8/17/2019 delusi6

1/13

Grandiose delusions: A review and theoretical integration of cognitive andaffective perspectives

Rebecca Knowles ⁎, Simon McCarthy-Jones, Georgina Rowse

University of Shef  eld, UK 

a b s t r a c ta r t i c l e i n f o

 Article history:

Received 24 August 2010

Accepted 23 February 2011Available online 5 March 2011

Keywords:

Grandiose

Grandeur

Delusion

Bipolar

Mania

Schizophrenia

Psychosis

Grandiose delusions (GDs) are found across a wide range of psychiatric conditions, including in around two-

thirds of patients diagnosed with bipolar disorder, half of patients diagnosed with schizophrenia, as well as in

a substantial proportion of patients with substance abuse disorders. In addition, over 10% of the healthy

general population experience grandiose thoughts that do not meet full delusional criteria. Yet in contrast to

other psychotic phenomena, such as auditory hallucinations and persecutory delusions, GDs have received

little attention from researchers. This paper offers a comprehensive examination of the existing cognitive and

affective literature on GDs, including consideration of the evidence in support of  ‘ delusion-as-defence’  and

emotion-consistent’   models. We then propose a tentative model of GDs informed by a synthesis of the

available evidence designed to be a stimulus to future research in this area. As GDs are considered to be

relatively resistant to traditional cognitive behavioural techniques, we then discuss the implications of our

model for how CBT may be modied to address these beliefs. Directions for future research are also

highlighted.

© 2011 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685

2. What are grandiose delusions? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685

3. The epidemiology of grandiose delusions and beliefs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685

3.1. Prevalence of clinically relevant grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685

3.2. Demographic variables, culture and grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686

3.4. Prevalence of grandiose delusion-like beliefs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686

3.5. Diagnostic specicity of grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687

4. Grandiose delusions, persecutory delusions and depression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687

5. Affect and grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687

5.1. Affect in “delusion as defence”  models of GDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 688

5.2. Affect in emotion-consistent models of grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 688

6. Anomalous experiences, their appraisal, and GDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 689

7. Cognitive styles and grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 690

7.1. Jumping to conclusions bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 690

7.2. Attributional style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 690

7.3. Modality of thought . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6907.4. Thinking about thinking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691

8. The dynamic nature of grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691

9. Developing a model of grandiose delusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691

10. Implications for treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693

11. Future research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694

Clinical Psychology Review 31 (2011) 684–696

⁎   Corresponding author at: Clinical Psychology Unit, Shef eld University, Western Bank, Shef eld S10 2TN, UK. Tel.: +44 114 2226577; fax: +44 114 2226610.

E-mail address: r.knowles@shef eld.ac.uk (R. Knowles).

0272-7358/$ –   see front matter © 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.cpr.2011.02.009

Contents lists available at   ScienceDirect

Clinical Psychology Review

http://dx.doi.org/10.1016/j.cpr.2011.02.009http://dx.doi.org/10.1016/j.cpr.2011.02.009http://dx.doi.org/10.1016/j.cpr.2011.02.009mailto:r.knowles@sheffield.ac.ukmailto:r.knowles@sheffield.ac.ukmailto:r.knowles@sheffield.ac.ukhttp://dx.doi.org/10.1016/j.cpr.2011.02.009http://www.sciencedirect.com/science/journal/02727358http://www.sciencedirect.com/science/journal/02727358http://dx.doi.org/10.1016/j.cpr.2011.02.009mailto:r.knowles@sheffield.ac.ukhttp://dx.doi.org/10.1016/j.cpr.2011.02.009

8/17/2019 delusi6

2/13

1. Introduction

Recent approaches to psychopathology have shifted away from a

diagnostically driven approachtowards a focus on transdiagnostic efforts

to understand individual symptoms and processes (Bentall, 2006).

Dedicated cognitive models have been developed for many of the

experiences typically associated with psychosis, such as auditory

hallucinations (e.g.,   Beck & Rector, 2005; Bentall, 1990; Horowitz,

1975) and persecutory delusions (e.g.,   Bentall, Corcoran, Howard,Blackwood, & Kinderman, 2001; Freeman, Garety, Kuipers, Fowler, &

Bebbington, 2002), including integrativeframeworksthat seek to account

for the positive symptoms of psychosis together (e.g.,  Garety, Kuipers,

Fowler, Freeman & Bebbington, 2001; Morrison, 2001). Grandiose

delusions (GDs), by contrast, have received relatively little theoretical

or empirical attention. Indeed, one of the few recent studies to have

addressed theseexperiences directly concluded that “much remains to be

determined in understanding the formation and maintenance of 

grandiose delusions” (Smith, Freeman, & Kuipers, 2005, p. 486).

Achieving a better understanding of the onset and maintenance of 

GDs is likely to be benecial for a number of reasons. Models of 

persecutory delusions have informed the development of focused

cognitive behavioural interventions (Freeman & Garety, 2006), and

similarly tailored interventions for GDs are likely to be useful. It has

also been suggested that GDs may play a role in the development of 

persecutory delusions (Lake, 2008) and a better understanding of GDs

may thus contribute to the development of more effective interven-

tions for persecutory delusions. The present paper aims to review the

existing literature on the psychological mechanisms underpinning

GDs and to propose an integrated conceptualization of this phenom-

enon that is amenable to empirical testing. In particular, we aim to

summarise epidemiological   ndings, to evaluate the evidence for

delusion-as-defense and emotion-consistent accounts, and to consid-

er the role of cognitive biases in the development and maintenance of 

GDs. The present review is hence restricted to a consideration of 

cognitiveand affectivefactors. Whilstthereis a clear need for a review

of genetic, neurobiological and neurocognitive perspectives on GDs,

this is beyond the scope of the present review.

This reviewis informed by a systematic search of MEDLINE (1950–May 2010), PsychInfo (1967–2010) and Scopus (1823–May 2010)

databasesfor peer-reviewed articles on grandiosedelusions published

in English. The search string employed was  “(grandeur* OR grandi*)

AND (delus* OR belief)”. Each result was examined rst by inspection

of the title, and then, as required,the abstract and the full text. Studies

wereexcluded if theyfocused exclusively on grandiosity in the context

of personality disorders (e.g., Narcissistic Personality Disorder). The

reference sections and citation reports of papers identied by this

search were examined to identify further relevant papers. The

criterion for the inclusion of studies for the purposes of prevalence

estimates of GDswas a samplesize in excessof 50 participants, andfor

studies of psychiatric patients, that they be published after the

publication of DSM-III-R in 1994 in order to achieve some measure

of diagnostic comparability.

2. What are grandiose delusions?

GDs are dened as false beliefs about having inated worth,

power, knowledge or a special identity which are   rmly sustained

despite undeniable evidence to the contrary (APA, 2000). Some

examples are given in Table 1. GDs, like other delusional beliefs, are

multidimensional (Garety & Hemsley, 1994), varying with regard to

the degree of conviction and preoccupation, and the levels of distress

and dysfunction caused. GDs (alongside religious delusions) seem to

be held with the greatest conviction and tend to be associated with

less negative affect than other delusions (Appelbaum, Robbins, &

Roth, 1999). However, an apparent paradox given this increased

degree of conviction is Appelbaum and colleagues' report that GDs are

less likely to motivate individuals to act than other types of delusions.

This   nding is somewhat counterintuitive in the light of clinical

observations of patients engaging in risky and impulsive behaviour

fuelled by grandiose delusional beliefs.  Appelbaum et al.'s (1999)

nding may be a measurement artefact, since the  ‘action’  dimension

of the MacArthur-Maudsley Delusions Assessment Schedule is biased

towards the assessment of aggressive acts, meaning that respondents

would achieve low scores for not acting on their beliefs in an

aggressive or violent manner regardless of other behavioural con-

sequences. Because of their nature and content, we would not expect

the behaviours motivated by GDs to be of an aggressive or violent

nature, andso this assessment tool maybe inadequate forthe purposesof establishing links between GDs and any associated behaviour.

3. The epidemiology of grandiose delusions and beliefs

 3.1. Prevalence of clinically relevant grandiose delusions

GDsare among the mostcommonlyencountereddelusional beliefs.

In a study of 1,136 consecutively admitted psychiatric patients,

Appelbaum et al. (1999)   found that 43% of the 328 patients who

presented with delusional beliefs reported GDs. Only persecutory

delusions (78%) and delusions of body/mind control (60%) were more

commonly observed. Table 2 lists thestudies that have reported on the

prevalence of GDs in large (NN50) psychiatric samples since the

publication of DSM-III-R in 1994. Clear prevalence estimates are hardto establish since a wide range of self-report and clinician-rated

instruments with differing thresholds and cut-offs have been

employed. Prevalence comparisons are also made harder by the use

of both lifetime prevalence and point prevalence  gures (usually at

hospital admission), with some studies failing to report which of these

two methods was used.

