Criticizing and reassuring oneself: An exploration of forms, styles and reasons in female students

Abstract

Self-critical people, compared with those who self-reassure, are at increased risk of psychopathology. However, there has been little work on the different forms and functions of these self-experiences. This study developed two self-report scales to measure forms and functions of self-criticism and self-reassurance and explore their relationship to depression.
A self-report scale measuring forms of self-criticism and self-reassuring, and a scale measuring possible functions of self-criticism, together with a measure of depression and another self-criticism scale (LOSC), were given to 246 female students.
Self-criticizing vs. self-reassuring separated into two components. Forms of self-criticizing separated into two components related to: being self-critical, dwelling on mistakes and sense of inadequacy; and a second component of wanting to hurt the self and feeling self-disgust/hate. The reasons/functions for self-criticism separated into two components. One was related to desires to try to self-improve (called self-improving/correction), and the other to take revenge on, harm or hurt the self for failures (called self-harming/persecuting). Mediation analysis suggested that wanting to harm the self may be particularly pathogenic and is positively mediated by the effects of hating the self and negatively mediated by being able to self-reassure and focus on one's positives.
Self-criticism is not a single process but has different forms, functions, and underpinning emotions. This indicates a need for more detailed research into the variations of self-criticism and the mechanisms for developing self-reassurance.

Full text

31
British Journal of Clinical Psycholog y (20 04), 43, 31–50
2004 The Bri tish Psychological Society
Criticizing and reassuring oneself: An exploration
of forms, styles and reasons in female students
P. Gilbert
1
*, M. Clarke
1
, S. Hempel
2
, J.N.V. Miles
3
and C. Irons
1
1
Mental Health Research Unit, Kingsway Hospital, Derby, UK
2
Department of Psychology, University of Derby, Derby, UK
3
Department of Health Sciences, University of York, York, UK
Objectives. Self-critical people, compared with those who self-reassure, are at
increased risk of psychopathology. However, there has been little work on the
different forms and functions of these self-experiences. This study developed two self-
report scales to measure forms and functions of self-criticism and self-reassurance and
explore their relationship to depression.
Methods. A self-report scale measuring forms of self-criticism and self-reassuring, and
a scale measuring possible functions of self-criticism, together with a measure of
depression and another self-criticism scale (LOSC), were given to 246 female students.
Results. Self-criticizing vs. self-reassuring separated into two components. Forms of
self-criticizing separated into two compo nents related to: bein g self-critical, dwelling on
mistakes and sense of inadequacy; and a second component of wanting to hurt the self
and feelin g self-disgust/hate. The reasons/functions for self -criticism separated into two
components. One was related to desires to try to self-improve (called self-improving/
correction), and the other to take revenge on, harm or hurt the self for failures (called
self-harming/persecuting). Mediation analysis suggested that wanting to harm the self
may be particularly pathogenic and is positively mediated by the effects of hating the
self and negatively mediated by being able to self-reassure and focus on one’s positives.
Conclusions. Self-criticism is not a single process but has different forms, functions,
and underpinning emotions. This indicates a need for more detailed research into the
variations of self-criticism and the mechanisms for developing self-reassurance.
Billy Connolly, the famous Glaswegian comedian, suffered various forms of early abuse
and has struggled with depression and an alcohol problem (Stephenson, 2001). During
an interview with Tim Adams (Adams, 2001), Connolly recalls an important insight he
www.bps.org.uk
* Correspondence should be addr essed to Paul Gilbert, Mental Health Research Unit, Kingsway Hospital, Derby DE22 3LZ,
UK (e-mail p.gilbert@der by.ac.uk).

had whilst trying to relax in a flotation tank. He became aware of a kind of voice that is
there when he makes plans to do things. The inner voice says: ‘No, you’ll never do that.
No, you’re not good enough, not clever enough, you’ll never go there’ (p. 9, italics in
original). Connolly relates ‘this voice’ to the many who have beat and shamed him.
However, Adams (2001) says of Connolly, ‘he jokes about it now but he believes, too,
that the voice will never really leave him; he sees his life as a series of different
strategies for learning to live with it. Not least of these was the tried-and-tested one he
employed in the tank itself: ‘‘Away with you!’’ he yelled, ‘‘F . . . off’’ ’ (p. 9).
Many theorists have viewed self-devaluation, self-condemnation, and self-critical/
attacking feelings and cognitions as important components of psychopathology.
Indeed, for many centuries, depression has been linked to feelings of worthlessness and
self-devaluation (Radden, 2000). Freud (1917) formalized this experience by suggesting
that depression-linked self-devaluation and criticism arise from superego attacks on the
ego and efforts to protect a needed person from anger. In cognitive therapy, negative,
self-focused automatic thoughts have been seen as evaluative, condemning, and
blaming, and closely linked to depression (Beck, Rush, Shaw, & Emery, 1979). Other
therapists have suggested that self-criticism constitutes a specific vulnerability (different
from dependency) to depression (Blatt, Quinlan, Chevron, McDonald, & Zuroff, 1982;
Blatt & Zuroff, 1992). Kohut (1971, 1977) argued that as a result of early experiences,
people could internalize an ability to be self-reassuring and soothing in times of
personal failure, or to act out their frustration with narcissistic rage at the self.
Zuroff, Koestner, and Powers (1994) found that the degree of self-criticism in
childhood is a predictor of later adjustment. Zuroff, Moskowitz, and Cote (1999) found
that self-criticism is associated with depression and poor interpersonal relationships. A
study by Hartlage, Arduino, and Alloy (1998) suggests that self-criticism may be a trait
marker for depression. In a large study (n= 489), Murphy et al. (2002) found that self-
disparagement, marked by feelings of personal inadequacy, is strongly associated with
life-time risk of depression. Teasdale and Cox (2001) found that with a lowering of
mood, recovered depressed people become more self-critical than never depressed
people. Gilbert et al. (2001) found that for both self-critical thoughts in depression and
malevolent voices in schizophrenia, the degree to which individuals experienced these
‘internal attacks and put-downs’ as powerful and dominating was significantly
associated with depression. In a study by Mongrain, Vettese, Shuster, and Kendal
(1998), self-criticism was associated with feelings of being subordinate and inferior and
poor affiliative relationships with others. Gilbert (1992, in press), suggested that self-
critical thoughts and feelings can be seen as forms of ‘inner harassment’ that are
stressful. Rector, Bagby, Segal, Joffe, and Levitt (2000) found that self-critics tended to
have a poorer outcome with cognitive therapy than dependent people, but the extent
to which one could modify self-criticism affected outcome.
It is interesting that Billy Connolly refers to his own self-criticism as being like ‘an
inner voice’, and indeed we believe this is not an uncommon experience for some
people. Such experiences of self-criticism raise the proposition that there must be a part
of the self that ‘does’ the criticism and another part that ‘responds’ to it, i.e. an
interaction between different aspects of the self (Gilbert, 2000a; Gilbert et al., 2001).
Gestalt therapists have referred to this as topdog–underdog (Greenberg, 1979).
Evolutionary psychopathologists have suggested three important propositions that may
illuminate such a relationship (Gilbert, 1989, 1992). First, humans have evolved specific
competencies to be able to learn, understand and enact social roles (e.g. for attachment,
friend–enemy, dominant–subordinate, or sexual roles). Second, these competencies
32 P. Gilbert et al.

