Raffaela Giovagnoli Robert Lowe Editors

Studies in Applied Philosophy,Epistemology and Rational Ethics

Raffaela GiovagnoliRobert Lowe Editors

The Logic of Social Practices

Studies in Applied Philosophy, Epistemologyand Rational Ethics

Volume 52

Editor-in-Chief

Lorenzo Magnani, Department of Humanities, Philosophy Section, University ofPavia, Pavia, Italy

Editorial Board

Atocha AlisedaUniversidad Nacional Autónoma de México (UNAM), Mexico, Mexico

Giuseppe LongoCNRS—Ecole Normale Supérieure, Centre Cavailles, Paris, France

Chris SinhaSchool of Foreign Languages, Hunan University, Changsha, China

Paul ThagardUniversity of Waterloo, Waterloo, Canada

John WoodsUniversity of British Columbia, Vancouver, Canada

Studies in Applied Philosophy, Epistemology and Rational Ethics (SAPERE)publishes new developments and advances in all the fields of philosophy,epistemology, and ethics, bringing them together with a cluster of scientificdisciplines and technological outcomes: ranging from computer science to lifesciences, from economics, law, and education to engineering, logic, and mathe-matics, from medicine to physics, human sciences, and politics. The series aims atcovering all the challenging philosophical and ethical themes of contemporarysociety, making them appropriately applicable to contemporary theoretical andpractical problems, impasses, controversies, and conflicts. Our scientific andtechnological era has offered “new” topics to all areas of philosophy and ethics—forinstance concerning scientific rationality, creativity, human and artificial intelli-gence, social and folk epistemology, ordinary reasoning, cognitive niches andcultural evolution, ecological crisis, ecologically situated rationality, consciousness,freedom and responsibility, human identity and uniqueness, cooperation, altruism,intersubjectivity and empathy, spirituality, violence. The impact of such topics hasbeen mainly undermined by contemporary cultural settings, whereas they shouldincrease the demand of interdisciplinary applied knowledge and fresh and originalunderstanding. In turn, traditional philosophical and ethical themes have beenprofoundly affected and transformed as well: they should be further examined asembedded and applied within their scientific and technological environments so toupdate their received and often old-fashioned disciplinary treatment and appeal.Applying philosophy individuates therefore a new research commitment for the21st century, focused on the main problems of recent methodological, logical,epistemological, and cognitive aspects of modeling activities employed both inintellectual and scientific discovery, and in technological innovation, including thecomputational tools intertwined with such practices, to understand them in a wideand integrated perspective.

Studies in Applied Philosophy, Epistemology and Rational Ethics means todemonstrate the contemporary practical relevance of this novel philosophicalapproach and thus to provide a home for monographs, lecture notes, selectedcontributions from specialized conferences and workshops as well as selected Ph.D.theses. The series welcomes contributions from philosophers as well as fromscientists, engineers, and intellectuals interested in showing how applyingphilosophy can increase knowledge about our current world. Initial proposals canbe sent to the Editor-in-Chief, prof. Lorenzo Magnani, [email protected]:

• A short synopsis of the work or the introduction chapter• The proposed Table of Contents• The CV of the lead author(s)

Formore information, please contact the Editor-in-Chief at [email protected]. Indexedby SCOPUS, ISI and Springerlink. The books of the series are submitted for indexing toWeb of Science.

More information about this series at http://www.springer.com/series/10087

Raffaela Giovagnoli • Robert LoweEditors

The Logic of Social Practices

123

EditorsRaffaela GiovagnoliSocial Ontology and Social EpistemologyFaculty of PhilosophyPontifical Lateran UniversityRome, Italy

Robert LoweDivision of Cognition and CommunicationDepartment of Applied ITUniversity of GothenburgGothenburg, Sweden

ISSN 2192-6255 ISSN 2192-6263 (electronic)Studies in Applied Philosophy, Epistemology and Rational EthicsISBN 978-3-030-37304-7 ISBN 978-3-030-37305-4 (eBook)https://doi.org/10.1007/978-3-030-37305-4

© Springer Nature Switzerland AG 2020This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or partof the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmissionor information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilarmethodology now known or hereafter developed.The use of general descriptive names, registered names, trademarks, service marks, etc. in thispublication does not imply, even in the absence of a specific statement, that such names are exempt fromthe relevant protective laws and regulations and therefore free for general use.The publisher, the authors and the editors are safe to assume that the advice and information in thisbook are believed to be true and accurate at the date of publication. Neither the publisher nor theauthors or the editors give a warranty, expressed or implied, with respect to the material containedherein or for any errors or omissions that may have been made. The publisher remains neutral with regardto jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AGThe registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

The present volume includes the contributions of the workshop “The Logic ofSocial Practices” that took place at the UNILOG2018 World Conference in Vichy,https://www.uni-log.org/vichy2018 and some other invited ones.

Social practices refer to everyday practices which are routinely performed. Theyintegrate different types of elements such as bodily and mental activities, materialartifacts, knowledge, emotions and skills. The theme of the book aims to raise thequestion about the peculiar “logic” of social practices, and it is broadly conceivedso to allow different responses from various research fields. We can investigate therelationship/continuation with natural behavior and what are the characteristics ofhuman social behavior. Moreover, there are important results in the area of AI andcomputational where we find different approaches that aim at reproducing theprocesses underlying social practices.

From a philosophical perspective, the study of social practices originates fromAmerican Pragmatism (Dewey, James, Mead) and it is relevant to a new conceptionof experience. Experience means “practice,” namely the process that characterizeshumans in their continuous relation with their natural and social environment.According to Dewey, to be human means to be engaged in practices. This thoughtis at the center of the so-called practice theory (Schatzki and Reckwitz). It analyzesimportant features of the world we inhabit; insofar, it is routinely created usingtools, language and bodily activities. The nature of social practices is at the centerof the debate on collective intentionality (Searle, Tuomela, Bratman, Gilbert).Collective intentionality represents an important topic in sociology (Durkheim andWeber), and there are important studies also in the philosophical tradition (Scheler,Walther, Sellars). Collective intentionality is a mental capacity necessary to createthe social world. There are forms of it also in non-human animals, but the way inwhich we create habits, rituals, norms and institutions seems to require a specialresearch and also possibilities of criticism.

The present volume investigates several aspects of social practices, and it isintended in a transdisciplinary way. The first part focuses on the relationshipbetween collective intentionality and social practices. In the ambit of the debate oncollective intentionality, Michael Barber considers the genetic phenomenology

v

introduced by Alfred Schutz. It captures pre-theoretical, commonsense experienceto show how we-intentionality is a primitive experience in which individual con-sciousness is historically and biographically shaped. Eva Schwarz and JonnaLappainen offer a conceptual investigation of whether, and how, action containingjudgement can be understood as collective. Using an empirical example from thepolice profession, the article tries to capture the complexity of a real situation thatrequires both agency and judgement. The example is analyzed in relation to the ideaof collective judgement, both regarding ongoing debates in phenomenology con-cerning the possibility/need for a we-subject and in social epistemology concerningthe idea of judgement based on social evidence. Philip Larrey makes a comparisonbetween John Searle and Barry Smith about the notions of meaning and truth thatmake social transmission of knowledge possible. By way of contrast withneo-positivism and the analytical tradition as a whole, Barry Smith offers his notionof “truth-makers” which are not merely truth vehicles or conditions of truthfulness,but rather that which makes sentences true or false, in other words, using thegerundials, the being of that which is (in an empirical sense). Furthermore, such anaccount can be seen to dovetail well with Davidson’s account of the “triangular”nature of human natural language, where there is no need of an epistemological orlogical construct which would “hook” language onto reality (or on to a “fact” or onto an “event”).

The second part includes contributions that focus on fruitful relations betweenscience, ethics and social practices. Douglas Marshall argues that human behaviorhas evolved to be adaptive, and therefore adheres to a biological logic that is verydifferent from the logic of instrumental rationality. This thesis has importantimplications for individuals, social science and social policy’s governance. Alongthe line of an evolutionist approach, Paolo Crocchiolo examines the interplaybetween our complex neural-hormonal system and body’s integrative to enhanceour genes’ reproductive potential. The analysis brings to the interesting conclusionthat social media and networks can be seen as “surrogate tribes,” i.e., updatedinstruments typical of our contemporary social communities. Lorenzo Magnanielucidates a kind of “embubblement” that makes human violence invisible to us. Hewould also like to offer to the attention of researchers in sociology, psychology andpsychiatry the main features of the situations in which the awareness of humanviolent acts is in question. In these cases, a process of what can be called“autoimmunity” is at play. He contends the concept of moral bubble can provide anintegrated and unified perspective able to interpret in a novel way many situationsin which morality and violence are intertwined. Starting from an original inter-pretation of Amartya Sen’s “Social Choice Functions,” Gianfranco Basti, AntonioCapolupo and Giuseppe Vitiello propose a model of social and economic systems,based on the dissipative quantum field theory paradigm. The algebra–coalgebraapproach aims at making Sen’s SCF theory effective in its both mathematical andethical components. The recent developments in robotics claim for a philosophicalapproach to the current and future relationship between persons and robots. GiusyGallo and Claudia Stancati aim to look for some hints in the facets of the debate inorder to define the ontology of these entities which in the long run could make

vi Preface

decisions or act autonomously. First, we will question which features human beingsand artificial beings share, focusing on the concepts of “person” and “legal per-sonality,” investigating the philosophical work of thinkers such as Kelsen,Scarpelli, Radin, Wolff and Nèkàm. The urgency of the (philosophical) matter isconfirmed by the concept of “electronic person” proposed in the motion that theCommittee on Legal Affairs presented for a European Parliament resolution withrecommendations to the Commission on Civil Law Rules on Robotics in order toestablish rules to arrange the relationship between robots and human beings.

The third part concerns the notions of “habit,” “routine” and “ritual” as well astheir interplay. Pawel Kawalec investigates research routines by developing anevolutionary approach. He aims at identifying some general patterns in socialpractices. In particular, he shows three mechanisms underlying cognitive dynamicof research routines by using the case of microRNA study. In his chapter, RobertLowe discusses value-based habitual and goal-directed systems as studied in theanimal and human learning literature. It focuses on the means by which these twosystems might interact in knowledge transfer, particularly as it applies to sociallearning. Knowledge is conceived here in terms of types of logic computations asimplemented by neural networks. A discussion of dual-process-type structures inthe brain is provided as well as neural-dynamic implementations thereof and con-siderations for how a perspective of the brain as carrying out logic computationsmight be useful for developing the general cognitive capacities of artificial agents.Anna Kawalec considers rituals as they are constitutive of a wide range of socialpractices. She proposes a conception of ritual that goes beyond traditional patternsand disciplinary boundaries. An original “folk” conception aims to be functional bycrossing philosophy and anthropology. She examines selected categories andconcepts of philosophy and anthropology to substantiate her case: (i) the etymo-logical meanings in Sanskrit, (ii) the Hebrew origins, (iii) the ethnographic facts (onthe basis of Alfred Gell’s research on Muria’s “God playing” and some selectedtheses of cultural and anthropological theory), (iv) Jerzy Grotowski’s idea oftechniques of the performer’s breath and (v) some points from Karol Wojtyła’sphilosophical integrative view of the human person and agency. Habits and ritualsrepresent a relevant field to investigate human behavior also in comparison withnon-human animals and nature in general. Raffaela Giovagnoli moves from veryinteresting perspectives that cross philosophy and neuroscience to establish acontinuity between individual and social habits and to clearly isolate what makeshuman rituals peculiar in the field of social ontology. Therefore, she introduces ageneral and brief discussion on rituals in different disciplines before focusing on theactuality of the Aristotelian notion of habit in relationship with some studies inneuroscience that can be extended to the social context. The perspective ofwe-intentionality introduced by John Searle embeds the dimensions of habits thatground cooperation and is very useful to describe the peculiarity of human rituals.

Rome, Italy Raffaela GiovagnoliGothenburg, Sweden Robert Lowe

Preface vii

Contents

Collective Intentionality and Social Practices

Joint and Individual Intentionality: A Genetic, PhenomenologicalApproach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Michael Barber

Collective Phronesis? An Investigation of Collective Judgementand Professional Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Eva Schwarz and Jonna Hjertström Lappalainen

From Searle’s Speech Acts to Smith’s Truth-Makers . . . . . . . . . . . . . . . 37Philip Larrey

Science, Ethics and Social Practices

The Biological Logic of Human Action: On the (Considerable)Difference Between “Rational” and “Adaptive” . . . . . . . . . . . . . . . . . . . 49Douglas A. Marshall

Social Media: “Surrogate Tribes”? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69Paolo R. Crocchiolo

Moral Bubble Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Lorenzo Magnani

The Computational Challenge of Amartya Sen’s Social Choice Theoryin Formal Philosophy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87Gianfranco Basti, Antonio Capolupo and Giuseppe Vitiello

Could an Electronic Person Exist? Robots and PersonalResponsibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121Claudia Stancati and Giusy Gallo

ix

Habits, Routines and Rituals

Cognitive Dynamics of Research Routines: Case Studyof MicroRNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133Paweł Kawalec

Habit-Based and Goal-Directed Systems: Knowledge Transferin Individual and Social Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153Robert Lowe

Agency of Breath—Beyond Disciplinary Views on Ritual . . . . . . . . . . . 169Anna Kawalec

From Habits to We-Intentionality: Rituals as Social Habits . . . . . . . . . . 185Raffaela Giovagnoli

x Contents

Collective Intentionalityand Social Practices

Joint and Individual Intentionality:A Genetic, Phenomenological Approach

Michael Barber

Abstract There has been an extensive debate about we-intentionality, with somepositions demonstrating how it is derived from individual intentionality or othersarguing that it precedes and takes priority over individual intentionality. Recent con-tributions have come to recognize that the theoretical discussions ofwe-intentionalitydepend upon pre-theoretical, pre-analytic, and commonsense experiences of jointintentionality, the description of which might avoid some of the polarizations thatthe theoretical discussions cannot escape.One needs to consider the spectrumof posi-tions from the more individualistically inclined treatments (Bratman, Tuomela) tothose attempting to establish the irreducibility ofwe-intentionality (Searle,Gilbert) tothose de-emphasizing individual intentionality (Baier) or striving for a more refinedbalance (Schmid). In the end, Alfred Schutz’s genetic phenomenology, capturingpretheoretical, commonsense experience, can show how we-intentionality is a prim-itive experience in which individual consciousness is at the same time historicallyand biographically shaped. Schutz’s account holds in tension the individual andsocial dimensions of experience and avoids reducing we-intentionality to individualintentionality.

Keywords Joint intentionality ·We-intentionality · Collective intentionality ·Genetic phenomenology ·We-relationship · Biographical situation · History ·Consciousness

1 Introduction

Within analytic philosophy, there has been an extensive debate about the characterof we-intentionality, between positions in which we-intentionality is derivative fromindividual intentionality or in which the former takes precedence over the latter.Several of the discussants in this debate have come to recognize the importance ofpre-theoretical, pre-analytic, commonsense experiences of joint intentionality whose

M. Barber (B)Saint Louis University, St. Louis, MO 63108, USAe-mail: [email protected]

© Springer Nature Switzerland AG 2020R. Giovagnoli and R. Lowe (eds.), The Logic of Social Practices,Studies in Applied Philosophy, Epistemology and Rational Ethics 52,https://doi.org/10.1007/978-3-030-37305-4_1

3

4 M. Barber

consideration might enable one to escape some of the dilemmas and polarizationsengendered in theoretical discussions. Genetic phenomenology,1 particularly thatdeveloped by Alfred Schutz, can show how we-intentionality is experienced as aprimitive experience in which individual consciousness, at the same time histori-cally and biographically shaped, is developed. This genetic account holds in tensionthe individual and social dimensions of human experience as well as the possibilityof a we-intentionality that is not reducible to individual intentionality. The positionson joint intentionality vary from positions that stress the individual origins of we-intentionality (such as Michael Bratman’s and Raimo Tuomela’s) to those that seek,without complete success, to establish the irreducibility of we-intentionality (JohnSearle and Margaret Gilbert). By contrast, other have sought to de-emphasize indi-vidual intentionality (Annette Baier) or achieve a kind of balance (Hans BernhardSchmid) that would bemore fully accounted for through a genetic phenomenologicalapproach. First, I will consider those inclined toward the individual origins (Bratmanand Tuomela) as well as those who strive to uphold we-intentionality, even thoughthey still seem entrapped within a kind of individualism (Searle and Gilbert). Then Iwill turn to those who stress the independence and even priority of we-intentionality(Baier, Schmid).

2 Those Prioritizing Individual Intentionality Over JointIntentionality

For Raimo Tuomela, an individual actor (A) we-intends to perform an action X ifand only if A intends to perform his part of X because he believes that “joint actionopportunities” [1] with others for X obtain, and he believes that there is mutualbelief among those others about the obtaining of those joint action opportunities. Inaddition, the beliefs about the opportunities and about the mutual beliefs of theseothers about those opportunities function as partial reasons for why A intends toperform his part of X. Amember of such a group who intends to perform such a jointaction with others in a we-mode, acts for a collectively constructed group reasonthat is to be found among the group’s constitutive goals, values, and beliefs and thatmakes it possible for group members to think, emote, and act. One typically acts “asa group member” [2] functioning for the group (or adopting a “we-perspective”) insuch awe-mode, though it is also possible for one to participate in a group action (thatis, take up a “we-perspective”) while only typifying oneself as a private person, in

1“Genetic phenomenology” can have several senses. In this paper and in the work of Alfred Schutz,the “we relationship,” is something experienced, first of all, in the everyday life-world that is theoften-forgotten basis that higher-level theories arise out of and often forget and that we all inhabitbefore we undertake theorizing. In addition, it is within that everyday life we-relationship that thechild begins learning about how to live and maneuver in the life-world. In each sense, that of thelife-world and that of the child’s development, one is starting with something already established,the theoretical sphere or the experience of adults, and reconstructing experiences that geneticallypreceded those higher-level activities and that gave them birth.

Joint and Individual Intentionality: A Genetic … 5

the I-mode, thinking “I intend to contribute to the group’s goal.” [2]. In the we-modeof joint attention, one thinks of oneself as doing one’s part in tandem with others inthe group, one feels entitled to expect of others that they do their part, one expectsto be exposed to critique or to criticize others for any arbitrary withdrawals from thegroup action, one believes that these commitments are public and mutually known,one monitors others as they monitor oneself, and those engaged in pursuing in thewe-mode a collective goal express their collective acceptance of the group goal as“Our goal is p” [2] or “We will achieve p.” [2]. All these interactional entwinementscontrast with group dynamics when group members typify themselves as “privatepersons” with only aggregated private commitments [2].

Given the depth of involvement and the intersubjective intricacy that we-modejoint action calls for, in contrast to I-mode, one can see why Tuomela can claimthat his notion of “we” is not reducible to the conjunction “you and I” [2]2 andwhy his theory is “conceptually rather anti-individualistic, even if ontologically itis, if not individualistic, at least interrelationistic and eschews collective agents andactions in a literal ontological sense” [2].3 Further evidence of his less individualisticorientation is to be found in the fact for those in a we-mode the collective actiondilemma that one might find in highly individualistic prisoners’ dilemma scenariosdo not exist insofar as group members do not “act strategically with respect to eachother but do as the group reason specifies,” [2] as opposed to what might occur in theprivate or I-mode point of view. In addition, Tuomela criticizes Michael Bratman’sanalysis of shared intentions in which an “I” intends that a “we” engage in a jointaction and a “you” intends that “we” engage in this action and in which our subplanshappen to mesh and we have common knowledge of all these factors [3]. In contrastto Bratman’s explanation, Tuomela presents an account of joint action in which Iintend to and would be committed to do my part just as you are, in which both

2Another way in which Tuomela argues that a we-intention differs from and is irreducible to anindividual intention is by distinguishing an “aim-intention” in contrast to an “action intention.”He states, “An aim-intention can be satisfied without the aim-intending agent alone satisfying theintention, whereas in the case of an action intention the agent must believe, if rational at all, thathe can (with some probability) satisfy the intention by means of his own actions.” See Tuomela[1]. Aim-intentions are clearly not reducible to the action intentions of individuals and the “we”is not reducible to “you and I,” nevertheless the way aim intentions are developed depends uponindividuals taking account of, considering, or estimating whether others will be willing to an adoptan end and “do their part” to realize it [1].3Tuomela [1, 2]: One ought not to think that there can be no group, joint action if group membersdo not share a we-intention insofar as Tuomela allows for joint intention in a we-mode sense andan I-mode sense. The We-mode includes a sense of “for-groupness” and acting for a group reason,whereas the I-mode requires only private reasons for cooperation; and in the we-mode the grouphas authority whereas in the I-mode each person has full authority over whatever he or she is doing;in we-mode, the group members are committed to each other as opposed to the “I-mode” in whichthe participant is only committed to herself; in the We-mode actors are willing to do all parts ifothers do not agree but the I-mode would not do the same and instead weighs actions in terms ofindividual costs or private reasons. Tuomela insists that his analysis is conceptually non-reductivesince all these intertwined conditions must be in place as opposed to having individual intentionsmerely aggregated with each other, and he admits that his account is “ontically individualistic or,better, interrelational” [1].

6 M. Barber

of us are aware of each other’s willingness to do our parts, and in which thesemutually shared beliefs serve in part as reasons for my intending to do my part of ourcommon action [2]. In Tuomela’s opinion, for Bratman the agents have their personalintentions on the basis of their private reason and there is, in effect, no group reason(in accord with which each partner executes his particular part of the joint actionwith another). Bratman’s account represents simply an I-mode willingness to act,but without the kind of cooperation that a joint action would require, such that ifthe other person did something wrong one might step in (“do one’s part”) to ensurethat the action intended, to which both parties are committed, is realized [2]. Further,Tuomela allows that we-mode shared intentionality can be found among not yet fullydeveloped individuals, such as infants, or it can be produced in conformist, culturaltransmission among individuals who are not clearly individuated, differentiated, andmature, as if autonomous independent agents, fully in possession of themselves, arenot necessarily a pre-condition of joint action [2].

However, the fact that Tuomela’s standpoint includes a much more expansivewe-mode of group participation in contrast to Bratman’s I-mode does not implythat he is anti-individualist. His notion of we-mode joint intentionality seems topresuppose autonomous individuals who voluntary choose to enter into a we-modejoint action, and whether or not one will perform one’s part as part of a joint actiondepends on how such individuals typify themselves, that is, what role they take upwith regard to a group, for instance, as a group member or as a private person [2].While legitimately resisting any idea of that a group acts as an ontological agent [2],Tuomela, it would seem, places a priority on individuals from whom the level ofjoint attention is derivative. Hence he asserts that the “we mode is in the minds andactions of individuals,” [2] that “groups are agents and persons only in ametaphoricalsense,” [2] and that “Much of what is said in this book prima facie treats groups asreal but nonagentive social systems supervenient on the members functioning quagroup members.” [2].

Counter to the individualistic leanings of analytic accounts of shared intention-ality, such as Tuomela’s and Bratman’s, Annette Baier [4], while acknowledgingthat voluntary associations do in fact depend upon “pre-planned intentional ‘joining’(joint activities) or some intentional ‘sharing’ (shared activities), or by a ‘unification’of separate wills,” argues that we engage in many activities done with others beforedoing them alone. It is precisely this fact that Baier charges that analytic philosophersof action overlook insofar as they operate with the following assumptions:

One such assumption is that the analysis of singular-person action must precede that ofplural-person action. Another is that intentional actions that initiate something should beanalysed before intentional continuings of what is already started. The third is that “priorintentions” should be analysed before intention in action [4].

John Searle, as Baier [4] acknowledges, seeks to do justice to the idea of pluralsubjectivities, and posits a collective intentional behavior that is not the summationof individual intentional behavior, in which one’s intention is entirely independentof the intentions and behavior of others. Rather collective intentionality, as might befound in an orchestra, corps de ballet, or football team, is based on the we-intention


expressed as “We are doing act A,” from which individual intentionality, expressedby “I am doing act A,” is derivative [5]. Searle posits we-intentions as a primitiveform of intentionality and contends that analyses that reduce such intentionality toindividual intentionality (including mutual beliefs about each other’s intentionality)do not provide sufficient conditions for cooperation [5]. Just as Tuomela opposes anywe-subject, Searle insists that “society consists entirely of individuals, there cannotbe a group mind or group consciousness. All consciousness is in individual minds,in individual brains” [5]. He also allows that one could be mistaken about whatis actually happening—even to the point that two parties could share a collectiveintention if they were merely brains in a vats. Such a collective intention could beexpressed as “We intend that we perform act A” [5] and such an intention can existin the mind of each individual agent who acts as part of this collective and whoseindividual acts are part of the collective act [5].

Hans Bernhard Schmid criticizes Searle for restricting collective intentionality towhat is in individual minds, and he imagines an evil scientist whomight separate twofriends contemplatingMatisse’s “Dance,” place their brains in vats on different floorsof the museum in which they were located, and then provide appropriate input sothat both think that they are contemplating Matisse’s “Dance” together. But Schmidobjects,

Intentionality does not become shared intentionality just because completely independentlyof each other, two brains just happen to have appropriately ‘matching’ illusions. If sharedintentionality is not a matter of what goes on inside an individual head, it is not a matter ofwhat goes on inside different heads either… it is not enough to check only what is in theminds of two individuals … sharedness is a matter of the relations between minds [6].

In fact, Searle’s rejection of any group consciousness and his confinement ofall consciousness to individual minds or brains isolates one consciousness fromanother and blocks the development of a collective consciousness that involves a realrelationship with another, in which one is immediately in contact with the other’sbodily cues and gestures that solicit and promise cooperation, express agreement orcommitment, and confirm one’s anticipations. Moreover, Searle’s hypothesizing ofbrains in a vat runs contrary to what he, at the end of his essay “Collective Intentionsand Actions,” calls the “pre-intentional background conditions” that make collectiveintentionality possible and that are so fundamental and taken for granted that onerarely even expresses any beliefs about them. Such conditions are at play when onejoins in pushing the stranger’s car because one is experiencing a real (and not merelybelieved) relationship with the stranger in which the car is real, the stranger’s bodyis real, the earth one stands on is solid, and one responds to an array of soliciting andconfirming bodily cues from the stranger, even if the stranger and oneself exchangeno words. Like all of the thinkers surveyed in this discussion of joint attention,Margaret Gilbert refuses to “invoke any ontologically suspect kind of ‘social spirit’or ‘group mind,’” [7] but rather seeks to explain joint commitment as “a function ofthe understandings, expressive actions, and common knowledge of the parties” [7].On the one hand and despite this rejection of supra-personal agents, Gilbert movestoward a more collective understanding of what she calls the plural subject than

8 M. Barber

Bratman or Tuomela. She contrasts a “personal commitment,” [7] in which one isthe sole author of a commitment and has the authority unilaterally to rescind it, withjoint commitment, which is not an aggregate of personal commitments. Further, sucha joint commitment is not a matter of creating corresponding personal commitmentssimultaneously or a conjunction of two members in which each participant is theauthor of his or her part of the joint commitment. “Rather, they comprise the authorof the joint commitment together.” [7]. In addition, if the idea of two individuals eacharriving at separate, deliberate decisions to partake in a joint action might accentuatethe primacy of individuals over the group, then a more relaxed approach to jointcommitment, with less self-consciousness involved, could de-emphasize individualautonomy as a condition of a plural subject. Gilbert stresses the informal meansof coming to joint commitments [7], such as tacit or implicit agreements or therather unplanned ways in which people find themselves having “fallen into” [7]joint commitments [7]. Gilbert summarizes her leaning toward a more collectiveunderstanding of plural action when she states,

Note that it is not obvious that coming to use the pronoun “we” correctly in the pluralsubject sense presupposes prior mastery of the pronoun “I.” The plural subject theory doesnot obviously require self-consciousness or striving conceived of as personal to precede theunderstanding that one is a group member [7].

On the other hand, one can detect in Gilbert’s work a continued kind of emphasison individualism, as Baier [4] charges. Gilbert states that joint commitment typicallyconsists in two people “in face-to-face contact, mutually expressing their readinessto be jointly committed, in conditions of common knowledge” [7].4 In addition, forGilbert those “understandings, expressive actions, and common knowledge” [7] thatreplace any “group mind,” [7] play a role in upholding the importance of individualsin plural action, in particular, the obligations and entitlements partners experience inrelationship to each other in a shared intention.Within a joint commitment each part-ner is entitled to the corresponding partner’s obligated performance and is licensedto rebuke that partner if he or she violates such an obligation [7].5 That Gilbert meansthesemutual obligations and entitlements between those jointly committed to protectthe individualism of the parties, who are each held to be individually accountable toeach other, appears in the fact that she appeals to these obligations and entitlementsin countering what she takes to be the collectivism of Searle and Baier. CounteringSearle’s view that “We intends” are in the minds of individual human beings with herown view of how expressions of readiness yield a joint commitment, she proceeds toemphasize repeatedly that Searle needs to explain his idea of “We intends” throughthe mutual obligations and entitlements that are associated with shared intention andaction and that by the partners’ mutual answerability would help preserve the indi-viduality of each individual [7]. Likewise, in response to Baier’s objection to her andothers’ individualistic bias, she faults Baier for not explaining how doing things withothers, which is more basic than doing things alone, cashes out in terms of the mutual

4The italics in the quotation are mine.5Gilbert [7] is reluctant to call these obligations “moral.”


“accountability those acting together perceive themselves to have to one another” [7].Baier, in turn, substantiates her accusation about Gilbert’s individualism by citingher claim in On Social Facts that plural subjects “pool their wills” together [4, 17].

3 The Irreducibility and Priority of We-Intentionality

Having seen repeatedly Annette Baier’s critique of those who assign to we-intentionality a derivative status, it is not necessary to reiterate her arguments abovethat highlight the prioritization of individual intentionality. Instead, as an exampleof another position stressing we-intentionality, one can consider Bernhard Schmid’seffort to defend the priority of we-intentionality through his account of plural action.Schmid begins by distinguishing intentional autonomy, according to which indi-viduals perform their own actions, from intentional individualism or motivationalautarky, according to which each individual’s behavior “has to bottom out in his orher own volitions or pro-attitudes (rather than in some other individual’s volitions orpro-attitudes)” [6]. Schmid [6] argues that plural actions occur when several personsshare one goal (not just a similar goal), and he suggests that that intentional autonomydoes not preclude intentional hertararky, a kind of other-regarding motivation thatdoes not bottom out in one’s own volitions (or motivations) and that does not reduceoneself to falling under the remote control of another or to acting as an intentionalzombie. Schmid provides examples:

What is at stake here are simple patterns such as the following: holding open a door for astranger, spontaneously helping a stranger to lift a baby carriage into a train, or moving asidea little on the park bench so another person can find a seat too… These are social actions,because they require cooperation, and they are plural actions exactly insofar as the helper’sgoal is the same as the individual’s who is being helped [6].

At this point, favoring a “teamwork” model [6] and concurring with Baier that wehave been brainwashed by Cartesian individualism [4, 6], Schmid rejects Bratman’sindividualism as well as what he calls Searle’s subjective individualism that wouldpermit joint intentionality between brains, placed in vats, which are not really inrelationship with each other [6], favoring a form of collective intentionality that isirreducible to individual intentionality and yet involves real relationships betweenindividuals. Contrary to the positions of Bratman and Tuomela then, Schmid makesthe case that collective intentionality, or shared intentionality, actually takes prece-dence over any individual intending, such as one’s intending to do one’s part, anydecision to share intentionality, any agreement to share intentionality, or any under-taking of the kinds of entitlements or obligations that Gilbert discusses. For instance,if one, like Gilbert, posits tacit agreement between two individuals at the basis ofshared intentionality, Schmid points out that a pre-condition for entering into such anagreement is that a certain kind of shared intentionality already must be in place sothat the two parties can come together to agree. Agreeing itself is something peoplehave to do together [6]. For Schmid, it is possible to have shared intentional activities

10 M. Barber

without any obligations and entitlements being operative and hence shared intention-ality is not itself socially normative, though it inevitably leads to socially normativeconsequences. Hence, merely habitual social practices, like customs, already involveshared intentionality, but they usually issue in social norms. In Schmid’s opinion,having a collective intention, then, provides a reason to develop an appropriate per-sonal intention, such as to do one’s part in a common endeavor [6]. Schmid providesnumerous examples of people sharing intentionality in which one lacks a sense thatone participates as a distinct individual. One instance appears when two colleaguescollectively intend to go to a cafeteria and one of them becomes so absorbed in theconversation that she forgets that they are even going to the cafeteria or that she hadagreed to go to the cafeteria in the first place. The colleagues share an intentionalityin which at least one party’s individual intentionality has dropped from sight [6].Or at a public assembly the president of an organization might express “our” (thecompany’s) grief over the death of one of its officials, even though those attendingand being interviewed afterward, might admit to feeling little of the grief that thepresident spoke of. There is a shared intentionality even though some members donot experience themselves as individual participants [6]. To head off any concern thatsuch shared intentionality might imply a collective super-agent, Schmid urges read-ers not even to ask the question who the subject is who has the collective intention;the question itself is misplaced [6].

4 Beneath Theory: The Genetic Origins of CollectiveIntentionality

When, as Tuomela, one approaches the problemof collective intentionality beginningwith adult individuals typifying themselves in the we-mode of the we-perspective asa “group member” (as opposed to as a “private person” in the I-mode) and takingone’s belief that others will do the same as a reason for doing one’s own part as such a“group member,” one’s shared commitment appears derivative from one’s individualintentionality—and Bratman does this even more than Tuomela. Searle posits sharedintentionality as a primitive, however his localizing of all minds (and intentional-ity) in brains and allowing that two brains placed in vats might share intentionalitydespite their not have any real relationship seem to permit individualism to resurfacein a problematic way. Gilbert emphasizes the informal ways in which shared com-mitments are entered into and tends to diminish the polarization of individuals overagainst each other, but that individualism remains prominent insofar as she starts fromindividuals signaling to each other a readiness for shared intention and insofar sheemphasizes the obligations and entitlements that are constitutive of shared intentionsand that ensure that individuals remain accountable to each other. Granting a placefor the deliberately entered into shared intentionalities that Tuomela and Bratmantake as paradigmatic, Baier, by contrast, emphasizes shared intentions by focusingon actions we cannot do with others and continuing actions rather than actions we


initiate, though Gilbert accuses Baier of neglecting to clarify the individual that hersystem of entitlements and obligations reveals. Finally, Schmid posits shared inten-tionality as the fundamental fromwhich individual intentionalities emerge, graduallyincluding the kinds of normative social relationships that Gilbert accentuates, andhe explains how intentional autonomy need not preclude intentional heterarky. Ona spectrum one might plot Bratman and Tuomela at the individualist pole, withGilbert and Searle in intermediate position, and Baier and Schmid at the more col-lectivist pole. What becomes interesting, however, is the ways in which many of theauthors considered seek to return to a level of experience beneath that of their the-orizing, to seek to recover how actors experience their individual intentionality andshared intentionality, an example of which occurs in Searle’s above mentioned ideaof “pre-intentional background conditions.” For instance, Tuomela, at the end of hisessay “We-Intentions Revisited” responds to criticisms that his argument is circularbecause he seeks to define collective intention as each “doing his (or her) part” andyet the very idea of doing one’s part presupposes collective intention. Tuomela thenattempts to describe how pre-analytic, ordinary, common-sense agents might thinkabout joint intention even though they might not have in view at all the notion ofcollective intention that is central to the theoretician’s point of view, that is describedabove in terms of an individual actor we-intending to perform an action, and thatinvolves an intricate intertwined taking account of other’s willingness to do theirpart, etc. Such a common-sense actor “intends to perform X with the others in partbecause (of his belief that) the others intend to perform their parts of X and intendto perform X with the others” [1]. It should be noted that this description of thecommon-sense actor’s point of view seems more individualistic than ever since oneseems to take even less account of beliefs about others’ beliefs, and, at the same timeit consists a kind of individualistic calculation that one’s own action is conditionalupon the cooperation of others. Nevertheless, this effort to reconstruct how commonsense actors experience joint intentions converges, as I hope to show, with a kind ofgenetic phenomenology that Alfred Schutz provides and that synthesizes both thecollective and individualistic dimensions of joint intentionality.

Similarly, John Searle, at the end of “Collective Intentions and Actions,” afterproviding his own account of collective action, turns to discuss a pervasive “senseof community” [5] that collective intentionality itself presupposes and that includesthe kinds of background skills needed, for instance, to play football together or per-form in an orchestra together. Searle argues that such background features are noteasy to characterize, that old-time philosophers tried to get at them by describingus as social or political animals, and that they require a “pre-intentional sense of‘the other’ as an actual or potential agent like oneself in cooperative activities” [5].6

6It should be noted that when Searle speaks of a “pre-intentional” sense of the other, he is taking“intentional” in the typical way that analytic philosophy uses that term, namely as describingan aiming at a goal toward which action will be directed and as implying a certain amount ofdeliberateness. For Schutz and the phenomenological tradition, “intention” refers to any consciousact, such as desiring, thinking, memory, judging, etc. that is aimed at an object (and hence one isdesirous of, thinks of, has a memory of, or judges about), and often one is unaware of such intentionsuntil one reflects upon them.

12 M. Barber

About such background features, Searle posits that one just presupposes them andthat one does not have any special belief about them being there, in the same waythat one supposes without specifically believing that the ground beneath me is solidand that others are conscious agents. For Searle we cannot ultimately explain soci-ety through conversation or other collective behaviors since even they presupposea form of society for them to function at all. In my view, Searle’s forays into thisquestion precisely converge with the life-world that Schutz and Luckmann describeas that “province of meaning which the wide-awake and normal adult simply takesfor granted in the attitude of common sense” [9]. It is “unquestionable” [9] and “un-problematic until further notice” [9]. In addition, Schmid seeks to cast doubt on theidea that collective intentional activity is based in some commitment or dedicationto a cause on the part of “participating individual,” [6] and he seeks to recover “intu-itive, pre-theoretic understanding of what it means to share an intention” [6] “withoutimplying any appropriate individual contribution or contributive intentions, or anyindividual we-intending” [9]. As we have seen, Schmid [6] offers examples of sharedintentionality such as when one of the individuals has completely forgotten aboutthe intentionality with which the joint action was initiated or when a group shares anemotion despite the fact that some participants do not feel it—both of which illustrate“non-participatory uses of we” [6]. Given these examples and given Schmid’s obser-vation that “social norms arise out of merely habitual social practices” [6] (in whichone shares intentionality before finding oneself obligated or entitled as an individ-ual to certain actions), Schmid contends that “the analysis of shared intentionalitycannot be based on an analysis of what individuals personally intend when sharingan intention” [6]. This pointing toward non-participatory shared intentionality, asSchmid describes it, can indicate, I hope to demonstrate, the kind of experiencesthat Schutz’s account of the life-world brings into focus. Finally, the author whosethought most approaches a genetic phenomenology like Schutz’s is Annette Baierin “Doing Things with Others: The Mental Commons.” Baier’s ultimate premise isthat shared intentions lie at the root of mutually referring individual intentions andthat individual action is itself learned as a departure from common action. Criticalof Jon Elster’s individualism, she posits that social norms become mysterious if theyare supposed to be “evolved either by chance or by rational agreement between fullyfledged self-interested individualists” [4]. Allowing a role for individual intentions involuntary associations in which pre-planned intentional action might be important,Baier wonders why “first person singular should be deemed more fundamental thanfirst person plural” [4]. Instead, she focuses on unplanned actions rather than plannedones, on inevitably intersubjective institutional actions, on unreflective customs thatstructure unpremeditated acts, on activities that two people find themselves doingtogether (such meeting each other’s gaze, glaring at each other, traveling alongsideanother), on non-voluntary activities on which the voluntary are parasitic, and on theintentional continuation of activities that were not intentionally intended or plannedbeforehand. Instead of considering actions that one does on one’s own and mightadmit others to partake of (e.g. such as sharing one’s bank-account), she concentrateson actions one cannot do alone or learn to do alone such as war-making, wrestling,


disputing, dialoguing, reasoning, inferring (according to Robert Brandom), disagree-ing, playing, greeting, marrying (essentially done with others in contrast to paintinga house that is accidentally done with others), and all the kinds of activities whoseverbs commencewith the prefix “con.” The genetic character of her concerns becomeclear in the relationship between a child and its mother, a symbiosis, which includessuch as gestation and breast-feeding, in which amother cannot act alone and of whichElster exhibits “the usual philosophical amnesia” [4]. Concluding by recalling theshared existence mother and child share in pregnancy, she proposes that maybe thequestion should not be how collective action is possible but how individuality andindividual action is possible [4]. Baier sets the stage in this paper for our consider-ation of Schutz’s genetic-phenomenological account of the everyday life-world andthe development of individuals within it, in which the We-relationship is primordial,but at the same time, co-originary for Schutz is one’s biographical career—an aspectthat Baier to a degree underestimates, despite her brief acknowledgement that thereare voluntary, pre-planned individual intentions that that may be necessary in somecases for shared intentionality.

5 Genetic Origins: Social and Individual Intentionalityin Alfred Schutz’s Phenomenology

Alfred Schutz’s phenomenological approach to the social world synthesizes many ofthe diverse strands of the preceding discussion and presents a novel reconciliation ofthe social-individual polarization, including the individual/joint intentionality prob-lematic. Like Baier, Schutz locates through a kind of genetic investigation the originsof our higher level activities, individual and social, in what he calls a primordial We-relationship, which, once experienced in childhood, remains a kind of relationshipinto which one can enter at any time as an adult. Schutz writes,

As long as man is born of woman, intersubjectivity and the we-relationship will be thefoundation for all other categories of human existence. The possibility of reflection on theself, discovery of the ego, capacity for performing any epoché, and the possibility of allcommunication and of establishing a communicative surrounding world as well, are foundedon the primal experience of the we-relationship [10].

Schutz, as a philosopher, defines the we-relationship (in a way that the common-sense actor would never think of doing) as a matter of each partner to a relationshipattending in simultaneity to the ongoing processes of each other in a face-to-facerelationship, in which we share the same space, bodily presence to each other, andtime—or, as Schutz puts it, a “vivid present” [11]. Schutz demonstrates how sucha we-relationship is enacted in the simplest conversation in which one follows apartner’s building up the thought being communicated, through a process that unfurlswith each participant simultaneously coordinatingwith each other, step by step, wordbyword. Such a we-relationship, based on a sharing of space, time, and the unfoldingprocess permits both parties to say “We experienced this occurrence together” [11]

14 M. Barber

or to comment that “we grow older together” [11]. While one lives immersed in thewe-relationship with the other, one is absorbed with the development of the other’sconscious processes in the fullness of their concrete content in the vivid presenceone shares, and to bring that relationship into reflective focus, the partners must stopfocusing on the other and step out of the we-relationship [12].

The child, despite lacking much of a developed self or of consciousness, clarity,or determinateness about what is going on, engages in a we-relationship with adultswho conduct themselves as if they were in a we-relationship with the child. Onlyin this context, in this “reciprocal motivational context,” [9] does the child slowlycome to understand goals, means, attitudes, identity, and the group’s natural worldview, and to develop his or her own personal identity [9]. Furthermore, “the concretereality of the we-relationship is the social basis for the individual learning a languageas a system of meanings referring to ‘any reality whatever’” [9]. Accentuating theimportance of such a shared relationship, Schutz agreeswithMax Scheler’s objectionto the supposition of the analogy argument for the existence of the Oher that the firstthing given to us is our own self. Instead Schutz concurs with Scheler that the “We”is naively presupposed and given to each of us as prior to the I. In fact, for Schelerpsychological accounts of children and “primitives” indicate that they live in others’experiences before living in their own in such a way that one cannot differentiateone’s experiences from those of another. Schutz accepts that we live in others’ lifeinstead of our own but only insofar we are unreflectively directing our thoughts andacts toward the other in the “vivid simultaneity of theWe” [13].7 However, when oneturns in reflection to one’s experience and breaks the spell of the We-relationship, itbecomes clear that there is no such thing as an experience given to me that does notindicate to which individual stream of consciousness the experience belongs. Schutzagrees that children and “primitives” exhibit the immersion in an experience in whichour individual experiencesmay appear indistinguishable—aswe all dowhile living inthewe-relationship—but this is also in part because children and “primitives” acquirelater the capacity for reflection, in which I, like them eventually, once outside thebounds of the we-relationship, recognize my stream of consciousness as “throughand through the stream of my experiences” [13].

Having presented theworkings of thewe-relationship, Schutz andLuckmann raisethe question of how it is “constituted,” [9] and this technical, phenomenological termrefers, not to a kind of idealistic creation of something that really exists only in one’smind, but rather to the fact that, when one experiences an object as something of afinished product standing before oneself, say a house, it is always possible to inquireback into the many types and levels of experiences that are involved in experiencingthe object; thatwould go into the building up of finalmeaning of the object, as a house,for instance; that one may have forgotten or never even focused on; or that one mightreconstruct in attempt to give a systematic account of how the meaning of the object

7Of course, the use of the term “primitives” to describe cultures different from one’s own or not astechnologically advanced as one’s own is regrettable. The negative connotations of that term maylead one to ignore the psychological or spiritual dimensions of such cultures that may be richer thanthose in one’s own culture.


would be arrived at. For instance, to be involved in a we-relationship with another,one must have experienced the Other as an Other like me and in such an experienceadopted what Schutz calls a “thou orientation” in the other’s presence [9].8 Further,the constitution of theWe-relationship includes, as Schutz andLuckmann continuallypoint out when discussing the we-relationship, “mirroring,” [9] an “interlocking ofglances” [9].

A further essential component of this [we-] experience is that I grasp his attitude to me. Healso experiences my action not simply in an Objective interpretative context, but also as anexpression of my conscious life. Further, I apprehend that the fact that he experiences me assomeone who experiences his conduct as an expression of his subjectivity. In the we-relationour experiences are not only coordinatedwith each other, but are also reciprocally determinedand related to one another. I experience myself through my consociate, and he experienceshimself through me. The mirroring of self in the experience of the stranger (more exactly, inmy grasp of the Other’s experience of me) is a constitutive element of the we-relation [9].

Having discussed all these features constitutive of a we-relationship, Schutz andLuckmann insists that neither the complex refractions of mirroring nor the factthat one is involved in a we-relationship is explicitly grasped in reflection by acommon-sense actor [9]. Rather one immediately experiences a shared experiencewith another; it is the phenomenologist or philosopher who gives a reflective accountof what is supposed by such an experience.

One of the problems involved in discussing joint intentionality is that the phe-nomenological tradition defines intentionality differently than do many in the ana-lytic tradition, such as many of those whose positions were presented above. Forinstance, Bratman and Tuomela, define intentionality as a deliberate, self-conscious,autonomously decided-upon intention to engage in action, without or withoutanother, in pursuit of some purpose, with or without others. In this context, it wouldseem inevitable that joint intentionality appears derivate from individual intention-ality. However, in the phenomenological context, intentionality is any consciousactivity that is oriented toward an object, whether one is discussing awareness ofsomething, desire for something, fear of something, etc. Schutz’s account of the we-relationship, and in particularly of the mirroring processes, involves multiple inter-twined intentional activities, as when one is conscious of another’s consciousness ofone’s consciousness of her—an interlacing of intentionalities that Husserl describedas “iteration” [8] or that the American sociologist Charles Horton Cooley denomi-nated the “looking glass effect” [14]. Remarkably, one engaged in a we-relationshipprobably would not even be aware of this web of intentionalities. Hence, the we-relationship, in which one experiences being-with another and is hardly aware ofone’s individual conscious processes or individuality at all, precludes any derivationof joint intentionality from individual intentionality, in contrast to the individual “in-tentionality” by which one deliberately aims at a project and considers whether tocooperate with others to realize that project through a joint intentionality.

8There would be several indications that the other is “like” me, and the indications of the likenessand the likeness itself are not a matter of inference, but passive syntheses, by which I immediateassimilate the other to myself, as capable of sharing a we-relationship.

16 M. Barber

But the shared intentionality (in the phenomenological sense) in the we-relationship is not only a matter of the partners mutually immersed in each other’sunfolding conscious processes and the mutual coordination between them, but it isalso possible for them to jointly attend to an object in their environs. For instance,Schutz comments on how two partners can “grow older together” by simultaneouslyexperiencing together in a we-relationship a bird in flight. Schutz the phenomenolo-gist reflectively unpacks the constitutive underpinnings of this experience that eitherof us might report in terms “we have seen a bird in flight.” He observes, for instance,that “Perhaps, while I was following the bird’s flight I noticed out of the corner ofmy eye that your head was moving in the same direction as mine” [12]. And, thestatement about us seeing the bird would depend on many other intentional linkagesabout which neither of us may be explicitly conscious (e.g. recognizing that youare an Other like me, that you are not visually handicapped, that your body overallindicates that you are wide-awake and alert instead of sleep walking—all of which Imay be vaguely aware at the moment when we see the bird’s flight together). But inspelling out the underlying intentional anticipations (and their confirmation insofaras they continue to be borne out), we are only giving a constitutive analysis of theexperiences, of whose complexity both parties may not even be conscious and inwhich one may have no idea of the individuality of one’s intentionality as distinctfrom the other’s. The experience that Schutz analyzes consists in the jointly intendedperceptual act in which, as would be reported by both the participants, “we togethersee a bird a bird in flight.”

In addition to a shared perceptual intentionality with the same perceptual objectthat we are perceiving together, a “we” can jointly intend a project in which we areso immersed that we are not mindful of our individual activity and our individualcontribution, so taken up are with each other and what we are doing together. Schutzpoints to a series of activities in which this is the case: playing baseball or tennis,fencing, dancing together, making love together, or making music together [15].He articulates how a “we” intend together the last-mentioned project insofar aseach of the co-performers of a group undertake the project of producing a musicalpresentation, in which each

has, therefore, to take into account what the other has to execute in simultaneity. He has notonly to interpret his own part which as such remains necessarily fragmentary, but he has alsoto anticipate the other player’s interpretation of his—the Other’s part and, even more, theOther’s anticipations of his own execution. Either’s freedom of interpreting the composer’sthought is restrained by the freedom granted to the Other. Either has to foresee by listening tothe Other, by protentions and anticipations, any turn the Other’s interpretation may take andhas to be prepared at any time to be leader or follower. Both share not only the inner durée inwhich the content of the music played actualizes itself; each, simultaneously, shares in vividpresent the Other’s stream of consciousness in immediacy. This is possible because makingmusic together occurs in a true face-to-face relationship—inasmuch as the participants aresharing not only a section of time but also a sector of space. The Other’s facial expressions,his gestures in handling his instrument, in short all the activities of performing, gear intothe outer world and can be grasped by the partner in immediacy. Even if performed withoutcommunicative intent, these activities are interpreted by him as indications of what the Otheris going to do and therefore as suggestions or even commands for his own behavior. Any


chamber musician knows how disturbing an arrangement that prevents the coperformersfrom seeing each other can be [15].

In the we-relationship of musicians jointly intending to reproduce a piece ofchamber music, the predominant experience is one of we together intending andexecuting the performance, with their individual intentionality and the intricacies oftheir taking account of each other submerged and subordinated to their final purpose,however adept a theoretician like Schutzmight be about illuminating such subliminalsyntheses and intentionalities.

Schutz’s account, which depends on the submersion of individuality and the pas-sive syntheses and anticipations of which one might not even be aware, explains notonly all that is presupposed under the governance of a jointly intended project inthe we-relationship, but it can also explain how we often, as Gilbert suggests, “fallinto” other joint commitments, in the first place. For instance, given that we havebeen immersed from childhood in we-relationships pursuing joint projects that havedrawn on syntheses and associations (e.g. the relaxed look on a fellow musician’sface elicits one’s own softer, more delicate use of one’s own instrument) that exceedour ability to give an account of them, so when we find similar joint projects “in themaking as it were,” we find ourselves being passively drawn into joining with othersin pursuit of a common goal. Thus, for example, because we are so familiar withprojects inwe-relationships, whenwe see our car and another blocked by a fallen tree,we immediate immerse ourselves in a cooperative endeavor with the other car driverto move that tree, engaging in little if any verbal communication with that driver andrelying on a massive fund of interpersonal and physical anticipations and skills—all serving that we-project without our being much aware of the operations of thoseintentionalities at all. Our massive fund of experience of undertaking and completingprojects within we-relationships, from childhood on, disposes us to cooperation withothers. It is as though the very sight of a we-project in the making elicits our par-ticipation without our needing to deliberate and decide whether to partake, whetherwe are talking about helping a stranger lift a baby carriage onto a train, vacating asector of a park bench so that another can find shade, breast feeding, accompanying acolleague to the dining room, leaping into a dispute or dialogue, or merely engagingin the simplest conversation. This explanation of the we-relationship does not ruleout an individual’s reflecting and deliberately deciding to embark upon a joint actiononly if one sees that otherswill also bewilling to do their part and that this willingnessto cooperate is mutually known, as Tuomela portrays joint intention, and, as evenBaier admits when she acknowledges that such individual intentionality is requiredfor voluntary associations in which preplanning is needed. One wonders, though,whether even such a deliberate decision to embark upon a jointly intended actionmight only seem plausible or even attractive to someone who has from childhood onalready partaken successfully in joint projects within we-relationships and withoutmuch deliberation or even awareness of all the latent intentional processes at play.

Finally, one cannot simply neglect Gilbert’s question about whether Baier’semphasis on collective action provides sufficiently for the individual, whom Gilbertbelieves is called to accountability by the obligations and entitlements that pervade

18 M. Barber

relationships of joint intentionality. One cannot dismiss the individuality that comesto the fore when one decides to enter into the voluntary associations that Tuomelafocuses on or that appears in the notion of intentional autonomy that Schmid refusesto let go of, despite Schmid’s proclivity for “non-participatory” we-experiences thatconverge with Schutz’s position on “we-relationships.” One ought not overlook thepresence of individuality even within Schutz’s depiction of the we-relationship,despite the fact that such a relationship in a family subsumes within it childrenwhose self-identity is scarcely developed and whose parents in effect “force” [9]or “impose” [9] the reciprocity of the we-relationship on them. In addition, onecannot dispense with individuality despite the fact that the implementation of thewe-relationship among adults tends to be so focused on the other with whom oneshares a vivid present and on the other’s conscious processes coordinated with one’sown, that the individual self hardly becomes visible at all.

But in the case of Schutz,whose own thought derives fromHusserl’s, including thelatter’s emphasis on the transcendental ego individualized by its history as an ongoingstream of consciousness, one might expect that he would not neglect the importanceof individuality. Indeed, as Schutz himself observes, my subjective experience andthe other’s are such that in my understanding of the other’s experience I can neverunderstand it the way the other does. I might have been able to do so if I had livedthrough all her conscious states and intentional acts, down to the smallest details,impressions, and horizonal areas of protention and retention. But I would also have toremember all her experiences and to have lived through them in the order andwith theintensity that she did, paying attention to each of them as she did. As a consequence,such “‘Intended meaning’ is therefore essentially subjective and is in principle con-fined to the self-interpretation of the person who lives through the experience to beinterpreted. Constituted as it is within the unique stream of consciousness of eachindividual, it is essential inaccessible to every other individual” [12]. Although anyor all our interpreting concepts or thoughts rely on the socially transmitted and inher-ited typifications bestowed upon us within the various we-relationships to which wehave belonged from childhood on, they are also inflected and informed by the uniquehistory of the individual relying on them. Schutz insists, though, that this does notdeny that anyone can understand another person’s experience, but that, rather, themeaning I give to another’s experience cannot be exactly the same as the other’s [12].

Schutz and Luckmann recognize that the subjective system of interpretational andmotivations relevances, including plans, typifications, models of explication, and soon, have a “prior history” [9] that one brings into every we-relationship. They reit-erate over and over again that same events and objects never have exactly the samemeanings for one partner as they do for another because of the unique biographi-cal situation that each brings into any we-relationship [9]. Indeed, one can imaginethat even in one’s initial we-relationship, in the pregnancy that Baier references,should the mother’s pregnancy be tumultuous, the child might emerge to a degreeas more physically and psychologically intense than a child emerging from a moretranquil pregnancy. And this very intensity, bodily impressed on the child in thewe-relationship of pregnancy, will affect the next we-relationships into which thechild enters, e.g. in breast feeding or in encounters with one’s father or grandparents.


In effect, one’s individual biography and identity and one’s inescapable social con-nectedness and formedness are inextricably intertwined and one’s individuality andintersubjective relatedness are co-originary from the start. As Schutz suggests else-where, our biographically situation is an imposed relevance, it is there as given andwe cannot redo it, although we can come to terms with in, given the set of relevancesand typifications we have at hand, and of course that system of “intrinsic relevances”by which we come to terms with imposed relevances, are themselves the product ofour social and individual history that has preceded the present moment [16].

Schutz, in effect, has reconstructed a genetic account of joint intentionality—something sought inchoately by the philosophical interlocutors mentioned above—though, Schutz does so by understanding intentionality phenomenologically, that is,as including the wide spectrum of acts or orientation toward objects, persons, andevents that are often directed to their objects often without being consciously the-matized. He shows the foundational character of the we-relationship, for all othercategories of human existence, and the trajectory of one’s biography, which is co-originary with one’s earliest we-relationships. Schutz’s account then would implythat individuality and social pertinence are present in any social interactions, thoughone comes to the fore over the other in different contexts and in the considera-tion of different problems. For example, when one is self-consciously deliberatingabout whether to engage in a joint action with others, about whether her doing herpart will be matched by them doing their part, and about whether there is sufficientmutual knowledge for the joint action’s successful execution—all of which precedeembarkation upon a joint action in Tuomela’s view—it is clear that one’s individualintentionality emerges into prominence and the joint intention appears to be deriva-tive. Schutz himself has developed an account of how individuals intend a project byimagining in future perfect tense the project completed, choosing it, and setting inmotion the sub-acts leading to its realization—and such an account could be extendedto explain joint intentions. However, even in those moments when one deliberatesbefore a joint action, one’s understanding of and willingness to embark with otherson such an action depends on one’s massive past experience of we-relationships,however much this experience might lie on the horizons of one’s focus while onedeliberates. Likewise, Baier’s explanation of shared intentions, that Gilbert thinksoverlooks individual intentionality and that consist in such activities as pregnancy,breast feeding, playing a game, or painting a house, simply does not attend to the indi-viduating stream of consciousness that belongs to each member of a we-relationshipand that is co-originary with that member’s earliest experience of we-relationships.