The issue of whether patients are able to reliably report on their

own delusional beliefs is a further consideration. The absence of an

objective marker of delusional beliefs means that we are reliant on

patients' descriptions and clinicians' interpretations of these phe-

nomena, and a recent article by Lincoln, Ziegler, Lüllmann, Müller, and

Rief (2010)   suggests that patients with schizophrenia spectrum

disorders are reliably able to provide information about the presence

and type of their delusional beliefs.

 Table 1

Examples of grandiose delusions.

Example Source

“I was spitting on a light bulb, thinking if I watched the

saliva burn, the different colours and shapes, I could nd

the key to the cure to cancer”

Goodwin and Jamison

(1990, p. 26)

“I wouldwrite books on psychiatric theory… on theology. I

would write novels. I had the libretto of an opera in

mind. Nothing was beyond me… The major work which

would be based on this material would be accurate,provocative, and of profound signicance.”

Goodwin and Jamison

(1990, p. 29)

I can communicate and have a special relationship with

God. I am also the cousin of Tony Blair and I can  y.

Smith et al. (2005)

I am a special athlete and I run a national charity.   Smith et al. (2005)

I am God; I created the universe and I am son of Prince

Phillip. I am also a famous DJ. I have superman-type

powers.

Smith et al. (2005)

I have special luck and have wonthe lotteryfour times and

am owed £126 million.

Smith et al. (2005)

I am Roger Taylor from the rock group  “Queen”.   Smith et al. (2005)

I am the gang leader of drug dealers in Los Angeles. I have

great wealth.

Smith et al. (2005)

Believed he possessed the recording of a song he had

composed and performed that was   “worth millions of 

dollars.

Lake (2008)

Belief he had developed a   “Star Wars”   intercontinental

ballistic missile interceptor system, and had tried tophone Ronald Reagan to tell him.

Lake (2008)

685R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

3/13

In the only study to have examined the transdiagnostic prevalence

of GDs, Appelbaum et al. (1999) found that they were more frequent

in bipolar disorder (59%) than in schizophrenia (49%), substance

misuse disorders (30%) or depression (21%), although the statistical

status of these differences was not reported. Consistent with this

pattern, an earlier study by Junginger, Barker, and Coe (1992) found

signicantly fewer patients with schizophrenia (19%) than with

affective disorders (37%) presenting with GDs, although the failure to

apply a Bonferroni correction to the correlational analysis may have

yielded a false positive result. The   ndings of the other studies in

Table 2   are broadly in line with   Appelbaum et al. (1999), with

approximately two-thirds of patients with a diagnosis of bipolar

disorder and around half of patients with a diagnosis of schizophreniareporting GDs. Two studies, however, report much lower prevalence

rates.   Baethge, Baldessarini, Freudenthal, Streeruwitz, Bauer and

Bschor (2005) observed GDs in just 9% of patients diagnosed with

bipolar disorder, perhaps due to the use of point prevalence in a study

in which only 26% of patients presented with any form of delusion.

Breier and Berg (1999) also found a GD prevalence of only 9% in a

schizophrenia sample, which appears to result from the conservative

diagnostic criteria employed.

 3.2. Demographic variables, culture and grandiose delusions

There is some preliminary evidence of a relationship between age

of onset ofbipolardisorderandthe occurrence ofGDs. Carlson, Bromet,

and Sievers (2000) identied GDs in 74% of patients with early-onset(b21 years) bipolar disorder, but in just 40% of those who were aged

30 years or over at the time of their   rst episode, representing a

statistically signicant difference. There is no evidence that rates of 

GDs differ between men and women (de Portugal, González, Haro,

Usall, & Cervilla, 2010), although it is interesting to note that the

particular content of GDs may vary according to gender (Rudalevičienė,

Stompe, Narbekovas, Raškauskienė, & Bunevičius, 2008).

In a study of over 1000 individuals with a diagnosis of schizophrenia

from socioculturally diverse countries, grandiosity was found to be the

second most common delusional theme behind persecutory delusions

(Stompe, Karakula,Rudalevičiene, Okribelashvili, Chaudhry,Idemudia et

al., 2006). Stompe and colleagues also found that the prevalence of GDs

had remained broadly similar (38–44%) in Austrian patients over the

past 145 years despite the extensivesocietal changes during this period.

So GDs are evident across cultures, but there appears to be some cross-

cultural variation in the specic presentation (Suhail & Cochrane, 2002).

Several studies have compared GDs between European and Asian

patients.   Stompe, Bauer, Karakula, Rudaleviciene, Okribelashvili,

Chaudhry et al. (2007) reported signicantly higher frequencies of GDs

in patients with schizophrenia in Austria than in Pakistan, and they also

found that delusional grandiosity with a religious theme was especially

rare in Pakistan (2007). In contrast, when Suhail (2003) compared the

delusional beliefs of three groups of patients diagnosed with schizo-

phrenia  – a White British group living in Britain, a group of Pakistani

people living in Britain, and a third group of Pakistani people resident in

Pakistan – they found that the groups did not differ in thefrequencies of 

GDs. In fact, in this study, Pakistani people living in Pakistan were  morelikely to have a delusion about being a star/hero/famous person (32%)

compared to the other cultural/ethnic groups (b   10%). The authors

speculate that the large socio-economic disparities in Pakistan and the

dif culty in achieving upwardsocial mobility mayfuel delusional beliefs

about self-worth and achievements — a sort of self-defensive strategy.

There are also variations in the occurrence of GDs between ethnic

groups living in the same country. Yamada et al. (2006) studiedpatients

with psychotic disorders in the USA and found a greater prevalence of 

grandiose content in the delusions of European-American patients

(45%) compared to African-American (35%) and Latino (25%) patients.

Yamada and colleagues attempted to explain this pattern of ndings as

being due to the   “individualistic orientation associated with an

emphasis on uniqueness often associated with the Euro-American

culture”, whereas grandiosity might be   “culturally dystonic with thesocio-centric values of the Latino culture” (p. 165). It therefore appears

that cultural factors may inuence the prevalence and manifestation of 

GDs, although precise causal mechanisms are unclear.

 3.4. Prevalence of grandiose delusion-like beliefs

Grandiose beliefs which fail to meet full delusionalcriteria are found

in the general population. Indeed, as is the case for several other

psychotic phenomena (e.g., Johns & van Os, 2001), grandiose ideation

appears to exist on a continuum ranging from full-blown delusions

(held with conviction, resistant to change and causing signicant social

and occupational impairment) to more transient grandiose thoughts at

the other end of the spectrum. The most commonly used tool to assess

the presence of grandiose beliefs is the short-form version of the Peters

 Table 2

Prevalence of clinically relevant grandiose delusions (studies published post 1994, NN50).

Diagnosis Study Sample Diagnostic tool Delusion categorisation

Bipolar 59%   Appelbaum et al. (1999)   73 adults DSM-III-R DSM-III-R: Point prevalence

55%   Geller, Williams, Zimerman, Frazier,

Beringer and Warner (1998)

60 children

(7–16 yrs)

DSM-IV WASH-U-KSADS: Unspecied

9%   Baethge et al. (2005)   549 adults ICD-10 AMDP: Point prevalence

68%   Tillman, Geller, Klages, Corrigan,

Bolhofner and Zimerman, 2008

257 children,

(6–16 yrs)

DSM-I V WA SH- U- KSADS: Lifetime p reva lenc e

62%   Canuso, Bossie, Zhu, Youssef andDunner (2008)

515 adult patients DSM-IV PANSS score≥4 at time of study

88%   Conus, Abdel-Baki, Harrigan,

Lambert and McGorry (2004)

87 adult patien ts DSM-III-R RPMIP: point prevalence

47%   Goodwin and Jamison (1990)* 3801 manic patients various various

Schizophrenia 49%   Appelbaum et al. (1999)   138 patients DSM-IV DSM-III-R: Point prevalence

9%   Breier and Berg (1999)   1665 patients DSM-IV PANSS score≥5 at time of study

40%   Azhar, Varma and Hakim (1995)   270 patients ICD-9 Present State Examination: unspecied

Depression 21%   Appelbaum et al. (1999)   56 patients DSM-III-R DSM-III-R: Point prevalence

Alcohol or drug

disorders

30%   Appelbaum et al. (1999)   30 patients DSM-III-R DSM-III-R: Point prevalence

Alzheimer's   N1%   Hirono, Mori, Yasuda, Ikejiri, Imamura,

Shimomura et al. (1998)

228 patients DSM-IV DSM-IV: Point prevalence

6%   Migliorelli, Petraccam, Tesón, Sabe,

Leiguarda and Starkstein (1995)

103 patients DSM-III-R DSM-III-R: Point prevalence

Note: * This study was a review of 26 studies (published between 1922 and 1989) studies of manic patients, and is just included for reference as it is hard to make direct comparison

with present studies due to the range of diagnostic criteria and assessment tools these studies would have used.

