embodied-simulation theory and, 596as emotion, 5emotional states and, 308, 619empathy and, 440, 447–448evolution and, 115–116, 127, 131–132exchange approach and, 42experience and, 312expression and, 92expressive development and, 294, 299–

300facial expressions and, 213t, 215–217,

218t, 219–220, 220f, 222, 226, 399as feeling, 54fight-flight response and, 664–665functional significance of, 730–731gender and, 401grief and, 798groups and, 429health behavior and, 678–679, 683thierarchical model and, 349fimaging and, 249infants and, 365, 368, 371intergroup emotions and, 431–433Islamic thought and, 423tKarl Marx on, 32language and, 581–582morality and, 763movement and, 310music and, 107overview of, 730positive emotions and, 784during postpartum period, 387as primary emotion, 43prototype approach and, 593tpsychology and, 70–71, 74–76, 79–81Sanskrit emotions and, 411–412

semantic-primitive approach and, 592social context and, 338–339social function and, 457, 460somatic nervous system and, 187speech and, 200–202stereotypes and, 396structural approach and, 36

Anger Response Inventory, 339Anhedonia, 692–693tAnimal spirits, 8Animals, 711, 723f, 760–762. see also

PrimatesAnnoyance, 430, 516–517, 537Anomie, 32Anorexia nervosa, 692–693tANS. see Autonomic nervous systemAntagonists, 109Antecedent events, 414Anterior cingulate cortex, 55f, 171, 250,

258, 261, 272, 275f, 277–279, 280f,282–284, 522, 669, 718–719

Anterior insula, 250, 253–254, 258, 260,279–285, 280f

Anterior temporal cortex, 253Anterior thalamus, 159Anthropodenial, 50fAnthropomorphism, 50fAnticipation, 622, 684Antidepressant medication, 699–700Antisocial personality disorder, 692–693tAntivitalism, 51Anxiety

ANS and, 185avoidance and, 166complexity and, 516conclusions, 724–725disgust and, 767–769disorders, 723fdisturbances and, 696–697EFT and, 89emotional intelligence and, 538evolution and, 123facial expressions and, 222fear and, 721–724, 723ffear inhibition and, 164–165fight-flight response and, 663–665Freud’s theory of, 12gender and, 404health behavior and, 648–650t, 678–679,

683theuristics and, 654hierarchical model and, 349fhistorical context of, 20intergroup emotions and, 430–431, 433introduction to, 709–710language and, 581–582music and, 107negative emotions and, 780neurology and, 718–724olfaction and, 237, 245overview of, 710positive emotions and, 784, 788during postpartum period, 387regulation and, 93, 508self-reporting and, 397situational context of, 710–712as symptom, 692–693ttemperament and, 490, 493transdiagnositic approach to, 694treatment of, 699–701unconscious processes and, 712–718

Apathy, 5, 7Apology, 460Appeal function of expression, 205Appearance, 423tAppraisal

affect and, 578–580, 778appraisal and, 580–581children and, 359cognition and, 33, 38conclusions, 583–584, 638–639conscious, 75constraints on, 581–583critics of appraisal theory and, 628–629culture and, 414–417

emotional, 11emotional intelligence and, 535f, 536, 539emotional scripts and, 323–324emotional signals and, 104evolution and, 124expressive development and, 296, 298facial expressions and, 219grief and, 798heuristics and, 654fintergroup emotions and, 431, 434introduction of, 574–575, 628memory and, 583modal model of emotion and, 499fnature of, 631–632nature of emotion and, 629–631overview of, 576–578positive emotions and, 779–780, 782process of, 632–638psychology and, 71, 82regulation and, 503–504social context and, 333social function and, 460treatment and, 701

Approach-withdrawal motivation, 189Arousal

affect and, 49, 553anger and, 731–732ANS and, 182children and, 360circumplex model and, 349fcomplexity and, 522descriptive models of, 590fdevelopment and, 387disturbances and, 692–693tEFT and, 89emotion elicitation and, 78emotional, 33femotional states and, 307–309empathetic overarousal and, 442–443empathy and, 441–442evolution and, 119, 126–129exchange approach and, 41–42expression and, 90–92expressive development and, 294fear and, 714, 714fgender and, 402–403health behavior and, 646, 681interoception and, 284–285memory and, 603, 605–606, 608, 611negative, 32olfaction and, 237overview of, 348physical. see Physiologypositive emotions and, 779, 788psychology and, 70, 83reflection on emotion and, 95socioemotional selectivity theory and, 379speech and, 199–201subjective well-being and, 477vocalization and, 197

Arrogance, 413Art, 111. see also Literature; MusicAsperger’s disorder, 692–693tAssessment, 89–90, 563–564, 602, 767Asset integration, 140–141Associations, 239, 243, 441, 518, 550–551,

610, 632, 635Associative processing, 633Astonishment, 411Asymmetry, 284–285, 580Atherosclerosis, 665Attachment, 54, 56, 308, 338, 786Attack, 294Attention

affect and, 553–556, 579anger regulation and, 338–339anxiety and, 717appraisal and, 575–576, 633–634development and, 315disturbances and, 698emotion and, 169–171, 498, 587, 778evolution and, 118, 126fear and, 715–716gender and, 400intergroup emotions and, 432

Subject Index 833

Attention (cont.)memory and, 583, 606modal model of emotion and, 499fpositive emotions and, 784, 785fpsychology and, 69regulation and, 488–489, 500, 501f, 502–

503ritual approach and, 41self-soothing and, 343shifting of, 97subjective well-being and, 478

Attitudes, 13, 73, 106–107, 142, 181, 429,434–435, 552, 565

Attractor states, 300Attribution processes

ANS and, 185children and, 321cognitive-attributional theory and, 745–

748emotional scripts and, 324–325empathy and, 447–448exchange approach and, 42global/specific, 743health behavior and, 653internal/external, 743maltreatment and, 753–754self-conscious emotions and, 745, 747–

748sexual abuse and, 752fshame and, 751fsocial context and, 334–337structural approach and, 35–36

Audience, 34Auditory guilt, 744Auditory nuclei, 718Auditory stimuli, 160–161, 366, 721–722Auditory thalamus, 162fAutism, 222, 323, 359, 692–693tAutobiographical memory, 605–607Automacity, 632–633Automatic emotion regulation, 498, 506–

507Autonomic nervous system

affect and, 51, 578–579appraisal and, 584elicitation and, 306emotion and, 575emotional states and, 307expression and, 311facial expressions and, 220ffear conditioning and, 162ffear/anxiety and, 723fight-flight response and, 663–665intergroup emotions and, 432interoception and, 272–273positive approach and, 670psychology and, 76–77psychophysiology and, 181–186, 190

Autonomy, 341–342, 422, 476, 479, 744–745, 763

Availability heuristic, 648Avoidance

central nervous system and, 189disengagement and, 668disgust and, 763emotion awareness and, 90–91expressive development and, 294facial expressions and, 224fear and, 709gender and, 400health behavior and, 646intergroup emotions and, 429neurology of, 165risk and, 564theory of, 89treatment and, 701

Avoidant personality disorder, 692–693tAwareness. see ConsciousnessAwe, 778Axillary odors, 241

B cells, 667Baby FACS, 295Backward masking, 713BAS. see Behavioral activation system

Basal ganglia, 159, 165, 223, 253, 767Basal nucleus, 162fBasal telencephalon, 261Basic emotions, 104, 315, 334, 349f, 350,

410–412, 517, 592–594, 743–744, 754Bed nucleus of the stria terminalis, 55f, 252Behavior

acceptance/rejection and, 69affect and, 49f, 51–54, 58, 629. see also

Affectanger and, 5, 731anxiety and, 722children and, 350, 352, 360collective, 37consciousness and, 171culture and, 410, 414development and, 384–385differential-emotions theory and, 377disengagement and. see Disengagementdisgust and, 758economic. see Economic behavioremotion and, 498emotion as cause of, 630–631emotion elicitation and, 78emotion instances and, 74emotional-expressive, 339–341evolution and, 119–120, 127–128, 131expression and. see Expressionfacial expressions and, 221, 224, 226fear and, 135n, 722groups and, 429health and. see Health behaviorhistorical context of, 24intergroup emotions and, 434–435interoception and, 273, 278, 284–285personality and, 630positive emotions and, 778–779, 784–785prosocial, 61psychology and, 69, 71, 79regulation of, 250–251self-reporting and, 397temperament and, 486–487, 491treatment and, 701

Behavioral activation system, 478Behavioral economics. see Economic

behaviorBehavioral history, 29Behavioral neuroscience, 52–53Being and Nothingness, 9Beliefs, 13, 50, 322–326, 355, 359–360,

429, 498, 505–506, 575Benefit/cost analyses. see Cost/benefit

analysesBereavement, 692–693t. see also GriefBerlin Aging Study, 381–382Bhava, 68–69Bias

affect and, 566anger and, 736anxiety and, 716–717behavior and, 779development and, 386differential-emotions theory and, 378health behavior and, 648–651intergroup emotions and, 430, 433, 435–

436memory and, 607–608negativity, 185projection, 145self-reporting and, 398

Bioecology, 332, 343. see also Social contextBioevolutionary approach, 306Biological model, 13–14, 43, 79Biology. see also Neurology; Physiology

affect and, 48emotion and, 70–71emotion concepts and, 591environment and. see Bioecologyevolution and, 119, 126, 132–133fear and, 135nfight-flight response and. see Fight-flight

responseimmune system and. see Immune systemtemperament and, 486

Bipolar disorder, 699

Blame, 36, 400, 460, 754fBlindsight, 83, 720Bliss, 7Blood. see Bodily functionsBlood pressure, 163, 184–185, 504–505,

579, 583, 665Blood-oxygenated-level-dependent signals.

see BOLD responsesBlunting, 692–693tBlushing, 220fBNST. see Bed nucleus of the stria terminalisBodily functions, 745, 760–762, 764f, 765Bodily-homeostatic affects, 48Body language

communicative theory of emotions and,103–104

culture and, 414development and, 384–385dramaturgical approach and, 34emotion and, 631emotional states and, 307facial expressions and. see Facial

expressiongender and, 399, 401infants and, 366posture and. see Posturesociology and, 33f

Body odor, 237–238, 240–242. see alsoOlfaction

BOLD responses, 168–169, 263nBonding. see AttachmentBorderline personality disorder, 692–693t,

698Boredom, 80, 538Brain. see NeurologyBrain emotional affects, 48Brainstem, 55f, 250, 251t, 253, 257, 259,

261, 271, 277Breast cancer screening, 646–650t, 653f,

656–657Breathing, 93, 120Broaden hypothesis, 783–786, 785fBroaden-and-build theory of positive

emotions, 777, 782–783, 783fBrown v. Board of Education, 451Buddhism, 409, 414–418Build hypothesis, 783, 786–789Built-In-Critics, 620fBulimia nervosa, 699Bullying, 325

CAD triad, 763Calculation, 562–563Calmness, 411, 422Cancer, 646–650t, 652–653, 653f, 656–657,

679Candide, 81Cardiorespiratory activity, 278, 402, 665,

778, 788, 791Cartesian view of emotion, 10, 60Categorization, 307, 357–358, 514, 517–

518, 523, 550–557, 593, 632–633Catholicism, 28, 41Caudate, 253Causation, 27, 70, 75, 447–448, 472, 524,

651, 677–678, 677fCB. see CerebellumCBT. see Cognitive-behavioral therapyCC. see Corpus callosumCE. see Central amygdalaCentral amygdala, 162fCentral gray, 162fCentral nervous system, 188–190, 668, 671.

see also NeurologyCentromedian nuclei, 257Cerebellum, 55f, 251t, 258–259Change, 17–18, 21, 26–28, 90, 94–97Channels, 310Character, 423t, 539Cheerfulness, 17–18, 21, 24, 423tChemosignals, 237–238Child Language Data Exchange System, 353CHILDES. see Child Language Data

Exchange SystemChildren. see also Infants

834 Subject Index

2-year-olds and. see 2-year-oldsabuse and, 752–754affect and, 56, 62, 579anger and, 730–737appraisal and. see Appraisalcomplexity and, 523, 525conclusions children and, 328–329, 360–

361development and. see Developmentdisgust and, 765disorders in, 692–693telicitation and, 305–307emotional awareness and, 320emotional scripts and, 322–326emotional states and, 309emotional understanding of, 348–350empathy and, 443, 445–446evolution and, 121, 130expression and, 310–311expressive development and. see

Expressive developmentfacial expressions and, 221–222, 226historical context of childrearing and, 20–

22, 28–29individual differences regarding emotional

understanding and, 326–328music and, 104overview of emotional understanding of,

348positive emotions and, 782preschoolers and. see Preschoolerstemperament and. see Temperamenttoddlers and. see Toddlersverbalization of emotions and, 320–322