operate through specialized brain systems and can be affected by specific hormones
and neurotransmitters (Panskepp, 1998). Third, the role-forming competencies that
evolved for coordinating social roles with external others can be recruited for self-
evaluations and can play off against each other (Gilbert, 2000a). Hence, for example,
competencies and behaviours for the hostile dominating of others (associated with
vigilance to subordinate violations, derogating others, and issuing threats), and those for
acting as a threatened subordinate (fearful compliance, appeasement, submission, and
escape) can interact at the subjective (self-to-self) level (Gilbert, 2000a, 2000b; Gilbert
et al., 2001). A related proposition, derived from interpersonal and cognitive theory,
has been posited by Baldwin and his colleagues (e.g. Baldwin & Fergusson, 2001). They
have shown that self-evaluations and feelings are derived from internalized inter-
personal or relational schema.
There is clinical evidence that a dominant–subordinate self-to-self relationship can
indeed be acted out internally (e.g. with one part of the self issuing threats and shaming
put-downs while another part of the self submits and feels beaten down). For example,
using the Gestalt technique of the two chairs (Greenberg, Elliott, & Foerster, 1990;
Greenberg, Rice, & Elliott, 1993), depressed people can be asked to role-play their
critical thoughts by expressing them from one chair, and then to switch chairs and
explore their feelings and self-beliefs to these ‘attacks’. After delivering an attack on
him/herself, the depressed person often responds by agreeing with (submitting to) the
substance of an attack (e.g. ‘Yes it is true, I am an inadequate, worthless person’) and
quite commonly will take up submissive postures in the feeling chair. In a fascinating
study of responses to self-criticism, Whelton (2000) measured students’ levels of self-
criticism with the Depressive Experiences Questionnaire (DEQ). He then asked each
student to sit in one chair and spend 5 minutes imagining him/herself in the other chair,
criticizing themself. They were then invited to switch chairs and respond to the self-
criticism. Those high in self-criticism often ‘submitted’ to their own self-criticisms,
expressed shamed and sad faces, and felt weak and unable to counteract their attacks.
In other words, some people appear to feel beaten down and have to submit to their
own self-criticism and shaming attacks. Indeed, Greenberg et al. (1990) suggested that
it is the inability to defend oneself against one’s own self-attacks that results in
depression.
Despite the important role that self-criticism, self-condemnation, and self-denigration
play in many forms of psychopathology, and their prominence in many theories, they
tend to be treated as a single process that varies in terms of degree or severity.
However, Driscoll (1989) argued that self-condemnation has different functions, e.g. to
self-correct, to guard against future errors and maintain standards, or to elicit sympathy.
Gilbert (1997) suggested that self-blaming and self-criticisms could arise from efforts to
try to improve oneself and prevent errors, out of frustration (a lashing out at self), or
from self-hatred. Billy Connolly’s example of self-devaluation, that operated when he
thought of trying new things, suggests yet a different form, of preventing self from
taking risks and keeping one in a subordinate position (Gilbert, 2002).
Gilbert (2000a, 2000b) suggested that the forms and functions of self-attacking may
mirror (and be adapted from) evolved competencies that regulate external relation-
ships. For example, dominant animals attack and threaten subordinates to coerce them
to behave in certain ways; that is, to be compliant. In animals, attacks are designed to
stress subordinates and keep them in a ‘subordinate frame of mind’ (Gilbert & McGuire,
1998). Parents may threaten their children and punish them for mistakes or ‘not trying
hard enough’. They may threaten the child by pointing to losses, e.g. ‘If you don’t do
33Criticizing and reassuring oneself

‘‘X’’ or succeed at ‘‘Y’’ no one will like you; you won’t get on in the world’. Here,
threats and attacks are used to regulate the behaviour of the child (entice compliance),
but a parent may claim that this is for their (the child’s) own good (Bowlby, 1980).
People who internalize this form of self-regulation (dominant–subordinate) may say that
self-attacking and punishment are aimed at correcting behaviour (‘I am self-critical to
stop me making mistakes’ or ‘for my own good’), as a parent might say to a child. It is to
prevent bad things happening, by being obedient.
However, humans and animals can launch quite savage attacks on each other for
reasons other than the coercion of subordinates. This kind of attacking in interpersonal
relationships is more typical between ‘enemies’ or directed at the stigmatized where
there is active dislike or hatred. The evolution of this style of relating may have more to
do with persecuting out-groups and those seen as harmful to the self or group.
Consider, for example, forms of group persecution, racism, or ethnic cleansing, where
the aim seems to be active destruction of the other and their removal from the body of
the group. Here, submissive and appeasement behaviour may not work to deflect
attacks. The aim of the attacker is to purify and cleanse that seen as ‘bad/contaminating’
by distancing, excluding and/or destroying it (Shweder, Much, Mahapatra, & Park,
1997). Moreover, the motives and beliefs involved in this style of relating may be those
also underpinning forms of stigmatization, ostracism, and persecution (Kurzban &
Leary, 2001). The emotions recruited in these roles tend to be ones of disgust,
contempt, and hatred, and may have evolved from competencies regulating the
avoidance of noxious substances (Shweder et al., 1997). When a parent attacks a child
in this way, the child is constructed as ‘the bad’, a source of disgust or contempt, not
just the ‘subordinate other’. There is indirect evidence that some people can develop
this orientation to themselves. For example, self-hating and wanting to get rid of aspects
of the self is seen in people who self-harm or wish to ‘destroy’ parts of themselves
(Strong, 1998). The language that an individual may use about aspects of themselves
seems similar to that used to depict enemies and the stigmatized (i.e. they are [I am]
disgusting, contaminating, can destroy the fabric of our society—[the good me]).
Although self-attacking is typically activated when people feel they have failed in
important tasks, or if things go wrong, an alternative response to failure could be self-
support or compassion for the self (Gilbert, 2000a; Kohut, 1971)—to focus on what
one can do to try to reassure the self by (for example), focusing on one’s positives and
active coping/resisting. Whelton (2000) called this resilience and found that no matter
how harsh a self-criticism, students low in self-criticism (as measured by the DEQ) had
more resilience and could defend themselves against their own criticisms. It is believed
that a child adopts these self-reassuring responses (partly) from how parents have
behaved towards them (caring and non-shaming) at times of failure and disappointment
(e.g. Bowlby, 1980; Kohut, 1971); that is, through warmth and affection in early
relationships.
In brief, different classes of interpersonal relationship can be identified, such as:
dominant–subordinate relationships that function to coerce subordinates to behave in
certain ways; stigmatizing and persecuting relationships to eject, expel, reject, distance,
or destroy that which is seen to contaminate (and this form seems routed in a more
persecuting, destructive, sadistic, and cruel orientation to others); and affiliative, warm,
and forgiving styles of relating that grew out of attachment systems. Based on the idea
that the competencies that evolved for social role formation are often recruited into
self-evaluation, it is possible to suggest that people can adopt these various orientations
in self-to-self forms of relating (Gilbert, 2000a). Hence people can be self-critical to try
34 P. Gilbert et al.