6 Conclusion

In brief, Schutz recovers the originary, commonsense social bonds of the we-relationship through a genetic analysis, toward which Searle, Baier, and even

20 M. Barber

Tuomela gesture, based on an examination of children’s experience and the life-world that precedes theory and that has been a theoretical focus of the phenomeno-logical tradition for over a century. In addition, also following phenomenology’sguidance, Schutz presents a comprehensive account of intentionality, which includes,as one of its instantiations, the deliberating/deciding to act jointly that Bratman andTuomela treat in depth, but also the spontaneous anticipations, attendings, orienta-tions—in general aimings at—objects, persons, events, etc. that are operative, oftenwithout even being noticed, pervasively in interpersonal relationships, especially we-relationships.Also, the phenomenological tendency for careful description appears inSchutz’s elucidations of thewe-relationships, as is shownbyhis observations onmak-ing music together, as a particularly clear example of the many we-relationships intowhich adults plunge themselves, long after the identity-forming we-relationships ofchildhood. The important phenomenological methodology of “constitution” permitsone to distinguish the level of the we-relationship that is experienced as irreduciblycollective, as Searle, Schmid, and Baier would have it, and the multiple intersub-jective, constitutive processes and rich entanglements of attention, response, andintervention that usually remain anonymous. These processes and entangled inten-tionalities resemble some of the implicit obligations that Gilbert makes explicit asnecessarily bound up with joint commitments, whose frequently informal characterGilbert accentuates, an informality in which the independent individual conscious-ness recedes from prominence as it does in a Schutzian we-relationship. Moreover,the distinction between the grasp of the constituted we-relationship and the appreci-ation of the web of intentional activities that constitute it depends on the differencesphenomenology has seen between living in an experience and reflecting on it andbetween a simplifying reflection that interprets what one has just experienced mono-lithically, e.g. as being in a we-relationship with another, and a more thorough reflec-tion that captures its unnoticed, polythetic, many-faceted, intentional underpinnings.Finally, phenomenology’s reflection on the history and stream of unrepeatable, indi-vidual consciousness is not lost upon Schutz, however much that individual historymay elude one whose principal focus is on experiences of we-relationships. Phe-nomenology, then, affords various concepts and methodological tools that enable itto avoid or reconcile some of the polarizations within recent analytic discussions ofjoint intentionality.

References

1. Tuomela, R.: We-intentions revisited. Philos. Stud. 125, 3–369 (2005)2. Tuomela, R.: The Philosophy of Sociality: The Shared Point of View. Oxford University Press,

Oxford (2007)3. Bratman, M.: Faces of Intention: Selected Essays on Intention and Agency. Cambridge

University Press, Cambridge (1999)4. Baier, A.C.: Doing things with others: the mental commons. In: Alanen, L., Heinämaa, S.,

Wallgren, T. (ed.) Commonality and Particularity in Ethics, pp. 15–44. St. Martin’s Press, Inc.,New York (1997)


5. Searle, J.: Collective intentions and actions. In: Cohen, P.R., Morgan, J., Pollack, M.E. (ed.)Intentions in Communication, pp. 401–415. The MIT Press, Cambridge, MA and London,England (1990)

6. Schmid, H.B.: Plural Action: Essays in Philosophy and Social Science. Springer, Dordrecht(2009)

7. Gilbert,M.: Sociality andGroupResponsibility:NewEssays in Plural Subject Theory.Rowmanand Littlefield Publishers Inc., Lanham (2000)

8. Husserl, E.: Ideas pertaining to a pure phenomenology and to a phenomenological philosophy,vol. 1. General Introduction to a Pure Phenomenology (trans: Kersten, F.). Martinus Nijhoff,The Hague (1983)

9. Schutz, A., Luckmann, T.: The Structures of the Life-World, vol. 1 (trans: Zaner, R.M., TristramEngelhardt, H.). Northwestern University Press, Evanston (1973)

10. Schutz, A.: The problem of transcendental intersubjectivity in Husserl. In: Schutz, I. (ed.)Collected Papers 3: Studies in Phenomenological Philosophy, pp. 51–91. Martinus Nijhoff,The Hague (1966)

11. Schutz, A.: On multiple realities. In: Natanson, M. (ed.) Collected Papers 1: The Problem ofSocial Reality, pp. 207–259. Martinus Nijhoff, The Hague (1962)

12. Schutz, A.: The Phenomenology of the Social World (transWalsh, G., Lehnert, F.). Northwest-ern University Press, Evanston (1967)

13. Schutz, A.: Scheler’s theory of intersubjectivity and the general thesis of the alter ego. In:Natanson, M. (ed.) Collected Papers 1: The Problem of Social Reality, pp. 150–179. MartinusNijhoff, The Hague (1962)

14. Cooley, C.H.: Human Nature and the Social Order, rev. edn. Charles Scribner’s Sons, NewYork (1922)

15. Schutz, A.: Making music together: a study in social relationship. In: Brodersen, A. (ed.)Collected Papers 2: Studies in Social Theory, pp. 159–178. Martinus Nijhoff, The Hague(1964)

16. Schutz, A.: Collected Papers 5: Phenomenology and the Social Sciences. In: Embree, L. (ed.).Springer, Dordrecht (2011)

17. Gilbert, M.: On Social Facts. Princeton University Press, Princeton (1992)

Collective Phronesis? An Investigationof Collective Judgement and ProfessionalAction

Eva Schwarz and Jonna Hjertström Lappalainen

Abstract This paper is a conceptual investigation of whether, and how, action con-taining judgement can be understood as collective. Using an empirical example fromthe police profession, the article tries to capture the complexity of a real situationthat requires both agency and judgement. The example is analysed in relation tothe idea of collective judgement, both regarding ongoing debates in phenomenol-ogy concerning the possibility/need for a we-subject, and in social epistemologyconcerning the idea of judgement based on social evidence. The article claims thata richer account of action is needed in order to describe the collective aspects ofjudgements. Drawing on an Aristotelian understanding of action, the paper arguesthat an understanding of certain collective actions requires a notion of judgementthat involves both we-subjectivity, and social evidence.

Keywords Collective knowledge · Practical knowledge · Phronesis · Aristotle ·Judgement · Police

1 Introduction

With examples drawn from the context of professional action, this paper will discusswhether, and how, action containing judgement can be understood as collective.Professional action is of especial interest with respect to the broader question ofcollective agency since acting in one’s capacity as a professional often transcendsthe individual horizon of an agent in different ways; for example, as professionals,individuals represent a professional corps, meaning that they often act in relationto situations requiring the use of specific skills that define a professionally trainedgroup of persons. The question is how this can be understood.

E. Schwarz (B) · J. H. LappalainenCentre for Studies in Practical Knowledge, Södertörn University, 14189Stockholm, Swedene-mail: [email protected]

J. H. Lappalainene-mail: [email protected]


23

24 E. Schwarz and J. H. Lappalainen

Within phenomenology and social epistemology debates are ongoing surroundingthe possibility of collective agency. One central point in these debates concernswhether it is possible, if not necessary, to introduce the idea of a “we-subject” forwhom experiences, emotions and actions are shared [16] or a “group agent” [15]that performs a joint act in order to satisfy a certain goal irreducible to instancesof individual action or judgement. A further commonly discussed issue is whetherjudgment is possible based either on social evidence [7] or social emotions, or, forthat matter, feelings and embodied interactions [14].

In studying the everyday practices of professionals it is clear that complex casesare dealt with in which the very relation between collective action and judgementsis at stake. In order to aspire to and thus define a sufficient common goal (one thatmay itself change during the course of an action), both the quality of the action andthe judgement depend on one another. But can we take such professional cases asexamples of the irreducible and inextricable linkage between collective action andjudgement? In our investigation we take our point of departure in an Aristotelianunderstanding of action that enables us to deal with the complexity of professionalaction as both goal-directed and dependent on judgement, emotions, skills and knowl-edge. With the use of an actual example from the police, we will show that a jointactivity is best understood in collective terms. Yet, we will argue that, if we are togive a credible account of action then a more complex understanding of judgementis necessary, one requiring a form of collective judgement in a more comprehensivesense, involving both reasoning and embodied cognition.

Our paper is comprised of three parts. Firstly, we will introduce the Aristotelianconcept of phronesis and the particular understanding of action (praxis) this implies.Secondly, we will present an example from the professional praxis of police officersand discuss it within the framework of a contemporary philosophical debate on thepossibility of collective action. The empirical example is a text written by, and fromthe perspective of, a young police student who describes a difficult and potentiallydangerous situation in which, together with his (more experienced) colleagues, hemust exercise judgement. A beginner’s perspective on how a group is constituted, not(only) in an empirical sense (i.e. how a newcomer is inducted into a profession) butalso epistemically, is especially interesting since the newcomer’s perspective oftenrenders visible the constitutive structure of a group. The use of an empirical examplethus allows us to capture the complexity of a real situation that requires agency andjudgement in our conceptual analysis. In the third and final section, we will analyzeour chosen example in relation to the idea of collective judgement and finally drawsome conclusions on how professional action might be understood in relation to bothindividual and collective judgement.

2 Action and Practical Knowledge

In book six of his Nicomachean Ethics Aristotle famously introduces several formsof knowledge, one of which is “practical knowledge”. He goes on to distinguish two

Collective Phronesis? An Investigation of Collective … 25

different forms of human occupations or activities that require two different spheresof practical knowledge, namely poiesis and praxis. While poiesis is production,making or work, praxis is what we usually call action. Poiesis or work is, accordingto Aristotle, characterized by producing something in light of a pre-assigned goal.Work in this sense means producing a new entity that mimics a predefined product.Whether it is a good entity, or not, is determined by the product’s definition, i.e. by thegoal formulated in advance of production. The skill or capacity required for work istechne, art. According to Aristotle, art or techne, that is, the very process of making,does not demand any overview of the context or the purpose of the product. “Artis identical with a state of capacity to make, involving a true course of reasoning”[4]. Techne is the skill of know-how. This very process is often seen as monotonousor mechanical. It is not dependent on the fact a particular person is performing thetask, only that the agent possesses the requisite skills. It could, as Aristotle claims,be performed by a slave. In the case of poiesis the very process sometimes requiresseveral agents working together on a shared goal or product, one example of thismight be a group of ancient boat constructers merging the logs in order to form thehull of a boat. In terms of poiesis, collectivity does not seem so hard to grasp.

Praxis, or action, on the other hand, is introduced by Aristotle as the sphere ofinteraction between humans. It is a mode of engagement in which the very processis a goal where the goal is not external to the process. The implication here, though,is that the individuals involved in praxis have a different relation to the very process.It is not only the product that is valued, but also whether the ongoing actions areperformed in a good or skillful manner. In Politics Aristotle compares praxis withlife: “Life is action and not production” [4]. In life, and in praxis, humans do thingsout of habit. Yet, when something unexpected or new appears, the imperative isthat habits and sedimented behaviours are changed. Praxis or action, according toAristotle, requires phronesis or practical wisdom.While techne is apparent within theproduct (like a quality of the bread or the manufactured product) this is not the casewith phronesis. Phronesis shows itself in the very action, e.g. a surgeon’s capacityto tell a husband that his wife did not survive the operation or a teacher’s ability toconvince a reluctant pupil to engage in her maths classes.

According to Aristotle, phronesis is always related to a certain action carriedout by a specific person. It shows itself in the very doing. In modern language, weoften try to approach this ability in terms of judgement. This gives us a good pictureof the kind of knowledge to which Aristotle aspires. But, at the same time, wemust be aware that, when talking about phronesis, Aristotle has in mind somethingthat is much wider than epistemic judgement. It is not only the very capacity tojudge in a particular situation, it also involves handling and coping with the situationas a whole. Take the example of a surgeon; she shows her phronesis not only inreaching a certain judgement during a complicated operation, but, rather it is presentthroughout the process as a whole, in her way of using her art in order to preparefor certain moments, to remain attentive to the possibility of potential complicationsand so forth. Phronesis tends to reveal itself in unforeseen situations. One exampleof this might be if during an operation, a surgeon suddenly runs into a complicationsuch that the patient’s life is in danger and is thus under pressure to make a crucial


decision how to proceed. At these critical points, the surgeon’s confidence in herown knowledge and (technical) skills, information about the patient, as well as anoverview of the medical resources available, forms the basis for the judgement shemust make; the quality of the judgement made in this critical situation will showwhether she possesses phronetic knowledge. Phronesis is “knowing how”, but in adifferent sense to techne. The former is related to who the performer is—that is, withrespect to her educational history, previous experiences, her understanding as wellas the present interpersonal situation, e.g. the needs of the specific patient. Phronesisis always related to others. Moreover, while techne or art is displayed in products,phronetic knowledge is displayed in the actions of a particular individual, includingher judgements. Since phronesis is understood as something carried by a certainperson, it seems harder to understand it in collective terms, even if it takes place inthe interpersonal sphere of praxis.

Our claim is that phronesis might be understood as something that concerns thecollective. Thought in this way phronetic knowledge can be said to be displayedin actions that encompass the judgements of a group of agents irreducible to thejudgements of individual group-members. For example, if a group of physicians co-performing a complicated surgical operation, it is the group of agents that has toact phronetically. It is not enough for one individual to act in a way that displayspractical wisdom based on previous experiences. We can even ascribe to the groupa shared history, along with possible joint experiences and established interpersonalrelations.

In a basic sense, we understand an activity as collective when it is dependenton several agents. In some situations, as we have argued above, more than oneperson is required to perform an action. This is the case, for example, when a groupof police officers takes control over a riot. In other situations, it may be that anongoing action is handed over to, or supplemented by, another. Examples of thisare shift changes in healthcare or a team of teachers that seek to teach a groupof pupils about the parliamentary system in an inter-disciplinary way. One way ofexplaining the collective dimension of these activities would be to understand theindividuals involved as trained in a shared strategy and invested in a common goal.For example, the police’s ability to handle a riot could be the result of joint trainingas well as following distinct rules of engagement. The healthcare staff or teacherteam might be governed by common routines and jointly agreed goals and learning-outputs. Here the very possibility of forming a collective is based on the fact thatpersons share a certain formal knowledge, agreed rules or technical skills (techne).But these activities are also part of an interpersonal praxis that requires not onlyskillful interaction towards a predefined goal but also judgement regarding the veryprocess, experience and context of action [5]. Our claim is thus that these are activitiesin the sphere of praxis, for example when a group of policemen (implicitly) judgewhether they are facing the possibility of a riot or just sporadic aggression shown byparticular individuals. It is therefore our view that this activity is to be understood asoperating within the sphere of praxis and therefore as activities that call for a formof collective phronesis [12].


Can an Aristotelian understanding of phronesis be applied to a group of agents?Here the question is whether and if so how an action, which requires judgement inthe more complex sense of praxis, be understood as irreducibly collective. From theperspective of Aristotelian and Neo-Aristotelean thought phronesis is possessed byan individual (e.g. [2]). For example, according to Bela von Brandenstein, action, inthe Aristotelian sense, cannot be understood as collective, even if it takes place inthe interpersonal sphere of praxis. “Collective action” could only be understood asthe sum of individual actions or as a result of discourse.1 Actions that possibly canbe understood as collective actions, we-actions, are said to be parasitic on individualactions ([17], 679). Our concern is thus to maintain the complex Aristotelian under-standing of action, extending it in order to understand collective action as irreducibleto the acts of the individual agents involved.

3 Different Understandings of Collective Action

Within social epistemology there exists an ongoing debate about the possibility ofcollective agency. In this debate, one often speaks of (at least) two distinct under-standings of collective action, both of which are tied to the question of epistemicjudgement. In the first understanding, an individual must act upon a situation thatrequires judgement based on what is called “social evidence”, i.e. the judgements ofothers. Here, what is regarded as collective is not necessarily the concrete purportedact of judgement, but the character of the evidence. As stated above, in the field ofprofessional action, this often applies when different professions must collaborateto judge upon something, e.g. in the event of considering and deciding on a clinicaldiagnosis, when one physician will make a judgement on the basis of a range ofexpertise from other experts.

The second understanding of collective judgement considers situations where agroup of individuals act in concert with respect to a joint goal, and doing so in sucha way that seems to imply a common judgement. This act of judgement itself pre-supposes a form of agency that is irreducible to individual agents. Here the debatebroaches the question whether an ontological assumption of “group subjects” isimplied. On this, a variety of positions are adopted in the literature: eliminativismis an ontological position rejecting the idea of collective judgements by a “groupagent”; to talk about a collective agent is to do so only metaphorically. Instead,“anything ascribed to a group (…) can be re-expressed by reference to its members”(List/Pettit 2011, 3). A similar position, singularism, which is common in economictheory, criticizes group agency from a methodological perspective: according to thisposition, the introduction of “group agents” would imply the unnecessary introduc-tion of an entity to social ontology that is methodologically inaccessible. We cannot

1We will not further discuss the role of discourse as a presupposition for an understanding of We-Intentionality as it is e.g. discussed in relation to a Habermasian perspective, and as s expressedthrough the use of language in different social contexts (see [6]).


investigate collective acts without first relying on an analysis of individuals and theirjudgements. A further view, “non-eliminative individualism”, held by List and Pet-tit, combines both the abovementioned positions: Some uses of group agencies areindeed metaphorical, “but often the notion of a “group agent” expresses a correctobservation” (ibid. s. 4). For List and Pettit, we need to analyse judgements made bya group as if they had been made by individuals (methodological individualism) inorder do justice to the social world.

We cannot go any deeper into this quite complex and extensive debate. Whatis interesting for us is not the different positions adopted and solutions suggestedto the problem of group agency and social evidence but the very way the problemof collectivity is understood in the first place. In this sense, we share List/Pettit’smethodological approach. Both understandings of collective action—based eitheron social evidence or group subjects—are stated from a position exterior to thegroup that aims at a certain goal or object. Even “methodological individualism”treats action as a goal directed activity based on a judgement that can be describedon the basis of an observation.2 Yet we can imagine situations where a group ofagents, from an outside perspective are jointly related to a common goal and yet stilldo not act as a group. They act without an understanding of what it means to judgetogether, they lack a notion of each other’s position in relation to the situation inwhich they are involved. If we would ask a member of the “group”, she or he wouldnot be conscious about being part of it. We can equally imagine ourselves being partof a collective that might do things together, sharing certain activities, but all thesame we would not necessarily relate to such situations in a collective way. Insteadwe just follow the rules or habits of the group, doing the things the group usuallydoes or is supposed to do [6]. By taking our point of departure in examples fromwithin the everyday practice of professionals one can see that questions about thecollective are complex in a way that seems not to be fully captured by the differenttheoretical approaches we introduced above. In professional practice these differentcases of collectivity tend to intermingle. Let us take a closer look at an example froma professional practice.

4 A Night Shift

Our example is taken from an essay written by a police student, who gives a personalaccount of an experience from his work placement. As part of this account, thepolice trainee analyses his own reflections from different theoretical perspectives.3

2List and Pettit place emphasis on their commitment to realism, that is, to the world as an observableand describable object independent from the constitutive perspective of any subject (ontologicalrealism) (ibid. 26). At the same time, the “methodological individualist” is interested in contrasting“the methodological significance in detecting when agency is present, and when not” (ibid. 11).3This way of writing the final undergraduate paper is a method developed at the Centre for Studiesin Practical Knowledge. This specific essay-form is used at the Police education at SödertörnUniversity in Stockholm, Sweden, as well as at the Police University College in Bodø, Norway.


In our own analysis of the text, we will focus exclusively on the first part of hisessay. As a part of Swedish police education, students spend six months at a workplacement; during this period students are asked to work as police officers togetherwith their instructors and supervisors. The instructors are experienced police officerswhose responsibility is to introduce the student to practical police work, while thesupervisors are responsible for the students’ professional development. The studentdescribes his experience from a first-person perspective. Since we are dealing with abeginner’s perspective here, we are afforded especially interesting insights into howa trainee reflects on and questions established practices that more experienced policeofficers might otherwise take for granted. The essay also displays the police student’sdesire to be accepted as a fully-fledged police officer. For these reasons, this insideperspective makes the material interesting in light of the issue of the collective andcollectivity. The student wants to be part of the collective yet at the same time isuncertain of what this collective consists. The example we are drawing upon also isnonetheless limited, since we have access only to one person’s impression, and thusnether to the others in the group nor to the group.

The example describes an emotionally demanding police action that potentiallycould have turned into a violent and dangerous situation. This accentuates the ques-tion of phronesis. Situations involving a latent threat of violence, and thus possiblycan change its meaning, require professional judgement. An important part of thepolice’s professional knowledge is about acquiring an ability to read a situation,judging quickly with which situation they are dealing, and thus knowing how toact appropriately. It requires an ability to detect and manage threats of violence butalso to curb or respond to violence. Often these are situations where more than oneperson is required, which is why there seems to be a certain requirement not only forcoherence in the reading of the situation but also a demand for mutual trust betweencolleagues, both when it comes to reading possible threats of violence as well asthe ability to intervene and respond to any form of violence that may arise. This hasbeen described in research on police practices.4 The practical knowledge requiredfrom police officers in situations defined by risks of strong affections and violenceis a capacity to sense emotions, moods and signs of affection in different individuals[10, 11], but it is also a matter of communicating and sensing when, how and towhat extent authority or violence should be used [1, 10, 13]. These situations oftengive limited or no space for interpersonal deliberation or discussion. As Knutssonand Granér claim, it is impossible to have opinions on a colleague’s actions in anongoing violent situation [9]. The situation in the example below never turns intoa state of violence. Nonetheless, it does contain a risk of emotional volatility andviolence to which the police officers must relate.

The night shift starts with a meeting. The group leader describes the area where the oper-ation will take place and what spots we will focus on during the night. Some places aremore interesting for us than others, spots where drug-dealing and drug-abuse are common.

4Another crucial aspect it the professional capability to not provoke disputing persons to use violencebut instead to prevent the situation from becoming violent. This is discussed by the police officerMarkus Antonsson in an interesting text about how to communicate with groups of adolescents [3].


The group leader gives instructions of the places we shall focus on, such as car parks andbasements. He tells us that we are supposed to work very “aggressively”. We shall proceed“as usual”, seeking out, and frisk searching, suspicious-looking individuals.

Everyone sitting around the table seems to understand exactly what he is talking about,nodding affirmatively. To me everything sounds a bit vague, and I start thinking about whatI consider to be “suspicious,” and if this might be the same as my colleagues take it to be. Iwonder […] if I can live up to my colleagues’ expectations, what they expect me to know,and I especially wonder if I will manage to live up to my instructor’s expectations.

In the above excerpt we find some interesting aspects about the police trainee’sviews on collectivity and the collective. He claims that the instructions he receivesseem to presuppose knowledge he does not yet possess, and he expresses uncertaintyabout, for example, how he might determine whether or not an individual is sus-picious. We also see that he is anxious about not living up to his colleagues’ andinstructor’s expectations. He thereby shows that he has an idea, or a sense (at least)of some kind of existing collective. His insecurity could be described as a feeling ofinadequacy in relation to the knowledge and attitudes he ascribes to this idea of thecollective. What is obvious is that he aspires to be a full member of the collective.

Further in his essay the student describes how his squad is instructed to investigatean indoor car park often used by young people and adults to smoke cannabis. Hedescribes how his group organizes the search and closes off possible escape routes.Thereafter he gives a vivid picture of the actual search in the car park that turns outto be dark and thus difficult to survey in its entirety. After investigating this uneasyenvironment for a while, they run into two men:

In front of the back door, not more than one meter away from where my colleague is waitingat the other side, two people are sitting on red beer crates. One of them is wearing a suit,which immediately makes me wonder what he is doing here. The other person is wearingsportswear. I think to myself, biasedly, that [the second man] seems to be the sort of personone expects to meet down here. Neither of the two men make a big fuss when we show up.I show them my police badge and start talking to the two men while my instructor lets ourcolleague in from the backside. In a nearby trash bin I can see a packet of rizzla papers. Thisis a very common product used by cannabis-smokers. I am reflecting on the rizzla papers,the location, and the time, and whether these factors suffice to establish reasonable suspicionthat would warrant a frisk search.

In this part of the squad’s activity we can see that the ongoing police operation isessentially collective. The building could not have been both searched and securedby one individual solely. In a potentially dangerous situation they have to rely onthe fact that other persons are securing their part of the area and also that theyprovide adequate assistance in case of an outbreak of violence. Also in the situationdescribed above, the officers approach the two men and in doing so are expected toperform different tasks as the situation requires; the police officers are thus actinginterdependently. They are acting so to say as “one body”. This means that neitherone single policeman is able to handle the situation on his own, nor is the “operation”reducible to a cluster of individual actions performed by individual agents.

We also see that the police student is attentive to whatever he discovers or sees.Further he reflects on his impressions andwhatmight be expected of him.He displays


a strong attentiveness to the situation, its potential dangers and at the same time tohis colleagues’ actions and reactions. He relates to and reflects on numerous aspects:the group, the situation, his ideas of his colleagues’ expectations. This becomes evenmore evident in the following passage:

Both men are very cooperative and do precisely what we say. I think that it is a bit strangethat neither of them shows any signs of drug use. Can it be that I am missing obvious signs,or are they just not affected by drugs at all? While my colleague and I look after the men,my instructor is checking their ID over the phone. He tells us to come over to see what hehas found: absolutely nothing. Neither man shows up in the police register. My thoughtsare: “Ok, that’s good. These are nice guys.” My instructor asks me what I think we oughtto do with them. My mind is spinning: They show no signs of drug use, which speaks totheir innocence. What works against them is that there is rizzla papers close by. The time,as well as the place is also suspicious. I wonder what my instructor will think about meand my professional knowledge if I give the wrong answer. This makes it difficult for me todecide. I remember how earlier this evening the group leader told us that we should work“aggressively.” So, even if I don’t think we will find any drugs, I say: “Let’s frisk searchthem.”

Again, in this quote the collective aspects of the activity become evident. Twopolice officers “look after the men” while one officer checks their criminal records.The division of tasks does not seem to be preceded by any form of dialogue. Itseems as if they already have a common understanding about how the division isto be made. This is an example of how they collectively maintain their power andsuperiority in this public location that now has turned into their working place. Oncemoreweget the impression of themacting as one body. Then comes themomentwhenthe investigation of the location and the men reaches a critical point; the squad hassearched the area, confronted the two individuals,made an impression of them, talkedto them and received information about their potential criminal records. Preciselyat this point the squad must decide how to proceed. This is the moment where wecan see that the situation seems to involve a judgement (based on the precedingmoments in the whole operation), namely whether the two men ought to be labelledas suspicious or not.

We see this moment actualised by the instructor’s question to the student. He asksthe student if he thinks that the men ought to be frisk searched. The very fact that theinstructor poses this question shows that a judgement is required. We see from thestudent’s description that he perceives this as a test of his personal ability to judge.The police student then describes his own ambivalence: on the one hand, he does notbelieve there are sufficient grounds for a frisk search but, on the other, he runs intodoubts about his capability to read the situation correctly. We also see how importantit is for the student to appear as both a competent and a full member of the squad. Inhis description, he seems to have come to the conclusion that the other officers thinkthat the men ought to be frisk searched, and finally he concludes that the men shouldbe searched. Unsurprising, since he writes that he hopes to live up to his colleagues’expectations. This is nonetheless interesting; his account implies that while he hasan impression that the others concluded that the men should be frisk searched, hedid not. He thus ends up expressing what he believes to be the conclusions of thecollective.


The two men are thereafter searched and this very act seems to imply the judge-ment that the two men were acting suspiciously. We cannot know for sure that theact of frisk searching the two men is based on a judgement in an Aristotelian sense.However, what we can see is that if it were, it would be a collective judgement differ-ing from (at least) one of the individual judgements in the situation. If a judgementtakes place, it is not reducible to the sum of the individuals’ judgements.

5 Collective Phronesis?

The example contains an ambivalence with regard to different perspectives on thesituation: firstly, an outside perspective through which the reader sees the action andjudgement as collective appear; secondly, the inside perspective of the student thatshows the individual differences in relation to the collective action. Both perspectivesare relevant but they give us contradicting pictures of the situation. From an outsideperspective, the student tries to act “as a police officer”, he accepts the joint goalof the operation and acts according to it, as a part of a “we”. From the perspectiveof the young police student, he displays the activity of a lonely individual. Lackingany previous experiences of similar situations, the police trainee describes how heapprehends the goal of the operation as unclear, neither does he knowhow to achieve itnor how he is supposed to detect if a person is acting suspiciously or not. Therefore,the student as an individual does not possess the professional capabilities that hehimself ascribes to the collective: namely, in this context, the capability to judgewhether a person is suspicious and ought to be frisk searched.

The instructor’s question to the police student is a request for the student’s veryjudgement of the situation. Here it becomes evident that he is supposed to judge. Thestudent’s capacity to judge is put on trial. But as a trial of judgement the ambivalenceof the situation becomes explicit, since he is of the opinion that there seems to bea given answer. His impression is that he is supposed to judge in a way that iscoherent with the judgement of others. The novice officer feels obliged to adapt tohis colleagues’ judgements. From this individual, inside-perspective, the situation ischaracterised both by the student not being part of a group but also by the presenceof concrete others, others that could contribute with their expertise. The result of thisis that he does not trust his own interpretation of the situation. Instead, he ceases tojudge (and to include the others in his process of judging) but decides in favour ofthe supposed perspective of the group, and thereby paradoxically shows that he isnot part of a collective of competent agents. From an inside perspective, then, thestudent seems to rely on social evidence from the collective, though he could not besaid to take part in a joint act. When looked at from an inside perspective we alsosee his experience of an existing “we”, the existence of which greatly affects him.

Now, if we look at the situation from the outside, a different interpretation comesto light: the whole group of police officers can be regarded as acting collectivelyin a way that their action is based on something we could call a joint judgement,or rather a process of joint judging. After all, the squad as a whole did frisk search


the two men. If the operation leading up to and including the frisk search is to beregarded as an action in the Aristotelian sense, it implies that the men were judgedas suspicious individuals. It is through the judgement that the action (praxis) isdistinguished from goal-directed work (poiesis). In an Aristotelian understandingof action, the judgement cannot be reduced to a conclusion based on intentions andgoals set up in advance. In praxis, phronesismight emergewhen the individual comesto a judgement based on her knowledge and understanding of the actual situation.Judgement is in this sense essentially a trial. In order for a judgement to be collectivethis trial ought to imply an interpersonal aspect. What we know from this situationis solely the police student’s description of the dialogue with his supervisor and hisown sprawling thoughts on the issue. We have no way of knowing how the other twopolicemen thought or to what extent they were considering the situation. We thuscannot determine how they came to their own conclusions that serve to comprise thecollective judgement. What we do know is that the police student’s answer to hisinstructor is not an expression of his own judgement of the situation, but rather aneffort to fulfill a predefined task. The trial to which he responds is something otherthan the trial of the collective. If we get back to the distinction made by Aristotlebetween poiesis and practice, work and action, then, as a novice, the student displaysthat he is not in the sphere of praxis but in the sphere of poiesis. In an Aristoteliansense, he is thus not acting, he is just doing his work trying to accomplish his tasks,to fulfil a predefined goal. His answer to his instructor shows that his main purposeis to do his work as a good student. It is the product (i.e. giving either the right orwrong answer) that is the focus of his decision.

We have to acknowledge that it is not possible to know for sure if this collectiveactivity can properly be described as action in Aristotelian terms at all. It might bethe case that the two other police officers were just trying to fulfil the group leader’sorder and predefined tasks, and thus were not acting in a phronetic way at all. It mightalso be possible that the two other police officers ended up with identical individualjudgements. Therefore, it is inconclusive whether their activity is an example ofpoiesis or praxis. However, by highlighting the student’s trial, the actual meetingbetween the individuals’ perspectives becomes evident. Here at least there is spacefor the possibility of the formation of collective judgement.

For us, operating within an Aristotelean framework, there are at least three aspectsthat are important about the very possibility of collective phronesis: First, in an Aris-totelian understanding of action, the judgement is something attempted and formedin the act itself. Action entails intervening in an uncertain future, and this alwaysinvolves an element of risk, which manifests in the very act of passing judgement.And if we would like to understand the situation as an action containing phronesisand thus a judgement, it is apparent that the collective arrives at a judgement not inthe same way as individual persons. Here we run into a situation in which an actioncontaining a collective judgement cannot be explainedwith reference to participatingindividuals and their single judgements.

Secondly, if the example we have described above is understood in terms of anAristotelian conception of action then the police officer’s collective activity is shownin its complexity. Of interest, here is that the example from police practice shows that


collective judgement is not a cluster of individual judgements but rather the resultof a dynamic process. Collective action and judgement both emerge from within aplurality. This pluralitymight be explained as the police student not being a part of thecollective. But it might also be explained by the fact that a collective is dynamic in thesense that the individuals involved possess different levels of experience, knowledgeand goals. These different levels might be a part of the judgement.

A third aspect broaches methodology, specifically a question about the relationbetween different perspectives that arise within the situation. We see that none ofthe presented ideas of collective action suffice to explain the collective act in thisexample. The collective action cannot be explained solely as an instance of socialevidence, i.e., as a judgement based on the sum of plural judgements. The judgementof the police student might be said to rely on social evidence (or rather on his wish fortheir experience as deeper knowledge), but this is not the whole story about how hemakes his judgement. His judgement is rather determined by an aspiration to judgein line with his colleagues, and thus his decision seems to be strongly influenced byhis wish to be a part of the collective. This becomes clear once we have accountedfor the inside perspective of the student. From the outside perspective, we see acollective action that tends to imply a judgement. In that sense the action seems to bean example of group-agency. The fact that they decide to frisk search suggests thatthe squad has come to a collective judgement surrounding the suspicious men. Weknow that the student did not arrive at this judgement on his own. But as a collectivethis is the judgement implied in the action. The situation thus contains both group-agency and an idea of the importance of social evidence, as well as an emotionalunderstanding of the collectivity.

One difficulty with our example is that it can be interpreted in several ways. Itcould be interpreted as an example of corrupt police work in a gratuitous exercise ofpower. But it could also be interpreted as an example of the police student’s inabilityto detect or notice facets of the situation that were obvious for the more experiencedpolice officers. The empirical material we have analysed from the police student isinsufficient to determine whether the action is phronetic or ought to be understoodas an example of Aristotelian praxis. Nonetheless, what we have here is an exampleof an act containing a judgement that is irreducibly collective, or it is not as an actionat all, i.e. it is an example of work with a predefined goal.

6 Concluding Reflection

Wewould thus claim that by taking our point of departure in an actual example from aprofessional practice, action understood in terms of Aristotelian praxis might appearas irreducibly collective. We further claim that neither the framework offered bysocial epistemology nor by phenomenology can sufficiently explain how action thatincludes judgement can be considered collective. We suggest that an Aristotelianunderstanding of collective action must consider the dynamics of collectivity. Acollective consists of members with different skills, different levels of knowledge,


different emotional attitudes towards and understandings of collectivity and finallydifferent evaluations of a situation and other human beings. Collective judgementsform out of a plurality of perspectives.

References

1. Algård, J.: Det är inte olagligt att ta livet av sig [It is not illegal to commit suicide]. In: Ekstromvon Essen, U. (ed.) Den goda polisen. Sju essäer om reflekterad yrkeserfarenhet. [The VirtuousPolice. Seven essays on reflected professional experience]. PHS serie i polisiärt arbete, Solna(2008)

2. Annas, J.: Intelligent Virtue. Oxford University Press (2011)3. Antonsson, M.: En höna av en fjäder [Mountain of a molehill]. In: Ekstrom von Essen, U.

(ed.) Den goda polisen. Sju essäer om reflekterad yrkeserfarenhet. [The Virtuous Police. Sevenessays on reflected professional experience]. PHS serie i polisiärt arbete, Solna (2008)

4. Aristotle: The Nicomachean Ethics. Harvard University Press, Cambridge, Mass (1975)5. Collins, H.: Tacit and Explicit Knowledge. University of Chicago press, Chicago/London

(2015)6. Giovagnioli, R.: From habits to rituals: rituals as social habits. Open Inf. Sci. 181–188 (2018)7. Goldman, A.: Knowledge in a Social World. Oxford University Press, Oxford (1999)8. Goldman, A.: Experts. which ones should you trust ?. Philos. Phenomenological Res. 63(1),

85–110 (2011)9. Granér, R., Knutsson, M.: Perspektiv på polisetik [Perspectives on Police Ethics]. Studentlit-

teratur, Lund (2001)10. Heder, M.: Vad är en bra polis? [What is a Good Police Officer?]. In: Vad är praktisk kunskap?

[What is Practical Knowledge?], Södertörn Studies in Practical Knowledge, Huddinge (2011)11. Hoel, L.: Politiarbeid i praxis [Police Work in Practice]. Universitetsforlaget, Oslo (2013)12. Kinsella, E., Kinsella, A.: Phronesis as Professional Knowledge. Sense Publishers, Rotterdam

(2012)13. Krogh, U.: “Krona eller Klave”. In: Ekstrom von Essen, U. (ed.) Den goda polisen. Sju essäer

om reflekterad yrkeserfarenhet. [The Virtuous Police. Seven essays on reflected professionalexperience]. PHS serie i polisiärt arbete, Solna (2008)

14. Krueger, J.: Phenom. Cogn. Sci. 13, 533. (2014). https://doi.org/10.1007/s11097-014-9363-115. List, C., Pettit, P.: GroupAgency. The possibility, design, and status of corporate agents. Oxford

University Press (2013)16. Szanto, T., Moran, D.: Phenomenology of Sociality. New York, Routledge (2017)17. Von Brandenstein, BH.: Handbuch der Philosophischen Grundbegriffe, Bd. 2,München (1973)

From Searle’s Speech Acts to Smith’sTruth-Makers

Philip Larrey

Abstract Beginning with Searle’s meaning theory and his account of mental acts(“intentionality”), the article identifies the limits of Searle’s approach from the per-spective of common-sense reasoning. The second part explores the notion of “truthmakers” in the work of Barry Smith, inasmuch as it can be seen to complete a robustmeaning theory in the analytical tradition.

Keywords Meaning theory · John Searle · Truth makers · Barry Smith

1 Introduction

In the transmission of knowledge, a crucial passage is that which is analyzed by ameaning theory. In other words, the passage of understanding the meaning of aninterlocutor, on the basis of language.

Searle explains: “The hypothesis then of this work is that speaking a languageis engaging in a rule-governed form of behavior. To put it more briskly, talkingis preforming acts according to rules [1], p. 22.” Of course, rules are not enough,but they are necessary. And rules are by and large conventional, depending on thevarious linguistic communities. This is what Searle means by social structure, or theconstruction of social reality (likemoney, a driver’s license and the game of baseball).If all Italian language speakers agreed to convey the meaning of “rabbit” with theword “topo”, then when we spied a rabbit with friends from Florence we would say,“Ecco un topo”.

“To ask for the criteria for applying the name ‘Aristotle’ is to ask in the formalmode what Aristotle is; it is to ask for a set of identity criteria for the object Aristotle.‘What is Aristotle?’ and ‘What are the criteria for applying the name ‘Aristotle’?’ask the same question, the former in the material mode, and the latter in the formal

P. Larrey (B)Pontifical Lateran University, Vatican City 00120, Italye-mail: [email protected]


37

38 P. Larrey

mode of speech. So if, prior to using the name, we came to an agreement on theprecise characteristics which constituted the identity of Aristotle, our rules for usingthe name would be precise”.1

For most of the twentieth century in analytic philosophy, meanings were bannedor at least looked down upon. Quine’s lasting criticism of essentialism (where mean-ings separated from words become like Platonic essences separated from individualbeings) ushered in many decades of avoiding meanings and longing for a purelyextensional logic (because meanings were only operable within intensional log-ics). With Donald Davidson, we once again are able to speak about meanings andintensional content. Hilary Putnam’s quite famous article,Meaning of Meaning, [2]brought us full round as he maintains that the prephilosophical notion of mean-ing is almost completely mistaken. But even Putnam’s criticism of the “everyday’sunderstanding of meaning” has not held up entirely.

2 Meaning

A prephilosophical notion of meaning still has its merits. When we look at howeffective body language can be at transmitting meanings (independently of languageper sé) [3], we can conclude that there must be something that is at work between twointerlocutors who do not speak. Social Engineer, Inc., specializes in helping com-panies realize how information can be “stolen” or received by hackers through non-verbal communication (“…malicious hackers are able to exploit human weaknessesto obtain access to information and resources through manipulation and deceit”2).Recent hacks on Jeff Bezos’ cell phone,3 Sony,4 HB Gary5 and Lockheed Mar-tin6 all have non-digital components involved, which indicates that to some extent,meanings (and even information) can be separated from words and/or sentences. Wecan also recall how the Stuxnet virus was uploaded into the computer system at anuclear refining facility in Iran by the U.S. and Israeli military. This malware was adestructive hack which ruined many of the centrifuges that the Iranians were usingto produce enriched uranium. It infected the system because a plant worker found a

1Ibid., 172.2See Ibid., xvi.3See https://www.businessinsider.com/theories-on-how-national-enquirer-got-jeff-bezos-private-photos-2019-2.4See https://www.washingtonpost.com/news/the-switch/wp/2014/12/18/the-sony-pictures-hack-explained/?utm_term=.c0caf75333d3.5See https://arstechnica.com/tech-policy/2012/03/the-hbgary-saga-nears-its-end/. The saga nearsits end because almost all of the hackers were caught and put in jail.6See https://www.computerweekly.com/opinion/No-excuses-for-Lockheed-Martin-cyber-attack.Even more recently, it is quite apparent that the Chinese continue to get access to sensitive contentfrom the Defense Contractor. When we examine the Chinese new J-20, the resemblance to the U.S. F-22 is striking. See https://www.cnbc.com/2017/11/08/chinese-theft-of-sensitive-us-military-technology-still-huge-problem.html.

From Searle’s Speech Acts to Smith’s Truth-Makers 39

pen drive in the parking lot and unsuspectingly plugged it into his computer station.The very digital hack began in a very non digital fashion.

Slightly less philosophical could be the uncanny way that dogs (as well as otherdomestic pets) can actually grasp (in some non-linguistic sense) the meaning of theirowner, without understanding language. How does the dog realize what the ownerwants or how the owner feels, even though it has no access to linguistic structure?How can dogs react also to similar body-language signs in order to fetch, come, stay,or lay down? The emotional bond between certain types of dogs and their ownerstestifies to the presence of intensional content which is independent of normal verballanguage.

On Searle’s account of meaning,7 we have brain states and mental states whichcoincide. The brain causes meaning by causing mental states, yet there is reallynothing more to the mental state than the physical and chemical states in the brainitself. Because of this approach, Searle is often classified as a reductionist, reducingthe mental (or spiritual) to the physical components of the brain (cortex).

How do mental states get transmitted to our interlocutor? Usually through lan-guage. That is why each must understand the language of the other in order to convey“mental entities” to one another. If you understand the language of the other, thishappens naturally and easily. If not, it is more difficult but still possible.

Is it possible to communicate mental states without language? Yes, as we havementioned, through body language, pointing, sounds, and so forth. If those ges-tures are considered language, then the point is moot. However, the prephilosophicalnotion of meaning would seem to be defended, as the transmitting of meaningfulmental states between interlocutors (even in the case of the dog who comes and goesaccording to the wishes of the owner).

From a neurobiological point of view, all we can go on is the brain state, which wecan map quite successfully and even interpret what is happening in the brain. We cantake for example the work of Robert Knight and his colleagues at the University ofCalifornia, Berkeley (Neuroscience Lab). He writes, “We recorded electrical signalsdirectly from the human language areaswhen a person heardwords.We then decodedthese electrical signals and were able to turn them into sound files that reflected whatthe person heard, with remarkable accuracy”.8 Remarkably, the teamwas then able todecode speech when a person thinks of a specific word, from direct brain recordings.“The new techniques and mathematical processing of the brain signals got us closerto the details we need to extract the signals that are relevant for reproducing speech”.9

They are close to creating a prosthetic implant which can reproduce words on thebasis of the neurological activity in a person who is unable to speak.

7Above all, that found in his book, Intentionality. As Essay in the Philosophy of Mind. Also, hisshort work, The Mystery of Consciousness, his The Rediscovery of the Mind, and simply, Mind. ABrief Introduction.8See https://blog.frontiersin.org/2016/05/27/setting-free-the-words-trapped-in-our-heads/, byMonica Favre, Frontiers Science writer.9Ibid.

40 P. Larrey

With just a little imagination, we can extrapolate the example and discover thefollowing: a brain-to-brain connection which bypasses speech all together, and trans-mits the appropriate neurological activity between two interlocutors without the needto use spoken language.

Although this may sound like science fiction, it has already been tried. At theUniversity of Washington, Seattle, neurologists were able to establish direct brain-to-brain links, asking each participant a series of questions (without spoken language)and testing the results.10 The accuracy of correct answers was 90%. In addition, ElonMusk wants to have us all connected directly and has founded a company for thatpurpose, Neuralink, and has raised $27 million to fund it.11

A crucial distinction needs to be made at this point. Neurological activities andtheir patterns are not the same as thoughts. Thoughts are logical entities that areexpressed through words, and the brain states that correlate to thoughts are physicalprocesses and can therefore be mapped and transmitted through an implant. Thereis nothing science-fictional about that. But achieving that (which is a wonderfulthing for people who cannot communicate—think, for example, of the late StephenHawking) is not the same as achieving telepathy. If we do not reduce mental states tobrain states, then we conclude that the brain state is something more and immaterial.

3 Truth Makers

As a result of the elimination of the classical notion of propositio, W. V. Quinefaced the daunting task of coming up with a new logical entity that would replacethe propositio as vehicles of truth. According to the classical theory of logic, whatcarried truth valuewas not theword (or the sentence), but rather the immaterial logicalentity known as the propositio, which was understood to be the concept (or idea)which was not yet linguistically determined. Due to Quine’s criticism concerningthe lack of identity (“No existence without identity”), the idea of proposition had tobe eliminated. In so doing, Quine eliminated the traditional truth vehicle.

His novel response was to offer eternal sentences as proper truth-vehicles. Thesesentences were called eternal because they were not context dependent, i.e., theywere not sentences whose truth depended on certain contexts. His notable exampleinWord and Object was “George marriedMary andMary is a widow” [4]. Of course,here there are many contexts which are ambiguous: wasMary a widowwhen Georgemarried her? Is Mary a widow now (which would not please George too much)?The idea of eternal sentences for Quine overcomes the contingent difficulties of

10See http://www.washington.edu/news/2015/09/23/uw-team-links-two-human-brains-for-question-and-answer-experiment/.11See http://www.businessinsider.com/elon-musk-neuralink-raises-27-million-2017-8. This isMusk’s fourth company, and although the idea sounds absurd, people tend to take Musk seriously.


utterances whose truth value can change over time. Some critics of Quine have evensuggested that substituting eternal sentences for propositions turns out to be almostthe same thing.

From the point of view of the history of analytical philosophy, eliminating propo-sitions as truth vehicles required obtaining others. Barry Smith then suggested theconcept of truth-makers. “The fall from favour of logical realism brought with it acorresponding decline of interest in the ontology of truth. The notions of correspon-dence and indeed of truth itself first of all came to appear obscure and ‘metaphysical’.Then Tarski’s work, while rehabilitating the idea of truth, seemed to embody a rejec-tion of a full-blooded correspondence. (3) In the wake of Tarski, philosophers andlogicians have largely turned their attentions away from the complex and bewilderingdifficulties of the relations between language and the real world, turning instead tothe investigation of more tractable set-theoretic surrogates”.12

The reasonwhy truth-makers became important in the analytical tradition lieswiththe inheritance left by Tarski. The influence of his article on the semantic concept oftruth cannot be overstated. Simplifying that influence somewhat, it can be argued thattwo impressions were conveyed and affected almost all of subsequent trends in theanalytical tradition. The first impression was that in terms of a “correspondence the-ory of truth”, Aristotle’s analysis suffices to understand the prephilosophical notionof truth as the conformity between what is stated and what there is. This stancebecomes obvious from a common sense perspective: to say what is that it is, andwhat is not that it is not, is to speak the truth. If it is raining outside where you are,and you utter, “It is raining outside” means that the utterance is true. If it is notraining, and you say, “It is not raining outside” means that the utterance is true aswell. With almost all examples we can examine, Aristotle’s description adequatelydeals with what we intuitively understand as truth.

The second impression of Tarski’s game changing article is that the introductionof a “coherence theory of truth” no longer deals with real-world truth relations, butrather with truth for a formal language. In fact, the analytical tradition favored thecoherence theory of truth precisely because of the birth and development of formalsystems in the philosophy of logic. Although the formalization of language and logicssystems has its merit, it does not deal with real-world truth relations, and thereforedoes not concern itself directly with that which matters to those who seek the truthfrom a realist’s point of view.

Hilary Putnam concluded that “[I]f we want to account for truth, Tarski’s workneeds supplementing with a philosophically non-neutral correspondence theory”.13

12Reference [5]. The footnote used in the original article refers to Aristotle: “3. Aristotle’s famous

To say of what is that it is not, or of what is not that it is, is false, while to say of what is thatit is, or of what is not that it is not, is true

(Met., 1011b 32 ff.) is, as Tarski himself is anxious to claim (1944, p. 343), less than full-bloodedcorrespondence theory, but Aristotle is elsewhere (op. cit., 1027b 22, 1051b 32 ff.) prepared tospeak of truth reflecting ‘combinations’ of subject and attribute in reality.”13Ibid., 288.

42 P. Larrey

Such is the task that Barry Smith and his colleagues set out to accomplish withthe notion of truth-makers. Truth makers are more than truth-bearers (or in Quine’svocabulary, truth vehicles) because they are those rational entities in virtue of whichsentences or statements are true. The truth vehicles simply carry the burden of thetruth predicate (i.e., “is true”), whereas the truth-makers are those logical constructsthat determine the truth value of propositions.

Barry Smith’s initial point has to do with what he calls “Moments”. Momentsare objects that require other objects in order to exist at all; they are not ontologi-cally autonomous, but rather dependent on other objects for their existence. Someinteresting examples include a human smile (which requires a human face in orderto be) or a duel, which requires duelers in order to exist (although the “duel” is notreducible to the duelers).

Such a distinction can also be traced back to Aristotle, with his now classical dis-tinction between substantial (not in a subject) and accidental (in a subject). Althoughthere are nuances which differ between “moments” and “substances”, the notion isclose enough to be useful for the analysis of truth-makers. “Substantial” is not in asubject, but rather the subject itself; while “accidents” adhere in a subject and aretherefore not substantial. Such a distinction runs through the entire scholastic period(especially in Aquinas) as well as modern philosophy with Descartes, Locke andHume.

Smith summarizes: “[…]We shall here consider the theory obtained from the viewthat what makes it true that Socrates died is Socrates’ death [example borrowed fromRussell], what makes it true that Amundsen flew to the pole is his flight [exampleborrowed from Davidson], what makes it true that Mary is smiling is her (present)smile, and so on. Or, in other words, that for many simple sentences about spatio-temporal objects the truth-makers for these sentences are the moments picked outby gerundials and other nominalized expressions closely related to the main verbsof the sentences in question”.14

The notion of “gerundial” is crucial to the theory of what makes a sentence true.Simply put, what makes something true is it’s being thus. In the above examples,we can slightly modify the verbs by converting them into gerundials (or gerunds).Thus, we have “what makes it true that Socrates died” is Socrates’ dying, not inthe sense of an ongoing action or happening (i.e., not in the sense of the process ofSocrates’ death—“dying”—but rather in the sense of “Socrates having died”). Smithattempts to capture the action verb form in order to render the noun more dynamicand therefore more clearly a “maker” than just a truth vehicle.

The notion of truth makers extends beyond the normal boundaries of both theanalytical tradition and neo-positivism in general. One of the central tenants of theneo-positivist traditionwas verificationism, where themeaningof a statement impliedunderstanding the conditions under which the statement was true. Even in the twen-tieth century, such a theory came under fire and faced harsh criticism. One line ofcriticism (still valid today) is that I can understand the meaning of a phrase withoutknowingwhether it be true or false. However, a valid response from the neo-positivist

14Ibid., 296/7.


is that I cannot understand the meaning of an utterance without knowing what wouldmake it true or false, i.e., the truth conditions (from an empirical point of view).

Smith concludes: “A knowledge of truth-conditions takes us at most one steptowards reality: one can, surely, envisage understanding a sentence (knowing itsmeaning), whilst at the same time having only partial knowledge of the nature of itspossible truth-makers”.15 “Partial knowledge” signifies “given our current stage ofcomprehension”. Due to the constant progress of scientific knowledge, our knowl-edge is always partial compared to future knowledge. When Tiberius sat in his villaon Capri Island in the first century and pointed to the bay of Naples describing suchwaters as “magnificent and profound”, he had no idea that he was referring to H2O.His knowledge of chemistry did not get that far. Yet, what he meant and to whathe referred are very clear. Having only partial knowledge of what makes a sentencetrue does not lead us to skepticism, but rather takes us down the path of a full-bodiedepistemology in the realist tradition. Even though today we can substitute H2O for“waters of the bay of Naples” and the statement comes out true, such a substitutionsalva veritate was not possible for the Emperor two thousand years ago.

Aside from the neo-positivistic limitation in terms ofmeaning and truth, within theanalytical tradition the notion of conceptual schemes exerted an enormous amountof influence in the second half of the twentieth century. Thanks not only to Quine,but also Thomas Kuhn and others, truth became relative to one’s conceptual scheme.Quine referred to this as “ontological relativity”, in which what there is is deter-mined by the conceptual scheme. For Kuhn, who popularized the notion with histerm “paradigm” (and associated term “paradigm shift” describing how we abandonone paradigm in favor of another), what there is is determined by the paradigm. Ifreality is dependent on the conceptual scheme, so too are truth values (and even truthconditions). As Donald Davidson writes in his classic essay, On the very idea of aconceptual scheme16 “Reality itself is relative to a scheme: what counts as real inone system may not in another”.

If reality is relative to a scheme or paradigm, certainly truth is also. What istrue in one scheme may not be in another. Although this relativity of truth can beadjudicated within a realist tradition (where some truths are dependent on culture,history or even linguistic communities), that which is completely objective is themeaning of the truth predicate. What is true for a Chinese speaker may not be fora Spanish speaker, but what “is true” means is the same for everyone. Such is theobjectivity of the truth predicate across all natural languages.

From the perspective of truth makers, Quine’s ontological relativity cannot besustained. Davidson’s devastating critique of conceptual schemes concludes withone of his most famous paragraphs and sets up the need for a conversation concern-ing truth-makers. Davidson writes: “Nothing, however, no thing, makes sentencesand theories true: not experience, not surface irritations, not the world, can makea sentence true. That experience takes a certain course, that our skin is warmed orpunctured, that the universe is finite, these facts, if we like to talk that way, make

15Ibid., 299.16See [6], pp. 183–198. The quoted line is on page 183.

44 P. Larrey

sentences and theories true. But this point is put better without mention of facts. Thesentence ‘My skin is warm’ is true if and only if my skin is warm. Here there is noreference to a fact, a world, an experience, or piece of evidence”.17

Because of this paragraph, Davidson is often misunderstood. When he writes that“Nothing, no thing makes sentences true”, he remains a realist in his conceptionof truth. That which he attempts to critique is the effort to place an intermediarybetween a speaker’s language and reality she is describing (i.e., a conceptual scheme,a paradigm, a framework and so on). There is ‘no thing’ which renders a phrase trueaside from the truth-makers (as in Smith’s account), which simply put means thatreality is as described by the speaker. Davidson crystalizes the point: “In giving upthe dualism of scheme and world, we do not give up the world, but re-establishunmediated touch with the familiar objects whose antics make our sentences andopinions true or false”.18

InDavidson’s account,we do not need an epistemological constructwhich “maps”language onto reality, or fixes language to somenon-linguistic reality. The reasonwhythere is no need for such a construct is that language does not need to be “hooked onto reality” (as Hilary Putnam would suggest) precisely because language arises fromreality as described by a speaker (of any natural languagewhatsoever). Davidson alsopreserves objectivity by examining speakers of a language as a community and notas solipsistic individuals only. Speakers of a given language understand each otherbecausewith their words, they refer to the same reality (assuming they speak the samelanguage). This account is often referred to the “triangular approach” of Davidson,where the three points of the triangle are the two speakers and the denotated reality,already “hooked together” by a common language.

4 Conclusion

In conclusion, the analysis presented here traced the notion ofmeaning and intention-ality in John Searle as pertinent for the examination of truth and the truth predicate.By way of contrast with neo-positivism and the analytical tradition as a whole, BarrySmith offers his notion of “truth-makers” which are not merely truth vehicles orconditions of truthfulness, but rather that which makes sentences true or false. Inother words, using the gerundials, the being of that which is (in an empirical sense).Furthermore, such an account can be seen to dovetail well with Davidson’s accountof the “triangular” nature of human natural language, where there is no need of anepistemological or logical construct which would “hook” language onto reality (oron to a “fact” or on to an “event”). The notion of truth-maker adequately makesexplicit that which is implied by the Davidsonian account.

17Ibid., 194.18Ibid., 198.


References

1. Searle, J.R.: Speech Acts: an Essay in the Philosophy of Language. Cambridge University Press,Cambridge, MA (1969/1994) (17 reprintings)

2. Putnam, H.: Meaning of meaning. Minnesota Stud. Philos. Sci. 131–193 (1975)3. Hadnagy, C.: Unmasking the Social Engineer: the Human Element of Security. Wiley & Sons,

Indianapolis (2014)4. Quine, W.V.O.: Word and Object. MIT Press, p. 170 (1960)5. Smith, B., (University of Manchester), Mulligan, K., (University of Hamburg), Simons, P. (Uni-

versity of Salzburg): Truth makers. Philos. Phenomenological Research, vol. 44, pp. 287–321.(1984)

6. Davidson, D.: Inquiries into Truth and Interpretation. Oxford University Press, Oxford (1984)

Science, Ethics and Social Practices

The Biological Logic of Human Action:On the (Considerable) DifferenceBetween “Rational” and “Adaptive”

Douglas A. Marshall

Abstract Modern democratic and capitalist (i.e. classically liberal) societies areconstructed and justified on a foundational assumption that human action conformsto a rationalist logic, approximating, if not identical to, that of instrumental ratio-nality. This chapter argues that because human behavior has evolved to be adaptive,it adheres to a very different, biological, logic. After establishing the multi-layeredreasons that human beings could not, and did not, evolve to be rational, it out-lines the thoroughly social behavioral capacities and dispositions that could and didevolve, which together constitute the actual adaptive logic of human action. In anutshell, the resultant model replaces the rationalist model of individualist, single-minded, amoral, materially maximizing, informed, and calculating human actorswith an empirically and theoretically substantiable model of groupish, distractible,moral, Satisfycing, epistemologically conforming, and habitual actors. The chapterconcludes with a consideration of several significant implications of the improvedmodel for individuals, for social science, and for social policy/governance.