Chimpanzees. see PrimatesChoice, 14, 144–147, 564, 651. see also

Decision making; Economic behaviorChristianity, 6, 17Chronometry, 698–699Cigarette smoking, 678, 763Cingulate gyrus, 159, 274Circadian rhythm sleep disorder, 692–693tCircuit theory of emotions, 159Circular causation, 70Circulation, 116, 119Circumplex model of emotion, 349fClamminess, 220fClarity, 400Class system, 37Classification, 43, 307, 577. see also

TaxonomyCleanliness, 424CM. see Centromedian nucleiCoaching, emotion, 98–99, 734Coalitional aggression, 115–116Coercion. see ControlCognition

actions and, 72affect and, 47–48, 49f, 51–54, 60, 550–

557, 564–568amygdala and, 169–171anger and, 5, 664appraisal and, 71, 104, 631–632. see also

Appraisalbasis of emotions and, 14change and, 97children and, 322, 360cognitive-affective theory and, 380–381complexity and, 519decision making and. see Decision makingdisengagement and, 669disgust and, 765dramaturgical approach and, 34elicitation and, 306–307, 743embodied, 595emotion and, 8, 70, 81–82, 166–167, 540–

541, 575, 745, 778emotion awareness and, 90–91emotion concepts and, 587emotional intelligence and, 534–537, 535f,

539emotional states and, 307–310empathy and, 441evolution and, 124, 128experience and, 312–314

expressive development and, 294health behavior and, 651–652, 678, 682heuristics and. see Heuristicshypercognizing of emotion and, 9immediate/expected emotions and, 142Islamic thought and, 422–423limbic system and, 160memory and, 609music and, 107nature of, 12olfaction and, 236–237positive emotions and, 780, 784–785, 789psychology and, 69psychophysiology and, 181–182regulation and, 500, 501f, 503–504self-conscious emotions and, 742–743six-level model of mind and, 623f, 624social context and, 335social function and, 456–457sociology and, 33, 33f, 43symbolic-interactionist approach and, 34,

38–40temperament and, 491

Cognitive cuing, 322Cognitive neuroscience, 52–53Cognitive tuning view, 558Cognitive-affective developmental theory,

380–381Cognitive-affective processing model, 651–

652Cognitive-attributional theory of self-

conscious emotions, 745–748Cognitive-behavioral therapy, 89, 94, 789Cognitive-social health information

processing, 645, 652–653, 653f, 655–656

Collaboration, 622Collective behavior, 37Collective effervescence, 40–42Collectivism, 401–402, 473, 476, 479Colonial American period, 21, 23Coloration, 220fComfort-seeking, 296Common-sense model, 651–652Communication and Reception of Affect

Test, 228nCommunication processes

children and, 329culture and, 415, 418evolution and, 119, 124–127expressive development and, 299human affective signaling and, 204infants and, 371–372language and. see Languagemusic and, 283olfaction and, 240–242, 244outcomes of, 203reciprocity and, 336social communication and, 372social context and, 335–336

Communicative theory of emotions, 103–104

Community, 763Comparisons, 479–480Compassion, 80, 96, 341, 424. see also

EmpathyCompetence, 399–400, 459. see also

Emotional intelligenceCompetition, 440Complex decision making, 564Complex emotions. see Emotional

complexityCompliance, 36Computation, 107, 115–116Computers. see TechnologyConation, 70The Concept of Mind, 11Concepts

conclusions, 597–598descriptive models of affect and, 590fdimensions approach and, 588–589embodied-simulation theory and, 595–597emotion as, 68emotion categories and, 597evolution and, 118, 125–126

first-generation representational modelsand, 592, 594–595

overview of, 587–588prototype approach and, 588, 592–593psychology and, 73–74semantic-network approach and, 588,

593–594semantic-primitive approach and, 588–592substance vs. fuction, 74temperament and, 486

Conceptual system for emotion, 517–519Concern, 522Conditioned stimuli, 160–161, 162f, 164–

165Conditioning, 78, 441. see also Fear;

ReinforcementConduct disorder, 359, 736Confidence, 37, 778Conflict, 32, 77, 224, 379, 387, 435, 492,

734Confrontation, 338–339Confusion, 412Congruence, 594Consciousness

affect and, 56, 60–62amygdala and, 719appraisal and, 75, 637behavior and, 629blindsight and, 83children and, 320complexity and, 514culture and, 410development and, 316, 316f, 382, 385disconnection disturbances and, 695disturbances and, 697–698emotion and, 10–11, 90–91emotional signals and, 104emotional states and, 626experience and, 312–313facial expressions and, 225fear/anxiety and, 712–718feelings and, 171, 637–638goals and, 577of God, 425nhistorical context of emotion and, 19interoception and, 282–284language and, 180–181Levels of Emotion Awareness Scale and,

519–521limitations of, 619memory and, 601music and, 107–108olfaction and, 237–239reflection on emotion and, 94regulation and, 506–507of self. see Self-awareness; Self-conscious

emotionssix-level model of mind and, 623f, 624social function and, 459somatic marker hypothesis and, 281vocal signaling and, 205

Consequences, health behavior and, 651Consequentialist models of decision making,

138–139, 139f, 143Consonance, 104Conspiracy, 412Constructivist approach, 18–19, 25–27, 43,

52, 76, 94. see also Psychologicalconstructionist perspective of emotion

Consumerism, 27, 29Contagion, emotion, 205, 397, 434Contamination, disgust and, 757–758, 760–

761, 766Contempt

development and, 385emotion concepts and, 587facial expressions and, 217, 218t, 219grief and, 798groups and, 429health behavior and, 683thierarchical model and, 349fmorality and, 763Sanskrit emotions and, 413self-reporting and, 398social context and, 224

Subject Index 835

Contempt (cont.)social function and, 457, 460–461stereotypes and, 396universal facial expressions and, 213t,

215–216tContent, complexity and, 514Contentment, 200, 316f, 349f, 423t, 477,

782, 784, 790–791Context, 163–164, 226, 332, 334–335, 551–

552, 609, 710–712. see also Socialcontext

Contextual and Affective Sensitivity Test,228n

Contradiction, 622Control

culture and, 342expressive development and, 293fear and, 714gender and, 400–401health behavior and, 651Islamic thought and, 424olfaction and, 242parental, 339preschoolers and, 733regulation and, 502social function and, 460temperament and, 490–491

Control precedence, 72, 74Control shifts, 68–69Conversation, 328–329, 338, 365, 373. see

also Verbalization of emotionsCooperation, 115–116Coping resources

affect and, 549, 557depression and, 800emotion elicitation and, 78emotional intelligence and, 540–543expression and, 91fear/anxiety and, 724gender and, 400grief and, 798–799health behavior and, 678heuristics and, 654fpsychology and, 81regulation and, 502social context and, 338subjective well-being and, 480

Core affect, 76, 274, 431–433, 514, 522,630, 693, 697

Core disgust, 757, 759–761, 764f, 765Core self. see SelfCorpus callosum, 55fCorrugator supercilii, 187, 189Cortex, 159Cortical regions, 251t, 603, 611, 638Corticotropin-releasing factor, 55–56, 722Cortisol, 220, 661–670, 788Cost/benefit analyses, 551–552, 679Couple therapy, 89Courage, 423t, 424Coyness, 412Creative problem solving. see Problem

solvingCRF. see Corticotropin-releasing factorCRH. see Corticotropin-releasing factorCritical responses, 635–636Criticism, 692–693tCritics, emotional states and, 620f, 622Critic-selector model, 621–622, 621f, 623f,

626Critique of Judgment, 8Crohn’s disease, 182–183Crying

children and, 350, 358children’s verbalization of, 321development and, 304, 385as emotional expression, 310empathy and, 444expressive development and, 296, 298–

299facial expressions and, 220fhappiness and, 299social context and, 335

CS. see Conditioned stimuli

C-SHIP. see Cognitive-social healthinformation processing

Cues, 198, 522–523, 562Culture

affect and, 566anger and, 414–418, 731appraisal and, 581classification of emotion and, 43complexity and, 515–516, 520conclusions, 424–425core affect and, 693defining, 32–33disgust and, 757, 759–766disturbances and, 695–696dramaturgical approach and, 34–35elicitation and, 305emotional scripts and, 323empathy and, 451evolution and, 131experience and, 314expression and, 336expressive development and, 291–292,

297facial expressions and, 214, 225gender and, 395, 401–402health behavior and, 684–685historical context of emotion and, 25–26intergroup emotions and, 435–436Islamic thought and, 422–424language and, 324music and, 104–105olfaction and, 243positive emotions and, 781–782, 785–786,

785fpsychology and, 70, 76, 82, 409–410ritual approach and, 40–41Sanskrit emotions and, 411–414semantic-primitive approach and, 591shame and, 418–420social context and, 333–335, 341–342social function and, 457, 463social/moral context and, 421–422sociology and, 33f, 44somatic nervous system and, 187stereotypes and, 396structural approach and, 35–38subjective well-being and, 476, 479–480

Curiosity, 80

Danger avoidance, 119, 284–285, 299Death, phobias and, 711Deceit, 135nDecision making. see also Problem solving

affect and, 548–550, 561–568, 654–655cognitive flexibility and, 550–557complex emotions and, 104conclusions, 568–570consequentialist models of, 138–139, 139fexperience and, 313health behavior and. see Health behaviorrisk and, 140–144, 143f, 151nsadness and, 799

Decoding skills, 399Defense mechanisms, 312, 552, 559, 730.

see also Fight-flight responseDejection, 412Delayed gratification, 539Deliberate emotion regulation, 507Delight, 411, 596Delinquency, 754fDemand-withdrawal patterns of behavior,

336Dementia, 692–693tDemographics, 477, 549. see also CultureDenotative responses, 635Dependency, 54, 89Dependent personality disorder, 692–693tDepersonalization, 428–429Deployment, attentional, 502–503Depression

affect and, 56, 558anxiety and, 723fattachment failure and, 54central nervous system and, 189

cheerfulness and, 18circumplex model of emotion and, 349fcomplexity and, 522, 525descriptive models of, 590fdisengagement and, 666–667disturbances and, 696–697emotional states and, 620evolution and, 129–130, 133facial expressions and, 222fight-flight response and, 665gender and, 400, 404–405glucocorticoids and, 167grief and, 805health behavior and, 678–679, 681, 683theuristics and, 654imaging and, 260negative emotions and, 780olfaction and, 236positive emotions and, 788during postpartum period, 387primates and, 802regulation and, 508sadness and, 800–801self-reporting and, 397sexual abuse and, 752fshame and, 752–753, 752fsocial context and, 340structural approach and, 36as symptom, 692–693ttransdiagnositic approach to, 694transformation of emotions and, 95–96treatment and, 88–89, 699–700

De-repression of emotion, 380Description, 591tDesire, 7, 249, 322–326Despair, 75, 93DET. see Differential-emotions theoryDetachment, 692–693t, 754f. see also

AttachmentDetail, memory and, 607–611Detection, health and, 677f, 678–679Determiners, 591tDeterrence, 128Development

adult models of, 376–381anger and, 731–736appraisal and. see Appraisalcomplexity and, 520–521conclusions development and, 388–389development and, 381–387disgust and, 765–766in early childhood, 316felicitation and, 305–307emotional experiences and, 311–314emotional scripts and, 325emotional states and, 307–310of empathy, 444–447of expression. see Expressive developmentfacial expressions and, 226gender and, 403–404infants and. see Infantsmodel regarding, 314–317overview of, 304–305regulation and, 491–492reintegrating emotional life and, 314social context and. see Social contextsupport for theoretical models of, 387–

388temperament and, 485–486

Diagnosis, 677f, 682, 695–697Diagnostic Analysis of Nonverbal Accuracy

Scale, 228nDiagnostic and Statistical Manual of Mental

Disorders, 691, 696–697, 700, 710Dialectical-constructivist view, 94. see also

Constructivist approachDialecticism, 515–516, 525nDiencephalon, 250–251, 251t, 259, 271Differential-emotions theory, 292–295, 298–

299, 376–378Differentiation, 293–295, 309, 317, 341–

342, 349, 356, 367, 381, 633Diffuse attack, 294Digestion, 119–120

836 Subject Index

Dimensional approach to emotion, 202–203,588

Direct association, 441Directory of Tunes and Musical Themes, 106Disappointment, 339Disasters, 449–450Discernment, 423tDiscipline, 423t. see also ParentingDisclosure, 541–543Disconnection disturbances, 695, 699Discontent, 69Discounted utility, 144–147Discrete systems model, 309Discrete-emotions approach, 202–203, 334,