to correct their behaviour; or because they have an active dislike of, or hatred for the
self; or people can be relatively warm and reassuring to themselves.
There are a number of measures of self-criticism (e.g. Blatt et al., 1982, Blatt & Zuroff,
1992). Thompson and Zuroff (2000) have recently developed a measure of self-criticism
that relates to two evaluative domains of self-criticism: one resulting from negative
comparisons with others, and one related to criticism over failing to meet internal,
personal standards. However, current measures do not focus on the forms of self-
attacking. For example, they do not allow differentiation of self-to-self attacks that take
the form of dominant coercion of a subordinate to act in a certain way, in contrast to a
form of contemptuous attacking, more akin to enemies and aimed at harming or
wanting to destroy the self. This study, therefore, set out to develop a measure of forms
of self-attacking when things go wrong for people. We wanted to see if self-attacking
could be separated into those forms that are about drawing attention to failures and
inadequacies and things that need improving, in contrast to more aggressive/disgust
based, self-hating forms of self-attacking. Also, as there are no scales for measuring self-
reassuring, and self-reassuring may be an important counter to the effects of self-
criticism (Kohut, 1971; Whelton, 2000), we wanted to explore how self-reassuring
(ability to focus on one’s positives and be reassuring to self when things go wrong)
might be related to self-criticism/attacking and depression.
As noted above, attacking self or others need not imply rejection or a wish to harm or
destroy the self or other. Thus, a further aim of this study was to explore in more detail
people’s beliefs about the functions of their self-attacking; that is, what people see as
the reason(s) for criticizing or attacking themselves. We wanted to explore if some
reasons and functions of self-criticism/attacking are for self-correcting (as suggested by
Driscoll, 1989), whereas others are aimed at harming the self, feeling disgusted, and
wanting to reject and get rid of (parts of) the self, i.e. part of the self has become like an
‘enemy within’.
Method
Participants
A total of 246 female psychology undergraduate students were used in this study (mean
age 27.7 years, SD = 7.2). They completed a series of self-report scales, presented in the
same order at the beginning of a lecture.
Measures
Centre for epidemiological studies depression scale (CES-D)
Depression was measured with the CES-D, which was developed to measure depressive
symptomatology in non-psychiatric populations (Radloff, 1977). It is a 20-item scale
which measures a range of symptoms (such as depressed mood, feelings of guilt, sleep
disturbance), and respondents indicate on a 4-point scale (0–3) how often they have
had the symptoms in the past week. Scores range from 0 to 60, with higher scores
indicating greater depressive symptoms. Radloff (1977) found internal consistency
coefficients of greater than 0.84. This scale has been recommended for use in a general
population (Gotlib & Hammen, 1992).
35Criticizing and reassuring oneself

Levels of self-criticism scale (LOSC)
The LOSC is a new scale developed by Thompson and Zuroff (2000), derived from
considerable previous work on self-criticism by Zuroff and kindly made available for use
in this study by the authors. The LOSC measures comparative self-criticism (Com S.C.)
defined as a negative view of the self in comparison with others, and internalized self-
criticism (Int S.C.), defined as a negative view of the self in comparison with internal,
personal standards. Com S.C. items include ‘I am usually uncomfortable in social
situations where I don’t know what to expect’, and ‘I fear that if people get to know me
too well, they will not respect me’. Such items may reflect social anxiety concerns as
much as self-criticism. Int S.C. items include; ‘I get very upset when I fail’, and ‘When I
don’t succeed, I find myself wondering how worthwhile I am’.
The LOSC consists of 22 items (12 for Com S.C. and 10 for Int S.C.). Each item
consists of a statement, and participants rate how well the statement describes them on
a Likert scale from 1 (not at all) to 7 (very well). Thompson and Zuroff (2000) reported
good reliability with Cronbach’s alpha coefficients of 0.81 and 0.87 for Com S.C. and Int
S.C., respectively. They also report a moderate correlation between Com S.C. and Int
S.C. (r= 0.45). Cronbach’s alphas for this study are given in Table 3.
Forms of self-criticizing/attacking and self-reassuring scale (FSCRS)
The items of this scale were derived from clinical work with depressed people where
one of us (P.G.) had noted some typical thoughts depressed patients offered about their
own self-criticisms and ability to self-reassure. In the first instance, it is unlikely that
these are fully comprehensive, and other items could have been included, but we felt
that they offered a reasonable reflection of typical self-criticizing/attacking thoughts and
would provide sufficient variance to test our hypothesis. The original scale consisted of
24 items to examine how critical/attacking or how supportive/reassuring people are
when things go wrong for them. To a first probe statement: ‘When things go wrong for
me . . .’ participants respo nd on a 5-point Likert scale (ranging from 0 = not at a ll like
me, to 4 = extremely like me) on series of questions: e.g. ‘I am easily disappointed with
myself; there is a part of me that puts me down; I find it easy to forgive myself; I
encourage myself for the future’. The full scale with the principal-components analysis
is given in Table 1.
Functions of self-criticizing/attacking scale (FSCS)
Like items of the FSCRS, the items for the FSCS were derived from clinical work with
depressed people where one of us (P.G.) had noted some typical reasons and functions
that depressed patients have offered about their own self-criticisms. Again, it is unlikely
that these are fully comprehensive, and other items could have been included, but we
felt that they offered a reasonable reflection of typical reasons/functions people have
for self-attacking and would provide sufficient variance to test our hypotheses.
Moreover, the items seemed to offer a reasonable opportunity to see if desires to harm
the self would separate out from self-improvement or correction. Following a probe
statement, ‘I get critical and angry with myself’, there are 21 questions reflecting
possible reasons for self-attacking. Participants respond on a 5-point Likert scale
(ranging from 0 = not at all like me to 4 = extremely like me). Items include: ‘to make
sure I keep my standards up; to stop me being lazy’, through to more persecuting and
36 P. Gilbert et al.

self-harming reasons of ‘to destroy part of me; to harm part of myself’. The full scale
with the principal-components analysis is given in Table 2.
Results
Principal-components analysis of the forms of self-criticizing and self-reassuring scale
A principal-components analysis with oblimin rotation was conducted with a cut-off
point of 0.4 for the inclusion of an item in the interpretation of a component. Oblimin
rotation was chosen because the underlying components were hypothesized to be
related. The Kaiser–Mayer–Olkin measure of sampling adequacy statistic was 0.929.
Examination of the scree plot indicated a three-component solution, which
accounted for 58.32% of the variance. The eigenvalues were: 9.24, 2.13, and 1.46
(with the fourth being .98 and the fifth .85), respectively. Therefore, a principal-
Table 1. Component loading in the Forms of Criticism/Self-Attacking and Self-Reassuring Scale
Item no. Content of question Component 1
Inadequate self
Component 3
Reassure self
Component 2
Hated self
20 I think I deserve my self-criticism .80 .16
16 I remember and dwell on my failings .76
6 There is a part of me that feels I am not
good enough.
.73 ¡.18
1 I am easily disappointed with myself .73
7 I feel beaten down by own self-critical
thoughts
.72 ¡.13
2 There is a part of me that puts me down .72 ¡.19
19 I can’t accept failures and setbacks without
feeling inadequate
.72 .10
22 There is a part of me that wants to get rid
of the bits I don’t like
.58 .16
4 I find it difficult to control my anger and
frustration at myself
.51 .11 .33
8 I still like being me .77 ¡.15
3 I am able to remind myself of positive
things about myself
.75 .15
15 I find it easy to like myself ¡.21 .74
23 I encourage myself for the future .19 .73 ¡.24
13 I can feel lovable and acceptable .73 ¡.16
18 I am gentle and supportive with myself ¡.29 .60 .12
5 I find it easy to forgive myself ¡.29 .52 .24
21 I am able to care and look after myself .49 ¡.12
9 I have become so angry with myself that I
want to hurt or injure myself
.75
10 I have a sense of disgust with myself .15 ¡.16 .73
14 I stop caring about myself .16 ¡.17 .65
17 I call myself names .29 .60
24 I do not like being me .18 ¡.44 .43
Note. Items 11 and 12 were removed due to poor psychometric properties. Factor items in bold.
37Criticizing and reassuring oneself