Keywords Evolution · Rationality · Sociology

1 Introduction

At the very origin of sociology isDurkheim’s rejection of instrumental rationality as aplausible basis for human society. Contra the then-prevailing view of sociation as buta calculated means of maximizing individual utility, Durkheim understood that anysuch foundation would crumble under the weight of what we would today recognizeas a “collective actionproblem”.Despite this early dismissal on theoretical grounds—and no shortage of subsequently discovered empirical shortcomings—the belief that

This chapter is distilled from a longer, unpublished, manuscript, which interested readers can soonfind, along with a comprehensive bibliography, at the author’s Academia.edu page.

D. A. Marshall (B)University of South Alabama, Mobile, AL, USAe-mail: [email protected]


49

50 D. A. Marshall

means-ends rationality is an innate and defining characteristic of human beings—most formally manifest in Rational Choice Theory—has persisted both inside andoutside of social science. As evidence accumulated that human beings neither cannor do behave according to rationalist models, their advocates sought to preservethe rationalist ideal by diluting it, a project that culminated in attempts to assimilate“rationality” to the evolutionary virtue of “adaptivity” by claiming that, despite allevidence to the contrary, human actors must, as a matter of biological imperative,behave “as if” they are instrumentally rational.

Such determination to preserve the ideal of human rationality is of a piece withWestern society’s larger dependence upon it. Rationalist logic infuses the folk psy-chology that modern individuals use to predict and explain the world, while at thesocietal level, it is foundational to modern economic, legal, and political systems,since the perceived legitimacy of both capitalism and democracy hinge upon citi-zens’ purported ability and disposition to rationally pursue their own interests. Thus,instrumental rationality amounts to the prevailing “logic” of human action in themodern world, in the informal sense of logic as that which renders behaviors orevents consistent, comprehensible, and justifiable.

This chapter proposes an improved alternative to the rationalist logic of humanaction, one that—unlike attempts to assimilate “rational” to “adaptive”—takes bio-logical fitness and evolutionary theory seriously, in both critiquing rationalist logicand constructing its successor upon them. It is precisely because human behavioris adaptive that the underlying logic of human action is not that of instrumentalrationality, but one constructed upon the preeminently adaptive value of sociality.

1.1 The Rationalist Assumption

At about the same time that Durkheim was dismissing rationality as a foundationof society, Max Weber was preoccupied with its implications for society. He sawthe perfusion of instrumental (“zweck-“) rationality throughout modern society andculture as “the master trend of history” [1]. For him, and many subsequent observers,the rational actor is:

Individualist. The rational actor is exclusively interested in their individual out-comes. Those of others matter only to the extent that they impact the actor’s own.Colloquially, the rationalist assumption is that individuals are selfish.

Single-Minded. The rational actor is fixated on a single goal to the exclusion ofall competing or ancillary ends. Accordingly, rational actors are commonly describedas amoral, since questions of right and wrong (like every other competing desiderata)are irrelevant to them.

Maximizing. The rational actor is insatiable, in that for them, more is alwaysbetter. Like “right” and “wrong”, the concept of “enough” is meaningless, as theirappetite for their particular maximand is boundless.

The Biological Logic of Human Action: On the (Considerable) … 51

Materialist. In theory, the rational actor can seek to maximize any conceivableend, but most applications gravitate towards material, usually financial, ends, giventhe analytical need for quantifiable, calculable, outcomes.

Informed. The rational actor is presumed to have access to, and to avail themselvesof, any and all information of relevance to the pursuit of their goal.

Calculating. The rational actor uses that information to calculate the impact thatof each of the options available to them would have on their goal in order to selectthe optimal means of maximizing it.

The best-known shortcomings of the rationalist assumption concern the fact thatthe sheer amount of information and calculation it requires far exceeds human cog-nitive abilities. Such findings gave rise to “bounded” models which stipulated thatwhile actors might not be able to obtain all relevant information or to calculate allpossible consequences, they will at least avail themselves of whatever information,and perform whatever calculations, they can. But even this “thinner” rationality wasundermined by further discoveries that people regularly fail to make use of evenreadily accessible information, or to engage in any more calculation than they abso-lutely have to. We are, in short, “cognitive misers” [2]. Most damaging of all hasbeen evidence that human behavior diverges from rationality in ways irreducible toeither limits or laziness—that it is “predictably irrational” [3]. Thus, the primaryshortcoming of the rationalist logic of human action is that human beings behaveaccording to a different logic altogether.

This chapter first lays out the nested “negative” arguments as to why evolutionwas highly unlikely to ever produce the capacity or dispositions necessary to aninherently zweck-rational species, and then retraces those arguments to construct the“positive” case outlining the biological logic of human action.

2 The Unlikely Evolution, and Dubious Adaptive Value,of Rationality

The justification for assimilating adaptivity to rationality goes like this: Individualfitness is primarily a function of survival, and survival, in turn, is primarily a functionofmaximizingmaterial resources. Since rationality is by definition the optimalmeansof maximizingmaterial resources, and evolution invariably finds and follows optimalsolutions, organisms inevitably behave rationally. This chain of reasoning includesat least three dubious assertions:

1. That evolution inevitably arrives at optimum solutions.2. That survival is primarily a function of material resource maximization.3. That individual fitness is primarily a function of survival.

Let us address each in turn.

52 D. A. Marshall

2.1 Biology’s Blunt Tool

The idea that evolution is inherently optimizing was popularized by a subset ofsociobiologists called “panglossian adaptationists” by their critics [4]. Evolution’sability to produce ingenious solutions is easily confounded with the more rarifiedability to produce optimal solutions—consider such brilliant (but decidedly subop-timal) elements of human physiology as upright posture (chronic back pain/injury),bipedalism (perinatal mortality), and the conjoined trachea and esophagus thatmakesspeech possible (significant choking risk). It is, in fact, hard to imagine a less intelli-gent or intentional mechanism than evolution. Neither speedy nor subtle, it presumesso few preconditions that its operation is ubiquitous across genetic, biological, andeven non-biological systems, including neural development, the immune response,operant conditioning, the market economy, peer review, and the ecosystem of likesand retweets that is the currency of social media. Nevertheless, it remains a blunttool with which to shape organisms’ physiology, much less their behavior.

That evolution produces such sub-optimal structures is guaranteed by a handfulof its inherent features:

The Primacy of Variation. Despite its theoretical and chronological primacy,the “variation” part of “variation and selection” typically gets short shrift. But it isa key reason that rationality is unlikely to evolve. Put simply, an optimal solutioncannot be selected for unless and until it emerges, and since variation is random andunplanned, the odds against the single most optimal trait on any dimension emergingare considerable.More important than the lowprobability of any single variant arisingis the much lower cumulative probability of the entire series of variations necessaryto produce a given adaptation occurring in the proper order (i.e. its emergence is pathdependent). In evolution, sometimes you really can’t get there from here.

A Passive and Retrospective Process. The “selection” part of “variation andselection” implies an active choice between options. But the actual process is anegative one of “de-selection”, in which competing alternatives are instead elimi-nated. Such a process is blunt because its sole standard is whether the trait in ques-tion is “good enough” to avoid deselection. Thus, it is a fundamentally satisfycingrather than optimizing process. Moreover, it is a retrospective rather than prospectiveprocess, incapable of looking ahead to what might happen, only back at what hashappened.

A Blurry View of Genes. Panglossian adaptationism—the idea that whateverhas evolved must the best possible solution to some evolutionary problem—ulti-mately ran aground on the implausibility of genetic optimization.Becausemost genesare non-particulate and multi-focal, there is no simple one-to-one correspondencebetween genotypes and the phenotypes they code for. Between such complications aspleiotropism, spandrels, horizontal gene transfer, epigenetics, gene-culture coevolu-tion and sexual selection, evolution’s “view” of the genes it is acting upon, as filteredthrough the phenotypes they code for, is downright blurry [4]. The more generalized


problem is that evolution can’t even “see” (or select) any single gene’s phenotype,only entire organisms, the cumulative products of tens of thousands of interactinggenes.

Between the primacy of variation, the fact that evolution is a negative and ret-rospective process, and its blurry view of genes, it is extraordinarily unlikely thatthe specific collection of traits that together constitute rationality could ever haveevolved, even were rationality an optimally adaptive pattern of behavior—which itisn’t.

2.2 Rationality Is Orthogonal to Survival

The belief that instrumentally rational behavior necessarily optimizes fitness reflectsa fundamental misunderstanding of the adaptive problem faced by organisms, in thatit reduces fitness to a problem of material maximization—most often of calories. Buteven the proximate aim of survival, much less the ultimate goal of fitness, has littleto do with material maximization. At best, rationality is orthogonal to (uncorrelatedwith) survival, but more commonly antithetical to it. To see why, let us consider thedefining features of rationality on a case-by-case basis.

Individualism. The supposed survival value of individualism derives from a con-ception of evolution wherein the primary competition for survival is that amongindividuals. But interindividual competition is but one of many kinds of evolution-ary challenge. Darwin helpfully clarifies the field by distinguishing between “abi-otic” and “biotic” competition [5]. Abiotic competition refers to non-living thingsthreatening one’s fitness—gravity, storms, heatstroke, hypothermia, dehydration, andstarvation. Biotic competition (living threats to one’s fitness), further breaks downinto: Inter-specific competition—threats posed by members of other species, and;Intra-specific competition—threats posed by other members of one’s own species.This latter category further breaks down into inter-group (e.g. raiding, warfare) andintra-group (i.e. inter-individual) forms.

Of these many varieties of competition, only one subset—biotic, intra-specific,intra-group competition—pits the individual against other individuals in their group.In every other instance, the individual’s own survival is positively correlatedwith thatof the othermembers of their group, since their ability to overcome such challenges ishighly dependent upon the presence and cooperation of their fellow group members.All those extra eyes and ears increase foraging efficiency and vigilance, while allthose extra arms and legs augment the individual’s capacity for self-defense in theface of both inter- and intra-specific challenges.

Most importantly, proximity to, and the goodwill of, others provides the individualwith a priceless buffer against the vagaries of life, since said others can provision theindividual with food, water, transport, and protection during those inevitable periodswhen injury, illness, childbirth, or other temporary disability means they can’t obtainthese for themselves. Without comrades able and willing to do this, such episodeseasily turn lethal.

54 D. A. Marshall

Consequently, individual survival is strongly and positively correlated with theexistence, health, and positive regard of other group members. Which is why, contrathe rationalist assumption, careful consideration of the problem of survival leads usto believe that human beings are more likely to have evolved dispositions to eschewindividualism than to engage in it.

Single-Minded. One of the more egregious faults of the rationalist understand-ing of survival is its conflation with caloric maximization. Although starvation is anot-uncommon cause of death across the natural world, it is hardly the only one.Dehydration, predation, hypo-/ hyper-thermia, accidental injury, infection, and con-tagious disease are no less lethal. For an organism to focus too closely on avoidingany one of these is to invite catastrophe due to another.

The notion that single-mindedness enhances survival is an overgeneralizationfrom economics, where money is a uniquely fungible maximand, convertible intoalmost any desideratum. But the world that humans evolved in and for is one ofstubborn unfungibility, such that a surplus in one domain conveys no advantageagainst threats from any other. Accordingly, human nature mitigates the such risksvia individuals’ notoriously short attention spans and easy distractibility. While theability to exert control over one’s own attention is so indispensable to success todaythat those lacking it are subject to medical and educational intervention, our limitedand fugitive attention is, evolutionarily speaking, a feature rather than a bug.

MateriallyMaximizing. Even if nutritionwere the sole decisive factor in survival,caloric maximization would not address the real problem thereof. Quantitativelyadequate nutrition is vitally important, but maximization proffers a one-dimensionalsolution to a multi-dimensional problem. It is the temporal, not quatitative, dimen-sions of nutrition that render maximization so inadequate. From the moment of for-age, food begins to spoil, losing its nutritional value and potentially becoming toxicand/or infectious. Preservation techniques are but labor-intensive delaying tactics ina perpetually losing battle.

Of course, preservationwould be unnecessary if not for the physical andmetaboliclimits on how fast organisms can ingest, and digest, food. Combined with food’s ten-dency to spoil, this creates a context inwhich there is little advantage to securingmorecalories than necessary at any given point in time, since any surplus may well go badbefore it can be consumed. Maximization can become downright disadvantageouswhen one adds in the costs of doing so in terms of energy expended and the increasedrisks of predation, accidental injury, and elemental exposure thereby incurred—notto mention the possibility of over-extracting a resource beyond its ability to replenishitself.

This issue of elective risk highlights an important difference between the logicof rationality and that of biology. In a rationalized domain like economics, rewardsare proportional to risks, so maximization is best pursued via brinksmanship, but inthe natural world the costs of underestimating risks can easily be catastrophic, so theoperative calculus is one of abundant caution, not brinksmanship. Such risk-aversivetendencies in humans appear throughout the cognitive bias literature where subjects’preference for maintaining status quo rather than maxmizing has posed an enduringchallenge to rationalist models [6].


The ultimate rejoinder to the idea of human maximizers is the familiar phe-nomenon of satiation. However cold, thirsty, tired, or hungry an individual maybe, once sufficient warmth, water, sleep, or food is obtained, further consumptionbecomes aversive. That all of humans’ homeostatic drives are similarly subject tosatiation is compelling evidence that we are not only not naturally maximizing, butinnately disposed against it.

Recognizing the temporal dimension of nutrition also brings the true nature ofthe challenge, and yet another essential distinction between rationalist and adaptivelogics, into focus: Though an organism may be unable to make use of any morecalories at time “A”, they will inevitably need more at some not-too-distant time“B”. Even in the narrow domain of nutrition, the essential problem is not one ofmaximization, but the substantially different and more complex one of continuity.The same applies to every other dimension of survival: Being warm or protectedtoday does nothing to negate one’s odds of freezing to death or being ambushedtomorrow. In conjunction with the aforementioned lethality of single-mindedness,this suggests that human behavior has evolved not to emphasize having as much aspossible of anything, but rather to always have enough of everything. In a nutshell,survival is a lot more like juggling than throwing shot-put.

Informed. The idea that humans make decisions in light of complete informa-tion has always been a straw argument, since it is obvious that crucial data is oftenunavailable, or exceeds human perceptual and memory capacities. But the biggerobstacle to a fully-informed individual actor is less one of ability than of motiva-tion. Ironically, it is the capacity for human rationality which insures that, left toour own devices, we seldom seek to acquire anything like “complete information”.Understanding why requires a bit of background:

Across the natural world, most forms of behavior are effectively under controlof the environment (i.e. correlated with daylength, temperature, etc.). But amonghumans, the evolutionary premium on ever-faster responses to changing conditionsproduced a capacity for flexible, internally instigated and directed behavioral control[7]. This is possible thanks to our capacity for constructing an internal model of theworld—a “schema” or “nomos” [8] that makes the internal simulation of potentialbehaviors and their consequences—in a word, calculation—possible.

The individual’s dependence upon that internal model creates competing epis-temic needs that regularly take priority over those for accurate information, as actorsbalance accuracy against the need to protect their nomos from epistemic threat. Putsimply, we’d rather believe we are right than be right, so individuals are typicallyhighly selective in their search for, and weighting of, information, so as to avoid ordownplay nomos-inconsistent data.

More generally, while rationality puts a premium on complete and accurate infor-mation, biological adaptation is much less persnickety, in that incomplete and inac-curate is often good enough for its purposes—and sometimes even superior. Forexample, as Harari points out, large-scale human sociality has long depended uponindividuals’ ability to subscribe to such patently imaginary constructs as nation andideology [9]—the implication being that our success as a species is a function of ourgenerally credulous and non-critical stance towards information.

56 D. A. Marshall

Calculating. Cogitation, like any other activity, is effortful, so human beingstend to be miserly with their thinking. Even when the necessary calculations lie wellwithin an individual’s ability, they generally won’t perform them unless they have to,relying instead upon habits and heuristics. Indeed, most human behavior is largelyautomated, with actual cognition engaged only in the face of a completely novelproblem or in the wake of an abject failure of habitual responses [10]. Such relianceon automaticity itself poses a fundamental challenge to the rationalist assumption,since purposiveness has long served as a last ditch definition of rationality. If nothingelse, being “rational” meant that people’s actions were consciously and intentionallytaken. That most day to day behavior is driven instead by automatic habit, tradition,and custom constitutes a major departure from even minimalist models of humanrationality.

2.3 Fitness Is Irreduceable to Survival

Fitness—an organism’s ability to leave viable copies of its genotype—is the undis-puted currency of evolution. Having elaborated the ways in which survival is ill-served by rationality, the task here is to demonstrate that even if rationality werean asset to survival, it still wouldn’t be an evolutionary imperative, since fitnessentails much more than survival—and its other elements are no better fulfilled byinstrumental rationality than is survival.

So just what is fitness? One venerable answer is that it is the ability to haveoffspring that can have offspring, a formulation which still well-captures the breadthand depth of the evolutionary challenge facing organisms. Thus, for most intents andpurposes, fitness comes down to an organism’s ability to make grandchildren.

How to Make Grandchildren. Making grandchildren, and thereby achievingfitness, is a complex and multi-dimensional task, But the problem is parsable into ahandful of sub-tasks. In roughly chronological order, these are:

Task 1—Survive to (and Through) Sexual Maturity. Though survival is not theentire challenge of fitness, it is something of a prerequisite. Given its elaborationabove, let us simply reiterate the point that, contrary to rationalist models, it isfundamentally an exercise in continuous, multi-dimensional satisfycing.

Task 2—Gain and Maintain Access to Potential Reproductive Partners. The issueof reproductive access has both quantitative and qualitative dimensions. Quantita-tively, the relatively low probability of conception per act of intercourse amonghumans implies a fairly linear relationship between an individual’s frequency andduration of access and their fecundity. Qualitatively, there is a premium on partnersin possession of particularly fitness-conveying genes and/or socio-cultural capital.Though males’ interests arguably lie more on the quantitative side while females’interests arguably lie more in the qualitative realm, both have an interest in gainingand maintaining access to the largest possible pool of potential partners, as it makesboth a greater number of, and a greater selectivity among, partners possible.


Task 3—Insure Survival of Mother-Child Pair Through Weaning. Assuming thatthe individual survives until sexual maturity and has access to a reproductive partner,copies of their genes may now reside in a fetus. At this point, the father’s fitness isno longer dependent upon his own health and survival, but both his and the mother’sfitness remains dependent upon hers. Thus, they each have a profound interest in themother-child pair’s ability to successfully survive the singularly perilous prenatal,perinatal, and early postnatal periods.

During the roughly nine months of gestation, the physiological changes of preg-nancy (morning sickness, compromised balance, connective tissue laxity) make itincreasingly difficult for the mother to forage, defend herself, and everything elsenecessary to maintain the pair’s health. Though the debility of pregnancy ends withchildbirth, that process is itself plagued by high fatality rates for both mother andchild among humans, due to the unfortunate conjunction of upright posture and largeheadsize. For mother-child pairs who survive gestation and birth, surviving the post-natal period through weaning is no less precarious. If anything, the transition fromcarrying the child internally to doing so externally makes everything more, not less,difficult for the mother.

Task 4—Insure Offspring Survive from Weaning to Sexual Maturity. After wean-ing, themother’s fitness, too, becomes less tightly coupled to her own health, but bothparents retain their evolutionary interest in the offspring’s survival to an age where it,too, can reproduce. During this period, a child gains greater degrees of mobility andfreedom from direct parental surveillance (since its mother may well be occupiedwith another infant)—but is thereby exposed to increased risks of injury.

Task 5—Equip Offspring with the Social and Cultural Capital Necessary to Sur-vive, etc. The decade andmore between weaning and sexual maturity among humansprovides a window for the transmission of the cultural tools of language, beliefs, andpractices that are the forte of our species, and the key to our evolutionary success.

An individual’s fitness—their ability tomake grandchildren—is a joint function oftheir success at each of these five tasks, none of which is advanced by the rationalistlogic of human action.More importantly, however, they all are advanced by behaviorsthat accord with a very different logic of human action—one whose outlines we cannow begin to sketch.

3 The Biological Logic of Human Action

Having worked our way through three discrete ways in which the assumptions ofevolved rationality are untenable, we now work our way back through these argu-ments in order to develop an alternative, theoretically and empirically credible logicof human action.

58 D. A. Marshall

3.1 Fulfilling the (Real) Tasks of Fitness

In stark contrast to the rationalist emphasis on individual resource maximization,the actual currency of fitness—making grandchildren—is eminently amenable to,and indeed dependent upon, social solutions. Each of the constituent elements offitness enumerated above is effectively impossible for any individual to fulfill alone,and is rendered tractable only through an individual’s membership in an intact andfunctional group.

That a single expedient like sociality could be so comprehensively advantageousto fitness is noteworthy, but unsurprising, given that the adaptive value of socialityis the unifying theme of biology on this planet. Every turning point in the history oflife comes down to some set of once-independent entities securing a fitness advan-tage over their competitors by suppressing inter-individual competition enough tojoin forces and compete collectively: Genes coalesced into collectively-reproducingchromosomes, differentially specialized prokaryotes combined to form comprehen-sively capable eukaryotes, uni-cellular microorganisms amalgamated themselvesinto multi-cellular organisms, and organisms—from crows and rhinos to homosapiens who subsequently organized themselves into murders, crashes, tribes andcorporations [11, 12].

To connect sociality’s overall adaptive value to the more concrete task of dis-cerning the logic of human action, let us consider what it contributes to each of theconstituent tasks of fitness.

Task 1—Survival. As already established, the major complications of survivalderive from the need to simultaneously and continuously meet a panoply of homeo-static requirements. For simplicity, let us further break this task down, starting withnutritional needs.

Individual forager success varies on a daily basis, but also adheres to larger pat-terns—good periods of recurrent surpluses, and bad periods of growing deficits,potentially culminating in a starvation spiral. The frequency, duration, and severityof bad periods can be ameliorated for every individual in a group via the simple expe-dient of collective foraging, whereby group members pool their daily yields. Overtime, as the identities of the successful and unsuccessful foragers change, the averagegroup member’s lifetime risk of death by starvation is estimated to drop from 27 to3% [13]. To be sure, sharing by itself may not completelymitigate starvation risks butjoint foraging also boosts provisioning efficiency, as having more, and more widelydispersed eyes and ears in the field increases the likelihood of gleaning whateverresources an area has to offer. Even greater advantages accrue from the advantagesof scale gained by cooperating in pursuit of larger game via cooperative hunting.Other abiotic threats are similarly ameliorated by group membership, e.g. extra eyesand ears are as advantageous in finding water and shelter as in finding food. Butthe most important way the group insulates the individual from abiotic threat is viatheir provision of support in the event of a fall, break, sprain, frostbite, burn, or otherepisodic debilitation, thereby preventing a temporary incapacity from escalating intopermanent lethality.


In terms of biotic threats, the group’s redundancy and wide dispersal of eyes,ears, and attention significantly decreases each individual’s vulnerability to macro-predators and conspecific attackers by making the detection of such threats morelikely [12]. For example, though circadian rhythms mean that each individual organ-ism is asleep, and thus singularly vulnerable, during some part of each day, inter-individual differences in sleep patterns within any group of conspecifics mean thatthe group as a whole has fewer and smaller lapses of vigilance, and thus a reducedvulnerability to surprise attack relative to any individual. Moreover, once the groupdetects a threat, it can address it more effectively by mounting a collective defense orcounterattack. This capacity for collective counter-aggression also deters the kind ofconspecific inter-group violence that characterized premodern human life [14]. Note,too, that each group’s ability to deter attack and defend itself is directly proportionalto its size, its solidarity, and the health of its individual members.

We come at last to the inter-individual competition which the rationalist modelassumes to be ubiquitous. Here, one’s fitness really might come at the expense ofanother’s, potentially evoking selective pressures for individualist behaviors.Withoutdenying that such competition exists, the pressure it exerts is insufficient to instillindividualism as a defining feature of human nature because it is, as noted, but onethreat to survival among many- the rest of which collectively present formidableselection pressures against it. Moreover, even inter-individual competition has acooperative dimension, in that success in such instances is more often decided bythe size and status of one’s allies than by one’s individual strength.

Task2—Reproduction. Group living is obviously advantageouswhen it comes tomaintaining access to a sizeable pool of potential partners. Among humans, with ourfrequent and recurrent periods of concealed fertility, near-continuous sexual receptiv-ity, and frequent recourse to polygamy/polygyny, more-or-less constant proximityto potential partners is integral to fitness. Beyond the mere quantity of potentialpartners, group living provides a context in which to evaluate their potential contri-butions to offspring, since some fitness-enhancing traits—e.g. charisma, status, intel-ligence—become apparent only upon the kind of extended observation that groupliving allows.

Task 3—Mother-Child Viability. Sociality is most decisive when it comes toinsuring that the mother-child pair survives their pre-natal, peri-natal and post-natalinterdependency. Pregnancy, childbirth, and its immediate aftermath represent anextended instance of temporary debility for the mother, during which other groupmembers become indispensable provisioners of their survival needs. Following birth,the child’s growing nutritional needs soon outstrip any individual’s ability to pro-vide them by herself, so the alloparenting (communal nursing, kiss-feeding) that thegroup provides proves indispensable to the pair’s survival, and thereby to her and thefather’s, fitness [15].

Task 4—Offspring Survival. Any offspring’s ability to survive from weaningto sexual maturity is complicated by the fact that humans are born “unfinished”,bereft all but a token arsenal of instincts [8]. The period during which offspring areacquiring their nomos (i.e. “childhood”) is one of ignorance-fueled vulnerability. Achild’s odds of surviving are, yet again, very much a function of their membership

60 D. A. Marshall

in an intact and formidable group that can protect them from said ignorance. Here,too, both parents’ fitness is dependent upon; (a) the existence of such a group and,(b) their offspring’s ability to join and remain a member of that group.

Task 5—Socializing Offspring. Because human beings are born “unfinished”,both parents’ fitness is tied to their offspring having acquired the knowledge andpractical skills necessary to their own success at all the tasks of fitness in theirparticular physical, social, and cultural contexts. This requires the kind of reiterationand repeated reinforcement, ubiqutous behavioral modeling, and extensive practicethat no parent, or parents, can provide by themselves.

In sum, it really does take a village to make grandchildren—and thereby forindividuals to achieve evolutionary fitness. But villages don’t occur by chance—theiremergence is dependent upon the existence and ubiquity of individuals possessingan extensive set of behavioral capacities and dispositions that make them possible.For such traits to be so invariant among human beings that tribes and societies canreliably be assembled from them suggests the influence of strong and consistentselection pressures throughout human evolution.

3.2 The Evolvability of Sociality

The lesson of “Biology’s Blunt Tool” in Part I is that however adaptive any particulartrait might be, there is no guarantee that it actually will evolve. In the abstract, this isas true of the traits of human sociality as of human rationality. But human socialitywas much more likely to evolve than human rationality for at least three reasons.

In the first place, sociality is a bigger target than is instrumental rationality. Thelatter involves the co-occurrence of several very specific capacities and dispositions,the probability of each of which is small, and the cumulative probability of all isvanishingly so. Sociality is a larger target in that there are many different ways ofbeing social—any of which are consistent with the hypothesis that some kind ofsociality would evolve. While the particular ways that humans are social very muchmatters, the biological model doesn’t try to deduce the particulars ahead of time, butinstead induces them from the empirical literature.

Secondly, relative to rationality, sociality has a much longer runway. Since onlyhumans have typically been ascribed the quality of rationality per se, one presumesthat its constituent traits would have had to evolve within the three hundred-thousandyear history of homo sapiens, or at most, the three million year history of the homogenus. Either figure is an implausibly short time for the several defining features ofinstrumental rationality to have arisen, been selected for, and dispersed throughoutthe species. By contrast, the pedigree of sociality stretches back hundreds of millionsof years. The net effect was that our species was evolutionarily primed for sociality,making further social adaptationsmore likely to emerge over time because the geneticpathways to them were already well-established [12].


Finally, sociality has a clear leverage advantage over rationality. As we’ve seen,sociality is advantageous in multiple ways across each of the tasks of fitness, increas-ing the number of ways and reasons a prosocial trait can be selected for once itemerges. Furthermore, it benefits from a positive feedback loop, in that the advent ofsociality brings new, socially-mediated selective forces like reputation and assortiveassociation into play.

But sociality has an evenbigger leverworking for it, in the formof inter-level selec-tion. Here—in the conflict between a lone individual and a collective competitor—iswhere the real fulcrum of sociality lies. In a population of competing individualorganisms, any two individuals who can mutually suppress their competition againstone another enough to cooperate, even minimally, against their solo competitors hasa decisive advantage over them in every dimension of fitness, rendering individualitya clear and ongoing liability.

3.3 Elements of a Biological Logic of Human Action

The overarching argument here is that belonging to a group is of such pre-eminentadaptive value that human behavior can be usefully understood as the product ofmyriad capacities and dispositions that evolved to enable individuals to achieve andmaintain membership in an intact and formidable primary group. More specifically,human beings have evolved so as to avoid the four basicmeans bywhich an individualmight become isolated: (1) Failure to join or stay with a group; (2) Being ostracizedfromone’s group; (3)Dissolution of the group fromwithin, and; (4)Destruction of thegroup fromwithout. Accordingly, we should expect human nature to be equipped andstrongly disposed to fulfill four social imperatives, to: (1) Firmly attach themselvesto one or more groups; (2) Avoid turning other groupmembers against the individual;(3) Strengthen the solidarity of the group as a whole, and; (4) Increase the numberand health of its other members. Unsurprisingly, these imperatives map closely uponthe empirically established elements of the biological logic of human action.

From Individualism to Groupishness. In addition to satiation, the over-eagerpursuit of individual ends is tempered by a host of countervailing forces, beginningwith our capacity for collective identification—for seeing ourselves as but part ofa larger social entity [16]. We are breathtakingly predisposed to divide the worldinto “us” and “them”, and to thereafter treat the resultant ingroup and outgroupsin profoundly different ways [17]. At the same time, we are both “pulled” intogroup membership by robust mechanisms of attachment, affiliation, and liking, and“pushed” away from individuality by our capacity for social pain, which hijacks thephysiology of physical pain to create acutely aversive responses (heartache, jealousy,loneliness) to any threat of social loss or isolation [18].We are, in short, neither selfishnor selfless, but “groupish”, a trait that insures that we join and stay with a group(Imperative 1) and creates a strong incentive to conform so as to avoid the pain ofostracism (Imperative 2).

62 D. A. Marshall

FromSingle-Mindedness toDistractibleMorality. Human beings diverge fromsingle-mindedness in amore specific and consequential way than the aforementioneddistractibility, thanks to a persistent alternative goal set—morality—which under-mines our ability to exclusively pursue any given maximand. With some (patholog-ical) exceptions, human beings are far from amoral, each equipped with a suite ofmoral capacities and dispositions. Among these are: The moral sentiments of empa-thy and reciprocity; The dispositions—hierarchy, norm following, peace-making,anti-bullying—collectively known as “community concern”, which helps to main-tain group solidarity and efficacy, and; Moral judgment, which proscribes “wrong”actions, and prescribes the performance of “right” ones [19].

Morality as a whole is perhaps the most evolutionarily consequential social adap-tation in the human repertoire. Empathy makes us good caretakers of our offspringand comrades (Imperative 3), while our commitment to reciprocity helps us to stayin the good graces of our fellow group members (Imperative 2) while also helping tomaintain order amongst group members (Imperative 3) and ultimately strengtheningthe group as a whole (Imperative 4). Community concern likewise does double dutywith regard to Imperatives 3 and 4, while moral judgment helps the individual toabide by group expectations (thereby avoiding expulsion—Imperative 2) and alsomotivates the individual to see that others do the same (Imperatives 3 and 4).

FromMaximization to Continuous andMultiplex Satisfycing. Human beings’propensity for continuous and multiplex satisfycing is fully explicable on survivalgrounds alone, but it is worth emphasizing that it also serves corollary social func-tions. By making “more” unappealing, satiation makes it easier for the individual toavoid selfish behavior, and thereby the opprobrium of other group members (Impera-tive 2), and more evenly distributes resources, so that the group as a whole is stronger(Imperative 4).

From Material to Obligational and Reputational Resources. While there is amaterial dimension to humanfitness, resources like obligations, reputation, and statusare more decisively important to it [20]. Thanks to themoral sentiment of reciprocity,assisting other group members helps insure that they will, in future, return the favor.Contributing to the group’s welfare (Imperatives 3 and 4) simultaneously buildsreputation and status, further endearing the individual to the group (Imperative 2).Unsurprisingly, then, humans find public affirmation of their reputation and statussingularly rewarding [18].

From Informed to Epistemic Conformity. Humans’ preference for nomos-consistent information over accurate informationmeans that individuals work hard toavoid nomos-inconsistent, and to obtain nomos-consistent, information. Left to theirown devices, they will do their best to construct an epistemic bubble. Because inter-subjectivity is a common proxy for truth, such insulation efforts extend to not allow-ing those around us to introduce information inconsistent with our shared nomos [6].Thus informational conformity is an important means of staying in the good gracesof one’s group (Imperative 2), but also buttresses members’ certainty in the group’sshared norms and narratives (Imperative 3).

From Calculation to Habit and Imitation. While human individuals are notkeen on calculation to begin with, the social context—for much the same epistemic


reasons discussed above—often provides further disincentive for calculation, in thatany innovations thatmight come of it might be construed as threatening to the group’sshared nomos, as anyonewho has ever lived in a small town iswell aware. It is usuallyless risky to stick with habit and imitation. As a result, individuals thereby maintaintheir membership and standing in the group (Imperative 2) by helping to preservethe group’s shared nomos (Imperative 3).

4 Conclusion and Implications

Having established that the evolved biological logic of human action is different from,and more often than not antithetical to, its rationalist predecessor, let us consider ahandful of the implications thereof, as regards individual experience, social science,social policy, and government.

4.1 Homo Duplex

The argument herein has not been that we, as humans, are incapable of instrumentalrationality, only that it is not a biologically evolved, hence innate and universal,characteristic of human beings. Obviously, we possess some nominal capacity for it,since we would otherwise have no means recognizing the ways and degrees to whichwe deviate from it. Moreover, though it did not evolve per se, our (limited) capacityfor rationality is nevertheless a by-product (or spandrel) of other capacities that didevolve—and it certainly has been a significant asset to human fitness in some waysat some times. But to argue that it constitutes the underlying logic of human behaviorwould be tantamount to calling our species homo accordionus based on the fact thatsome of us, sometimes, can play the accordion.

This (limited) capacity for instrumental rationality is a recent addendum to amucholder andmore robust biologically adaptive system.Muchof the tragedy andgrandeurof human existence ultimately comes down to the fact that we are homo duplex [21]—composed of and subject to the competing demands of two different control systems:The evolved, non-conscious, and automatic “System 1” (i.e. the biological logicestablished here), and; The newer, cultivated, conscious, and controlled “System2” [4, 22]. The tension between these is loaded with implications for human lifeand history [23, 24], but for present purposes, we will confine our attention to thesimple observation that rationality is, like flight (and accordion playing) not a humanproperty, but merely a human potential, one which comes neither naturally nor easilyto us. The older System 1 is primary to System 2 in every sense of the term. It is thedefault logic of human action, overcomeable only with much effort and attention.The obvious question of where and when individuals manage to realize what degreeof rationality brings us to a second implication.

64 D. A. Marshall

4.2 Rationality as a Dependent Variable

For all of its implausibility as the logic of human action, instrumental rationalityremains an important concept in social science—as a Weberian ideal-type againstwhich to contrast observed behavior, as a normative assumption of Western cul-ture, and, again as per Weber, as a (the?) defining feature of modern society. Mostimportantly, though instrumental rationality fails as an independent variable—anassumption with which to describe or explain human behavior—it is a promisingdependent variable—an outcome to be explained as a function of the problem athand, of individual motivation, and of the cultural and social contexts in which itoccurs [25].

As per the “massive modularity" (i.e. “swiss army brain”) hypothesis, the humanmind is an assembly of discrete special-purpose calculators, rather than a singlegeneral-purpose processor [26], so our ability to deal with a problem rationallydepends upon its falling into a domain that both lacks competing non-rational heuris-tics, and for which our calculation skills are evolutionarily primed (e.g. deceptiondetection [26]). Additionally, the odds of a problem evoking rationality are increasedwhen it is highly circumscribed, such that the option set is limited and explicit,the relevant probabilities are pre-established, and the relevant information is highlydomain-specific and readily available. At the same time, some individuals are morelikely to engage in the effort necessary to move beyond the ready answers providedby System 1 than are others as a result of personality, experience with “critical think-ing”, and geographical context (i.e. rural fixity vs. urban mobility), among otherfactors [27].

Thepotential for rationality is also a function of the cultural context of the problem,in terms of the “tools” it does or doesn’t provide to support rational decision-making,from stopwatches to software packages, double-entry bookkeeping, and t-tests. Lessobvious social technologies—i.e. the scientific method and peer-review—exploitlow-moral-density (loosely-coupled) social structures to dilute and mitigate manyof the individual obstacles to rationality. As for the social context of a problem,though instrumental rationality was maladaptive for most of the history of humanevolution, it is not necessarily so today. Many of the adaptive challenges that oncemade it so antithetical to fitness are much less relevant in the modern world, wherethe capitalist economy has shifted more of human ambitions into the fungible andinsatiable domain of money. Instrumental rationality is potentially less maladaptiveand more profitable than it used to be, which sets the stage for a new domain of (nowcultural) evolution—one effectively decoupled from biological fitness and biologicalevolution and focused instead on the acquisition of resources.

4.3 A Systematics of Social Actors

One implication of this chapter is that the individual person is something of anartificial construct, given our species’ thoroughly social nature and lifestyle. Thus,


there are good reasons to at least supplement the idea of the individual social actorwith what is arguably the more naturally-occurring unit of human action—the tribe.If anything, such tightly-coupled groups tend to amplify individual human traitsand exacerbate non-rationalist tendencies by compounding individual biases anddispositions. While vestiges of such tribes still abound (e.g. families, squads, cults),they are eclipsed in the modern era by a third kind of social actor—the loosely-coupled organization. Unlike the other two actors, the organization is distinguishedby its instrumentally rational nature.

Under the constraints of an emergent imperative to maximize their acquisition ofmaterial resources, organizations have (culturally) evolved to leverage the finite ratio-nality of their constituent persons into potentially infinite manifestations of zweck-rationality. Lacking a physiology of its own, the organization is free frommany of thecompeting needs and limitations that beset individuals. It imposes an incentive struc-ture that aligns different individual goals with its own, and parses decision-makinginto myriad highly specialized and routinized domains, thereby concentrating themeager potential for rationality of its individual members to produce instrumentalrationality at scale. Even as it concentrates cognitive power, it, by a combination ofhappenstance and design, dilutes human morality by diffusing individuals’ sense ofresponsibility for their actions both horizontally and vertically [25].

The emergent rationality of organizations explains how instrumental rationalitycame to be the master trend of the modern world even as the individual human beingswho are its original constituents remain largely unable and unwilling to engage init. It also alerts us to a potentially era-defining social problematique of the modernworld—the uneven distribution of rationality. In a world in which the biologicallogic no longer serves fitness much of the time, and in which profit is the prevailingmaximand, the old, naturally-occurring social actors of individuals and tribes aremassively outmatched by their new, organizational, competitor. Witness the extentto which organizations are presently accruing unprecedented economic, political,and cultural power at the expense of individuals and tribes, in the form of wideninginequalities, subzero tax rates for large companies, and corporations’ increasingarrogation of the rights and privileges of human citizens.

4.4 Democratic Institutions—Weaponized Tribalism

One of the most disturbing mysteries of the early twenty-first century has been thefailure of democracy to secure benefits for voters because those voters repeatedlyvote against their own interest.As per one popular treatment, “What’s thematter withKansas?” [28]. From the perspective of this chapter, this pattern is fully predictable.Democracy, as noted in the introduction, is predicated on the existence of a moreor less rational electorate—one that is, at minimum, self-interested enough to seekaccurate and complete information, and to engage in the calculation necessary tochoose the best way to pursue those interests. But as we’ve seen, human beings havenever truly fit that description. Following their advent and ascent, organizational

66 D. A. Marshall

actors, consciously or not, discovered that the biological logic of human action canbe deployed against the individuals who make up an electorate, and over time, theyhave perfected the technologies of doing so. Though voters may think of themselvesas informed, calculating, and self-interested, they are easily manipulated into actingagainst their own interest by a combination of fearmongering and manufacturedtribalism which plays upon their needs for belonging, their fear of ostracism, andtheir need for collective identity (e.g. racism, nationalism).

That this strategy is only now coming to fruition has much to do with the rise oftechnologies which made possible, for the first time, the mass construction of thekind of fully-insulated epistemic bubbles that truly effective propaganda requires[29]. Thus, we close this chapter with an illustration of one of the dangers inherentin the rationalist assumption: The internet was supposed to usher in the next phaseof humanity by democratizing access to information and calculation power—butsuch rosy assumptions were predicated on false conception of individuals as rationalconsumers of information and calculating maximizers. We were thus blindsided bythe reality that human actors were never going to use it that way, opening the wayfor organizational actors—from Fox news to “Fancy Bear” to Facebook—to exploithuman nature in order to seduce voters into behaving in ways that are profoundlyantithetical to their own self-interests by using their biological logic against them.

References

1. Weber, M.: Economy and Society. University of California Press, Berkeley CA (1978 [1922])2. Fiske, S., Taylor, S.: Social Cognition, 2nd edn. McGraw-Hill, New York NY (1991)3. Arieley, D.: Predictably Irrational: The Hidden Forces that Shape Our Decisions. Harper, New

York NY (2009)4. Gould, S.J., Lewontin, R.: The spandrels of SanMarco and the panglossian paradigm: a critique

of the adaptationist programme. Proc. R. Soc. Lond. Ser B Biol. Sci. 205, 581–598 (1979)5. Darwin, C.: The Origin of Species: 150th Anniversary Edition. Signet Reprint (2003 [1859])6. Kahneman, D.: Thinking, Fast and Slow. Farrar, Strauss, Giroux, New York, NY (2011)7. Bonner, J.: The Evolution of Culture in Animals. Princeton University Press, Princeton, NJ

(1980)8. Berger, P., Luckmann, T.: The Social Construction of Reality: A Treatise in the Sociology of

Knowledge. Anchor Books, Garden City, NY (1967)9. Harari, N.Y.: Sapiens: A Brief History of Humankind. Harper Perennial, NewYork, NY (2018)10. Bargh, J., Chartrand, T.: The unbearable automaticity of being. Am. Psychol. 54, 462–479

(1999)11. Nowak,M.: Supercooperators: Altruism, Evolution, andWhyWeNeed EachOther to Succeed.

Free Press, New York, NY (2011)12. Wilson, E.O.: Genesis: The Deep Origins of Societies. Liveright, New York, NY (2019)13. Kaplan, H., Hill, K., Hawkes, K., Hurtado, A.: Food sharing among Aché hunter-gatherers of

Eastern Paraguay. Curr. Anthropol. 25, 113–115 (1984)14. Bowles, S., Gintis, H.: ACooperative Species: Human Reciprocity and Its Evolution. Princeton

University Press, Princeton, NJ (2011)15. Hrdy, S.B.: Mother Nature: Maternal Instincts and How They Shape the Human Species.

Ballantine Books, New York, NY (2000)


16. Diener, E., Lusk, R., DeFour, D., Flax, R.: Deindividuation: effects of group size, density,number of observers, and group member similarity on self-consciousness and disinhibitedbehavior. J. Pers. Soc. Psychol. 39, 449–459 (1980)

17. Tajfel, H.: Experiments in intergroup discrimination. Sci. Am. 223(5), 96–102 (1970)18. Lieberman, M.: Social: Why Our Brains are Wired to Connect. Crown Publishers, New York,

NY (2014)19. DeWaal, F.: Primates and Philosophers: How Morality Evolved. Princeton University Press,

Princeton, NJ (2006)20. Weisner, P.: Taking the risk out of risky transactions: a forager’s dilemma. In: Salter, F. (ed.)

Risky Transactions. Berghahn Books, Oxford UK (2002)21. Durkheim, E. The dualism of human nature and its social conditions. In Wolff K (trans.) Emile

Durkheim: Essays on Sociology and Philosophy, pp. 325–340 (1964 [1914])22. Chaiken, S., Trope, Y.: Dual Process Theories in Social Psychology. Guilford Press, NewYork,

NY (1999)23. Marshall, D.A.: Behavior, belonging, and belief: a theory of ritual practice. Sociol. Theor.

20(3), 360–380 (2002)24. Marshall, D.A.: Sacralization and the throes of modernization: witch trials, world war, and the

WTC. New York J. Sociol. 1, 26–90 (2008). www.newyorksociology.org25. Marshall, D.A.: Beyond a Rational Choice Sociology: A Sociology of Rationality. Doctoral

Dissertation. Retrieved from Proquest Dissertation Publishing (2003)26. Barkow, J., Cosmides, L., Tooby, J.: The Adapted Mind: Evolutionary Psychology and the

Generation of Culture. Oxford University Press, New York NY (1992)27. Stanovich, K.: Who is rational?: Studies of Individual Differences in Reasoning. Psychology

Press, London UK (1999)28. Frank, T.: What’s the Matter With Kansas?. Henry Holt, New York, NY (2004)29. Ellul, J.: Propaganda: The Formation ofMens’Attitudes. VintageBooks, NewYork, NY (1973)

Social Media: “Surrogate Tribes”?

Paolo R. Crocchiolo

Abstract According to recent research, our complex neural-hormonal system(NHS)was shapedby evolution in functionof its ability to protect our body’s integrity,thus ultimately enhancing our genes’ reproductive potential. Our “mind-body units”may hence be viewed as NHS-endowed “survival and reproduction machines” ofour genes. In this context, social media, and networks, may be viewed as a sort of“surrogate tribes”, i.e., updated instruments of a contemporary version of the evolvedhuman social communities.

1 Introduction

The senses in turn may be interpreted as prostheses of their carriers’ respectiveNHS, as they are capable of converting electro-magnetic waves, air vibrations, orother environmental signals, into luminous, acoustic, or other sensations to whichwe arbitrarily assign specific qualities that they in themselves do not possess.

Communicating instruments developed in the recent past (radio, TV, telephones,internet etc.) are becoming like “second degree” prostheses of our previously evolvedNHS prostheses, i.e. the senses.

One of our main NHS’ emergent properties is to distinguish between positive andnegative hedonistic tones (pleasure and pain), which generate emotions and instincts;the latter consisting of gene combinations that were selected as they codified neural-hormonal patterns of response to environmental challenges inducing behaviours bestgranting, in the average, the reproductive success of their phenotypic carriers [1].

According to recent neuroscientific views [2], thewhole neural-hormonal networkhas evolved as an extremely fine-tuned re-elaboration of our primordial ancestors’cell membrane, basically serving the same purpose of protecting the integrity ofits underlying genes’ “phenotypic envelop”, i.e., the body. In fact, our mind has

P. R. Crocchiolo (B)American University, Rome, Italye-mail: [email protected]


69

70 P. R. Crocchiolo

evolved as a complex machinery constantly monitoring our inner state. The “emo-tional colouring”, which is generated by the NHS, and characterizes all activitiesperformed by our mind, thus appears to be indispensable for the preservation of thevital processes generating that “well-being” state.

Any “species”, including ours, can be viewed as a “DNA consortium”, an involun-tary and unconscious way of survival and reproduction that allowed the propagation,so far, of their respective genetic pools. We, “individuals”, aren’t, in this perspective,but physically separated variants of the shared broader DNA consortium, i.e. of ourcommon genetic pool.

Within this context, individuals were selectively advantagedwho best fit the needsof their troop/tribe’s social structure (in its turn emerged from a previous, in fact aparallel, selection), as they were carrying the tribal instincts that best granted theircommunity’s genetic pool’s survival and reproduction [3]. Thus, individual selectioncompounds in the long run with social selection.

In this context, social media, and networks, may be viewed as a sort of “surrogatetribes”, i.e., updated instruments of a contemporary version of the evolved humansocial communities.

According to evolutionary biology, our very complex individual phenotypes arethe end-results of an equally sophisticated gene-network, i.e., of genes unconsciouslyinteracting with each other in such a way that their end-product made them survivein preference to the end-product of other networks/patterns, the latter being outnum-bered because of the former’s higher reproduction rate in their given natural andsocial environment.

In this context, both our body and our mind seem to be the result of biologicalevolution by natural selection [4, 5]. More specifically, both physical and psycholog-ical traits of our phenotypes appear to represent the end-results of a complex chainof processes triggered by our genotypes’ proteinopoietic activity.

In fact, our body which, as a sort of “envelop”, harbours and protects our genes,and ourmind, whose neural correlate is our complex neural-hormonal system (NHS),appear to have evolved (by selective differential reproduction rate of their phenotypiccarriers) in function of protecting our body’s integrity, thus ultimately enhancing ourgenes’ reproductive potential [2].

In this perspective, mind-body units may be interpreted as “survival andreproduction machines” of their underlying genes [6].

2 Emergent Properties of Our Neural-Hormonal System

2.1 The Senses

Representing the biological interfaces with the outer world, the senses appear tohave evolved as prostheses of our neural systems. Senses, in fact, may be defined

Social Media: “Surrogate Tribes”? 71

as transductors, capable of converting electro-magnetic waves potentially perceiv-able as luminous, or air particles’ vibrations potentially perceivable as acoustic, intosensations to which we arbitrarily and conventionally assign, within each speciesor broader biological category, particular qualities that they in themselves do notactually possess [1].

Within the whole spectrum of electro-magnetic waves and air vibrations, therange “captured” by the evolved brains and senses of each species has been tailoredby natural selection to the one that, in the past generations, turned out to effect thehighest average reproduction rate of their phenotypes in that given environment,and hence the widest spread of their underlying genotypes. Even size-wise, humansand their closest related species have evolved brains and senses that do not perceivemacro- or microcosmic, but only “meso-cosmic” entities, i.e. entities which ourancestors necessarily came across, and had to deal with, in the course of their lives[7].

2.2 Hedonistic Tones

One of the main emergent properties of our NHS is its fundamental capacity todistinguish between positive and negative hedonistic tones (pleasure, pain, and everysensation in between the two ends of the perceivable emotional spectrum) [1].

The latter capacity rests upon an elaborate network connecting certain inner brainareas, mainly located in the “limbic system”, with specialized groups of nervous cellssecreting neurotransmitters, such as dopamine, serotonin, etc., and neuro-hormones,such as endorphins, oxytocin, etc. All these molecules, when released upon recogni-tion of an appropriate triggering signal by the cells where they are normally stored,travel to, and impact on, their specific receptors situated on their corresponding targetcells.

2.3 Reflexes and Instincts

Natural selection rewarded those phenotypes endowed with a NHS that associateda positive or a negative hedonistic tone, respectively, with situations and eventsenhancing or inhibiting their survival and reproduction, thus originating reflexesand instincts. The latter may be defined as genetically transmissible associationsof neural-hormonal rewarding/penalizing mechanisms with their resulting specificsurvival- and reproduction-enhancing skills/potentials. Instincts, in other words, maybe viewed as a kind of “unconscious bridles” that guide us in our thoughts and inour actions.

72 P. R. Crocchiolo

2.4 Emotions and Feelings

In addition to reflexes and instincts, the hedonistic tone system also originatesemotions and feelings.

Emotions may be defined as amplifying modulations [1] of the basic, positiveor negative, DNA-imprinted hedonistic tones, allowing their carrier organisms tospecifically adapt their behavioural responses to the great variety of natural andsocial environmental challenges they are exposed to. Feeling, in man, represents afurther fine-tuning capacity resting upon the development of autobiographic self-consciousness [2].

Why modulations? Qualitatively, organisms developed the capacity to select,within a generally positive or negative hedonistic background, a number of differentneural-hormonal patterns appropriately fitting different situations to cope with (e.g.,fear, jealousy or rage predisposing to different behaviours within a generally negativehedonistic tone).

And why amplifying? Quantitatively, organisms developed the capacity to tailorthe intensity of their emotions to the biological impact of the triggering events.Amplification of the emotional colouring roughly reflects the increase, or decrease,in biological fitness of the carrier’s organism in relation to the initial situation.

The broader the range of possible modulations and the ampler the intensity of thepositive or negative emotion evoked, the higher the likelihood of the underlying genecombinations to survive, reproduce, and spread within the subsequent generations.

2.5 Genetically Predisposed Fine-Tuning of EmotionalResponses

While reflexes and instincts are phylogenetically selected and consolidated by adap-tation to a relatively constant environment, emotions and feelings are ontogeneticallylearned and potentially adaptable to individual experience. Neural-hormonal mech-anisms, such as the “mirror neurons” system [3], grant the necessary flexibility toadapt the organism’s emotional response to a virtually infinite variety of triggeringevents and situations it may be exposed to during the course of its life. Thus, organ-isms gradually acquired the ability to qualitatively and quantitatively fine-tune theappropriate responses tailoring them to the specific situations encountered, choos-ing from a more basic repertoire of reflexes/instincts directly tied to survival andreproduction, and made available by ancestral, previously consolidated structuresand functions [2].

In this respect, memory plays an analogous role: qualitative and quantitative fine-tuning of the response is related to a learning process whereby previous individualexperiences exhibit a booster effect, biasing the organism’s selection and modulationof the most appropriate emotional response, based on the previously retained/storedexperience.


Thus, natural selection rewarded, and spread through the next generations, thosegenetic blends that codified neural-hormonal patterns of response to environmentalchallenges which induced behaviours best granting, in the average, the reproductivesuccess of their phenotypic carriers [8].

2.6 The “Self” as a Biological Construct

All these genetically predisposed phenomena, generated by our NHS in response tothe outer and inner signals it comes across, ultimately serve the same purpose as ourprimordial ancestor’s cell membrane, of which it seems to represent the ultimate,extremely sophisticated re-elaboration: protecting the integrity of the body, i.e., ofits underlying genes’ “phenotypic envelop” [2].

Remarkably, the human “self” is not originated by the outer world, but by theinner body, as our mind has evolved as a complex machinery constantly monitoringour inner state. In fact, it is this very inner state (the “milieu intérieur” according toClaude Bernard) [9] that hierarchically overrides our NHS, due to its vital importancein protecting our genes’ survival, and their reproductive potential.All visual, acoustic,and other images impacting on us, are thus elaborated, and attributed a qualitativevalue (emotional colouring) which is closely connected with the preservation of thevital processes that generate that state. The body integrity is the prerequisite for allthis. In short, it’s the mind serving the body, and not the other way around [2].

The sense of individual identity seems to have been rewarded—and consoli-dated—by natural selection, as an extremely sophisticated development of the outerenvelope protecting the body’s integrity, and its continuity throughout time. In thisconnection, the “autobiographic, extended consciousness” appears to be the endresult of a long evolutionary trail that, startingwith the “proto-self”, has led, stepwise,to the “core self” and, finally, to the stage of the “autobiographic self” [2].

In this connection, it is important to note that, according to Metzinger [10], nosuch things as selves exist in the world, and the phenomenal “self”-model, i.e., thevery sensation of being an “I”, is not a thing, but an integrated process and, ultimately,an illusory construction of our “biosystem”, a sort of metaphorical representation ofour body’s integrity and continuity throughout time.

2.7 Evolutionary Psychology

As the NHS is not a static phenomenon, but a dynamic process based on a continuousturnover, it changes with time. In fact, our gene-networks’ activity unfolds in acomplex feedback-regulated time sequence.

According to evolutionary psychologists [4], neuro-embryological developmentand, subsequently, neural-hormonal phases of development characterizing the courseof our life after birth, seem to be genetically imprinted and pre-programmed as a result

74 P. R. Crocchiolo

of cumulative natural selectionmechanisms advantaging individuals whose resultingsequential behavioural patterns exhibited a fine-tuned adaptation to their natural andsocial environmental set-up.

3 Human Communities and Social Media

3.1 The Species as a Comprehensive “Phenotypic Envelop”of Its Genetic Pool

Any “species”, including ours, can be viewed as a “DNA consortium”, an involuntaryand unconscious way of survival and reproduction of its shared genetic pool, thatallowed its propagation so far in the environmental niches it was exposed to all alongthe course of its evolutionary history. Thus individual selection, in the long run,appears compounded with what Gould calls “species sorting”, i.e., species selection[11]. In this perspective,we, the “individuals”, aren’t but physically separatedvariantsof our broader DNA consortium, i.e., of our shared genetic pool.

3.2 The Evolution of Human Communities

Within this context, but at a subordinate level, those individuals were selectivelyadvantaged, who best fit the needs of the troop/tribe’s social structures, which in themeantime had emerged from a previous, in fact from a parallel, selection. Accordingto sociobiologist Edward O. Wilson [12], troop- and tribal identities, i.e., the senseof belonging to their community, as well as troop- and tribal hierarchies, were thenselected among our ancestors as survival- and reproduction enhancing instinctualadaptations, to the extent they were protecting the integrity of the collective “en-velop”. The selected individuals were the ones carrying the tribal instincts that bestgranted the success of their respective community’s acquisition, and defence fromother tribes and predators, of the food and space necessary for their genetic pool’ssurvival and reproduction.

Genes were selected that predisposed their phenotypic envelops to the sharing oftribal behaviours, beliefs, practices and emotional involvements, including collectiveexcitements such as playing, dancing, singing, etc. These instinctual behaviours allcontributed to inclusion of their individual organisms into their surrounding, mainlynext of kin, community [13]. Adherence to the shared “biased thinking” (tribal prej-udices, taboos, commonly accepted stereotypes, etc.) turned out to be advantageousfor their carriers, hence psychologically rewarding (Facebook’s “I like it” may soundlike a pale vestige of one of our ancestral reward-system arousal factors). Isolation,self-exclusion, and marginalization instead, all resulted in selective disadvantages,


leaving less chances to survive and reproduce to the carriers of those alternativeexclusion-inducing instinctive traits, and hence of their underlying genes.