348–349, 382–383, 514, 682–683Discretion, 423tDiscrimination, 667–668. see also BiasDiscussion. see ConversationDisease. see IllnessDisengagement, 662, 666–667Disgust

affect and, 59animal-nature, 761–762, 764fANS and, 185as basic-level emotion, 350central nervous system and, 188–189children and, 356, 357fcomplexity and, 522components of, 758–759conclusions, 770–771core disgust and, 759–761cultural differences regarding, 766cultural evolution of, 763–765defining, 757–758development and, 316felicitation and, 764f, 770embodied-simulation theory and, 596emotion concepts and, 587evolution and, 115–117, 130facial expressions and, 213t, 215–217,

218t, 219–220, 220f, 224, 226health behavior and, 683thierarchical model and, 349fimaging and, 249individual differences regarding, 766–767interpersonal, 762, 764fmoral, 762–763, 764fneurology and, 767–768overview of, 757pleasure and, 769–770as primary emotion, 43psychopathology and, 768–769Sanskrit emotions and, 411, 422self-reporting and, 397social context and, 224somatic nervous system and, 187

Disgust Propensity and Sensitivity Scale-Revised, 767

Disgust Scale, 766–767Disillusionment, 411–412Disintegrative disorder, 692–693tDislike, 440Disney World, 17Disorganization, 71Display rules, 396, 730, 734Displeasure, 349f, 360, 516–517Disposition, 477–478. see also TemperamentDissonance, 104, 106Distancing, 457, 460–461Distortions, 607Distraction, 312, 400, 503, 541–542, 555–

556Distress

vs. anger, 730circumplex model of emotion and, 349fdescriptive models of, 590fdevelopment and, 316fdisturbances and, 692–693temotional states and, 626empathy and, 446–447expressive development and, 294, 297,

299infants and, 308–309intergroup emotions and, 431

novelty, 490social context and, 334

Disturbances. see also Psychopathologycausal/maintaining processes of, 697–699conclusions, 701constraining emotion/affect and, 691–694descriptive psychopathology and diagnosis

of, 695–697overview of, 691social function and, 457–458as symptom, 692–693ttransdiagnositic approach to, 694–695treatment/evaluation of, 699–701

Divinity, 413, 422, 476, 763Divorce, 476DMT. see Dorsomedial thalamusDopamine, 52, 56–57, 260–261, 568Dorsal pons, 259Dorsal posterior insula, 274, 275f, 277–279Dorsal preoptic area, 55fDorsolateral prefrontal cortex, 169, 171,

280fDorsomedial parafascicular region, 56, 258Dorsomedial thalamus, 55fDoubt, 744–745Dramaturgical approach, 34–35Dread, 592Drive, 71, 624fDrug addiction. see AddictionDrug use. see Substance useDS. see Disgust ScaleDSM-IV-TR. see Diagnostic and Statistical

Manual of Mental DisordersDU. see Discounted utilityDual-aspect monism, 52, 60–61Dual-process models, 149–150, 507, 629,

633–634, 636–637, 803, 805Duchenne laughter, 218tDuchenne smiles, 215–219, 218t, 221–223,

227, 242–243, 372Duration, 651Durga, 413, 419–420Dynamic model of affect, 804Dynamical-systems perspective of expressive

development, 297–300Dysphoria, 397

Eating disorders, 692–693t, 697, 699–700,780

Ecology. see EnvironmentEconomic behavior

conclusions, 149–150decision making under risk and, 140–144,

143fevolution and, 121immediate/expected emotions and, 138–

140intertemporal choice and, 144–147social preferences and, 147–149

Education, disgust and, 767EEA (environment of evolutionary

adaptedness), 120–121, 128, 132–133EEG. see ElectroencephalogramEffect indicator model, 631Effort, 129–130, 506–507EFT. see Emotion-focused therapyEgo, 32, 413, 444–445, 624f, 744Elaboration, 312, 550–551Elation, 200Electroencephalogram, 181–182, 188–189,

223–224, 260, 667Electromyography, 187–188, 243, 397, 759Elevation, 790Elicitation of emotions

anger and, 732development of, 306–307disgust and, 757, 764femotion concepts and, 588emotional scripts and, 323emotional states and, 308expressive development and, 292, 294,

297health behavior and, 684–685overview of, 78, 305–306

role of self and, 743–744scripts and, 348self-conscious emotions and, 742, 745,

746fsocial context and, 334

Embarrassmentculture and, 419development and, 316felicitation and, 742–743as emotion, 11–12emotional intelligence and, 538facial expressions and, 220f, 222, 226health behavior and, 678, 683tself-conscious emotions and, 750–751self-reporting and, 397social function and, 457stereotypes and, 396structural approach and, 36

Embodied cognition, 82Embodied-simulation theory, 595–597Emergence of emotion. see DevelopmentEMFACS. see Emotion Facial Action Coding

SystemEMG. see ElectromyographyEmotion, see also specific emotions

action and, 165–166vs. affect, 578–580, 657naffect and. see Affectamygdala and, 160appraisal of. see Appraisalbasic. see Basic emotionscharacterization of, 123–125coaching and, 98–99cognition and. see Cognitioncommunicative theory of, 103–104competence and. see Competencecomplexity and, 384concepts. see Conceptscontagion, 205, 397, 434culture and. see Culturedefining, 4, 10, 68, 373n, 629, 662dimensions of, 575distinguishing of, 75–77disturbances. see Disturbanceselicitation of. see Elicitation of emotionsepisodes, 74evolution and. see Evolutionary

psychologyexperience and. see Experiencefacial expressions and. see Facial

expressionfunction of, 80–81, 680–681groups and. see Intergroup emotionsheuristics and. see Heuristicshierarchical model of, 349fhistory of. see Historical contexthomeostatic, 274, 279–281, 284–285. see

also Interoceptioninstances, 74–75limitations of, 619maladaptive. see Maladaptive emotionsmanagement of, 34metatheory of emotions and, 619–621mixed/ambivalent, 326modal, 76modal model of, 499fmotivation and, 77nature of, 629–631negative, 780–781, 788. see also Positive

emotionsneurology of, 159–160, 167–171number/types of, 43olfaction and. see Olfactionorigins of, 79–80overview of, 498–499positive. see Positive emotionsprimary. see Primary emotionsprinciples of working with, 90–97programs, 121–123psychology and, 70–73. see also

Psychologyreason and, 14, 73regulation of. see Regulationrole of, 3–4

Subject Index 837

Emotion (cont.)scripts and, 322–326self-reporting and, 397–398six-level model of mind and, 623fsociology and. see Sociologystates of. see Emotional statesstimuli and, 172fstudy of weak, 43–44traits, 387treatment and. see Psychotherapyvocalization and. see Vocalization

Emotion coaching, 98–99, 734Emotion congruence, 594Emotion Facial Action Coding System, 221–

222, 227nThe Emotion Machine, 619–622, 625–627,

627nEmotional Awareness Questionnaire, 523Emotional complexity

complex emotions and, 104, 384emotional intelligence and, 537measurement of, 524overview of, 513–514propositional knowledge and, 519–523self-characterizations of, 523–524self-reported experiences and, 514–519

Emotional granularity. see GranularityEmotional intelligence

conclusions, 543coping strategies/self-regulation and, 540–

543facial expressions and, 225framework for, 535–538, 535fgender and, 399–400information/disinformation and, 534–535measurement of, 538–540overview of, 533–534

Emotional literacy, 539Emotional states

appraisal and, 431“critic-selector” model and, 621–622,

621f, 623fculture and, 410–411development of, 308–310The Emotion Machine and, 625–627emotional intelligence and, 536experience and, 312facial expressions and, 311metatheory of emotions and, 619–621overview of, 307–308, 618–619representation of, 627fsix-level model of mind and, 622–625

Emotion-focused therapy, 88–89, 93–94, 99,700

The Emotions: Sketch of a Theory, 9Emotivism, 9, 14Empathetic overarousal, 442–444Empathy

causal attribution and, 447–448as complex emotion, 104defining, 440development and, 316fdevelopment of, 444–447disturbances and, 692–693temotional scripts and, 325empathetic overarousal and, 442–443future research regarding, 452gender and, 448–449law and, 446–448, 450–452limitations of, 449literature and, 109–110modes of empathetic arousal and, 441–

442moral motivation for, 440–441overview of, 440regulation and, 93socialization and, 448

Encephalization, 277Encoding, 606Endowment effect, 145, 152nEndurance, 423t, 786Energy, 129–130, 198, 411, 431Engagement, 13, 590f, 716Enjoyment, 778

Enlightenment, 7–8, 18Entity theories, 506Environment

appraisal and, 632biology and. see Bioecologyculture and, 415, 417depression and, 800development and, 315evolution and, 120–121, 128, 132–133expressive development and, 296feelings and, 103–104person-environment relationship and, 778positive emotions and, 779psychology and, 69, 76, 81regulation and, 492–493social context and, 335, 344temperament and, 489–493

Envy, 28, 316fEOA. see Empathetic overarousalEpinephrine, 166, 788Equity, 42ERN. see Error-related negativityError-related negativity, 488Errors, forecasting, 144–145Ethical considerations, 5, 9, 150n, 421–422,

763Ethics, 5Ethics of autonomy, 763Ethics of community, 763Ethics of divinity, 763Ethnicity. see CultureEU. see Expected utilityEuphoria, 692–693tEvaluation. see also Judgment

affect and, 559–560, 579appraisal and, 635, 638dimensions approach and, 589disturbances and, 699–701emotion and, 575evaluative judgment and, 12fear and, 714fself. see Self-evaluationself-conscious emotions and, 746–747,

754semantic-primitive approach and, 591ttransdiagnositic approach to, 694

Evaluative need, 182Event coding, 414Evolution

affective adaptations and, 59–60disengagement and, 666disgust and, 759, 763–765, 770–771elicitation and, 306emotional states andexpressive development and, 291facial expressions and, 211–212fear conditioning and, 163fear/anxiety and, 710–712grief and, 801interoception and, 283negative emotions and, 782, 792nolfaction and, 243origins of emotions and, 79positive emotions and, 779psychology and. see also Evolutionary

psychologyEvolutionary psychology

adaptationist foundations of, 119–123anger and, 131–132approach and, 115–116characterization of emotion and, 123–125fear and, 118–119internal regulation and, 130–131mobilization of emotions and, 125–130overview of, 80, 114–115recalibrational emotions and, 132–134theory of emotion and, 116–118

Exchange approach, 34, 41–43Excitement, 42, 71, 80, 95–96, 106, 123,

349f, 590f, 692–693tExercise, 678Expectancy motivation, 559–560Expectancy violation, 299Expectation states theory, 36

Expectations, 5, 316, 333, 714fExpected emotions, 138–140, 148–149,

150nExpected utility, 140–144Experience. see also Feelings

affect and, 804children and, 361complexity and. see Emotional complexitydevelopment and, 387–388development of, 312–314disturbances and, 693emotion and, 498emotional intelligence and. see Emotional

intelligenceof emotions, 82–83vs. expression, 92health behavior and, 653infants and, 371–372meta-analysis of, 268–269music and, 104–105narratives and, 383–384overview of, 311–312vs. perception, 259–260, 263npleasant vs. unpleasant, 260–261psychology and, 82–83self-reporting and, 397stimuli and, 172fsubjective, 217–219symbolic-interactionist approach and, 40

Experimental Psychology, 237Explanation, 338–339Exponential discounting, 145–147Exposure, 91, 750–751Expression

anger and, 731–732, 733–734appraisal and, 575–576, 635arousal and, 90–92children and, 360children’s understanding of, 326complexity and, 518, 522culture and, 336, 401–402, 414depression and, 801development and, 387–388development of. see Expressive

developmentdifferential-emotions theory and, 377disgust and, 759disturbances and, 693elicitation of, 297as emotion, 10emotion as cause of, 630–631emotion concepts and, 587emotion instances and, 74emotional intelligence and, 535fof emotions, 11evolution and, 126–127experience and, 312facial. see Facial expressionfear acquisition and, 168of feelings, 69vs. feelings, 326gender and, 401, 403historical context of, 25infants and, 365–366, 368–370psychology and, 77, 80regulation and, 505sadness and, 801self-reporting and, 397–398social context and, 333, 339–341social function and, 457, 459transformation of emotions and, 97vocal. see Vocalization

The Expression of the Emotions in Man andAnimals, 212, 291, 534, 757

Expressive developmentchildren and, 293, 301differential-emotions theory and, 292–293differentiation theories of, 293–295dynamical-systems perspective of, 297–

300functionalist view of, 296–297future research regarding, 300–301in infancy, 292ontogenetic view of, 295–296

838 Subject Index

overview of, 291–292, 310–311sociocultural internalization model of, 297

Externalization, 400, 402–403, 490, 492–493, 629, 743, 754f

Exteroception, 280fExtinction process. see AnxietyExtraversion, 387, 476, 478Extrinsic nature of emotion regulation, 500Eye movements, 366