686   R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

4/13

Delusion Inventory (PDI-21: Peters, Joseph,Day, & Garety, 2004). Scores

on these items from non-clinical populations are shown in Table 3.

It can be seen from Table 3 that rates of endorsement of grandiose

beliefs are higher in student samples than in the general population.

Given thatgeneral population samples have a signicantly higher mean

age than student samples, this difference may be explained by the

nding of Verdoux, van Os, Maurice-Tison, Gay, Salamon and Bourgeois

(1998) who reported a negative correlation between age and scores on

the grandiose subscale of the PDI-21. This would also be consistent withCarlson et al.'s (2000)  nding of a greater prevalence of GDs in early-

onset bipolar disorder. This pattern may be driven by the feelings of 

uniquenessand indestructibility (Elkind, 1967) that have been found to

peak in adolescence (Enright, Shukla, & Lapsley, 1980).

 3.5. Diagnostic speci city of grandiose delusions

There have been no direct attempts to establish whether GDs in

bipolar disorder and schizophrenia share a common aetiology and

phenomenology. One factor that may distinguish between GDs in the

psychotic and in the affective disorders is the widely-held but poorly-

evidenced assumption that GDs are mood-incongruent in the former,

and mood-congruent in the latter. If true, this could be taken to

suggest that GDs occurring in patients diagnosed with bipolar

disorder and schizophrenia, respectively, may have different phe-

nomenologies, aetiologies, and affective/cognitive/behavioural ante-

cedents, and thus require separate maintenance models.  Junginger

et al. (1992) reported a strong positive correlation between mood and

the presence of GDs in a mixed sample of psychiatric patients with

manic mood states being associated with GDs. However, the authors

didnot specify whether this correlation was signicant in both patient

groups or just those with affective disorder diagnoses.

Current psychiatric nomenclature retains Kraepelin's original dis-

tinction between the diagnostic categories of schizophrenia and bipolar

disorder (APA, 2000), but the clinical reality suggests that there is

extensive symptomatic overlap and comorbidity between the two

presentations (e.g.,  Laursen, Agerbo, & Pedersen, 2009) which means

thatit may not be meaningful to talk in terms of diagnostic distinctness.

Indeed, the diagnostic category of schizoaffective disorder (APA, 2000)captures those individuals whomeet criteria for both schizophrenia and

bipolar disorder. We have no reason to suspect that the origins and

maintenance of GDs will differ between individuals with different

psychiatric diagnoses, and so for the purposes of this review it will be

assumed that the same underlying psychological processes are of 

interest independent of diagnostic classication.

4. Grandiose delusions, persecutory delusions and depression

Qualitative studies of delusions have noted   “a complex set of 

interconnected themes drawing uniquely from several possible

domains”  (Rhodes, Jakes, & Robinson, 2005, p.383) and GDs appear

unlikely to exist in a pure and isolated form. They tend to occur most

frequently alongside persecutory delusions (PDs), although large-

scale studies are lacking.   Raune, Bebbington, Dunn, and Kuipers

(2006)  found that of 39 psychiatric patients who presented with

delusions, 54% reported PDs only, 10% reported GDs in isolation, and

33% reported having both PDs and GDs. Another recent study found

that of 14 patients diagnosed with non-affective psychosis, 68% had

 just PDs, 16% had just GDs, and 16% had both PDs and GDs ( Jolley,Garety, Bebbington, Dunn, Freeman, Kuipers et al., 2006). The high

comorbidity between GDs and persecutory delusions might lead to

the assumption that GDs in patients with schizophrenia diagnoses are

mood-incongruent (see   Section 4).   Lake (2008)   suggests that the

potential of a GD to contribute to the development of persecutory

ideation and thus to low mood may disguise the fact that it was

originally associated with positive affect. He argues that the strength of 

people's beliefs in their extraordinary possessions, powers or talents as

reected in typicalGDs, in conjunction with the broader effects of other

unusual attentional and reasoning processes, means that individuals

worry that otherswillwishthemillor try to steal their ‘gifts’ from them,

which then leads to the development of persecutory ideation. Such an

account is consistent with the co-occurrence of GDs with both

persecutory delusions and depression and highlights the need for

longitudinal studies of the course and emergence of delusional beliefs.

In the absence of larger-scale clinical studies, information about the

relationship between GDs and PDs can be obtained from analogue and

factor analytic studies.   Fowler, Freeman, Smith, Kuipers, Bashforth,

Coker et al. (2006) found that levels of grandiose beliefs were predicted

by levels of paranoia in a non-clinical student sample. However, this

correlational study was unable to establish whether this reected an

individual's tendency to experience delusion-like thoughts per se, or

whether there was a causal relationship between the two types of 

beliefs. Several factor analytic studiesalso show an association between

PDs and GDs. Bedford & Deary's (2006) analysis of 713 participants'

delusional symptoms found that grandiosity formed a distinct factor

whichcorrelatedsignicantlywith a separate persecutory beliefs factor.

Other factor analyses have concluded that GDs and PDs are at least

partially independent from one another (e.g.,   Kitamura, Okazaki,Fujinawa, Takayanagi, & Kasahara, 1998) and that PDs and GDs have

some non-shared causes (Freeman, 2007), although what these might

be remains poorly understood.

5. Affect and grandiose delusions

Contemporary psychological models of GDs reject Berrios's (1991)

claim that delusions are meaningless speech acts and propose instead

that GDsare related to past/current emotional concerns. Onefamily of 

modelssuggests that GDs arise from an individual's attempt to defend

themselves against negative affective states, which Freeman, Garety,

 Table 3

Grandiose beliefs in general population.

Study Sample Mean age

(SD, range)

% endorsing PDI-21 item  “Do you

ever feel that you are a very special

or unusual person?”

% endorsing PDI-21 item  “Do you ever

feel as if you are, or destined to be

someone very important

Verdoux et al. (1998)   462 attendees of a GP with no

psychiatric disorder

52 years (18, 18–95) 12% 8%

Peters, Joseph and Garety (1999)   2 72 gener al p op ulat ion 3 7 year s (1 0, 1 9–75) 43% data not reported

Armando, Nelson, Yung, Ross,

Birchwood, Girardi et al. (2010)

1,777 high school and university

students

18 years (4, 15–26) 75% 65%

 Jones and Fernyhough

(2007, unpublished data)

4 93 university stu dents 1 9 year s (2 , 18–54) 44% 36%

Scott, Chant, et al. (2006)   10,641 general population not reported 3.4% of people answered the question [which the authors aimed to

address the presence of grandiose delusions]  “Do you have any special

powers that most people lack?” in the af rmative.

687R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

5/13

Fowler, Kuipers, Dunn, Bebbington et al. (1998)   call   “delusion-as-

defense” (DAD) accounts.A secondgroupof modelstermed “emotion-

consistent” proposes that GDsemerge out of current positive affective

states (Smith et al., 2005, p. 486).

5.1. Affect in   “ delusion as defence”  models of GDs

Beck and Rector (2005) have argued that GDs  “may develop as a

compensation for an underlying sense of loneliness, unworthiness, orpowerlessness” (p. 588) and note from their clinical experience that

many patients with GDs   “have experienced prior life crises charac-

terized by a sense of failure or worthlessness” (p. 588). There is some

preliminary evidence that early traumatic life-events may be

associated with GDs (Read, Agar, Argyle, & Aderhold, 2003) but in a

sample of patients with psychosis,   Mason, Brett, Collinge, Curr, and

Rhodes (2009) only detected a trend towards an association between

grandiose beliefs and a composite measure of childhood trauma.

Neale's (1998) manic defense hypothesis – by which grandiose beliefs

(with other symptoms of mania) serve the function of keeping

distressing thoughts out of consciousness   –   is similar to   Beck and

Rector's (2005) argument. Some preliminary evidence in support of 

the hypothesis that GDsmay be understood to compensate for failure/

dissatisfaction with life can be found in qualitative study of delusional

patients, which concluded that   “a link can be made between

delusional themes and themes from personal goals”   (Rhodes &

 Jakes, 2000, p.221).

While the qualitative evidence is limited by the small number of 

studies, there is a greater volume of quantitative research in this area.

