Available online at www.sciencedirect.

com
Consciousness
and
Cognition
Consciousness and Cognition 17 (2008) 725740
www.elsevier.com/locate/concog

Dierentiating shame from guilt q

Fabrice Teroni *, Julien A. Deonna
Department of Philosophy, University of Geneva, 2 rue de Candolle, 1205 Geneva, Switzerland

Received 4 September 2007

Available online 28 April 2008

Abstract

How does shame dier from guilt? Empirical psychology has recently oered distinct and seemingly incompatible
answers to this question. This article brings together four prominent answers into a cohesive whole. These are that (a)
shame diers from guilt in being a social emotion; (b) shame, in contrast to guilt, aects the whole self; (c) shame is linked
with ideals, whereas guilt concerns prohibitions and (d) shame is oriented towards the self, guilt towards others. After pre-
senting the relevant empirical evidence, we defend specic interpretations of each of these answers and argue that they are
related to four dierent dimensions of the emotions. This not only allows us to overcome the conclusion that the above
criteria are either unrelated or conicting with one another, it also allows us to tell apart what is constitutive from what
is typical of them.
2008 Elsevier Inc. All rights reserved.

Keywords: Shame; Guilt; Self; Emotion; Philosophy; Norms; Values

1. Introduction

The study of shame and guilt has become a fundamental part of the psychology of what are called the self-
conscious emotions.1 One of the most debated questions within the literature is how shame diers from guilt.
Distinct and original answers to this question have recently been defended. In what we hope is an interesting
interaction between psychology and philosophy, the aim of this paper is to oer a systematic response to this
question.
There are however two connected problems related to this enterprise. First, such an interaction may seem
methodologically ill-founded. Philosophy tends to look for necessary and sucient conditions for being a cer-

q
We are grateful to Otto Bruun, Tjeert Olthof, Kevin Mulligan, Raaele Rodogno, Klaus Scherer and two anonymous referees for
Consciousness and Cognition for their helpful comments on previous versions of this paper. This paper was presented at ESPP Belfast,
SOPHA Aix-en-Provence, the Universities of Fribourg and Geneva and we thank the audiences at these venues for the stimulating
discussions. This paper was written with the support of the Swiss National Centre for Competence in Research (NCCR) in Aective
Sciences.
*
Corresponding author. Fax: +41 22 3791131.
E-mail addresses: Fabrice.Teroni@lettres.unige.ch (F. Teroni), Julien.Deonna@lettres.unige.ch (J.A. Deonna).
1
For a brief history of the psychological research on self-conscious emotions, see Fischer and Tangney (1995).

1053-8100/$ - see front matter 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.concog.2008.02.002
726 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

tain kind of entity. By contrast, psychology tends to establish correlations between dierent entities, thereby
revealing what is often called typicality eects. Second, it is not easy to understand how the distinct and seem-
ingly incompatible (e.g. Fontaine, Luyten, de Boeck, & Corveleyn, 2006; Olthof, Ferguson, Bloemers, & Deij,
2004) or disconnected answers proposed by psychologistsindeed often oered as statistically relevant corre-
lationscan give rise to a unied theory.
By distinguishing various dimensions of the emotions, we believe that we can nd an appealing way out of
these problems. For two emotions may dier in that they (a) arise in distinct kinds of contexts; (b) take dif-
ferent kinds of particular objects; (c) have dierent formal objects; or (d) are characterized by dierent action-
tendencies.2 If it can be shown that the distinct answers oered by psychologists to dierentiate shame from
guilt speak to these dierent dimensions and do so in a way that allows us to tell apart what is constitutive from
what is typical of them, a unied account might be in the ong.
Our attempt at showing that this is the case proceeds as follows. Section 2 presents some key ndings in the
empirical literature, which support four major criteria for distinguishing shame from guilt. Sections 36 defend
specic interpretations of these criteria. Finally, Section 7 argues that the chosen interpretations are compat-
ible with one another and develops a unied account of shame and guilt.

2. Empirical evidence

We succinctly introduce four groups of experiments and the four criteria they can give rise to. Some words
of caution are in order. First, the criteria are phrased in a way which is more clear-cut than probably intended
by most of their advocates and thus in need of thorough elaboration. This strategy is justied inasmuch as our
aim is to tell apart which of these criteria, when properly interpreted, are apt to qualify as constitutive as
opposed to statistically signicant criteria. Second, the focus is on emotional episodes, i.e. short-lived mental
states which wax, wane and then disappear, unless we specify otherwise. There is of course a very important
debate on shame proneness and guilt proneness in the literature (see e.g. Ferguson, 2005; Tangney & Dearing,
2002), to which we will allude to when appropriate, but which we cannot engage in this article.

2.1. Social vs. personal

The rst opposition one might use to distinguish shame from guilt is the best known and the least empir-
ically tested. It is the idea that guilt is the emotion of internal sanction, shame the emotion of social sanction.
This thesis can be formulated as follows:

Criterion (1): Shame is, as opposed to guilt, a social emotion.

Anthropologists Benedict (1947) and Mead (1937) have popularized the idea that these emotions function
as dierent means by which behavior is sanctioned. Shame sanctions what is socially undesired, guilt what is
perceived as outing private norms. Since these seminal contributions, philosophers (e.g. Harre 1990; Wil-
liams, 1993; Wollheim, 1999) and psychologists (e.g. Ausubel, 1955; Crozier, 1998; Lewis, 1971) have, on lin-
guistic or conceptual grounds as well as on the basis of phenomenological considerations and patients
narratives, insisted on the importance of other peoples judgements and criticisms in shame, and consequently
on the role of this emotion in social regulation and cohesion (Sche, 1988). This allegedly contrasts with the
personal nature of guilt.
This idea has been supported by large-scale cross-cultural studies conducted by Wallbott and Scherer
(1995) in 37 countries. They asked subjects to rate remembered personal experiences of shame and guilt
according to phenomenological, behavioral, and situational variables. Prominent among them was whether
subjects ascribe the cause of their experiences to external or internal factors. Shame experiences were elicited
signicantly more often by other people or external sources than guilt experiences which were to a very large
extent attributed to the self. A recent study conducted by Fontaine et al. (2006) in Peru and Belgium found

2
This should not be understood as implying that these are the only distinctions worth making with respect to the emotionswe mention
here only those we are going to make use of.
 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740 727

similar results. The conclusion is that guilt is caused by internal sanctions, whereas shame is caused by exter-
nal sanctions emanating from other people or institutions (Wallbott & Scherer, 1995, p. 474).