F0. see Fundamental frequencyFace recognition, 116Facial Action Coding System, 295, 759Facial Affect Scoring Technique, 214Facial electromyography. see

ElectromyographyFacial expression

anger and, 315, 731ANS and, 183–184, 220fcentral nervous system and, 188–189children and, 350–352, 354f, 355–356,

358children’s understanding of, 326–327communicative theory of emotions and,

103complexity and, 523conclusions, 227culture and, 414development and, 316, 384–385development of. see Expressive

developmentdisgust and, 758–759, 765, 769disorders and, 692–693telectromyography and, 187emotion and, 181, 310, 778emotion as cause of, 631emotion concepts and, 588emotional scripts and, 323emotional states and, 307evolutionary perspective on, 211–212experience and, 313fear and, 715, 719, 721future research regarding, 225–227gender and, 399grief and, 802–803infants and, 333, 365, 368–371language and, 321linkages with subjective experience and,

217–219musculature and, 311as part of a coherent package of emotional

responses, 219–222psychology and, 77research regarding, 215–216tSanskrit emotions and, 411self-reporting and, 397signaling behaviors and, 296social context and, 333–334, 336, 340social function of, 224–225somatic nervous system and, 186–188universality of, 212–217, 213t, 222–224,

226–227FACS. see Emotion Facial Action Coding

System; Facial Action Coding SystemFailure, 746f, 747Fairness, 148, 429False-belief tasks, 355False-consensus effect, 149Families, 27, 240, 327–329, 340, 342, 401,

509, 732–734. see also Attachment;Marriage; Parenting; Social context

FAST. see Facial Affect Scoring TechniqueFDA. see Food and Drug AdministrationFear

acquisition of, 167–168affect and, 53, 657nANS and, 184–185anxiety and, 710, 721–724, 723fappraisal and, 632, 634as basic-level emotion, 350behavior and, 135nbuilt-in critics and, 620fchildren and, 353, 355, 357f, 358children’s verbalization of, 321

complexity and, 517–518, 522conclusions, 724–725conditioning of, 161fcore affect and, 630culture and, 402, 414development and, 315, 316fdifferential-emotions theory and, 377dimensions approach and, 589discrete-emotions approach and, 383disgust and, 768disorders, 723fdisturbances and, 692–693t, 701as emotion, 5emotion categories and, 597emotion concepts and, 587–588emotion elicitation and, 78, 181emotional states and, 308–309, 619evolution and, 118–119, 128–130experience and, 313expression and, 315, 630–631expressive development and, 294, 297facial expressions and, 213t, 215–217,

220, 220f, 222, 226, 310as feeling, 54feelings and, 171–172fight-flight response and, 663–664gender and, 402–403grief and, 798groups and, 429health behavior and, 646–648, 678, 683thierarchical model and, 349fhistorical context of, 20, 29imaging and, 253infants and, 371intergroup emotions and, 431–432introduction to, 709–710language and, 581–582memory and, 611neural circuitry of, 162fneurology and, 718–724neurology of, 160–167olfaction and, 242–243overview of, 710perception and, 169–170positive emotions and, 792nas primary emotion, 43prototype approach and, 592psychology and, 70–71, 79, 84regulation of, 168–169Sanskrit emotions and, 411–413self-reporting and, 397semantic-primitive approach and, 592situational context of, 710–712social function and, 456–457somatic nervous system and, 187speech and, 200stereotypes and, 396structural approach and, 36–37unconscious processes and, 712–718

Fear effector system, 718Fear-drive reduction model, 646Fear-potentiated startle, 721Feces. see Bodily functionsFeedback

autonomic/skeletal response and, 76–77feelings and, 172gender and, 403–405heuristics and, 654positive emotions and, 788psychology and, 75social context and, 334speech and, 199–200temperament and, 491

Feelings. see also Emotion; Emotional states;Experience

actions and, 72–73affect and. see AffectANS and, 182appraisal and, 635Aristotle on, 5bodily/environmental, 103–104children and, 353consciousness and, 637–638control shifts and, 68–69

culture and, 410–411, 414–417emotion and, 10–11, 73evolution and, 130–131vs. expression, 326expressive development and, 301homeostatic, 53–54instinct and, 59intergroup emotions and, 430–431interoception and. see Interoceptionlanguage and, 322limitations of, 619motivation and, 557–558neurology and, 171–172physical. see Physiologypsychology and, 69, 71, 75, 80rules and, 34transformation of emotions and, 96

Feldstein Affect Judgment Test, 228nFFS. see Fight-flight responseFidelity, 121–123, 424Fight-flight response, 119, 242, 478, 578–

579, 662–665Films, 182–184Filtering, 198Financial success, 476, 479First-generation representational models,

592, 594–595Fitness, biological, 1155-HTT, 800Fixed-effects models, 263nFlashbulb memory, 605, 607Flattening, 692–693tFlexibility, 550–557, 567, 782, 785, 789Floral odors, 235–236, 240, 242–244Fluoxetine, 699fMRI, 167–170, 190, 249–250, 257f, 261,

268–271, 273, 402–403, 763. see alsoImaging

Foaming, 220fFocus of attention. see AttentionFood and Drug Administration, 699Foods, 128–129, 421, 757–761. see also

Eating disordersForaging, 116Forbearance, 424Forebrain, 284–285Forecasting errors, 144–145, 151nForgiveness, 96, 424Form, 105–106Formidability, 122, 132, 424Fortitude, 423tFree recall, 583Free-floating anxiety, 12Freezing behavior, 163, 334, 578–579, 732Frequency. see Fundamental frequencyFriendship, 56, 115–116, 558Frowning, 350, 358Frustration, 12, 294, 299, 488–489, 517,

732Fulfillment, 783Function concepts, 74Functional imaging. see fMRI; ImagingFunctional magnetic resonance imaging. see

fMRIFunctional perspective of fear/anxiety, 712–

713Functionalist view of expressive

development, 296–297Fundamental expressions, 127Fundamental frequency, 198–200, 202, 205–

206Fury, 349f, 581

GAD. see AnxietyGains, 36, 150n, 414Gambling, 692–693tGanglia, basal. see Basal gangliaGaze, 226–227, 299, 412, 578–579, 668,

751GCR. see GlucocorticoidsGender considerations

abuse and, 753, 753fanger and, 736complexity and, 516, 520

Subject Index 839

Gender considerations (cont.)conclusions, 405culture and, 401–402disturbances and, 695–696emotional competence and, 399–400empathy and, 448–449etiology of gender differences and, 403–

405facial expressions/nonverbal

communication and, 399health behavior and, 679, 684–685historical context of, 24–25, 27–28neurology and, 402–403nonverbal decoding skills and, 399overview of, 395–396physiological arousal and, 402self-conscious emotions and, 746–747self-reporting and, 396–398situation/relationship specificity and, 401stereotypes/display rules and, 396structural approach and, 38symbolic-interactionist approach and, 38verbalization of emotions and, 398–399

Genderalized anxiety disorder. see AnxietyGeneralization, 42, 402, 411, 684–685Generalized anxiety disorder, 710. see also

AnxietyGenerosity, 424, 558Genetics, 486–487, 722. see also BiologyGentleness, 423t, 424Gestures. see Body languageGist, memory and, 610Global self-attributions, 743, 746f, 747–748Globus pallidus, 253Glucocorticoids, 166, 669Goals

accessing, 97action, 76affect and, 579anger and, 308, 730appraisal and, 576–578, 580–581, 635basic emotions and, 103–104defining, 577development and, 315disgust and, 770elicitation and, 306emotion and, 498emotional experience and, 582–583emotional intelligence and, 542–543emotional scripts and, 324–325emotional states and, 621evolution and, 118, 122, 125facial expressions and, 225Freudian theory and, 624fhealth behavior and, 646memory and, 602motivation and, 77positive emotions and, 779regulation and, 505–506self-conscious emotions and, 742significance of, 575social function and, 456–457socioemotional selectivity theory and, 379subjective well-being and, 478–479

The Go-Between, 109God-consciousness, 425nThe Golden Bough, 759GP. see Globus pallidusGranularity, 516–519, 523Gratification, delayed. see Delayed

gratificationGratitude, 5, 131–134, 152n, 424, 778, 790Greatness, 423tGreek philosophy, 5Grief, 24–25, 54, 80–81, 130, 133, 217,

478, 792n, 797–799, 806Grooming, 217Groups, 32, 446–447, 770. see also

Intergroup emotionsGuilt

children’s understanding of, 326culture and, 421development and, 316fdifferential-emotions theory and, 377elicitation and, 742–743

emotional scripts and, 325–326evolution and, 131–134Freudian theory and, 744grief and, 798health behavior and, 683thistorical context of, 23, 25intergroup emotions and, 430, 432–433psychology and, 80self-conscious emotions and, 317, 746f,

748–749self-reporting and, 398social function and, 459–460, 463stereotypes and, 396symbolic-interactionist approach and, 38

Gustation, 278Gyrus, cingulate. see Cingulate gyrus

Habituation, 69, 161f, 350, 580Handbook of Child Psychology, 730Happiness

affect and, 56ANS and, 185as basic-level emotion, 350central nervous system and, 188–189children and, 350, 353, 355–356, 357f,

358children’s verbalization of, 321complexity and, 516–517crying and, 299culture and, 420development and, 317, 385discrete-emotions approach and, 383elicitation and, 307emotional intelligence and, 542evolution and, 123, 129experience and, 313facial expressions and, 213t, 215–216t,

220f, 226facial expressions/nonverbal

communication and, 399gender and, 403heart rate and, 183hierarchical model and, 349fincome and, 476infants and, 365, 368–369, 371intergroup emotions and, 431language and, 581–582motivation and, 557music and, 107olfaction and, 242–243physiological response and, 662as primary emotion, 43risk and, 563Sanskrit emotions and, 412–413semantic-primitive approach and, 591social context and, 340social function and, 457somatic nervous system and, 187speech and, 200stereotypes and, 396subjective well-being and, 475–476, 478,

480–481Hassam, 76Hatred, 6–7Head position, 226, 299Health

behavior and. see Health behaviorconclusions, 685discrete-emotions approach and, 682–683emotion and, 190emotion elicitation and, 684–685function of emotion and, 680–681literature review and, 676–680mechanisms of emotion/health link and,

681–682mental. see Mental healthnegative emotions and, 780–781olfaction and, 236organizational model of, 677foverview of emotion and, 676physical. see Physical healthpositive emotions and, 783, 788–789

Health behavioraffect and, 654–655cognitive-affective processing and, 651–652

cognitive-social health informationprocessing and, 652–653, 653f

early studies on, 646–647fear and, 647–650t, 656–657heuristics and, 648–651, 653–654integration of affect functions and, 655–

656organizational model of, 677foverview of, 645–646parallel-processing model and, 647

Health belief model, 645Healthy People 2010, 647Heart rate, 116, 183–184, 189, 220, 285,

522, 579, 583Hebrew dietary system, 760Hedge, 591tHedonism, 76, 253, 309, 473, 500, 562, 580Helpfulness, 35, 400, 424, 558Heroism, 411Heuristics, 564–568, 646, 648–651, 653–

654, 784Hierarchical approach, 77, 349fHigh reactors, 579, 583Hinduism, 409–411, 418–420, 422, 759Hippocampus, 159–160, 166, 252, 259,

603–605, 609–610, 669Historical context

development of emotion and, 18–21disgust and, 763–765expressive development and, 291–292key changes of Western society and, 23–25maturation of field of study and, 21–23olfaction and, 236overview of, 17–18philosophy of emotion and, 4–9problems/responses of historical approach

and, 25–27range/theory and, 23strengths of historical approach and, 27–

28The History of Manners, 762Histrionic personality disorder, 692–693tHIV, 677, 680HLA. see Human leukocyte antigenHMS Pinafore, 106Holistic theory of emotion, 9, 12Homeostasis, 273, 276–277, 285Homeostatic affects, 48Homeostatic emotions, 274, 279–281, 284–

285Homesickness, 123Honesty, 423tHope, 93, 581–582, 787fHormone release, 163, 166, 778, 789. see

also specific hormonesHorror, 592Hostility, 664–665, 683t, 798HPA axis, 661–670Hubris, 746f, 749Human affective signaling, 203–204Human Genome, olfaction and, 236–237Human leukocyte antigen, 240Humiliation, 93Humility, 8, 424Humor, 338–339, 411, 769Humours, 6Hunger, 135n, 273–274, 278, 298, 620f, 778Huntington’s disease, 692–693t, 767Hygiene, disgust and, 757, 761–762Hyperactive avoidance motivation, 189Hyperbolic time discounting, 146Hyperventalation, 127Hypoactive approach motivation, 189Hypochondriasis, 692–693t, 768Hypothalamus, 159–160, 250–251, 261,