component analysis with direct oblimin rotation was conducted, specifying a three-
component solution. Component 1 correlated with components 2 and 3 at 0.399 and
¡0.436, respectively, and component 2 and component 3 correlated at ¡0.276.
Two items (11 and 12) were omitted during the analyses. Neither item 11 ‘I try to
boost myself up’ nor item 12 ‘I think I deserve it’ correlated significantly with the other
items. Item 12 was aimed to tap feelings of ‘deserving to fail’ but was inappropriately
worded, and some participants left it blank or with a question mark by it. Therefore,
these two items were omitted from the analysis as recommended by Field (2000).
Factor items are given in bold.
The first of the three specified components accounting for the largest percentage of
the variance consisted of nine items which we labelled ‘inadequate self’. It captures a
sense of feeling internally put-down and rendered inadequate by failures and setbacks.
The highest loading item was ‘I think I deserve my self-criticism’, with the second
highest item being ‘I remember and dwell on my failings’. Other items included ‘I am
easily disappointed with myself’.
A second component consisted of seven items and could be labelled ‘reassured self’.
The highest loading item was, ‘I still like being me’, indicating a positive warm
disposition to the self, with the second highest item being, ‘I am able to remind myself
of positive things about myself’. Other items included, ‘I encourage myself for the
future’. Self-reassuring came out as a separate component, and the items of self-
reassuring seem to form a coherent component, though, again, more detailed research
might find different types of self-reassuring in different contexts. It would also be
Table 2. Component loading in the Functions of Self-Criticism/Attacking Scale
Item no. Content of question Component 1
Self-correction
Component 2
Self-persecution
14 To make me concentrate .83 ¡.16
20 To remind me of my responsibilities .82
5 To stop me being lazy .80 ¡.12
7 To keep myself in check .77
17 To prevent future embarrassments .74
19 To kep me from making minor mistakes .72 .13
16 To stop me becoming arrogant .70
3 To show I care about my mistakes .70
11 To stop me getting overconfident .70
1 To make sure I keep up my standards .66
18 To remind me of my past failures .56 .37
15 To gain reassurance from others .52 .20
12 To stop me being angry with others. .51 .15
10 To take revenge on part of myself .91
9 To cope with feelings of disgust with myself .87
13 To destroy a part of me .10 .86
6 To harm part of myself .82
4 Because if I punish myself I feel better .82
21 To get at the things I hate in myself .82
8 To punish myself for my mistakes ¡.24 .69
2 To stop myself being happy .52
Factor items in bold.
38 P. Gilbert et al.

interesting to study those who can be caring, reassuring, or supportive of others but not
themselves.
A third component consisted of five items and could be labelled as ‘hated self’. The
highest loading item was, ‘I have become so angry with myself that I want to hurt or
injure myself,’ with the second highest item being, ‘I have a sense of disgust with
myself’. This seemed to capture a more destructive, disgust-based response to setbacks
(rather than inadequate) characterized by self-dislike and an aggressive/sadistic/
persecuting desire to hurt the self. It is possible that additional items directly measuring
self-hatred and self-contempt would have strengthened this component. We were
surprised that the item, ‘There is a part of me that wants to get rid of the bits I don’t
like’ loaded on the inadequate component rather than the hated-self component.
Although the reasons for this are unclear, it is possible that in this young, student,
female population, they may have had in mind relatively minor self-attributes (e.g.
aspects of physical shape or appearance) or that they saw such desires to be in the
service of self-improving rather than self-hating. As a preliminary result, it would seem
possible to distinguish self-criticism that is focused on feeling disappointed and
inadequate, from self-criticism of anger and disgust with the self. Indeed, it is probably
possible to feel inadequate, and even a failure at times, but still basically like the self.
The self-reassuring component could have emerged as an artefact of having
positively worded items. One negative item, ‘I do not like being me’, loaded on both
self-reassuring and the third component of hated-self. For this analysis, given the
wording of the item, we retained it in the self-hating component, even though it loaded
negatively, marginally higher on the self-reassuring component.
Principal-components analysis of the functions of self-criticizing/attacking scale
What about the reasons for, and perceived function of, self-attacking? If there are
different underlying mechanisms in self-attacking, and especially differences in
contempt–anger–disgust with self, we would expect these to show up in reasons/
functions for self-attacking, which in fact is the case.
A principal-components analysis with oblimin rotation was conducted with a cut-off
point of 0.4 for the inclusion of a variable in the interpretation of a component. Oblimin
rotation was chosen because the underlying components were hypothesized to be
related. The Kaiser–Mayer–Olkin measure of sampling adequacy statistic was 0.921.
Examination of the scree plot indicated a two component solution, which accounted
for 57.75% of the variance. The eigenvalues were 9.01 and 3.12 (with a third
component being 1.09, and the fourth being .91). When we looked at a three-
component solution, the composition of the items was hard to interpret, whereas a
two-components solution appeared more reasonable and intelligible. Therefore, a
principal-components analysis with oblimin rotation was conducted, specifying a two-
component solution.
The first of the two specified components for the functions of self-criticism was
labelled self-improving and self-correcting (self-correction for short). Thirteen of the
21 items loaded on this component. The highest loading item was ‘to make me
concentrate’. Other items included, ‘to remind me of my responsibilities’, and ‘to stop
me being lazy’. Many depressed people offer such reasons for self-attacking, fearing
they might become inadequate or lazy unless they keep criticizing themselves and
coercing themselves into action. In a sense, this may be felt as a kind of self-coercion
‘for one’s own good’.
39Criticizing and reassuring oneself

The second component was labelled ‘self-persecution’ and consisted of eight items.
The highest loading item was ‘to take revenge on part of myself’. Other items included
‘because if I punish myself I feel better’ and ‘to cope with feelings of disgust in myself’
and ‘to destroy part of myself’. We call this persecution because it denotes a level of
aggression to the self, which is about hurting, harming, and taking revenge.
Taken together, these components are suggestive of there being different forms and
functions to self-criticizing/attacking (Gilbert, 2000a). This invites more specific
research on how these different forms develop, are related to different psychopathol-
ogies and respond to treatment.
Descriptive statistics of measures
Table 3 shows the means, standard deviations and Cronbach’s alpha coefficients for
depression, the LOSC (internalized self-criticism and comparative self-criticism), the
three components of the ‘forms of self-criticizing and self-reassuring scale’ (FSCRS) and
the two components of the ‘functions of self-criticizing/attacking scale’ (FSCS).
Correlations between measures
In order to explore how these self-criticism/attacking measures are interrelated, how
they performed against other self-criticism scales (Thompson & Zuroff, 2000), and how
they are related to depression, a Pearson product-moment correlation matrix was
calculated. Table 4 shows the correlation between age, depression, internalized and
comparative self-criticism, the three components of the FSCRS, and the two
components of the FSCS.
There was a small but significant effect of age, with younger people being slightly
more self-hating than older people. The reasons for this are unclear. All measures of self-
criticism had significant associations with depression, with the aggressive type of self-
attacking (hated self) and self-persecution having particularly high correlations.
Interestingly, although self-improving/correcting functions of self-criticism were
correlated with depression, it is only at a moderate level of r= .30.
Table 3. Mean, standard deviation, and Cronbach alpha levels
MSD ¬
CES-D (n= 246) 18.20 11.99 0.93
LOSC-Int (n= 200) 42.04 12.48 0.90
LOSC-Comp (n= 200) 41.64 10.11 0.71
FSCRS F1 (n= 246) 16.75 8.44 0.90
FSCRS F2 (n= 246) 3.86 4.58 0.86
FSCRS F3 (n= 246) 19.81 5.92 0.86
FSCS F1 (n= 245) 19.27 11.10 0.92
FSCS F2 (n= 245) 4.80 6.43 0.92
Note: CES-D = Depression; LOSC (Int) = Internalized Self-Criticism; LOSC (Comp) = Comparative
Self-Criticism; FSCRS F1 = Inadequate Self; FSCRS F2 = Hated Self; FSCRS F3 = Reassured Self; FSCS
F1 = Self-Correction, FSCS F2 = Self-Persecution.
40 P. Gilbert et al.