This body of work arose from the success of a paradigm for testing a

defensive account of persecutory delusions (PDs).   Bentall (1994)

proposed that PDswere a means of protecting against low self-esteem

and depression by preventing awareness of discrepancies between

actual and ideal self-concepts. This leads to the prediction that there

will be a measurable discrepancy between overt (explicit) and covert

(implicit) self-esteem in individuals with current PDs, a hypothesis

which has received some empirical support (e.g.,   Lyon, Kaney, &

Bentall, 1994) although results are mixed (e.g., Martin & Penn, 2002).

A similar paradigm has been applied to the evaluation of a defensiveaccount of GDs. As Smith et al. (2005) note, a DAD model makes more

intuitive sense in the context of GDs which appear to be better

candidates for protecting positive self-esteem than PDs. As they put it,

“believing yourself to be, for example, a famous talented individual is

more likely to reduce low self-esteem than thinking that the

neighbors are spreading distressing rumors about you and are

plotting to have you evicted” (p. 480).

To date only Smith et al. (2005) have attempted to test the DAD

theory of GDs by comparing explicit and implicit self-esteem in a

mixed group of 21 patients. Explicit self-esteem was measured by

self-report questionnaire, while implicit self-esteem was assessed

using an emotional Stroop task and the Self-Referent Incidental Recall

Task (Lyon, Startup, & Bentall, 1999). No evidence of low implicit self-

esteem in the patients with GDs was found. Raune, Bebbington, Dunn,and Kuipers (2005)   have also reported evidence that appears

inconsistent with a DAD model. If the DAD account is correct, then

events that threaten the individual's self-esteem would be expected

to be associated with the development of GDs which would arise as a

psychological defense. However, the authors found that GDs in arst-

episode psychosis sample were actually   negatively   associated with

loss events and they also failed to   nd a relationship between the

onset of GDsand recent humiliating events. It is worth noting that key

prospective studies have reported an increase in manic symptoms

after routine-disrupting and goal-attainment events, while negative

life events precede depressive episodes but do not appear to predict

symptoms of mania ( Johnson, 2005a,b; Johnson, Sandrow, Meyer,

Winters, Miller, Keitner et al., 2000; Johnson, Cuellar, Ruggero,

Winnett-Perlman, Goodnick, White and Miller, 2008). Although

these studies did not report specically on GDs, the  ndings further

undermine the plausibility of a DAD model of GDs.

However, these investigations are open to criticism. As the authors

themselves concede, Smith et al. (2005) only tested a defense against

verbally mediated thoughts, but a defense against negative feelings

might take a different form from the traditionally conceived verbal

delusional belief. In our view (see  Section 7.3), there are important

reasons to believe that a defense may need to counteract visual

thoughts or mental images which are commonly found in patientswith bipolar disorder — the clinical population in whom GDs are most

common (Tzemou & Birchwood, 2007).

Empirical tests may also have failed to  nd support for the DAD

model of GDs because of the stipulation that the defense operates

specically to protect self-esteem. Self-esteem is non-relational

concept, dened as an individual's perception of themselves.

However, delusional beliefs are typically interpersonal and embedded

in a social context. Hence, a more appropriate candidate for what is

being defended by the emergence of GDs may be  “social self-esteem”

(Heatherton & Polivy, 1991) or social rank (Gilbert, 1992).  Smith,

Fowler, Freeman, Bebbington, Bashforth, Garety et al. (2006) argue

that negative beliefs about others held by individuals with GDs may

serve to increase social rank. It has been suggested that grandiosity

could increase the social status of the individual who expresses it by

generating an impression of enhanced access to resources which

favors social success (McGuire & Troisi, 1998). That is, GDs may

represent the exaggeration of an adaptive coping mechanism. The

qualitative work of   Rhodes and Jakes (2000),   who found that an

individual's GD was related to a real-life theme of failure might be

usefully interpreted in this light.  Birchwood, Trower, Brunet, Gilbert,

Iqbal and Jackson (2006)  argue that when individuals without the

social power to protect themselves are alerted to their low relative

rank they may activate internal defensive emotions and strategies, of 

which GDs are an example. If this is the case then we would expect

individuals with GDs to display a fear of negative evaluation by others,

which may also be linked to the development of persecutory

delusions. Furthermore, since social rank theory predicts that shame

and outsider statusis associated with socialanxiety (Gilbert & Trower,

2001) we might also expect these features to be associated with GDs.Given that only one study has directly examined the relationship

between GDs and implicit/explicit self-esteem discrepancies, it is

important to consider other potential sources of evidence. Explicit

self-esteem has been found to exceed implicit self-esteem in currently

manic and remitted patients (Lyon et al., 1999) and in an analogue

study of hypomania (Bentall & Thompson, 1990). Bentall, Kinderman,

and Manson (2005) examined self-discrepancies in manic, depressed

and remitted bipolar patients as well as a group of healthy controls.

They found that manic patients rated their actual self as being closer

to their ideal self than any of the other groups, why they interpreted

as being consistent with the proposal that a defense was operational

in the manic state to prevent the negative affective consequences of 

actual-ideal self-discrepancies.

5.2. Affect in emotion-consistent models of grandiose delusions

Analternativeto theDAD pathway is thesuggestionthatdelusional

beliefs arise from current concerns (Freeman et al., 2002), referred to

as an  “emotion-consistent” account (Smith et al., 2005, p. 486). This

model suggests that grandiose beliefs are built on   “existing or

preserved raised areas of self-esteem”   (Smith et al., 2005, p. 481).

These positive beliefs about the self may become exaggerated against

the backdrop of the positive affective state, and may be uncritically

accepted due to cognitive and information processing biases (see

Section 7.1). Thepositive affective state mayalso be ampliedby other

processes such as the occurrence of mood-congruent mental imagery

(see  Section 7.3). Smith et al. (2006)  have further proposed that a

combination of elevated mood and positive views of the self alongside

688   R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

6/13

negative evaluations of others may promote a social position that

sustains positive self-beliefsand a rejection of socialcues,whichcould

in turn maintain GDs.

The limited available evidence is broadly in support of this

emotion-consistent model. Two separate clinical studies have

reported that levels of grandiose beliefs were correlated with higher

explicit self-esteem and lower depression scores (Moritz et al., 2010;

Smith et al., 2006). Smith and colleagues suggested that individuals

with GDs tend to make negative evaluations of others in order to“foster a social position that maintains positive self-evaluations”  (p.

183). Although there was no correlation between negative evalua-

tions of others and levels of GDs, a logistic regression showed that

both low depression scores and negative beliefs about others were

independently associated with GDs (Smith et al., 2006). An analogue

study also found that levels of grandiose beliefs were predicted by

positive (but not negative)views of theself (Fowleret al., 2006). None

of these studies is able to establish the potential direction of causation

in the relationship between explicit self-esteem and grandiose beliefs:

higher explicit self-esteem may facilitate the development of mood-

congruent grandiose beliefs, but a DAD interpretation of these

ndings could be that grandiose beliefs arise as a defense against

low self-esteem, resulting in increased positive self-esteem.  Raune

et al.'s (2005) study which found fewer negative loss events in the

recent history of those individuals with GDs offers some basis for

favouring the emotion-consistent interpretation of the results,

although the (limited) evidence discussed in   Section 5.1   linking

traumaand GDs leavesthis question open. We note that at least oneof 

the predictions of an emotion-consistent account appears to be borne

out by empirical research. Given that feelings of uniqueness,

indestructibility and heroicness peak in adolescence, the emotion-

consistent model would predict a higher incidence of GDs in younger

adults than in older people  —  a pattern that has been observed.  Lake

(2008) has argued for a key role of positive mood in the aetiology of 

GDs, and interestingly, the clinical vignettes in his study appear to

support a relatively underdeveloped facet of the emotion-consistent

account of GDs.   Freeman and Garety (2003) suggest that GDs may

build on “pre-existing inated, or accurate, perceptions of the self ” (p.

938, italics added), and Smith et al. (2005, p.481) note that GDs maybuild on   “existing or preserved raised areas of self-esteem”. This is

consistent with Lake's examples of the chemistry graduate who

believed that he possessed a formula to make synthetic narcotics, and

the patient with a background in rocket engineering who believed

that he had a Star Wars missile design. Lake describes these beliefs as

building on a  “thread of truth” (p. 1153).

GDs can be very specic beliefs related to particular areas of 

positive self-esteem expertise or achievement. However, as is clear

from Table 1, they may also be much broader and less grounded in

reality, and the differences between the two types of beliefs is unclear.