2.2. Self vs. behavior

Another distinction between shame and guilt made prominent by H. B. Lewis (1971) is the following:

Criterion (2): Shame is distinctively related to the entire self, guilt is tied to some specic behavior.

Researchers inuenced by H. B. Lewis have experimentally supported this criterion. Lindsay-Hartz, De

Rivera, and Mascolo (1995) used a methodology relying on an interactive process between subjects and psy-
chologists. Subjects are asked to recount personal stories involving either shame or guilt. Next, the experi-
menter asks them questions, makes suggestions and comments. The goal is to lead them to a better
understanding of the way they implicitly distinguish between these emotions.3 Lindsay-Hartz and colleagues
next checked these results against stories told by other subjects and derived abstract structures of shame and
guilt. They conclude that shame is about the entire self, guilt about a specic behavior.
The subjects of Tangney (1992) experiment had to describe situations in which they were more likely to
experience guilt and situations in which they were more likely to experience shame. Analysis of the short sto-
ries, she concludes, supports the claim that the shame experience [is] more global, more primitive, and less
verbal than the guilt experience (Tangney, 1992, p. 205, see also Ferguson, Stegge, & Damhuis, 1990).
In an original study, Niedenthal, Tangney, and Gavanski (1994) rst asked subjects to describe shame- and
guilt-eliciting situations. Next, they asked them how they would try to undo these situations. Subjects were
more ready to undo their behavior when claiming to feel guilty (e.g. if only I hadnt done so and so), whereas
they were more ready to undo their traits when claiming to feel shame (e.g, if only I werent so and so). Thus,
subjects experiencing shame tend to focus on themselves, subjects experiencing guilt on their behavior.
Supporting the same kind of conclusions, Barrett et al. experiments (1993) consist in giving children special
toys to scrutinize their reactions after they break. According to the authors, two main groups of children can
be distinguished on this basis. Avoiders tend to avert the gaze, their bodies seem to collapse, and they remain
quite passive. Amenders, while also averting the gaze, do not exhibit bodily collapse and try to x the toy.
This conrms, according to M. Lewis, that the attempt at reparation, the focus upon the toy rather than the
self, constitutes the most important behavioral distinction between shame and guilt (Lewis, 1992, pp. 2526).

2.3. Ideal vs. prohibition

The next way of distinguishing shame from guilt can be traced back to the Freudian opposition between
ego-ideal and superego. Shame would arise as a failure to live up to standards set forth by the ego-ideal, guilt
as a result of violating prohibitions imposed by the superego (Lynd, 1956; Piers & Singer, 1953). The latter is
nothing other than external (especially parental) authority internalized as a gure exerting normative control
over the subject, while the former is an idealized gure the subject wishes to resemble. In contemporary empir-
ical psychology, this contrast nds a noteworthy expression in self-discrepancy theory (Higgins, 1987; Higgins,
Klein, & Strauman, 1985; Morretti & Higgins, 1990), which predicts that distinct negative emotions arise from
perceived lack of congruence between actual selves and dierent self-guides. According to this theory, the way
one is potentially conicts with the ways the self or some other(s) perceive how one ought to be or what one
aspires to be. That is, specic disharmonies between the various domains of the self on the one hand and the
various points of view on the self on the other hand generate distinct negative aects. Specically, self-discrep-
ancy theory predicts that actual/own vs. ideal/other perceived discrepancies result in vulnerability to shame
experiences, actual/own vs. should/own perceived discrepancies in vulnerability to guilt experiences.
This hypothesis was rst tested by Selves questionnaire (Higgins, 1987) designed to measure the quality and
intensity of self-discrepancy through free-responses: Subjects have to report with their own words perceived
dierences between actual, ought and ideal selves. This test contrasts with forced-responses questionnaires

3
This is a qualitative case study analysis. Further empirical support of this kind can be found in Lewis (1971) and Lindsay-Hartz (1984).
728 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

used in subsequent tests of the hypothesis. Although it is far from clear that the results of these experiments
conrm it (see Ozgul, Heubeck, Ward, & Wilkinson, 2003 and Tangney, Niedenthal, Covert, &, Barlow, 1998
for serious doubts), self-discrepancy theory has become part of mainstream theorizing about shame and guilt.
Although the idea of not living up to ones ideals is one way of understanding what happens in shame,4 it is
a misleading one. For shame is related to a distinctive kind of failure with respect to ideals. The failure in ques-
tion has to be distinctive since, for shame to occur, it is not enough that the trait or behavior which triggers the
emotion does not live up to ones idealsfor this will more likely elicit self-disappointment or anger at oneself
-, it has to be perceived more dramatically as undermining the ideal in question5 (for empirical support, see
Lindsay-Hartz et al., 1995).
This can be summarized as follows:
Criterion (3): Shame is linked to those failures that undermine ones goals or ideals, guilt to failures with
respect to prohibitions.

2.4. Self-oriented vs. other-oriented

The work of Baumeister, Stillwell, and Heatherton (1995) suggests a fourth criterion. On the basis of a
methodology that analyses autobiographical stories and delineates their most common features, it is suggested
that guilt is distinctively interpersonal. Other studies, focusing on shame- and guilt-proneness, suggest that, as
dispositions, guilt is, in contrast to shame, linked with other-oriented behavior (e.g. Tangney, 1993). This ech-
oes an idea still in currency in the literature according to which shame is ultimately about punishment, a
defensive and self-focused emotion which leads at best to concealment (Gilbert, 2003). By contrast, guilt
is outward-focused, is about responsibility and caring for others (Tangney & Dearing, 2002).6 These contri-
butions hint at a fourth criterion:
Criterion (4): Shame is self-oriented, whereas guilt is other-oriented or communal.