Table 4. Correlation between all components
Age CES-D Int S.C. Com S.C. Inadequate Self Hated Self Reassured Self Self-Correction
CES-D ¡.05
LOSC Int S.C. .02 .48**
Com S.C. .01 .49** .53**
Inadequate Self ¡.12 .52** .77** .63**
Forms Hated Self ¡.14* .57** .45** .55** .68**
Reassure Self .09 ¡.51** ¡.45** ¡.63** ¡.57** ¡.56**
Self-Correction ¡.12 .30** .46** .21** .50** .34** ¡.12
Functions Self-Persecution ¡.11 .52** .49** .48** .65** .80** ¡.49** .51**
CES-D = Depression; Int S.C. = Internalized Self-Criticism; Com S.C. = Comparative Self-Criticism.
*p<.05; **p<.01.
41Criticizing and reassuring oneself

On the whole, these new measures performed well against other measures of self-
criticism (the LOSC; Thompson & Zuroff, 2000), although it should be noted that the
LOSC measures different aspects of self-criticism/attacking. Interestingly, self-criticism
arising from internalized standards (Int.S.C. taken from the LOSC) has a very high
correlation with inadequate self (r= .77) and significantly higher (p< .001) than with
hated self (r= .45). This may indicate that people can feel inadequate, as judged by
their own self-evaluative standards, and can be disappointed in themselves, but not
necessarily hate themselves or like being ‘themselves’. Also, interestingly, self-criticism
based on feeling inferior to others (Com S.C. from the LOSC) had a significant but
relative low correlation with self-correction (r= .21). Presumably, one can be self-
critical to try to improve oneself (e.g. as a sports person might do to gee themselves
up—‘come on—concentrate, focus’) without necessarily feeling inferior to others.
Generally, when the function of self-criticism is to harm and persecute the self, this
appears to be more pathogenic than when self-criticism is for self-improvement. For
example, testing for significant differences between correlations we found that
compared with self-correction, the correlations between self-persecution with
depression are higher (r= .30 vs. r= .52; p< .0001) and with hated self much higher
(r= .34 vs. r= .81, p< .0001). Again, this may indicate that self-correction and self-
improvement need not imply active dislike of, or disgust with, the self. However, as
noted in our discussion (the case of Clare), these two forms of self-criticism can have
complex interactions that may change with affective state.
Self-reassuring was negatively correlated with depression and the other forms of self-
criticism/attacking. It would appear that people who self-criticize and attack
themselves when things go wrong may have a relatively poor ability to self-reassure
or focus on their positives. However, note that although the self-improving/correction
function has a negative association with self-reassuring, it is low and non-significant.
This may imply that one can be self-critical, but when the function is to self-improve/
correct, one can also be self-reassuring and supportive.
Relationships between forms of self-criticism/attacking and self-reassuring, and the
functions of self-criticizing/attacking
The pattern of correlations suggests complex relations amongst these self-to-self
processes. To examine further the structure of the functions and forms of self-criticism/
attacking, and to analyse the usefulness of the presented measures and their impact on
depression, multiple regression analyses were applied to carry out a path analysis. A
mediational model, as described by Baron and Kenny (1986), was tested. The model
postulated that the two functions and reasons why people self-criticize and attack
(namely here to self-improve/correct or to harm/persecute the self) would determine
the forms criticism can take (i.e. to focus on inadequacy in the self, or hatred of the
self). Also, we postulated that if one can reassure the self and focus on one’s positive
this will mediate the effects of self-criticizing/attacking on depression.
1
Brief results of
the analysis are presented in Fig. 1, with full details presented in Table 5.
The first stage of the analysis was to examine the effect of the two functions on the
two forms of self-criticism/attacking and self-reassuring. As the proposition was that the
1
We wer e guided to this formu lation and analysis by Marc Fournier, who reviewed a previous d raft of this paper, an d to whom
we expre ss our gratitude.
42 P. Gilbert et al.

functions of self-criticism/attacking determine the specific form it takes, a four-step
analysis, checking for a mediational model, was conducted.
The first test of a mediational model is to determine if the predictor variables do have
an impact on the dependent variable. The first regression analysis examined the
perceived functions of self-criticism/attacking and their impact on the level of
depression (Regression Analysis 1). Using both functions a significant model,
F(2, 244) = 45.25, p< .001, to predict levels of depression with R
2
= .52 was obtained
(The parameter estimates for each of the analyses are shown in Table 5; note that the
regression weight for self-correction is small and does not achieve statistical
significance. We chose to retain this variable in the analysis because of its significant
correlation with CES-D scores and because self-correction in the model has an affect on
the regression weight for self-persecution).
The second step is to ensure that the independent variable can predict the
hypothesized mediators. The second step therefore involved the test of the ability to
predict the form self-criticism/attacking can take through the function these criticisms
serve. The three regression analyses showed significant results for: inadequate self:
Ana lysis 2a: F(2, 244) = 104.55, p< .001, R
2
= .46; hat ed self, Analysis 2b:
F(2 / 2 44 ) = 2 21 .9 , p< .0 0 1 , R
2
= .6 4; a nd r e as s ur e s e l f, A n a ly si s 2 c:
F(2, 244) = 42.12, p< .001, R
2
= .26.
Self-attacking with a perceived persecuting function for the self has a particular
impact on the form ‘hated self’ (standardized beta estimate = .85, p< .001) but also
strong implications for inadequate self (standardized beta estimate = .54, p< .001) and
reassure self (standarized beta estimate = ¡.57, p< .001). If the perceived function of
self-criticizing/attacking is being used for self-correction, the greatest impact appears on
the form of inadequate self (standardized beta estimate = .22, p< .001).
The third analysis step examines whether the hypothesized mediators do impact
upon the dependent variable. Therefore, the third test (Regression Analysis 3) involved
tested whether the forms self-criticism takes and the ability to self-reassure can predict
Figure 1. Path analysis showing the effect of functions mediated by forms, with standardized
parameter estimates labelled. (Statistically significant paths at p< 0.05 are shown in bold)
43Criticizing and reassuring oneself