It may be that if an individual does not have a clear current/past

strength on which to draw, then a GD may involve global self-esteem

being elevated. Alternatively, it may be that broad GDs start off as

more specic beliefs that escalate and expand due to the effects of various cognitive mechanisms as discussed in Section 7.

In conclusion, the current empiricalevidenceappears to offer more

support for an emotion–consistent account of GDs than for a DAD

model, although a series of clinical vignettes and the one qualitative

study in this area suggest the DAD model may also be applicable,

highlighting the need for further longitudinal research into the onset

and development of GDs.

6. Anomalous experiences, their appraisal, and GDs

Psychological models of delusional ideation typically argue that

the search for meaning following an initial anomalous experience

leads to the emergence of a delusional explanation ( Maher, 1988),

and an account of this type may be applicable to GDs. One candidate

for the initial anomalous trigger experience is an individual's internal

state.  Mansell, Morrison, Reid, Lowens, and Tai (2007) have argued

that a range of mental disorders may be characterized by intrusions

into awareness of information (in the form of body state information,

affect, thoughts and images or external sensory input, for example)

which is appraised in an unusual manner. Mansell et al. (2007)further

note that the particular appraisal made about an intrusion is

inuenced by existing beliefs resulting from individual life experi-

ences. Mansell and colleagues have applied this idea primarily to theemergence of manic mood states, but it may also be relevant to the

development of GDs.

An example of the problematic appraisal in action may look some-

thing like this. An initial event (e.g., an exciting life event, or stimulant

use) leads to a more positive mood, greater physiological arousal, or

altered cognition. This experience is interpreted by an attribution of 

extreme personal meaning, rather than an external or situational

attribution (Mansell et al., 2007). For example, experiencing one'srate

of thinking as being increased might be interpreted as a sign of great

intelligence, wit and intuition (a personal attribution), or it could be

interpreted as a consequence of managing too many competing tasks

(a situational attribution) (Mansell et al., 2007; Jones, Mansell, &

Waller, 2006). Similarly, an individual who suddenly feels alert, active

and nds they have a reduced need for sleep may make sense of this

sensation by assuming that it is due to dispositional characteristics

such as their underlying dynamism or ability as opposed to being the

result of excessive stimulation from the environment ( Jones et al.,

2006). The tendency to make extreme personal attributions has been

found to be more common in patients with bipolar disorder (who are

vulnerable to hypo/mania and grandiose ideation) than in healthy

controls ( Jones et al., 2006).

An individual may reach for an internal personal appraisal of their

unusual experience because of an image they hold of an aspirational

“imminent possible self ”  (Mansell et al., 2007, p. 523). For example,

individuals who are vulnerable to mania exhibit higher levels of 

aspirations for fame, wealth and political inuence ( Johnson & Carver,

2006), and the greater the extent to which patients with bipolar

disorder value and perceive themselves as dynamic, creative and

successful, the more likely they are to relapse into hypomania ormania (Lam, Wright, & Sham, 2005). Similarly, hypomania scores in

students have been found to predict their expectations of academic

and career success (Meyer & Krumm-Merabet, 2003). Drawing on

studies of clinical mania (e.g., Meyer, Johnson & Carver, 1999) these

ndings may be interpreted as evidence of heightened responsiveness

of the behavioural activation systems in individuals. Additionally,

thinking in terms of the DAD model discussed earlier, if an individual

has a strongly positive ideal self, then any attempt to reduce the

discrepancies between this and the actual self should lead to the

adoption of positive self-representations. Hence, a positive imminent

or ideal self may act as a risk factor for GDs.

Mansell et al.'s (2007) ascent behaviour construct helps to explain

how these initial ideas may escalate into full-blown GDs. Ascent

behaviours aim to either enhance or control anomalous internalstates,such as feeling unusually alert or active. Mansell and colleagues list a

number of possible ascent behaviours including  “extended wakeful-

ness; increased rate of activity; generating multiple ideas and goals,

the seeking of social stimulation; and the dismissing of others’

attempts to moderate behavioural changes”   (p. 523). They give an

example of thelatter behaviour in thecase of individual whoappraises

his racing thoughts as a manifestation of his superior intelligence, and

who subsequently dominates social interactions and ignores negative

feedback from others. This particular ascent behavior, ignoring

negative feedback from others, has been further examined in relation

to bipolar disorder by  Mansell and Lam (2006)  who reported that

individuals with current symptoms of bipolar disorder were less likely

than remitted participants to use advice from others to inform their

decisions after having undergone an experimental mood elevation.

689R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

7/13

The internal state may also be amplied resulting in a GD through

the process of goal-setting and the impact of goal-attainment. Again,

there are at present no studies that pertain directly to GDs, but a

number of suggestive  ndings come from studies of bipolar mania.

Individuals with a history of mania have a tendency to set themselves

highergoals than control participants( Johnson, 2005a,b), and current

levels of hypomania have been found to predict a greater positive

affective response to achievinga goal as well as thesubsequent setting

of more extreme goals ( Johnson, Ruggero, & Carver, 2005). Thesendings are consistent with evidence that it is goal-achievement life

events rather than positive events more generally that are associated

with increased manic symptoms ( Johnson et al., 2000). This is

interesting in the light of  Lake's (2008) case examples which describe

the emergence of grandiose beliefs about achievements in particular

elds of personal expertise.

Another candidate for the initial anomalous triggering experience

might be the perception of unexpected, unsolicited or undue attention

from others. If other people are experienced as paying more attention to

the individual than usual (perhaps leading to comorbid delusions of 

reference), impaired theory of mind (ToM) may lead an individual to

explain this behaviour by inferring that it must be because s/he is special

and worthy of this additional attention. ToM performance in the context

of persecutory delusions has been the subject of extensive empirical

research (see  Freeman, 2007, for a review), but GDs have not been

considered in the same way. ToM has, however, been studied in the

contextof bipolarmania.Kerr, Dunbar,andBentall (2003)found thatToM

was impaired in currently manic patients as well as bipolar depressed

patients, but not in a remitted group. It remains unclear whether ToM

becomes impaired as a result of a manic mood state, or whether ToM

decits are a risk factor for mania and hence perhaps for GDs.

Not only may individuals with GDs misinterpret the interest and

attention of others, but they may also react more strongly to such

perceived attention. Evidence from studies of patients with bipolar

affectivedisorder reveals increasedneuralactivity in response to human

smiles in brain areas associated with positive affect and reward ( Rolls,

2000). Chen,Lennox, Jacob,Calder, Lupson, Bisbrown-Chippendale et al.

(2006) reportevidenceof abnormal brain activation in patients in manic

and depressed states of bipolar disorder in response to moodincongruent affective stimuli. Manic patients were found to have

abnormal neural activation in response to images of sad faces, and

interestingly theyshowed a differentiated patternof neuralactivationto

implicit and explicit presentations of sad face, unlike the unipolar and

bipolar depressed patients. Specically, Chen et al. found that a number

of areas were underactivated by explicit processing of sad faces but

overactivated by implicit processing of these faces in the currently

manic patients. These were the amygdala, insula, and superior and

middle temporal gyri (involved in emotional arousal and perception),

and the hippocampus, dorsal anterior cingulate and medial superior

frontal gyrus (involved in emotional regulation and sensitivity to

affective incongruency). This pattern is consistent with the nding that

bipolar patients in a manic episode are worse at identifying negative

facialaffect (Lembke& Ketter, 2002).ThismaybeonewayinwhichGDsare maintained, through the failure to consciously process incongruent

emotionalstates in others. Giventhe apparent failure of individuals with

mania to attend to negative social feedback, it would be particularly

interesting to replicate this study to examine whether there are

differences in the implicit and explicit processing of disapproving

faces in individuals with GDs.

7. Cognitive styles and grandiose delusions

7.1. Jumping to conclusions bias

Freeman (2007) asserts that if delusions are incorrect/uncorrected

beliefs, then we must understand the reasoning processes involved in

the formation and maintenance of these false beliefs. The presence of 

reasoning biases such as  ‘ jumping to conclusions’ (JTC) is well docu-

mented in association with delusional beliefs in general (see Garety &

Freeman, 1999), although to date, no studies have investigated this

bias in relation to GDs in particular. Previous research has tended to

combine patients with GDs and patients with persecutory delusions

(PDs) into a mixed delusional group for comparison with a non-

delusional control group (e.g., Dudley, John, Young, & Over, 1997) and

analysis by type of delusion is unfortunately often prevented by the

small sample sizes. However, the majority of research into decision-making processes and delusions has explored purely cognitive

mechanisms without considering the impact of emotion.   Mansell

and Lam (2006) found that individuals with bipolar disorder showed

impaired decision-making (characterized by failure to utilize social

feedback) in an induced elevated mood state, but there has been

almost no research into how positive affect might inuence cognitive

biases and decision-making.