2.5. Taking stock

This concludes this selective presentation of the empirical evidence. Four criteria to tell shame and guilt
apart have been located. More ndings, less directly correlated with any of these four criteria, will be intro-
duced when appropriate. These criteria draw various boundaries within the shame and guilt territory; they
seem supported by the data and appear prima facie plausible. But are they compatible with one another? If
guilt is a communal emotion (criterion (4)), is it not social rather than personal as criterion (1) suggests? If
shame is about personal ideals (criterion (3)), then how can it be thought to be the emotion of social sanction
(criterion (1))? Moreover, because it is often dicult to assess the theoretical import of the empirical results,
we are left with interesting truths concerning the domain, but with no obvious way of arranging them in a
coherent whole.
The aim in what follows is to dissolve potential tensions of the type just alluded to by oering an account
that locates the dimensions of the emotions to which these criteria speak and in such a way that it will be pos-
sible to demarcate the typical from the constitutive. The originality of the account we oer lies in the way the
dierent components of the theory hang together, which should constitute a worthwhile move forwards in the
debate. Criteria (1)(4) are discussed in Sections 36, their links evaluated in Section 7.

4
This thesis is shared by some philosophers. See for instance Rawls (1971).
5
What does undermining an ideal, a goal or a value, as opposed to not living up to them mean? While the latter is intuitively
compatible with still having the ideal in questionafter all, having an ideal does not imply that one always (or even most of the time) live
up to it-, the former appears not to be so compatiblein this case the subject crosses a threshold below which she will wonder whether she
still qualies as having the ideal.
6
Some evolutionary accounts of shame take a much more balanced view in this respect, see Fessler (2001, 2004).
 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740 729

3. Personal vs. social

According to criterion (1), shame is, as opposed to guilt, social. What does this exactly mean? The answer is
complex, since the criterion can be, and has been, located at the level of the respective particular objects, gen-
esis, cognitive antecedents or contexts of shame and guilt. These dierent theses should be evaluated separately;
since we argue elsewhere that none is convincing (Deonna & Teroni, 2008), we will only briey review the rel-
evant arguments here.
By far the most common way of unpacking the claim that shame is social is through the defence of a thesis
about the contexts in which it occurs. It is allegedly distinctive of shame to arise in the context of audiences.
First, this claim is evaluated (Section 3.1), next (Section 3.2) another interpretation of the thesis that shame is
social is assessed, and the nal Section 3.3 explains the grain of truth in criterion (1).

3.1. Social contexts

It is often claimed that shame regulates our image or standing with others. We allegedly feel it when what
we are or what we do triggers critique, laughter or contempt by an audience. Guilt, by contrast, is usually felt
independently of such reactions. Of course, shame does not require a physically present audience;7 imagining
it, many have claimed, is enough (e.g. Benedict, 1947; Mead, 1937; Taylor, 1985; Williams, 1993). This is cer-
tainly true, but is imagining an audience necessary for shame?
Here are two reasons against this claim. First, the claim that we always imagine somebody or other in
shame should be faithful to its phenomenology. Given that imagination is subject to the will and that we
are in a privileged position to know what we are imagining, the claim that we always imagine audiences in
shame is unwarranted. For we are often unaware of any audience. Must the shame I feel of not being able
to apologize in a given situation be accompanied by an episode of imagining someone looking me down,
be it the person I oended or anyone else? And this is true even if shame is indeed often accompanied by such
imaginings. Imagination plays two crucial roles that explain its regular occurrence. It sometimes serves to x
ones emotional reaction when one has doubts, at other times to check whether ones emotion would be eval-
uated as adequate by an audience. But none of this is required for shame and thus constitutive of it. For these
reasons, the (real or imagined) audience thesis is at best sometimes true of shame8.

3.2. Shame, reputation and honor

If this is correct, the contexts of shame and guilt do not constitutively dier. Is there nevertheless a grain of
truth in this widely-shared claim? The best way to answer is to evaluate the other social thesis alluded to
above. This thesis is that shame diers from guilt in that we often experience it without agreeing with the crit-
icism (or praise) that triggers it.9 Guilt is a personal emotion, regulating ones behavior through ones own
standards. By contrast, shame is social since it regulates ones conduct through other peoples standards. It
is important to evaluate this interpretation - rarely explicitly discussed in the empirical literature10for (a)
the position we defend depends on its denial and (b) it will allow us to explain why social theses about shame
only point to typicality eects while accommodating the empirical data.
Now, there are many reasons to think that this distinction will not do. For although it is frequent to feel
shame because of a slander while disagreeing with the standards motivating it, accounting for this fact should

7
Tangney et al. studies (1996) show that subjects claim that more of their shame than of their guilt episodes are experienced privately.
The gures are 18.2% of shame experiences and 10.4% of guilt experiences (pp. 12591260).
8
First, note that there are more sophisticated readings of the imagined audience claim. A prominent one depends on psychoanalytical
premises with which many empirical psychologists will disagree (Wollheim, 1999). We cannot assess it here. Second, we do not intend these
remarks and the ones that follow to be enough for rebuking the audience interpretation of criterion (1), which has been a cornerstone from
Benedict (1947) to Leary (2007) in the attempts to distinguish shame from guilt. But they should be enough to place the onus of the
argument on the advocates of social theses about the context of shame. For an in-depth treatment of this issue, see Deonna and Teroni
(2008).
9
In philosophy, claims to this eect are prominent in Calhoun (2004) and Wollheim (1999).
10
See Crozier (1998) for an exception.
730 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

still meet two constraints. First, if one is to feel shame on those occasions, there must be something one cares
about which is threatened. The obvious candidates here are reputation, honor, prestige and/or privacy. If this
is true, shame triggered by mockery with which we disagree does not structurally dier from guilt in that it
ultimately occurs as a result of a perceived threat to personal standards. Second, loss of reputation or invasion
of privacy are only some of the possible reasons for shame. As will be seen (Section 4), shame puts into question
ones overall perception of ones value. Now, what people think of us, independently of whether we agree with
them, is often of paramount importance to us. For most of us, honor and prestige are important values and
their being threatened constitutes paradigmatic triggers of shame.

3.3. A modest conclusion

The fact that honor and privacy are important personal values is the grain of truth behind the above social
theses. It explains the temptation to claim that shame does not depend on ones sharing the standards accord-
ing to which one is judged. It also explains why social theses about the context of shame have been widespread,
since reputation and privacy of course depend on audiences. If it is true that loss of reputation or privacy are
paradigmatic, but not the only, triggers of shame, then the empirical evidence is accommodated by claiming
that social theses point to typicality eects. For this evidence merely reects the importance of these cases
without implying that shame is, by contrast to guilt, constitutively social on any of the above interpretations11.