3.3 Language

We have the potential to use to their full extent the instruments made available tous by our NHS-endowed “survival and reproduction machines”. Among all theseinstruments, language plays a paramount role to communicate, both rationally andemotionally, within human societies. In fact, since perceptions are automatically andrapidly translated into language, we have evolved as verbal creatures [14].

3.4 Tribal Instincts Showing Through the Social Media

In addition to language and other previously evolved means of communicationbetween humans, all communicating tools developed in the recent past, such astelephones, radios, televisions, computers, e-mails, the internet, and the whole cor-nucopia of their network “derivatives”, are swiftly becoming a sort of “seconddegree”prostheses of our anciently acquired NHS-prostheses, i.e., the senses. Humans havedeveloped the technical tools that offer them the possibility to exchange verbaland visual messages at any distance. Conversations, debates, gossips, photographs,videos, voice recordings, etc., all possibly contribute to the strengthening of the tiesbetween members of the same “tribal” community.

3.5 The Role Played by Sexual Selection

In the sort of “ping-pong” game called sexual selection, gene combinations wereselected that—via NHS—“suggested” attraction to certain physical or psychologicaltraits of their potential partners’ phenotypes. Peacocks, for instance, that instinctivelyexhibited their magnificent tails (a powerful proxy standing for “good genes”), wereinvolved in the reproductive success of peahens that in turn had been selected to theextent they were impressed and attracted by those very tails, which ultimately hadgranted a higher reproduction rate to their peacock carriers in the first place.

Through the media, in general, and the social media in particular, a possibility isoffered to the “human peacocks, and peahens” to exhibit, as a kind of “ornament”,their very “intelligence” [15], or other mental or physical attributes, to their potentialpartners in the web. The social media thus convert, also from this point of view, intoconvenient instruments to convey all kinds of messages (not only “friendship”) tothe other members of the modern tribal community.

76 P. R. Crocchiolo

4 Conclusion

In the long history of evolution, the genes that exhibited best survival and repro-duction chances were those that, in addition to mono- or multi-cellular “individu-al” skills, evolved particularly strong “social bonds”. In fact, the same cooperativeinteractions observable in all living beings, starting from bacterial strains, socialinsects, flocks of birds, mammals’ troops, etc., seem to be amenable to the samegene-preserving driving force that generated the social bonds characterizing humantribes, subsequently evolved into today’s complex human societies. The gradual evo-lution of more and more sophisticated “emotional brains” [8] in multicellular animalorganisms, obviously played a fundamental role in this context.

Thus, in conclusion, socialmediamay be seen as updated instruments of a contem-porary version of the human social groups, a sort of “surrogate tribes”, instinctivelydictated, in ultimate analysis, by our genes.

References

1. Johnston, V.S.: WhyWe Feel: The Science of Human Emotions. Perseus Books, Reading, MA(1999)

2. Damasio, A.R.: The Feeling of What Happens: Body and Emotion in the Making ofConsciousness. Harcourt Brace, New York (1999)

3. Rizzolatti, G.: Mirrors in the Brain: HowOurMinds Share Actions, Emotions, and Experience.Oxford University Press, Oxford (2008)

4. Cosmides, L., Tooby, J.: Cognitive Adaptations for Social Exchange. The Adapted Mind:Evolutionary Psychology and the Generation of Culture. Oxford University Press, New York(1992)

5. Damasio, A.R.: Self Comes toMind: Constructing the Conscious Brain. Pantheon Books, NewYork (2010)

6. Dawkins, R.: The Selfish Gene, 3rd edn. Oxford University Press, Oxford (2006)7. Boncinelli, E.: Minuscoli universi chiamati “uomini”. Corriere Della Sera, Milan, 28 Nov

(2003)8. Le Doux, J.E.: The Emotional Brain: TheMysterious Underpinnings of Emotional Life. Simon

& Schuster, New York (1998)9. Bernard, C.: An Introduction to the Study of Experimental Medicine. Dover Publications, New

York (1957)10. Metzinger, T.: Being No One: The Self-model Theory of Subjectivity. MIT Press, Cambridge,

MA (2003)11. Gould, J.S.: The Structure of Evolutionary Theory. Belknap Press of Harvard University Press,

Cambridge, MA (2002)12. Wilson, E.O.: Consilience: The Unity of Knowledge. Knopf, New York (1998)13. Ridley, M.: Nature Via Nurture: Genes, Experience, and What Makes Us Human. Harper-

Collins, New York (2003)14. Pinker, S.: The Language Instinct: How the Mind Creates Language. HarperCollins, New York

(1994)15. Miller, G.: The Mating Mind: How Sexual Choice Shaped Human Nature. Vintage, New York

(2001)

Moral Bubble Effect

Violent and Unaware of It: How Difficultiesin Recognizing One’s Own Violence Lead to Disregardthe Inflicted Harm

Lorenzo Magnani

Abstract To undertake a serious investigation of violence calls for a discrete amountof courage and sincerity: as human beings we can wishfully ignore our own violence,thanks to a kind of “embubblement” I am illustrating in this article. I would also liketo offer to the attention of researchers in sociology, psychology, and psychiatry themain features of the situations in which the awareness of human violent acts is inquestion, as indicated in the subtitle “Violent and Unaware of It: How Difficulties inRecognizing One’s Own Violence Lead to Disregard the Inflicted Harm”. In thesecases a process of what can be called “autoimmunity” is at play. I contend the conceptof moral bubble can provide an integrated and unified perspective able to interpretin a novel way many situations in which morality and violence are intertwined.

Keywords Moral bubble · Epistemic bubble · Embubblement ·Morality ·Violence · Autoimmunity

1 Epistemic and Moral Bubbles

In 1999 JustinKruger andDavidDunning published in the Journal of Personality andSocial Psychology the seminal paper “Unskilled and Unaware of It: HowDifficultiesin Recognizing One’s Own Incompetence Lead to Inflated Self-Assessments” [10],which explained what later on has been called Dunning-Kruger effect. In the presentarticle I would like to offer to the attention of sociological, psychological, and psychi-atric research a similar effect that regards the absent (or at least extremely unstable)awareness of human violent acts, as indicated in the subtitle “Violent and Unawareof It: How Difficulties in Recognizing One’s Own Violence Lead to Disregard theInflicted Harm”. I have already sketched this effect in my book Understanding Vio-lence. The Intertwining of Morality, Religion, and Violence: A Philosophical Stance[14, chapter three]. I am detailing here some more hypotheses that ask for further

L. Magnani (B)Department of Humanities, Philosophy Section and Computational Philosophy Laboratory,University of Pavia, Pavia, Italye-mail: [email protected]


77

78 L. Magnani

theoretical deepening and empirical study. This common psychological mechanismcalled “embubblement” is indeed at the basis of the process of rendering violenceinvisible and of condoning it. Human beings are prisoners of what I call moral bub-bles (see below the following sections), which systematically disguise their violenceto themselves: this concept is also of help in analyzing and explaining why so manykinds of violence in our world today are treated as if they were something else. Sucha disseminated and disguised violence, which is at the same time in various wayslegitimated, leads to the central core of my commitment to the analysis of variousrelationships between morality and violence, as I have illustrated in the book quotedabove.

In thefirst two chapters of this book I described the so-called coalition enforcementhypothesis—in a paleoanthropological perspective—together with a philosophicalview emphasizing the inherent moral (and simultaneously violent) constitution oflanguage (and of abductive and other hypothetical kinds of cognition and reasoningswhich are carried by linguistic propositions). Following this point of view naturallanguages can be seen as fundamentally rooted in a kind of military intelligence, ascontendedbyThom [20]. Indeed it is through coalition enforcement that our ancestorsbuilt collectives in which cooperation was granted by primitive rules that we can nowcall “moral”, a cooperation at the same time preserved by violence against free-ridersinside the group, and against collectives that carry different moral (or proto-moral)rules.1

First of all let us focus on the concept of epistemic bubble that is a precondition forthe explanation of my proposed moral bubble effect, I will explain in the followingsections. Woods contends that “A cognitive agent X occupies an epistemic bubbleprecisely when he is unable to command the distinction between his thinking thathe knows P and his knowing P” and, further, that “When in an epistemic bubble,cognitive agents always resolve the tension between their thinking that they knowP and their knowing P in favour of knowing that P” [22]. In sum, we know lessthan we think we do. It is also important to add that epistemic bubbles change, thedifferentiation between simply apparent correction and truly successful correctionexceeds the agent’s control and so a cognitive agent, from his own first-person view,is guided to consider what happened as genuinely sound correction.2

At last, a citation by Woods is remarkable and clear:

1Moral rules and related violence are first of all performed thanks to natural language—imaginefor instance the military/violent status of hateful, racist, homophobic speech—but also by motoractions and emotions: it is well-known that patent hostility in emotions is a chance to start violentactions.2Studying the related problem of “feeling of knowing” Burton [4, p. 12] contends that it is in generalrecognized by the absence of knowledge itself (that is when we do not know something we feelwe should know), and it is also connected to the well-known situation of the so-called “cognitivedissonance”: “In 1957, Stanford professor of social psychology Leon Festinger introduced the termcognitive dissonance to describe the distressing mental state in which people find themselves doingthings that don’t fit with what they know, or having opinions that do not fit with other opinions theyhold”.

Moral Bubble Effect 79

As before, when in an epistemic bubble, cognitive agents always resolve the tension betweenonly the apparently true and the genuinely true in favour of the genuinely true. Corollary12a reminds us of the remarkable perceptiveness of Peirce […] that the production of beliefis the sole function of thought. What Peirce underlines is that enquiry stops when belief isattained and is wholly natural that this should be so. However, as Peirce was also aware,the propensity of belief to suspend thinking constitutes a significant economic advantage. Itdiscourages the over-use of (often) scarce resources [22].

Hence, truth is “fugitive” and the process of embubblement stresses the importanceof corrigibility of our beliefs, especially the ones that present unwelcome effectsduring the processes of problems solving and decision-making: Woods correctlycontends that the process of embubblement is not corrigible in itself, unless theagent dismisses a bubble to adopt a new one. The issue about corrigibility seems tohave something in commonwith another issue, namely, de-biasing. Indeed epistemicbubbles share with biases the fact that both present a resistance to correction.3

2 Embubblement and Autoimmunity

The process of embubblement is also self-supporting since it favors the enhance-ment of the agent’s knowledge without rendering her aware of her self-delusion.It generates a cognitive autoimmune system [1, 2, 16], since the agent exploits thesame cognitive mechanisms to acquire knowledge and control its validity. The termautoimmunity has been first applied byWoods [6, 22] when he described the curiouscharacter of the so-called epistemic bubbles, I introduced in the previous section.Here I am proposing an extended version of the term in order to describe the ubiq-uitous and the self-supporting nature of the process of “embubblement”. Indeed, theanalysis of the human agent’s capacity to deal with her deficiencies is the first motifwhy “autoimmunity”, a term with clear negative aspects, can be presented in thisframework as a universal and fundamental concept.

In the bio-medical disciplines the term “autoimmunity” regards the impairedwell-being of the agent generated by a corrupted reaction of the immune system. Sinceit delineates a set of diseases, it is adopted in order to indicate an abnormal eventaffecting the standard health of the individual. On the other hand, in the frameworkof this article—nurtured by the aims of the so-called “naturalization of logic” andunder an eco-cognitive perspective (for further details, see my books on abductivecognition [13, 15])—the term does not concern the characterization of a diseasebecause human cognition is not principally delineated as epistemologically sane.Certainly the term still regards a problematic and unintentional reaction of the agentagainst herself, but it does not indicate an anomalous state. The concept of cognitiveautoimmunity, indeed, refers to an unavoidable condition of the human epistemicstatus that jeopardizes her capacity to acknowledge her own fallacious reasonings

3Even though, during the last decades, studies regarding decision-making have increased knowledgeabout biases and fallacies, de-biasing inferential processes are still relatively uninvestigated. Somemore notes about this issue are illustrated in [14, chapter three].

80 L. Magnani

and eradicate her own misconceptions. Thanks to the autoimmune mechanism theagent can enter a status that is less problematic as possible, for example in case ofdecision-making procedures and emotional reactions. The intertwining between theepistemological status of the agent and her related cognitive and emotional state isindeed at the basis of the concept of cognitive autoimmunity. 4

3 Moral Bubble Effect: How Difficulties in RecognizingOne’s Own Violence Lead to Disregard the InflictedHarm

3.1 Language, Fallacies, and Moral Bubbles

The studies on logic, informal logic, and rhetoric have demonstrated that fallaciesperformed thanks to human language are often characterized by their “military” (inThomian sense) effects in terms of softness and gentleness, because one of their mainvirtues is the capacity to disguise errors. Epistemic bubbles are complicit in theseprocesses of concealment, and being unaware, in a constitutive and spontaneousway, of our mistakes is in many cases intertwined with the self-conviction that weare not at all violent and aggressive in the arguments we are advancing (and inour possible subsequent acts). We have to remember that humans use language andthe so-called fallacies embedded in it to the aim of producing positive and vitalresults5—even if they can at the same time carry violent outcomes: very often theeventual target agent feels violently affected by fallacious expressions. If fallaciesare, so to speak, eco-cognitively fruitful, it is difficult to call them again “fallacies”,at least if we use the term fallacy in its standard negative connotation. I contend thatthe example of the role of fallacies embedded in human argumentation favors thehypostatization of the more general6 phenomenon of what I called moral bubbles[15], which are homomorphic with the epistemic bubbles. In summary: unawarenessof our mistakes is often intertwined with the absence of awareness concerning thedeceptive/aggressive status of our speech (and behavior).

4We have to remember that the cognitive autoimmunity of the agent is also related to the psycho-logical problems of the so-called “epistemic feelings” [17, 18], that is feelings that are linked to theepistemic situation of the agent, such as in the case of feeling of knowing, feeling of forgetting, andtip on the tongue feeling. They can be seen as the neurological and cognitive detonators of cognitiveautoimmunity [2].5Just to make an example, research on informal logic has illustrated that a fallacy such as hastygeneralization, that is an induction based of one or few examples, can be positive and favor thefitness of the reasoner, even if it is a violence against the correct reliable way of seeing inductions,that it is unacceptable in the light of good rational criteria for a good induction [21].6This phenomenon not only concerns natural language, but also emotions, actions, and all kindsof other non propositional cognitive activities, such as the ones based on models(icons, diagrams,visualizations, simulations, etc.).


In this perspectivewe can say thatmoral bubbles are a perfect psychological devicethat allow human beings to legitimize and the same time to dissimulate violence.

3.2 Moral Viscosity and Consistency

Moral philosophers have stressed that feature of morality defined as “viscosity”,which in my opinion is strictly related to the genesis of the moral bubbles. Lahtiand Weinstein [12] propose the concept of viscosity to explain the spread that thereis between the abstract idea of absolutism of morality and the empirical fact hatmoral rules are often contravened with no considerable negative outcomes neitherfor the whole moral apparatus, nor for the subjects who disregard the rules. Moralityis viscous because is glue-like, that is capable to be distorted, distressed, weakenedand restored without appreciable negative effects and harm with respect to its ownstability: of course this need of stability also reflects the evolutionary adaptive role ofmorality (and consequently of cooperation) as “group stability insurance”. Viscosityexplains that a moral agent can disregard moral rules, for example stealing preciousobjects, and still consider robbery to be morally wrong. In general, he can performviolence and yet preach non-violence as the way to felicity and redemption. This isnot mere hypocrisy but one of the many facets of the moral bubbles we are trappedin, which in this case permits the compensation of the cognitive inconsistency thatwould be generated by the gap between our behavior and our moral convictions.When, morally speaking, a person feels she is “inconsistent”, she ends up in a moralbubble, so that the inconsistency itself is not actually “sensed”.

Burton says: “The more committed we are to a belief, the harder it is to relinquish,even in the face of overwhelming contradictory evidence. Instead of acknowledgingan error in judgment and abandoning the opinion, we tend to develop a new attitudeor belief that will justify retaining it” [4, p. 12]. Indeed, we trust a belief even if wehave clear knowledge which contradicts it: I can add that, within our moral bubbles,we assume an axiological conviction even if we know its possible or actual violentoutcomes, that at the same time disappear from awareness. The actions that derivefrom firm moral beliefs are at the same time characterized by a strong intellectualworth because they are and remain andwe agree with them: possible violent outcomepossess a secondary worth and, consequently, they can easily vanish into an unawaredimension.

3.3 Structural Violence

Moral bubbles, that typically characterize individual agents, can also easily charac-terize entire groups of people that share common tasks. Usually structural violenceoften appears to the agents asmorally legitimate: parents, relatives, policemen, teach-ers, and other agents inmany cases perform physical or invisible violence on the basis

82 L. Magnani

of legal and/or moral reasons. Actually these legitimizing reasons cannot cancel thereality of the violence committed and the fact that some subjects feel offended, butsurely can conceal the perpetrated violence to the actors. The concept of moral bub-ble is consequently able to explain why so many types of current everyday violenceare considered as if they were something else.

3.4 Moral Hazard

Dowd [5] says that moral hazard is happening when one party is responsible forthe interests of another having an incentive to let his interests come first. A famouscase of moral hazard is the one regarding the subprime scandal and the relatedfinancial crisis of fall 2008. Banks were selling their mortgages disregarding theprobabilities of default, so they misperceived all the risks at stake. I think that moralhazard basically consists in a deceptive/faking dimension in which cooperation maystop. To explain this aspect the following example can be useful. Humanitarianmilitary help to prevent genocides is considered good inwestern countries and byUN.Kuperman [11] expressed a doubt about some aspects of the impact that humanitarianintervention can have. He said that some minority leaders may have an incentive toact without responsibility, and so to abandon cooperation, for instance provokingbloody events to produce a good opportunity for the military intervention of theinternational community on their side. Indeed, Kuperman notes that in this case ithappens that the intervention arrives too late, and the costs of the whole process aretoo high. In this case the free-riding aspect of the moral hazard remains concealedto those who have it: the minorities think they are cooperating when their are not,because they perverted the cooperation, still favoring a moral embubblement.

3.5 Banality of Evil and Decent Embubbled People

Hannah Arendt splendidly noted that normal, banal, decent people perpetrate violentand more or less atrocious acts, and this seems

[…] to go against one’s basic understanding of the world. Some exceptional explanationmust be required, because it seems that evil deeds should be done by evil people, and yetmany such deeds are committed by people who do not conform to the stereotypes of evil.Yet these stereotypes are one of the major obstacles to understanding evil. This is ironic,because the myth of pure evil was constructed to help us understand evil – but it ends uphampering that understanding. The myth is a victim’s myth, and there is often a wide, almostimpassable gap between the viewpoints of victims and perpetrators [3, p. 379].

Moral bubbles are still at play. Arendt’s banality of evil is a another facet of thethem, acting at both the individual and collective level. Evil is banal because it is notperceived as such.


3.6 Mobbing and the Scapegoat Mechanism

Seeing mobbing behavior in action immediately shows that civil and decent peoplewho usually follow modern morality, inspired by legal/civil rules—that of courseexclude mobbing—can imperceptibly adopt a mobbing behavior, simply and spon-taneously. Mobbing is related to a kind of protomoral mechanism that Girard [7, 8]wonderfully described as the “scapegoat mechanism”. In this case moral bubblesrender the mobber able to avoid self-contradiction and to stay calm and satisfiedwith having adopted the supposed morality—momentarily considered as such—ofgossiping, obviously felt as exempt of violent possible effects.

3.7 Narcissism

The psychoanalyst Racamier [19] illustrated narcissism as a typical “perverse”,“healthy”, mechanism. In many cases narcissists are perceived as aggressive andviolent by their human targets, they feel themselves, to adopt a Kantian concept,“treated like means”. In this case too the narcissist is embedded in a moral bubble:she thinks her behavior right and moral, and even when advised about violent effectson other human beings, she is not able to become aware of it.

3.8 Sacrifices

It is well-known that in sacrificial ritual events in which human beings and animalswere in various cases killed or harmed, violence was disguised thanks to the activerole played by the accompanying moral bubble in which the related collective wasentrapped. Indeed sacrifices are usually endowed with primary and fundamentalmoral or moral/religious strong worths.

3.9 Moral Disengagement

When explaining above the mobbing behavior we said that it happens thanks to akind of moral disengagement: civil and decent people who follow modern morality,inspired by legal/civil rules, become violent perpetrators of possible violentmobbing.These “disengagers” are still embedded in a moral bubble in the following sense: asI already observed in my book [15, chapter five] the disengagement of morality—that is at the same time the reengagement of another moral system of rules—isperformed to the aim of making possible actions that would result violent in thelight of the previous morality. In this perspective the disengagement of morality is

84 L. Magnani

accompanied by a lack of awareness of the deceptive/aggressive character of theconsequent behaviors, because the new reengaged moral rules that trigger violenceand punishment are exactly chosen to conceal an condone the possible actually“violent” consequences of the previous ones.

3.10 Last Cases

Other cases could be illustrated, but here there is no room to further detail them. Ican list the most important ones.

Kant himself in [9, pp. 58–59] adumbrates the role of what I called moral bubbles(and so I contend he corroborates my perspective) when he sees human moral situa-tions in which the moral agent falls into a perfidious and self-deceitful reengagementin a new decisional framework where evil is simply supposed to be good, and somorally justified:

This is how somany human beings (conscientious in their own estimation) derive their peaceof mind when, in course of actions in which the law was not consulted or at least did notcount the most, they just luckily slipped by the evil consequences; and [how they derive]even the fancy that they deserve not to feel guilty of such transgressions as they see otherburdened with, without however inquiring whether the credit goes perhaps to good luck, orwhether, on the attitude of mind they could well discover within themselves, they wouldnot have practiced similar vices themselves, had they not been kept away from them byimpotence, temperament, upbringing, and tempting circumstances of time and place (thingswhich, one and all, cannot be imputed to us) (p. 60).

Other cases of moral bubbles in action can be seen in the extreme behavior ofthe violent “pure-evil” perpetrators, in confabulation, in the so-called fatal love, inviolent ideologies, in the so-called fascist state of the mind, in the psychological ideaof hidden trauma, in religious cognition, in the politically correct environments, andin the moral strategies of silence, I have sketched in my book [15] and that I willdeepen later on in a further article.

4 Conclusion

As I said at the beginning of this article a serious investigation of violence callsfor a discrete amount of courage and sincerity: as human beings we can wishfullyignore our own violence, thanks to that kind of “embubblement” I have illustrated.I also aimed at offering to the attention of researchers in sociology, psychology andpsychiatry the main features of those situations in which the awareness of humanviolent acts is in question, as indicated in the subtitle “Violent and Unaware of It:HowDifficulties in Recognizing One’s OwnViolence Lead to Disregard the InflictedHarm”, and in which a process of what can be called “autoimmunity” is at play. Ihave sketched various cases in which both the legitimizing and the dissimulation


of violence typical of moral bubbles are at stake. I tried to show that the conceptof moral bubble provides an integrated and unified perspective able to shed novellight on various situations in which morality and violence are intertwined: from theinvisibility of structural violence to the problem of the so-called moral viscosity,from the smartness of moral hazard to the difficulties of the identification betweenviolence and pure evil, to the issues regarding banality of evil, mobbing, narcissism,sacrifices, and moral disengagement.

Acknowledgements Parts of this article were originally published in chapter three of L. Magnani,Understanding Violence The Intertwining of Morality, Religion and Violence: A PhilosophicalStance, Springer, Heidelberg, 2011.

References

1. Arfini, S.: Ignorant Cognition: A Philosophical Investigation of the Cognitive Features of Not-Knowing. Studies in Applied Philosophy, Epistemology and Rational Ethics, vol. 46. Springer,Switzerland (2019)

2. Arfini, S., Magnani, L.: Cognitive autoimmunity: knowledge, ignorance and self-deception.Logic J. IGPL 24(4), 612–627 (2016)

3. Baumeister, R.: Evil. Inside Human Violence and Cruelty. Freeman/Holt Paperbacks, NewYork (1997)

4. Burton, R.A.: On Being Certain. St. Martin’s Press, New York (2008)5. Dowd, K.: Moral hazard and the financial crisis. Cato J. 29(1), 141–166 (2009)6. Gabbay, D.M., Woods, J.: The Reach of Abduction. A Practical Logic of Cognitive Systems,

vol. 2. North-Holland, Amsterdam (2005)7. Girard, R.: Violence and the Sacred [1972]. Johns Hopkins University Press, Baltimore, MD

(1977)8. Girard, R.: The Scapegoat [1982]. Johns Hopkins University Press, Baltimore, MD (1986)9. Kant, I.: Religion within the Boundaries of Mere Reason [1793]. Cambridge University Press,

Cambridge (1998) (translated and edited by Wood, A., di Giovanni, G.: Introduction by R. M.Adams)

10. Kruger, J., Dunning, D.: Unskilled and unaware of it: how difficulties in recognizing one’s ownincompetence lead to inflated self-assessments. J. Pers. Soc. Psychol. 77(6), 1121–1134 (1999)

11. Kuperman, A.J.: The moral hazard of humanitarian intervention: lessons from the Balkans. Int.Stud. Q. 52(1), 49–80 (2009)

12. Lahti, D.C., Weinstein, B.S.: The better angels of our nature: group stability and the evolutionof moral tension. Evol. Hum. Behav. 2, 47–63 (2005)

13. Magnani, L.: Abductive Cognition. The Eco-cognitive Dimensions of Hypothetical Reasoning.Springer, Heidelberg/Berlin (2009)

14. Magnani, L.: Understanding Violence. The Intertwining of Morality, Religion, and Violence:A Philosophical Stance. Springer, Heidelberg/Berlin (2011)

15. Magnani, L.: The Abductive Structure of Scientific Creativity. An Essay on the Ecology ofCognition. Springer, Cham, Switzerland (2017)

16. Magnani, L., Arfini, S., Bertolotti, T.: Int. J. Adv. Intell. Paradigms 8, 327–342 (2016)17. Muñoz, S.A.: The nature of epistemic feelings. Philos. Psychol. 27(2), 193–211 (2013)18. Muñoz, S.A.: Metacognitive feelings, self-ascriptions and mental actions. Philos. Inq. 2(1)

(2014)19. Racamier, P.C.: L’inceste et l’incestuel. Éditions du Collège de psychanalyse groupale et

familiale, Paris (1995)

86 L. Magnani

20. Thom, R.: Esquisse d’une sémiophysique. InterEditions, Paris (1988) (trans: Meyer, V.: SemioPhysics: A Sketch. Addison Wesley, Redwood City, CA (1990)).

21. Woods, J.: The Death of Argument. Kluwer Academic Publishers, Dordrecht (2004)22. Woods, J.: Errors of Reasoning. Naturalizing the Logic of Inference. College Publications,

London (2013)

The Computational Challengeof Amartya Sen’s Social Choice Theoryin Formal Philosophy

Gianfranco Basti, Antonio Capolupo and Giuseppe Vitiello

Abstract A significant chapter of the short history of formal philosophy is relatedwith the notion and the theory of the so-called “Social Welfare Functions (SWFs)”,as a substantial component of the “social choice theory”. One of the main uses ofSWFs is aimed, indeed, at representing coherent patterns (effectively, algebraic struc-tures of relations) of individual and collective choices/preferences, with respect toa fixed ranking of alternative social/economical states. Indeed, the SWF theory isoriginally inspired by Samuelson’s pioneering work on the foundations of mathe-matical economic analysis. It uses explicitly Gibbs’ thermodynamics of ensembles“at equilibrium” based on statistical mechanics as the physical paradigm for themathematical theory of economic systems. In both theories, indeed, the differencesand the relationships among individuals are systematically considered as irrelevant.On the contrary, in the mathematical theory of “Social Choice Functions” (SCFs)developed by Amartya Sen, the interpersonal comparison and the real-time informa-tion exchanges among different social actors and their environments—different—ethical values and constraints, included—play an essential role. This means that theinspiring physical paradigm is no longer “gas” but “fluid thermodynamics” of inter-acting systems passing through different “phases” of fast “dissolution/aggregationof coherent behaviors”, and then staying persistently in far from equilibrium condi-tions. These processes are systematically studied by the quantum field theory (QFT)of “dissipative systems”, at the basis of the physics of condensed matter, mod-eled by the “algebra doubling” of coalgebras. This coalgebraic modeling is highly

G. Basti (B)Faculty of Philosophy, Lateran University, 00184 Rome, Italye-mail: [email protected]

A. Capolupo · G. VitielloDepartment of Physics, University of Salerno, 84084 Fisciano, Italye-mail: [email protected]

G. Vitielloe-mail: [email protected]


87

88 G. Basti et al.

significant for making computationally effective Sen’s SCF theory, because bothbased on a dynamic and not statistical weighing of the variables for interacting sys-tems, respectively in the physical and in the social realms.

Keywords Formal philosophy · Social choice theory · Quantum field theory

1 Introduction

1.1 The Social Choice Theory: A Short History

A significant part of the mathematical theory of social, political, and economicalsciences, specifically in the context of welfare politics and economy, is related withthe development of the notion and the theory of the so-called social choice theory,in which the theory of “social welfare functions (SWFs)” occupies historically aprimary place. Social choice theory, as far as a formal theory, constitutes today abranch of the so-called formal philosophy.

This is a newborn discipline that, despite the name, started within the theoreticalframework of computer sciences, essentially for the stimulus of the AI research, formigrating during these very last years from the computer science departments intothe philosophy departments too (see [1] for an updated introduction). Substantially,it consists in the formalization of philosophical theories using the axiomatic methodapplied not only to mathematical logic—as it was since the beginning of the secondhalf of XIX cent.—but also, to philosophical logic.

Indeed, different philosophical theories (ontological, epistemological, ethical, so-cial) consist ultimately into different models or semantics of the axiomatic modallogical calculus (which is an extension of the mathematical logical calculus), re-lated to different interpretations of the possibility/necessity operators of modal logic.Think, for instance, at the fundamental distinction, not only in philosophy, but also inall social sciences, known as the “Hume problem”, between the logical necessity andthe ethical oughtness, on which the common-sense distinction facts-values formallydepends. The “social choice theory” consists therefore in the formalization of socialphilosophy, in strict relationship with the economic philosophy.

In a recent paper with the significant title Logic and Social Choice Theory, UlleEndriss so synthesizes the core of the social choice theory, as a chapter of formalphilosophy:

When a group needs to make a decision, we are faced with the problem of aggregating theviews of the individual members of that group into a single collective view that adequatelyreflects thewill of the people. How arewe supposed to do this? This is a fundamental questionof deep philosophical, economic, and political significance that, around the middle of 20thcentury, has given rise to the field of social choice theory ([2], p. 333).

The different theories depend then on the different sets of data on which the choicesare performed and, overall, on the different aggregation criteria of individual choices

The Computational Challenge of Amartya Sen’s Social Choice … 89

into one only social choice (see Sect. 3.2). Even though the precursor of the theoryof social choice is universally identified with Marie Jean Antoine Nicolas de Caritat,the Marquis de Condorcet (1743–1794), and with his famous paradox of voting (seebelow and especially Footnote 4), the social choice theory started in the last centurywith the pioneering contributions of Abram Bergson and of the Nobel Laureate inEconomics, Paul Samuelson leading to the so-called “Bergson-Samuelson SWF” [3,4], but it received a substantial improvement by the contribution of another NobelLaureate in Economics, Kenneth Arrow, who, for the first time, gave theory anaxiomatic form, demonstrating some fundamental theorems [5], and then finallywith a further Nobel Laureate, Amartya Sen, who gave the formalism of socialchoice theory an essential contribution [6, 7] (see Sect. 3.2).

Arrows SWF is intended as a function ranking social states as less, more, orindifferently desirable, for every pair of them, with respect to individual welfarepreferences. In the standard formalism inaugurated by Arrow, we have to do withlinear orders because based on the standard notion of transitivity that is a linearrelation.1 Formally, following Endriss ([2], p. 338):

Definition 1 (Social Welfare Function) Let L(X ) denote the set of all linear orderson X . A profile R = (R1, . . . , Rn) ∈ L(X )N is a vector of linear orders (i.e., pref-erences), where Ri is the linear order supplied by individual i . We write {N R

(x>y)} todenote the set of individuals that rank alternative x above alternative y under profileR. (…) A SWF is a function F : L(X )N → L(X ) that maps a profile of preferenceorders to a single (collective) preference order.

This strict relationship so instituted byArrowbetween the social choice theory andthe order theory in logic—as Endriss and Sen noticed—historically depends on thefact that Arrow followed in 1940, as an undergraduate student, a course with AlfredTarski. This justifies the usage of the lexicographic (see Footnote 1) symbolism insocial choice theory, like in order theory of logic and mathematics. This formalismis characterized by the usage of alphabetical ordering or of numerical ordering in thesymbolism,where, however, numbers are properly ordinal and not cardinal numbers.From that, the usage of the symbol ≻ and derivatives (“higher than”) instead of >(“greater than) and derivatives, since we have to deal here with “topological” andnot “geometrical” algebraic relations, because in principle independent of a given,shared metrics, on which only a quantitative ordering relation can be defined.

1Indeed, using the standard ideographic symbolism, e.g., the alphabetic letters, x, y, z, for repre-senting relations R, the ordering relation among objects using the transitive relation is representedas follows: ((x Ry ∧ yRz) → x Rz). As we see, in such a way the transitive relation is linear, andthen in set-theoretic ordering it supposes a total ordering among sets. That is, for all sets the or-dering relation ≥ holds—because admitting gaps (in our case, between x and z) among sets thatcan be indefinitely long. Another possible transitive relation is the so-called “Euclidean relation”:((x Ry ∧ x Rz) → yRz ∨ zRy) that has a tree structure in which no gap among sets is admitted,and in which therefore set total ordering is not necessary, but only partial orderings are sufficient,because in this case—and this is the essential point!—transitivity is not linear. As we see, this dis-tinction between total and partial orderings in social choice rankings is essential for understandingthe novelty of Sen’s approach to social choice theory as to the classical ones based on a liberalapproach in economy.

90 G. Basti et al.

Given this definition of SWF as the starting point of the social choice theory, andemphasized its relationship with formal logic and then with formal philosophy, wecan outline the scheme of our contribution.

1.2 The Scheme of This Contribution

In the Second Section of our contribution we illustrate the passage from the SWFsto the Social Choice Functions (SCFs), in the framework of social choice theory,in relationship with Sen’s “SCF theory”, because the passage from SWFs to SCFsdepends largely on Sen’s original contribution.

Indeed, another theoretical background to take into account in our examination ofthe early social choice theories as SWF theories, is their development in the frame-work of an approach to the study of social and economic systems at equilibrium (or inclose-to-equilibrium) conditions, as the standard formalism of statistical mechanicsin physics requires. The so-called classical and neo-classical economic theories, inwhich early SWF theories were developed, are all inspired, indeed, by Samuelson’sgeneral approach to the mathematical economics in his seminal handbook [4]. It usesindeed as its reference physical paradigm for modeling economic theories Gibbs’statistical mechanics, and gas thermodynamics, to which the first two chapters ofSamuelson’s handbook are significantly dedicated. This is because this mathematicaltheory is naturally consistent with the liberal individualistic vision of economy andsociety. In both approaches indeed, statistical mechanics and statistical economy inits liberal interpretation, the relations among individuals, either physical particlesor economic actors, are considered as ultimately irrelevant. And significantly, Sen’scriticism to Samuelson’s approach starts just from this point.

Unfortunately, indeed, a fundamental unexpected and undesired consequence ofArrow’s SWF theory is the famous “Arrow’s impossibility theorem” [5] demonstrat-ing formally—following the efficacious synthesis of Amartya Sen—the “impossibil-ity of the Paretian liberal” [8] for a formally consistent social choice theory. In fact,one of the more troubling corollaries of Arrows impossibility theorem is the incon-sistency for democratic systems of social choices based on the majority decisions(see Sect. 2.3) on a liberal basis. That is, any liberal approach to SWF theory likeArrow’s one takes as an axiom the so-called Pareto’s efficiency principle, for whichall the individual preference rankings have the very same possibility of determiningthe social one: an abstract principle that is evidently unrealistic in photographing theeffective situation in social sciences. On the other hand, seen from the standpointof social choice theory, the voting democratic principle is nothing but a way foraggregating individual preferences into one only social ranking using as aggrega-tion criterion the majority principle. The fierce debate triggered by Arrows resultsin the next fifty years contributed definitely to the development of the formal andmathematical theory of social choices, as a fundamental mathematical discipline forsocial, political, and economical sciences. Not casually, an impressive testimony ofthis debate is given by the bibliography of the second enlarged edition, published in


2017 [7], of Amartya Sen’s fundamental book Collective Choice and Social Welfare,firstly published in 1970 [6].

Now, as we discuss briefly below, one of the most significant contributions ofSen’s theory, determining its immediate diffusion among scholars and the widerpublic, was precisely the formal demonstration that the onlyway for avoidingArrowsimpossibility results is to enrich Arrows theory with two new features.

1. Before all, by enriching SWF’s ranking (ordinal) measures with (cardinal) mea-surements and comparisons of capabilities, resources, and utilities among in-dividuals, so to transform SWFs in as many and more realistic social welfarefunctionals.

2. Secondly (and definitely), by introducing into the model the interpersonal com-parison of goods and utilities—and, more generally, the information exchangesamong persons constituting homogeneous groups—, on the contrary consideredas irrelevant in the classical (liberal) economic theory, as well as in standard SWFtheories, like Arrows one.

In the Third Section of this contribution we discuss therefore the main characters ofSen’s “social choice function (SCF) theory”. Indeed, the two just remembered im-provements as to Arrows SWF theory allowed Sen to introduce into themathematicalmodeling of his SCF theory comparative principles of social and economic equityor fairness, using as aggregation principle the ethical maximin principle that givespriority to the interests of worst-off persons used before by one of Sen’s teacher,John Rawls, in his celebrated political theory of “justice as fairness” [9]. The maindifference, as we see, is that, for Rawls’ neo-Kantian normativism the maximin prin-ciple supposes again—if applied to social choice theory—an untenable in our pluralsociety, because absolute and then abstract, “total ordering of the distributive justicegrading. On the contrary, this limit does not hold in Sen’s SCF theory, where onlypartial orders are allowed, but where it is possible to merge different local gradingsof justice into one shared ranking. In such a way, the maximin becomes a powerfulaggregation criterion of comparative distributive justice, not supposing any unreal-istic reference to an absolute criterion, for unveiling and solving concrete injusticesituations. This possibility transforms Sen’s SCF Theory into a comparative theoryof social choices, which considers as intrinsic to the models also the ethical and theinterpersonal constraints, because of Sen’s wider and systematic usage of the max-imin principle as a local equity principle we illustrate below. On this regard, from thestandpoint of the formal philosophy of ethical choices, the fundamental contributionof Sen was the formal demonstration that an effective mathematical modeling ofethical constraints in economy cannot be based on abstract and not-computable op-timal choices defined on a unique, because supposed as complete (=total) grading ofdistributive justice for social/economical states in a society, but on concrete criteriaof maximal choices relative to the different contexts, that is, based on interpersonalequity criteria. This mathematical modeling of SCFs can constitutes also the basisfor more suitable measures of social well-being and welfare than GDP, recently de-veloped in social, political and economic sciences with the passionate and excitingpublic discussions that followed, to which Sen gave a fundamental contribution [10].

92 G. Basti et al.

Finally, in the Fourth Section of this contribution we want to discuss the computa-tional challenges of Sen’s SCF theory. Indeed, as Sen himself emphasizes in manyplaces, precisely the difficulty of applying his theory using the classical statisticaltools of the economic analysis is the main “technical” obstacle—apart from the ideo-logical ones—against its systematic application. The problem is well synthesized bySen saying that an approach like his SCF theory, all based on the interactions amongpersonal and collective social actors, requires from the mathematical standpoint adynamic weighing of the variables that no statistical weighing function can grant inprinciple.

Moreover, also from the computer science standpoint, we are faced in AImachinelearningwith the very same problem.Aswe see, indeed, it is absolutely unfittingwiththe required dynamicweighing of the variables the statisticalmachine learningmodelin AI systems, based on the so-called “deep-learning” approach of artificial neuralnetworks (ANN) in AI (see Sect. 4.1). This approach, indeed, maintains the classicaldistinction of the statistical analysis, between the determination of the statisticalweights of the variables, and their further application on the whole data set. In theANN modeling, this distinction becomes between:

1. The necessarily automatized—because of the presence of inner higher order cor-relations among huge amounts of data that makes impossible the determination oftheweights by hand—(deep) learning phase for fixing theweights of the variableson a representative sample of the data set, followed by

2. The testing phase, consisting in the application of this statistical weighing toanother different representative sample of the data set, and then, if this testing issuccessful, to the whole data set.

To stress the relevance of this point, let us consider the same problem of the infor-mation exchange among social/economical actors from another but complementarystandpoint as to Sen’s political theory. In fact, the real-time information exchangeamong communicating agents that internet today allows, determines the fast aggre-gation/dissolution of interest groups in a world-wide environment. For an effectiverepresentation of this phenomenon evident to everybody, think, for instance, at ourworldwide currency and stock-exchange markets, and at the streams of data, withalways changing inner correlations, which internet produces in real-time,2 beforeon the computer screens, and then on the consequent choices of the financial opera-tors, distributed over the five continents. As we know, this problem of “infinite datastreams” in Theoretical Computer Science (TCS) belongs to the wider class of the so-called “big-data issue”, determining the growing interest to innovative computationalsolutions, the controversial quantum computer architectures included.

Indeed, “infinity” is here related with the continuous, sudden aggregation/disaggregation of coherent patterns of behavior of the financial agents (humans andcomputers), in markets distributed all over the world, dramatically represented by the

2E.g., think at Bloomberg data streaming, updating continuously the values of shares and of curren-cies on the worldwide financial market, active all over the day because of the different time zones,and that are changing in average every 10s.


(apparently) erratic oscillations of multiple economical indices on the many screensof each financial operator. The higher order inner, long-range correlations amongthese inter-related oscillating statistical quantities correspond, indeed, to as manyphase transitions among indefinitely many phase coherent behaviors of markets, be-cause of the real-time information exchanges among operators, both humans andmachines.

Coming back to the ANN deep learning approach in AI, making “dynamic”the weighing of variables for locking the network dynamics onto the continuousdata oscillations like Sen’s theory—but also like the modeling of data streaming—require would signify to design ANN architectures in continuous learning—just as,in parentheses, the natural NNs of our brains are. This, however, makes immediatelychaotic—in the technical sense of the dynamic system theory notion of “determinis-tic chaos”—and then absolutely unpredictable using the classical statistical methods,the overall network dynamics.

To sum up, from the precedent discussion it is evident that, as fa as we considerthe interactions among social and economic actors, we cannot any longer considergas thermodynamics as the reference physical paradigm, but fluid thermodynamicshaving in the quantum field theory (QFT) of dissipative systems—that is, of “opensystems” in continuous interaction with their environments—its fundamental theory.It is not casual that, by the principle of the so-called doubling of the degrees of freedomcharacterizing the QFT of dissipative system, we can satisfy the requirement of adynamic determination of the statistical expectations—or “dynamic weighing” ofthe variables involved. That is, what we are searching for solving the computationalchallenges of Sen’s social choice theory and/or of data streaming, as far as bothcharacterized by the continuous change of the inner correlations among data.

Therefore, we propose in this contribution the path toward an original solutionfor this problem, based on the so-called “algebra doubling” of coalgebras, by whichmodeling interactive systems, in computational, and then in physical, cognitive andnow also social systems [11]. What is relevant, in our proposal—even though its fulldevelopmentwould require awider discussionwe cannot develop here—is that such amodal coalgebraic semantics is able to implement in the same computational model,not only the mathematical component of Sen’s SCF theory, but also its intrinsicdeontic component based on Sen’s theory of “comparative justice as equity” [12]. Atheory that is consistent on his revolutionary idea of “development as freedom” [13].Namely, based on Sen’s theory of the fair and concrete distribution of capabilities(i.e., of effectively available choices among different possibilities of “being” and“doing”) and of functionings (i.e., of concrete ways of “being” and “behaving”) ofpersons and groups, in order to fulfil their free self-achievement as persons, accordingto the different value systems they choose to share, in a pluralistic ethical panoramasuch as ours.

On the other hand, the logical consistency and the computational effectivenessof a theory of comparative justice as equity is fundamental for supporting, not onlymore complex and effective measures of development and of well-being for nationsand groups than the obsolete GDP [10], but also for overcoming, by the formal andthen universal comparability of different moral systems, the moral relativism that

94 G. Basti et al.

negates the universality of human rights, and so offends the weaker persons andpeoples in our society, without, on the other hand, supposing an absolute ethicalsystem of reference that would be again another form of oppression.

Indeed, by substituting the abstract universalism/normativism of a theory of jus-tice, such as, for instance the Kantian one, with a concrete and computationallyeffective theory of “comparative justice as equity”, because based on criteria ofequity among persons and groups sharing different value and moral systems, it ispossible to make effectively and universally comparable the local solutions of globalproblems for different contexts in our open society.

This is a tremendous task [14] for which the aid of suitable AI systems for sup-porting the social choices of policy and decision makers in economics and societyis unavoidable, according to a notion of collective wisdom, including necessarilyhumans and machines [15, 16]. All this emphasizes the relevance of proposing com-putationally effectivemodels in the field of social choice theory. And these reflectionswill constitute the Conclusions of our contribution.

2 Social Welfare Functions as Social Choice Functions

2.1 Two Main Types of Social Welfare Function

Generally, in economic literature, there are two main types of Social Welfare Func-tions (SWFs), as far as they are defined respectively, either on

1. Some support (domain-codomain) of real valued economicalmagnitudes (definedon cardinal numbers) for one only social group, or

2. Orders, i.e., domains-codomains of rankings (ordered sets) of preferences, andof rankings (ordered sets) of social states, both defined obviously on ordinalnumbers.

Now, when we are speaking about orders, effectively we are considering a SWFlike a particular type of Collective Choice Function (CCF), relating ordered setsof individual preferences/utilities, with ordered sets of social/economical states foreach individual. This means that, mathematically, we are moving from real valued(cardinal) functions, to set-theoretic logic functions—effectively we are moving toset-theoretic semantics (“order theory”) applied to social entities. The initial rep-resentatives of these two types of SWF theories generally quoted in literature are,respectively, “Bergson-Samuelson SWF” and “Arrow SWF”.


2.2 Bergson-Samuelson Social Welfare Functions

Abram Bergson first introduced in economics the SWF notion as real-valued dif-ferentiable functions, aimed at formally representing “the conditions of maximumeconomic welfare” for the society as a whole. Bergson’s SWF then includes as func-tion arguments several quantities of different commodities produced and consumed,and of resources, labor included [3].

The fundamental contribution of the 1970 Nobel Prize in Economics PaulSamuelson—founder of the prestigious “MIT School of Economics”, countingamong its members an impressive list of Nobel Prizes—is synthesized in what iscommonly called the “Bergson-Samuelson SWF”. It aims at representing (in themaximization calculus) all real-valued economic measures of any belief system—“that of a benevolent despot, or a complete egotist, or all men of good will, a mis-anthrope, the state, race, or group mind, God, etc.”—required to rank consistentlydifferent feasible social configurations as “better than”, “worse than”, or “indifferentto” each other in an ethical sense (see [4], p. 221).

What is essential for our aims is that Samuelson’s modelling of the equilibriumstability for economic systems explicitly depends on Willard Gibbs statistical me-chanics interpretation of thermodynamic systems, as Samuelson explained in the firsttwo chapters of his masterpiece Foundations of Economic Analysis, of which eighthchapter is dedicated to welfare economics and then to his SWF interpretation in sucha theoretical framework (see [4], pp. 203–256).

2.3 Arrow Social Welfare Functions and the “ImpossibilityTheorem”

The second type of SWF is related with the work of Kenneth J. Arrow, awarded withthe Nobel Prize in economics on 1972. Since the first version of his theorem (1948),indeed, he transformed effectively Bergson’s SWF into a collective choice function(CCF), as the same title of his fundamental book Social Choice and Individual Valueexemplifies [5]. That is, whereas the Bergson-Samuelson SWF rules the mappingfrom any set of individual orderings of preferences/utilities into one only set ofordering of social states, Arrow SWF rules the mapping from any set of individualorderings of preferences/utilities into a set of social states, among many alternativeones (see [5], pp. 22–25). As Arrow himself emphasizes the interpretation of SWF asa CCF is effectively a restriction over SWF, because it requires that for any individualordering, for some sufficiently wide but finite range of them, the SWF “give rise toa true social ordering” among finitely many ones.

Arrow’s CCF effectively poses other conditions to SWF giving rise to the famousArrow Impossibility Theorem. Following the later version of Arrow’s theorem (see[5], pp. 46–61) these conditions can be synthesized as follows.

96 G. Basti et al.

Let A be a set of alternatives of social states, N a number of individuals or pref-erences i, and L(A) the set of all linear orderings of A. The SWF, intended as anindividual preference aggregation rule, is a function F : L(A)N → L(A) which ag-gregates individual preferences Ri on N into a single order on A. The N-tuple(R1, . . . , RN ) of Ri is called an “(individual) preference ordering”.

The “Arrow Impossibility Theorem” states that for N at least of 2 individuals,and for A at least of three alternatives the follow four conditions are incompatible—particularly the fourth one as to the others, and overall as to the second one, as Senwill demonstrate:

1. U*: Unrestricted Domain. The domain of the function (rule) F must include alllogically possible individual orderings for a finite set of them;

2. P*: Pareto Efficiency or Unanimity. All individual orderings have the same possi-bility of determining the social state ordering, i.e., if alternative x is ranked strictlyhigher than y for all orderings R1, . . . , RN , then also for F(R1, . . . , RN );

3. I*: Independence of Irrelevant Alternatives. F depends strictly on the pairwiserelations associating subsets of A and N. I.e., if for two different individual pref-erence orderings R and S the alternatives x and y have the same order in R andR′, then also in F(R1, . . . , RN ) and F(S1, . . . , SN );

4. D*:Non-Dictatorship. There is no individual i ∈ 1, . . . , N such that for∀(R1, . . . ,

RN ) ∈ L(A)N , x ranked strictly higher than y implies that x is ranked strictlyhigher than y, for all x, y.

To sum up, the “Arrow Impossibility Theorem” states that there is no SWF satisfyingsimultaneously the conditions U*, P*, I*, and D*. Roughly speaking, given thecondition P*, for any individual set ordering, there must be one pivotal individualthat orders the alternatives. Now, because of the transitivity and asymmetry of thestrict individual orderings supposed inArrow’s construction (see below), it is possibleto demonstrate that this pivotal individual is unique for all individual orderings, soviolating the condition D* of non-dictatorship.

The theorem had during the last fifty years a tremendous impact overall in po-litical sciences because, if taken for granted, it would demonstrate the irrationalityof vote systems in democracy, as far as based on the principles of “majority” and of“representativeness”, which suppose evidently that the conditions P* and D* holdsimultaneously.3

3There exists evidently, and many Authors noted it, a relationship between Arrow’s impossibilitytheorem and the famous Condorcet’s paradox on voting stated in his Essai sur l’Application del’Analyse à la Probabilité des Decisions Rendues à la Pluralité des Voix published in 1785 [17],and considered unanimously as the official birth-date of the modern, mathematical theory of socialchoices. However, Sen demonstrated that Arrow’s impossibility theorem cannot be reduced toCondorcet’s paradox, before all because it would be hard to argue that majority rule would really bea plausible way of aggregating preferences in welfare economics (see [7], p. 275). See also [7], pp.395–419, and overall the concluding chapter of Sen’s book ([7], pp. 453–472), in which he offers adeep analysis of what a modern comprehensive, because rational theory of social choice requires inour “Communication Age”, and that is far beyond any oversimplified and oversimplifying “majoritycriterion”.


Effectively, two are the main ways for avoiding the outcome of Arrow’s theorem:

1. Working on infinite sets of preferences—i.e., on finite sets of ever-changing pref-erences and aggregation rules—, and not on finite (static) ones like in Arrow’sSWFs. In such a case, however, as far as we suppose a total ordering relationof sets, the aggregation rules seem to be of limited interest because in standardset-theoretic logic they are based on ultrafilters,—i.e., maximal filters on given“partially ordered sets” or posets (see Rel. 9 below), supposing the total order-ing (see Rel. 8 below) of all partial orders. Namely, this means a supposition ofcompleteness in the definition of the social rankings that is simply illusory tosuppose in complex and global societies like ours. “Ultrafilters”, in other terms,are non-constructive mathematical objects because they suppose the (infinite)set total ordering—i.e., their definition requires a non-computable higher orderlogic—, so to require practically as many “invisible dictators” [18]. And, in fact,such aggregation rules violate Turing computability [19], so to result practicallyuseless in mathematical social choice theories, and in social computing appliedto economics and to social sciences.

2. Working exclusively on set partial orders in social choice theory, instead of thestrict orders of Arrow’s SWF (see Sect. 2.4 the list of the main set orderingrelations), so to define another type of rational CCFs, the SocialDecisionFunction(SDF). For SDFs Arrow’s impossibility theorem, indeed, does not generally hold,even though it holds for SWFs as special cases of SDFs. This approach was firstlypro-posed by another Nobel Prize in Economics (1998), Amartya Kumar Sen,who gave in this way an essential contribution of clarification to our discussion[6, 7].

We can anticipate, however, that also in Sen’s approach the challenge of infinity andhence of effective computability on infinite sets, e.g., data streams, represents itselfas an unsolvable computational problem within the Turing Universality paradigmin computations. This holds also in the case of the “probabilistic” Turing Machine”architecture, artificial neural networks (ANN) included, as far as based on a purelystatistical weighing of the variables, through the machine learning principle—i.e.,the so-called “deep learning” of the actual AI systems—of the stochastic gradient de-scent, aswe discuss briefly in theThird Section of this contribution. These limitations,indeed, hold till we are working in standard set theory—e.g., the Zermelo-Fraenkel(ZF) set theory—and not in non-standard ones, like the theory of Non-Wellfounded(NWF) sets, requiring the “Algebraic Universality” in computations, and not the“Turing Universality” that is a subset of the former one, as we hint below. Unfortu-nately, a satisfying discussion about this fundamental point would require anotherspecific study that we cannot develop in this contribution.

98 G. Basti et al.

2.4 Amartya Sens Logic of Preferences and His Notionof “Social Choice Function”

Formally, Sen’s contribution to the discussion consists in a rigorous clarification ofwhat “preferences/evaluations among alternatives” (e.g., “better than”, “as good as”,“at least good as”, etc.) means in the framework of the logic (algebra) of relations, forsolving ambiguities and inconsistencies that often appear in SWF theories interpretedas CCFs. We illustrate Sen’s contribution essentially following his outstanding bookCollective Choice and SocialWelfare, in its first edition published in 1970 and then in1979 [6], in which he outlines a mathematical logic of preference/utility in economicand social sciences. As we anticipated in the Introduction, Sen recently published in2017 an enlarged edition of this fundamental work including the first one, plus somefurther extension we discuss below [7].

To summarize briefly the core of Sen’s formal approach to social choice theory,we have to recall before all the basic fact that mathematically whichever prefer-ence/evaluation binary relation R over a set S consists in specifying a subset R ofthe Cartesian Product S × S, defined as the set of all ordered pairs (x, y) such that(x, y) ∈ S. Following the standard order theory in set-theoretic logic, Sen definesrigorously the following list of relationsR over a set S in order to offer a lexicographictheory of social orderings:

1. Reflexivity: ∀x ∈ S: x Rx2. Totality (Completeness): ∀x, y ∈ S: (x = y) → (x Ry ∨ yRx)3. Transitivity: ∀x, y, z ∈ S: (x Ry ∧ yRz) → x Rz4. Anti-symmetry: ∀x, y ∈ S: ((x Ry = yRx) → x = y) ∨ ((x Ry = yRx) → x =

y)5. Asymmetry: ∀x, y ∈ S: x Ry → ¬(yRx)6. Symmetry: ∀x, y ∈ S: x Ry → yRx

For logical and computational aims, it is worth to notice that anti-symmetry includesasymmetry, and not vice versa, so that the implication connective is logically correctonly in the case of anti-symmetry. Indeed, in the case of equivalence—which is adouble, symmetrical implication—anti-symmetry but not asymmetry holds, sinceasymmetry holds in the case of the simple implication only. This observation is es-sential for understanding the logical notions of partial and total orders, in all theirpossible versions, where anti-symmetry plays an essential role. Effectively, Sen de-fines also and newly in a standardway, different types of ordering relations, accordingto the different 1–5 basic relations they satisfy:

7. Pre-ordering: 1, and 3;8. Total (Complete) pre-ordering: 1, 2, and 3;9. Partial ordering: 1, 3, and 4;10. Total (Complete) ordering: 2, 3, and 4 (in the supposition that 2 implies 1);11. Strict ordering: 2, 3, and 5.

Sen defines thus the ordering relations 7, 8, and 9 as “quasi-orderings that is thekey-notion of his logic of preferences, evidently based on the notion of “pre-orders”


(see [6], p. 10). Then, Sen notices, in the case of Arrow’s SWF, the ordering musteffectively satisfy 1, 2, and 3, but it is, per se, irrespective of 4 and then of 5 (see[6], p. 8). In such a way, we can avoid, under some conditions explicated by Senin the rest of the book that we only summarize here, the uniqueness of the pivotalindividual for all social orderings, as depending on some rational CCF on social stateshe defined as social decision function (SDF). This opens the possibility of satisfyingsimultaneously the conditions P* and D* of Arrow’s theorem as characterizing anyCCF, and then a SDF too, even though at the price of an organization of social statesmuch more fluid, than in Arrow’s vision.

In other terms, the core of Sen’s approach consists in demonstrating formallythat we can obtain suitable CCFs such as SDFs, satisfying all Arrow’s conditionswithout falling into his “impossibility result”, by supposing quasi-orderings andoverall partial orderings instead of strict orderings like in Arrow’s approach. Werefer to ([7], pp. 313–337), for a complete discussion of Sen’s formal treatment ofthe CCFs and SDFs in social choice theory.

Anyway, to conclude this subsection, let us recall, another interesting consequencederiving from the finiteness of SDFs—and of the Social Choice Functions (SCFs)we discuss below.

That is, the possibility of representing in it also judgements about values, sharedby a homogeneous social group because characterized by strong interpersonal ex-changes of information. These topics were the main object of a more recent essayof Sen, with the significant title: The Informational Basis of Social Choice [20], andthey are the leit-motiv of the new part added to the Second Expanded Edition of [6],even though largely discussed also in ([6], pp. 89–117), where he introduced thetransition from a SDF to a SCF. We return in the next subsection on this fundamentaltransition.