7.2. Attributional style

In addition to the reasoning processes discussed above, it has been

suggestedthat GDsmay be associated with a self-servingattributional

style, according to which individuals tend to make more internal

attributions for positive events (Freeman et al., 1998). However,

research on the specic relationship between attributional style and

GDs in isolation from other types of delusional belief is very limited.

This may be partly due to the signicant comorbidity between GDs

and PDs and the tendency for researchers to recruit mixed groups of 

patients for comparison with a healthy control group (e.g.,   Fear,

Sharp, & Healy, 1996; Sharp, Fear, & Healey, 1997). Jolley et al. (2006)

have come closest to a specic investigation of attributional style in

GDs in their study of 71 patients with a diagnosis of non-affective

psychoses. They predicted that patients with persecutory beliefs

would form grandiose and depressed subgroups displaying self-

serving and depressive attributional styles, respectively. Jolley and

colleagues found that higher levels of grandiose beliefs were

associated with a greater self-serving attributional bias, but they

failed to apply a Bonferroni correction for the number of correlations

calculated. The only other specic association with grandiose beliefswas detected in a small subsample of patients (N=16) who showed

an externalizing bias for negative events. In this group, there was a

correlation between the extent to which people made external

attributions of negative events and the severity of their grandiose

beliefs.

There has also been a lack of research into how elevated mood

affects attributional style, and whether elation enhances the self-

serving bias. However, one study has examined how goal attainment

affects attributional style (Stern & Berrenberg, 1979). Students who

scoredhighlyon a measure of hypomania were more likelyto attribute

their apparent success to internal factors. After an initial success,

students also exaggerated their likelihood of correctly guessing the

outcome of a coin toss. Attributional style maybe thereforebe affected

by goal attainment in the context of GDs.

7.3. Modality of thought 

Given the key role of positive affect in GDs, factors that amplify

positive emotion arelikely to play a role in the escalation of mood and

grandiosity as well as in the maintenance of GDs. The role of ascent

behaviors has already been discussed, and one further factor is the

modality of thought used. Holmes and Mathews (2005) have argued

that emotional processing in the brain is particularly sensitive to

visual mental imagery — more so than to verbal thought. They suggest

that imagery susceptibility (the tendency to be a  “visualizer”  rather

than a   “verbalizer”) may be a neglected risk factor for psychiatric

disorders due to the amplifying effect of imagery on emotion and

several authors have shown that visual mental imagery has a greater

690   R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

8/13

amplifying effect on positive affect than verbal thought (e.g.,  Holmes,

Lang, & Shah, 2009).

Although no empirical studies have examined the role of mental

imagery in GDs, its potential for a role in hypomania is suggestive. For

example, intrusive images of positive future events are commonly

observed in hypomania (Gregory, Brewin, Mansell & Donaldson,

2010), many of which involve   “goal-oriented events that were

interpersonal in nature, for example  ‘being in charge of a project, in

an all-powerful situation with people underneath me’  or 'self as agreat business man, people looking up to me’. Interestingly, depressed

patients in the same study reported intrusive goal-orientated mental

images of a negative emotional valence. This suggests that GDs might

arise from depressive mood statesif self-esteem shiftsand thecontent

of mental imagery changes. This proposal is consistent with the

nding that self-esteem in bipolar patients is unstable compared to

healthy controls and patients with unipolar depression (Knowles, Tai,

 Jones, Higheld, Morriss and Bentall, 2007).

The tendency to use mental imagery may interact with the goal-

setting and achievement patterns characteristic of individuals who

are susceptible to mania and which may play a role in the

development of GDs. If goals are simulated as mental images then

this could stimulate mood elevation according to the imagery-as-

amplier hypothesis (Holmes & Mathews, 2010). The   “pre-experi-

encing”  of the future via imagining positive goal attainment could

amplify mood and contribute to the development of GDs. This

tendency to pre-experience goals or behaviours may help to explain

the high level of conviction with which GDs are typically held

(Appelbaum et al., 1999): given the commonalities in neural

activation between perception and mental imagery, the representa-

tion of GDs in imagery may enhance people's judgments of their truth

value.

7.4. Thinking about thinking 

The apparent role of mental imagery in the amplication of 

emotion means that it is important to consider the role of both the

modality and method of thinking in relation to GDs. A distinction is

made between rumination and reection as cognitive processingstrategies. Rumination describes a repetitive focus on the content,

causes and consequences of events (Gruber et al., 2009). It often

involves taking a rst-person perspective, resulting in the experience

of being immersed in mental events (Kross, Ayduk, & Mischel, 2005).

In contrast, reection has a cognitive distancing effect as events are

viewed more objectively from a third-person perspective. In PTSD,rst-person perspective memories are associated with greater levels

of affect than are third-person memories (McIsaac & Eich, 2004), and

when asked to recall a happy autobiographical memory, both patients

with bipolar disorder and healthy controls show greater positive

affect, positive thoughts and faster heart rates when they are

instructed to think about the memory in a ruminative manner as

opposed to a reective manner (Gruber et al., 2009). Although

rumination is more commonly associated with verbal thought thanwith mental imagery (McLaughlin, Borkovec, & Sibrava, 2007),

individuals who have a tendency to think ruminatively using   rst-

person imagery may be at greater risk for the development of GDs.

Furthermore, the   nding that ruminating about past events often

leads to rumination about futureevents (McLaughlin et al., 2007) may

be linked to the future- and goal-orientated thinking often found in

mania.

In addition to cognitive style, meta-cognitive style may also play a

role in the development and maintenance of GDs. Although studies of 

meta-cognition and grandiosity have not been conducted in clinical

samples, one analogue study has examined grandiose ideation in the

general population.   Larøi and Van der Linden (2005)  found that

cognitive self-consciousness (the tendency to focus attention on and

monitor one's own thinking) predicted levels of grandiose ideation.

Interestingly, other forms of metacognitive beliefs were the best

predictors for proneness to all other types of delusion-like beliefs. The

authors concluded that this could be interpreted to mean that   “the

presence of such intrusive thoughts and beliefs (i.e. having grandiose

ideas) may not be incompatible with the subject's metacognitive

beliefs (therefore not creating an aversive state of arousal). Indeed,

the presence of such experiences may even lead to a state of positive

affectivity” (p. 1437). In this way, people experiencing GDs may not

only appraise internal state changes by inferring extreme personalmeaning (Mansell et al., 2007), but they may do so by thinking in a

modality   that amplies affect (imagery), in a  manner  that amplies

affect (1st person rumination), with more positive  content  (future-

orientated goal-directed thoughts), and they may also be more

consciously aware of such thoughts when they occur.

8. The dynamic nature of grandiose delusions

One factor that has been overlooked in much of the research into

delusional beliefs is their dynamic nature.  Appelbaum, Robbins, and

Vesselinov (2004) studied the stability of patients’ delusions over the

course of a year in a mixed group of 405 delusional patients by

interviewing them every 10 weeks. Although they did not report data

on GDs in particular, a number of   ndings are of interest. They

reported that only 15% of patients were delusional at every follow-up

interview. Furthermore, there was only a 61% chance that a patient's

primary delusion type would remain the same between any two

interviews. Appelbaum et al. concluded that  “[d]elusions appear to be

more dynamic and uid over relatively short periods of time than has

been suggested by many classic descriptions and contemporary

formulations”  (p. 323). One underlying cause for both the primary

type of delusion and whether an individual endorses a delusion at all

might be the instability of a number of cognitive and affective states,

such as the  uctuation in paranoia that results from  uctuations in

self-esteem for example (e.g.,  Thewissen, Myin-Germeys, Bentall, de

Graaf, Vollebergh and Van Os, 2007; Thewissen, Bentall, Lecomte, van

Os, & Myin-Germeys, 2008).

Integrating the instability of delusional beliefs and the association

of unstable self-esteem with paranoia, we might hypothesize that notonly are GDs unstable, but they are so due to instability in self-esteem

or social rank. There is already evidence that patients with bipolar

disorder (diagnostic category most strongly associated with GDs)

have greater day-to-day uctuations in self-esteem than both healthy

controls and patients with unipolar depression (Knowles et al., 2007).

This hypothesis remains in need of testing more specically in groups

of patients with GDs, and we suggest that ESM would be an

appropriate methodology. We would expect GDs to be associated

with momentary positive  uctuations in self-esteem and social rank.