4. Self vs. behavior

According to criterion (2), shame is distinctively related to the self, whereas guilt is focused on behavior.
This criterion is clearly intended to refer to the particular objects of these emotions, that is, to the things
referred to in answering questions such as What are you afraid of? or What does Paul regret? Intuitively,
shame puts the self into question, whereas guilt puts actions into question. While the distinction is prima facie
plausible, can we clarify its import? To answer, we will rst (Section 4.1) examine the clues provided by the
respective grammars of shame and guilt and rule out two interesting but inadequate understandings of crite-
rion (2). Second (Section 4.2), we oer and develop a provisional interpretation of it in terms of a contrast with
respect to the evaluative focus of emotions.

4.1. Grammatical constraints

The grammar of shame is compatible with the claim that this emotion is always of ones self. Although the
of myself can be left out in any sentence like I am ashamed of myself because of x, it is always correct to
have it in there. Call the action, trait, etc. because of which one feels shame the secondary object of shame, the
self its primary object.12 By contrast, the grammar of guilt only requires a primary object, which is paradig-
matically an action. One is guilty of some action, never of oneself. Grammar thus suggests that the self has
a distinctive role in shame. Should we thus take literally the claim that, in shame, the focus is on the self rather
than on the behavior?
No, for there are two reasons against this idea. A closer look at the grammar reveals the rst. Shame has
both a primary and a secondary object. That is, being about the self is not the whole story about a given shame
episode, since one is generally ashamed for a specic reason which might be, as in guilt, some behavior. One is
ashamed because of some action or because of some property of oneself. Shame, like guilt, has its reasons. Sec-
ond, the self is of course not wholly absent in guilt. One does not feel guilty over any action, but only over
those in which one thinks one is somehow implicated. This distinguishes guilt from regret or hope.13 The
object of guilt, like that of shame, is conceived as implicating the self in some way. Take Sam who feels guilty
over being rich. This only makes sense if we understand him as feeling guilty as a result of thinking that he

11
Recent philosophical discussions of guilt have stressed the fact that it is social (Deigh, 1983; Velleman, 1999, 2003). This is not the place
for evaluating these claims, but note that, if correct, they give further reason to reject the above interpretations of criterion (2).
12
The phenomenon of shame for someone else is briey discussed in Section 7.3.
13
We come back to this point at the end of the discussion of this criterion.
 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740 731

could do something about it, but does not, or that he is somehow responsible for it14. So, if shame and guilt
dier as regards their particular objects, it is not in terms of the contrast between presence and absence of the
self.
Another tempting way of understanding criterion (2) should also be set aside from the outset. Prima facie,
an appealing and familiar way of understanding what is meant by implication of the self in an emotion is in
terms of responsibility or control. In this sense, the more one is responsible for or in control of an action or a
trait that triggers an emotion, the more one could be said to be implicated in its eliciting conditions. This, how-
ever, will not do for contrasting shame with guilt. The self is in this sense no more implicated in shame than in
guilt for one can be ashamed of traits one has no control over. The picture is quite the reverse, in fact. It is
commonly thought, although it is far from obvious (Williams 1993), that guilt, not shame, has intimate links
with responsibility and control (e.g. Lamb 1983). So, we still lack an understanding of implication of the self
that is relevant to the distinction we are after15.

4.2. Evaluative foci

So, how should we interpret criterion (2)? It was claimed above that the object of guilt must be understood
as related to the self, but that, grammatically speaking, its object is never the self. This contrasts sharply with
shame. Does the noted grammatical point really mirror an important contrast as regards the way the self fea-
tures in these emotions?
Going back to one of the main inuences behind the theoretical interpretations of the relevant experiments,
the picture put forward by H. B. Lewis (1971), proves helpful. She expresses her main thesis16 as follows:

The experience of shame is directly about the self, which is the focus of evaluation. In guilt, the self is not
the central object of negative evaluation, but rather the thing done or undone is the focus. In guilt, the
self is negatively evaluated in connection with something but is not itself the focus of evaluation (Lewis
1971, p. 30).

In shame, one is in a distinctive emotional relation with oneself, which one apprehends as globally negative.
In guilt, by contrast, one apprehends negatively only ones behavior. Lewiss thesis is that, in guilt, we evaluate
ourselves negatively only insofar as we have done something (I was bad to do so and so and not I am bad).
To convey its import, think about Mary, who has stolen money from her mothers bag. As a result, she might
feel shame or guilt. According to Lewis, Mary is ashamed if her behavior leads her to apprehend herself as
globally negative: She must construe it as having a negative import on herself as a whole. By contrast, if
she feels guilt, she apprehends her behavior in a negative light without taking a global negative stance in rela-
tion to herself.
We can pursue Lewis important insight by illustrating the relevant distinction with other emotions. If I
enjoy a lm, the particular object of my emotion is the lm and not the lm in relation to myself watching
it. Guilt, by contrast, requires the above-mentioned link with the self; this is why you cannot feel guilty about
a lm. Finally, there are emotions where the self features centrally, as when you are proud of yourself. These
examples support a contrast between emotions which take the self as object insofar as it is a constituent of their
specic particular object, and emotions which take the self as object independently of such a restriction. The
important point is that something may be among the particular objects of an emotion without being evaluated
by it. For instance, if I feel happy about having kissed Mary, this is not a positive evaluation of either Mary or
me; we are only constituents of a positively evaluated kiss.
Applying these points to shame and guilt claries the present interpretation. According to it, guilt is an
evaluative stance towards ones behavior, the self featuring as a restriction on the objects one can feel guilty

14
We return to problematic cases in Section 7.3.
15
Not surprisingly, there is considerable evidence that guilt is more correlated with self-ascription of responsibility and control than
shame is (e.g. Weiner, 2000). Controlability, however, will not do as a constitutive criterion, as of course one can be ashamed for
something one has or had total control over.
16
This thesis has many advocates in contemporary psychology. See Caplovitz Barrett (1995), Barrett & Campos (1987), Lewis (1992,
2000), Niedenthal et al. (1994), Tangney (1993).
732 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

about. In shame, by contrast, the self does more than restrict the objects one can be ashamed about: It is the
evaluative focus of the emotion. The contrast between partial and global self can thus be cashed out as follows.
Guilt is partially about the self in virtue of being about ones behavior; the self is there as the doer of some-
thing one apprehends negatively. In shame, we have an evaluative stance towards oneself (primary object)
because we have an evaluative stance towards something else (secondary object, which may be behavior).
On this interpretation, the self is an intentional object of the specic evaluative stance which is shame, but
not of the specic evaluative stance which is guilt. This is constitutive of these two emotions. Shame is thus
characterized by a specic negative appearance of oneself to oneself, an appearance one might perfectly well
know to be misleading, and so refuse to endorse.17
This is the most appealing interpretation of criterion (2) although, as we shall see, it is not without raising
potentially important objections (see Section 7.3). Moreover, it stands in need of further elaboration. For what
is the nature of the self it appeals to? Investigating criterion (3) will help answer this question.