280f, 661–670

IAT. see Implicit Association TestId, 624fIdeals, 624fIdentity, 24, 38–39, 625–627, 668. see also

Groups; SelfIdeo-affective structures, 378IET. see Intergroup emotionsIfaluk, 323

840 Subject Index

If-Do rules, 621fIFG. see Inferior frontal gyrusIgnoring, 338–339Ikari, 76IL-2. see Interleukin-2 cytokine productionIllness, 5, 129, 711, 760, 777, 787f. see also

HealthImagery, 97, 182–184, 522Imagination, 104, 609Imaging. see also fMRI; Positron emission

tomographyactivation coordinates and, 257fbrain regions and, 251tcomparisons of PET/fMRI studies and,

261experience vs. perception and, 259–260facial expressions and, 223–224fear and, 164–165, 167–168gender and, 402–403interoception and, 273, 278–279, 280flimitations of studies regarding, 262nmemory and, 603meta-analysis of, 254–257, 256t, 268–271neural reference space and, 250–254, 257–

259overview of, 249–250, 261–262pleasant vs. unpleasant experiences and,

260–261positive emotions and, 789subjective feelings and, 280–282

Immediate emotions, 138–140, 142–144,150n

Immune systemdisengagement and, 662, 665–670emotions and, 777fight-flight response and, 662–665olfaction and, 240overview of, 661–662positive approach and, 662, 670–671psychophysiology and, 190summary of, 671

Implicit Association Test, 638Implicit emotion regulation, 507Implicit memory, 601, 611Impression management, 538Improvisation, 107–108Impulsivity, 490, 493, 507, 692–693tIncest, 115–116, 121, 130Incidental emotions, 143–144, 147–149,

152nInclusion, 41Income, 476, 479Indebtedness, 790Independence, types of, 472Indexical cues, 198Indignation, 75, 349f, 413, 423tIndividualism, 401–402, 473, 476, 479Individuals, 34, 429, 461, 516, 766–767Induction, 448Industrialization, 24–25Infants. see also Children

anger and, 731–733appraisal and, 633conclusions intramodal processes of, 372–

373elicitation and. see Elicitation of emotionsemergence of emotion and. see

Developmentemotional states of, 308–309empathy and, 441, 444evolution and, 130experience of emotion and, 371–372expressive development and. see

Expressive developmentgoals and, 577–578multimodal dynamic presentations and,

366–368music and, 104overview of emotional processes of, 364–

365perception of emotion and, 365–366, 371perception of facial-vocal expressions and,

368–370plasticity of, 333regulation and, 489

responses to tactile information and, 370single-modality presentations and, 371temperament and. see Temperamentunderstanding of emotion and, 350–352

Inference systems, 119, 128, 587Inferior frontal gyrus, 253–254Inferotemporal cortex, 720Inflammation, 665Information-processing, 69, 115, 118, 125,

549–550, 654–655, 713. see alsoCognitive-social health informationprocessing

Ingroups. see Intergroup emotionsInhibitions, 188, 490Initiation, 678Innate releasing mechanisms, 307–308INS. see InsulaInsolence, 5Instincts, 51, 59, 624fInstructed fear. see FearInstrumental behaviors, 366Insula, 148, 171, 250, 261, 718–719, 768.

see also Agranular insula; Anteriorinsula; Middle insula

Integral immediate emotions, 139, 148Intelligence, 327, 424. see also Emotional

intelligenceIntensity, 400, 402, 591tIntentionality, 10–13Interaction

exchange approach and, 34, 41–43facial expressions and, 224–225imaging and, 280–282infants and, 366–368psychology and, 70ritual approach and, 34, 40–41sociology and, 33f

Interdisciplinary context, 25Interest

complexity and, 517development and, 315, 316f, 385exchange approach and, 42expressive development and, 299health behavior and, 683tpositive emotions and, 782, 791during postpartum period, 387speech and, 200two-systems approach and, 562

Interferon-a, 670Intergroup emotions

affect and, 431–433appraisal theory and, 431behavior regulation and, 430bias and, 430, 433, 435–436depersonalization and, 428–429differences from individual-level emotions

and, 433disgust and, 770foci of, 430–431group identification and, 429–430, 433–

434implications for emotion theory and, 436–

437implications for study of, 436introduction to, 428regulation of attitudes/behavior and, 434–

435social function and, 461–464social sharing and, 434theory of, 429

Interleukin-2 cytokine production, 664, 670Intermittent explosive disorder, 692–693tIntermodal matching, 350–351Internal regulation. see RegulationInternalization, 402–403, 629, 743Interoception

defined, 273feelings/agency/time and, 282–284homeostatic/emotional asymmetry in the

forebrain and, 284–285imaging and, 278–279, 280foverview of, 272–274primates and, 274–278temperature sensation and, 279–280

Interoceptive cortex, 275f

Interpersonal relations, 458–461, 757, 762.see also Relationships

Interpersonal therapy, 89Interpretation, 312, 410Intertemporal choice, 144–147Intervention, 89–90, 678, 787f, 789Intragroup relations, 462Intrepidity, 423tIntrinsic motivation, 560–561Intrinsic nature of emotion regulation, 500Intrinsic welfare tradeoff ratio. see Welfare

tradeoff ratioIntrospective aspect of emotion, 10Intuition, 82, 562–563Invalidation, 89Investment, 122, 129Involuntary compliance, 36IPT. see Interpersonal therapyIRMs. see Innate releasing mechanismsIrrationality. see ReasonIrritation, 430, 433, 517, 537, 692–693tIslam, 409, 422–424, 423tIT. see Inferotemporal cortexItch, 273–274, 278Iterative model, 638

JACBART. see Japanese and Caucasian BriefAffect Recognition Test

Japanese and Caucasian Brief AffectRecognition Test, 225

Jealousy, 24, 38, 75, 109, 115–116, 457,537

Jorvik Viking Centre, 239Joy

affect and, 56–58, 630appraisal and, 632complexity and, 517culture and, 402development and, 315, 316fdifferential-emotions theory and, 377evolution and, 129expression and, 315expressive development and, 294health behavior and, 683tIslamic thought and, 423tmovement and, 310olfaction and, 239positive emotions and, 781–782, 791as primitive passion, 7psychology and, 79–80structural approach and, 36transformation of emotions and, 95–96universal facial expressions and, 213t

Judgmentaffect and, 562, 655cognition and, 12cognitive flexibility and, 550–551cognitive-affective theory and, 380development and, 317disengagement and, 668emotion and, 5emotion concepts and, 588emotional scripts and, 324experience and, 312facial expressions and, 223, 225morality and, 763music and, 104–105stereotypes and. see Stereotyping

Judicial empathy, 449Justice, 42, 421, 424, 429, 447–448

Kali, 413, 419–420Katharsis, 111Kindness, 558Kinship, 115–116, 130, 133–134Kleptomania, 692–693tKluver-Bucy syndrome, 160Knowledge

complexity and, 519–523dramaturgical approach and, 34emotional intelligence and, 535f, 539emotional states and, 626preschoolers and, 735psychology and, 82

“Kubla Khan”, 108–109

Subject Index 841

LA. see Lateral amygdalaLa Mer, 103Labeling, 76, 353–357, 357f, 404, 653Lajja, 418–420Lamina 1, 276–278, 276fLanguage. see also Verbalization of emotions

appraisal and, 575–576changes in word means and, 26children and, 320–322, 353–355complexity and, 521consciousness and, 180–181constraints on, 581–583culture and, 324defining emotion and, 4development and, 315dramaturgical approach and, 34emotion and, 18–19, 180–181emotion concepts and, 587emotion vocabularies and, 33fempathy and, 441–442first-generation representational models

and, 594goals and, 577good/bad distinction and, 69memory and, 606metaphor and. see Metaphornonverbal communication and, 320–321prototype approach and, 592–593psychology and, 76Sanskrit emotions and, 412semantic-primitive approach and, 591social context and, 337theoretical perspectives on, 361

Latent-trait model, 630–631Lateral amygdala, 162f, 163–164Laughter, 56–58, 205–206, 217–219, 218t,

321, 337, 367, 385, 411, 759Law enforcement, 128Law of hedonic asymmetry, 580Law of similarity, 761Law of sympathetic magic, 761Lay theories, 506Learned associations, 632Learning systems, 119, 128–129, 582–583LEAS. see Levels of Emotion Awareness

ScaleLegal system, 446–448, 450–452Lek, 420LESSNS. see Localized electrical stimulations

of specific neural systemsLevels of Emotion Awareness Scale, 519–

521, 523–524Liberality, 423tLife circumstances, 695–696. see also

Subjective well-beingLife expectancy, 777Life satisfaction. see SatisfactionLifespan emotional development. see

DevelopmentLimbic system

affect and, 50, 53circuit theory and, 159critique of, 159–160economic behavior and, 152ngender and, 402–403imaging and, 257interoception and, 272, 274, 277, 283memory and, 610

Linear oscillator model, 805Listeners. see VocalizationLiteracy. see Emotional literacyLiterature, 103, 108Localist models, 637Localization, 256–257Localized electrical stimulations of specific

neural systems, 49–50, 54, 55f, 60, 62Location, 591tLocomotion, 310Logical positivism, 50–52Loneliness, 398Longitudinal Study of Aging, 387Long-term average spectrum, 198Long-term memory. see MemoryLonliness, 77Loss. see also Grief

affect and, 54, 56, 559culture and, 414economic behavior and, 150n–151nevolution and, 130orientation, 803oscillation/resilience after, 805–806psychology and, 81structural approach and, 36subjective well-being and, 478

Loveaffect and, 54, 56anger and, 5children’s verbalization of, 321culture and, 414emotional states and, 620evolution and, 115–116facial expressions and, 226historical context of, 22, 27Islamic thought and, 424meaning and, 26metaphors regarding, 182positive emotions and, 782, 791as primitive passion, 7psychology and, 74, 80Sanskrit emotions and, 411social function and, 457

Loving-kindness meditation, 786Low reactors, 579, 583LTAS. see Long-term average spectrumLung lang, 414–418Lutheranism, 19Lying, 325

Macrostructural analysis of emotions, 37–38, 40

Magical transformations of the world, 9Magnificence, 423tMaintenance, 678Major histocompatibility complex, 240Maladaptive emotions, 89–90, 700Malaise, 129Mammillary bodies, 160Management of emotion. see RegulationMania, 222, 692–693tManliness, 423tMarriage, 19–20, 22, 399, 401, 476Masking, backward, 713Master and slave as metaphor, 3Mate choice, 116Maturation, 308, 487, 490MAX facial coding system, 293, 295, 299–

301, 350, 730Mayer-Salovey-Caruso Emotional

Intelligence Test, 539–540McDonald’s, 17MD. see Mediodorsal nucleusMDvc. see Medial thalamic relayMeaning, 333, 410, 498, 798. see also

Symbolic-interactionist approachMedia, 17–18Medial frontal cortex, 57, 275fMedial geniculate body, 161Medial prefrontal cortex, 164–165, 250,

253, 522Medial subcortex, 259Medial temporal lobe, 250, 252Medial thalamic relay, 277Mediators, 337–339, 607Medications, 699–700, 789Mediodorsal nucleus, 251Meditation, 34–35, 786, 787fMelody, 106–108Memory

affect and, 548–549, 556complexity and, 518–519, 525conclusions, 611–612consciousness and, 171detail and, 607–611emotional events and, 583emotional influence on, 169–171evolution and, 119, 126experience and, 312expression and, 92glucocorticoids and, 166–167implicit, 611

olfaction and, 237–240overview of, 601–602psychology and, 69quality of, 602–607regulation and, 504self-reporting and, 398stimuli and, 172fsubjective well-being and, 478transformation of emotions and, 97

Menstruation, 240–241, 760Mental health, 222, 329. see also

Disturbances; Health; Health behaviorMental mechanisms, 70, 75Mental states, 625. see also Emotional statesMentalities research, 20, 24Metaphor, 109, 181–182Metapsychology, 51Meter, 106METT. see Micro Expression Training ToolMGB. see Medial geniculate bodyMHC. see Major histocompatibility complexMicro Expression Training Tool, 225Microelectrode recordings, 273Microstructural analysis of emotions, 35–37Middle Ages, 6, 19Middle class, 24Middle insula, 280fMiddlemarch, 110Mimicry, 107, 441, 458Mind, 3Mindfulness, 93, 787fMiranda v. Arizona, 446–447Mirror neurons, 372, 441Mirroring, 297Mirth, 411Misery, 590fMixed emotions, 326Mnemonic processes, 604–605Modal emotions, 76Modal model of emotion, 499f, 509Modeling. see AppraisalModerators, 337–339Modernization, 24Modesty, 423t, 424Modulation, 500, 501f, 504–505Momentary independence, 472Monism, 52, 60–61Monkeys. see PrimatesMood

affect and, 655complexity and, 523defining, 373n, 662emotion and, 9, 74, 181, 778emotional intelligence and, 536–538, 541evolution and, 123, 129–130mood-as-information view, 558music and, 104olfaction and, 239–244positive emotions and, 788psychology and, 75ritual approach and, 41