However, when we study social and economic systems, taking into account alsothe motivational social forces acting within them, overall if spread in real time bysocial media, the challenge of infinity represents itself becausemaking untenable thateither a SDF or a SWF can range over a fixed finite set of social state alternatives,chosen by individual preference orderings.

Nevertheless, the situation changes radically when we pass from individual pref-erence orderings (rankings), to social choice functionals, because expressed by ho-mogeneous groups of persons, that is, by groups of individuals characterized by astrongmutual information exchangemaking them an unique social subject: this is thecore of the passage from SDFs, to SCFs. Using a thermodynamic metaphor, this isthe core of the passage from a “liberalist” vision of society seen as a “gas” or “vapor”of individuals, to a “pluralist” vision of society characterized by the “condensation”of individuals into homogeneous but reciprocally different groups—like as manydifferent “drops”, for continuing in the thermodynamic metaphor. In such a case,the “infinity challenge” represents itself in a different way, because of the boundarycondition of the social media influence, often on a world-wide dimension—thinkat internet, before all. The influence of social media indeed, in specific contexts,acts like sudden changes of “temperature” in the human environment, determiningimmediate rarefaction/condensation processes of old/new groups inside the human

100 G. Basti et al.

community, often in a world-wide, super-national dimension. In a word, such a “liq-uid” situation of our society and of our economy requires a paradigm shift in themathematical approach to economic, social, and political sciences. That is, a passageto a modelling based on fluid thermodynamics that at the fundamental physical levelis modeled in the quantum field theory (QFT) of dissipative systems, as we discussbriefly in the Fourth Section, being the common coalgebraic modeling, the formalbridging between mathematical physics and mathematical logic.

3 Sen’s Theory of the Social Choice Functions and ItsAnthropological Background

In this section we develop our analysis in several steps. After a general definition ofwhat is a social choice function (SCF) in social choice theory, we examine, beforeall, Sen’s social choice function definition, and the consequent “SCF possibilitytheorem”.

Afterward, we discuss the relevance of the information basis in SCFs. Then, weexamine John Rawls’ theory of “justice as fairness” in the context of social choicetheory because this prepares Sen’s comparative theory of justice as equity, on whichhis SCF theory is based, and that constitutes the core of its originality and complexity.

The examination of the anthropological background of this theory constitutesparticularly the object of Sect. 3.4 concluding this section. Indeed, Sen’s SCFs theorystrictly depends on his original notions of the so-called personal capabilities andfunctionings, as the key-concepts of his comparative theory of “justice as equity”.

3.1 Sen’s SCF Theory and the “SCF Possibility Theorem”

Generally, as we have seen, SCF in axiomatic social choice theory is a functionF : L(X )N → 2X \{∅} mapping profiles of linear orders L on alternatives X , tonon-empty sets of alternatives.

Intuitively, given a profile of declared preferences, F will choose the best alternatives [2], p.341.

Therefore, according to Sen, the possibility of eschewing the notion of individual orsocial preferences, toward a theory of social choice functions received considerableattention in the literature of social choice theory in recent years. Indeed, in the caseof SCF theory, it is systematically avoided the necessity of defining a set of binarypreference relations over one only ranking (ordering) of alternatives. A non-binaryformulation of social choice, indeed, works directly with a choice function C for asociety, defined over the alternative rankings that can be chosen by individuals andby groups, i.e., over a choice set C(S) for each menu (set) of alternatives S.

From this Sen’s general definition of a SCF follows ([7], p. 313):


Definition 2 (Definition of Social Choice Function) The functional relation betweenthe set of alternative menus S and the corresponding choice set C(S) can be calledthe social choice function (SCF), and this needs not invoke (…) any binary relationof social preference.

Therefore, a functional collective choice rule (FCCR) maps n-tuples of individualpreference relations Ri to a social choice function:

C(•) = f ({Ri }) (1)

It is evident that a SCF escapes from Arrow’s impossibility results simply becauseArrow conditions are essentially binary relational demands (i.e., individuals vs. onesocial state ranking). In fact, in the present case, what is required is not a SWF (orSDF) (see Sect. 2.4), but a SCF:

g (S, {Ri }) such that g (S, {Ri } ⊆ S for all non-empty subsets S ⊆ X (2)

This is essentially equivalent to making the value of the function f (Ri ) a finitely completechoice function C(•) for each society, and not – as with SWFs or SDFs – a social preferencerelation R [7], p. 315.

In thisway, for such a FCCR: f (•), the SCF theory requires a re-definition ofArrow’sconditions U, P, D and relative axioms (see Sect. 2.3), in terms of restrictions of n-ary choices over pairs only. That is, U , P, D, plus the extension of non-dictatorshipcondition D to a stronger condition of “anonymity” A. I.e., if Ri is a permutation of!R

′i

", then f ({Ri }) = f

#!R

′i

"$.

Unfortunately, we cannot discuss here this significant point, and we refer to [7], p.316 for further particulars. Now, if we combine these conditions to the other Arrow’scondition I of “independence of irrelevant alternatives” that was defined in functionalterms, i.e., as an I A by Arrow himself, we can state the following theorem [7], p.316:

Theorem 1 (Choice-functional Positive Possibility Theorem) For H-ariety ≥ 2,there is a FCCR satisfying conditions U , I A, P, D and A.

In the following discussion [7], pp. 316–336, Sen emphasizes the relevance ofhaving an ariety (number of function arguments) H -ariety ≥ 2, where H is the setof individuals, in a social choice functional interpretation of SWFs and SDFs. Indeed,also in the case of restrictions of choices over pairs only, a FCCRmust satisfy anywaythe condition of transitive closure x B∗y of a sequence of at least a triple of binaryrelations xBy, having necessarily a maximal choice set C(Smax) including the factorBA of B, defined as “xBy and not yBx”, which is an asymmetry condition that closesthe sequence, avoiding any cyclicity. We return in Sect. 3.4 on the centrality of thismaximality condition for emphasizing the computability and then the effectivenessof Sen’s SCF theory also if we inserted in it ethical personalistic components.

Anyway, this can be seen straightforwardly also in the more elementary case ofCondorcet’s paradox of voting, interpreted as a choice function. In fact, the transitive

102 G. Basti et al.

closure intrinsically associated with a maximality condition avoids the cyclicity ofpreference orderings, from which the paradox derives.4 In a word, it is the finitecompleteness of a choice function, because including necessarily an intrinsic max-imality criterion, the core of a SCF theory. And this, without supposing, either themaximization of some extrinsic “optimization function”, or some criterion of inter-nal “consistency” among the different menus of social alternatives that are, on thecontrary, independent of each other, as it is the case of a “plural society” such asours.

3.2 SCF Theories and Their Information Bases

However, it is not in this “possibility theorem” extended also to a theory of votingand then fundamental for a formal justification of a democratic regime based on themajority principle in a pluralistic society, the main advantage of Sen’s SCF withrespect to Arrow’s SWF (see on this point [7], pp. 220–250; 395–419). Indeed, moregenerally, what is systematically lacking in Arrow’s social theory is the possibil-ity of an interpersonal comparison of people well-beings, using several differentquantitative indices, or the possibility of any other comparative data about peoplerespective advantages. Effectively, as Sen noticed, Arrow was following in this thegeneral consensus among scholars emerged in the ’40s of last century that “the in-terpersonal comparison of utilities has no meaning and no relevance in economictheory mathematical modeling ([5], p. 9). A prejudice that is formally related to agas-like or individualistic (liberalist) social theory, as we said many times. On thecontrary, if it possible to allow in a mathematical social/economic theory the com-parison of social/economic states based on “interpersonally comparable informationabout different “persons” (i.e., “individuals-in-relation”), many possibilities of mak-ing systematic and realistic social welfare judgements open up. Particularly, we caninsert in the social welfare theory “ethical concerns such as equity, or removal ofdeprivation, or enhancement of peoples freedoms, etc.” ([7], p. 338).

For this aim, in the light of the above Eq. (1) defining a FCCR with respect to agiven SCF, Sen examines, in a very original and intriguing way, different theories ofsocial justice in terms of their “informational bases”. That is, in terms of the presenceof information exchanges about states, achievements and opportunities of the people,which are central to assessing justice and injustice and their grading in a givensociety according to the examined theory, at the same time ruling out informationof other types considered as irrelevant. For instance, the utilitarian theory of justice

4Indeed, the paradox derives from the fact that voter 1, 2, and 3 are ranking the three alterna-tives in the orders: x, y, z; y, z, x; and z, x, y, respectively. In this way, there is a preferencecycle, and each alternative is defeated in a majority vote by another alternative. Now, if we donot consider preference maximization, but maximality in terms of the transitive closures of pref-erence relations, we have the following choices: C ({x, y}) = {x}; C({y, z}) = {y}; C({z, x}) ={z} ; and C {(x, y, z)} = {x, y, z}. Of course, this choice function is not binary, but satisfies all theArrow axioms in their choice-functional version, of which at Theorem1 (see [7], p. 316).


attaches relevance to the economical utilities of individuals, and it does not considerinformation about fulfilment or violation of rights or of liberties, levels of incomesor affluence that people enjoy—if not for the indirect effects that such variables haveon individual utilities. On the contrary, a libertarian theory of justice stresses overindividual liberties and rights, and it gives no direct relevance to the levels of utilitiesor of incomes, and of their interpersonal comparison.

To sum up, Sen defines the informational bases of different theories of justice forinserting them in the functional structure of a SCF in terms of (see [7], pp. 339–340):

1. The basal space of the theory, i.e., the class of variables on which the justiceassessment of that theory depends (e.g., in utilitarian theories, the combinationsof the individual utilities);

2. The aggregation system of those variables, for allowing a discriminating use ofthese variables. E.g., in utilitarian theory, the individual utilities are simply addedtogether to arrive at a sum-total, serving as basis for the assessment of the so-cial state, without paying attention, for instance, to some measures of dispersionor inequality. In this context, the aggregation systems of theories of justice thatinclude as their essential components “distributive criteria” of commodities, utili-ties, and opportunities among individuals become particularly relevant. These arethe so-called theories of justice as fairness or equity like Rawls’ and Sen’s ones.

3.3 SCF Theory and Rawls’ Theory of “Justice as Fairness”

Effectively, in all his work and not only in his Collective Choice book in both itseditions (see [6], Chaps. 9 and 9*, and [7], pp. 187–219; 347–364), Sen pays aparticular attention to the Theory of Justice as Fairness of one of his former teachers,Rawls [9, 21–24].

This theory influenced deeply not only Sen’s theory, but the political philosophyof XX cent. overall in US, in the sense of a “moderate liberalism”, if not in the senseof a “social-democratic liberalism”, even though such labels are not able to do justiceof the relevance, the richness and the originality of this theory.

Indeed, Rawls theory of “justice as fairness”, because it introduces a fundamentalcriterion of “equity” or “fairness”—that is, of interpersonal comparison of welfare—as constitutive of the idea of justice, establishes also a substantial point of distinctionas to the Kantian formalist theory of justice. Even though Rawls wants explicitly tooffer by his theory an updated version of Kant’s ethics in the direction of a theoryof distributive justice, without losing, however, the unconditional character of themoral principles characterizing Kant’s theory.

Not casually, indeed, in his systematic treatment of Rawls’ theory in both editionsof his book on Collective Choice (see, for instance, [7], pp. 187–209), Sen startsprecisely from the quotation of the fundamental statement of I. Kant’s Grundlegungzur Metaphysik der Sitten (1785) [25], synthesizing his formal principle of “dutyfor duty” as the unconditional basis of the “moral oughtness”, as distinct from the

104 G. Basti et al.

unconditional “logical necessity”. Namely, this principle reads: “act always on sucha maxim as thou canst at the same time will to be a universal law”.5 As we know, inthis way Kant wanted to satisfy the so-called “Hume’s law” distinguishing betweenlogical necessity and ethical oughtness—sometimes denoted also as the reciprocalirreducibility facts/values—in philosophical (modal) logic. At the same time, how-ever, he wanted to preserve the universality of moral laws by granting them the sameunconditional or absolute character of the logical laws. That is, by granting to morallaws the same tautological character of logical laws by the principle of the “duty forduty”, without confusing them. Historically, in this way, Kant wanted to criticize theAristotelian foundation of justice on equity made proper by the Scholastic Tradition,which seems to imply a factual comparison of what is effectively right for differentindividuals in similar circumstances.

Not casually the modern re-proposals of the theory of equity or fairness beforeRawls, as Sen recalls, tried systematically to reconcile the notion of equity with theabsolute universality of the Kantian moral law, by making irrelevant the personaldifferences. On this regard, Sen quotes Henry Sigdwick’s approach that is paradig-matic in this sense because perfectly compatible with Kant’s principle (see, [7], pp.187–188).

…Whatever action any of us judges to be right for himself, he implicitly judges to be rightfor all similar persons in similar circumstances. Or, as we may otherwise put it, if a kind ofconduct that is right (or wrong) for me is not right (or wrong) for someone else, it must beon the ground of some difference between the two cases, other than the fact that I and heare different persons. A corresponding position may be stated with equal truth in respect ofwhat ought to be done to – and not by – different individuals ([27], p. 374).

In the formalization of Sigdwick’s theory of equity firstly proposed by Hare [28, 29]the universality of moral judgements consists ultimately in inserting their invariancefor permutation of persons within the similar circumstances recalled by Sigdwick,everything else remaining the same. Of course, Sen notices, there was a lot of crit-icisms to Hare’s modelling of equity theory as a formal basis of a social choicetheory, because it requires that a moral judgement be held from every position thata person can occupy in a society through interpersonal permutations. The referenceto Sidgwick’s and Hare’s theory of equity is useful to Sen, however, for introducingthe originality of Rawls’ theory about justice as equity, for which no interpersonalpermutation of positions is required since for Rawls “the principles of justice arethose which would be accepted in a fair situation in the original position” ([7], p.192).

In a paper of 1958 that anticipates his famous book of 1971, but it is often ig-nored by the scholars, Rawls defined, indeed, the original position as “an initialposition of equal liberty (which is defined by the pattern of rights and duties, pow-ers and liabilities, established by a practice)” ([21], p. 166), from which any theoryon the principles of justice must start. In Sen’s words, “the principles of justiceare those which would be accepted in a fair situation in the original position” ([7],p.192). Effectively, this position is for Rawls a purely hypothetical and a-historical

5Quoted in [7], p. 187, from [26], p. 66).


situation—and in this sense it is not comparable with J. J. Rousseau’s primordial“state of nature” of humans. It is “hypothetical” in the sense that the principles tobe derived from the original position are what the individuals would agree to, undergiven legitimating conditions, not what they have agreed to—and in this sense it isnot comparable with any social contract theory. It is “a-historical” in the sense thatit is not supposed that such an agreement has ever been or could ever have been. Inother terms, the “original position” is where individuals choose principles in a stateof basal equality without any reference to their own actual place in different socialstates (see [9], p. 23). In this situation, these principles would satisfy the criterion of“fairness”, being the result of a fair agreement with no vested interests, also becauseof a “voluntary veil of ignorance” of individuals about their actual personal differ-ences. In Saul Kripke’s modal logic jargon of “possible worlds” semantics appliedto deontic logic (see Sect. 4), the Rawlsian original position is therefore an “ideallygood possible world” that never can become an “actual world”.

To sumup, themain twoprinciples of distributive justice deriving from the originalposition are for Rawls the following ([9], p. 266):

First Principle: Each person is to have an equal right to the most extensive total system ofequal basic liberties compatible with a similar system of liberty for all. Second Principle:Social and economic inequalities are to be arranged so that they are both: (a) to the greatestbenefit of the least advantaged, consistent with the just savings principle, and (b) attachedto offices and positions open to all under conditions of fair equality of opportunity.

These two principles are interpreted by Sen as constituting the information basisof Rawls theory of justice. That is, respectively, as determining the basal space ofthe main variables characterizing the theory (First Principle), and as determiningthe “aggregation criteria” of these variables (Second Principle), mainly the maximinprinciple (a).

In this framework, according to Sen ([7], p. 348ff.), the basal space ofRawls theoryis constituted essentially by the “basic liberties”, but in reference to the “primarygoods” that each individual as citizenmust hold, and this is afirst point inwhichRawlsdifferentiates himself from the libertarianism. Indeed, Sen emphasizes, the range ofthese liberties is narrower than in libertarianism—e.g., it does not include propertyrights or rights of exchange or of bequeathal. Primary goods are, in fact, general-purpose resources that are useful for the pursuit of objectives that the individualmay have so that, in Rawls words, they are things that citizens need as free andequal persons, and claims to these goods are counted as appropriate claims ([30],p. 257). “Primary goods” are then “things that every rational man is presumed towant”, and include, without any pretension of exhaustiveness, beside the “basicliberties”, “income andwealth”, “freedomofmovement” and “choice of occupation”,“equality of opportunities”, etc. The reference to human rationality as characterizingall human individuals, makes evident that, inmoral philosophy, Rawls primary goodscorrespond to moral goals, which are proper to humans as such, and then that eachindividual universally (naturally) wants, so that each person has the right to pursuethem in whichever society. Namely, they correspond to what in political and lawphilosophy are defined as “human rights” that naturally must have their place in

106 G. Basti et al.

any social choice theory (see [7], pp. 420–452). On the other hand, the fundamentalaggregation principle of Rawls theory is, according to Sen ([7], p. 350), themaximinprinciple. Indeed, Sen emphasizes ([7], pp. 193–194), the component (a) of theSecond Principle of justice in the “original position we quoted before and by Rawlshimself defined as theDifference Principle, because it requires that social inequalitiesbe arranged to make the worst-off best-off, is essentially a maximin criterion. In fact,it requires that the minimal element in the set of individual welfare’s be maximized.Now, despite Rawls main focus is on the theory of type of institutions to be chosenin order to satisfy at best his two principles of justice, and not to propose a socialchoice theory, nevertheless the maximin principle could be used in principle alsoto order social states based on individual orderings, and so within a social choicetheory framework. For instance, for any social state, we may order the individualsin terms of their welfare and pick on the worst-off individual. This can be comparedwith the welfare of the worst-off individual in another social state, and so on, so tohave recursively, at least in principle, a complete social ordering ([7], p. 193).6

Nevertheless, Sen rightly emphasizes, the maximin principle interpreted as a sortof choice procedure in a social choice theory is not aSWF in the sense ofArrows, sincethe maximin procedure imposes that a worst-off individual i in a social state x goesup for every alternative so to change his situation. This means that the social orderinginvolving x, being based on a different individual welfare’s, becomes different. Andthis is incompatible with any SWF (see [7], p. 193). Neither the maximin procedureis aCollective Choice Rule (CCR), since a CCR specifies a social preference relationbased on the set of individual orderings of social states, each from one’s own pointof view. In other terms, a CCR is based on n orderings of m elements, whereas aRawlsian maximin choice mechanism is based on one ordering of mn elements (see[7], p. 194).

At this pointmanyproblems raise in the possibility of interpretingRawls’maximinprinciple as an effective decision rule in a social choice theory—as unfortunatelyit has been done too many times by policy makers, and in recent times, by theso-called “populist movements”. Indeed, if it is absolutized as a decision rule insocial choices, the maximin criterion becomes easily a factor of social injustice andeconomical regression, given that it oversimplifies complex situations. And in fact, itis often accompanied by misinformation campaigns on the media, which are, in fact,a perverse way of implementing the Rawlsian principle of the “veil of ignorance”over the real causes of the actual differences, making this veil not “voluntary” but“imposed”.

Anyway, some of the main limitations of using the Rawlsian maximin principleas a decision rule in a social choice theory are summarized by Sen in ([7], pp.195–198). Ultimately, they reduce themselves to the fact that Rawls criterion offairness, as far as referring to a hypothetical original position that, for definition, isirrespective of the historical, actual personal differences (covered by a voluntary veilof ignorance), becomes a too severe condition - in some proper sense, “ascetical”,

6Effectively, Sen offered in [6], since its first printing in 1970, a lexicographic formal version ofthe maximin principle that Rawls himself included in his book A Theory of Justice of 1971 [9].


because this requires that individuals renounce voluntarily to take into account theirpersonal differences, for being used in a theory of decisions about effectively fairsocial choices.

Indeed, we repeat, it risks becoming a false strategy of curing the symptoms ofan illness without removing the causes—or because there is the precise will of notremoving them, as it is the case of the social media manipulators behind populistmovements. Nevertheless, it remains true, as Sen emphasizes, that in several institu-tional questions that is the proper focus of Rawls theory of justice, the appeal of theRawlsian maximin criterion is well demonstrated and it works. Precisely becauseit cures one—if not the principal one—cause of a social disease: the existence in asociety of some unjust institutions, because not aimed, each in its proper way (e.g., asan educational, or as an occupational, or as a political institution, etc.) at maximizingthe social states of worst-off individuals.

However, what Sen notices is that the maximin principle suggested by Rawlscan be used also independently of his theory of justice, in a social choice theory.Particularly, in a theory of justice as equity like that implemented in Sens SCFtheory (see Sect. 3.4). It, indeed, starts from the actual personal—also subjectiveand not only objective—differences among individuals, the situations of manifestinjustice or iniquity, without veiling them, before all. That is, it starts from therelevance in a theory of justice as equity of the subjective differences (including theethical and cultural differences and preferences, the religious beliefs, or even thesimple personal tastes) in the interpersonal comparison of welfare states, becausedetermining concretely different subjective ways and grades of enjoying what are,objectively, the same commodities, and/or of profiting by what are objectively thesame utilities.

Indeed, precisely such differences among individuals and groups in a societymakeuntenable the Pareto unanimity axiom, we discussed in Sect. 2.3, so to transformArrow’s “impossibility theorem” in Sen’s “impossibility of the Pareto liberal” [8],because incompatible with the definition of equitable social rankings. Indeed, to sumup the example discussed at length by Sen in ([7], pp. 207–209), if the increasingof GDP in a given nation, e.g., in India, depended on the growth of production ofbeef, the impact would be different over Muslims and Christians welfare states, withrespect to Indus one. Just as, on the contrary, it would happen for the welfare states ofIndus and Christians, as to Muslims one, if GDP increasing depended on the growthof production of pork.

In a word, in Sen’s theory of justice as equity, the interpersonal comparison ofstates and preferences for defining a common ground of equality, on which applyingthe maximin criterion, is ethically based not on a voluntary “veil of ignorance” overthe differences, but on the “ethical “principle of extended sympathy in the form ofplacing oneself in the position of another” ([7], p. 290]). The starting point is, inthis case, not the usual individual ordering Ri , but an extended individual orderingRi, j of individuals i, j, and states x, y, one for each individual: (x, i), (y, j), wherei = j . That is, the “extended ordering” is an ordering where the subjects are notmere individuals i, only numerically distinct, but persons i.e., different individuals.

108 G. Basti et al.

Starting from this extended ordering, the principle of “extended sympathy” thatSen borrowed from Adam Smith, as we see, is formalized by our Author within hisSCF theory in terms of two axioms of identity. The first one reads: “each individualj in placing himself in the position of person i, takes on the tastes and preferencesof i ([7], p. 214). This ethical subjective principle can be objectively extended to thesecond axiom of “complete identity”, so to give a formal criterion of grading justicefor the “interpersonal comparison and permutation” among social states of differentpersons in a SCF theory.

That is, the first axiom can be extended to the identification between differentpreference rankings J of the social states, x, y, for any pair of different persons i, j ,i.e., x Ji y = x Jj y, where x Ji y reads: “x is more just than y according to person i”([7], p. 210).

What is fundamental for the logic of Sen’s SCF theory, is that, while each pref-erence ranking Ji in P. Suppes’ theory of “grading justice” [31], from which Senderived his own, is a strict partial ordering over the set {X} of possible social states,if we insert the possibility of making equivalent two or more of preference rankingsoriginally different, each Ji becomes a “simple” partial ordering7 Oi , which is suf-ficient for extending the Rawlsian maximin principle, interpreted as an aggregationprinciple, to Sen’s SCF Theory (on this essential point, see Chap. 9* of both [6, 7]).

In a word, what in Rawls justice theory is a hypothetical condition of fairnessamong individuals, hiding under a voluntary veil of ignorance the objective andsubjective actual differences among persons, in Sen theory such differences are thestarting point. In thisway, fair judgements and solutions of unjust inequalities becomethe final point of a process of interpersonal comparability and information exchange,in the limit leading to an ethos of dialogue among different persons and groups, in amulti-cultural society.

Complementarily, for any assumption of real-valued (statistical) measurabilityand interpersonal comparability of personal welfare states, in a SCF theory basedonto Sen’s theory of grading justice as interpersonal equity, the axiom of completeidentity becomes operationally an equivalence between vector spaces V, V ′ of ⟨m|n⟩alternatives. That is, of m states for n individuals for two different persons i, j , ontowhich calculating the statistical expectations in a quantitative theory of real-valuedwelfare functions (see [7], pp. 210–220 for the formal treatment of these notions).8

7We recall here what we illustrated in Sect. 2.3. Namely, that a “strict” partial ordering is an order-ing R(≤) among sets, in which only the asymmetric case ((x Ry = yRx) → x = y) holds of theantisymmetric relation that, on the contrary, holds for the partial ordering as such. That is, wherealso the symmetric case, ((x Ry = yRx) → x = y), holds.8The usage of the ⟨bra|ket⟩ notation for the matrix representation of Sen’s doubled vector states—a notation made famous by the quantum formalism, even though not limited to it—is an implicitsuggestion that the more effective approach to make computationally effective the two axioms ofidentity of Sen’s SCF theory is the usage of the quantum state superposition principle. In the frame-work, however, of the doubling of the degrees of freedom formalism for calculating dynamicallythe statistical expectations for entangled interacting systems, typical of the “many-body physics”approach to QFT. That is the fundamental physics of condensed matter physics (see Sect. 4).


To conclude this section, let us discuss briefly what is he proper of Sen’s theoryof justice as equity among different persons, characterized by different capabilitiesand functionings, intrinsically related, objectively, to different social contexts andrelated opportunities, and, subjectively, to different goals and value systems, as afundamental ethical component of Sen’s SCF theory.

3.4 Sen’s Theory of Social Choice Function Based on HisComparative Theory of Distributive Justice

Sen’s theory of personal “capabilities and functionings”. What characterizesSen’s theory of justice [12, 13] and of his many followers (see, for instance [31–34])as to the classical liberal and utilitarian theories, on the one side, and as to Rawlstheory of justice as fairness, on the other side, is synthesized by Sen himself in thefollowing passage

In recent years there has been considerable discussion on an approach to justice that con-centrates on peoples capability to lead the kind of life they have reason to value - the thingsthat they can do, or they can be. The roots of the approach can be traced to the ideas ofAristotle, and, to some extent, Adam Smith; it concentrates on the opportunities that peoplehave to lead valuable and valued lives. Aristotle saw this achievement in terms of “humanflourishing”. Among other things, he pointed out, in Nicomachean Ethics, that wealth isevidently not the good we are seeking for it is merely useful for something else ([7], p. 356).

In other terms, the core of Sen’s theory of justice as to liberal and utilitarian theoriesconsists in considering commodities and utilities as means for reaching those goodsor “valuable ways of being and of doing”, which a person—on the basis of her/hissystem of values—considers as fundamental for living “a valuable and valued life”,that is, a personal achieved life.

A personal achieved life can be seen as a combination of functionings (i.e., of doings andbeings), and, taken together, can be the basis for assessing that persons quality of life. Thefunctionings on which human flourishing depends include such elementary things as beingalive, being well-nourished and in good health, moving about freely, and so on. It can alsoinclude more complex functionings, such as having self-respect and respect of others, andtaking part in the life of the community (including appearing in public without shame), onwhichAdamSmith in particular presented an extraordinarily insightful analysis in hisWealthof Nations ([7], p. 357).9

Therefore, the combination of the different “functionings”, i.e., of different thingsto do, and/or of different ways of behaving and of being, constitutes the focus ofpersons life, and the extent of their concrete achievements represent the degree ofeach person self-achievement. On the other hand, persons capabilities consist

9Apart from the reference to Smiths masterpiece, Sen quotes in note another less famous book ofA. Smith, The Theory of Moral Sentiments, of which Sen himself cured a re-edition with his ownPreface [35].

110 G. Basti et al.

in the set of combinations of functionings from which the person can choose any one com-bination. Thus, the capability set stands for the actual freedom of choice a person has overthe alternative lives that he or she can lead ([7], p. 357).

To sumup, the “basal space” of Sen’s theory of justice is characteristically constitutedby “what people are able to be or are able to do”, rather than in terms of the means orresources they possess. In this way, for the construction of a suitable “social choicetheory based on these personalistic principles,

individual claims are to be assessed neither just by the incomes, resources or primary goodspeople respectively have, nor only with reference to the pleasures or utilities they enjoy, butin terms of the freedom they have to choose between different ways of living they can havereason to value ([7], pp. 357–358).

Then, in a social choice theory, of which information basal space is given in termsof “capabilities and functionings”, the following question becomes fundamental.Which is the aggregation principle of the different variables and of the differentinterpersonal comparisons, able to take note of variations of individual conditionsand circumstances, in which even the underlying personal values and preferencesplay an essential role?

Aswe anticipated, as far as we introduce in Rawls’maximin aggregation principlethe two axioms of extended identity among different welfare rankings, this principlecan be extended as well to Sen’s SCF theory, as far as we can suppose a suitablereal-valued quantitative parametrization of the welfare aggregates involved (see [7],pp. 210–220 for a formal treatment).

Indeed, by these axioms we can formalize, on the one side, the ethical principleof the interpersonal/intercultural dialogue as “extended sympathy” among differentpersons—that is, the ability of posing him/herself in the situation of the other one,taking on also his/her personal and ethical preferences and tastes. On the other side,by the same axioms, we can formalize the formation/comparability of homogenousgroups of persons in a heterogeneous society, because sharing the same set of personalconditions/preferences. In both case, indeed, the formal condition consists in thepossibility that there exists in our modeling an one-to-one correspondence fromthe set of persons H to H itself, such that (the objective/subjective situation of) aperson j can be mapped onto (the objective/subjective situation of) a person k, thatis, k = ρ( j), despite originally k = j .

Of course, Sen’s theory of SCF depending on a theory of justice as equity, whosebasal information space is given in terms of capabilities and functionings, posessignificant logical and then computational problems for their modeling.


4 Conclusion: The Computational Challenge of Sen’sTheory

4.1 The Computational Challenge: The Dynamic Weighingof Variables

Sen well synthesizes the deep problems related with variable weighing in statisticalmodeling of his SCF theory in the two following passages of his book ([7], pp.358–359).

The heterogeneity of the components in the basal space, such as different functionings, pointsinevitably to the need to weigh them against one another. This applies to all approaches thatrespect plurality in one form or another, including the Rawlsian focus on primary goods, orthe Aristotelian focus on functionings and capabilities (which is present also in other theoriesthat take note of different aspects of the quality of life).

For instance, we can “close our eyes” by simply assuming that there exists somethinghomogenous like “the income by which everyone’s advantages can be judged andimpersonally compared, but this cannot solve the problem rather it avoids it. Indeed,

real income comparisons involve aggregation over different commodities, and in judgingcomparative individual advantages there is the further problem of interpersonal comparisons,taking note of variations of individual conditions and circumstances. It is, of course, possibleto reflect these variations in values of “adjusted income” that can be appropriately defined,but that is only another way of stating the same problem, requiring that attention be paid tothe valuation of heterogeneous factors, though expressed in the indirect space of equivalentincomes. Measurements in the direct space (e.g. quality of life, or capability indicators) andthose in the indirect space (e.g. equivalent incomes) would have a tight correspondence witheach other, related to the underlying values on which both the normative exercises are based.One way or another, the issue of valuation and weighing has to be faced.

Of course, all this means that the usual mathematical and computational tools ofstatistical analysis used in social and economic sciences appear inadequate, as far asthere are based on the necessity of supposing a fixed and then unique set of weightsfor evaluating the variables. Indeed, the problem depends on the high variability ofthe circumstances that make in principle impossible to arriving at the definition of aunique and fixed set of weights, because of the consequent ever changing characterof the inner higher order correlations in the data set. The same problem that computerscientists are faced when dealing with data streams. It is therefore not casual that themain criticism to Sens SCF theory is that

it is not practically useable because of the weighing problem. In fact, however, diversityof ingredients is only an invitation to address the problem of relative values and weights,not an admonition to resign and go home. Indeed, when the valuation of inescapably di-verse concerns is done implicitly as in the measurement of GDP through the use of marketprices for distinct commodities we simply withhold our responsibility to evaluate, goinginstead for the mechanical use of some relative values that may have been fixed without anyrelevance to normative assessment (market prices can be hugely inappropriate for ethicalevaluations, when there are externalities, asymmetric information and big inequalities inincome distribution) ([7], p. 369).

112 G. Basti et al.

On the other hand, Sen emphasizes another fundamental factor making simply ab-surd the argument sometimes proposed that the solution for making usable and evenoperational the capability approach only if it comes with a set of given weights ondifferent functionings in some fixed list of relevant capabilities—like for instancein Martha C. Nussbaum’s approach (see ch. 2 of [32]). This factor is strictly re-lated with our communication age in which valuation and weighing are in real timesystematically influenced by the public discussion and scrutiny on it

This need not, however, derail evaluation of public policy, or disable welfare-economic eval-uations. The capability approach is entirely consistent with reliance on partial rankings andlimited agreements. The main task is to get the weights – or ranges of weights – appropriatefor the comparative judgements that can be reached through reasoning, and if the result is apartial ranking, then we can make precisely those judgements that a partial ranking allows.There is no obligation to feel compelled to opine on every comparative assessment that canbe proposed ([7], p. 369. Italics are ours).

4.2 The Computational Challenge: The Necessity of AISupport and the Issue of Dynamic Weighing in MachineLearning

It is precisely at this point that the relevance of the AI support to social choice the-ory becomes evident. Indeed, if the problem mainly consists in getting “the weightsappropriate for the comparative judgements that can be reached through reasoning”,because of the complexity of these judgements and the necessity to be continuouslyupdated often in real time, the support of AI to the human reasoning becomes essen-tial.

On this regard, for concluding this contribution, it is essential to discuss brieflythe core of the actual discussion about the differentmachine learningmodels10 in AIsystems because their strict relationship with the problem of the statistical variable

10By machine learning computer scientists generally intend the ability of a machine of simulat-inghuman brain ability of learning from its past experience without any further interven-tion of theprogrammer. Namely, in classical or symbolic AI computational systems such as those based onM.Minsky’s frame theory [36]—which is at the basis of contemporary “object-oriented programming”technique—, when the machine encounters a new situation it selects from memory a given data-structure or frame for representing a stereotyped situation, whose details can be adapted to fitreality. Effectively, a frame is a network of nodes and relations, whose top levels are fixed becauserepresenting what is supposed to be always true (effectively, true with the highest probability) ina given situation, and lower levels have many terminals or slots to be filled with specific instancesor data, according to specified conditions its assignments must meet. Of course, “frames” areprogrammed data-structures. Machine learning using ANN deep learning wants precisely to avoidasmuchas possible this humanprogrammer intervention in the definitionof thememorized referencepatterns. In this sense, it is at the basis of the so-called non-symbolic approach to AI, also becausein the case of the “big data” this human programmer intervention is effectively impossible, becauseof the complexity of the data structures involved, characterized by millions and ever billions ofparameters to be taken into account.


weighing that emerges therefore as the true challenge both for the natural and theartificial reasoning in any realm of contemporary sciences.

Effectively, the actual state of art of machine learning in AI systems is strictlyrelated with the different deep learning algorithms inspired by the so-called artificialneural networks (ANN) approach in the so-called non-symbolic AI, necessary fordealing with large sets of statistical data and of their inner and variable correlations(see Footnote 11), as it is the case of our social choice theories.

Without entering here into the technicalities of ANN approach in machine learn-ing, and despite the demonstrated utility of the different and more evolved learningalgorithms in solving specific problems with “big data” (see [37], for an updatedreview), it is significant for us for understanding why the actual deep learning ap-proach in AI-decisional support for human reasoning is certainly useful but not yetsufficient in principle for addressing to the problems emphasized by Sen of a truedynamic weighing of the variables in social choice theory—and not only in it.

Indeed, all the actual deep learning models in AI systems are based on a statistical(not dynamic) approach to machine learning. This is essentially a modeling in whichthere exists a sharp division between the “learning phase” onto some representativesample of the database for fixing the proper set of weights representing the higherorder correlations among the variables in the dataset, and the “testing phase” forapplying fruitfully the learned weights configuration onto another representativesample of the database, and finally, if the test is positive, onto the whole database.Therefore, it is evident that this approach cannot work when we have to deal inmachine learning not only with huge, even though static bases of data, but withdynamic streams of data, in which there exists an indefinite and unpredictable—because continuously changing—number of possible correlations among an anywayfinite set of data.

That is, we are faced again with the infinity challenge in computational systems.This means that such an approach cannot work in the actual social and economiccontexts, in which the fast information exchange among the social and economicactors distributed all-over the world determines the sudden formation/destructionof long-range correlations among their behaviors. In such cases, like, on the otherside, in Sen’s modeling of SCFs in which economic and social actors are directlyinvolved in the process to be modeled, the weights cannot be fixed in any way, as Senemphasized many times. In other terms, in these cases we need ANNs in continuouslearning, just it happens in biological neural networks inwhich no absolute separationbetween the learning and testing phases occurs. The brain neurons, constituting thehidden layers of biological NNs, indeed, not only are always oscillating, otherwisetheir EEG would become flat, but the EEG traces in healthy brains are characterizedby the highest variability, without any recurrent periodicity, because a living brain isphysically a dissipative or open system in continuous exchange with its environment.

Unfortunately, for the actual statistical approach to ANNs that are based on statis-tical mechanics and then must model ANNs as closed systems non interacting withtheir environments, “continuous learning” means that the non-linear behavior of thenet becomes immediately chaotic and then in principle intractable for whicheverstatistical learning algorithm such as, for instance, the well-known gradient descent

114 G. Basti et al.

algorithm that, with several updating and corrections, is the actual standard paradigmin AI machine learning, since 80s of the last century [37–39].

According to this approach, the deep learning in themany layers of the hidden neu-rons between the input and the output layers of the net, and whose dynamics is gov-erned by a non-linear sigmoid function σ ,11 consists in an optimization process by thestochastic gradient descent algorithm. That is, it consists in the calculus of the mini-mum of an error-function (loss function), reducing recursively the distance betweenthe actual and the desired output of the net. This happens by “back-propagating”recursively the error for redefining the statistical weight configuration of the connec-tions between the input and the hidden neurons of the net, till the absolute minimumof the error function is reached [39]. This means that the back-propagation learningalgorithm is a supervised learning algorithm since the optimization process supposesthe existence of a learning target to be reached, depending on a teacher, effectivelythe human programmer.

The separation between the learning and the testing phases in this statistical stan-dard paradigm of ANNs depends then theoretically on the necessity of definingthrough the learning phase one only fixed set of neuron weights, and therefore oneonly “energy landscape” for the dynamics of the net, because determining one only“potential function” of the error to be minimized. This, therefore, requires that thesystem be conceived as stable at equilibrium, and then as isolated from any otherfurther information exchange with its environment, otherwise the dynamics wouldlose its “structural stability”. Namely, its weights would be continuously changing

11The non-linear sigmoid function σ (x) = 11+e−x , and overall its immediate relative the hy-

perbolic tangent function tanh x =%ex−e−x

ex+e−x

&= sinh x

cosh x are what characterizes the learning algo-rithms of the actual ANNs, instead of the linear step-wise 1/0 activation function of Rosenblatt’searly “linear perceptron” [40]. This activation function is, effectively, the Heaviside function:H(x) = d

dxmax {x, 0} for x = 0, whose values are obviously “zero” for negative arguments and“one” for positive arguments, so to make linear—namely, a summation of the neuron weights eachrepresenting a variable of the dataset—the activation function of the “inner” or “deep” neuronsbetween the “input” and “output” and neurons. To sum up, by using the sigmoid as the activationfunction, the overall neuron output is no longer 0/1, but any real value between 0 and 1, so toallow to calculate the probability that some event (output) happens, given some conditions (input),instead of the simple no/yes answer of the linear ANNs. The non-linearity is then fundamentalfor calculating higher-order or long-range correlations among data. Indeed, mathematically, theTaylor series expansion of tanh includes several higher order correlations, much more than sinhand cosh alone, of which tanh is the ratio, so to answer in principle to the main criticism made byM. Minsky and S. Papert at MIT [41] to the early linear ANNs structures, such as McCulloch’sand Pitt’s neural net [42] and Rosenblatt’s linear perceptron [40], who first demonstrated that aneural net is able to calculate in a desirable parallel way any Boolean logical function like a TuringMachine. Effectively, because of the prestige of M.Minsky in US computer science community, hiscriticism was able to block any further research in ANNs, from 60s to 80s of the last century. In 80sa non-linear multilayer perceptron with a sigmoid as activation function for the hidden or “deep”layer neurons [39], and then characterized by the gradient descent learning algorithm [38], answeredMinsky’s main criticism to early linear ANNs. However, only during the last twenty years, with thelarge availability of ever more powerful processors—essentially arrays of GPUs—for reckoningwith the computational weight of this type of ANN architectures, we have the actual developmentof AI-systems based on deep machine learning algorithms (see [37] for an updated synthesis).


like its overall error energy landscape. Now, this non-stationary character of a dy-namics is precisely what characterizes the notion of deterministic chaos in the theoryof non-linear dynamic systems, ANNs included [43, 44].

This finally explains because these statistical models of machine learning mightwork well with big data in static bases of data, but not in dynamic streaming of data.Moreover, the all-with-all connectivity between input and hidden neurons explains,on the one side, why this computational architecture effectively does not answer—theoretically—the other main objection of Minsky against ANNs. Namely that theyare not properly parallel computation architectures (for being really so, differentsubsets of the hidden neurons would have to be linked only with subsets of theinput neurons and not all-with-all, so to compute different parts of the complex inputsimultaneously). On the other hand, this total connectivity explains why this typeof machine learning algorithm are so computing-power (but also electric-power, ofcourse) consuming. Indeed, with the full connectivity, the number of the possiblecombinations to be computed grows exponentially with n!, where n is the number ofneurons of the net.

This explains why this machine learning statistical approach spread effectivelyonly during the last twenty years, when computationally powerful GPUs becameavailable at chip prices on the market. Even though this approach became known tothe wide public only few years ago, effectively after the successful application byHinton in 2012 [45] to the ImageNet database of millions of images. An applicationthat earned him the front page on the New York Times. On the other hand, this “com-putational weight” of deep learning algorithms explains why the future of machinelearning in AI is strictly related with the development of muchmore computationallypowerful quantum computers. It is therefore highly significant that the dynamics ofthe biological neural nets of our living brains is chaotic [46]. Living brains are evi-dently open systems in “continuous learning”, because in continuous exchange withtheir environments—otherwise they would be dead brains in containers fulfilled withformalin of some neuroanatomy lab. However, if this chaotic character make in prin-ciple despairing any attempt of a statistical approach to NNs in continuous learning,this opens the way to a dynamic modeling of this type of deep learning, effectivelyalready developed for biological NNs, but that can be applied also to ANNs. Indeed,the connection with our previous discussion is that such a chaotic dynamics is thedemonstrated macroscopic manifestation of an underlying microscopic dynamicsof the many-body field theory of dissipative quantum systems, characterizing thefundamental physics of the condensed matter, biological systems included [47, 48].

Moreover, the reference to dissipative QFT of many-body interacting systems weanticipated many times in this paper as the proper physical paradigm to which alsoSen’s social theory of the many interacting groups in society must refer, in the caseof biological NNs, has another proper justification. It is, indeed, the only possibleexplanation of the long-range correlation waves proper of the oscillatory behaviorof mammalian neocortex, as well as of the mammalian thalamocortical interactionsduring the action-perception cycles characterizing their intentional behavior. This,on the one hand, suggests an intriguing hypothesis about a dynamic mechanism ofmultilayered long-term memory constitution (“dynamic deep learning” in quantum

116 G. Basti et al.

computing) in the mammalian brains, using the powerful QFT principle of the QV-foliation [47, 48]. On the other hand, the operator algebra formalism underlyingthis model [49] suggests the development of a new architecture of quantum opticalcomputer based on dissipative QFT in physics, and on Category Theory in logic[11, 50], by which approaching from a new dynamical and no longer statisticalpoint of view the challenges of an unsupervised machine learning in AI [11, 51], fordealing with data streaming and with dynamic bases of data, both characterized bythe sudden formation-dissolution of long-range correlations (field phase coherences)among data.

Effectively, the same G. Hinton more recently stated that the future of machinelearning in AI is related with the development of unsupervised learning algorithmsin ANNs. In an interview with Steve Levin on September 15th, 2017 the interviewerso summarized Hintons new position:

In back-propagation labels or weights are used to represent a photo or voice within a brain-like neural layer. The weights are then adjusted and readjusted, layer by layer, until thenetwork can perform an intelligent function with the fewest possible errors. But Hintonsuggested that, to get to where neural networks are able to become intelligent on theirown, what is known as unsupervised learning “I suspect that means getting rid of back-propagation. I don’t think it’s how the brain works he said. “We clearly don’t need all thelabeled data.12

This reference of G. Hinton to the computational notion of labeling is particularlyprecious for understanding how brain works, according to the QFT modeling weillustrated in many papers [11, 47, 48]. Indeed, the notion of “label” pertains tothe elementary notions of theoretical computer science, where any automaton, theTuringMachine (TM) included, can bemodeled as a “labeled state transition system”(LTS).Thedifferentmodels inAImachine learning canbe thenusefully characterizedaccording to the different interpretations of the notion of “label”.

1. In the early classical or symbolic AI machine learning models, such as Minsky’sframe theory (see above Footnote 11), the label is an instruction of the program,more precisely a function of the logical Boolean calculus determining a compu-tational step or “state transition” in the machine, according to the TM symbolicparadigm in computer science.

2. In the non-symbolic but statistical AI machine learning of the ANNs, the labelis a given configuration of the statistical weights of the transition probabilitymatrix representing a specific ANN architecture, according to different neuronactivation functions. In the case of supervised models such as “back-propagation”and similar, for configuring the weights they use the “gradient descent algorithm”.As we know, these machine learning models, representing the actual state of theart in AI, are “supervised” because, even though the weight configuration doesnot depend directly on the programmer like in the statistical TMs, neverthelessthe target representation for calculating the right weight configuration by “back-propagating the error” is put by hand, i.e., it supposes a teacher. Of course, it

12For a more articulated criticism to “supervised” machine learning, see [52].


is possible to think also at statistical unsupervised models of ANNs, which areat the moment purely speculative because generally lacking a suitable selectioncriterion among weight configurations (see however Footnote 13 below).

3. In the non-symbolic but (thermo-)dynamic quantum models of AI machine learn-ing, the “label” is dynamically given by a given condensate of NG-bosons, eachidentifying univocally a long-range correlation among oscillating neural fieldsand their (inner/outer) environments, both in biological, and in artificial NNs [11,47, 48]. This “dynamic labeling” is observer-independent, i.e., absolutely unsu-pervised, and act as a “control parameter” of a given phase coherence domain,constituting physically in QFT computing, or socially in Sen’s theory (in bothcases, “coalgebraically”) the logical reference (coalgebraic semantics, includingby its “algebra doubling” (QFT) or “vector space doubling” (Sen) the system andits physical/social environment) making true the relative proposition of a Booleanalgebra calculus. This class of biological and artificial NNs represent therefore aQFT instantiation of a “thermodynamic LTS”, where effectively the “states” aredifferent phases of matter in physical, biological and social realms, and/or dif-ferent controlled phases of interfering coherent light waves (photon condensates)in the two branches (the principal and the reference one) of an interferometer,in our quantum optics implementation in computer science of the same principle[11]. In this case, because we work with “open” systems, in far-from-equilibriumconditions and therefore we are outside Gibbs statistical mechanics paradigm,differently from QM, we can pass through different entangled phases (states),using each time the minimum free-energy function as a generalized (universal)dynamic selection criterion among states (phases), linking systematically free-energy dissipation (entropy) with information.13

Finally, for concluding this contribution, we want to recall the continuity of thisthermodynamic interpretation of machine learning with Lorenzo Magnanis recentcharacterization of deep learning processes in humans, with respect to the actualAI systems—and specifically with the “AlphaGo” program, taken as a testbed ex-ample [54]. This specificity consists in the continuous interaction of humans withtheir environments, defined as the eco-cognitive openness or the unlocked characterof human creative abductive processes, with respect to the locked character of theAlphaGo abductive strategies, and generally of the AI systems based onto statistical

13At this point, it is significant to emphasize that in US, Todd L. Hylton and his collaboratorsrecently launched the Thermodynamic Computer Project (see https://knowm.org/thermodynamic-computing/). For instance, the so-called “thermodynamic NNs” are trying to approach a type ofunsupervised solution inmachine learning using theminimum free energy (in Helmholtz formaliza-tion within the statistical mechanics approach) as a selection optimization criterion. They remain,however, in the paradigm of the statistical interpretation of thermodynamics, also for modelingneurons interacting with (their representation of the) environment, using an ingenious, but anthro-pomorphic logic of “ question/answer” [53]. In this way, the dynamics of such a “thermodynamicneural network” is a sort of stroboscopic approximation (i.e., limited to one phase or to a small num-ber of close phases) of what in QFT is a continuous process of phase transitions (macroscopically, achaotic dynamics) and then, when implemented computationally, a universal thermodynamic LTS.In this limited sense, this class of ANNs is straddling (2) and (3) types above of unsupervisedmachine learning algorithms.

118 G. Basti et al.

deep-learning algorithms. The dissipative QFT interpretation of mammalian and hu-man neural dynamics [46, 47], and the consequent quantum logical calculus on acoalgebraic basis we can implement in it [11] might be then considered as a model-ing of Magnanis eco-cognitive abductive processes also because they share the samesemiotic and then algebraic background in logic [15].

References

1. Hansson, S.O., Hendricks, V.F. (eds.): Introduction to Formal Philosophy. Springer, Berlin &New York (2018)

2. Endriss, U.: Logic and social choice theory. In: Gupta, A., Van Benthem, J. (eds.) Logic andPhilosophy Today, pp. 333–377. College Pubblications, London (2011)

3. Bergson, A.: A reformulation of certain aspects of welfare economics. Q. J. Econ. 52(2),310–334 (1938)

4. Samuelson, P.A.: Foundations of Economic Analysis. Harvard University Press, Cambridge,MA (1947)

5. Arrow, K.J.: Social Choice and Individual Values, 2nd edn. Yale University Press, New Haven& London (1963)

6. Sen, A.K.: Collective Choice and Social Welfare. Elsevier, Amsterdam (1979)7. Sen, A.K.: Collective Choice and Social Welfare, Expanded edn. Penguin Ltd., Kindle Edition,

London (2017)8. Sen, A.K.: The impossibility of a Paretian liberal. J. Polit. Econ. 78(1), 152–157 (1970)9. Rawls, J.: A Theory of Justice. Oxford University Press, Oxford, UK (1971)10. Stiglitz, J.E., Sen, A.K., Fitoussi, J.-P.: Mis-measuring our lives.WhyGDP doesn’t add up. The

report by the Commission on theMeasurement of Economic Performance and Social Progress.With a foreword by Nicolas Sarkozy. New Press, New York (2010)

11. Basti, G., Capolupo, A., Vitiello, G.: Quantum field theory and coalgebraic logic in theoreticalcomputer science. Prog. Biophys. Mol. Biol. 130, 39–52 (2017)

12. Sen, A.K.: The Idea of Justice. Penguin Book, London (2010)13. Sen, A.K.: Development as Freedom. Anchor Books, New York (2010)14. Tirole, J.: Economics for the Common Good. Princeton University Press, Princeton, NJ (2017)15. Basti, G.: The post-modern transcendental of language in science and philosophy. In: Delic,

Z. (ed.) Epistemology and Transformation of Knowledge in a Global Age, pp. 35–62. InTech,London (2017)

16. Mulgan, G.: Big Mind. How Collective Intelligence Can Change Our World. Princeton Uni-versity Press, Princeton, NJ (2018)

17. de Condorcet, M.: Essai sur lApplication de lAnalyse à la Probabilité des Decisions Renduesà la Pluralité des Voix. L’Imprimerie Royale, Paris (1785)

18. Kirman, A., Sonderman, D.: Arrow’s theorem, many agents, and invisible dictators. J. Econ.Theory 5(2), 267–277 (1972)

19. Mihara, H.R.: Arrow’s theorem and Turing computability. Econ. Theory 10(2), 257–276 (1997)20. Sen, A.K.: The informational basis of social choice. In: Maskin, E., Sen, A.K. (eds.) The Arrow

Impossibility Theorem, pp. 67–100. Columbia University Press, New York (2014)21. Rawls, J.: Jusitice as fairness. Philos. Rev. 67(2), 164–194 (1958)22. Rawls, J.: Justice as fairness: political not metaphysical. Philos. Public Aff. 14, 223–252 (1985)23. Rawls, J.: Political Liberalism. Columbia University Press, New York (1993)24. Rawls, J.: In: Kelly, E. (ed.) Justice as Fairness. A Restatement. The Belknap Press of Harvard

University Press, Cambridge, MA & London, UK (2001)25. Kant, I.: Grundlegung zur Metaphysik der Sitten. Johann Friedrich Hartnock, Riga (1785)26. Kant, I.: Fundamental Principles of theMetaphysics of Ethics (trans: Abbott, T.K.). Longmans,

London (1907)


27. Sidgwick, H.: The Methods of Ethics, vol. 3. McMillan, London (1874)28. Hare, R.M.: Freedom and Reason. Oxford University Press, Oxford, UK (1963)29. Hare, R.M.: The Language of Morals. Oxford University Press, Oxford, UK (1963)30. Rawls, J.: The priority of rights and the idea of good. Philos. Public Aff. 17(4), 251–276 (1988)31. Suppes, P.: Some formal models of grading principles. Synthese 16(3–4), 284–306 (1966)32. Nussbaum,M.C.: Human functioning and social justice: in defense ofAristotelian essentialism.

Polit. Theory 20(2), 202–246 (1992)33. Nussbaum, M.C.: Creating Capabilities. Harvard University Press, Cambridge, MA (2011)34. Alkire, S.: Valuing Freedoms: Sens Capability Approach and Poverty Reduction. Oxford Uni-

versity Press, New York (2002)35. Smith, A., : In: Sen, A.K. (ed.) The Theory ofMoral Sentiments, 1776. Penguin, London (2009)36. Minsky, M.: Frame theory. In: Johnson-Laird, P.N., Wason, P.C. (eds.) Thinking: Reasings in

Cognitive Science, pp. 355–376. Cambridge University Press, Cambridge, MA (1977)37. Nwankpa, C.E., Ijomah, W., Gachagan A., Marshall, S.: Activation Functions: Comparison

of Trends in Practice and Research for Deep Learning. https://arXiv.org/arXiv:1811.03378v1(2018, 8 Nov). Accessed 30 Jun 2019

38. Werbos, P.: Beyond regression: new tools for prediction and analysis in the behavioral sciences.Ph.D. thesis. Harvard University Press, Boston, MA (1974)

39. Rumelhart, D.E., Hinton, G.E., Williams, R.J.: Learning representations by back-propagatingerrors. Nature 323(6088), 533–536 (1986)

40. Rosenblatt, F.: Principles ofNeurodynamics. Perceptrons and the Theory ofBrainMechanisms.Cornell University Press, Buffalo, NY (1961)

41. Minsky, M., Papert, S.: Perceptrons. An Introduction to Computational Geometry, 2nd edn.MIT Press, Cambridge, MA (1987)

42. McCulloch, W.S., Pitts, W.H.: A logical calculus of ideas immanent in nervous activity. Bull.Math. Biophys. 5(4), 115–133 (1943)

43. Hilborn, R.: Chaos and Non-linear Dynamics. Oxford University Press, Oxford (1994)44. Viana, M.: What’s new on Lorenz strange attractors. Math. Intell. 22(3), 6–19 (2000)45. Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional

neural networks. In: Proceedings of the 25th International Conference on Neural InformationProcessing Systems, vol. 1, pp. 1097–1105. CurranAssociates Inc., Lake Tahoe, Nevada (2012)

46. Freeman, W.J.: How Brains Make Up Their Minds. Columbia University Press, New York(2001)

47. Freeman,W.J., Vitiello,G.:Dissipation and spontaneous symmetry breaking in brain dynamics.J. Phys. A: Math. Theor. 41(30), 304042 (2008)

48. Freeman, W., Vitiello, G.: Matter and mind are entangled in two streams of images guidingbehavior and informing the subject through awareness. Mind Matter 14(1), 7–24 (2016)

49. Blasone, M., Jizba, P., Vitiello, G.: Quantum field theory and its macroscopic manifestations.Ordered Patterns and Topological Defects. Imperial College Press, London, Boson Condensa-tion (2011)

50. Rasetti, M., Marletto, P.: Quantum physics, topology, formal languages, computation: a cate-gorial view–an homage to David Hilbert. Perspect. Sci. 22(1), 98–114 (2014)

51. Rasetti,M.,Merelli, E.: Topological field theory of data:mining data beyond complex networks.In: Contucci, P., Giardinà, C. (eds.) Advances in Disordered Systems, Random Processes andSome Applications, pp. 1–42. Cambridge University Press, Cambridge, UK (2017)

52. Marcus, G.: Deep learning: a critical appraisal. https://arxiv.org/ftp/arxiv/papers/1801/1801.00631.pdf. Accessed 15 Jul 2019

53. Fry, R.L.: Physical intelligence and thermodynamic computing. Entropy 19, 107 (2017)54. Magnani, L.: AlphaGo, locked strategies, and eco-cognitive openness. Philosophies 4, 8–24

(2019)55. Alkire, S., Foster, J.E., Seth, S., Santos,M.E.,Roche, J.M.,Ballon, P.:Multidimensional Poverty

Measurement and Analysis. Oxford University Press, Oxford, UK (2015)56. Pattanaik, P.K.: Social choice and voting. In: Hansson, S.O., Hendricks, V.F. (eds.) Introduction

to Formal Philosophy, pp. 839–853. Springer, Berlin & New York (2018)

Could an Electronic Person Exist?Robots and Personal Responsibility

Claudia Stancati and Giusy Gallo

Abstract The recent developments in robotics claim for a philosophical approachto the current and future relationship between persons and robots. Our aim is tolook for some hints in the facets of the debate in order to define the ontology ofthese entities which in the long run could make decisions or act autonomously. First,we will question which features human beings and artificial beings share, focusingon the concepts of ‘person’ and ‘legal personality’, investigating the philosophicalwork of thinkers as Kelsen, Scarpelli, Radin, Wolff and Nèkàm. The urgency of the(philosophical) matter is confirmed by the concept of “electronic person” proposed inthe motion that the Committee on Legal Affairs presented for a European Parliamentresolutionwith recommendations to theCommission onCivil LawRules onRoboticsin order to establish rules to arrange the relationship between robots and humanbeings.