5. Ideals vs. prohibitions

As noted above, criterion (3)shame is related to ones goals or ideals, guilt to prohibitionsechoes the
Freudian contrast between ego-ideal and superego. Superego refers to a psychological mechanism internal-
izing prohibitions so as to regulate behavior and thought, ego-ideal to the person one wants to be. The
important point is that ego-ideal refers to what we deem valuable, superego to binding norms. Thus, the
most neutral way of cashing out these expressions is in terms of the contrast between values and norms.
According to criterion (3), I feel shame when I perceive myself as being or acting in ways that undermine
my values, guilt when I violate binding norms. In both instances the values and the norms have of course
to be construed as being adhered to.18
What is at stake here? Section 5.1 introduces important dierences between values and norms. We then
assess two interpretations. The rst (Section 5.2) is in terms of underlying evaluative attitudes, the second (Sec-
tion 5.3) in terms of formal objects. We accept the second. Finally, Section 5.4 explains how to make sense of
criterion (3) in practical terms.

5.1. Norms and values

There are, among others, esthetical values (beauty), intellectual values (intelligence), and moral values (hon-
esty). Examples of norms are: You shall not kill, it is forbidden to park here. At rst sight, values and norms
are very dierent sorts of things, at least insofar as norms are phrased in a deontological vocabulary. For this
discussion, a thorough investigation of the properties that tell them apart is not required. But, since this dis-
tinction is relied on, at least three dierences should be emphasized. First, values admit of degrees, norms do
not. Thus, it makes sense to ascribe more of a given value to one object than to another, whereas it makes no
sense to speak of degrees with respect to norms. A consequence of this is that living up to a value is a matter of
degree, whereas one either outs a norm or one does not. Second, there is a well-known contrast between thick
(the courageous) and thin (the good) values, but no thick and thin norms. Although norms can and often are
hierarchically ordered, this does not mirror the internal relations between thick and thin values. Third, super-
erogatory acts are best understood as acts that are good but not required.19

17
That is, an appearance of oneself to oneself is not a negative judgement about oneself. Note that this is incompatible with doxastic
theories of emotions with which we disagree.
18
This holds even if we sometimes do not, in shame, agree with the specic content of the criticism that has triggered it. As noted above
(Section 3.2), the values adhered to are in this case connected to reputation, honour or privacy.
19
For further distinctions between values and norms, see Konrad (2000), chapters 1 and 2, Livet, 2006; Mulligan, 1998.
 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740 733

5.2. Evaluative attitudes

We cashed out the initial contrast between superego and ego-ideal directly in terms of norms and values. To
which dimension of the emotions does this distinction speak to?20 In the psychoanalytical tradition and its
subsequent use in psychology the superego and the ego-ideal were conceived of as distinct psychological mech-
anisms, respectively, subtending norm compliance and value-driven behavior. And this easily lends itself to an
understanding of the distinction in terms of distinct evaluative attitudes on which the emotions of guilt and
shame would, respectively, be grounded. This is the rst interpretation of criterion (3) we shall discuss. Guilt
would allegedly dier from shame because the evaluative attitude subtending the norms that trigger it would
dier from the evaluative attitude characteristic of ones relations to values triggering shame. The prospects
for this interpretation depend, rst, on the possibility of locating distinctive evaluative attitudes and, second,
on whether we can deploy them to distinguish shame from guilt.
The most appealing distinction in this connection is one that Helm expresses as that between caring and
valuing. In caring, something is important to one, although not fundamental to ones sense of oneself. In val-
uing, by contrast, the thing is relevant to ones self-worth. As he writes, values are [. . .] deeper than cares inso-
far as they provide a framework of meaning in terms of which one can understand who one is (. . .) (Helm,
2001, p. 101). Applying this distinction to our problem would take it that guilt depends on the attitude of car-
ing, shame on that of valuing, irrespectively of any question about norms and values. For there is no more
reason to think that the attitude of care only subtends norms than to think that the attitude of valuing only
subtends values. Shame occurs with respect to whatever one values, guilt when one merely cares for
something.
This path faces, however, three diculties. First, it is not clear that care, as understood here, is enough for
guilt. Is guilt not also subtended by the attitude of valuing? Second, even conceding that care is enough for
guilt, are there not cases of guilt elicited in connection with what we value? Think about norms the respect
of which you perceive as relevant to your self-worth. It may well happen that if you act against them as a result
of inattention, you feel guilty for having so acted, but not ashamed of yourself. The third diculty takes the
shape of a dilemma. The contrast caring vs. valuing relies on the existence, on the one hand, of evaluative atti-
tudes making something relevant independently of self-worth (for guilt) and, on the other, of attitudes making
something relevant to self-worth (for shame). Now, this talk of self takes us back to the discussion of criterion
(2)self vs. behavior. This generates the rst horn of the dilemma. For guilt and shame are probably required
to distinguish between self-relevant and self-irrelevant underlying attitudes. If this is the case, then the present
suggestion relies on the interpretation of criterion (2) retained above and thus cannot be an independent basis
for telling apart these emotions on pain of circularity. If criterion (2) is in place, to unpack the contrast caring
vs. valuing in this way is unproblematic. It means that shame arises with respect to what is relevant to oneself,
relevance being here understood as the focus of a self-reective emotional network. If caring and valuing are
evaluative attitudes, and if evaluative attitudes depend on emotions, this is sound but does not give rise to an
independent distinction between shame and guilt. Trying to remove the circularity without relying on criterion
(2) on the other hand takes us to the second horn of the dilemma. For the contrast caring vs. valuing would
need to be unpacked without appealing to that between self-reexive and non-self-reexive emotional dispo-
sitions. But we can see no way of doing this. These three diculties motivate the search for another interpre-
tation of criterion (3).21