Mood Awareness Scale, 523Mood disorders, 692–693t, 701Moral context, 421–422, 440–441, 757,

762–763Moral sentiment theory, 9Motility, 13Motivation

action and, 332affect and, 557–561, 565–566appraisal and, 635central nervous system and, 189depression and, 800–801differential-emotions theory and, 292, 377disengagement and, 668emotion and, 540–541empathy and, 440–441evolution and, 118, 125, 130–131exchange approach and, 41expectancy, 559–560expression and, 291expressive development and, 297health behavior and, 655intrinsic, 560–561psychology and, 72, 77–78

842 Subject Index

Sanskrit emotions and, 413social context and, 335, 344

Motive states, 72, 75Motor function, 299mPFC. see Medial prefrontal cortexMSCEIT. see Mayer-Salovey-Caruso

Emotional Intelligence TestMTL. see Medial temporal lobeMultifactor Emotional Intelligence Scale,

539Multimodal perception, 366–368Musculature, 213t, 227n, 299, 377, 397,

778. see also specific musculatureMusic, 102–108, 180, 283–284, 312“My Oedipus Complex”, 109“My wife and mother-in-law”, 186f

NAACP, 451NAC. see Nucleus accumbensNarcissistic disorder, 749Narratives, 93–94, 383–384, 451–452, 582,

627f, 632Narrowing, 357–358National Association for the Advancement

of Colored People, 451National Comorbidity Survey, 694Nativism, 52Natural disasters, 449–450Natural killer cells, 663–664, 667Natural selection. see EvolutionNatural-kinds perspective of emotion, 513Nature, vs. nuture, 79Natyasastra, 409, 411–414, 422, 759Nausea, 758NCS. see National Comorbidity SurveyNeeds, accessing, 97Negative affect. see AffectNegative arousal, 32Negative emotions. see specific emotionsNegativity bias, 185Negotiated exchanges, 42Neocortex, 50, 53–54, 60, 165Neocultural theory, 70Neo-Jamesian theories, 9–11Neologisms, 26Nervousness, 403, 592, 692–693tNetwork models, 637Networks, 33f, 81–82Neural reference space, 249–254, 257–259Neurochemical modulators, 56–57Neuroendocrinology. see Immune systemNeuroimaging. see ImagingNeurology. see also specific structures

affect and, 47–48, 49f, 568, 579anxiety and, 718–724appraisal and, 638brain regions and, 251tcentral nervous system and, 188–189complexity and, 522differential-emotions theory and, 377disgust and, 763, 767–768emotion and, 11, 159–160emotional states and, 307–308emotional systems and, 167–171evolution and, 115facial expressions and, 222–224, 227fear and, 160–167, 162f, 597, 718–724feelings and, 171–172gender and, 402–403imaging and. see Imaginginfants and, 367–368interoception and. see InteroceptionLESSNS and, 49–50memory and, 602–611olfaction and, 237peripheral nervous system and, 182–188positive emotions and, 789psychology and, 69, 76, 79, 82–84psychophysiology and. see

Psychophysiologyregulation and, 505structure of nervous system and, 181temperament and, 486–487typical brain and, 619f

Neurons, counting, 263n

Neuropeptides, 61Neurophilosophy, 11Neuroticism, 382, 387, 476, 767Neutrophils, 667Nicomachean Ethics, 5Nobility, 423tNodes, semantic-network and, 593Noise, 119Non-human primates. see PrimatesNonverbal communication. see Body

languageNonverbal decoding skills, 399Nostalgia, 21Novelty distress, 490Nucleus accumbens, 253, 260Nurture, vs. nature, 79Nutrition, 123

OB. see Olfactory bulbObjective correlative, 109Objects, evocation of feelings and, 69Objects of emotion, 11Obligations, 422Obsessive-compulsive disorder, 692–693t,

699, 711, 768–769OCC. see Occipital cortexOccipital cortex, 258–259Odors. see OlfactionOFC. see Orbito-frontal cortexThe Old Curiosity Shop, 105Olfaction

communication and, 240–242, 244disgust and, 761, 765floral odors and, 235–236, 240, 242–244future research regarding, 245Human Genome and, 236–237hypothesis regarding, 244–245memory and, 238–240overview of, 235–236semiochemicals and, 237–238

Olfactory bulb, 55fOn the Genealogy of Morals, 8On the Passions of the Soul, 6Ontogenetic view of development, 295–296Openness, 225, 785Opiates, 56–57Opinion, 423tOpponent-process theory, 788Opportunity cost, 146–147Oppositional defiant disorder, 692–693tOptimal adjustment model, 803Optimism, 424, 476, 539, 778, 784Optimization and selectivity theories, 378–

380Oral incorporation, 759–760Orbicularis oculi, 187, 189, 213t, 243, 596Orbito-frontal cortex, 159, 171, 223, 253–

254, 260, 280f, 609–610, 669Organicity, 333Organization, 94, 115, 120, 333–334, 517–

518, 542, 550–557, 730Orgasm, 54, 312Orientation, loss/restoration, 803Oscillation, grief and, 802–806Others, symbolic-interactionism and, 38Outcomes. see Consequentialist models of

decision makingOutgroups. see Intergroup emotionsOutputs. see BehaviorOverarousal, empathetic, 442–444Overeating, 678Oxytocin, 56, 59, 223, 789

PAG. see Periaqueductal grayPain

affect and, 59anterior insula and, 148emotion and, 5emotion concepts and, 587emotion elicitation and, 78emotional states and, 625–626emotions and, 777experience and, 312expressive development and, 299–300facial expressions and, 222

imaging and, 258interoception and, 273–274, 277–278, 284psychology and, 71, 75

Panalogies, 626PANAS. see Positive and Negative Activation

SchedulePanic, 54–56, 93, 117, 592, 664, 692–693t,

723fPanpsychic view of universe, 58Papez circuit, 159–160Parahippocampal gyrus, 608Paralimbic cortex, 253–254Paralimbic region, 251t, 261Parallel distributed processing models, 637Parallel-processing model, 645, 647, 654fParanoid personality disorder, 692–693tParasympathetic nervous system, 181, 189,

285, 769Paraventricular nucleus, 669Parenting

affect and, 56, 59, 61anger and, 736attachment and. see Attachmentcaregiver style and, 340, 342complexity and, 517emotional intelligence and, 537empathy and, 448evolution and, 130historical context of, 19–22, 26–29psychology and, 80self-conscious emotions and, 753–754,

754fsymbolic-interactionism and, 38temperament and, 490–493

Parkinson’s disease, 236Paroxetine, 699Partonomy, 591tPassion, 4, 6, 48, 68, 72Passions de l’âme, 68, 83Passive associations, 635Passivity, 68, 71Pathèma, 68Patience, 423t, 424Paxil, 699PCC. see Posterior cingulate cortexPeak density analysis, 254–255Peer interactions. see Social contextPerception

anger and, 736appraisal and, 580–581, 631children and, 351complexity and, 522–523emotion and, 6–7emotion concepts and, 587emotional influence on, 169–171emotional intelligence and, 535f, 536evolution and, 118, 126vs. experience, 259–260, 263nexperience and, 312health and, 677f, 678imaging and, 261immediate/expected emotions and, 142of infants, 365–366, 368–371intergroup emotions and, 430meta-analysis of, 269–270psychology and, 83speech and, 200–203

Perception of Affect Test, 228nPerformance Theory, 411Perfumes, 236. see also OlfactionPeriaqueductal gray, 55f, 250–251, 257,

259–261, 277, 664–665, 718–719Peripheral nervous system, 182–188Peripheral physiology, 498Perseverance, 411, 413Persistence, 539Personal cues, 198Personality, 75, 222, 378, 476–478, 618,

630. see also TemperamentPersonality disorders, 89, 692–694, 697Person-correlations, 516–517Perspective, 442, 555Persuasion, 538pgACC. see Pregenual anterior cingulatePharmacology, 56–57, 60

Subject Index 843

Phase shifts, 298Phenomenological aspect, 377The Phenomenology of Spirit, 8Pheromones, 237–238, 240–242. see also

OlfactionPhilosophical Investigations, 11Philosophy, 4–14Phobias, 82, 96, 164, 508, 692–693t, 710–

712, 712f, 723f, 768. see also FearPhonation, 198Phylogentic origin of emotions, 80Physical arousal. see ArousalPhysical health, 222, 329, 342, 542, 665. see

also Health; Health behaviorPhysiology. see also Biology;

Psychophysiologyanger and, 5anxiety and, 722–724appraisal and, 575, 584, 635central nervous system and, 188–189conclusions, 189–190culture and, 414developmental changes and, 310disgust and, 758–759disturbances and, 693elicitation and, 305emotion and, 6–8, 498emotional states and, 307–308experience and, 312–313expression and, 630expressive development and, 298facial expressions and. see Facial

expressionfeelings/emotion and, 10immune system and. see Immune systeminfants and, 367–368intergroup emotions and, 432introduction to, 180–181“my wife and mother-in-law”, 186folfaction and. see Olfactionoverview of, 181–182peripheral nervous system and, 182–188positive emotions and, 789psychology and, 69

PIC. see Proinflammatory cytokinePick’s disease, 692–693tPiloerection, 220fPitch, 105–106Pituitary function, 250–251, 661–670Pity, 5Planned behavior, 645Plans. see GoalsPlasticity, 333–335, 344Play, 50, 54, 56–58, 115, 123, 217, 550,

733, 782Pleasantness, 589, 590f, 591, 779Pleasure

as accompaniment to emotion, 5affect and, 54children and, 360circumplex model of emotion and, 349fdescriptive models of, 590fdisgust and, 769–770exchange approach and, 42expressive development and, 294, 297hierarchical model and, 349fimaging and, 252infants and, 308–309psychology and, 71Sanskrit emotions and, 413sensory, 778subjective expected pleasure theory and,

651vocalization and, 197

Plessy v. Ferguson, 448Political legitimization, 37Politics, 106Positive affect. see AffectPositive and Negative Activation Schedule,

779, 791nPositive approach response, 662, 670–671Positive emotions. see also specific emotions

appropriateness of, 781–782broaden-and-build theory of, 777, 782–

783, 783f

broadened attention and, 785fconclusions, 791defining, 778–780as distinct from negative, 780–781future research regarding, 789–791intervention results and, 787fmeasurement of, 781oscillation in grief reactions and, 802–806overview of, 777–778short- and long-term effects of, 783–789

Positivity offset, 779, 782Positron emission tomography, 190, 249–

250, 257f, 261, 268–271, 279–280,402–403. see also Imaging

Posterior cingulate cortex, 258, 261Posterior insula, 279, 284Postpartum period, 387Posttraumatic stress disorder, 89, 91, 167,

664, 692–693t, 698–699, 723f, 724Posture, 226–227, 310Potency, 589, 635Poverty, 5Power, 34–37, 40, 42, 89, 413, 460Power-status theory, 36Practice. see PsychotherapyPrayer, 35, 476–477Predator avoidance, 115–116, 128–129, 308Predicates, 591tPreferences, 115–116, 148–149, 577Prefrontal areas, 171, 505, 568, 606, 666Pregenual anterior cingulate, 253, 257Prejudice. see BiasPrereflective processes, 51–52Preschoolers, 348, 356–360, 733–735Prevention, health and, 677–678, 677fPride, 5, 226, 307, 310, 316–317, 396, 430,

742–743, 746f, 749–750, 753f, 778Pride and Prejudice, 103Primary control, 502Primary dimensions of meaning, 589Primary emotions, 89–90, 315, 316f, 743–

744, 754Primary expressions, 127Primary-process affects, 48, 50Primates

disgust and, 758empathy and, 441grief and, 802interoception and, 273–274, 280, 282–

283universal facial expressions and, 213t,

214–217, 223, 227nPriming, 611Principal-agent framework, 150Principle of incompatible responses, 788Prioritization, 124Prisoner’s Dilemma game, 221Prizes. see RewardsProbability, 563–564, 655Probability-weighting function, 143Problem solving, 225, 400, 552–553, 799.

see also Decision makingProblem-focused coping, 502Process analysis, 75Process model, 497–498, 500–501, 501fProcessors, 593Productive exchanges, 42Profile of Nonverbal Sensitivity, 227nProgramming, 69, 121–123Proinflammatory cytokine, 665–669Projection bias, 145Prolactin, 56Proletariat, 32Prominence, 38Propositional knowledge, 519–523Props, 33f, 34Prosocial behaviors, 61Protagonists, 109Protest, 19Protestant Reformation, 17, 27Protestantism, 28. see also Protestant