levels of depression. The independent influence of the forms ‘inadequate self’, ‘hated
self’, and ‘reassured self’ were able to predict the level of depression, F(3, 244) = 52.69,
p< .001, R
2
= .40).
The final analysis step is to examine whether the hypothesized mediators do mediate
the relationship between the predictor variables and the outcome variables. The
question is whether the mediation is partial or complete. The fourth test (Regression
Analysis 4) therefore involves a test of the final model using the two functions (self-
correction and self-persecution), with the two forms of self-criticizing/attacking
(focusing on self-inadequacies and hating the self) and self-reassuring. If the form that
criticism takes fully mediates the functions of self-criticism, these functions should not
show any significant effect on depression if the effects of the forms are controlled.
Interestingly, self-correction shows a standardized beta weight of .10 (p> .05) when
mediation effects are controlled for. Similarly, the variable self-persecution has a non-
significant relationship to depression when the mediation effects are controlled.
In other words, it is the context of not liking the self and self-disgust that harming/
persecuting the self may be most pathogenic. Perhaps, this implies that it is possible to
Table 5. Path analysis steps
Regression analysis 1:
DV: Depression F(2, 244) = 45.3, p< .001, R
2
= 0.27
DV Standardized estimate tp
Self-Correction 0.05 0.79 0.43
Self-Persecution 0.49 7.78 <0.001
Regression analysis 2a
DV: Inadequate Self, F(2, 244) = 104.55, p< .001, R
2
= .46
Self-Correction 0.22 4.09 <0.001
Self-Persecution 0.54 9.91 <0.001
Regression analysis 2b
DV: Hated Self, F(2, 244) = 221.9, p< .001, R
2
= .64
Self-Correction ¡0.09 ¡2.00 0.046
Self-Persecution 0.85 19.11 <0.001
Regression analysis 2c
DV: Reassured Self, F(2, 244) = 42.12, p< .001, R
2
= .26
Self-Correction 0.17 2.66 0.008
Self-Persecution ¡0.57 ¡8.92 <0.001
Regression analysis 3
DV: Depression, F(3, 243) = 52.69, p< .001, R
2
= .40
Inadequate Self 0.15 2.12 0.035
Hated Self 0.34 4.75 <0.001
Reassured Self ¡0.24 ¡3.80 <0.001
Regression analysis 4
DV: Depression, F(5, 241) = 32.47, p< .001, R
2
= .40
Self-Correction 0.10 1.62 0.106
Self-Persecution 0.05 0.52 0.602
Inadequate Self 0.08 1.05 0.300
Hated Self 0.29 3.18 0.002
Reassured Self ¡0.27 ¡4.08 <0.001
44 P. Gilbert et al.

feel very angry with oneself and even want to hurt or take revenge on oneself (maybe
out of frustration), but if underneath this, one basically likes oneself, then it has less
impact on depression. Because our scale does not allow us to explore impulsive lashing
out at self (e.g. out of frustration), we cannot say for sure. An example of this may be a
case treated by P.G., who, when he became frustrated with himself, had smashed his
fist into a wall, breaking a bone in his hand but when not ‘frustrated with himself’
thought he was a ‘reasonably good and worthwhile’ person. He also preferred to ‘take
his anger out on himself’ rather than on others, which he saw as a ‘good’ trait. So there
is still much to do to develop our understanding and methodologies in this area.
These analyses show that the level of depression in students can be predicted with
the functions self-correction and self-persecution, the two forms of self-attacking
(inadequate self and hated self), and the ability to reassure the self and focus on one’s
positive, F(5, 241) = 32.47, p< .001, R
2
= .40. It also suggests that the three different
ways of treating the self were significant mediators of the two different functions, self-
correction or self-persecution. Put another way, the perceived functions of self-
criticism/attacking serve to determine the specific form that self-criticism/attacking
takes. Furthermore, the analyses allow differentiation of the individual influence of each
aspect of self-criticism and attacking. Figure 1 illustrates the independent contributions
of the variables on the level of depression. Especially powerful in this model is the form
of self-hatred (standardized beta estimate = 0.29, p< .01) and the protective effects for
the ability to reassure the self (standardized beta estimate = ¡.27, p< .001).
Comparing this analysis with the correlation matrix (Table 4) shows that using a self-
persecution attacking strategy acts as a suppressor variable for the self-correction
function. The bivariate correlation between self-correction and reassured self was
¡.118 (negative, non-significant), but when a regression analysis is carried out,
including the variable self-correction, the regression estimate is 0.17, a change from a
negative, and now a significant effect. The effect of self-correction on reassuring is
positive, when controlling for self-persecution. What this implies is that there is an
aspect of self-correcting that reflects a more positive orientation to the self when one
controls for self-hating. Hence, as indicated above, people can be self-critical at times,
but this does not mean that they dislike themselves or have self-disgust (hate-self). They
may like themselves but think that being harsh on themselves and dwelling on their
mistakes will improve them—and presumably they are worth trying to improve. Also, as
noted, people may be impulsively aggressive with themselves but necessarily hate
themselves.
Conclusion
This study set out to see if there are different forms and functions of self-criticizing/
attacking. To date, most cognitive-based theories tend to see self-criticizing/attacking as
a single process (e.g. Beck et al., 1979; Blatt et al., 1982), and current measures (e.g. the
DEQ) do not allow discrimination of types of self-criticizing/attacking. We found,
however, that feeling inadequate could be separated from more hateful feelings for the
self. Moreover, we found that the perceived functions of self-criticizing/attacking could
be separated into self-correction and distinguished from desires to persecute or hurt the
self for failing. We also found that the forms of self-criticizing/attacking mediate the
functions; that is, the effects of self-correction and self-persecution on depression are
mediated by the forms of self-hating vs. self-reassuring.
These are early days for this type of research, but our data may suggest that therapists
45Criticizing and reassuring oneself

may need to explore in detail the functions of self-criticizing/attacking and for some
people illuminate the possible aggressive/disgust feelings and emotions that can come
with some forms of self-criticizing/attacking. Some forms of self-criticizing/attacking
seem sadistic. Further, there is a need to explore how people respond (to defend
themselves) to their own attacks (Greenberg & Paivio, 1997; Whelton, 2000). Indeed, it
is believed that in some forms of personality disorder (e.g. borderline personality
disorder), the degree of hatred and disgust with the self is one reason for physically
attacking the self, i.e. self-harm (Strong, 1998). It is also interesting that for patients with
psychosis who hear voices, the voices are often malevolent, aggressive, insulting, and
threaten harm (Chadwick & Birchwood, 1994; Nayani & David, 1996). People can have
strong desires to escape such voices (Gilbert et al., 2001). Taken together, such findings
indicate a need for more research to explore and understand the hostile and harmful
intent of internal self-criticizing/attacking, their developmental origins (e.g. in early
abuse), how they are implicated in different disorders, and ways to work with these
affect-based experiences (Gilbert, 2000a).
We caution in assuming that these are different trait-like vulnerabilities. For example,
a patient, Clare, treated by P.G. felt that her mother had been very critical of her because
she wanted her to succeed and be special. Her mother instilled a very competitive
orientation to life, and Clare often used self-criticism to make herself work hard(er) to be
special. However, as she felt increasingly unable to live up to these harsh standards and
became depressed, she noted that her form of self-criticism changed. She became, she
said, ‘horrified by the state she had got into’. Her self-criticizing/attacking took a far
more sadistic form of wanting to harm her and tell her ‘she would be better off dead; she
was a waste of space’. She had images of wanting to cut herself and felt that if she did,
her blood would be black and all the badness would flow out. Moreover, while she
wanted to get rid of her self-hatred, she was unsure about changing her self-improving
form of self-criticism because she felt that it helped her succeed and become lovable.
Teaching depressed people skills to re-evaluate their self-criticisms/attacks often
involves a form of self-reassuring such as focusing on one’s positives and reducing black
and white thinking (Beck et al., 1979). This might work well for those with self-
improving or self-correcting types of self-criticism, especially if they basically like
themselves. It is unclear if they will work as well for self-persecutors and self-haters.
Another way of working with self-criticizing/attacking is to help people de-centre
themselves from their negative self-directed thought/feelings by teaching mindfulness
techniques (Segal, Williams, & Teasadale, 2002). Teasdale (1999) has argued that as
people become depressed, they access negative (self-attacking) thoughts that easily
spiral them downwards. This study suggests that there may be different types of self-
attacking thoughts/feelings that ‘come on line’ and that it may be accessing hostile
affect (self-disgust and hatred) that is particularly pathogenic. Based on the idea that
self-reassuring and developing affiliative relationships to the self operate through
systems that evolved for affiliative and affectionate-attachment, Gilbert (2000a) has
recently suggested value in helping people develop inner warmth and compassion for
the self as a counter affective response to self-disgust and hatred. Clare had not
internalized an ability to be empathic to her distress or reassure herself when feeling
low or failing at things. She ‘hated’ feeling weak, depressed, and vulnerable. Developing
abilities for compassionate and warm re-evaluations of the self and tolerating negative
feelings with empathic support had a major impact on her recovery. At the end of
therapy, she reflected that trying to focus on her positives (as we had in the early days
of our work together) had felt like self-criticism as if she was telling herself ‘You must
46 P. Gilbert et al.