20
We cannot, of course, locate it at the level of the particular objects of shame and guilt, respectively. For while, as seen above, it makes
sense to claim that one feels guilty of a behaviour and ashamed of oneself, it does not make sense to say in this sense that one is guilty of
norms or ashamed of values.
21
The well-known contrast between rst-order and second-order desires (Frankfurt, 1988, Blackburn 1998, p. 66) will not fare better.
Pursuing the strategy would have it that guilt depends on desires towards norms, shame on second-order desires towards values. But
counter examples abound. A fashion model turned social worker may feel ashamed of her physical appearance even if she no longer desires
desiring to look beautiful. Flouting norms one desires desiring to conform to may lead to guilt and not to shame. Think of John who has
parked his car in the wrong spot, thus outing a norm he desires desiring to conform to; he might well feel guilt for having done so, but this
is not ipso facto a case of shame.
734 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

5.3. Formal objects

According to the interpretation we favour, the contrast norms vs. values should be understood via the
notion of formal objects: The formal object of guilt is a outed norm, that of shame an undermined value.
To evaluate this interpretation, the notion of a formal object has rst to be claried. The best way to do
so is to contrast formal objects with particular objects and to explain two important roles the former are meant
to play. Consider fear. It always takes a particular object: One is always afraid of something, for instance a
dog or ones future ruin. The notion of a formal object is a way of locating a common feature of all instances
of a given emotion kind. For instance, all particular objects of fear are intimately linked with danger. Further-
more, and in direct connection with this, formal objects are meant to play at least two roles. (1) That of pro-
viding a rational link between the way the object is given and the action-tendencies characteristic of a given
emotion type. Appeal to danger explains the ight/ght actions typical of fear as being motivated by perceiv-
ing the object as dangerous. (2) Formal objects are used in evaluating whether emotions are correct. Thus, fear
is correct if the dog is dangerous.22
The present interpretation of criterion (3) holds that the relations of guilt to norms and of shame to values
mirror these points. Guilt is the perception of something as breaching a binding norm (as forbidden) and is
appropriate when norms are outed. Shame is the perception of something as undermining ones values and
is correct if this is the case. This is on the right track. Central cases of guilt can indeed be glossed by saying
that one perceives something as forbidden. In religious guilt, for instance, one is guilty for having committed a
sin, that is, for having outed a divine command. This analysis also satisfactorily deals with guilt for having
outed social norms. For shame, appeal to undermined values as formal objects seems an appealing move too.
Central cases of shame are situations where one perceives something, not as forbidden, but as devoid of value
or instantiating a negative valuethe formal object of shame is often thought to be the degraded or the
worthless (Roberts, 2003). In shame for ones physical appearance, for instance, nothing is construed as for-
bidden, but traits are construed as ugly. Moreover, in the cases alluded to above where something triggers
shame when witnessed by others, the subject does not perceive something as forbidden, but as undermining
one of his values. These relations of guilt to norms and of shame to values are constitutive of these emotions.23
We scrutinized two interpretations of criterion (3): First, that values and norms are subtended by distinct
underlying aective attitudes, and, second, that undermined values and outed norms are the respective for-
mal objects of shame and guilt. It has been argued that only the second is illuminating. Let us now show how
to make sense of it in less abstract terms.

5.4. Making sense of the distinction in practical terms

It is dialectically useful to start by an objection to the favoured interpretation of criterion (3). Having dis-
missed the idea that shame diers from guilt in terms of underlying aective attitudes, it might be thought that
the possibility of making sense of any dierence between them in terms of values and norms has been ruled
out. For, if (a) shame and guilt are reactions to failings with respect to what we value norms for guilt
and values for shame and if (b) for every norm we adhere to, a value can be found that subtends it and every
value can give rise to norms that promote it, then (c) the favoured way of understanding criterion (3) might
seem unpromising. But this worry is grounded in the claim that all commitments to norms are commitments to
some value and all commitments to values commitments to some norms. Is this true?
Although the answer might well be positive,24 important points have to be brought to the fore with regard
to the dierences between valuing norms and valuing values. If I abide by the norm of not parking my car in a

22
For a detailed presentation of the various objects of emotions, see De Sousa (1987), chapter 5. For a thorough discussion of the notion
of formal objects, see Teroni (2007). In psychology, formal objects often go under the label of core relational themes, see Lazarus (1991),
chapter 3.
23
We discuss cases of guilt and shame that might be thought of as going against these points in Section 7.
24
Note that a positive answer should be restricted to the norms and values we are committed to. This is fully compatible with our
previous remarks as to the fundamental contrasting features between values and norms (Section 5.1). Furthermore, this does not imply
that there is a oneone mapping between given norms and values as the next paragraphs make clear.
 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740 735

forbidden spot, I do probably live up to some of my values, but not ones to be understood in terms of for-
bidden parking spaces. Perhaps I care about the equitable sharing of public space, and adhering to this norm
is a way of acting accordingly. That is, for most of the norms we casually abide by, values might be found that
ground our norm-conforming behavior. This link, however, might be more or less obvious or accessible when
we act. The same is true of the reverse link between values and norms. Although the importance I attach to,
say, solidarity, does translate into norm-conforming behavior, these norms entertain only loose relations with
solidarity. I might be someone who helps those in need that cross my path, and thus conform to the norm that
one should help the needy, but the link between this norm and my attachment to solidarity is tenuous. What
the latter dictates must often be pondered in the light of novel circumstances.
So, although all norm-conforming behavior might involve acting according to ones values and every action
driven by ones values might be thought of as conforming to norms, we do not usually construe everyday
actions in terms of a congruence between our values and our norms. Behavior can be seen as violating a norm
although its relation to values is presently obscure. Behavior can also be seen as the direct expression of a
value, but how it might be linked to any norm is unclear. And sometimes, I can see that I have violated a
norm, and because I realize that the norm is the direct expression of an important value, I construe my act
both as undermining one of my values and as violating a prohibition. Shame and guilt would in this case ensue.
All this is captured by the opted-for interpretation of criterion (3). Construing ones action in negative
axiological terms is being vulnerable to shame, construing it as violating a prohibition is being vulnerable
to guilt.