As discussed above, these may not be global   uctuations in self-

esteem but may instead be specic to a particular   “island”  of self-

worth, such as one's skill as a entrepreneur or prowess at football for

instance. These   uctuations in self-esteem may also inuence

cognitive processes like the JTC bias. We would predict that negativeuctuations in mood and/or self-esteem would be associated  rstly

with a reduction in the degree of conviction in the GDs, and that

further falls in mood/self-esteem might be associated with the

development of persecutory delusions.

9. Developing a model of grandiose delusions

The factors discussed above may be congured into a tentative

model of the development and maintenance of GDs. Various empirical

ndings have indicated that persecutory and grandiose delusions are

the result of distinct yet related psychological processes but at the

moment there is insuf cient evidence to determine whether GDs

differ from other types of delusional beliefs with respect to key

aetiological and maintaining factors and processes. And so, while this

691R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

http://dx.doi.org/10.1016/j.brat.2010.04.005http://dx.doi.org/10.1016/j.brat.2010.04.005http://dx.doi.org/10.1016/j.brat.2010.04.005http://dx.doi.org/10.1016/j.brat.2010.04.005

8/17/2019 delusi6

9/13

tentative model may be adaptable to other classes of delusional

beliefs, it has been congured specically with GDs in mind. We

would expect the content, direction and valence of attributional

biases, mental images, life events, internal state changes and self-

esteem   uctuations to vary between different types of delusion,

although the over-arching framework might still apply. It is clear from

the evidence discussed above that discrepancies between implicit and

explicit psychological states are common, and contemporary theories

of affect (e.g., Cacioppo & Bernston, 1994; Watson & Tellegen, 1985)allow for the simultaneous occurrence of contrasting emotional

experiences. The proposed model is principally concerned with the

explicit versions of affect and cognition that are overtly accessible and

amenable to self-report.

The balance of the current evidence is slanted in favour of an

emotion-consistent account of GDs, although there is some evidence

consistent with a DAD model (Beck & Rector, 2005; Rhodes & Jakes,

2000), and both routes are therefore reected in the model. We

propose that the research to-date has been overlyfocused on tryingto

determine whether   either   an emotion-consistent   or   a delusion-as-

defense account of delusional beliefs is ‘true’. In fact, these may not be

mutually exclusive mechanisms or pathways, and indeed they might

each imply the other's internal logic. What we present in Fig. 1 is an

attempt to integrate the two theoretical accounts of the development

of GDs in such a way as to illustrate the potential interplay between

the routes as well as being amenable to empirical testing.

A personwho is experiencinglow mood, lowself-esteem/social rank

is likely to be motivated to modify these feelings and may therefore be

primed to detect any sign or evidence of change/improvement. They

may detect an externally generated   ‘chink of light’   in the form of a

positive event or an internal state change (e.g., achieving a goal,

physiological/ cognitive changes due to drug use or fatigue) which they

seek to make sense of in dispositional terms, their internal positive

attributions being inuenced by earlier life events or cultural factors, or

perhaps by their desire to feel better and motivation to seize

opportunities to modify their negative mood. In this way, the process

might be both emotion-consistent (the initial grandiose thought iscongruent with an initial positive change in mood) and defensive

(occurring within a broader context of negative affect and self-esteem

that the individual is motivated to change). The initial grandiose

thought might then develop into a full-blown GD by the operation of 

cognitive biases, thefailure to process socialfeedback, and possible ToM

impairments. Throughout this phase, a number of factors are likely to

play a role in the further amplication of positive affect and the internal

state change, including the use of ruminative   rst-person mental

imagery relating to future goal attainment.

From a DAD perspective, grandiose thoughts are proposed to arise

as a deliberate defense against a perceived decrease in mood, self-

esteem or social rank. The precise mechanisms underlying such a

process remain unclear, but may involve ruminating on past

achievements or on autobiographical memories of episodes when

self-esteem or social rank were more positive (perhaps again using a

mental imagery modality), leading to small improvements in mood

and self-esteem and internal state changes which may then be

appraised in positive dispositional terms, initiating the same mood-

 Delusion as

defenceroute

 Emotion-consistent route

Earlier life

events

Ruminative,

1st

person

visual mental

imagery

Cognitive biases,

ToM impairment,

and ascent

behaviours

Precipitating event(e.g., goal achievement, substance use)

Positive internal state change(e.g., positive mood change, increase

self-esteem or social rank, change in

cognitions, such as speeded up

thoughts) amplification

Search for meaning

Appraisal

Positive internal attribution

Range of possible

future selves

Cultural

factors

Seeds initial grandiose thought

Grandiose delusion

Persecutory delusion

Unstable self-esteem:

Negativefluctuation

Unstable self-esteem:Positive

fluctuation

Previously documented routes topersecutory delusion formation

(see Bentall et al., 2001; Freeman2007)

Negative life

events

involving

threatened

self-esteem

or social

rank

Fig. 1. A preliminary model of grandiose delusions.

692   R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

10/13

congruentprocess as describedabove. In this way, thetwo supposedly

contrasting pathways into grandiosity may in fact operate alongside

each other, feeding into each other rather than being mutually ex-

clusive. Clearly, not everyone who makes internal attributions for

their positive internal state changes develops GDs so this is a nec-

essary but not suf cient part of the process of onset, and a complex

interplay of situational, cognitive and affective factors contributes to

the emergence of GDs in a vulnerable individual.

The model allows for two ways in which persecutory ideation mayrelate to GDs, centring aroundunstableself-esteemand socialrank.As

a result of negative   uctuations in self-esteem and/or social rank,

individuals may believe that a facet of their (grandiose) worth is

coveted by others, resulting in secondary persecutory delusions(PDs).

Conversely, GDs may also emerge from existing PDs, since positive

uctuations in self-esteem or social rank may encourage the appraisal

that the negative intentions of otherstowards them aredue to thefact

that they have a special talent/value/wealth that others wish to steal.

It is clear that at present, due to a lack of empirical studies in this area,

this model is tentative and in need of extensive empirical testing.

However, we believe that it offers a fruitful range of hypotheses that

researchers may wish to explore. A number of hypotheses can be

derived from this model, but just three specic examples are given

here. First, this model would predict that internal state changes

precede the onset of GDs. This could be tested using Experience

Sampling Methodologies (ESM: Hurlburt & Heavey, 2006) to examine

the affective and cognitive precursors to grandiose beliefs, as well as

the role of appraisals and attributional style. The model also predicts

that the relationship between GDs and persecutory delusions may be

mediated by unstable self-esteem. ESM could be used to examine

whether  uctuations in self-esteem precede the transition between

GDs and persecutory delusions. A third hypothesis derived from the

model is that the development and maintenance of GDs would be

associated with impaired ToM, which could be tested by comparing

the performance of GD, non-GD delusional and control groups on

standard ToM measures.

10. Implications for treatment

GDs present a particular problem for psychological intervention,

predicting poor clinical outcome up to ten years later (Thara & Eaton,

1996) as well as being negatively associated with medication compli-

ance (Appelbaum & Gutheil, 1980). Cognitive Behavioural Therapy

(CBT) has been found to be effective in ameliorating the dif culties

associated with schizophrenia-spectrum disorders (Turkington, Dudley,

Warman, & Beck, 2004), but the results for bipolar disorder are less

encouraging (Scott, Paykel, Morriss, Bentall, Kinderman, Johnson et al.,

2006; Lam, 2006). A number of factors have been suggested to account

forthe somewhat disappointing evidence forthe effectiveness of CBT for

bipolar disorder (such as the number of previous episodes, Scott, Chant,

et al., 2006; Scott, Paykel, et al., 2006), but theprevalence of GDs may be

an additional consideration. Although CBT is effective for treating the

positive symptoms of psychosis (Zimmerman, Favrod, Trieu, & Pomini,2005), littleworkhas focused on specic psychoticexperiences. In oneof 

thefewsuchstudies,persecutorydelusions and unusual thought content

but not grandiosity were highlighted as being most reduced following

CBT interventions(Kuipers et al., 1997). Indeed, littleis known about the

specic application of CBT to GDs.

CBT appears to be less effective at preventing manic episodes than

at preventing depressive relapse in bipolar disorder (Lam, Hayward,

Watkins, Wright, & Sham, 2005), and the sense of a   “hyperpositive

self ” –   the belief that one is creative, dynamic or entertaining for

example –  predicts a worse response to cognitive therapy in patients

diagnosed with bipolar disorder (Lam, Wright, & Sham, 2005).