learn to focus on your positives and not think in black and white!’ She noted that
compassionate re-evaluation felt very different and was much harder for her.
We note that one does not need an evolutionary approach to interpret these
findings. For example, they could be seen as forms of internalization of relationships
and scripts, as outlined in the relational schema approach (e.g. Baldwin & Fergusson,
2001) and many other therapies (e.g. Greenberg et al., 1993; Kohut, 1971). Individuals
come to treat themselves as others have treated them. We do not see these approaches
as alternatives, however. What the evolutionary approach can add is the suggestion that
there may be specific brain systems that allow for specific forms of relating (e.g.
affectionate attachment, dominant–subordinate, enemy–stigmatized) and thus their
recruitment into self-to-self relationships. Interestingly, Clare (noted above) did not
think (consciously at least) that her mother hated her (but wanted to see her as a special
child). However, Clare came to hate herself when she became depressed because she
could not live up to these standards. Because her loss of energy and low mood made it
even harder to succeed (and hence feel lovable), she wanted to destroy this ‘weak–
depressed–needy and [as she saw it] pathetic part of herself’. Hating and the wish to
destroy, eject, or reject and get rid of that which contaminates and damages the self
may be rooted in our evolved systems for dealing with (what we see as) contaminants
(Shweder et al., 1997). Self-compassion may be one way of working with such
difficulties.
There are a number of weaknesses to this study. First, our participants come from a
particular social group (students), age, and gender. It remains to be seen how far our
findings replicate in other populations, such as older people, men, and different types
of patients. Second, we recognize that our measures remain uncomfortably crude
compared with the complexity of these issues but hope this encourages others to
develop more compressive measures of self-attacking that focus on differences in
internal, self-to-self-relationships. The data should be seen as suggestive, requiring more
rigorous measure development and methods. Third, our data do not distinguish
between those who become aggressive with the self out of frustration, but when not
frustrated have a more benign self-relationship, from those who have a more chronic
sense of self-dislike. Fourth, although the hated-self component works well
psychometrically, we feel that it could be significantly improved in content, and it is
questionable how much ‘hate’-based emotion it actually taps.
Fifth, although we have shown that the ability to self-reassure is inversely related to
self-attacking, we are not able to say if there are any differences in self-reassuring. For
example, when things go wrong, some people may focus on previous personal
accomplishments, while others may use coping strategies that are more directly
focused on emotional control and relationships, e.g. by using actual (or memories of)
attachment objects (e.g. Bowlby 1980; Kohut, 1977). We suspect that self-reassuring is
itself a complex process with different components such as the ability to remind
oneself of one’s positives, past successes, and abilities, the capacity to tolerate
disappointment and feeling vulnerable, and the ability to have compassion for the self.
Some forms of reassuring may be about encouragement and ‘geeing oneself up—
energizing’, while others may be soothing and calming oneself down. However, given
the increasing recognition of the importance of the ability to self-reassure, even when
one is self-critical or disappointed with oneself (e.g. Kohut, 1971, 1977; Whelton,
2000), illuminating the types/forms, capacity, origins, and ability to develop self-
reassuring (and reduce aggression to the self) may be an important therapeutic task
(Gilbert, 1997, 2000a).
47Criticizing and reassuring oneself

Finally, we would like to draw attention to the fact that we have not addressed
another potentially very important function of self-criticism, captured by our opening
reference to Billy Connolly. Here, criticism and self-devaluation are not the result of
failures but arise when trying to do new things; that is they are about not taking risks or
not trying to improve the self (e.g. ‘You’ll never be able to do that; this is beyond you’).
While this may arise from internalizing the shaming from others, the evolutionary
function may be to inhibit behaviour and maintain (non-conscious) subordinate (low
risk) strategies in a world that is potentially punitive. We also note that an ‘attacker’s’
put-downs (e.g. a parent on a child) can be stimulated by envy (Gilbert, 1992). For
example, one patient recently reflected that ‘‘My mother used to put me down a lot, I
think because she could not bear the thought that I could do things better than her, she
needed to keep reminding me that she was better than me, I was the child, and I
needed her. Maybe she was jealous of my abilities—I don’t know’. Helping people
understand the complex origins, styles, and functions of their self-criticism may be
important in helping people change them. Here, we hope to have shown that self-
criticism is a complex multifaceted experience that we have yet to fully understand—it
is far more than just negative self-evaluation, and interventions that work for one type
may not work for another.
Acknowledgements
We would very much like to thank Dr Andrew MacLeod (as editor for this paper) and two
anonymous reviewers for their careful reading of earlier versions of this paper and their detailed,
insightful, and helpful suggestions for improving our analysis and conceptualizations. This
research was undertaken by Derbyshire Mental Health Services NHS Trust who received funding
or a proportion of its funding from the NHS Executive: the views expressed in this publication are
those of the authors and not necessarily those of the NHS Executive.
References
Adams, T. (2001, September 23). Connolly on the couch. The Observer Magazine, pp. 9–24.
Baldwin, M. W., & Fergusson, P. (2001). Relational schemas: The activation of interpersonal
knowledge structures in social anxiety. In W. R. Crozier & L. E. Alden (Eds.), International
handbook of social anxiety: Concepts research and interventions to the self and shyness
(pp. 235–275). Chichester. Wiley.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator distinction in social psychology.
Journal of Personality and Applied Social Psychology, 51, 1173–1182.
Beck, A. T., Rush, A. J., Shaw, B.F., & Emery, G. (1979). Cognitive therapy of depression. New
York.
Blatt, S. J., Quinlan, D. M., Chevron, E. S., McDonald, C., & Zuroff, D. (1982). Dependency and
self-criticism: Psychological dimensions of depression. Journal of Consulting and Clinical
Psychology, 50, 113–124.
Blatt, S., & Zurof, D. (1992). Interpersonal relatedness and self-definition. Two prototypes for
depression. Clinical Psychology Review, 12, 527–562.
Bowlby, J. (1980). Loss: Sadness and depression. Attachment and loss, Vol. 3. London: Hogath
Press.
Chadwick, P., & Birchwood, M. (1994). The omnipotence of voices: A cognitive approach to
auditory hallucinations. British Journal of Psychiatry, 164, 190–201.
Driscoll, R. (1989). Self-condemnation: A comprehensive framework for assessment and
48 P. Gilbert et al.