Keywords Legal entity · Legal personality · Electronic person · Person · Robot ·Driverless car

The authors have equally contributed to the ideas and content of this article. Claudia Stancati isresponsible for Sects. 2 and 3. Giusy Gallo is responsible for Sects. 1, 4 and 5. The first versionof this text has been presented at AISB 2017 Convention and it has been published in AISB 2017Proceedings.

C. Stancati · G. Gallo (B)Department of Humanities, University of Calabria, Rende, Italye-mail: [email protected]

C. Stancatie-mail: [email protected]


121

122 C. Stancati and G. Gallo

1 Introduction

The term “electronic person” appeared for the first time in the European Parliamentresolutionwith recommendations to theCommission onCivil LawRules onRoboticsin 2017. What does “electronic person” mean and why it is relevant to the ethicaland legal discussion on the relationship between (human) persons and robots to cointhat term?

The urgent need of a political consideration marks the necessity to find an answerto the question about the features human beings and artificial beings should share.While it is a simple task to enucleate the features of artificial beings, we maintain thephilosophical urgency to begin to think the concept of person again. For this reason,the starting point is the meaning of the term persona in Roman law and the way theconcept has changed in Western law, including the various notions of legal person.The main theme in question is personal responsibility, not only for human actionsbut also for robotic ones, since robots—for example, social robots—have started toshare daily actions with human beings.

The European report mentioned above focused the attention on the developmentsin robotics paying attention to the preservation of human safety and dignity startingfrom the fact that juridical systems still today are not adequate to face situations inwhich a robot is “guilty”.

2 Electronic Personality: The Issue

In “Alienable rights” (1995), Minsky [1] plays out a dialogue between two aliensspeaking of the rights to be granted to the human beings. He wonders if we need toallow that the world is ‘pollute’ with ‘redundant intelligences’.

We are ever closer to themoment inwhichwemust seriously respond to the crucialquestion on Artificial Intelligence: which features human beings and artificial beingsshare? Recently, Floridi has expressed his concern about this point: “the issue isnot to decide whether the robots, as autonomous agents which are able to learn, aresuch a kind of person or not, but we should understand that the question is badlyformulated” [2].

Actually, the question is particularly complex not only becausewe are consideringthe humanbeing close to the artificial one, but because the term“person” is everythingexcept that defined in an univocal way. Our aim is to look for some hints in the facetsof the debate on this notion in order to define the ontology of these entities which inthe long run could make decisions or act autonomously. This is the symmetrical caseto the one which claims for the concepts of the extended mind and the electronicbody, which have been made actual by the new technologies and have reshaped theborders of the old philosophical debate of the mind-body problem [3].

The semantic adventure of the term ‘person’ led the “mask” to become what isessential: the persona ficta has become what is most essential of a subject.

Could an Electronic Person Exist? … 123

The concept of the person has now entered the common language and it definesevery individual in its unique peculiarity. Philosophically speaking, the concept of theperson has a long history which also crosses the religious tradition and particularlyChristianity; today this concept has landed at the center of the discussion in bioethics,in cognitive sciences, passing through sociology, from Marcel Mauss onwards [4].In this story, ‘individual’ and ‘person’ have been considered different concepts butonly in relatively recent times they have been opposed one another: the first one hasbeen managed such as a naturalistic category and the second one such as an ethicalconcept or a notion which can define a center of action.

The concept of the person, superimposed on the problem of subjectivity andstrictly tied to the theme of the values and their intention acceptance, is not clearin its ethical and epistemological implications. We would mention the differencesamong the notion of person forged by Peter Singer or Martha Nussbaum and the roleplayed by the ability to act, the biological nature and biography in defining thoseideas.

The different notions of the person refer to different ontological levels and becauseof them we would like to consider here a concept of person drawn in legal field inorder to verify whether this one, once opportunely adjusted, could be the right toolto define the status of robots and androids.

The crucial point of the question is to recollect a definition of the person suitable tothe electronic personalitywith reference to the attempt of replicating humanbehavior,emotion and way of thinking [5], facing the issue of the imputation of liability, sincethe future could lead us in a dimension of co-working with robots.

3 The Legal Definition of Person

In Roman law, the term ‘persona’ is equivalent to homo; then it evolves to an abstractmeaning to designate who has a juridical role. This is a tendency which lead to theconcept of the persona ficta or rapraesentata like the “fictum positum pro vero” [6,D-48, ii9, 16, 10].

A very long process of doctrinal elaboration, which is parallel to the history ofthe Western law, particularly but not only the Roman juridical tradition,1 lead todissociate the concept of the juridical person from the legal entity, the imputationof liability from the conferring of the legal personality. The last concept does notnecessarily refer anymore to the man but it’s linked to the powers and the interestsof the man, which are concentrated in accordance with predicative structures andnormative schemes, or it refers to complex subjects to whom have been conferredfeatures suitable to action both in the field of private law and public law.

This sort of persons is distinguished from natural elements, or rather the juridicalperson arises as opposed to the natural element as persona ficta, bothwith reference to

1The term ‘person’ is not present in the Code Napoléon, but one can find it in the common law, asit shows the Judicial Dictionary by Stroud [18].


a single individual and to a group of persons. Law breaks down bodies and subjects tocompose persons namely masks which could personam sustinere o gerere, personamvicem gerere. The same individual can be invested with different personae; manysubjects can represent one unique persona, whose unity is not superimposed to theunity of a physical or psychological subject. Moreover, in some subjects there can bethe rights in the absence of a dead physical or unborn person, and on the other handthe concept of physical person renames, for the purposes of the law, the biologicalbodies (e.g. let’s consider bioethics).

The notion of legal person adapts to the definition of the artificial subjects whichare allowed on a merely functional level and, then, to every artificial individual.

In his Reine Rechtslhere Hans Kelsen maintains that: “one might consider (…)not individuals such as, but only their actions and omissions set out in legal rules”[7] so that the physical person “is not a man but the personified unity of legal rulesthat give rights and duties to the same man. It is not a natural reality but a legalconstruction” in order to describe “legally relevant facts” [7].2

This way Kelsen, who obviously in the time he wrote can apply law only tohuman beings, makes a legal person without biological constraints and he rejects theanthropomorphic view of the concept of the person.

Both the physical and the legal person are social realities such as the same lawallow us to understand that whether “the law does not create the persons” [7],3

society can regulate a new reality described by technology using this kind of conceptof person.

Uberto Scarpelli, who has introduced the analytical philosophy in Italy, reachedthe thought of Kelsen; according to Scarpelli, the condition of use of the concept ofthe legal person “is in the normative forecast of a plurality of actual and possiblefacts, mere facts or actions, a connection among them,which consists in the referenceto a same legal system and the attribution of a juridical relevance to such facts interms of relationships intervening among them” [8].4

We can use this “limitedly technical” [9]5 concept to regulate new experiencesand new realities the way it arose to answer to the new situations in which “individualsubjects find multiplied and widened their limited life in a lot of new realities” [9].6

In Anglo-saxon legal literature we also find similar positions. According to Radin,legal personality is “a convenient symbol, a short-hand expression by which a greatmany factors can be designated by a single word or a single phrase” [10, p. 625, 643]and the concept of person is “ameans ofmaking reference to complex situations easy,and is only an inexperienced person that is deceived by it, or fancies that calling the

2Translation from the Italian: H. Kelsen, Reine Rechtslhere, 1960, trad. it. La dottrina pura deldiritto, Torino, Einaudi, 1952, p. 193.3Translation from the Italian: La dottrina pura del diritto, Torino, Einaudi, 1952, p. 198.4Translation from the Italian: U. Scarpelli, Contributo alla semantica del linguaggio normativo,Milano, Giuffrè, 1959, p. 142.5Translation from the Italian: G. Capograssi, Il problema della scienza del diritto, n. ed. ac. P.Piovani, Milano, Giuffrè, 1962, p. 81.6Translation from the Italian: Il problema della scienza del diritto, n. ed. ac. P. Piovani, Milano,Giuffrè, 1962, p. 82.


series a, b, c, …, n, by the name of x, add a new item to the series” [10, 665]. Wolffspeaks about abbreviation and a formula of fiction, starting from the natural extra-legal concept of personality and entities considered such as persons despite the lackof human dignity [11, 494, 505].

Rules are a fundamental mechanism of connection between the social and themental by the means of the so-called “multiagent structure” that expresses itself inthe plurality of the recipient of the majority of legal rules. This structure allowsan education which falls beyond the one imagined by Kant, for whom from theconcept of legal entity emerges only and necessarily the relationship between freeand equal people and for this reason, according to him, legality cannot connote neitherour relationship with animals nor the one with God or, we can say today, with theandroids. However, the process of de-psychologizing, abstraction, de-naturalizationthat features the rule of laws allows to them to refer to different models of socialorganization and it shows how they are semantically multipotent thanks to the abilityto use natural elements in an artificial context, in which sometimes by the means ofthe rule “x counts as y in the context c” [12] it occurs the transition from the physicalobject to the social one.

According to Nèkàm the feature that makes a legal entity is the fact that thecollectivity considers it as a unity which has interests that need or deserve socialprotection: “and it is completely irrelevant that such entity exists as an objectivereality, that has its will, and that it possesses or less a natural or artificial, imaginedor proven personality” [13, 21]. Following this perspective, we can think to look forthe theoretical grounds to reshape the new legal challenges that Artificial Intelligencebegins to set.

4 Could Robot’s Electronic Personality Exist?

The theoretical issue that we are facing concerns the chance that artificial systems,which could at least show a certain degree of human agency in case of work sharingbetween robots and humans, would benefit of their own juridical status.

Starting from the first studies, the research in the field of Artificial Intelligencefocused on some classical philosophical issues: learning, perception and the rela-tionship between mind and body. Despite this beginning, it is not at all obvious thatthe current developments in Artificial Intelligence, particularly in robotics, claim fora philosophical approach to the newborn (and future) relationships between humanbeings and robots.

The continuous online life we are experiencing is shaping our way of living in theinformation society claiming—as Floridi [14] underlines—for a new ethical frame-work, in which we should ward off the anthropocentric view over robots allowingto accept them as co-workers, including the huge amount of legal, social and ethicalissues at the stake.


Despite the fight between cyber-optimists and cyber-pessimists [15], it is no longerpossible to image our life without Artificial Intelligence: algorithms follow and sup-port us in a series of daily actions (withdraw from an ATM, watching a TV series onNetflix, booking a doctor’s appointment, applying online at university and so on).If we accept these kinds of examples of Artificial Intelligence, generally speaking,we are skeptics about robots and androids. Because of science fictions films and TVseries, one might conceive human world invaded by robots, condition which presup-poses a fight for supremacy won by artificial systems created by the loosers humanbeings. Obviously, this is a naïve and false view on Artificial Intelligence, since weare not facing an invasion, but those artistic perspective could give us some hints inorder to imagine the new challenges for the future infosphere [16].

Science fiction showed us the chance to have driverless cars (or any other trans-port). Now, the project of a driverless car is under construction: Google, Tesla, Uberand Nissan are testing their prototypes of driverless or autonomous car.

Two severe crasheswhich involved autonomous car occurred inMarch2018. In thecrash occurred in Arizona, an Uber car in self-driving mode killed a pedestrian. Theseverity of the situation and the anticipation of some future scenarioswith damages tohuman beings triggered a lot of questions and doubts.Without keeping in consideringthis accident, whose causes are being realised, other events should be taken intoaccount. In 2016 and 2017, during previously testing phases, there have been someaccidents, without serious consequences: in those cases, the autonomous cars showedthe capability to prevent bigger accidents, like the one prevented in December 2016by a Tesla car running for testing along a Dutch highway.7 Even though not lethal,those accidents confirm the urgency of the revision of some philosophical aspects ofthe concept of responsibility.

Let’s consider for a while, for the sake of argument, that in the same situation thereoccurred a malfunction of the ‘Forward Collision Warning’ onboard the Tesla car:instead of preventing the accident, the autonomous car provokes it, investing a carcarrying three people, including a child. The result of the accident is little damages tothe three people and huge damage to their car. Who is responsible for this accident?The question, that is only apparently provocative, is one of the dilemmas philosophyof artificial intelligence has to solve. There is no need to be afraid of robots, since theycould help us with machine learning, overall in medicine for application in healthsystems for improvement in cancer and other rare diseases research, but usefulnessdoes not exclude a reflection on the future “responsibility from algorithm” [17].

The proposed example shows that we need to think the relationship we would liketo experiencewith those artificial systems. Following this perspective, theCommitteeon Legal Affairs (Rapporteur Mady Delvaux) presented a motion for a EuropeanParliament resolution with recommendations to the Commission on Civil Law Ruleson Robotics8 in order to begin to establish rules and a kind of ethical and legal codeto arrange relationships between human beings and Artificial Intelligence artifacts.

7Link: http://www.carscoops.com/2016/12/teslas-autopilot-anticipates-huge-crash.html.8Link: http://www.europarl.europa.eu/committees/en/juri/subject-files.html?id=20170202CDT01121.


First of all, we should consider the issue about the protection of the persons whoshare a daily experience with a robot in order to avoid potential damages to them. It isthe case of social robots. A social robot is an autonomous or semi-autonomous robotthat is involved in some programmed and scheduled practices which are structuredtaking into account the role they play in interactions with human beings and/orother social robots. Care robots and butler robots are two kinds of social robots. Thecare robot is programmed to help elderly persons in their everyday life while thebutler robot acts like a servant. One critical juncture is the chance of disruptionsand malfunctions of the social robot and eventual damages to the persons he has totake care of. Who will be responsible for those damages? The mentioned Europeanmotion is an attempt to find an answer, facing legal, ethical and social issues. Floridi[14] offered us a clear analysis of the motion, stating that its most relevant featuresare the issues of unemployment and liability. We could add to these issues that

developments in robotics and artificial intelligence can and should be designed in such a waythat they preserve the dignity, autonomy and self-determination of the individual, especiallyin the fields of human care and companionship, and in the context of medical appliances,‘repairing’ or enhancing human beings [18, 5].

This idea covers the chance that developments in robotics and AI could be helpfulto people but their application need to be constantly checked since one cannot ignorethe users are not simply individuals but persons which sometimes have a reducedautonomy but they have the right to preserve their dignity.

Back to the two features stressed by Floridi, we should recall the example of thedriverless or autonomous car. Let’s say, for example, that within a couple of years wecould travel with a Google or Tesla car. It makes sense that taxy service companiesmight fire their drivers, just like bus and train companies since there will be no moreneed to pay people for a job they will not do anymore. So that there might be a lotof unemployers. If this scenario should happen, this means that the social relapsewill be dramatic, so that today there is already the need to think about an economicsupport for people who will lose their job because of robots and algorithms. Then,companies will save money and will reduce complaints, but the social systems ofsome Nations will be troubled by this new developing situation. The only solutionis to tax robots, which means to tax the companies which replace a lot of humanworkers with robots.

Getting back to the matter of responsibility, what if an autonomous car crashesagainst another car drawn by a human being? Who is legally liable for the accident?Apparently, nobody or, at least, the company who has produced the car if there is amalfunction or the owner who made a mistake in running the system. However, themain question concerns whether our legal systems are ready to face these situationsand what are the consequences on the notion of the person.

The previous paragraph showed that there is not an univocal notion of the person.Since concepts such as rights, responsibility, ethics, privacy have been ascribed topersons and not to material things such as robots, how to shape the concept of theperson?


The motion proposed to the European Parliament seems to solve the problemcreating the notion of ‘electronic person’, a new legal status to be given to autonomousrobots:

creating a specific legal status for robots in the long run, so that at least themost sophisticatedautonomous robots could be established as having the status of electronic persons responsiblefor making good any damage they may cause, and possibly applying electronic personalityto cases where robots make autonomous decisions or otherwise interact with third partiesindependently [18, 19].

Although the motion highlights the relevance of the attribution of the juridicalstatus, it does not clarify the contexts of its application. Unfortunately, the term‘electronic person’ appears only once in the whole motion (the last version is adocument of 68 pages). This recommendation to the European Parliament is notdeeply explained or investigated.

It seems that the legal commission, taking into account the difficulties in inves-tigating the notion of the person as it is clear in the history of the philosophy oflaw, has proposed the notion of ‘electronic person’ in order to begin to think to theconsequences of robots on human beings.

The purpose of the commission which uses the term ‘electronic person’ seems amistake: since an ‘electronic person’ is a robot, then it is an artifact, which is some-times, in particular, an Information Technology artifact, and it has to be consideredan artificial system. Unfortunately, this artificial system could be invested of rights,duties, legal liability. If this should happen, then it will make sense to start think-ing about robots such as Isaac Asimov’s Bicentennial Man, ascribing to robots—infuture—also emotions, according to the most wide meaning this word can cover.

The anthropomorphic view risks to make us think the problem of the relation-ship between robots and persons—the persons we are—following the path whichsometimes we run thinking about the relationship between human beings and ani-mals. However, there is a big question: animals are not artificial systems, but robots’behavior is always governed by rules imposed by their creators.

We can say, agreeing with Floridi, that the relevance of the issue needs to beconsidered changing the point of view. So then, we need to reshape the notion ofelectronic person, giving sense to the terms electronic and person in a philosophicalperspective. In particular, in the previous paragraphs, we showed that the philosophyof law of Scarpelli, Kelsen andNèkàm, for example, ward off the biological constrainwhich generally is ascribed to the person, trying to describe a new kind of society inwhich there is a widespread co-working with robots and androids.

5 Conclusion

The remarks on the term “electronic person” are not conclusive by the very nature ofthe whole issue. First of all, there is no chance to strictly define the electronic personwithout living room for future reviews by reason of the developments in robotics


and Artificial Intelligence. Paragraph 2 showed that there is not a univocal conceptof person and legal person. Ancillary concepts such as responsibility and ethics getinvolved also in the issue of electronic personality that is an artificial personality.How to make consistent a system of qualities, generally ascribed to human beings,recently at stake because they could be also ascribed to artificial beings? One couldsimplify the question such a new sort of opposition between natural and artificial.On the contrary, we are facing the risk of an anthropomorphic view, while we need anew ethical and philosophical perspectivewhich follows over the year the prospectiverelationship between robots and persons.

References

1. Minsky, M.: Alienable rights. In: Android Epistemology, pp. 307–312. MIT Press, Cambridge,MA (1995)

2. Floridi, L.: L’ultima legge della robotica. La Repubblica (2017)3. Gallo, G., Stancati, C.: Where is the mind? The extended mind reloaded. In: Proceedings of

Annual Convention of the Society for the Study of Artificial Intelligence and the Simulationof Behavior (AISB 2016), pp. 242–249. Curran Associates, New York (2017)

4. Carruthers, M., et al. (eds.): The Category of the Person. Anthropology, Philosophy, History.Cambridge University Press, Cambridge (1985)

5. Minsky, M.: The Emotion Machine. Commonsense Thinking, Artificial Intelligence and theFuture of the Human Mind. Simon & Schuster, New York (2006)

6. Bartolo di Sassoferrato: Commentarium in Digestum Novum, c. D. 48, ii9, 16, 107. Kelsen, H.: Reine Rechtslhere. Reine Rechtslehre. F. Deuticke, Wien (1934)8. Scarpelli, U.: Contributo alla semantica del linguaggio normativo. Giuffrè, Milano (1959)9. Capograssi, G.: Il problema della scienza del diritto, new edition. Giuffrè, Milano (1962)10. Radin, M.: The endless problem of corporate personality. Columbia Law Rev. 32, 643–667

(1932)11. Wolff, M.: On the nature of legal persons. Law Q. Rev. 54(4), 494–521 (1938)12. Searle, J.: Mind, Language and Society. Basic Books, New York (1998)13. Nékàm, A.: The Personality Conception of the Legal Entity. Harvard University Press,

Cambridge, MA (1938)14. Floridi, L.: Roman law offers a better guide to robot right than sci-fi. Financial Times (2017)15. Lovink. G.: Network Without a Cause. Polity Press, London (2012)16. Floridi, L.: Robots, jobs, taxes and responsibilities. Philos. Technol. (2017). https://doi.org/10.

1007/s13347-017-0257-317. Ruffolo, U.: Per i fondamenti di un diritto della robotica self-learning: dalla machinery pro-

duttiva all’auto driverless: verso una “responsabilità da algoritmo”? In: Ruffolo, U. (ed.)Intelligenza Artificiale e responsabilità, pp. 1–30. Giuffrè, Milano (2017)

18. Committee on Legal Affairs (RapporteurMadyDelvaux): European Parliament resolutionwithrecommendations to the Commission on Civil Law Rules on Robotics. http://www.europarl.europa.eu/committees/en/juri/subject-files.html?id=20170202CDT01121

19. Stroud, F.: The Judicial Dictionary of Words and Phrases Judicially Interpreted. Sweet &Maxwell, London (1890)

Habits, Routines and Rituals

Cognitive Dynamics of ResearchRoutines: Case Study of MicroRNA

Paweł Kawalec

Abstract The paper develops evolutionary approach to study research routines inorder to identify some general patterns in this prevalent area of social practice.Three mechanisms underlying cognitive dynamics of research routines are iden-tified and examined using the case of microRNAs research. Their mutual interactionis attributed to propagation of shock impulses. Some inherent limitations of alterna-tive scientometric approaches, such as the gradient of flow vergence, are examined.They presumably arise because of the particularistic orientation of the dominant sci-entometric approaches. The paper concludes by advocating processualist approachto dynamics of scientific research and postulating a more detailed examination of thenature of shock impulses.

Keywords Breakthrough research ·MicroRNAs · Processualism · Researchroutine · Science of science · Scientific discovery

1 Introduction

The idea to systematically investigate science as a social practice was first intro-duced in the 1920s and 1930s by the founders of “the science of science” move-ment [26] Znaniecki [57], Ossowska and Ossowskis [41, 42] and independently byBernal [8]. Maria and Stanisław Ossowskis elaborated on the former work of Flo-rian Znaniecki and identified three core disciplines that constitute scientific study ofscience: epistemology and philosophy of science, psychology of scientific creativityand anthropology or sociology of science. They argued that understanding scienceas a social practice is necessary in order to grasp the dynamics of science, which can-not be uncovered by studying its “immanent factors” alone [42, p. 75]. The overallargument presented by Ossowskis was meant to support the idea of an autonomousdiscipline of “the science of science” and, in consequence, give credit to the plansof launching a dedicated research institute in the late 1930s. These plans, however,

P. Kawalec (B)Faculty of Philosophy, John Paul II Catholic University of Lublin, Lublin, Polande-mail: [email protected]


133

134 P. Kawalec

did not materialize because of the outbreak of WWII and the post-war communistregime in Poland.

An alternative proposal was advanced by John Bernal in his book The SocialFunction of Science (1939). It was, at least partly, motivated by “disillusion” dueto negative effects of science development, such as unemployment, inequality ornew weapons threatening “personal security”. Moreover, as Bernal observed, therewas “a strange coincidence, the disturbing events of the Great War, the Russianrevolution, the economic crises, the rise of Fascism, and the preparation for newerand more terrible wars have been paralleled inside the field of science by the greatestchanges in theory”, such as emergence of mathematical logic, relativity and quantummechanics, biochemistry and genetics [8, p. 2]. So, he set out to elaborate on “thefact that science is both affecting and being affected by the social changes of ourtime”, emphasizing that “this is a social and economic rather than a philosophicalinquiry” [8, p. 9]. For Bernal, then, overall “Scientific research and teaching are infact small but critically important sections of industrial production” [8, p. 10].

Admittedly, he was first to provide ample, but necessarily incomplete, evidenceestablishing the link between the economic growth and intensity of R&D. In fact, hesignificantly contributed to the UK almost doubled increase in R&D public expendi-ture of the post-war period. Bernal’s unwillingness to pursue the underlying mech-anism of the effectiveness of economic regularities related to R&D spending wasinherited by the post-war mainstream economists like Paul Samuelson, KennethArrow, Richard Nelson and Robert Solow.

Slightly later, in the early 1950s, Derek de Solla Price initiated the turn of “thetools of science on science itself” (1951). He explicitly motivated it by an analogywith physics: “Themethod to be used is similar to that of thermodynamics, inwhich isdiscussed the behavior of a gas under various conditions of temperature and pressure.One […] considers only an average of the total assemblage in which some moleculesare faster than others, and in which they are spaced out randomly and moving indifferent directions. On the basis of such an impersonal average, useful things canbe said about the behavior of the gas as a whole, and it is in this way that I want todiscuss the analysis of science as a whole” [17, p. xiv].

However, it was only after WWII that this idea really took off. Although themain perspectives, such as philosophy and history, economics, scientometrics andsociology significantly differed, they seem to have converged on a common topic ofthe growth of knowledge.1

In the remainder of this paper I focus on the philosophical and descriptive per-spective on the cognitive aspect of this dynamics, leaving aside its social aspect.Important modern contributors, such as F. Bacon or Descartes, predate the contem-porary philosophers of science in their search for “New Organon”, “the logic ofinduction” or “the logic of discovery”. During the early decades of the 20th cen-tury the analysis of scientific discoveries was predominantly “logical” [44, p. 3],which reflected the then recent canonical formulation of the mathematical logic.Nonetheless, already forty years ago the narrowness of this approach was forcefully

1More fully discussed in [27].

Cognitive Dynamics of Research Routines: Case Study of MicroRNA 135

questioned [21, p. 120] by, among others—what was later dubbed as—“the friendsof discovery” [40]. In effect, it paved the way for a new research on science dynamicsthat has given rise to a plethora of different approaches.

Following [22] I overview the main threads in these accomplishments and char-acterize the contribution of the present paper within this thriving research field. Thedeductive approach develops the traditional approach of logical empiricists, in par-ticular Popper’s. The main claim is that while science progresses by constructing“hypotheses, or systems of theories” and testing “them against experience by obser-vation and experiment”, it is the task of deductive logic to reconstruct “this procedure”and “analyse the method of the empirical sciences” [44, p. 3]. The obvious objec-tion is that deductive logic is non-ampliative, so deductions can merely make moreexplicit the information that was already presumed in the premises. Emiliano Ippolitidebunks the main counter-arguments and demonstrates that the deductivist approachcannot account for generation of hypotheses as “the crucial step in discovery” [23].

The cognitive approach provides an alternative perspective on scientific discovery.It was initiated by Simon [47] and characteristically claims that scientific discoveriesas well as “creative products of all sorts are brought about by our ordinary cogni-tive processes, such as those involved in our day-to-day problem-solving activities”[55, p. xi]. Further, it is claimed that these cognitive activities are rule-governedand transparent to cognitive agents themselves, so, in principle, it is possible toimplement them in any kind of information-processing systems, such as computers[31]. Nevertheless, three objections to this approach still seem to hold. Firstly, theimplementations, such as the most famous software BACON, are not capable of per-forming the crucial step in the generation of a new hypothesis, namely identificationof relevant variables and data. Secondly, by analogy to the deductivist approach, theymerely reconstruct the historical processes and, at best, discover a new solution toa problem that has already been recognized and solved independently. Thirdly, thiscognitive system is incapable of generating new conceptualizations that would drivesubsequent formulations of problems and their solutions.

Evolutionary approach provides yet another way to conceive of scientific discov-ery. In its traditional formulation [14] itmimics biological processes in claiming blindgeneration of hypotheses that are subsequently subsumed to processes of selectiveretention during acquisition of new data or new theoretical models. One of the objec-tions to this approach is that the process of scientific discovery is not intentional andis not guided by methods or rules. The other objection is that it is, in principle, notpossible to articulate the aim of science as an endpoint of an evolutionary process.

While my approach presented in this paper takes the criticisms and limitationsof the three aforementioned approaches for granted, it builds on the evolutionaryapproach. In my earlier publications [25, 27, 28] I elaborated heuristics-driven vari-ation + institutional selective retention (HDV + SR) model of two-fold cognitive-social dynamics of science. It claims, roughly, that the hypotheses are generated asprojections of the previously established body of scientific knowledge that are furtherexplored in order to ground them empirically [53]. If successful, the process may ini-tiate a growing institutional set-up that promotes diffusion of the hypotheses among

136 P. Kawalec

the members of scientific community and may bring about a new research routineas a recognizable social practice of research sharing a symbolic representation.

HDV+SRmodel has several characteristics that stand out from the traditional BV+SRmodel. First, it takes into account amore complex ontology of scientific change,including not only ideational generations and re-combinations, but also an associatedsocial ontology [19]. Second, it takes the growth of knowledge to constitute anintentional process that is rule-governed. Third, it propounds a processualist approachto the growth of knowledge [39]. The traditional model of “scientific discovery”presumes that there are some special “turning points” that orient the developmentof the whole research field. In contrast, I favor the perspective that takes the growthof knowledge to be a continuous process which, of course, exhibits variations in itsdynamics that are due to interaction of different underlying mechanisms and shockimpulses. The present paper uses the case of one of the major recent breakthroughsin molecular biology, namely microRNA, to elaborate in greater detail on the latterpoint and to strengthen the processualist aspect of HDV+ SRmodel by identificationand examination of the aforementioned mechanisms.

2 The Conception of Research Routines Appliedto the Discovery of MicroRNAs

It has to be remarked that the terminology used to discuss the issues related toscientific discovery and its various kinds, such as “breakthrough”, “novel” or “trans-formatory”, is far from clear. The use of the very term “scientific discovery” hasranged frommeaning a single outcome to referring to the whole procedure of inquiry[46]: “In the narrowest sense, the term ‘discovery’ refers to the purported ‘eurekamoment’ of having a new insight. In the broadest sense, ‘discovery’ is a synonymfor ‘successful scientific endeavor’ tout court”. The latter use, however, has steadilydisappeared from modern philosophy of science with its gradual divergence fromscientific research.

Apparently, the term “scientific discovery” still connotes the kind of approach inphilosophy of science which I would call here “particularist” as it tends to focus onpinning down particular moments of radical change in science. Most often they areunderstood as manifested by key publications that are indicated as turning points inthe history of science and are recognized by prestigious scientific awards. BecauseI would like to avoid the particularist presumption here, I am inclined to use analternative and more processualist-oriented terminology that allows me to carve outmore complex wholes within the process of inquiry.2 It appears that research processis the obvious candidate for such a unit, granted a processualist approach, as itdelivers the outcomes which are then disseminated in publications and it is relatively

2I adopt here the terminological distinctions developed in the debate between proponents ofparticularist and processualist anthropology (see esp. [10, 11]).


well discriminable from other research processes by its unique research question, theadopted methodology, its specific research design, the collected data and results.3

As argued in my earlier publications, my preferred way to approach the dynamicsof research processes is in terms of research routines. By “research routine” I meana repeated and recognizable pattern of research practices of a scientific communitywhich share a symbolic representation, such as a model or theory. Research routinepresumably constitutes themain source of the heuristics-driven variance for each timeit is used in a particular research process it is enacted in amodifiedway due to specificresearch question and research design. In effect, it is accordingly used to project newhypotheses. Nevertheless, in most research processes research routines remain stableand the respective changes do not affect the core of the symbolic representation, butonly expand it as they become more specialized and more widely applicable. Onlyoccasionally some components of the symbolic representation undergo a signifi-cant modification and eventually a new research routine emerges. However, it is notaccomplished in a single research process, but lasts for an extended period of timeand remains continuous between the old and the emergent routine as exemplified bythe phases in the development of microRNAs [28].4

As mentioned in Sect. 1, research routines result from the twofold dynamics:cognitive and social. The cognitive progress is evidenced by the appearance of newconcepts, new distinctions and links between them, new areas of application, etc.The social dynamics is evidenced by the growing institutionalization of a givenresearch routine, such as getting from an informal group of researchers throughacquiring large-scale research projects, establishing standards (e.g. terminology,databases, protocols), setting up dedicated graduate educational programs and pub-lication of monographs, and new formal institutions (e.g. chairs, departments, insti-tutes, research laboratories etc.), up to formation of subdisciplines, specialized con-ferences, learned societies and journals. Although these two dynamics are integratedin the growth of research routines, for the purpose of the present paper I try to singleout the cognitive dynamics and determine its underlying mechanisms.

3 The Cognitive Dynamics of MicroRNAs

A quantitative approach to scientific discoveries within the BV+ SRmodel has beenelaborated by Simonton [48–50]. His model uses three variables which correspond tothe criteria of theUSPatentOffice that were elaborated in view of patent assessments.Non-obviousness or surprisingness is intended to measure the extent to which a newdiscovery was unexpected given the state-of-the art. Novelty is concerned with the

3A general and forceful case for the processualist turn in the philosophy of life sciences has recentlybeen presented in [39].4If one adopts John Losee’s broad classification of philosophical approaches to scientific progressinto incremental and discontinuous [37, p. 1], then obviously the general processualist approachinclines towards the incremental one that would presumably accommodate research routines as thefavorable unit of analysis.

138 P. Kawalec

issue of how new knowledge expands the existing body of knowledge, such as newclassifications (e.g. subcategories of microRNAs) or new relationships between theexisting categories (e.g. microRNAs as tumor suppressors). And, finally, utilitymea-sures the extent to which the new idea is useful. Simonton set the scope of values foreach of the three variables in analogy with probabilistic measure ranging between0 and 1. The product of the three variables constitutes the measure of scientificcreativity.

As argued elsewhere [25, 27, p. 133], for epistemological reasons I do not followSimonton’s procedure for parametrization of his creativity model.5 Nonetheless, Ifind his conceptualization useful in constructing the argument to support the pro-cessualist approach to the growth of scientific knowledge. Thus, I use these threedimensions: surprisingness, novelty and utility to identify three separate interact-ing mechanisms which underlie the process of new research routine formation. Thereminder of this section presents detailed examination of these three mechanismswith regard to microRNAs as a new research routine. Before setting on this task,however, I embark on a succinct characterization of this newly discovered kind ofRNA particles and overview the main phases that led to establish the correspondingnew research routine.

On the basis of the so called “Crick’s dogma” it has been taken for granted since1950s [24, p. 1] that DNA is transcribed into RNA, which, in turn, is translated intoprotein. At the time preceding the discovery of microRNAs “standard wisdom” [6,p. 133] was that there existed only three major kinds of RNAs: messenger (mRNA),transfer (tRNA) and ribosomal (rRNA). The recognition of a new kind of RNAparticles that were not involved in protein translation—so called non-coding RNA—and acknowledgment of their occurrence in a rich class of organisms was a slow andlong-time process.

MicroRNAs are approximately 20 nucleotides in length that regulate gene expres-sion by inhibition of translation or mRNA degradation. MicroRNAs may regulateup to 30% of the human protein-coding genome and control the expression of genesinvolved in key biologic processes, including apoptosis, proliferation, differentiation,and metastasis [2, 4, p. 1]. The microRNAs database—miRBase (v.22.1, October2018) records 38,589 different microRNAs occurring in 271 kinds of organisms,including Homo Sapiens with 1917 identified kinds of microRNA.

There are several phases that led to the discovery of microRNAs and their fur-ther exploration [28]. The relevant line of research has started in 1980 when Vic-tor Ambros, joined later by Garry Ruvkun, became a post-doc team member ofRobert Horvitz’s team at his laboratory in MIT. Initially, they tried to identify “hete-rochronic” geneswhichwere responsible for developmental mutations of a nematodeCaenorhabditis elegans (C. elegans), a model organism used in studies of mutations.By 1989 they were able to identify the functional and later on also the molecularmodel. In 1993 Ambros with collaborators reported what is now recognized as thefirst microRNA, namely the gen lin-4, which “does not encode a protein”, but rather“regulates lin-14 translation” [34, p. 843]. A complimentary paper was published by

5For the details concerning how the values of Simonton’s measures are determined see e.g. [49].


Ruvkun’s research team [56], who in 2000 demonstrated moreover that the secondmicroRNA, namely let-7, is highly evolutionarily conserved and—in contrast to lin-4that occurs only in a small class of organisms—is prevalent in a wide class of species,including humans [43]. With this result came the recognition that non-coding RNAsform an important class of RNA particles that regulate expression of other genes.Right after the standard terminology of “microRNAs” (or “miRNAs”) was coined in2001 [30, 32, 33] researchers started to explore the role of microRNAs in oncogen-esis [12]. In 2007 cell-free microRNAs (circulating miRNAs) have been detected inplasma and serum [52], demonstrating high therapeutic and diagnostic potential ofmicroRNAs as non-invasive biomarkers.

A synthetic way to illustrate the evolution of microRNA as a research routine ispresented in Fig. 1. In 1993 the giant component in the author collaboration network,where nodes indicate authors who are linked to their co-authors of their joint pub-lications, included only 26 authors, who constituted 11% percent of all the authorswho published papers on issues related to microRNAs. In contrast, in 2018 the giantcomponent included 70% of the authors totaling 27,176 individuals. It not only indi-cates the widespread diffusion of microRNAs among researchers, but, more impor-tantly, it demonstrates the fact that microRNA biology has become established as asub-discipline of molecular biology of its own.

Prima facie, it appears that the growth of microRNA routine follows a typicaldiffusion curve following the inverted S-shape (Fig. 2). In the remainder of thissection I demonstrate that it indeed is a more complex phenomenon by pointingout three interacting mechanisms related to: breakthrough, novel and utility-orientedresearch. As it turns out, the diffusion curve may only be taken to characterize justone of these mechanisms related to novel research.

While the difference between breakthrough and novel research is well establishedand has been epitomized by ThomasKuhn’s distinction between “revolutionary” ver-sus “normal” stages for the sake of the research on microRNA as a research routine

Fig. 1 The evolution of microRNA. The two snapshots are based on analysis of MEDLINE datafor microRNA and its MeSH synonyms as MeSH keywords

140 P. Kawalec

Fig. 2 MicroRNA publications. It presents the annual number of publications on microRNA assubject in Web of Science (WoS) database

I adopt the following distinctions. Breakthrough research creates the new symbolicrepresentation that later becomes the organizing principle for the expanding socialpractices—starting in a single research team and diffusing steadily among otherresearch laboratories. As demonstrated in my earlier paper [28] microRNA in theperiod 1984–2000 has become established as a new symbolic representation, in con-trast with the earlier predominant view of RNA as coding proteins. Novel researchhas expanded this symbolic representation by advancing knowledge of microRNAbiogenesis, complimentarymolecularmechanism (RNAi), prediction of target genes,identification of microRNAs particles and their occurrence in various kinds of organ-isms, etc. Utility-oriented research applies the knowledge concerning the genesis,structure, function or types ofmicroRNAs to practical problemsof diagnosis and ther-apy in other areas of research (see also Fig. 6). As mentioned above, utility-orientedresearch has commenced as early as 2002 with the publication of [12].

The next section proceeds with a more detailed examination of each of thesemechanisms. This investigation will provide ample evidence against the overly sim-plistic view that research routine dynamics follows S-shaped diffusion curve. Letme remark here that the latter view seems to be widely entrenched as it was force-fully propounded by the founder of scientometric research, namely by de Solla Price[15–17]. Using an analogy with a growth process that can be observed in multipleorganisms he claimed that the growth of a scientific discipline as evidenced by theaccumulation of the number of publications pertaining to the discipline in questioncan be divided into phases each of which instantiates S-shaped curve. In order tomodel it, Price used the logit growth model. However, as pointed out recently [5] deSolla Price’s original idea is more adequately captured by the Gompertz function.6

Now, let me briefly make some remarks in order to explain why I did not decide touse this function to model the development of microRNA routine. The application ofthe Gompertz function, like de Solla Price’s preferred logit growth model, consists incutting the relevant dataset into overlapping periods each of which taken separately

6The interested Reader may consult [5] for details of the Gompertz function.


Fig. 3 Gompertz curve fitting for microRNA publications. Annual numbers of publications formicroRNA MeSH synonyms in the MEDLINE database. The dates indicate opening and closingdates for the respective periods. Note that given this partition of the dataset the publication numberfor the remaining part, starting in 1997/1998, grows exponentially

is taken to follow a Gompertz function. This kind of procedure is sometimes dubbed“contingent empiricism” as it is based on somewhat arbitrary approach to fittingthe original curve with handedly marking the cutting points for data subsets. Ineffect, we obtain a whole series of differently parametrized Gompertz functions.They seem to be introduced ad hoc for a particular dataset and lack predictive power.The development of a research routine generates different kinds of data andmanuallyfitting just one of them does not seem to be lead to a more comprehensive view ofthe whole process as manifested in other datasets. Moreover, the set of Gompertzfunctions, each of which matches just one phase of the whole process, does not allowone to explain the mechanisms underlying it. These problems are illustrated in Fig. 3.

4 The Breakthrough, Novel and Utility-Oriented Phasesof the MicroRNAs Cognitive Dynamics

The detailed analysis of the three mechanisms underlying the research practice inmolecular biology, as exemplified here with the dynamics of the microRNA researchroutine, proceeds in accordancewith the threemechanismsdescribed inSect. 3. Theseare: breakthrough, novel and utility-oriented research. I start with a description ofeach of these mechanisms separately, and only then comment on issues related totheir interaction.

142 P. Kawalec

For the sake of the present paper I understand breakthrough research as the onethat establishes the symbolic representation of an emergent research routine. In thecase of microRNA this phase of research can be rather precisely identified. It beginsin 1980, when Victor Ambros and Gary Ruvkun join the research team of RobertHorvitz at MIT and start the research on heterochronic gens of C. elegans.7 Theclosing date for this phase is 2000, when Ruvkun’s own research team demonstratedhigh conservation of let-7, the second microRNA. Of course, the early researchon heterochronic mutations of C. elegans was continuous with the prior research,most importantly Horvitz’s contributions in isolating and complete characterizationof cell-lineages of C. elegans mutants and identification of the role of lin-4 in theprocess [20].

Undoubtedly, it is cumbersome, if not impossible, to apply the typical sciento-metric tools to analyze the breakthrough phase of research. None of the two keypublications: Lee et al. [34] and Wightman et al. [56] is captured by typical searchbased on a keyword selection. Table 1 provides the complete keyword descriptionof both publications in MeSH database. Note that none of the keywords on the lists

Table 1 MeSH keywords forthe first two breakthroughpublications

Lee et al. [34] Wightman et al. [56]

Common keywords

Animals

Base sequence

Caenorhabditis/embryology

Caenorhabditis/genetics

Gene expression regulation

Genes, helminth

Helminth proteins/genetics

Molecular sequence data

Protein biosynthesis

RNA, antisense

RNA, messenger/genetics

Sequence alignment

Sequence homology, nucleic acid

Different keywords

DNA primers/chemistry Morphogenesis

Hydrogen bonding Species specificity

Nucleic acid conformation Time factors

Phylogeny

Time factors

7In 1984 both Ambros and Ruvkun moved out to head their own laboratories.


Fig. 4 Citations of [34]. Citations are grouped in accordance with WoS categories for researchareas, where “basic research” includes: basic biochemistry, cell biology and genetics; “appliedresearch” includes: applied oncology, medical research experimental, cardiac cardiovascular sys-tems, hematology; and “commercial research” includes: biotechnology applied microbiology andpharmacology pharmacy

corresponds to MeSH synonyms for “microRNA”.8 The reason is that—as observedin Sect. 3—the standard term was first introduced as late as 2001 and, in fact, itis when the publication record for “microRNA” as a topic starts in most datasets,including Web of Science (WoS).

Moreover, the analysis of citation data, which embraces all the relevant publica-tions, is not very helpful either. As Fig. 4 makes clear, the citation record for thepaper [34] that provided evidence for the regulatory function of the first micrRNA,in the relevant period of 1993–2000 is rather insignificant. Its recognition is slowand initially starts only in disciplines related to basic research, while its citationrecord in applied and developmental research is lagged by approximately 4 yearsand begins after the period of breakthrough research. It explains, perhaps, why thecomprehensive study of the emergent topics in [51] wrongly targeted 2006 as theyear of emergence of the topic “microRNAs&cancer”.

A closer look at the process of unfolding the recognition of lin-4 as the firstmicroRNA makes a pointed attribution of the 1993 “discovery” more problematic.The gen lin-4 was already identified in the early 1980s by Horvitz and his collabora-tors [20]. The functional model identifying the relationship between lin-4 and lin-14was fully described by 1989 [3, 28, 36, 136–137]. Admittedly, the molecular modelthat was to be published in 1993 was known at around 1991–1992 [18]. Still, theunderstanding of the role of non-coding RNAs came only in 2000 with evidence ofhigh evolutionarily conservation and the evidence of their widespread prevalence in

8“MeSH” stands for “Medical Subject Headings” and is systematically developed and updated forindexing articles in PubMed. It identifies 16 synonyms for “microRNA”.

144 P. Kawalec

different species. So, the quantitative analysis of publication record is significantlycompromised in its potential to uncover the initial breakthrough phase of a researchroutine emergence.

Once we acknowledge the inherent limitations of scientometric and quantita-tive analyses of breakthrough research, the recurrent question remains: Is there anypattern specific to this kind of research? To know the answer would help us to re-focus research and likely speed up the creativity of scientific knowledge. After manyfutile attempts, undertaken in the course of working on this publication, in order toidentify endogenous factors or exogenous shocks, one observation stands out. Asalready mentioned, there were several important research outcomes preceding the1993 publications, but no major breakthrough in the relevant research areas can benoted, including cell biology and molecular biology. Moreover, the experimentaltechniques reported in [34] were well-established (with dates of origin in parenthe-ses): restriction endonuclease (1962), Southern blotting (1975), Northern blotting(1977), Sanger DNA sequencing (1977), chromosome walking (1978), restrictionfragment length polymorphisms (1978), transformation rescue (1982) and RNAaseprotection (1982). Neither was I able to identify institutional (newcomer researchers,new research team, project or lab) or financial factors (no significant change in theNIH founding levels can be observed), as external shocks, which could provide anexplanatory background. Thus, the first observation is that we need to uphold thewell-worn idea of stochastic nature of the breakthrough phase and retreat to theposition that breakthroughs are in principle unpredictable.

The next phase in microRNA research started in 2001. As mentioned in Sect. 3, itinitially was focused on elaboration of microRNA biogenesis, which was thoroughlyexpanded with a complimentary discovery of the mechanism of RNA interference(RNAi), and on the prediction ofmicroRNAs target genes. The number of recognizedmicroRNA particles grew rather quickly as this task was greatly aided by estab-lishment of miRBase in 2002, which also helped in standardization and reportingoccurrences of different microRNAs in different species.

In contrast to the breakthrough phase, novel research seems to be well captured bythe typical scientometric data, including the annual publication records, the numberof authors and their affiliations (Figs. 4 and 5). Note that the WoS database is able totrace the records related to microRNA only up to 2001, when the symbolic represen-tation for the research routine in question has in principle been established. Thus, animportant corollary to de Solla Price’s original idea is that the characteristic invertedS-shape diffusion curves, exhibited in Figs. 2 and 5, can only be recognized at a laterphase when the respective research routine has already emerged and established itsspecific symbolic representation.

The final aspect concerning microRNA routine as a social research practice isrelated to utility. Asmentioned in Sect. 3, I adopt Simonton’s understanding of utilityas concerned with intellectual creativity. Thus, a more adequate name to representthis concept would, perhaps, be “epistemic utility” as it is taken as an incentive tostart a new line of research in other research areas. Apparently, microRNAs haveturned out to be very useful as the number of research areas that apply this categorygrows logarithmically.


Fig. 5 Novel research and diffusion curves (2000–2018) for a the number of affiliations, b thenumber of publishing authors, and c the number of papers published each year. Analysis of WoSdataset for “microRNA” as topic

Different explanations are possible. It may turn out that this is an artefact ofhow WoS codes the articles. If, on average, each paper is assigned multiple WoScategories, it may well be that the number of categories will, initially, grow muchfaster the number of publications. However, a more plausible explanation is that theconfirmation of the existence and the recognition of high conservation ofmicroRNAsin 2000 “led to intense research aimed at identifying new members of this family.This resulted in the discovery of multiple miRNAs across different species of plantsand animals” [9, p. 759–760].

However, a remarkable question remains: How these three mechanisms interact?One explanation would be that they are largely independent. To some extent thisanswer may be supported by the fact that the phase of novel research apparentlykicks-off just right after the breakthrough phase. Yet the same does not hold for therelation between novel and utility-oriented research as theywere advanced in parallelsince 2002.

A more plausible explanation seeks some sort of interaction between the threephases as obviously developments in each affected the others. A typical linear modelwould have it that the outcomes of breakthrough research fed into novel researchand that, in turn, provided input to utility-oriented research. However, this kind ofsimple accumulative view is not able to account for distinctly different periods ofrapid growth such as 2008–2012 for novel research (Fig. 2) and 2004–2008 forutility-oriented research (Fig. 6). Therefore, I would propose here to append the

146 P. Kawalec

Fig. 6 Number of different research areas using microRNA. Analysis of WoS database for“microRNA” as topic and the WoS categories for research areas

above mentioned explanations by offering an understanding of some of the majordevelopments within these phases in terms of shock impulses.

The most perspicuous example of a shock impulse is the publication of the break-through discovery of high conservation of the second microRNA let-7 [43]. It marksthe very beginning of publication record on microRNAs as topic in most databasessuch as WoS or MEDLINE, which—as argued here—is tantamount to the start ofthe phase of novel research. Another conspicuous example of shock impulse fornovel research is the development of genomic sequencing, which is relatively highlyranked on John Avise’s PS-scale (where PS stands for “Paradigm-Shift”) receiv-ing 7 points for 2001 on the 10-point scale. Figure 7 presents all points on Avise’sPS-timeline 1980 onwards. Apparently, the period of rapid growth of utility-orientedresearch in 2004–2008 corresponds to null achievements on Avise’s PS-timeline. Forthe utility-oriented research a likely shock impulse came from the publication of themost influential review papers [4, 7], which are among the 5 top cited papers in themicroRNAs citation dataset (containing 390,000 publications), which was extractedfrom the MEDLINE database.

In order to capture the difference in the dynamics of shock impulses for novel andutility research mechanisms in the microRNA routine I performed an extension ofFourier series analysis according to Formula (1):

δ(x − a) = 2L

∞!

n=1

sin(nπaL

)sin(nπxL

). (1)


Fig. 7 PS-score for main breakthrough discoveries in evolutionary genetics [6]

The results of this analysis are presented in Fig. 8. Apparently, some of theseshock impulses, although with varying effects, affect both kinds of research in asimilar way (e.g. in 2005, 2008, 2010, 2013, 2015), while others (such as 2011,2014, 2016) have adverse effects on each of the two mechanisms. Thus, if one yieldssome explanatory power to shock impulses in the dynamics of research routine, thenthese results support the view that there are two different mechanisms operating inthe course of the development of microRNA routine.

Fig. 8 Shocks for novel and utility research. Analysis of microRNAs as topic WoS dataset bymeans of extension of Fourier series as given by Formula (1)

148 P. Kawalec

5 Concluding Remarks

The arguments presented in the previous sections support the view that researchas a social practice has far greater complexity than assumed by the majority ofscientometric studies. In order to push this point a little further, I proceed to indicatesome pitfalls of the recent proposal to introduce a measure of “paradigm-shifting”papers. Prabhakaran et al. [45] proposes to use a measure of “the gradient of theflow vergence” as the sole indicator of the papers, which have transformative impactin a given research area. This parameter is defined for a network of paper citations,where indegree and outdegree of a paper i measure respectively the number of itscitations by other papers and the number of other publications cited in the paper i.Additionally, it takes into account the weight of the citations by considering the socalled eigenvalue of the paper i, which depends on the number of citations it getsfrom papers that are themselves highly cited. Thus, flow vergence for a paper i isdefined as the ratio of the difference between incoming and outgoing citations ofa given paper i to the total number of its incoming and outgoing citations and isappended with its eigenvalue:

WFV i =indegi − outdegiindegi + outdegi

+ eigi . (2)

The gradient of flow vergence from paper j to paper i is, then, defined as adifference between their respective flow vergences:

△FV i j = WFVi − WFVj . (3)

As could be expected, the application of (3) as the measure of the gradient offlow vergence to citations of papers on microRNA fails to identify the breakthroughpapers, namely [34, 43]. Table 2 presents the list of 5 top ranked papers that resultfrom application of (3) to the MEDLINE dataset of paper citations on microRNAs.Perhaps the most perspicuous misidentification is the last paper [35], because itis not even a research paper, but a review one that does not report any new results.Among the other 4 papers that are rather highly cited (given the completemicroRNAscitations dataset contains 390,000 cited papers) the only one that belongs to basicresearch is Kota et al. [29], which reports new small RNA-regulated genes in plants(note that the main line of research concerns animals, and humans, in particular).The other three papers apply microRNAs to oncological research [1, 13, 38], but aspointed in Sect. 3 it was a much earlier paper of George Calin and his collaborators[12] that is commonly recognized as opening this new line of applied research.

Moreover, the ranking based on the gradient of flow vergence misrepresents therole of the key two “paradigm-shifting” papers [34, 43] that reported, respectively,the first and the second microRNAs and the latter’s high evolutionarily conservation.Table 3 reports their respective rankings.


Table 2 Top rank papers in gradient scale. The numbers in parentheses indicate a position of thepaper among 500 top cited papers

Gradient based ranking (500 topcited ranking)

Paper source Title

1 (128) Kota et al. [29] Therapeutic microRNA deliverysuppresses tumorigenesis in amurine liver cancer model

2 (143) Allen et al. [1] microRNA-directed phasingduring trans-acting siRNAbiogenesis in plants

3 (22) Calin and Croce [13] MicroRNA signatures in humancancers

4 (5) Lu et al. [38] MicroRNA expression profilesclassify human cancers

5 (144) Lee and Dutta [35] MicroRNAs in cancer

Analysis based on MEDLINE dataset for “microRNA” as MeSH keyword

Table 3 Position of the breakthrough papers in the gradient based ranking

Paper Position in gradient based ranking

Lee et al. [34] 128

Pasquinelli et al. [43] 375

Analysis based on MEDLINE dataset for “microRNA” as MeSH keyword

These failures of the gradient of flow vergence is just an illustration of a moregeneral problem discussed in this section. All the attempts to provide a single scien-tometric measure of such a complex social practice as scientific research are perme-ated by serious shortcomings. In [28] I discussed in more detail problems of anotherapproach adopted in [51] that used one of the largest datasets of full-text publica-tions. Nevertheless, it came out with the wrong identification of [54] as the sourceof the newly emerged topic “microRNAs&cancer”. The obvious problems with thisidentification are: “the year of emergence” of the new topic 2006 was in fact pre-ceded by 13 years of basic research on microRNAs, the paper by Stefano Volinia etcollaborators is a review paper that follows their earlier published research papers.Moreover, it is ranked 14 among top cited papers (out of 390,000) and 36 (out of500) by the gradient of flow vergence.

The three cognitive mechanisms, pertaining to creative social practices, illus-trated here with the development of the phases of breakthrough, novel and utility-oriented research on microRNAs demonstrate that research processes display com-plex dynamics. Apparently, there are inherent difficulties for the scientometricapproach that impede the elaboration of an adequate measure of this development. Amore comprehensive approach is needed, such as, for instance, the kind of analysispresented here that attempts to identify the mechanisms underlying the dynamics ofscientific knowledge. It would, however, require a recognition of processualist nature

150 P. Kawalec

of the research dynamics and, as a consequence, would call for an elaboration of newanalytic tools that would be more adequate with regard to more complex units ofanalysis, such as research routine propounded in this paper. Such advances may belargely aided by a detailed scrutiny of shock impulses, their sources and interactionswith different mechanisms underlying the dynamics of research routines.

Acknowledgements The author gratefully acknowledges the support of the Polish NationalScience Centre (NCN) under the grant no. UMO-2014/15/B/HS1/03770.