6. Self-oriented vs. communal emotions

According to criterion (4), shame is self-related, guilt communal. Which dimension of the emotions does it
speak to? Because guilt is associated with moral transgressions and moral reparation, it might be thought that
the receiving end of these must be others. This would mean that criterion (4) speaks to the level of particular
objects: Guilt would centrally involve others (Baumeister et al., 1995), shame would not. But read as a
constitutive claim about these emotions, nobody should be tempted by it. For there are clear instances of both
emotions with both kinds of particular objects. On the one hand, one may feel guilty for not having worked. It
is also clear that one may feel ashamed of particular objects the specication of which does include others.
This is so when what causes shame is loss of reputation. So, this interpretation is unappealing. Guilt is
typically elicited by situations involving others, but this does not dene it and cannot set it apart from
shame.

6.1. Action-tendencies

A more appealing interpretation is to deploy this criterion at the level of action-tendencies. The thesis is
that guilt motivates other-directed behavior typically reparation, whereas shame motivates self-directed
behaviortypically concealment and rumination (Tangney, 1993).
There is no doubt that, typically at least, guilt motivates communal or pro-social behavior (because
Sam feels guilty about having hurt Jim, he visits the hospital with owers), whereas shame does not.
But this will not do as a constitutive dierence between these emotions. Indeed, if Sally feels guilty about
not having exercised daily, this will not motivate her to act in any way interestingly connected with others.
She will ceteris paribus be motivated to resume her solitary exercises. And although shame typically results
in self-focused activities or rumination, it often motivates reparative and conciliatory behavior (Fessler,
2004).
An option at this stage is to focus on the functions that might be served by action-tendencies rather than on
the tendencies themselves. For it is one thing to establish, as we have just tried to, that guilt can be followed by
self-directed behavior and shame by other-directed behavior, it is another to conclude that these emotions are
not associated with distinctive functions. It might indeed be the role or function of guilt to foster communal
behavior even though this function is served in a given occasion by self-oriented behavior and the role of
shame to promote self-eacing behavior even though this function is sometimes served by other-oriented
behavior.
736 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

Table 1
Location of the criteria
Context Particular object Formal object Action-tendencies
p
Criterion (1): Personal vs. social
p
Criterion (2): Self vs. behavior
p
Criterion (3): Ideal vs. prohibition
p
Criterion (4): Self- vs. other-oriented

Now, apart from the general misgivings that one might have regarding the accounts of the self-reective
emotions in evolutionary terms,25 we have specic reasons to resist them here. Focusing on shame will be
enough. First, shame has been variously associated with action-tendencies such as hiding, submissive behav-
ior, appeasement of others, self-improving behavior, repression of positive aect, passive disengagement, pro-
tective behavior, reappraisal of status and value. And true enough, it is not dicult to nd examples where
shame is connected with any one or more of these action-tendencies. Although there are interesting connec-
tions between them, there is no clear sense in which they serve a function related to the self, be it self-eace-
ment or self-improvement, one that would clearly tell it apart from the function served by guilt. Second, the
present functional interpretation of the distinction forces an implausible reading of many examples. For
instance, it compels us to a counterintuitive understanding of reparative and conciliatory behavior elicited
by shame. When so elicited, why say that these kinds of behavior serve a wholly distinct functionself-eace-
ment or self-improvementthan when they are elicited by guilt?

7. Unity regained

7.1. Taking stock

According to our conclusions, two of the four criteria to distinguish shame from guilt we introduced are
especially relevant: Criterion (2), which appeals to the contrast between self and behavior and criterion (3),
which appeals to that between undermined values and outed norms. By contrast, no constitutive truth has
been disclosed by the examination of criteria (1), the claim that shame is, by contrast with guilt, social and
(4), the claim that guilt is, by contrast with shame, other-directed. There is, with respect to these latter criteria,
only typicality eects. Where do these conclusions leave us? Some psychologists have claimed that these cri-
teria conict with one another, or support dierent predictions (Fontaine et al., 2006). Some specic interpre-
tations do indeed have this consequence, but this conclusion is not warranted in the light of our discussion.
Each criterion was located at the dimension of emotions it speaks to: Criterion (1) concerns the contexts; cri-
terion (2) concerns particular objects; criterion (3) concerns formal objects, and criterion (4) concerns action-
tendencies (Table 1).
We will now argue, rst, that these criteria do not conict with one another, but give rise to a coherent and
illuminating picture of these emotions (Section 7.2). Second, possible objections are discussed (Section 7.3).

7.2. The criteria combined

The favoured interpretations of criteria (2) and (3) are not only compatible, but tailored for each other.
Why? According to criterion (2), guilt has some behavior as particular object and criterion (3) states that
its formal object is a outed norm. Putting these two together means that guilt is directed towards behavior
construed as an infraction to a norm. This is a happy result, since only behavior (construed broadly, see Sec-
tion 7.3) is in accordance or in conict with norms. As regards shame, criterion (2) states that its particular
object is a self, criterion (3) that its formal objects are undermined values. What is the result of combining
these two criteria? To come back to the contrast between primary and secondary object (Section 4.1), note

25
Evolutionary explanations of guilt and shame are very polpular because of the role of these emotions in social cooperation and
cohesion. See Fessler (2001, 2004), Frank (1988), Gibbard (1990, 2003) and, for important criticisms, Prinz, 2004.
 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740 737

Table 2
Interpretations of the criteria in the unied theory
Context Particular object Formal object Action-
tendencies
Criterion (1): Personal vs. social Reputation as
a typical value
for shame
Criterion (2): Self vs. behavior Self and behavior as
constitutive evaluative
foci of shame and guilt
Criterion (3): Ideal vs. prohibition Undermined values and outed
norms as constitutive formal
objects of shame and guilt
Criterion (4): Self- vs. other-oriented Reparation as
typical action-
tendency for guilt

that values play a dual role in shame. A behavior or traitsecondary objectis construed as undermining a
value (say, ones nose as undermining ones aesthetical value) which motivates a negative apprehension of one-
selfprimary objectin terms of values. The link between self and value is less tight than that between behav-
ior and prohibition, since selves are not the only potentially worthless things. But it was said above that
criterion (3) illuminates criterion (2). How so?
There is an important link between having undermined values as formal objects and having a self as par-
ticular object which helps clarify the whole self vs. partial self contrast. For what is the notion of self that
should be appealed to in connection with shame? It is what some philosophers call a persons self conception
(e.g. Velleman 2001). And ones self conception is a function of ones values, it is this sense of oneself specied
in terms of the values one adheres to. This is why one may be said to identify with values, but not with
norms.26 This, together with the chosen interpretation of criterion (3) in terms of values as formal objects, cre-
ates the following picture. In shame, a secondary object is given as undermining one of ones values; since
these values are constitutive of ones self conception, what triggers shame ipso facto threatens ones identity,
this identity being the primary particular object of shame27.
We also said that criteria (1) and (4) point to typicality eects. Thus, it was agreed that guilt is often con-
nected with other-directed behavior. The empirical data are satisfactorily accommodated by appealing to
norms. For the most common norms regulate communal behavior. The fact that guilt is typically linked with
such behavior is therefore what is to be expected. But the defended conception of guilt is compatible with the
claim that it can be about behavior transgressing norms having no links with others. This is for example the
case of Sam feeling guilty about his weight. We have also argued that the typicality eects linked with criterion
(1) derive from reputation, honor and privacy being values most of us hold dear28 (Table 2).