treatment. Psychotherapy: Theory, Research, Practice, Training, 26, 104–111.
Field, A. P. (2000). Discovering statistics using SPSS for Windows: Advanced techniques for the
beginner, London: Sage.
Freud, S. (1917). Mourning and melancholia. In J. Strachey (Ed. & Trans.) Completed
psychological works, Vol. 14 (standard ed.). London: Hogarth Press.
Gilbert, P. (1989). Human nature and suffering. London: Erlbaum.
Gilbert, P. (1992). Depression: The evolution of powerlessness. Erlbaum.
Gilbert, P. (1997/2000). Overcoming depression: A self-help guide using cognitive behavioral
techniques. London: Robinsons and New York: Oxford University Press.
Gilbert, P. (2000a). Social mentalities: Internal ‘social’ conflicts and the role of inner warmth and
compassion in cognitive therapy. In P. Gilbert & Bailey K. G (Eds.), Genes on the couch:
Explorations in evolutionary psychotherapy (p. 118–150). Hove: Psychology Press.
Gilbert, P. (2000b). Varieties of submissive behaviour: Evolution and role in depression. In L.
Sloman & P Gilbert (Eds.) Subordination and defeat. An evolutionary approach to mood
disorders (pp. 3–46): Hillsdale, NJ: Erlbaum.
Gilbert, P. (in press). Depression: A biopsychosocial, integrative and evolutionary approach. In M.
Power (Ed.) Mood disorders: A handbook for scientists and practitioners. Chichester: Wiley.
Gilbert, P. (2002). Evolutionary approaches to psychopathology and cognitive therapy. In P.
Gilbert (Eds.), Cognitive psychotherapy: An international quarterly (Special Edition:
Evolutionary Psychology and Cognitive Therapy), 16, 263–294.
Gilbert, P., Birchwood, M. Gilbert, J., Trower, P., Hay, J., Murray, B., Meaden, A Olsen, K., & miles,
J.N.V. (2001). An exploration of evolved mental mechanisms for dominant and subordinate
behaviour in relation to auditory hallucinations in schizophrenia and critical thoughts in
depression. Psychological Medicine, 31, 1117–1127.
Gilbert, P. & McGuire, M. (1998). Shame, social roles and status: The psychobiological continuum
from monkey to human. In P. Gilbert & B. Andrews (Eds), Shame: Interpersonal behavior,
psychopathology and culture (pp. 99–125). New York: Oxford University Press.
Gotlib I. H. & Hammen, C. (1992). Psychological aspects of depression: Towards a cognitive-
interpersonal integration. New York: Wiley.
Greenberg, L. S. (1979). Resolving splits: Use of the two-chair technique. Psychotherapy, Theory,
Research and Practice, 16, 316–324.
Greenberg, L. S., Elliott, R. K., & Foerster, F. S. (1990). Experiential processes in the
psychotherapeutic treatment of depression. In C. D. McCann & N. S. Endler (Eds.),
Depression: New directions in theory, research and practice (pp. 157–185). Toronto: Wall
& Emerson.
Greenberg, L. S., & Paivio, S. C. (1997). Working with emotions in psychotherapy. New York:
Guilford Press.
Greenberg, L. S., Rice, L. N. & Elliott, R. (1993). Facilitating emotional change: The moment-by-
mement process. New York: Guilford Press.
Hartlage, S., Aduino, K., & Alloy, L. B. (1998). Depressive personality characteristics: State
dependent concomitants of depressive disorders and traits independent of current depression.
Journal of Abnormal Psychology, 107, 349–345.
Kohut, H. (1971). The analysis of the self. New York: International Universities Press.
Kohut, H. (1977). The restoration of the self. New York: International Universities Press.
Kurzban, R., & Leary, M. (2001). Evolutionary origins of stigmatisation: The functions of social
exclusion. Psychological Bulletin, 127, 187–208.
Mongrain, M., Vettese, L. C., Shuster, B., & Kendal, N. (1998). Perceptual biases, affect, and
behavior in relationship of dependents and self-critics. Journal of Personality and Social
Psychology, 75, 230–241.
Murphy, J. M., Nierengberg, A. A., Monson, R. R., Laird, N. M., Sobol, A. M., & Leighton, A. H.
(2002). Self-disparagement as a feature and forerunner of depression: Findings from the Stirling
County study. Comprehensive Psychiatry, 43, 13–21.
49Criticizing and reassuring oneself

Nayani, T. H., & David, A. S. (1996). The auditory hallucination: A phenomenological survey.
Psychological Medicine, 26, 177–189.
Panskepp, J. (1998). Affective neuroscience. New York: Oxford University Press.
Radden, J. (2000). The nature of melancholy: From Aristotle to Kristeva. New York: Oxford
University Press.
Radloff, L. S. (1977). The CES-D scale: A new self-report depression scale for research in the
general population. Applied Psychological Measurement, 1, 385–401.
Rector, N. A., Bagby, R. M., Segal, Z. V., Joffe, R. T., & Levitt, A. (2000). Self-criticism and
dependency in depressed patients treated with cognitive therapy or pharmacotherapy.
Cognitive Therapy & Research, 24, 571–584.
Segal, Z. V., Williams., J. M. G., & Tesadale, J. D. (2002). Mindfulness based cognitive therapy for
depression: A new approach to preventing relapse. New York: Guilford.
Shweder, R. A., Much, N. C., Mahapatra, M. & Park, L. 91997). The ‘‘big three’’ of morality
(autonomy, community and divinity) and the ‘‘big three’’ explanations of suffering. In A. M.
Brandt & P. Rozin (Eds.), Morality and health, (pp. 119–169). New York: Routledge.
Stephenson, P. (2001). Billy. London: HarperCollins.
Strong, M. (1998). A bright red scream: Self-mutilation and the language of pain. New York:
Penguin Group.
Teasdale, J. D. (1999). Emotional processing: Three modes of mind and the prevention of relapse
in depression. Behaviour Research and Therapy, 37, 29–52.
Teasdale, J. D., & Cox, S. G. (2001). Dysphoria: Self-devaluative and affective components in
recovered depressed and never depressed controls. Psychological Medicine, 31, 1311–1316.
Thompson, R., & Zurof, D. C. (2000). The Levels of Self-Criticism Scale. Poster session: APA:
Assessment and diagnosis, Washington Convention Centre.
Whelton W. J. (2000). Emotion in self-criticism. Unpublished Ph.D. thesis, University of York,
Canada.
Zuroff, D. C., Koestner, R., & Powers, T. A. (1994). Self-criticism at age 12: A longitudinal study of
adjustment. Cognitive Therapy & Research, 18, 367–385.
Zuroff, D. C., Moskowitz, D. S, & Cote, S. (1999). Dependency, self-criticism, interpersonal
behaviour and affect: Evolutionary perspectives. British Journal of Clinical Psychology, 38,
231–250.
Received 23 April 2001; revised version received 28 November 2002
50 P. Gilbert et al.