References

1. Allen, E., Zhixin, X., Gustafson, A.M., Carrington, J.C.: MicroRNA-directed phasing duringtrans-acting siRNA biogenesis in plants. Cell 121(2), 207–221 (2005). https://doi.org/10.1016/j.cell.2005.04.004

2. Almeida, M.I., Rui, M.R., Calin, G.A.: MicroRNA history: discovery, recent applications,and next frontiers. Mutat. Res. 717(1–2), 1–8 (2011). https://doi.org/10.1016/j.mrfmmm.2011.03.009

3. Ambros, V.: A hierarchy of regulatory genes controls a larva-to-adult developmental switch inC. elegans. Cell 57(1), 49–57 (1989). https://doi.org/10.1016/0092-8674(89)90171-2

4. Ambros, V.: The functions of animal microRNAs. Nature 431(7006), 350–355 (2004). https://doi.org/10.1038/nature02871

5. Andersen, J.P., Hammarfelt, B.: Price revisited: on the growth of dissertations in eight researchfields. Scientometrics 88(2), 371–383 (2011). https://doi.org/10.1007/s11192-011-0408-8

6. Avise, C.: Conceptual Breakthroughs in Evolutionary Genetics: A Brief History of ShiftingParadigms. Elsevier, Amsterdam (2014)

7. Bartel, D.P.: MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116(2), 281–297 (2004). https://doi.org/10.1016/S0092-8674(04)00045-5

8. Bernal, J.D.: The Social Function of Science. Routledge, London (1939)9. Bhaskaran, M., Mohan, M.: MicroRNAs: history, biogenesis, and their evolving role in ani-

mal development and disease. Vet. Pathol. 51(4), 759–774 (2014). https://doi.org/10.1177/0300985813502820

10. Binford, L.R.: Archaeology as anthropology. Am. Antiq. 28(2), 217–225 (1962). https://doi.org/10.2307/278380

11. Buettner-Janusch, John: Boas andMason: Particularism versus generalization. Am. Anthropol.59(2), 318–324 (1957)

12. Calin, G.A., Dumitru, C.D., Shimizu, M., Bichi, R., Zupo, S., Noch, E., Aldler, H., et al.:Frequent deletions and down-regulation of micro-RNA genes miR15 and miR16 at 13q14 inchronic lymphocytic leukemia. Proc. Natl. Acad. Sci. 99(24), 15524–15529 (2002). https://doi.org/10.1073/pnas.242606799

13. Calin, G.A., Croce, C.M.: MicroRNA signatures in human cancers. Nat. Rev. Cancer 6(11),857–866 (2006). https://doi.org/10.1038/nrc1997

14. Campbell, D.T.: Blind variation and selective retentions in creative thought as in otherknowledge processes. Psychol. Rev. 67(6), 380–400 (1960). https://doi.org/10.1037/h0040373

15. de Solla Price, D.J.: Quantitative measures of the development of science. Arch. Int. d’HistoireSci. 14, 85–93 (1951)

16. de Solla Price, D.J.: Networks of scientific papers. Science 149(3683), 510–515 (1965). https://doi.org/10.1126/science.149.3683.510

17. de Solla Price, D.J.: Little Science, Big Science—And Beyond. Columbia University Press,New York (1986)


18. Gitschier, J.: In the tradition of science: an interview with Victor Ambros. PLoS Genet. 6(3),1–4 (2010). https://doi.org/10.1371/journal.pgen.1000853

19. Hacking, I.: The Social Construction of What?. Harvard University Press, Cambridge, MA(1999)

20. Horvitz, H.R., Sulston, J.E.: Isolation and genetic characterization of cell-lineage mutants ofthe nematode caenorhabditis elegans. Genetics 96(2), 435–454 (1980)

21. Hoyningen-Huene, P.: Context of discovery versus context of justification and Thomas Kuhn.In: Schickore, J., Steinle, F. (eds.) Revisiting discovery and justification: historical and philo-sophical perspectives on the context distinction, pp. 119–131. Springer, Dordrecht (2006).https://doi.org/10.1007/1-4020-4251-5_8

22. Ippoliti, E.: Scientific discovery reloaded. Topoi (2017). https://doi.org/10.1007/s11245-017-9531-3

23. Ippoliti, E.: Building theories the heuristic way. In: Danks, D., Ippoliti, E. (eds.) BuildingTheories, pp. 3–20. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-72787-5_1

24. Jarroux, J., Morillon, A., Pinskaya, M.: History, discovery, and classification of lncRNAs. In:Rao, M.R.S. (ed.) Long Non Coding RNA Biology, pp. 1–46. Springer Singapore, Singapore(2017). https://doi.org/10.1007/978-981-10-5203-3_1

25. Kawalec, P.: Towards an evolutionary model of science dynamics: generation and productionof scientific knowledge. Zagadnienia Naukoznawstwa 53(4), 405–428 (2017). https://doi.org/10.24425/118035

26. Kawalec, P.: Philosophical perspectives: the science of science—from inception to maturity.In: Friedrich, C., Kleeberg, B. (eds.) A New Organon: Science Studies in Interwar Poland,pp. 521–535. Mohr Siebeck, Tübingen (2020)

27. Kawalec, P.: Integrative methodology: a study of dynamics of scientific knowledge. [in Polish].Wydawnictwo KUL, Lublin (2018)

28. Kawalec, P.: Transformations in breakthrough research: the emergence of Mirnas as a researchroutine in molecular biology. Open Inf. Sci. 2(1), 127–146 (2018). https://doi.org/10.1515/opis-2018-0010

29. Kota, J., Chivukula, R.R., O’Donnell, K.A., Wentzel, E.A., Montgomery, C.L., Hwang, H.-W.,Chang, T.-C., et al.: Therapeutic microRNA delivery suppresses tumorigenesis in a murineliver cancer model. Cell 137(6), 1005–1017 (2009). https://doi.org/10.1016/j.cell.2009.04.021

30. Lagos-Quintana, M., Rauhut, R., Lendeckel, W., Tuschl, T.: Identification of novel genes cod-ing for small expressed RNAs. Science 294(5543), 853–858 (2001). https://doi.org/10.1126/science.1064921

31. Langley, P. (ed.): Scientific discovery: computational explorations of the creative processes.MIT Press, Cambridge, MA (1987)

32. Lau, N.C., Lee, L.P., Weinstein, E.G., Bartel, D.P.: An abundant class of tiny RNAs withprobable regulatory roles in Caenorhabditis elegans. Science 294(5543), 858–862 (2001).https://doi.org/10.1126/science.1065062

33. Lee, R.C., Ambros, V.: An extensive class of small RNAs in Caenorhabditis elegans. Science294(5543), 862–864 (2001). https://doi.org/10.1126/science.1065329

34. Lee, R.C., Feinbaum, R.L., Ambros, V.: TheC. elegans heterochronic gene lin-4 encodes smallRNAs with antisense complementarity to lin-14. Cell 75(5), 843–854 (1993). https://doi.org/10.1016/0092-8674(93)90529-y

35. Lee, Y.S., Dutta, A.: MicroRNAs in Cancer. Annu. Rev. Pathol. 4(1), 199–227 (2009). https://doi.org/10.1146/annurev.pathol.4.110807.092222

36. Liu, Z.C., Ambros, V.: Heterochronic genes control the stage-specific initiation and expressionof the dauer larva developmental program in Caenorhabditis elegans. Genes Dev. 3(128),2039–2049 (1989). https://doi.org/10.1101/gad.3.12b.2039

37. Losee, J.: Theories of Scientific Progress: An Introduction. Routledge, London (2003)38. Lu, J., Getz, G., Miska, E.A., Alvarez-Saavedra, E., Lamb, J., Peck, D., Sweet-Cordero, A.,

et al.: MicroRNA expression profiles classify human cancers. Nature 435(7043), 834–838(2005). https://doi.org/10.1038/nature03702

152 P. Kawalec

39. Nicholson, D.J., Dupré, J. (eds.): Everything Flows: Towards a Processual Philosophy ofBiology. Oxford University Press, Oxford (2018)

40. Nickles, T.: Scientific discovery as a topic for philosophy of science: some personal reflections.Topoi (2018). https://doi.org/10.1007/s11245-018-9566-0

41. Ossowska, M., Ossowski, S.: Nauka o nauce. Nauka Polska 20, 1–12 (1935)42. Ossowska, M., Ossowski, S.: The science of science. Minerva 3(1), 72–82 (1964)43. Pasquinelli, A.E., Reinhart, B.J., Slack, F., Martindale, M.Q., Kuroda, M.I., Maller, B., Hay-

ward, D.C., et al.: Conservation of the sequence and temporal expression of let-7 heterochronicregulatory RNA. Nature 408(6808), 86–89 (2000)

44. Popper, K.: The Logic of Scientific Discovery, 2nd edn. Routledge, London (2002)45. Prabhakaran, T., Lathabai, H.H., George, S., Changat, M.: Towards prediction of paradigm

shifts from scientific literature. Scientometrics 117(3), 1611–1644 (2018). https://doi.org/10.1007/s11192-018-2931-3

46. Schickore, J.: Scientific discovery. In: Zalta, E.N. (ed.) The Stanford Encyclopedia ofPhilosophy, Summer 2018 edn. https://plato.stanford.edu/archives/sum2018/entries/scientific-discovery

47. Simon, H.A.: Models of Discovery and Other Topics in the Methods of Science. Springer,Dordrecht (1977)

48. Simonton, D.K.: Creativity and discovery as blind variation: Campbell’s (1960) BVSR modelafter the half-centurymark. Rev. Gen. Psychol. 15(2), 158–174 (2011). https://doi.org/10.1037/a0022912

49. Simonton, D.K.: Combinatorial creativity and sightedness: Monte Carlo simulations usingthree-criterion definitions. Int. J. Creat. Probl. Solving 22(2), 5–18 (2012)

50. Simonton, D.K.: On praising convergent thinking: Creativity as blind variation and selec-tive retention. Creat. Res. J. 27(3), 262–270 (2015). https://doi.org/10.1080/10400419.2015.1063877

51. Small, H., Boyack, K.W., Klavans, R.: Identifying emerging topics in science and technology.Res. Policy 43(8), 1450–1467 (2014). https://doi.org/10.1016/j.respol.2014.02.005

52. Valadi, H., Ekström, K., Bossios, A., Sjöstrand, M., Lee, J.J., Lötvall, J.O.: Exosome-mediatedtransfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells.Nat. Cell Biol. 9(6), 654–659 (2007). https://doi.org/10.1038/ncb1596

53. van Fraassen, B.C.: Modeling and measurement: the criterion of empirical grounding. Philos.Sci. 79(5), 773–784 (2012). https://doi.org/10.1086/667847

54. Volinia, S., Calin, G.A., Liu, C.-G., Ambs, S., Cimmino, A., Petrocca, F., Visone, R., et al.:A microRNA expression signature of human solid tumors defines cancer gene targets. Proc.Natl. Acad. Sci. U.S.A. 103(7), 2257–2261 (2006). https://doi.org/10.1073/pnas.0510565103

55. Weisberg, R.W.: Creativity: Understanding Innovation in Problem Solving, Science, Invention,and the Arts. Wiley, New York (2006)

56. Wightman, B., Ha, I., Ruvkun, G.: Posttranscriptional regulation of the heterochronic genelin-14 by lin-4 mediates temporal pattern formation in C. elegans. Cell 75(5), 855–862 (1993).https://doi.org/10.1016/0092-8674(93)90530-4

57. Znaniecki, F.: Przedmiot i zadania nauki o wiedzy. Nauka Polska 5, 1–78 (1925)

Habit-Based and Goal-Directed Systems:Knowledge Transfer in Individualand Social Learning

Robert Lowe

Abstract The present chapter discusses value-based habitual and goal-directed sys-tems as studied in the animal and human learning literature. It focuses on the meansby which these two systems might interact in knowledge transfer, particularly as itapplies to social learning. Knowledge is conceived here in terms of types of logiccomputations as implemented by neural networks. A discussion of dual-process typestructures in the brain is provided as well as neural-dynamic implementations thereofand considerations for how a perspective of the brain as carrying out logic compu-tations might be useful for developing the general cognitive capacities of artificialagents.

Keywords Logic · Habits · Dual process · Associative learning · Neural networks

1 Background

The relationship between habits and goal-directed systems has been the subject ofmuch research in the animal (and human) learning community over the last century.A variety of animal and human learning studies and models thereof have describedthe link between habits and cognition (e.g. [1, 2]). Indeed, habits in themselves cancomprise ‘higher’ cognitive phenomena, e.g. habits of thought [3]. Nevertheless,habits are typically characterized as learned phenomena and “inflexible, slow orincremental, unconscious, automatic, and insensitive to reinforcer devaluation” [3].To this end, they havemuch in commonwith learning and processing within standardfeedforward neural networks of the type that learn through target (supervised) orreinforcer feedback. In the animal learning literature, on the other hand, habits aretypically framed as concerning associations between stimuli and responses as theyare reinforced by rewarding or punishing feedback (e.g. [4]).

In this chapter the role of habitual and associative processes are considered withrespect to their role in utilizing or shaping logical functions in the neural networks of

R. Lowe (B)University of Gothenburg, Forskningsgången 6, Gothenburg 41756, Swedene-mail: [email protected]


153

154 R. Lowe

the brain. Firstly is discussed the relationship between habits, goal-directed systemsand value functions of the type that are formalized in reinforcement learning models(Sect. 2). Then in Sect. 3 the role of habits (and associative processes) in knowledgetransfer, in relation to boolean logic constructed knowledge, is discussed as it appliesto individual and social learning. In Sect. 4 the relationship between goal-directedsystems and predicate logical function is discussed, again with reference to constitu-tive associative and habitual processes and how bottom-up and top-down ‘systems’may facilitate one another. Finally Sect. 5 offers concluding remarks and how appre-hension of neural network implementation of habit-based and associative structuresmight be utilized in artificial, including robotic, agent controllers.

2 Habits, Goal-Directed Systems and Value Functions

Animal habitual behavior is characterized by inflexibility, unconscious automaticityand insensitivity to reinforcer devaluation (e.g. [3, 5]). Habitual responding entailsa relative persistence for a behavior when a contingent condition for a given, typ-ically reinforcing, outcome is changed (e.g. devalued). Such contingencies includechanges in stimulus-response pairings that precede the acquisition of the reinforceror a change in the relationship between either the stimulus or response and outcome.The standard modus operandi for engendering habits is stimulus-response training[1, 6, 7] where over a number of successive trials, an animal is presented a stimulus(e.g. light flash or sound) followed by two or more response options (e.g. two leversplaced adjacent to each other requiring one or other to be pushed). Finally, whenthe correct response option is chosen, a reinforcer (e.g. food pellet) is presented tothe animal. Repeated correct responses in the presence of a particular stimulus leadto an increasingly reinforced behavior manifesting in a higher probability of thefuture choice of that particular behaviour. Habits occur when the reinforcer-yieldingresponse is ‘over-trained’ and becomes resistant to a change in the reinforcing out-come contingencies that habitual responses yield. In animals (but also in humans [8]),overtraining manifests when the animal produces a particular response to a stimulus(typically rewarding) over a large number of trials such that when the response isdevalued (e.g. no longer leads to the presentation of a reward), the animal perseveratesin the response (at least for a certain number of trials).

It has been suggested that the traditional view of habits is that they are based onstimulus-response pairings that are free from considerations of value, i.e. represen-tations of reinforcement [9]. Miller et al. [9] have compared the traditional view withReinforcement Learning models of dual ‘habit’ and ‘goal-directed’ systems whichare considered to imbue habits with value. In Fig. 1, we show the relationwith respectto the traditional view (left) and a Reinforcement Learning perspective (right) that isfocused on anActor-Critic like architecture. Q-Learning and Sarsamodels (e.g. [10])use state-action valuations and updates based on reinforcing outcomes. Actor-Criticarchitectures (e.g. [11–13]), on the other hand, update their state-action valuationsbased on a Critic that is ignorant of the action selected. Figure 1 (right) shows this

Habit-Based and Goal-Directed Systems … 155

Out

com

eEx

pect

ancy

Stim

uli

Critic

Out

com

e

TD

er

ror

Valu

e

Inpu

tTr

aini

ng

sign

al

HHaa bb

iitt ssSSttii mm

uulluu ss

--RReRReesspp

o onnssee

--RRP PEE

AAs

AAss ss

ooccii aa

t tii oonnss

Act

ions

Envi

ronm

ent

Goa

lsAct

or

Sti m

uli

HHaabb

iitt ssSStt iimm

uulluu ss

--RReRR eess pp

oonnssee

--AAs

AAssss

oocciiaa

tt iioonnss

Ac t

ions

Goa

ls

Go a

l -D

irec

ted

Sti m

ulu s

- Act

ion-

Out

c om

eAss

ocia

ti on s

Fig.1

Habits,goalsandvaluefunctio

ns.Left.Traditional

view

ofhabits,connectedto

agoal-directedsystem

,as

value-free

(adapted

from

[9]).Right.

ReinforcementL

earning(A

ctor-Critic)view

ofhabitsas

temporald

ifference

(TD)value-based—

thearchitecturehere

hasbeen

adaptedto

includeOutcome

Expectancies(potentially

modulated

byGoalinputs)that,unliketraditionalActor-Critic

architectures,directly

interfaceCritic

andActor.T

hetraditionalActor-

Critic

architecture,on

theotherh

and,isinterfaced

only

throughtherewardpredictio

nerror(RPE

)thatformspartof

athree-factor

learning

ruleto

updatebo

thCritic

(Stim

ulus

Value

estim

ates)a

ndActor

(Stim

ulus-A

ctionassociations)

156 R. Lowe

Actor-Critic relation and the Stimulus-Action associations that are formed througha three-factor learning rule that requires reward prediction error (RPE) as comparedto the considered traditional view that entails pure Hebbian stimulus-response asso-ciations. In Fig. 1, outcome expectancies (Fig. 1, right) may be considered simply asreward-based goals (as in [14]). If the outcome expectancies are not explicitly repre-sented, the depiction is comparable to the Actor-Critic architecture of [15, 16]. In thisparticular figure, however, the outcome expectations permit representation of differ-ential outcomes in a manner theorized by Associative Two-Process (ATP) theory—an animal learning theory that in fact subscribes to three factor stimulus-responselearning (see Fig. 2).

The stimulus-response (S-R) associations of ATP form a value-based habitualprocess insofar as they are slow to devaluate relative to goal-directed processes [18].The significance of the S-R process in the context of ATP theory is the view that thevalue-constrained learning of both the S-R process and the outcome/goal-directedprocessing interlinks the two types of systems. Habits and goal-directed processesmay not be seen as working separately on this view.

When considered within the context of dual (or multi)-process learning mecha-nisms such as the Associative Two-Process, value-based habit learning (inflexible,slow to respond to devalued stimuli) should not be considered a lesser or vestigial

Fig. 2 Associative Two-Process theory and its underpinnings, adapted from [17]. a Standardbehaviourist perspective where S–R associations are learned via the outcome expectation repre-sentation (value). b Outcome expectation can cue responding following S–E, E–R learning. c Non-differential outcomes expectations do not provide extra information for cueing responses. d Differ-ential outcomes expectations provide additional information to stimuli for cueing responses. Hereλ denotes non-differential outcomes; λ1 and λ2 are differential


aspect of learning [19]. Many models of animal learning theorize the existence ofa dual process of memory wherein habit-based learning is embedded. It has beensuggested that the habit system takes control of responding late in learning while thegoal-directed process dominates during early learning that constitutes a period of rel-ative uncertainty [20]. Suchmodels posit a complementary process variably entailingexpectation, motivational or goal-directed processes (e.g. [4, 21–27]). Habit-basedlearning routes may be critical for detailed learning of the relationship between per-ceptible stimuli or objects and actions [1]. They are also efficiently learned relative togoal-directed routes that require the learning ofmultiple associations that link stimulito goals (or outcomes) and goals to responses [28]. Furthermore, they may providecrucial confirmation/facilitation of learned associations that underlie goal-directedprocesses [29].

Addressing the question of whether and how habits can guide the learning ofmore cognitive-motivational behavior (relevant to goal-directed activity) that hasbeen posited to develop under dual processes may further clarify the importanceof learning according to habit-based processes. Such cognitive phenomena includeclassification-based/category learning [24, 30], affective learning [31–33, 23], flex-ible switching of behavior [3] and inferential learning [2, 24, 17]. A challenge tounderstanding how habitual processing can bring to bear on cognitive processes isin finding an appropriate level of description of the learning and interactive pro-cesses that underlie a given dual process system. In the context of an Associativetwo-process, habits (Stimulus-Response route) may: (i) compete with, (ii) substitutefor, (iii) facilitate, the action selection influence of goal-directed processing [29] andvice versa.

3 Habits and Knowledge Transfer in Social Learning

A classic approach to functionally framing habitual-associative neural activity hascome in the form of logical computation. Indeed, the field of Neural Networks devel-oped under the premise that layers of artificial neurons could model biological neu-rons and that their ‘all-or-nothing’ spiking activations allowed for the learning oflogic, including boolean logic, functions [34–36]. Since the implementation of thesefunctions, logic gates, are used as the fundamental units of computation in digitalcomputers, neural activity was thereby conceived as performing digital computa-tions. Furthermore, multi-layered artificial neural networks were conceived as uni-versal function approximators [37], i.e. they could, with the appropriate tuning ofthe weights and provided that sufficient (hidden) neural units are used in the model,approximate any function to arbitrary precision.

As concerns value-based habit processes, neural networks are often used as stateand action function approximators for reinforcement learning algorithms (e.g. [38,39]).Multi-layered perceptrons with soft boundary activation functions (e.g. sigmoidfunctions) whose outputs are values between 0 and 1 compute degree of certaintythat input fall within a particular class or other and thereby implement a type of

158 R. Lowe

fuzzy logic [40]. Deep neural networks (3+ hidden layers) can realize increasinglycomplex nested logic functions (fuzzy boolean functions, or Zadeh operators, [41]).Fuzzy logic thereby provides, for neural network computations, a generalizationof boolean logic to account for the non-linear activations used in deep(er) neuralnetworks. State function approximators, when implemented by neural networks,extract critical information from the environment relevant for the agent task. Deep-Q type neural networks (e.g. [42–44]) typically involve supervised training of thestate function approximator, for example, an autoencoder or convolutional neuralnetwork, whose stimulus/state outputs are then valuated with respect to the viableactions.

Notwithstanding the early enthusiasm for conceiving of artificial neural networksas logic computation models of the brain, the more dominant perspective to neuralnetwork function in modern times is to view them as capturing statistical regulari-ties among related data sets. Deep learning approaches break down ‘big data’ intofeatures that reliably recur in portions of the data. Multiple layers within the deeplearning neural network permit the capturing of feature compounds whose level ofabstraction scales with increasing receptive field depth. Nevertheless, underlying thisfeature extraction view of neural network functionality is the phenomenon of hierar-chical decomposition of the data set in order to simplify non-linear classification andregression problems. This is reminiscent of the classical approach to resolving non-linearly separable (classification) problems, e.g. when constructing neural networksfor resolving linearly non-separable input-output functions (e.g. as for the famousXOR problem).

Aside from using nested boolean logic function to approximate habit components(state, actions), dual process reinforcement learning algorithms may be imbued withpredicate logic functions. For example the dual routes (S-R, habit-based; S-E-R,outcome/goal-based) of the Associative Two-Process (ATP) model, through beingempirically validated by a set of procedures known as transfer-of-control, reveal thepotential for implicit1 transitive inference through which knowledge, in the form ofstimulus inputs and responses classifiable by the outcomes to which they are associ-ated might be transferred to novel stimulus-response contexts. By way of example,Fig. 3 schematizes a transfer-of-control scenario. As is typical for the paradigm,there are three phases. Each of these phases consists of a number of independent tri-als for learning: presentations of a stimulus, response options, and then a non-negative‘outcome’ if the correct response is chosen.

The phases break down as follows: Firstly, there is an initial instrumental learningphase where the two components (S-E and E-R) of the ‘goal-directed’ route can belearned as well as the S-R (‘habit-enabling’) route. Secondly, a pavlovian (contin-gency change) learning phase is presented where new S-E associations are made.Finally, a second instrumental phase is utilized, which uses previously experienced

1Halford et al. [70] suggests that a transitive inference process that is not based on ‘omnidirection-ality’, i.e. that ‘outputs’ of an otherwise feedforward process can be used to infer ‘inputs’ is onlyimplicit and demonstrative of a lower order cognitive process.


Fig. 3 Transfer-of-control paradigm. The conditioning consists of three phases: Phase 1 (Discrim-ination Training)—an initial instrumental phase where different stimulus-response (S-R) pairings(S1-R1, S2-R2) yield different outcomes (O1, O2); Phase 2 (Pairing)—a pavlovian learning phasewhere new stimuli are presented and associated with previously experienced outcomes; Phase 3(Transfer Test)—an instrumental transfer phase where the stimuli from phase 2 are re-presentedas are the response options from Phase 1. ATP theory predicts that responding in the transfer test(phase 3) will be based on already existing S-E and E-R associations learned from the first twophases where the theorized preferred selections (underlined Rs) are shown in the left diagram andthe S3 → R1 choice process is schematized on the right leading to rewarding outcome 1 (λ1).Diagram taken from [45]

stimuli and responses but introduces novel stimulus-response pairings.2 This servesas a test of transfer of the knowledge of the components (S-E and E-R) learned inthe first two phases that provide the relevant building blocks for the S-E-R processto select the ‘correct’ response in phase 2.

In the specific transfer-of-control example given in Fig. 3, over the first two phasesoutcomes (O1 and O2) are common to S1 and S3, and S2 and S4, respectively (giventhat in phase 1 the correct responses are made to obtain those outcomes). As a result,when Phase 3 (transfer test) occurs, since the animal/human has learned S1 andS3 according to the same outcome (O1)—that is, it has formed S1-E1 and S3-E1associations—S3 automatically cues the response associated with E1 (learned inPhase 1), in this case R1 substituting for the external stimulus. No new learningis required for this in spite of the fact that the subject has not been exposed to theparticular (external) stimulus-response pairing (S3-R1) previously.

ATP postulates, therefore, that by way of a (dual-route) structured learning pro-cess, a type of transitive inference is possible to find correct responses in the testphase without the requirement of learning. S3-R1 associations have not been learnedat the beginning of the test phase, but previous experience allows for a transitiveperformance of the form A → C (S3-R1) derived from A → B (S3-E1), B → C(E1-R1).

The above demonstrates a transfer of knowledge by way of predicate logic of theform succeeds (A, B) imbuing implicit transitive inference as it concerns individuallearning outcomes. In subsequent work [45, 46], we have postulated and empiricallytested the existence of such a form of learning applied to social settings. In [45] we

2The first and second phase of the transfer of control paradigm can, in fact, be presented in anyorder though more standardly the initial instrumental phase is used first.

160 R. Lowe

suggested that viewing the Pavlovian (second) stage as a form of passive observationin a social context would allow to test hypotheses as to whether participants cantransfer knowledge learned from their own instrumental experience (Phase 1) andalso knowledge learned within the passive social context (Phase 2) to the standardtest phase (Phase 3). Drawing on a review of how value systems are representedfor self and other in joint activities [45], we hypothesized that subjects would beable to vicariously learn the value observed for the other where the context is non-competitive. This is to say that participants experience (the presentation of) other’sstimuli and outcomes as if they were their own and update their value functionaccordingly.

Such a ‘social’ transfer of knowledge might be particularly useful during jointactivities or actions [47], particularly if the actions require fast, non-verbal move-ments [48]. In such situations, knowledge of how the cooperating agent carries outhis/her action may be of secondary utility to information concerning the stimulusand outcomes that are perceptible to the observing agent. An ATP-based transfer ofknowledge would thereby alleviate a correspondence problem—imitating another’sactions to achieve an end goal requires knowledge of how one’s own morphologymaps to another. In the above-presented scenario this would not be necessary. Takingthe example in Fig. 3, the observing agent could perceive that S3 → O1, based onwhat it can observe, affects the cooperating agent and transfer that knowledge (S3→ E1) to its own behavior (responses) according to the past relevant associationsmade (E1→ R1). The perceiving agent that can then utilize its own transitive bridge(goal-directed process; S3 → E1 → R1) to correctly infer the S3-R1 (habit-based)relation that may be utilized by the cooperating agent.

Naturally, the abovediscussionhas centred on imbuingneural networks (asmodelsof the brain) with logic computations based on feedforward mechanisms. Goal-directed behavior, however, may entail top-down effects wherein perceptual stimuli(or states) are not so much passively received but actively searched according to theparticular goals of the time.

4 Goal-Directed and Relational Knowledge Transferin Social Learning

In the previous sections, focus has been placed on feedforward neural activity of thetype that imbues associative learning and might be likened to what [49] have referredto as System 1 processing that are just one part of a dual process (System 1–System 2)cognitive system in humans. Halford et al. [50, 51] proposed that cognitive capacitiesthat cover broader dual systems, i.e. System 1-like and System 2-like processes,can be ranked according to their relational representational complexity and indexedaccording to developmental maturation. Seven ranks were posited from ranks 0 to 6inclusive.


Ranks 0 and 1 concerned feedforward neural network architectures with eitherno representations (no hidden layer) or a “functionally structured” process (singlehidden layer). These ranks were considered System 1-like. These could be viewedas covering associative (habit-based) processing, i.e. feedforward, automatic pro-cessing of responses following perception of stimuli. Ranks 2–6 on the other handentailed logico-relational representations implemented as symbolic neural networksordered from unary, e.g. Fido is a dog, to quinary, e.g. as concerns requisite represen-tations to solve the Tower of Hanoi problem. These ranks were considered System2-like. The neural networks, forming the STAR [52, 53]—Structured Tensor Ana-logical Reasoning—architecture, are, as the name suggests based on n-dimensional(n = representational rank −1) tensor representations. Halford et al. [51] suggestedthat whilst Rank 0 and 1 architectures can potentially produce logico-relational andinferential outputs, such knowledge is considered implicit as it lacks the key propertyof omnidirectionality—the ability of the network to access any relational componentor (vectorized) components from any other relational component(s).

An alternative ANN approach to STAR (to which [51] attribute omnidirection-ality) is that provided in the LISA/DORA framework [54–56]. When processingvector patterns of inputs in a neural network—that might be constitutive of seman-tic or non-semantic features—the number of higher level classes (localist neuralrepresentations) comprising these features will be limited by the extent to whichvector patterns overlap. Class-constitutive patterns with strong overlap will tend toactivate multiple classes, therefore. In the LISA/DORA architecture the problem ofcapacity limitations of the brain (representing indefinitely large numbers of relations)is partially alleviated by the use of overlapping activation of constituent elementsto predicate component representations. Activation of each sub-component (and itsconstituents) of a predicate (e.g. John loves; Mary is loved of the predicate state-ment John loves Mary) occurs one at a time. So, all constituents of John loves aresynchronously activated and this pattern inhibits theMary is loved pattern. The Johnloves pattern, through self-inhibition, loses activation and simultaneously disinhibitsthe Mary is loved pattern [57]. A high frequency oscillation (relative to the neuralstates representing full binary predicate statements) between the two patterns is saidto allow for role-filler binding (of the “sub-propositions”) so as to input a stablepattern to the compound neural representation that encodes the full predicate state-ment/proposition John loves Mary that is of representational rank 3 [51]. Role-fillerbinding constituents may overlap in their representation—they might even involvethe same objects (fillers, e.g. John,Mary) and predicates (roles, e.g. loves, is loved)but LISA/DORA exploits time (temporary representations involving re-use of com-ponents) so as to bind position-sensitive (and thereby meaning-sensitive) role-fillersto predicate statements. Hummel et al. [58] has suggested that a critical feature thatLISA (and DORA) possesses that STAR lacks is role/filler independence—that isto say that the roles (predicates), e.g. loves, and fillers (object, subject), e.g. Johnand Mary, that are used in a relational statement should not be dependent upon theparticular statement. John could be both ‘the lover’ and ‘the beloved’ and should notbe re-presented in these different roles, which is a requirement for STAR.

162 R. Lowe

The above-mentioned logico-relational systems though implemented as neuralnetworks, do not allow for the learning of features constitutive of the roles and fillersof the predicate statements (and their components). As a result these architectures arepurely offline and in themselves do not help us address how habits and goal-directedsystems in learning agents might be exploited.

In [59] we discuss how, potentially, such roles and fillers might be groundedthrough the use of parallel distributed networks that provide featural components tothe LISA/DORA architecture. The key point is that the localist-symbolic neuronsthat compute predicate statements (e.g. John loves Mary) and their constituents (e.g.John loves, Mary is loved) may interface with feedforward neural networks of thetype referred to in Sect. 3 by way of top-down generative processing. This mightbe conceived as an agent obtaining a goal based on a predicate statement (or setthereof), e.g. box covers cup, wherein the agent then searches for a box and a cup insequence. Active search for the objects required to achieve the goal might even refinethe learning of the stimulus features or nested logic computations [60] such that top-down and bottom-up (feedforward) neural processing are inseparable. In this sense,components, at least, of habits (stimuli and action representations) may be partlyconstructed as a result of forming goal-directed predicate statements. Goal-directedpredicate knowledge may also facilitate social transfer of knowledge where agentsare explicitly aware of the relations between stimuli, responses and outcomes. Forexample, in reference to the social transfer-of-control example given in the previoussection, the cooperating agent might produce exaggerated emotional responses tooutcomes to facilitate knowledge transfer, e.g. if those outcomes are only partially,or not at all, observable to the observing agent in the joint task.

Empirical tests that serve to tease apart the effects on the logical constructs ofhabit-based an goal-directed based systems of bottom-up/feedforward and top-downactivations require hypotheses concerning how behavioural performance (or neu-rophysiological activity) should be affected by emphasis on one or other process(or system). By way of example, [61] showed that their associative based learningmodel could capture data from a paradigmatic learning task designed to require atransitive inference solution. Their model, nevertheless, predicted a profile of choiceperformance consistent with exposure to reinforced outcomes, i.e. a non-transitiveinferencememory-based solution. Notwithstanding, profiles of performance differedin subjects who reported understanding of the rules of the task suggesting that amore relational strategy was being employed in these cases (one which apparentlyimproved performance). Such differing profiles of behavioural performance provideclues as to which type of learning system is being used, e.g. the slower, more habitualprocess, would typically manifest in a different behavioural performance than themore goal-directed process that places certain expectations on outcomes. Experimen-tal procedures able to tease out how and when system 1 and system 2 like knowledgeis used provide a key means for furthering understanding of how associative learningbrings to bear on relational cognition and vice versa. In [46]we put forward a researchagenda for teasing out such effects of the two purported systems of logico-relationalknowledge in social settings for both human-human and human-robot interaction.


5 Conclusion

In this chapter, we have considered habit-based and goal-directed learning systems,how they may be implemented using neural networks and the sense in which suchthese neural networks produce logic-based computation.We have also discussed howhabit-based and goal-directed systems might combine in the service of transfer ofknowledge (based on logic computations) from one context to another or betweenindividuals (social) participating in joint activities.

It was mentioned that viewing brain activity, when learning habits or processingaccording to a goal-directed system, as carrying out logic computation is not nec-essarily the fashionable perspective. The influential Deep learning perspective onneural network computation emphasizes the role of feature extraction a means ofseparating data into classes. A further contention regarding the link between logicand neurobiological activity is the notion of what constitutes the basic unit of infor-mation. McCulloch and Pitts [34] alluded to ‘all-or-nothing’ spikes of single neuronsproviding digital values amenable to logic gate computation. However, neural infor-mation conveyance may be better understood in terms of spike trains, i.e. severalspikes occurring over a particular time window [62]. A single spike may have littleimpact on a neuron’s activation to which it projects as this neuron will synapse within the order of 10,000 other neurons a subset of which projecting output at any giventime. Thereby, neural activation might be considered closer to analog than digitalsince spike trains can be graded regarding their number and the patterns that theyconstitute. Furthermore, even if activation patterns are in themselves to be consideredas all-or-nothing phenomena, and that brains could be, in principle, at this level ofabstraction implemented by a combination of neural logic gates, it does not meanthat this is how the brain would operate or that it would constitute an efficient meansof operation as tapped by evolution.

Irrespective of the universal relevance to brain functioning, the notion that thebrain (at least sometimes) produces logic computations has gained some groundin recent years. In relation to deep neural networks, for example, embedding logiccomputations as a form of modularized prior knowledge has been shown to improveperformance of deep belief networks under certain conditions [63]. The use of binaryweights as well as input values, permitting logic functions, has also been deployed indeep convolution networks—a so-called XNOR-net—providing a means for moreefficient and computationally less costly processing [64]. Relatedly, in relation todeep belief networks, [65] have found similar benefits in using (sparse) binaryconnections.

In standard neural networks, apprehending neural network function in terms oflogic computation depends on the extent to which weights between neurons can belearned that serve to approximate a logic-gate input-outputmapping. The dependenceof such logic gate ‘emergence’ may, in many neural ensembles, also be dynamicallysensitive in other ways. Work by [66]—see also [67]—has shown that in biologi-cal neurons, particular circuits can transition from one logic gate implementation to

164 R. Lowe

another subsequent to a history of activation of the neurons. To provide a simpli-fied description, the decay parameter increases as a function of stimulation, i.e. thestimulus-response delay increases, and the results of this is that concurrent activa-tion of ‘input’ neurons may become more or less synchronized at an output neuronleading to transitions from OR to XOR and back to OR, for example. The research,stimulating neurons in vitro, led to the uncovering of all manner of dynamic logic-gates including generalized (more than two-input) logic gates. Interesting questionsare put forward by the authors as concerns the use of dynamic logic-gate for buildingmore complex (logical) structures that would give rise to cognitive phenomena evenimplementing, and going beyond, Universal Turing Machines [66].

The trend for producing ever large Deep neural networks to meet benchmark per-formance but to deprioritize transparency of process represents a potential bottleneckfor progress for modern Artificial Intelligence developments insofar as they can helpexplain natural cognition. Viewing deeper networks according to the types of logiccomputations that they can produce in the service of cognitive-behavioural phenom-ena such as habits, goal-directed activity, affective systems [68, 69] and the socialtransfer of knowledge, may be of great long term importance to imbuing artificialagents with more general cognitive capabilities.

References

1. Seger,C.A.:Howdo thebasal ganglia contribute to categorization?Their roles in generalization,response selection, and learning via feedback. Neurosci. Biobehav. Rev. 32, 265–278 (2008)

2. Phillips, S., Wilson, W.H., Halford, G.S.: What do transitive inference and class inclusion havein common? Categorical (co) products and cognitive development. PLoS Comput. Biol. 5(12)(2009)

3. Seger, C.A., Spiering, B.J.: A critical review of habit learning and the basal ganglia. Front.Syst. Neurosci. 5 (2011)

4. Davidson, R.J., Irwin, W.: The functional neuroanatomy of emotion and affective style. TrendsCognit. Sci. 3(1), 11–21 (1999)

5. Dickinson, A.: Actions and habits: the development of behavioural autonomy. Philos. Trans.R. Soc. Lond. B Biol. Sci. 308, 67–78 (1985)

6. Hirsh, J.: The hippocampus and contextual retrieval of information from memory: a theory.Behav. Biol. 12, 421–444 (1974)

7. Seger, C.A.: The involvement of corticostriatal loops in learning across tasks, species, andmethodologies, In: Groenewegen, H.J., Voorn, P., Berendse, H.W., Mulder, A.B., Cools, A.R.(eds.) The Basal Ganglia IX, pp. 25–39. Springer, New York (2009)

8. Daw,N.D., O’Doherty, J.P.:Multiple systems for value learning. In: Neuroeconomics: DecisionMaking and the Brain, 2nd edn, pp. 393–410. Academic Press, London (2014)

9. Miller, K.J., Shenhav, A., Ludvig, E.A.: Habits without values. Psychol. Rev. (2019)10. Lowe, R., Ziemke, T.: Exploring the relationship of reward and punishment in reinforcement

learning. In: 2013 IEEE Symposium on Adaptive Dynamic Programming and ReinforcementLearning (ADPRL), pp. 140–147 (2013)

11. Lee, G., Lowe, R., Ziemke, T.: Modelling early infant walking: testing a generic CPG archi-tecture on the NAO humanoid. In: 2011 IEEE International Conference on Development andLearning (ICDL), vol. 2, pp. 1–6 (2011)


12. Li, C., Lowe, R., Ziemke, T.: Crawling posture learning in humanoid robots using a natural-actor-critic CPG architecture. In: Artificial Life Conference Proceedings, 13, pp. 1182–1190.MIT Press, Cambridge, MA (2013)

13. Li, C., Lowe,R., Ziemke, T.:Anovel approach to locomotion learning:Actor-Critic architectureusing central pattern generators and dynamic motor primitives. Front. Neurorobotics 8, 23(2014)

14. Rolls, E.T.: On the brain and emotion. Behav. Brain Sci. 23(2), 219–228 (2000)15. Suri, R.E., Schultz, W.: Learning of sequential movements by neural network model with

dopamine-like reinforcement signal. Exp. Brain Res. 121(3), 350–354 (1998)16. Suri, R.E., Schultz, W.: Temporal difference model reproduces anticipatory neural activity.

Neural Comput. 13(4), 841–862 (2001)17. Peterson, G.B., Trapold, M.A.: Effects of altering outcome expectancies on pigeons’ delayed

conditional discrimination performance. Learn. Motiv. 11, 267–288 (1980)18. Daw, N.D., Niv, Y., Dayan, P.: Uncertainty-based competition between prefrontal and

dorsolateral striatal systems for behavioral control. Nat. Neurosci. 8(12), 1704 (2005)19. Sun, R.: Interpreting psychological notions: a dual-process computational theory. J. Appl. Res.

Memory Cognit. 4(3), 191–196 (2015)20. de Wit, S., Dickinson, A.: Associative theories of goal-directed behaviour: a case for animal–

human translational models. Psychol. Res. PRPF 73(4), 463–476 (2009)21. Mowrer, O.H.: On the dual nature of learning: A reinterpretation of “conditioning” and

“problem-solving”. Harv. Educ. Rev. 17, 102–148 (1947)22. Trapold,M.A.: Are expectancies based upon different positive reinforcing events discriminably

different? Learn. Motiv. 1, 129–140 (1970)23. Cardinal, R.N., Parkinson, J.A., Hall, J., Everitt, B.J.: Emotion and motivation: the role of the

amygdala, ventral striatum, and prefrontal cortex. Neurosci. Biobehav. Rev. 26(3), 321–352(2002)

24. Urcuioli, P.: Behavioral and associative effects of differential outcomes in discriminatinglearning. Learn. Behav. 33(1), 1–21 (2005)

25. Lowe, R., Humphries, M., Ziemke, T.: The dual-route hypothesis: evaluating a neurocompu-tational model of fear conditioning in rats. Connect. Sci. 21(1), 15–37 (2009)

26. Lowe, R., Billing, E.: Affective-associative two-process theory: a neural network investigationof adaptive behaviour in differential outcomes training. Adapt. Behav. 25(1), 5–23 (2017)

27. Lowe, R., Almér, A., Billing, E., Sandamirskaya, Y., Balkenius, C.: Affective–associative two-process theory: a neurocomputational account of partial reinforcement extinction effects. Biol.Cybern. 111(5–6), 365–388 (2017)

28. Braver, T.S., et al.: Mechanisms of motivation–cognition interaction: challenges and opportu-nities. Cognit. Affect. Behav. Neurosci. 14(2), 443–472 (2014)

29. Urcuioli, P.: Some relationships between outcome expectancies and sample stimuli in pigeons’delayed matching. Anim. Learn. Behav. 18(3), 302–314 (1990)

30. Urcuioli, P.: Stimulus control and stimulus class formation. In: Madden, G.J., Dube, W.V.,Hackenberg, T.D., Hanley, G.P., Lattal, K.A. (eds.) APA Handbook of Behavior Analysis, vol.1, pp. 361–386. American Psychological Association, Washington, DC (2013)

31. Amsel, A.: The role of frustrative nonreward in noncontinuous reward situations. Psychol. Bull.55, 102–119 (1958)

32. Amsel,A.: FrustrationTheory:AnAnalysis ofDispositional Learning andMemory.CambridgeUniversity Press, Cambridge (1992)

33. Overmier, J.B., Lawry, J.A.: Pavlovian conditioning and the mediation of behavior. Psychol.Learn. Motiv. 13, 1–55 (1979)

34. McCulloch, W.S., Pitts, W.: A logical calculus of the ideas immanent in nervous activity. Bull.Math. Biophys. 5, 115–133 (1943)

35. Rosenblatt, F.: The perceptron: a probabilistic model for information storage and organizationin the brain. Psychol. Rev. 65, 386 (1958)

36. Rumelhart, D.E., Hinton, G.E., McClelland, J.L.: A general framework for parallel distributedprocessing. In: ParallelDistributedProcessing:Explorations in theMicrostructure ofCognition,vol. 1, pp. 45–76 (1986)

166 R. Lowe

37. Hornik, K., Stinchcombe, M., White, H.: Multilayer feedforward networks are universalapproximators. Neural Netw. 2(5), 359–366 (1989)

38. Navarro-Guerrero, N., Lowe, R.J.,Wermter, S.: The effects on adaptive behaviour of negativelyvalenced signals in reinforcement learning. In: 2017 Joint IEEE International Conference onDevelopment and Learning and Epigenetic Robotics (ICDL-EpiRob), pp. 148–155 (2017)

39. Navarro-Guerrero,N., Lowe,R.J.,Wermter, S.: Improving robotmotor learningwith negativelyvalenced reinforcement signals. Front. Neurorobotics 11, 10 (2017)

40. Wierman, M.J.: An Introduction to the Mathematics of Uncertainty, 149–150. CreightonUniversity (2010)

41. Zadeh, L.A., et al.: Fuzzy Sets, Fuzzy Logic. World Scientific Press, Fuzzy Systems (1996)42. Lange, S., Riedmiller, M., Voigtländer, A.: Autonomous reinforcement learning on raw visual

input data in a real world application. In: The 2012 International Joint Conference on NeuralNetworks (IJCNN), pp. 1–8 (2012)

43. Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A.A., Veness, J., Bellemare, M.G., … Petersen,S.: Human-level control through deep reinforcement learning. Nature 518(7540), 529 (2015)

44. Yang, Z., Xie, Y., Wang, Z.: A theoretical analysis of deep Q-learning. arXiv preprint arXiv:1901.00137 (2019)

45. Lowe, R., Almer, A., Lindblad, G., Gander, P., Michael, J., Vesper, C.: Minimalist social-affective value for use in joint action: a neural-computational hypothesis. Front. Comput.Neurosci. 10 (2016)

46. Lowe, R., Almér, A., Gander, P., Balkenius, C.: Vicarious value learning and inference inhuman-human and human-robot interaction. In: 2019 8th International Conference onAffectiveComputing and Intelligent Interaction Workshops and Demos (ACIIW), pp. 395–400. IEEE(2019)

47. Bratman, M.E.: Shared cooperative activity. Philos. Rev. 101, 327–341 (1992)48. Vesper, C., Butterfill, S., Knoblich, G., Sebanz, N.: A minimal architecture for joint action.

Neural Netw. 23, 998–1003 (2010)49. Kahneman, D., Tversky, A.: Rational choice and the framing of decisions. J. Bus. 59(4), 251–

278 (1986)50. Halford, G.S., Wilson, W.H., Phillips, S.: A conceptual complexity metric based on repre-

sentational rank. Psychology Computer Science & Engineering, University of QueenslandUniversity of New South Wales, Queensland 4072, Australia; Sydney, NSW 2052 Australia(1999)

51. Halford, G.S., Wilson, W.H., Andrews, G., Phillips, S.: Categorizing Cognition: TowardConceptual Coherence in the Foundations of Psychology. MIT Press (2014)

52. Halford, G.S.,Wilson,W.H., Phillips, S.: Processing capacity defined by relational complexity:implications for comparative, developmental, and cognitive psychology. Behav. Brain Sci.21(6), 803–831 (1998)

53. Halford, G.S., Wilson, W.H., Phillips, S.: Relational knowledge: the foundation of highercognition. Trends Cognit. Sci. 14(11), 497–505 (2010)

54. Hummel, J.E., Holyoak, K.J.: Distributed representations of structure: a theory of analogicalaccess and mapping. Psychol. Rev. 104(3), 427 (1997)

55. Hummel, J.E., Holyoak, K.J.: A process model of human transitive inference. In: SpatialSchemas in Abstract Thought, pp. 279–305 (2001)

56. Doumas, L.A.,Morrison, R.G., Richland, L.E.: Individual differences in relational learning andanalogical reasoning: a computational model of longitudinal change. Front. Psychol. 9 (2018)

57. Knowlton, B.J., Morrison, R.G., Hummel, J.E., Holyoak, K.J.: A neurocomputational systemfor relational reasoning. Trends Cognit. Sci. 16(7), 373–381 (2012)

58. Hummel, J. E., Holyoak, K. J., Green, C., Doumas, L. A., Devnich, D., Kittur, A., Kalar, D.J.: A solution to the binding problem for compositional connectionism. In: Levy, S.D., Gayler,R. (eds.) Compositional Connectionism in Cognitive Science: Papers from the AAAI FallSymposium, pp. 31–34 (2004)

59. Lowe, R. Almér, A., Balkenius, C.: Bridging connectionism and relational cognition throughbi-directional affective-associative processing. Open Inf. Sci. 3, 235–260 (2019)


60. Hinton, G.E., Salakhutdinov, R.R.: Reducing the dimensionality of data with neural networks.Science 313(5786), 504–507 (2006)

61. Frank, M.J., Rudy, J.W., Levy, W.B., O’Reilly, R.C.: When logic fails: implicit transitiveinference in humans. Memory Cognit. 33(4), 742–750 (2005)

62. Amari, S.I.: Dynamics of pattern formation in lateral-inhibition type neural fields. Biol. Cybern.27(2), 77–87 (1977)

63. Tran, S.N., Garcez, A.S.D.A.: Deep logic networks: inserting and extracting knowledge fromdeep belief networks. IEEE Trans. Neural Netw. Learn. Syst. (2016)

64. Rastegari, M., Ordonez, V., Redmon, J., Farhadi, A.: Xnor-net: Imagenet classification usingbinary convolutional neural networks. In: European Conference on Computer Vision, pp. 525–542. Springer International Publishing (2016)

65. Zhou, X., Li, S., Qin, K., Li, K., Tang, F., Hu, S., … & Lin, Z.: Deep adaptive network: anefficient deep neural network with sparse binary connections. IEEE Trans. Neural Netw. Learn.Syst. (forthcoming)

66. Goldental, A., Guberman, S., Vardi, R., Kanter, I.: A computational paradigm for dynamiclogic-gates in neuronal activity. Front. Comput. Neurosci. 8 (2014)

67. Vardi, R., Guberman, S., Goldental, A., Kanter, I.: An experimental evidence-based computa-tional paradigm for new logic-gates in neuronal activity. EPL 103, 66001 (2013)

68. Kiryazov, K., Lowe, R., Becker-Asano, C., Randazzo, M.: The role of arousal in two-resourceproblem tasks for humanoid service robots. In: 2013 IEEE ROMAN, pp. 62–69 (2013)

69. Lowe,R., Philippe, P.,Montebelli,A.,Morse,A., Ziemke,T.:Affectivemodulationof embodieddynamics. In: The Role of Emotion in Adaptive Behaviour and Cognitive Robotics, ElectronicProceedings of SAB Workshop (2008)

70. Halford, G.S., Andrews, G., Wilson, W.H., Phillips, S.: Computational models of relationalprocesses in cognitive development. Cogn. Dev. 27(4), 481–499 (2012)

Agency of Breath—Beyond DisciplinaryViews on Ritual

Anna Kawalec

Abstract In this article, I focus on a particular ritual that is constitutive for a widerange of social practices. I examine it with the intention to develop a conception ofritual that goes beyond traditional patterns and disciplinary boundaries. The proposedconception of ritual is meant to reveal interdependence between the realm of humanexistence and actions, on the one hand, and the realm of the world that is autonomouswith regard to humans, on the other. The conception of ritual as advanced here isneither subsumed to the dominant categories of Western culture, nor derived fromsophisticated theoretical discussions. For the purpose at hand, I will refer to thisconception as “folk”. However, because the proposed approach has the ambitionof universal validity, I examine selected categories and concepts of philosophy andanthropology to substantiate my case: (i) the etymological meanings in Sanskrit,(ii) the Hebrew origins, (iii) the ethnographic facts (on the basis of Alfred Gell’sresearch on Muria’s “God playing” and some selected theses of cultural and anthro-pological theory), (iv) Jerzy Grotowski’s idea of techniques of the performer’s breathand (v) somepoints fromKarolWojtyła’s philosophical integrative viewof the humanperson and agency.

Keywords Etymology of ritual · Performer’s breath · Hylemorphism · Logic ofritual · Alfred Gell · Jerzy Grotowski · Karol Wojtyła

1 Introduction

In this article, I intend to formulate a conception of ritual that goes beyond tradi-tional patterns and disciplinary boundaries [1]. The conception of ritual as advancedhere is neither subsumed to the dominant categories of Western culture, nor derivedfrom sophisticated theoretical discussions [12]. For the purpose of this article, I willrefer to it “folk”, although I claim that it has a much wider applicability and is morecomprehensively functional than theoretical speculations determined by the subject

A. Kawalec (B)Faculty of Philosophy, John Paul II Catholic University of Lublin, Lublin, Polande-mail: [email protected]


169

170 A. Kawalec

matter and methodology of science framed by the confines of the Western world[2]. However, because the proposed approach has the ambition of universal valid-ity, selected categories and concepts of philosophy and anthropology will serve asmaterial for the course of my investigation (Sect. 2).

One of the most entrenched and, at the same time, schematic Western categoriesis the opposition “nature versus culture” [4]. Contemporaneity proliferates with thetrends of cultural thought, forcing the thesis of the absolute universality of humaninfluence and outputs. Humans interfere with the microcosm and the cosmos andtransform their own bodies, sex and mentality, leaving traces of this interferenceas “culture”. Moreover, humans are convinced that they have developed the poten-tial to transform everything that exists in the surrounding world. This convictionis further enhanced by innovative technological advances, recurrent scientific dis-coveries and theories [13] as well as shocking forms of art [29]. Modern Westernpeople attribute to “culture” the power and governance structure of the entire world.“Culture” thus fulfils an ideological function, motivating the exceptional activitiesof Western humans, who consider themselves the autonomous creators of “culture”,which is also ingrained in everyday attitudes of Western societies.

The schematic pair of categories “nature-culture” (also in the version “nature isculture”) delineates the conceptual framework within which the Western concept ofritual is located. However, I will argue that although it has become (especially owingto anthropology) the integral component of Western culture, it was not created byour culture. What matters is that “ritual” is not merely a concept but is the locusof injunction of a law of nature and the phenomenon that the people of Westerncivilization call “culture”, although its sources stem from a realm that transcendshuman reality. It is against this background that I attempt to demonstrate that ritualis the acting law of nature,1 on the assumption that nature means a primordial, priorand autonomous realm with regard to human acts of transformation.

I advancemy argument by elaborating on the claims of anthropologists, especiallyAlfred Gell, as well as the concept of agency created by the philosopher KarolWojtyła. Gell proposes to focus on the fundamental dimensions of human life (bodily,mental, social), removing abstract issues as “cultural”.Wojtyła, in contrast, examinesthe basic dimension of personal life: the ontological integrity of the person on thebasis of interacting psychosomatic determinations.

The starting point of my discussion of “ritual as the acting law of nature” is theassumption that the only law of nature is the one that determines the order of thematerial and immaterial (the “invisible”—irrespective of the research tools used)world, and the place of injunction of the material and immaterial dimensions of theworld is constituted by breath of acting beings. My holistic (mainly anthropological)point of view emphasizes the thesis that the law of nature is a primary and uniquelaw of the world, in which human being is embedded. The obviousness of this thesistoday is radically undermined by the influence of the Western conceptions of thenature of human being and the behavioral attitudes of the Western communities [9].

1The term “acting” is used here in analogy with its meaning in K. Wojtyła’s title phrase “the actingperson”.

Agency of Breath—Beyond Disciplinary Views on Ritual 171

The attitudes of non-Western societies or those that distance themselves fromWestern attitudes2 share a fundamental connection with the thought style that couldbe referred to in the Western terminology as “naturalistic”. This term would not beclear, however, without noting that the concept of nature in the non-Western worldincludes the above mentioned aspects: the material and immaterial dimensions ofthe world as well as the world of values. The “naturalism” of their perception of theworld, however, consists of acceptance of the primacy of the world of nature andconsistent life attitudes, particularly respect for nature [20, 30].

To delineate this complex issuewithout imposing overly theoretical presumptions,for the purposes of “folk thinking”, I refer to simple images of the world of theAmerican Indians. In everyday life, the Indians are peaceful people, living accordingto the laws of their pueblo. They move gently, flexibly and surprisingly deliberately.They are immersed in activities they perform: when they look, they look; whenthey give a hand, they give a hand; when they sit down, they sit down; and whenthey look at the world around them, they see the creator. Most importantly, eachaction is always envisioned with the same purpose: to sustain spiritual connectednesswith the surrounding world. This holistic attitude permeates every aspect of theirlives: when they ponder how rocks became, the answer is immediately forthcoming,and their attitude and gaze express their pride and gratitude for every moment oftheir existence rooted in the fraternal environment. Not by chance, one of today’sindigenous sculptors, Allan Houser, carved out the proud attitude and face of theIndian (Fig. 1), as in the series “Warm Springs Chiricahua Apache” and the sculptureReady to Dance (1988, Museum Hill, Santa Fe).

The following are some examples of the holistic and connected attitudes expressedin the words of the representatives of pueblos, for example, concerning a house:“Made of stone and mud, the foundation, the four corners. Made of earth and wood,the roof, the center. All anchored with prayers” ([17], p. 46). A composer and musi-cian of the Hopi tribe expressed the purpose of his activities as follows: “As a noviceHopi composer, I am consistently reminded that the words of a song are a humblepetition to the spiritual forces for their blessings. My father and grandfather taughtme that my corn plants are my children. By singing to the corn, I help them tomature.Indeed, the best place to contemplate and compose a song is in the solitude of mycornfield. The women would sing grinding songs to the rhythm of the metate andmano as they ground the corn into flour. Our traditions say that the stones becomehappy and honored when they hear the grinding songs” ([17], p. 43). Similarly,“native communities are about connections because relationships form the whole.Each individual becomes part of the whole community, which includes the hills,mountains, rocks, trees, and clouds” ([17], p. 45).

The above assumptions and outlooks invalidate the traditionalWestern oppositionof nature versus culture. In contrast, they presume inherent spiritual connectednessof human beings with their natural environment that is subsumed under the laws ofnature that humans are to observe. The theoretical premise of the existence of the

2I apply this term also to the ethnic groups that share the territory under the influence of theWesternworld. I particularly refer to the selected pueblos in New Mexico, US.

172 A. Kawalec

Fig. 1 Allan Houser’s sculpture of Indian (Santa Fe) Photograph Anna Kawalec

world of nature governed by autonomous laws is followed by the normative principleof the adequate conduct of beings belonging to this world. This order is essential andnot subject to change by humans.

2 “Ritual”—Reality and Disciplines

In this section I overview selected ideas of ritual. However, their relevance is notregarded within particular research disciplines, as it apparently would presume oneof the orders of Western culture, but rather they pertain to a new understandingof ritual advocated here. Thus, I propose to accordingly modify the meaning of“ritual” adopted in Western studies on culture towards a more universal, “realis-tic” one. My argument in the reminder of this section focuses on breath as thefundamental human ritual. I use this paradigmatic example to modify the estab-lished meaning of “ritual” by referring to (i) the etymological meanings in Sanskrit,(ii) the Hebrew origins, (iii) ethnographic facts and conceptions (on the basis ofAlfred Gell’s research on Muria’s “God playing” and some selected theses of cul-tural and anthropological theory), (iv) Jerzy Grotowski’s idea of techniques of theperformer’s breath and (v) some ideas ofKarolWojtyła’s philosophical interpretationof person and human agency in the context of integration of human beings. Despiteapparent diversity of these sources, they share a common feature, namely they allindicate the universal character of breath as the fundamental human ritual.