ReformationProtoconversations, 365, 373Prototype approach to emotion, 588, 592–

593, 593t

Prototypes, 226–227, 297, 300–301, 591t,747

Proust phenomenon, 239, 241–242Proximity, 226Prozac, 699Psycho, 105Psychoanalysis, 744, 761Psychoanalytic variants of symbolic

interactionism, 39–40Psychohistory, 19. see also Historical contextPsychological constructionist perspective of

emotion, 514Psychology

affect and, 60conclusions, 83–84cultural, 409–410defining emotions and, 70–73distinguishing of emotions and, 75–77emotion concept and, 73–74emotion elicitation and, 78emotion instances and, 74–75emotion/cognition and, 81–82emotion/motivation and, 77evolutionary. see Evolutionary psychologyfunction of emotion and, 80–81historical context of emotion and, 19, 25imaging and, 262origins of emotions and, 79–80overview of, 68–69philosophy and, 4psychophysiology and. see

Psychophysiologytask of, 69–70

Psychopathology, 82, 498, 507–508, 695–697, 751–754, 768–769. see alsoDisturbances

Psychophysiology. see PhysiologyPsychotherapy, 88–99, 700Public welfare tradeoff ratio. see Welfare

tradeoff ratioPull function of vocal expression, 205Punishment, 128, 325, 478, 493, 753–754,

754fPure music, 102–103. see also MusicPVN. see Paraventricular nucleusPyromania, 692–693t

Qualia, 12, 75–76, 83, 125, 759Qualifiers, 591t

Race. see CultureRage, 117, 420, 517, 537, 596Random-effects models, 263nRange, 106Range and Differentiation of Emotion

Experience Scale, 523Rank, 36–37Rasadhyaya, 409–414Rasas, 110Rationality, 14. see also ReasonrdACC. see Rostral dorsal anterior cingulateRDEES. see Range and Differentiation of

Emotion Experience ScaleReactions, 71, 623f, 624, 732Reactive emotions, 89–90Readiness, implications of, 72Reality-monitoring, 609–610Reappraisal, 503–504Reason, 3, 14, 19, 73, 80, 82, 307, 424, 587.

see also CognitionRebellion, 70Recalibrational emotions, 132–134Recall, 583Reciprocal exchanges, 42, 336, 404, 692–

693tRecognition ability, 225, 350–351, 587Reconciliation, 338–339Recreations, 26–27Recrimination, 150–151nRecuperative response, 662Reddening, 220fRedirection, 338–339Reductionism, 50–52Reflection, 90, 93–95, 507, 623f, 624–625Reflexes, 128

844 Subject Index

Reflexive responses, 635Reformulation, 621Register, 106Regret, 141–142, 150–151n, 317, 459, 746fRegulation

anger and, 131–132, 338–339, 730–735children and, 352children’s understanding of, 326cognitive-affective theory and, 380complexity and, 515conclusions, 508–509culture and, 414defining, 499–500development and, 386–387, 491–492disturbances and, 698–699emotion, 90–93emotional intelligence and, 535f, 537–538,

540–543vs. emotions, 487–488environment and, 492–493evolution and. see Evolutionary

psychologyof fear, 168–169future research regarding, 505–508gender and, 400health behavior and. see Health behaviorheuristics and, 653–654, 654fintergroup emotions and, 430, 434–435introduction of, 497–498olfaction and, 241oscillation in, 803–805overview of, 498–500social adaptation and, 488–489social context and, 334, 343socioemotional selectivity theory and, 379strategies for, 500–505, 501ftemperament and, 343treatment and, 701

Reinforcement, 51–52, 60, 78, 625Reinstatement, 633Rejection, 69, 71, 667–668, 692–693t, 735–

736Relapse, 678Relational cohesion, 42Relational-functionalism, 335Relationships. see also Social context

emotion coaches and, 98–99emotional intelligence and, 537, 539–540evolution and, 115–116, 132–134facial expressions and, 224gender and, 401, 403–404group identification and, 433–434Islamic thought and, 424marriage and. see Marriagepeer, 327positive emotions and, 787fsocial context and, 337–339socioemotional selectivity theory and, 379therapeutic, 96–97

Relativity, 14Relaxation, 590fRelief, 581–582Religion, 8, 17, 19, 27–28, 34–35, 40–41,

409–420, 422–424, 423t, 476–477, 759Remission, 423tRemorse, 349fRepresentational approaches to emotional

expression, 204–205Representational format, 514Representations, 654f. see also Concepts;

SymbolismRepression, 38Reproduction, 115The Republic, 4Repulsion, 349fResentment, 37, 429Resilience, 783, 788–789, 805–806Resonance, 373Resources, 33f, 41–42, 78, 115, 627nRespect, 424Response, 499f, 578–580, 635–636Response hierarchy, 310Responsibility, 746–747, 780Restoration orientation, 803Retention of standards/rules, 316f

Revenge, 80Revolution, 37Rewards, 42, 54, 60–61, 78, 143, 152n, 164,

253, 493Rhetoric, 5Rhythm, 105–106Rhythmic synchronization, 41Rights, 421Risk, 140–144, 143f, 151n, 560–561, 651–

657, 657n, 752f, 762Ritual approach, 34, 40–41Roe v. Wade, 451Role playing, 33fRoles, gender. see Gender considerationsRoman philosophy, 5Romanticism, 8Rostral dorsal anterior cingulate, 253, 257Routines, 567Rule-based processing, 633Rules, 316f, 325, 422–424, 742. see also

Display rulesRumination, 400, 503, 541–543, 800

Sacrifice, 24, 424Sadness

affect and, 54–56, 55fanger and, 735ANS and, 183–185as basic-level emotion, 350children and, 353, 355–356, 357fchildren’s verbalization of, 321complexity and, 516–518conclusions, 806core affect and, 630culture and, 414depression and, 800–801development and, 316f, 317, 385differential-emotions theory and, 377discrete-emotions approach and, 383elicitation and, 307empathy and, 104evolution and, 123expressive development and, 299–300facial expressions and, 213t, 215–217,

218t, 220f, 226facial expressions/nonverbal

communication and, 399as feeling, 54function of, 799–800gender and, 404grief and, 797–799, 801–802health behavior and, 679, 682, 683thierarchical model and, 349finfants and, 365, 368intergroup emotions and, 432language and, 315, 581–582music and, 107neurology and, 223overview of, 797as primary emotion, 43as primitive passion, 7psychology and, 80Sanskrit emotions and, 411self-reporting and, 397semantic-network approach and, 594semantic-primitive approach and, 592social context and, 340social function and, 457, 460somatic nervous system and, 187speech and, 200stereotypes and, 396structural approach and, 36subjective well-being and, 478transformation of emotions and, 96

Salience heuristic, 648Salivation, 220fSanskrit, 409–414Sarcasm, 336Satisfaction, 36, 42, 787f. see also Subjective

well-beingSchemas, 69, 81–82, 90, 96, 139f, 565–566Schizophrenia, 222, 692–693t, 697–699Science, 4Science, 183Scorn, 5

Screening, health behavior and, 677f, 678–679, 681, 684. see also Cancer

Scripts, 322–326, 348, 358–359, 593tSCRs. see Skin responseSDO. see Social dominance orientationSecondary emotions, 89–90Secrecy, 412Seeking urges, 57Selective serotonin reuptake inhibitors, 700Selectivity, 378–380Selectors, 621f, 623fSelf, 7, 33f, 38, 40–43, 61–62, 743–744Self-awareness, 313, 316–317, 697–698. see

also ConsciousnessSelf-blame, 338–339Self-comforting. see Self-soothingSelf-concept, 379–380Self-conscious emotions. see also

Embarrassment; Guilt; Pride; Shamecognitive-attributional theory of, 745–748conclusions, 754defining, 744–745development and, 317embarrassment/shyness and, 750–751emotional states and, 624, 626expression and, 310guilt and, 748–749hubris and, 749overview of, 742–743pride and, 749–750psychopathology and, 751–754role of self and, 743–744self-reporting and, 397shame and, 748structural model for, 746f

Self-control, 24–25, 424Self-criticism, 89, 96Self-effacement, 341Self-empathy. see EmpathySelf-esteem, 75, 476Self-evaluation, 307–308, 423t, 746–747,

754Self-injury, 780Self-interest, 8, 148–149Self-management, 414–415, 417–418Self-meaning, 38–39Self/performance focus style, 747–748Self-preservation theory, 668Self-regulation. see RegulationSelf-reporting, 396–398, 514–519, 538–539Self-soothing, 93, 343Semantic function, 237, 239, 627fSemantic-network approach to emotion,

588, 593–594Semantic-primitive approach to emotion,

588–592, 591tSemiochemicals, 235, 243–245. see also

OlfactionSensation. see Physical arousalSensitivity, 493, 508, 732–733, 766–767Sensory affects, 48Sensory cortex, 163fSensory processing systems, 171, 172f, 279–

280, 720, 778–779. see alsoInteroception; Olfaction

Sensory thalamus, 163fSentiments, 73, 77Separation, 6, 54–56Septal nuclei, 159Sequential nature of appraisal, 635–636Serenity, 349f, 411, 413, 422Serotonin, 487, 800Sertraline, 699SES. see Socioeconomic statusSET. see Social evaluative threatSet switching, 789Sex differences. see Gender considerationsSexual abuse, 752Sexuality

affect and, 59, 61cultural taboos and, 421disgust and, 758evolution and, 115–117, 121–123, 127–

131, 133experience and, 312

Subject Index 845

Sexuality (cont.)facial expressions and, 217historical context of, 25literature and, 109music and, 104olfaction and, 237–238, 240–241positive emotions and, 790psychology and, 80Sanskrit emotions and, 411–413sensory pleasure and, 778

sgACC. see Subgenual cingulateShame

culture and, 418–420development and, 316f, 385differential-emotions theory and, 377discrete-emotions approach and, 382disengagement and, 667–670EFT and, 89elicitation and, 742–743embarrassment and, 11–12emotional states and, 308, 625Erikson on, 744–745evolution and, 133facial expressions and, 222, 226Freudian theory and, 744gender and, 403health behavior and, 683thistorical context of, 23, 25Islamic thought and, 423tlanguage and, 315maltreatment and, 753–754, 753fpsychology and, 76, 80psychopathology and, 751–754regulation and, 93Sanskrit emotions and, 413self-conscious emotions and, 310, 317,

746f, 748sexual abuse and, 752fsocial function and, 459stereotypes and, 396structural approach and, 36–37symbolic-interactionist approach and, 38transformation of emotions and, 96trauma and, 751f

Sharing, 542Shock, 71, 129, 144, 349fShort-term dynamic therapy, 89Short-term memory. see MemoryShyness, 382, 750Siamangs. see PrimatesSignaling, 125, 296, 386, 534–535. see also

Vocal signalingSimilarity-difference judgment, 550–551Simplification, 621–622Simulation, 622Simulators, 595Sin, 17Sincerity, 424Single-modality presentations, 371Single-network model, 637SIP. see Social information processingSituated constructions, 629–630Situational cognitive appraisal, 38, 631–632Situational considerations, 123, 326, 499f,

500–502, 501f. see also ContextSiva, 419–420Skeletal response, 76–77Skin response, 184, 712f, 715Slave and master as metaphor, 3Sleep, 116, 590f, 692–693tSlights, 5Small-diameter afferents, 275Smelling. see OlfactionSmiling

behavior and, 224children and, 350, 358children’s verbalization of, 321development and, 385disappointment and, 339expressive development and, 294, 298–

299facial expressions and, 218t, 221–222gender and, 404–405infants and, 367, 371–372olfaction and, 242–243

social context and, 335, 337social function and, 459universal facial expressions and, 215–219,

218tvariants of, 227

SN. see Substantia nigraSNS. see Sympathetic nervous systemSocial anxiety disorder. see AnxietySocial communication, 372Social context

affect and, 61anger and, 5, 12, 730, 732–736attributes of, 334–337children and, 327cognitive-affective theory and, 380communicative theory of emotions and,

103–104complexity and, 521conclusions, 343–344, 464–465constructivist theory and, 27culture and, 341–342, 414–415, 418, 420–

422disengagement and, 668disgust and, 770dramaturgical approach and, 34economic behavior and, 147–149emotional intelligence and, 537, 539–540emotional states and, 310expression and, 310, 339–341expressive development and, 297facial expressions and, 224–225gender and. see Gender considerationsgroups and, 461–464. see also Intergroup

emotionsinfants and, 371–372interpersonal relations and, 458–461micro-level bias regarding, 44olfaction and, 243overview of, 332, 456–458phobias and, 711, 723fplay and, 57positive emotions and, 783, 785–786psychology and, 69–70, 75regulation and, 488–489relationship processes and, 337–339sadness and, 799significance of, 333–334socioemotional selectivity theory and, 379sociology and, 33fstereotypes and, 396structural approach and, 34–38subjective well-being and, 479–480symbolic-interactionist approach and, 38temperament and, 342–343, 493