Furthermore, given that GDs are associated with greater conviction

than many other forms of delusions (Appelbaum et al., 1999), the

initial success of cognitive approaches to modifying the belief and

engaging the patient in certain behavioural strategies such as reality

testing may be limited. Early collaboration and working towards a

shared understanding are crucial to successful therapeutic outcomes

in CBT (Gilbert & Leahy, 2007) and such therapeutic factors are thus

perhaps even more salient when working with GDs, as is the

consideration of the emotional valence of these beliefs. GDs are

typically associated with positive explicit affect and positive self-

beliefs (e.g., Smith et al., 2006) which may hinder the development of 

a shared rationale for a therapeutic intervention.Based on the model proposed in Fig. 1 we might expect a number

of cognitive and behavioural strategies to be of use in addressing the

dif culties associated with GDs. The proposed role for adverse early

life events in the evolution of GDs suggests that trauma-focused

interventions should be considered. This therapeutic approach would

also be expected to have an impact on self-esteem, mood and coping

styles which may also moderate the role played by GDs in an

individual's presentation. These therapeutic targets need not just be

addressed as part of trauma-focused work, but may also constitute

individually tailored, discrete interventions. If there is a role for ToM

impairments to lead to mistaken attributions of others' intentions,

then we might expect patients to benet from interventions which

specically target this domain, as used successfully in other popula-

tions (e.g.,   Ashcroft, Jervis, & Roberts, 1999). Although current

evidence is limited, other cognitive biases that may be implicated in

GDs (such as attributional style and the JTC bias) might be amenable

to the application of CBT techniques.

According to themodel, attributionalpatterns to account forinternal

state changes may be involved in the development of GDs which may

emerge as overly positive and personal explanations for these affective,

cognitive and physiological experiences. Therefore by applying cogni-

tive restructuring approaches to these unhelpful appraisals of mood-

related phenomena, we might expect patients to be able to work

towards the ability to generate more balanced explanations for these

experiences. This in turn should lead to a reduction in the striving and

excessivelygoal-orientedascent behavioursthat aretypically associated

withmanicepisodes.The identicationand recognitionof earlywarning

signs has been shown to be a valuable component of psychological

interventions aimed at preventing the onset and relapse of bothpsychosis (Spencer, Birchwood, & McGovern, 2001) and bipolar

disorder (Perry, Tarrier, Morriss, McCarthy, & Limb, 1999). We would

thereforeanticipatethat similar approaches to theidenticationof early

warning signs or relapse signatures alongside the development of 

‘staying well plans’   (Gumley & Schwannauer, 2006) might also be

effective in managing the impact of GDs.

The importance of social feedback and reinforcement in regulating

beliefs is acknowledged (Bandura, 1986) but if people are socially

isolated or purposefully withdraw from others then they are less able

to benet from the moderating effects of these interactions. Inter-

ventions which enhance people's access to a social network might

thereforebe expected to be usefulas an adjunct to augment theeffects

of individual therapy. Recent conceptualizations of mental imagery as

an emotional amplier (Holmes, Geddes, Colom, & Goodwin, 2008)suggest that harnessing that modality of thought as an additional

strategy for affect regulation might also be of use. This could involve

rescripting existing images or promoting the development of 

alternative images to initiate more adaptive cognitiveand behavioural

responses.

11. Future research

This review and the proposed model have generated a number of 

questions and hypotheses which need to be addressed by future

empirical work. The need for longitudinal investigations of GDs has

clearly emerged. This type of research would enable a better

understanding of the onset, development and stability of GDs over

time, as well as how their presentation and impact on functioning is

693R. Knowles et al. / Clinical Psychology Review 31 (2011) 684–696 

8/17/2019 delusi6

11/13

inuenced by changes in self-esteem, mood and social rank.

Longitudinal research designs would also lend themselves to answer-

ing questions about whether a delusion-as-defence or an emotion-

consistent paradigm (or a combination of both) best accounts for the

development and maintenance of GDs, as well as clarifying the factors

governing the hypothesised relationship between persecutory and

grandiose delusions.

There are also a number of important questions to be addressed

using cross-sectional designs, such as whether or not there aremeaningful ways of classifying or subdividing GDs, whether/when

GDs tend to occur in conjunction with or in isolation from PDs, and

whether the phenomenology of GDs allows clinicians and researchers

to differentiate between diagnosticcategories such as bipolar disorder

and schizophrenia. Much work remains to be done to understand the

potential role of specic cognitive biases in the aetiology and

maintenance of GDs and how this might inform targeted cognitive

behavioural interventions .An additional area of interest centres on

the variation of the specicity of GDs, and how they might relate to

real areas of achievement, esteem or ability. There seems,therefore, to

be an important role for qualitative research to explore the

relationship between life events and GDs and how preserved islands

of self-esteem might contribute to the development of these extreme

beliefs. In summary, there remains much to be done in developing our

understanding of GDs, and in turn developing effective psychological

interventions.

References

American Psychiatric Association (2000).   Diagnostic and statistical manual of mentaldisorders (4th edition). Washington, DC: American Psychiatric Association.

Appelbaum, P. S., & Gutheil, T. G. (1980). Drug refusal: A study of psychiatric inpatients. American Journal of Psychiatry, 137 , 340−346.

Appelbaum, P. S., Robbins, P. C., & Roth, L. H. (1999). A dimensional approach todelusions: Comparison across delusion types and diagnoses.   American Journal of Psychiatry, 156 , 1938−1943.

Appelbaum, P. S., Robbins, P. C., & Vesselinov, R. (2004). Persistence and stability of delusions over time. Comprehensive Psychiatry, 45, 317−324.

Armando, M., Nelson, B., Yung, A. R., Ross, M., Birchwood, M., Girardi, P., et al. (2010).Psychotic-like experiences and correlation with distress and depressive symptomsin a community sample of adolescents and young adults.  Schizophrenia Research,119, 258−265.

Ashcroft, A., Jervis, N., & Roberts, C. (1999). A theory of mind (TOM) and people withlearning disabilities: The effects of a training package. Journal of Applied Research inIntellectual Disabilities, 12, 58−68.

Azhar, M. Z., Varma, S. L., & Hakim, H. R. (1995). Phenomenological differences of delusions between schizophrenic patients of two cultures of Malaysia.  SingaporeMedical Journal, 36 , 273−275.

Baethge, C., Baldessarini, R. J., Freudenthal, K., Streeruwitz, A., Bauer, M., & Bschor, T.(2005). Hallucinations in bipolar disorder: Characteristics and comparison tounipolar depression and schizophrenia.  Bipolar disorders, 7 , 136−145.

Bandura, A. (1986).   Social foundations of thought and action.   Englewood Cliffs, NJ:Prentice Hall.

Beck, A. T., & Rector, N. A. (2005). Cognitive approaches to schizophrenia: Theory andtherapy. Annual Review of Clinical Psychology, 1, 577−606.

Bedford, A., & Deary, I. (2006). The British inventory of mental pathology (BIMP): Sixfactored scales. Personality and Individual Differences, 40(5), 1017−1025.

Bentall, R. P. (1990). The illusion of reality: A review and integration of psychologicalresearch on hallucinations. Psychological Bulletin, 107 , 82−95.

Bentall, R. P. (1994). Cognitive biases and abnormal beliefs: Towards a model of persecutory delusions. In A. S. David, & J. Cutting (Eds.),   The neuropsychology of schizophrenia (pp. 337−360). London: Lawrence Erlbaum.

Bentall, R. (2006). Madness explained: why we must reject the Kraepelinian paradigmand replace it with a complaint orientated approach to understanding mentalillness. Medical Hypotheses, 66 , 220−233.

Bentall, R., Corcoran, R., Howard, R., Blackwood,N., & Kinderman, P. (2001). Persecutorydelusions: a review and theoretical integration.   Clinical Psychology Review,   21,1143−1192.

Bentall, R. P., Kinderman,P., & Manson, K. (2005). Self-discrepanciesin bipolar disorder:Comparison of manic, depressed, remitted and normal participants.  British Journalof Clinical Psychology, 44, 457−473.

Bentall, R. P., & Thompson, M. (1990). Emotional stroop performance and the manicdefence. British Journal of Clinical Psychology, 29, 235−237.

Berrios, G. E. (1991). Delusions as  “wrong beliefs”: A conceptual history. British Journalof Psychiatry, 159(Suppl. 14), 6−13.

Birchwood, M., Trower, P., Brunet, K., Gilbert, P., Iqbal, Z., & Jackson, C. (2006). Socialanxiety and the shame of psychosis: A study in  rst episode psychosis.  Behaviour Research and Therapy, 45, 1025−1037.

Breier, A., & Berg, P. H. (1999). The psychosis of schizophrenia: Prevalence, response toatypical antipsychotics, and prediction of outcome.   Biological Psychiatry,   46 ,361−364.

Cacioppo, J. T., & Bernston, G. G. (1994). Relationship between att