2.1 Etymological Meanings in Sanskrit

In the Vedic religion, R. ta (Sanskrit r. tam. , which means order, truth) is the rule ofnatural order. R. ta manages the operation of the universe in the micro and macrodimensions and manages the functions of holistic order: the natural, moral and sac-rificial. In the hymns of the Vedas, R. ta, as “one of the most important religiousconceptions of the Rig Veda”, indicates that “from the point of view of the history ofreligious ideas we may, in fact we must, begin the history of Hindu religion at leastwith the history of this conception” ([3], pp. 12–13). R. ta is derived from the Sanskritverb root r. , which means “to go, move, rise, tend upwards”, and the derivative nounr. tam is defined as “fixed or settled order, rule, divine law or truth”. R. ta is translatedliterally as “that which has moved in a fitting manner”, abstractly as “universal law”or “cosmic order”, or simply as “truth”.3 In the context of Vedic religion, the featuresof nature that either remain constant or occur on a regular basis were seen to be amanifestation of the power of R. ta in the physical cosmos [18]. In the human sphere,R. ta was understood to manifest itself as the imperative force behind both the moralorder of society and the correct performance of Vedic rituals.4 The Sanskrit sourcepoints to the fundamental role of the natural order of universe, while human lawssupervene on that order.

2.2 Hebrew Practical Origins

The sources of Hebrew orthopraxis, which also influenced Christianity, include someinteresting practical indications5:

1. Take a breath and breathe it out.2. Do it again, slowly, and try to mean it.3. Breathing—of all things, maybe we take it most for granted.

The question, “Do we ever wonder why we are built this way, this soft machineof ours always pumping oxygen in and out? In sadness, we breathe heavy sighs. Injoy, our lungs feel almost like they will burst. In fear, we hold our breath and have tobe told to breathe slowly to help us calm down. When we’re about to do something

3“AŠA, “truth” in Avestan, from Indo-Iranian *r. tá-, a neuter noun having the same meaning. Theword is attested in Old Persian as a. rta and in Old Indian as r. tá-”. See ([24], pp. 168–182): (1) *r. támman- “to think (of) truth” (Y. 31.19 and Rigveda); (2) *r. tám vazh- “to drive the truth” (Y. 46.4and Rigveda): “The religious poet drives the true words of his hymn as the charioteer the horses”.For more information see http://www.iranicaonline.org/articles/asa-means-truth-in-avestan. (Lastaccessed 17 November 2017).4Ibidem.5https://rabbitroom.com/2011/08/is-the-name-of-god-the-sound-of-our-breathing/. (Last accessed10 April 2017).

174 A. Kawalec

hard, we take a deep breath to find our courage. When I think about it, breathinglooks almost like a kind of praying”.6

Hebrews ask, “What if the name of God is the sound of breathing? ‘In-out’”.They conclude, “How generous of God to choose to give himself a name that wecan’t help but speak everymomentwe’re alive.All of us, always, everywhere,waking,sleeping, with the name of God on our lips. When a baby is born—newly arrivedon planet earth, must they take their first breath, or rather speak the name of Godif they are to be alive here? On our deathbed, do we breathe our last breath? Or isit that we cease to be alive when the name of God is no longer on our lips?” Forthem, each negation of the existence of God affirms his existence. God is everywhereand always, in the universe and inside me: “Breathe in. Breathe out. ‘He does ourpraying in and for us, making prayer out of our wordless sighs… the word that savesis right here, as near as the tongue in your mouth’…” (Romans 8:28, 10:8, TheMessage). These assumptions are not confined to the Hebrew religious traditions—they form the foundation of religious experiences of other religious communities,such asChristians and the aforementioned Indian.They express the constitutive lawoftheir existence, nonseparable from the laws of nature as materially and symbolicallymanifested in the ritual breath.

2.3 Ethnographic Facts and Conceptions

Richard Schechner propounds the cultural and performative meaning of ritual draw-ing upon the accomplishments of Victor Turner [27] and Milton Singer [25]. As aresult, Schechner depicts the evolution of ritual and its genre as a tree [22]. Branchesand twigs mean ritualization among people: social ritual (everyday life, sport, pol-itics), religious ritual (rites, liturgy), esthetic ritual (codified forms, ad hoc forms);branches—social ritual (primateswithout humans); boughs—genetic and free behav-ior (birds, mammals); trunk—genetically determined behavior (fish, insects), androots—ritualization.

Schechner thus localizes ritual between evolutional origins and social activities.He concludes that every activity of the elements of the symbolic tree of nature ofritualized behaviour expresses the law of ritualization. However, we do not knowthe genre of this law. For Schechner at that time of writing The Future of Ritual itmay have been a social rule. For the later Schechner, however, it may have been aperformative rule,7 according to his words: “They [rituals] are ambivalent symbolicactions pointing at the real transactions even as they help people avoid too directa confrontation with these events. Thus rituals are also bridges—reliable doings

6Ibidem.7See [23]. “Being” is existence itself. “Doing” is the activity of all that exists, from quarks tosentient beings to supergalactic strings. “Showing doing” is performing: pointing to, underlining,and displaying doing. Explaining ‘showing doing’“ is performance studies” (p. 28).


carrying people across dangerous waters”.8 The law of ritualization, according toSchechner, concerns the symbolic sphere of human action, while its function is asocial adaptation.

Roy A. Rappaport pointed to the significance of ritual that transcends the realmof the social world [21]. Initially, he understood ritual as “the performance of moreor less invariant sequences of formal acts and utterances not entirely encoded by theperformers” [21] and “nothing more than a label stretched to cover phenomena ofsuchdiversity” ([21], p. 25). Later on, however, hemodified it by connecting ritual andecology. Rappaport’s appreciation of the objectively existing ecological environmentwas a consequence of the appreciation that human rituals (especially religious ones)fulfill an important function in human adaptation to the natural environment ([21],pp. 406–425).

A naturalistic interpretation of ritual was inspired by the anthropologist VictorTurner. In his work Body, Brain, and Culture [28],9 he refers to the standpoint ofJulian Huxley [14], who emphasized the value of the link between the environmentand the biology of human beings. He wrote, “The cultural rituals which seem mostto embody something resembling Huxley’s definition of ‘ritualization’ are seasonal,agricultural, fertility, funerary, and healing ones, because theymake explicit the inter-dependence of people with their physical environments and bodies”.10 For Turner, ashe concludes, “The meaning of that living macrocosm may not only be found deepwithin us but also played from one mind to another as history goes on-with ever finertuning-by the most sensitive and eloquent instrument of Gaea the Earth-spirit-thecerebral organ” ([28], p. 243).

Turner’s naturalistic interpretation of ritual would find its natural counterpart inthe field of neurobiological investigations of human ritual behaviors. Note that itfocuses on the role of endorphins, which affect the quality and intensity of ourexperiences. They activate after hard work, at night, or during rhythmic music ordance. The principal function of endorphins is to inhibit the transmission of painsignals. Theymay also produce a feeling of euphoria very similar to that produced byother opioids. Animal and human endorphins are endogenous opioid neuropeptidesand peptide hormones, and the central nervous system produces them. Endorphinsare biological elements of living organisms and point to human relationships withtheir environments.

Following Turner, one could argue that endorphins constitute the intersection ofthe biological with the cultural. Turner’s interpretation does not, however, underlinethe acting role of the law of nature; rather, it inscribes rituals into the naturalisticconception of the world which presumes the law to constitute a self-evolving mecha-nism. Neither endorphins nor Turner’s conception of ritual (1983) explain the agencyof acting law of nature.

8([22], p. 223). For more on the subject as the agent of the performative act, see [15].9Turner’s idea is further elaborated in [5]. Interestingly, the paper has the same title as Turner’s butdoes not list it among the references.10Turner follows here [10].

176 A. Kawalec

2.4 Marilyn Strathern’s Idea of Constructivism: “CultureAbove Nature”

The Western conceptions of ritual, as well as of nature, tend to emphasize the domi-nant role of human beings, and recognize ritual itself as amere expression of a humanor social order. Marylin Strathern perceives British sociality as the constructive factin her monograph After Nature: English Kinship in the Late Twentieth Century [26].While she initially formulates a claim about society as the construction of the worldand person, the main thesis of the book is conveyed by the formula, “The nature isthe construction of the human being” ([26], pp. 192–193). According to Strathern,there is currently a person-social construct and the family as the social construct;these are the subject of anthropology and social science.

Strathern seems to have committed the error suggested by Gell for Durkheim [7]:she redefines social constructions, in this case leading to culture. Moreover, it seemsthat she reverses the order of social relations (for which the ontological basis is realpersons) and their products or their institutional forms [12].

2.5 Alfred Gell’s Explication of Ritual

Perhaps the nearest approximation to the idea of ritual as the acting law of nature,propounded in this paper, can be found in the publications of a British anthropologistAlfred Gell concerning “swinging gods”. He observed that at least some of thetraditions we are already acquainted with contain the motifs of dancing gods orswinging and rocking deities (Fig. 2). Moreover, this theme is common to all Hindutraditions since at least 900 A.D. and its entrenchment is manifested by cultivationof a festival at which the gods are ritually “swung”. This festival lasts approximately10 days, held in May/June, and takes place in the middle of territory of India, socalled “pen karsana” ([6], pp. 219–248).

Among the modes of “play” there is “hook swinging”, comparable to otherforms of self-mortification (flagellation with chains, piercing of the body, pickingup the eyes, fire-walking). All swinging relates to the experience of vertiginous play(between bodily equilibrium and disequilibrium and states of consciousness). Inall swinging, there is a common element of self-surrender to a loss of individualequilibrium. It seems that human beings thereby imitate the behavior of their tradi-tional gods. During swinging, gentle rocking induces peaceful repose, and energeticrocking motions induce exhilaration, ecstasy and compulsion. Thus, the swing is anartefact whose use is mainly for the modification of mental states.

Gell advanced an original explication of the tradition of swinging. There are threeinterpretations of allied states (trans states) of possessed persons: (1) generalized“sensory overload”, (2) chemical changes in the body as a result of over-breathingand (3) the result of rhythmic stimulation by sound or light pulses at a frequencyapproximating that of natural alpha brain rhythms [6]. According to Gell, there are


Fig. 2 Panel a Shiva dancing. India. PhotographG. Solecki, A. Pietak. Courtesy of the Directorateof the National Museum in Szczecin, Poland; panel; b Medium on the swing. Photograph AlfredGell (The Gods at Play). Courtesy of Simeran Gell

two basic elements of trance induction: first, the modification of normal posture andmuscle tone (attention while standing rigidly, extension of the neck, stiffening ofotherwise flexible fingers and arms, staring and yawning); second, voluntary andlater uncontrollable shaking and trembling, shaking of the head, rapid movementsof the forearms up and down, hand-flapping, dancing, and twirling in the air.

TheBritish anthropologist participated in aHindu festival and noted that theMuriawere possessed without specific sounds and light, without visible modifications ofbreathing, and without special stimulus. Therefore, Gell proposed the interpretationof possession as the result of material incentives (as above) but also as immaterialincentives: as the intention of a person to be possessed.11 According to Gell, theMuria trance is much more complex than this material possession in that it is thebody itself. He saw the trance becoming a vehicle (like a journey by horse, a swing,or a ride on an elephant). Through this trance, the subject can rediscover (across thetrance-gap) between intention and experience. Gell explained that the Body is thehorse, and the Soul is the rider. We can ask, what is the world in this metaphoricstructure?

11The category of intention is central in Gell’s conception of nexus. See [8, 16]. In the context ofGell’s theory, it seems that the intention is the intermediary material and immaterial component ofthe (human and environmental) world.

178 A. Kawalec

Gell’s final conclusion on the phenomenon of possession is scientific: all thesemethods of lost equilibrium (physical and psychic) are means to the deautomatiza-tion of the experience of daily functional routines. Therefore, they could be con-sidered essentially playful (like the games of children). According to Gell, ecstatictrance combines with a re-structuring of self/world relationships by “equilibriumsense”.12 In addition to the sensory organs, which constitute the bodily/material fac-tor in the trance process, Gell conceives of the immaterial (oriented towards inten-tions/volitional) factor, which actually brings about the trance state as the input-output modulation (sensory-input) and brain activity result in the effector (output).This injunction initiates bodily actions (when the subject is becoming the agent ofacting).

Alfred Gell was a social anthropologist and he understood the person only as a“dividual” person. “Dividual”, as in the genealogical context of the life of the socialgeneration of the tribe and dividual, means interrelations between the subject and allelements around her. Therefore, ritual, for Gell, is the commonplace of the objectiveworld and the subjective person. It is the act of nature, an intentional nature.

2.6 Performative Example of the Integral Functionof Breathing (Physical and Intentional)

In the remainder of this section, I shortly present conceptions of ritual advanced byperformative artists andphilosophical anthropologists that seem tobe alignedwith theproposed approach to ritual as the acting law of nature. Jerzy Grotowski was a Polishdirector of theatre and actors’ practices. In his early work Towards a Poor Theatre[11], he proposed “Actor’s Training” activities (1959–1962), including a techniqueof voice: respiration. He distinguished among three types of respiration: (1) upperthoracic or pectoral respiration, prevalent in Europe, especially among women; (2)lower or abdominal respiration, in which the abdomen expands without the chestbeing used at all (this is the type of respiration usually taught in theatre schools),and (3) total (upper thoracic and abdominal) respiration, with the abdominal phasedominant. This is the most hygienic and functional type and is found in children andanimals.

According toGrotowski, total respiration is themost effective type for a performer.However, actors must not be dogmatic about this particular type: “Every actor’sbreathing varies according to his physiologicalmake-up, andwhether or not he adoptstotal respiration should be dependent upon this” ([11], p. 148). Grotowski noted thatthere also exists a difference between the respiratory possibilities of men andwomen.

12Gell underlined this “equilibrium sense” as consisting of the gravity- and movement-sensingorgans in the inner ear (semicircular canals in the vestibule), which are connected to a number ofsites in the central nervous system, which together constitute the vestibular system. The vestibularsystem monitors equilibrium and controls the muscles of the head, neck and upper body as well asthe posture.


Fig. 3 From Hatha Yoga:inhalation: 4 s—Hold thebreath: 12 s—exhalation: 8 s

In women, correct respiration has a definite abdominal phase, although the upperthoracic element is slightly different than in men. The actor should practice differenttypes of respiration in various positions and physical actions (such as acrobatics). Itis necessary to accustom oneself to total respiration. This helps a person to controlherself and to be ready to control the functioning of the respiratory organs.

The most common actors’ practice is breathing. Grotowski very much appreci-ated the Hindu tradition and its daily practices and rituals.13 It is well known thatthe different schools of yoga—including Hatha Yoga—demand the daily practiceof respiratory techniques to control and exploit the biological function of breath-ing. These practices have become automatic. Grotowski claimed that an actor needsmany exercises for the respiratory process and proposed some ways to verify totalrespiration. First, respiration in the horizontal position (Fig. 3).

Next, from classical Chinese theatre, when standing, place the hands on the twolowest ribs. Inspiration must give an impression of beginning in the very spot wherethe hands are placed and, continuing through the thorax, produce a sensation thatthe air column reaches right up to the head. This means that when breathing in, theabdomen and lower ribs dilate first, followed in smooth succession by the chest.During (correct) total respiration, “Do not store up or compress too much air!” ([19],p. 150).

In his work with actors, Grotowski used various techniques of breathing so thateach individual discovered the proper place of crossing of the biological and thepsycho-cultural determinations. It seems that for Grotowski—“the teacher for Per-former”—breath was a bridge over the hiatus of the external and internal humanenvironments, the physiology, the social dimension of the artist and her spirituality.It was in breath that he attempted to discover the law of nature exemplified in theindividual.

2.7 Karol Wojtyła on the Integral (Physical, Psychic,and Spiritual) Idea of the Person

Arguably, among the philosophical conceptions of ritual as the acting law of naturethe one that seems most pertinent here is the most integral philosophical conception

13On the Hindu and other influences on Grotowski, see esp. [19].

180 A. Kawalec

of the human person advanced by Karol Wojtyła.14 In his research, he used mainlyphenomenologicalmethods,whichmeans that he took detailed description of sensoryexperiences and objects as the starting point for his analyses. He was essentiallyinterested in human activities. Wojtyła understood human person as the only actingbeing. He proposed an idea of a human being as a person, a being who is integratingphysical, sensual and spiritual dimensions, with the latter (embracing volitional andconscious character) being ontologically prime.

Wojtyła proposed systematic research on the person as the way from her expres-sions to the inner realm, i.e., to the source of these expressions. According to Woj-tyła, there is no human act without its agent/subject, and vice versa: there is no agentwithout her outer expression. Wojtyła connects the ontological basis of the humanas person with the understanding of the relation: person-acting. He wrote of “themost deeply hidden roots of the dynamism of man, and more particularly the roots ofefficacy that is the key for the understanding” of the person ([31], p. 76). Accordingto Wojtyła, nature defines the subjective basis of acting. However, this “nature” hastwo meanings: nature as biological life and nature as the essence of a specific indi-vidual being. In Wojtyła’s philosophical context, Aristotle’s hylemorphism seems tobe transcended by the concept of the spiritual essence of the human being. Wojtyła,as aWestern philosopher, underlined the specificity of human existence among otherkinds of being: rationality and self-consciousness, voluntarity and the resulting skillsof the subject, self-governance, self-possession, and self-determination.

Wojtyła indicates that the acting of the integral human being (vegetative-sensual,covetous, voluminous and rational) is the realm of the crossing of the world of natureas the material and immaterial world, while the individual act of the human personwitnesses this realm. Thus, the law of the existing world is—according to Wojtyła—a harmonious interaction between the acting human person (as a free and rationalagent) and the acting nature (as an objectively existing reality).

2.8 A Brief Comparative Note

Etymologically, the term “nature” is derived from the Latin verb “nascor” (to beborn). It denotes literally everything that is going to be born or is contained in thefact itself of birth as its possible consequence. This literal meaning is basic in Gell’sapproach to the difference of nature-culture. For him, the notion of culture is the highabstractive notion specific to theWestern tradition. “Culture” includes everything thatwe may imagine, and what we want. Therefore, the person is also the culture.

The anthropological research approach of Alfred Gell indicates that a tribe’sthinking and understanding of the world, the self, and the environment has an integralcharacter (for example, totemic thinking). Therefore, the soul and body, ancestors,and elements of the environment are integrated. This is the individual person [8].

14Since 1978, he has been known as the Pope John Paul II. Wojtyła’s main philosophical work isThe Acting Person [31].


According toWojtyła, this is nature, which determines the ways of acting (vegeta-tive: the plants; vegetative and sensual: the animal; vegetative, sensual, rational, andvolitional: the person). Rational and volitional dimensions are determined by others;mutually, the vegetative and sensual determine the higher dimensions (however, theperson can self-determined only by the volitional-rational dimension of the subject).According to Wojtyła, all dimensions—physical, psychic, and spiritual—create thehuman acting.

According to Wojtyła and Gell, the person is the agent and performer in thesame process of action, which is the dynamic action-motion synthesis. For Wojtyła,various forms of the disintegration of the person may be defined as purely somatic.Such disintegration may take the form of the absence of a certain organ or memberof the human body, resulting in the inability to make certain motions, or it may bedue to an absence of some purely somatic reactions or skills leading to impedimentsor disorders of human action. For example, “A man who has lost an arm or breath…is subject to very definite limitations or difficulties in his acting […] BUT theseare external in the sense that they do not in themselves distort his consciousness orprevent his self-determination. On the contrary, very often a human being with a highdegree of somatic disintegration may represent a personality of great value” ([31],p. 215). Thus, according to Wojtyła, the person is the integration of the somatic,psychic and spiritual powers, but at the same time, the person is self-determined(only by subjective power), whereas the breath of humans is the experienced andunknown commonplace external and internal world, the injunction of the somatic,psychic, and spiritual power of agency.

3 Concluding Remarks

If we suspend the superstition concerning the distinction between culture and nature,ritual will lose its symbolic character (imposed upon it by theWestern people). Ritualwill become an expression of nature (a repetitive repetition of the breath), determiningthe actions of the breathable creatures. Ritual (in the sense of R. ta derived from theSanskrit), in this sense, is therefore the commonplace of the external reality forbreathable beings. Ritual as R. ta justifies the natural and inviolable character of thewhole reality irrespective of the labels given to it by various civilizations, religions,and disciplines.

The presented examples of various ideas aimed to support this conclusion, butthey may be less (Gell, Rappaport, Wojtyła) or more (Schechner, Strathern) limitedby the Western distinction of nature-culture. As argued in the previous section, it isSanskrit that instantiates the idea of the most realistic conception of ritual. The pre-sented examples of various ideas come from various fields of life of human being andsocial. But all of them are integrated by the above mentioned theses and pictures thatstem from different areas of human and social life. Their all share a non-scientisticperspective, and foremost, before taking the fact of human individual and social exis-tence as correlated with (intangible) law of nature. This correlation can be perceived

182 A. Kawalec

in continuously experienced in the material-spiritual breath. However, the questionremains: who and to what extent perceives it?

Acknowledgements The author gratefully acknowledges support from the grant NCN2013/11/B/HS1/03961, awarded by the National Science Centre in Poland.

References

1. Agier, M.: Epistemological decentring: at the root of a contemporary and situational anthro-pology. Anthropol. Theory 16(1), 22–47 (2016). https://doi.org/10.1177/1463499616629270

2. Apter, A.: Ethnographic X-files and Holbraad’s double-bind: reflections on an ontological turnof events. HAU: J. Ethnogr. Theory 7(1), 287–302 (2017). https://doi.org/10.14318/hau7.1.021

3. Bloomfield, M.: The Religion of the Veda: the Ancient Religion of India: from Rig-Veda toUpanishads. Adamant Media Corporation, Boston (2005)

4. Descola, P.: Beyond Nature and Culture. The University of Chicago Press, Chicago (2013)5. Geertz, A.W.: Brain, body and culture: a biocultural theory of religion. Method & Theory in

the Study Relig. 22(4), 304–321 (2010). https://doi.org/10.1163/157006810X5310946. Gell, A.: The Gods at play: vertigo and possession in Muria religion. Man 15(2), 219–248

(1980)7. Gell, A.: The Anthropology of Time: Cultural Constructions of Temporal Maps and Images.

Berg Publishers, Oxford (1992)8. Gell, A.: Art and Agency: an Anthropological Theory. Oxford University Press, Oxford (1998)9. Goodman, A.H., Heath, D., Lindee, M.S. (eds.): Genetic Nature/Culture: Anthropology and

Science Beyond the Two-Culture Divide. University of California Press, Berkeley (2003)10. Grimes, R.L.: Beginnings in Ritual Studies. University Press of America, Washington, DC

(1982)11. Grotowski, J.: Towards Poor Theatre. Simon & Schuster, New York (1968)12. Holbraad, M., Pedersen, M.A.: The Ontological Turn: an Anthropological Exposition.

University Press Cambridge, Cambridge (2017)13. Hornborg, A.: The political economy of technofetishism: agency, amazonian ontologies, and

global magic. HAU: J. Ethnogr. Theory 5(1), 35–57 (2015). https://doi.org/10.14318/hau5.1.003

14. Huxley, J.: Introduction. Philos. Trans. R. Soc. Lond. Ser. B, Biol. Sci. 251(772), 249–271(1966)

15. Kawalec, A.: Comic agents: from a poetic to an anthropological paradigm of comedy. Agathos:Int. Rev. Humanities & Soc. Sci. 7(1), 7–23 (2016a)

16. Kawalec, A.: Person and Nexus: Alfred Gell’s anthropological Theory of Art (in Polish).Wydawnictwo KUL, Lublin (2016b)

17. O’Donnell, J.K. (ed.): Here, Now, and Always: Voices of the First Peoples of the Southwest.Museum of New Mexico Press, Santa Fe, NM (2001)

18. Oldenberg, H.: Die Religion des Veda. W. Hertz, Berlin (1894)19. Osinski, J.: Jerzy Grotowski. Zródła, inspiracje, konteksty. Słowo/obraz terytoria, Gdansk

(2010)20. Owings, A.: Indian Voices: Listening to Native Americans. Rutgers University Press, New

Brunswick, NJ (2011)21. Rappaport, R.A.: Ritual and Religion in theMaking of Humanity. Cambridge University Press,

Cambridge (1999)22. Schechner, R.: The Future of Ritual:Writings on Culture and Performance. Routledge, London

(1995)23. Schechner, R.: Performative Studies: an Introduction. Routledge, London (2013)


24. Schlerath, B.: Awesta-Wörterbuch. Vorarbeiten II: Konkordanz. Otto Harrassowitz:Wiesbaden(1968)

25. Singer, M. (ed.): Traditional India: Structure and Change. The American Folklore Society,Philadelphia (1959)

26. Strathern, M.: After Nature: English Kinship in the Late Twentieth Century. CambridgeUniversity Press, Cambridge (1992)

27. Turner, V.: The Ritual Process: Structure and Anti-Structure. Routledge, London (1969)28. Turner, V.: Body, Brain, and Culture. Zygon 18(3), 221–245 (1983). https://doi.org/10.1111/j.

1467-9744.1983.tb00512.x29. Vitz, R.: Reforming theArt of LivingNature, Virtue, andReligion inDescartes’s Epistemology.

Springer, Cham (2015)30. Wall,D.,Masayesva,V.: People of the corn: teachings in hopi traditional agriculture, spirituality,

and sustainability. Am. Indian Q. 28(3), 435–453 (2005). https://doi.org/10.1353/aiq.2004.0109

31. Wojtyła, K.: The Acting Person Trans. A. Potocki. Reidel, Dordrecht (1979)

From Habits to We-Intentionality:Rituals as Social Habits

Raffaela Giovagnoli

Abstract Habits and rituals represent a relevant field to investigate human behavioralso in comparison with non human animals and nature in general. We move fromvery interesting perspectives that cross philosophy and neuroscience to establish acontinuity between individual and social habits and to clearly isolate what makeshuman rituals peculiar in the field of social ontology. Therefore, we introduce ageneral and brief discussion on rituals in different disciplines before to focus onthe actuality of the Aristotelian notion of habit in relationship with some studiesin neuroscience that can be extended to the social context. The perspective of We-Intentionality introduced by John Searle embeds the dimensions of habits that groundcooperation and is very useful to describe the peculiarity of human rituals.

Keywords Habits ·We-Intentionality · Rituals · Cooperation

1 Introduction

Rituals represent a fundamental natural behavior characterizing individual and sociallife and—from a sociological point of view—they can be considered as sets of actsor normatively codified practices forming cultural patterns of a given society repre-senting values and standards of the transmission function, the institutionalization ofroles, identity and social cohesion [1]. Recently, Schilbrack edited a very interestingbook [2], where different philosophical perspectives are considered by starting fromclassical thought in order to consider rituals as source of knowledge. Several impor-tant anthropological views show that rituals are part of our nature which entails alsoa cultural dimension [3].

We aim to capture the nature and function of rituals starting from processesinvolved in habitual behavior. First, we would like to refer to the recent study ofJuergen Habermas about the nature of rituals in the context of his post-metaphysicalthought. We think that Habermas offers an interesting view about the relationship

R. Giovagnoli (B)Faculty of Philosophy, Pontifical Lateran University, Vatican City, Italye-mail: [email protected]


185

186 R. Giovagnoli

between the anthropological research conducted by Tomasello and the intersubjec-tivity that characterizes human cooperation in different social contexts. In order toface the challenges coming out from our environment both in its natural and socialdimensions, we need to grasp the aspects of human behavior that make our dailylife easier and more pleasant together with those that give rise to social practices.Cooperation is the keyword to understand both the nature and function of rituals inhuman life.

Second, we try to grasp the connection between habits and rituals moving fromsome important studieswhich cross philosophy and neurobiology (broadly intended),which recall important Aristotelian notions. The core theoretical point is that theyinvolve processes on the dimensions of mechanical behavior and goal-orientedactivity.

Third,we propose a possible extension of habitual behavior to cover also collectiveagency, by reference to the classical distinctions introduced by Aristotle and revisedaccording to important results in neuroscience. Consequently, we aim at describingrituals in terms of We-intentionality; a notion that entail cooperation on differentlevels of behavior (namely routine and goal-directed activity), as the famous exampleof “marriage” introduced by John Searle shows.

2 Post-Metaphysical Thought and Rituals

Habermas conception of ritualsmoves from the analysis of the normativity character-izing the background of knowledge and action. Life-world is an important key-notioncharacterizing Habermas post-metaphysical thought, namely the choice for the onto-logical primacy of ordinary language instead of consciousness or any principle basedon the so-called first-philosophy [4]. It represents the horizon of human perceptiveand cultural experience, namely the background from which we raise our validityclaims and exchange our reasons for knowing and acting. Life-world reveals itself inthe dialogical situations where the pragmatic structure of communication acquires itsfundamental role. Language must be pragmatically interpreted, namely starting fromthe communicative praxis of the subjects who, while trying to dialogically solve theirquestions, are also able to learn from their experiences. Communication possessesa constitutive role: it continuously forces the participants to undertake a positionsfor or against available validity claims. So, they find themselves always running therisk to be contradicted and so, they have the possibility to correct their opinions in aautonomous and rational manner [5].

As Husserl himself has shown in his latest thought, life-world has a double nature,namely it is both the epistemological and ontological background for science andhuman activities in general. Habermas proposes to intend lifeworld as structured bylinguistic practices, so the problem is how we can intend these practices that openand project a world as something that happens in the world. It si a fact that we canproducing advances in our daily life only by facing the contingency of the events.Consequently, there exist a relation between our practices and these very events,

From Habits to We-Intentionality: Rituals as Social Habits 187

which we can easily grasp in case of failures of our projects. Beyond the traditionalprimacy of the phenomenological consciousness or a religious interpretation of theworld, Habermas proposes a third solution: the transcendental spontaneity of theKantian Ego manifests itself in the very practices where the reproduction of life-world is connected from the inside with our mundane processes of learning. Theselearning processes take place by virtue of an interactive exchange among intentionalreferences to the world, reciprocal assumption of perspectives, use of a proposition-ally differentiated language, instrumental reasoning and cooperation. We find here aclose connection between human agency and world through communication, whichrepresents the epistemological dimension characterizing our learning and cognitivedevelopment. Nevertheless, It is important to stress that Habermas does not suggesta merely “cultural” solution because he recognizes the surprising results of the bio-sciences, psychology and cultural sciences that could play a central, heuristic role inthe development of the natural history of human life.

In his latest thought Habermas maintains that human agency acquires normativitynot only from ordinary communication but also from extra-ordinary communica-tion. Life-world can be also described as a “space of reasons” (to use the metaphorof Sellars) which can acquire a social structure as Robert Brandom has shown in hisimportant work [6]. Habermas aims at describing the intersubjectivity that is typicalof human beings who are able to have collective intentionality in the form of cooper-ation. Humans have their unique way to interact so, even though non-human animalshave their rituals, humans have also their peculiar way to constitute and to do theirparts in ritual practices.1 According to Tomasello: “The understanding of personsas intentional agents at one to two years of age—in ways that show sharpness andperspective—thus inaugurates the development of uniquely human skills of socialcognition” [7]. This understanding involves appreciation of the “original normativ-ity” of actions because an intentional agent’s action—either self or other—may bejudged as successful or unsuccessful. One-year-olds have thus entered, at least ina nascent and implicit way, the space of reasons involving normative judgements.But it is even more relevant for our discussion that children of this age also come toappreciate that shared intentionality and collective practices create “derived norma-tivity”. This is a more deeply social sense of normativity, which entails the use ofsymbols, artifacts, and other culturally constituted entities. These entities are investedwith normativity through the actions of intentional agents and their attitudes: this isthe way “we” use a symbol or tool. This is the way it “should” be used; this is its“function” for us, its users.

Derived normativity is most readily apparent in the use of linguistic symbols andmaterial artifacts such as tools and toys. As Lorenzo Magnani shows, the role of

1Habermas is interested in the continuum of the development of rituals as a result of communicationhabits in living organisms—animals and humans alike. Gordana Dodig Crnkovic wonders what theessential difference between animals and plants results in animals (with humans) having ritual,patterned, repeating, symbolic, indirectly purposeful communication procedures while we cannotdetect rituals in plants—even though plants exhibit rich forms of communication. Can it be thatanimals (including humans)move, and thatmovement as a part of communication plays an importantrole in rituals?

188 R. Giovagnoli

habits, rituals and symbols in human cognition can be elucidated by the use of twoconcepts: “manipulative abduction” and “disembodied distributed mind”. In thissense, human mind can delegate to external representations suitably stored in theenvironment important cognitive functions. These representations:

(…) can be merely mimetic (when they represent something already internally present in themind, for example a triangle depicted in a dashboard), but also poetic, when external repre-sentational aspects are manipulated to the aim of discovering something new by exploitingthe specific constrains of the materiality of the external “mediators” chosen. [8]

Magnani introduces an eco-cognitive perspective, which is grounded on the relation-ship between internal and external representations fromwhichwe derive the notion of“manipulative abduction”. Manipulative abduction embeds cognitive action, whichgenerates information not yet available to the aim of selecting or generating inter-esting hypotheses. A clear example is represented by science, where the explorationof external tools, devices and instruments is widespread [9].

According to Habermas, rituals possess a cognitive aspect as they embed showsinterpretations of our common environment. Rituals represent the form of extra-ordinary communication which embeds the strong normativity of the cultural back-ground because symbolization goes beyond the mere embodiment of the semanticcontents as it shows the “bounding effect” produced by traditions, roles and institu-tions. Rituals possess the characteristic of auto-referentiality as they do not refer tosomething in the world. The absence of the referent put ritual communication in anextra-ordinary dimension:

Let’s think to a ritual dance. The rhythmic movements of the dance and procession expressthe comeuppance of intentions, where the reciprocal imitation of gestures stimulates thereciprocal exchange of perspectives and, with strong evidence, the sharing of an “internalexperience”.

Evidently, the dimension in which we look for the lacking referent cannot but be that dimen-sion of social cohabitation produced by the new evolutive stage of the symbolic communi-cation. In other words, I intend the ritual behavior as a response to the problems created bythe socialization of individuals on this level of communication. From the point of view of thesocial pragmatics, gestural communication represents a new form of (evolutively fruitful)intelligent cooperation. However, this socialization of intelligence cannot but modify thecoordination of conducts. The first gesture through which Ego and Alter institute an identi-cal meaning, frees also the individual consciousness from the egocentric shell that encasedit. The symbolization of meanings produces, at the same time, “externalization” and “so-cialization” of the contents of individual consciousness. So - in the companions of species- the monads of subjective consciousness open themselves to each other. This entrance inthe public world of symbolic forms does not only mean an overcoming of the cognitive ego-centrism; this socialization of intelligence produces also - and mostly - a revolution of thecoordination of conducts. By converting to symbolically mediated interactions, the membersof the group will be forced to face in a cooperative manner the contingencies of the worldstarting from the horizon of their common life-world, a cognitive challenge that proceedsparallel to a psycho-dynamic one [5].

As Durkheim has shown, rituals reveal their function in the re-generation of soli-darity as well as in the self-thematization of the communitarian identity. Rituals areself-reflexive practices that deserve to stabilize the compactness of the social group.


Durkheim has challenged the interpretation of rituals as a mere sequence of actionsexecuted in a mechanical way like routines [10]. He points out that the peculiarityof rituals resides in their being in relation with an order of socially shared meanings.Social shared meanings are what distinguishes the basis of rituals from other formsof repetition in animal behavior or psychological diseases. On the one side, we mustunderscore the social character of rites, on the other, they refer to a particular orderof meanings, namely an extra-ordinary dimension that goes beyond the triviality ofdaily existence.

Durkheim point on two aspects of this meaning that can be extended beyond thereligious sphere: first, the self-thematization of the society and, second, the produc-tion of obligations for the normative behavioral expectations. By performing rituals,societies express their present social structures and accept them again by recogniz-ing the obligations of their forms of common life. Arnold van Gennep enriched theworks of Durkheim in his famous analysis on the rituals that regulate the passagesfrom a social status to another (passage rites). For example, an adolescent is allowedto enter in the group of adult males and he learns the social roles connected withthis news status. By following certain ritual prescriptions, he enters, so to say, a newsegment of society, by acquiring at the same time the capacities to obey to corre-spondent expectation. The initiation aims at preventing the risk that—in the net ofthe generation changes—the continuity of social integration is broken and the forcesof normative bounding are paralyzed.

Starting from some important anthropological questions, which point on thenecessity to find a language through which humans can thematize extreme situationslike borne and death, unstable coexistence, depletion of natural resources, vulnera-bility of soul and body, Rituals are a form of reacting to vulnerability entailed by newcommunicative forms of socialization. Habermas makes the example of the gesturalcommunication. The way of responding to the presence of fire by expressing theword “fire!” with the corresponding linguistic force is a simple example of cooper-ative behavior which inform others that we are in a dangerous situation. Rituals arecommunicative forms that have the same function of gestural communication. Theydeserve to elaborate those vulnerable situations that characterize passages from asocial status to another and cooperation to face challenges from the external world.

3 Habits Beyond Routines

In order to grasp the relationship between habits and rituals, we need to presenta plausible sense for the notion of habit, which goes beyond the mere repetitivebehavior or routine. Starting from Latin, there are twomeanings for the English word“habit”. The first is Habitus, that entails a deliberate disposition to act; the secondis Consuetudo, that implies the constant repetition of an event or behavior withoutdeliberation. The traditional philosophical sense of habit (Habitus) is introduced byAristotle to characterize the notion of “virtue”. Virtue can be considered as a habit inthe sense of a disposition to deal with good or bad emotions and tendencies. Aquinas

190 R. Giovagnoli

moves from the Aristotelian view and maintains that habit is not a mere potency (i.e.a capacity) because it makes us able or unable to act in a good or wrong way. Deweymaintains that it is a human activity influenced by previous activity: we can say thatit is acquired in the process of socialization and entails a certain order and a certainsystem of minor elements of action. Habits can be seen as dynamic dispositions thatwork in a subordinate form when they are not the dominant activity at a certain time.

The meaning of the term habit (Consuetudo), as constant repetition of an event ora behavior, refers to a mechanism that can be described in the fields of natural andsocial sciences. Standardly, we assume that this mechanism mostly develops fromthe repetition of acts and behaviors and, in the case of human events, it is acquiredthrough training. Aristotle, conceived this notion of habit as a mechanism that isanalogous to natural mechanisms, and somehow guarantees the uniform repetitionof facts, acts, or behavior by eliminating or reducing effort and fatigue and so bymaking them pleasant. Habit as repetition without reasoning is also exemplarilydescribed by Pascal and Hume. Bergson presented a relationship between habits andmoral obligations: moral obligations can be interpreted as social habits to favor sociallife and social order. Metaphysical interpretations of the notion of habit are offeredby Main de Biran, Hegel and Ravaisson. In this case, we can observe that habithas a strong connection with religious views. According to Hegel, habit is the mostessential thing relevant to the existence of spirituality of the individual subject; thesubject can exist as a concrete subject and as ideality of the soul, namely the religiouscontent can belong to himself (with his own soul). The metaphysical perspective ofRavaisson presents a notion of habit as a law of grace that has the important resultto consider nature as spirit and spiritual activity, because it demonstrates that spiritcan turn into nature and vice versa.

A habit is not only a mere automatism or a repetitive behavior, but also a stabledisposition for action (practical skill), that implies the relationship between automa-tism and flexibility. We can recall the famous aristotelian difference between techneand praxis, which entails important consequences for collective agency [11]. Thedifference between habits and automatism or simple routines is that the former givecontrol over actions, while the latter don’t [12]. According to this view, that crossesphilosophy and neurobiology, the habit is a “stable disposition for self-development”.


We can observe habits according to an individualmode or I-mode because they areidiosyncratic as regards personal behavior. Habits have a very important function inindividual life because they reduce the complexity of daily life; they make our dailylife easier and pleasant. Naturally,we can control habits concerning the satisfaction ofour basic natural needs. Depending from natural and social environment, we developdifferent habits which organize the way to satisfy our human needs.

The mechanism of habit is something we share with nature; it becomes highlyvisible if we are in touch with animals like dogs, cats, birds, horses among others.Habits can appear also in a We-mode as in the case of a person who regularly meeta friend for jogging in a certain day at a certain hour. We do not say that the friendshave a ritual, but we say that they share a habit. We can have many different habitswe share with people and also with animals. For example, if I go out with my dog ata certain hour early morning, me and my dog share a habit.

There are some characteristics of habits relevant for understanding how they canbe extended to the social dimension. Bernacer and Murillo [13] note the explicitinvestigation of habits in neuroscience is quite recent despite the fact that “we actaccording to our habits, from the time we rise and go through our morning routinesuntil we fall asleep after evening routines” [14]. Graybiel proposes a fruitful rela-tionship between habits and goals because goals are explicitly present during actionevolution and selection and they increasingly blur the more an action is repeated.Along this line, we find interesting studies in the ambit of neural-dynamic logic[15]. We have examples of habits as fixed action patterns namely complex repeti-tive behavior in non-human animals and repetitive behavior and thoughts in humanpathological conditions [14]. She concludes that a habit completely disengaged froma goal becomes either a stimulus-response pair for a non-human animal or a patholog-ical trait for human beings. Her theoretical contribution resides in the classificationof habits as “neutral”, “good” or “bad” where good habits seem to be those choose toguide our behavior and bad habits those that powerfully take control on our behavior.They conclude that this categorization seems to make possible to include goals asdrivers of habits. They intend “good” habits as those rationally directed to a goal,and “bad” habits as behavioral depositions to perform rigidly defined actions uncon-trolled by cognitive processes. Graybiel also maintains that habits play an importantrole in social life, as in this case they are “shaped” as mannerism and rituals.

192 R. Giovagnoli

Williams James’ notion of habit [16] is central to the study of neurobiologicalfoundations of motor routines, the relation of consciousness with habits, the mecha-nism of instrumental learning in animals and the implication of these phenomena inhuman disorders. But, because of the associationist view grounding this research, itseems far from explaining the complexity of human habits. Consequently, Bernacerand Murillo [13] underscore three important results from a deep study of the Aris-totelian analysis of habits in Nicomachean Ethics. An acquired habit is an acquireddisposition to perform certain types of actions; this disposition, usually acquired bymeans of repetition of one or more actions, makes the exception of these actionsprompter, more spontaneous and autonomous from continuous supervision, all ofwhich generally leads to a better performance. If the habit increases cognitive con-trol of actions, it can be termed a habit-as-learning; on the contrary, if it increasestheir rigidity, it is a habit-as-routine. Habits-as-routines are fundamental for the cog-nitive enrichment of actions entailed by a variable amount of practice (efforts arerequired to engage in activities and performances). Differently, habits-as-learning arenot merely acquisition of a way of acting; they entail a cognitive capacity connectedto the habit that can flexibly used in different contexts.

For example the singer, who can easily reproduce notes (a mostly theoreticalhabit) and whose voice appropriately responds to the reading (a technical habit), isable to give her own interpretation. In the case of behavioral and technical habits,they imply the availability of motor skills for complex activities, as well as themodulation of tendencies and desires to positively respond to conscious and rationalgoals. Therefore, they entail the performance of habits-as-routines, but their criticalcharacteristics go beyond their motor aspects.

Habits-as-routines and habits-as-learning have a different relation to conscious-ness. Habits-as-routines represent a fully unconscious performance. Habits-as-learning reduce or eliminate consciousness of basic elements of the action in order toconcentrate on higher goals, while preserving at all times the possibility of recoveringthem for conscious attention.

It is worthy to underline the contribution of the Aristotelian distinction betweengood and bad habits, that intends good habits as those enhancing the agent’s controlto reach certain goals. Consequently, we can clarify the relation between habits andemotions. The habits-as-learning entail control and for this reason they are funda-mental to reach personal goals. This is the process that favors the agent’s pleasureand happiness.

Some authors intend the idea of “habit learning” as the performance of an action,previously learned after many repetitions namely in an unconscious manner, and


whose execution is inflexible and independent to the outcome [17]. This perspectiverequires an integration with other perspectives that recognize the importance of thesensitivity to the outcome and of different levels of flexibility and feedback [18].According toLomboandGiménezAmaya, a neurobiological viewof “habit learning”and recent experimental contributions (especially those of Graybiel) are consistentwith the Aristotelian notion of “habit”. Human habits are essentially based on twoaspects: (a) the stable character of an acquired quality; and (b) the capacity for newactions that arises from that quality [19].

The integration between routines and goal-directed behavior is at the center ofimportant study in social psychology. Wood observes that successful long-term self-regulation involves habitually engaging in actions that correspond with valued long-term goals; people exert self-control to inhibit unwanted habits, and they repeatgoal-directed actions in stable contexts so as to form desired habits [20]:

Self-regulation of habits, even habits of addiction, may proceed as a result of people shiftingtheir goals and intentions to be consistentwith the habits theyperform.Although the inferencethat a habitual behaviorwas intended is largely erroneous in termsof psychological processes,these inferences have an intuitive plausibility for high-frequency behaviors. Such inferencesare further promoted by the switching costs of deviating, the fluency of habit performance,and the potential benefits of habit performance for well-being.

To schematize, we could term behavior either habitual or goal-directed, but this isactually an oversimplification. According to Woods, in reality, habits integrate withgoal pursuit in guiding behavior and the integration of habits and more goal-orientedbehaviors provides a number of advantages for action control [20].

Perhaps most importantly, habit knowledge does not shift readily and is therefore retainedeven when people change their goals and plans. Also, by outsourcing action control toenvironmental cues, people have a ready response when distraction, time pressure, loweredwillpower, and stress reduce the capacity to deliberate about action and tailor responsesto current environments. Furthermore, although people might believe in the effectivenessof effortful goal pursuit over relying on habits, habit systems are also smart. By actinghabitually, people efficiently capitalize on environmental regularities, even ones of whichthey may not be consciously aware.

4 Rituals as Social Habits

Goal-directed behavior is observable in humans and other animals. Recent studiesfrom cognitive neuroscience, biology and psychology show converging perspectiveson the organization of goal-directed, intentional action in terms of (brain, compu-tational) structures and mechanisms. They conclude that several cognitive capabil-ities across the individual and social domains, including action planning and exe-cution, understanding others’ intentions, cooperation and imitation are essentiallygoal-directed [21]. To form habits we need goal representations both in individualand social contexts. They have a crucial role in the planning and control of action;moreover, action understanding and imitation are performed at the goal rather than the

194 R. Giovagnoli

movement level [22]. The researches indicate a close link between (socio)cognitiveabilities and situated action and try to enlighten their neural underpinnings andmech-anisms. In this context, it has been revealed that themotor system is highly engaged inanticipatory, simulative and generative processes. Some authors introduce an interest-ing speculative perspective, and make the case that the same predictive mechanismsprovide both a ‘linkagewith the future’ required for taking goal-directed action, and a‘linkage with others’ required to act socially [23]. We observe that recent researcheshave lead to a profound rethinking of basic concepts in cognitive and behavioralsciences, and a common theoretical view—a motor-based (or action-based) view ofcognition—is emerging across disciplines [24]. They provide a description of theabilities of action execution, its planning, and understanding of others’ intentions asessentially goal-directed and served by the same representations, which are action-oriented and involve deeply the motor apparatus.

Routines and goal-directed behavior characterize habits both in the case of indi-vidual and social behavior. We can observe ritual behavior also in non-human ani-mals, but, in the human case, they have their expression in a different form of We-Intentionality. Moreover, they need conventional devices, that are mutually recog-nized through ordinary communication. Rituals have the important function to createsocial spaces in which individuals can share emotions, experiences, values, normsand knowledge. The function to share experiences is fulfilledwhen there exist a socialspace created by cooperation for reaching certain goal. In this context, cooperationis a kind of intersubjectivity typical of human beings who, differently from apes,are able to have “Collective Intentionality” [7] i.e. the basic intention to cooperateand therefore to reach together a certain goal. If we want to get a positive resultabout the extension of habits in the social dimension we need to move from a sort ofgoal-directed activity that we can perform together. There is a contemporary livelydebate on the nature and structure of Collective Intentionality, as a necessary notionto researches in the field of social ontology (the pioneers in this area are John Searle,Raimo Tuomela, Margareth Gilbert, Michael Bratman and Philip Pettit).

We can observe that human beings (but also other species) have the capacityto impose a function to an object so that the object acquires a function dependenton the peculiar scope of the agent. The continuity between individual habits andrituals (social habits) is thus showed by the fact that humans create these “agentivefunctions” (in Searle’s terminology) in a wide variety of situations. Also non humananimals have their form of creating functions for objects but there is a fundamentaldifference in the concept of “function” in the human case. I think that this point isclearly explained by Searle and it is worthy to be quoted for our argument [25]:

(…) when we discover functions in nature, what we are doing is discovering how certaincauses operate to serve certain purposes where the notion of purpose is not intrinsic to mind-independent nature, but is relative to our sets of values. So, we can discover that the heartpumps blood, but when we say that the function of the heart is to pump blood, we take it forgranted that life, survival, and reproduction are positive values, and that the functioning ofbiological organs sees these values (…) a function is a cause that serve a purpose. And thepurposes have to come from somewhere; in this case, they come from human beings. In thissense, functions are intentionality-relative and therefore mind-dependent.


We create social habits in the form of rituals by using the “status function”,which is a peculiar kind of function from which we create the social world. Ritualsare created by intentionality-relative functions and are characterized by two specialfeatures. They require (a) a shared intentionality namely “collective intentionality”and (b) collective imposition and recognition of a status.

The “constitutive rule” is essential to the process of constitution of institutions ingeneral. The canonical form introduced by Searle is:

Status Function = X counts as Y in C

For instance, a certain expression counts as promise in a certain context C. So, itis fundamental to assign functions to objects and persons. We use ordinary languageto represent state of affairs and norms, namely to understand what are the condi-tions of satisfaction of different speech acts (assertions, commands, promises etc.).Beyond the classical dimensions of syntax, compositionality and generativity, thereis a fundamental dimension which generates public norms i.e. “deontology”, whichis characterized by the speech act of “declaration”. For example, if we say “This ismy house” or “This is my coach”, we not only represent a state of affairs, but wecreate a deontology which manifests itself in rights, obligations and duties as well asin the acceptation of the corresponding speech acts from the part of the interlocutors.This is the process by which a public deontology is created namely public reasonsfor acting that are desire independent. Language not only describes something butcreates and “partially” constitutes what it at the same time describes and creates.Representations that are partially constitutive of institutional reality, the reality ofgovernment, private property, money, universities and cocktail parties, are essentiallylinguistic. What must be clarified is the sense of this “partially” constitutes, becauselanguage use works on the basis of a prelinguistic dimension which embeds back-ground capacities such as capacity to cooperate, to act as a “We”. Moreover, thisvery capacity is fundamental to share and constitute rituals.

We pointed out the fundamental process of assigning functions to objects or tosome non- physical entities, which is a form of symbolization aiming at creatinginstitutional reality. This process is at the basis of the institutionalization of ritualsand works in every community even though social practices in general are culturallycharacterized. Status Function apart, there are other two basic notions that occur inthe explanation of a successful functioning and stability of social institutions. Thefirst is “cooperation” as a “strong” form of Collective Intentionality and the secondis “collective recognition” as a “weak form” of it. We think that these two forms ofintentionality correspond to the notion of “flexibility”, which imply the voluntarycontrol over our actions and to the notion of “rigidity”, which characterize the merefollowing rules in the sense of routinely behavior.

For example, in an actual transition when I buy something from somebody and put moneyin their hands, which they accept, we have full-blown cooperation. But in addition on thisintentionality, we have prior of the transaction and continuing after the transaction an attitudetoward the piece of paper of the type I am placing in the hands of the seller, that we bothrecognize or accept the pieces of paper asmoney and indeed, we accept the general institution

196 R. Giovagnoli

of money as well as the institution of commerce. As a general point, institutional structuresrequire collective recognition by the participants in the institution in order to function, butparticular transactions within the institution require cooperation of the sort I have beendescribing [25].

Similarly, we can describe a very famous example of a ritual (Searle often refersto), namely “marriage”. First, we need to be moved to act in a certain way. We-Intentionality works when we want to do something together (we have a collectiveintention) so that we can cooperate to achieve our common goal. As we alreadyanticipated, Collective Intentionality presents a weak form (collective recognition)and a strong form (cooperation). Both are crucial for rituals, in our case marriage.Now we can see how a social transition from one status to another is performedthrough an institutionalized ritual:

(1) We have “collective recognition”, which means that the couple simply acceptsthe institution of marriage prior to actually getting married.

(2) But, the actual marriage ceremony is an example of active cooperation, inwhich the couple enters in a new social situation acquiring new social statusesconsequently.

(3) This fact obtains by the performance of the speech act of promise.(4) The social context requires also the speech act of declaration from the part of

the institutional figure who has the suitable deontic powers to celebrate the riteand to ascribe the new status to the couple.

5 Conclusion

We briefly introduced the notion of habit, which is at the basis of a lively con-temporary debate. We think that a research that crosses philosophy and neuro-science/neurobiology could offer some important clarification of the functioning


of habits and can extend to the social sphere of rituals and their function in individ-ual and interpersonal contexts. Habits and rituals require routines and goal-directedactivity even though the context is different. Rituals requires also a kind of symbol-ization that can be represented in different forms: the attribution of a symbolic valueto certain objects, animals and procedures. The object acquires a “status function”i.e. counts as something that can be recognized to mean something else.

For instance, the ceremony to award diplomas requires the students to dress therobe which means the passage to a higher level of education and a potential accessto a prestigious university.

In bullfight, people assign a symbolic value to the bull and to the peculiar uniformof the toreador. Also in the case of rituals we can have “good” and “bad” practicesand the recognition of them implies the agents’ control on habitual behavior (habits-as-learning) so that they can consciously choose to take part to ritual practices or tochange their choices. For example, to kill a bull is not considered a good practicein every culture or for anyone (like eating lamb for Easter). Differently, to acquire ahigher level of education can be generally considered a good practice.

198 R. Giovagnoli

Language (written or spoken) is not always present in rituals, even though as wehave seen, symbolization is required to establish a social practice like a rite. Let’sthink to dance as a performance largely present in rituals. Similarly to the case ofmarriage, in the tea ceremony the ritual focuses on a codified set of sentences as wellas on some specifically objects used to prepare tea, so that they acquire a certainvalue and on the meaning of the very ritual.

References

1. Giovagnoli, R.: Life-world, language and religion. Habermas’Perspective. Aquinas, LVIII, 1–2, 253–260 (2015). Giovagnoli, R.: Communicative life-world and religion. Asian Res. J. ArtsSoc. Sci. 1(3), 1–7 (2016); Giovagnoli, R.: Habits and rituals. In: Proceedings of the IS4SI 2017Summit Digitalisation for a Sustainable Society, Gothenburg, Sweden, 12–16 June (2017).https://doi.org/10.3390/IS4SI-2017-03983; Giovagnoli, R.: Lifeworld, we-intentionality andrituals. In: Salis, P., Seddone, G. (eds.) Mind, Collective Agency, Norms. Essays on SocialOntology. Shaker Verlag, Aachen, pp. 22–32 (2017)

2. Schilbrak,K.: Thinking throughRituals: Philosophical Perspectives.Routledge, London (2004)3. Kawalec, A.: Agency of breath: beyond disciplinary views on ritual (forthcoming in this

volume)4. Giovagnoli, R.: Agire comunicativo e Lebenswelt. La prospettiva di Juergen Habermas.

Carocci, Roma (2000)5. Habermas, J.: Nachmetaphysiches Denken II. Suhrkamp, Frankfurt (2012)6. Brandom, R.: Making It Explicit. Cambridge University Press, Cambridge (1994)7. Tomasello, M.: Why We Cooperate. The MIT Press, Cambridge (2009)8. Magnani, L.: Ritual artifacts as symbolic habits. In: Giovagnoli, R., Dodig-Crnkovic, G. (eds.):

Habits and rituals, Special Issue, Open Information Science/De Gruyter, 2, 147 (2008)9. Magnani, L.: Abductive Cognition: The Epistemological and Eco-cognitive Dimensions of

Hypothetical Reasoning. Springer, Heidelberg/Berlin, cap 1 (2009)10. Dei, F., D’Orsi, L.: Emile Durkheim a hundred years later. In: Giovagnoli, R., Dodig-Crnkovic,

G. (eds.) Habits and Rituals, Special Issue, Open Information Science/De Gruyter, vol. 2,pp. 115–126 (2008)

11. Schwarz, E., Lappainen, J.: Collective phronesis: an investigation of collective judgment andprofessional action (forthcoming in this volume)

12. Lombo, J., Gimenez-Amaya, J.: The unity and stability of human behavior: an interdisciplinaryapproach to habits between philosophy and neuroscience. Front. Hum. Neurosci. 8, 607 (2014).https://doi.org/10.3389/fnhum.2014.00607

13. Bernacer, J., Murillo, J.I.: The Aristotelian conception of habit and its contribution to humanneuroscience. Front. Hum. Neurosci. (2014)

14. Graybiel, A.: Habits, rituals and the evaluative brain. Annu. Rev. Neurosci. 31, 359–387 (2008)15. Lowe, R.: Learning and adaptation: neural and behavioralmechanisms behind behavior change.

Connect. Sci. 50(Issue 1) (2018)16. James, W.: The Principle of Psychology. Dover, New York (1950)17. Seger, C.A., Spiering, B.J.: A critical review of habit learning and the Basal Ganglia. Front.

Syst. Neurosci. 5, 66 (2011)18. Lombo, J.A., Giménez Amaya, J.M.: La Unidad de la Persona. Aproximación Interdisciplinar

Desde la Filosofía y la Neurociencia. Pamplona, EUNSA (2013)19. Millán-Puelles, A.: Léxico Filosófico. Rialp, Madrid (2002)20. Wood, W.: The role of habits in self-control. In: Vohs, K., Baumeister, R. (eds.) Handbook of

Self-regulation, p. 104. Guilford, New York (2016)


21. Castelfranchi, C., Pezzulo, G.: Thinking as the control of imagination: a conceptual frameworkfor goal-directed systems. Psychol. Res. 73(4), 559–577 (2009)

22. Iacoboni, M., et al.: Grasping the intentions of others with one’s own mirror neuron system.PLoS Biol. (2005)

23. Pezzulo, G., et al. : The Challenge of Anticipation. Springer, Basel (2008); Pezzulo, G.,Castelfranchi, C.: The symbol detachment problem. Cogn. Process. 8(2) (2007)

24. Jeannerod, M.: Motor Cognition: What Actions Tell the Self. Oxford University Press, Oxford(2006)

25. Searle, J.: Making the Social World, p. 59. Oxford University Press, Oxford (2010)

Raffaela Giovagnoli Robert Lowe Editors

Documents

Transcript of Raffaela Giovagnoli Robert Lowe Editors