Social distancing function of emotion, 457,460–461

Social dominance orientation, 435Social dysfunction of emotion, 457–458Social evaluative threat, 668–669Social information processing, 736Social Interpretations Test, 228nSocial learning, 168Social movements, 37Social self-preservation theory, 668Social Skills Inventory, 227nSocialization, 21, 309–310, 314, 448, 732,

734–735, 737Social-structural theory, 37–38. see also

Structural approachSociocultural internalization model of

expressive development, 297Socioeconomic status, 476, 479, 492, 685,

767Socioemotional selectivity theory, 379Sociology

dramaturgical approach to, 34–35exchange approach to, 41–43overview of, 32–34problems/future prospects regarding, 43–

44ritual approach to, 40–41structural approach to, 35–38symbolic-interactionist approach to, 38–

40Solidarity, 41

Somatic experience, 307, 414–417, 522,653–654

Somatic marker hypothesis, 53, 172, 281,763

Somatic nervous system, 181–188Somatoform disorders, 692–693tSomatosensory cortex, 162fSomatosensory representation, 53, 56Somatovisceral illusions, 185–186Sorrow, 349f, 411, 413Soul, 4, 423tSound, 203–205. see also VocalizationSpecific self-attributions, 743, 746f, 747–748Speech. see Verbalization of emotionsSperm competition, 122Spite, 5SRG. see Standards, rules, goalsSSRIs. see Selective serotonin reuptake

inhibitorsStandards, rules, goals, 742–743, 745–748,

746fStartle, 349f, 579, 721States, emotional. see Emotional statesState-Trait Anger Expression Inventory, 339Statistical inference, 255–256Status, 33f, 34–37Stealing, 325Stereotyping, 202–203, 396, 399, 436, 481,

516, 520, 549, 565–566, 784Steroids, 241STG. see Superior temporal gyrusStigmatization, 667–668Stimuli. see also Touch

affect and, 559–560appraisal and, 634complexity and, 522conditioned. see Conditioned stimulicorrection of internal troubles and, 778defining, 601development and, 387disgust and, 769elicitation and, 305emotion and, 629–630emotion elicitation and, 78as emotional elicitors, 305emotional states and, 307–308experience and, 172f, 312–313facial expressions and, 187, 214, 222–223fear acquisition and, 167fear and. see Fearfear processing and, 160–161, 161f, 163ffear response and, 169grief and, 798health behavior and, 647heuristics and, 654finfants and, 365–366memory and, 602, 606, 609–611olfaction and, 243–245psychology and, 70–71, 79, 82psychophysiology and, 189regulation and, 487speech and, 201subjective well-being and, 478tactile, 370unconditioned. see Unconditioned stimuli

STN. see Subthalamic nucleusStoics, 5, 7Stonewalling, 399Stories. see NarrativesStrength. see FormidabilityStress

central nervous system and, 188emotional intelligence and, 541–542expression and, 311facial expressions and, 220fight-flight response and, 665health behavior and, 645hormones and, 166language and, 581–582medial prefrontal cortex and, 165memory and, 606positive emotions and, 787–788, 791social context and, 334as symptom, 692–693tverbalization of emotions and, 399

846 Subject Index

Stress-illness heuristic, 653–654Stria terminalis, 55fStriatum, 253Structural approach, 34–38STS. see Superior temporal sulcusSubcortical systems, 53, 61, 163, 251t, 252–

253, 271, 603, 634–635, 720–721, 798Subgenual cingulate, 253, 257Subjective essentialism, 10Subjective expected pleasure theory, 651Subjective well-being

causes of, 477–479correlates of, 476–477culture and, 479–480defining, 471–473emotion and, 498facial expressions and, 217–219future research regarding, 481happiness and, 475–476, 480–481measurement of, 473–475overview of, 471

Submission, 460, 491, 667–670Substance concepts, 74Substance use, 222, 539–540, 678, 692–

693tSubstantia nigra, 59Substantives, 591tSubthalamic nucleus, 257Success, 746f, 747Suffering, 69, 626Suicidality, 222, 780Superego, 624f, 744Superior colliculus, 368Superior temporal gyrus, 258Superior temporal sulcus, 258Suppression, 400, 505Surprise

attractor states and, 300as basic-level emotion, 350children and, 356, 357fcomplexity and, 522development and, 315–316, 316fexpressive development and, 299facial expressions and, 213t, 215–216t,

226health behavior and, 683thierarchical model and, 349fas primary emotion, 43Sanskrit emotions and, 413social context and, 334somatic nervous system and, 187stereotypes and, 396

Survival. see Evolution; Evolutionarypsychology

Swann’s Way, 238–239SWB. see Subjective well-beingSweating, 220fSymbolic-interactionist approach, 34, 38–40Symbolism, 81–82, 93–94, 410, 414–415,

418, 581–583, 626, 633. see alsoMetaphor

Symmetry rule, 653Sympathetic nervous system, 181, 189, 275,

284–285, 663, 759Sympathy, 80, 110, 226, 321, 400, 421, 445,

801The Symposium, 4Symptom sensitivity, 677f, 678Symptoms, emotion-based, 692–693tSynapses, counting, 263nSynchronization, 73Systematic desensitization, 788Systems of interaction, 332

T cells, 667Taboos, 421, 624f, 760Tactile stimulation. see TouchTaqwa, 424, 425nTask focus style, 747–748Taxonomy, 5, 48–49, 76, 336–337, 348,

355–356, 591t, 695–697TC. see Temporal cortexTearing, 220f, 335, 412, 762TEC. see Test of Emotion ComprehensionTechnology, 94, 107, 655

Teeth, 291, 366, 646Telencephalon, subcortical, 252–253Telic theories, 478Temper, 26Temperament, 6, 222, 333–334, 342–343,

395, 485–494, 733–737. see alsoEmotional intelligence; Personality

Temperance, 423tTemperature control, 184, 273–274, 276f,

279–280, 623fTempo, 106Temporal cortex, 253–254Temporal synchrony, 368–369Temporary nature of positive emotions, 786Tenderness, 423tTension, 297Terror, 349f, 411Terrorism, 37Test of Emotion Comprehension, 326Test of Emotional Styles, 228nTesting, 161fTexture, 367Thalamo-cortical relay nucleus, 277Thalamus, 56, 159, 162f–163f, 259, 718Theory theory, 360–361Therapy. see also Psychotherapy

avoidance theory and, 89children and, 328cognitive-behavioral therapy and, 89couple therapy and, 89disturbances and, 700EFT and, 88–89emotion awareness and, 90–91emotional arousal/expression and, 91–92interpersonal therapy and, 89music and, 106reflection on emotion and, 93–94short-term dynamic therapy and, 89therapeutic relationship and, 96–97

Thinking. see CognitionThirst, 273–274, 278, 620f, 778Thornburg v. American College of

Obstetricians, 451Thoughts, 34–35Threats, 414, 431, 654f, 717–718. see also

Disengagement; Fight-flight response;Positive approach response

Thresholding, 255–256Tickling, 57–58Til Eulenspiegel’s Merry Pranks, 103Timbre, 106Time considerations, 144–147, 282–284,

335, 591t, 651, 698–699TNF-Symbol”a. see Tumor necrosis factor-

Symbol”aToddlers, 350–352, 444–445, 731–732Toileting. see Bodily functionsTolerance, 54, 341, 424Tone of voice, 196–197, 358–359, 371–372,

414. see also VocalizationTotems, 32Touch, 54, 226, 278, 367, 370Toxin avoidance, 758Traces, 593Tract tracing, 273Traditionalism, 25Trait Meta-Mood Scale, 523, 538–539Tranquility, 423tTransdiagnostic approach to emotion

disturbances, 694–695Transduction principle, 594–595Transformation, 90, 95–97, 783Transgressions, 325Transient nature of positive emotions, 786Transient rituals, 40Transsituational cognitive appraisal, 38Trauma, 91, 400, 443–444, 449–450, 581–

582, 711, 751–753, 751f, 803A Treatise of Human Nature, 7Treatment, 88–89, 677f, 679–680, 699–701.

see also specific modalities; TherapyTrichotillomania, 692–693tTrigeminal system, 235, 237–238. see also

OlfactionTriggers, 13, 398, 651

Tripartite soul, 4Trustworthiness, 424Truthfulness, 424Tumor necrosis factor-Symbol”a, 665, 669–

670Twinkling, 220fTwins, 336, 475, 7222-year-olds, 348, 353–355, 354f, 517, 731Two-systems approach, 562–563Tyranny, 70

Ultimatum game, 148, 763Uncle Tom’s Cabin, 446–447Unconditioned stimuli, 160–161, 162fUnconsciousness. see ConsciousnessUnderstanding. see Children; Emotional

intelligenceUnderstanding our Feelings Test, 228nUnits, semantic-network approach and, 593Unity assumption, 71Universality, 14, 79, 104, 212–217, 213t,

222–227, 411–412, 477Unpacking, 151nUnpleasantness, 589, 590fUnvoiced energy, 198Upset, 71Urine. see Bodily functionsUS. see Unconditioned stimuliUsefulness, 552Utility, 140–147, 563–564

V1. see Visual cortexValence

children and, 352, 360classification of emotion and, 577cognitive flexibility and, 551–552cognitive-affective theory and, 381fear and, 714memory and, 602, 606–607overview of, 348, 779positive emotions and, 784psychophysiology and, 181, 189reactions and, 71speech and, 200–201subjective well-being and, 472two-systems approach and, 562vocalization and, 197, 206

Values, 624fVariability, 524–525Variety seeking, 560–561Vengefulness, 37Ventral septal area, 55f, 260Ventral tegmental area, 59, 251, 257, 259–

260Ventromedial prefrontal cortex, 142, 253,

666, 669Verbalization of emotions. see also Language

appraisal and, 575children and, 327complexity and, 517, 521, 525disturbances and, 692–693temotion coaching and, 98–99, 734empathy and, 442expression and, 198–200gender and, 398–399perception and, 200–203sadness and, 799social function and, 460

Veridical associations, 197Veridical empathy, 445–446Vertical causality, 75Victim impact statements, 452Victorian period, 24, 26Violence, 24–25, 115, 121–122, 491, 754f,

780Virtue, 413, 422, 423t, 425nVisceral urgency, 273–274Visual attention, 784Visual cortex, 258–259, 720Visual memory, 608Visual shame, 744Visualization, 256–257Vividness, 601, 610–611VMPFC. see Ventromedial prefrontal cortexVMpo. see Thalamo-cortical relay nucleus

Subject Index 847

VNO. see Vomeronasal systemVNS. see Vomeronasal systemVocabulary. see LanguageVocal acoustics, 196–198. see also Speech;

VocalizationVocal signaling, 197, 203–206. see also

VocalizationVocalization. see also Verbalization of

emotionsaffect induction and, 203–206children and, 352–355conclusions, 206–207development and, 384–385differential-emotions theory and, 293as emotional expression, 310emotion-related vocal acoustics and, 198expression and speech and, 198–200infants and, 366, 368–370language and. see Languageoverview of, 196perception of emotion from speech and,

200–203play and, 57

self-reporting and, 397–398social context and, 334theoretical perspectives on, 196–197tone of voice and. see Tone of voiceverbalization of emotions and, 398–399

Voiced energy, 198Volition, 278Volume, 106Voluntarism, 14Voluntary compliance, 36Vomeronasal system, 235, 237–238, 241. see

also OlfactionVomit. see Bodily functionsVon Economo neurons, 282–283VS. see Ventral septal areaVTA. see Ventral tegmental areaVulnerability, 93, 397, 403

Wanting, 622fWar, 5Wariness, 294Warmth, 403, 459Waste products. see Bodily functions

Ways to Think, 620fThe Wealth of Nations, 8Weariness, 412Welfare tradeoff ratio, 130–131, 134Well-being. see Subjective well-beingWestern society. see CultureWisdom, 423t, 424, 508–509Wit, 423tWithdrawal, 35, 54–55, 96, 221, 284–285,

336, 668, 801, 803WMC. see Working memory capacityWonder, 7, 411, 413Work, 27, 34–35Working memory capacity, 519, 525Worry, 89, 412, 525, 692–693t. see also

AnxietyWorthlessness, 89Wounds, 190WTR. see Welfare tradeoff ratio

Zoloft, 699Zygomaticus major, 187, 189, 213t, 243,

596

848 Subject Index

HANDBOOK OF EMOTIONS - CiteSeerX

Documents

Transcript of HANDBOOK OF EMOTIONS - CiteSeerX