26
Interestingly enough, Olthof et al., 2004 found support for a similar view in connecting shame with unwanted identities and guilt with
involvement in the creation of a morally wrong outcome. Their view, however, diers from ours in two crucial respects. First, their notion
of identity is social (audiences play a crucial role). Second, they consider their evidence as incompatible with the self vs. behaviour
criterion, their main reason being the possibility to feel shame but not guilt over a given behaviour. Note, however, that the contrast norms
vs. values as formal objects (criterion (3)) puts a constraint on the way one has to construe behaviour in order for it elicit guilt which
explains why one may feel shame, but not guilt, as a result of behaving in a given way.
27
The connection between self and identity occupies central stage in the writings of Boonin (1983); Taylor (1985). The present discussion
shows how this connection is made plausible by the empirical data as well as how it is located within a philosophical account of these
emotions. This connection has important ramications in moral theory. None of what we said here has, however, immediate consequences
on whether shame or guilt is the more moral emotion.
28
Note that on the retained interpretations of criteria (1) and (4), there is no incompatibility between them. Guilt is personal in the sense
that it is sensitive to transgressions of norms one adheres to, it is often other-oriented in the sense that it is associated with reparative
behaviour.
738 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

7.3. Objections and replies

Let us nally discuss possible objections to this picture. First, problematic cases are discussed. The focus
next shifts to supercial shame and devastating guilt, and nally the way the present account might deal with
shame for other people is briey presented.
Prima facie, the most problematic case is so-called survivor guilt. What are the relevant norms and behav-
ior here? To answer this, what is meant by behavior should be elucidated. Guilt is felt over intentional actions,
obviously, but is also triggered by non-intentional actions and omissions. The relevant notion of behavior
must thus encompass these cases. In guilt for having unintentionally provoked an accident, my driving will
be construed as violating a norm: It outs, say, the norm that one should adjust ones driving so as to be able
to deal with the unexpected. Omissions: Oblivious to the fact that my silence would precipitate the unjust pun-
ishment of a colleague, I retrospectively construe this omission as violating a moral norm. The model can thus
handle guilt felt when one is simply part of a causal chain bringing about harm or disadvantage to another
(Deigh, 1999; Morris, 1988; Olthof et al., 2004), provided the subject construes the situation as suggested.
Now, can survivor guilt be so explained? There are two ways of arguing that it can. First, the subject may
irrationally construe his having survived as the result of his own previous actions or omissions, which he takes
to have violated a norm, in this case probably a moral one. Second, he may also feel guilt for the thoughts and
emotions his having survived lead to. For instance, he feels happy for having been the lucky one, an aective
reaction he construes as forbidden (Jager & Bartsch, 2006).
Further problems might be thought to arise from the fact that we have made shame, as opposed to guilt,
central to our identity. Indeed, what about supercial shame and devastating guilt? It seems that some shame
episodes (e.g. being caught naked) are of no great import to the subject, whereas some guilt episodes have a
devastating impact. Let us see how these cases fare with the present picture. Supercial shame can be explained
in dierent ways. For it is possible that (a) one realizes that no value of ours has been undermined; (b) others
despise us, but we believe that what we did was required by our values (loss of reputation is seen as supercial
when weighed against the defence of this other value); (c) one no longer identies with the values. So, the posi-
tion is compatible with the fact that dierent cases of shame will vary in depth. What about devastating guilt?
These are situations where the impact of an action is so devastating as to make ones life meaningless. Think
for instance of tragic heroes. What is dramatic in these cases has nothing to do with ones perceiving ones
worthlessness. Ones knowledge of the consequences of the action is enough to have a dramatic impact on
ones life. Such cases of guilt can give rise to a negative assessment of oneself when the misdeed is seen as
the mistake of a lifetime. Our only claim is that a feeling of having undermined ones values is no part of
the specic intentional relation which is guilt but distinctive of shame.
Another potential problem with the theory is its ability to deal with shame for other people, as when Sam is
ashamed of his racist father. For we have stressed the idea that shame is a reexive emotion. This issue is com-
plex, so let us emphasize two points and postpone a thorough investigation to another occasion. First, many
cases usually reported by saying that we are ashamed for someone else involve a negative apprehension of our-
selves. It may well be that Sam is ashamed of himself because his father is racist. Deepening our understanding
of these cases requires an investigation of the complex notion of identication, but they do not threaten the
account since a negative appearance of oneself to oneself features in them. Perhaps it is possible to treat all
cases of shame for someone else in the way just sketched. But our account is compatible with other options.
It was argued that the self involved in shame is ones self conception, i.e. the values one adheres to. This can
give rise to the following alternative explanation of shame for someone else. There are values we adhere to
directly, and values we adhere to indirectly because we love or deeply respect someone else. For instance,
one may care about privacy or honesty insofar as ones friend cares about them. These embedded values
may well give rise to shame for another, that is, to a negative apprehension of anothers self conception as
reected in these embedded values.
The position defended here can thus handle apparently problematic cases. In conclusion, it is interesting to
note how it helps answering an important question: Why does behavior sometimes lead to shame, sometimes
to guilt? We can now rephrase it as follows: Why is behavior sometimes construed as prohibited, sometimes
the occasion for a negative apprehension of oneself in terms of values? It is unwise to look for a general expla-
nation, but a strategy for answering this question in each specic case has been oered. For the argument was
 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740 739

that the norms a subject adheres to can more or less directly express the values she holds dear (Section 5.4).
Each time one does not perceive the relevance of this relation, when one is unclear about or not conscious of it,
one is likely to feel only guilt. Why? The fact is that ones awareness of the links between ones norms and
ones values is more or less transparent to one deserves specic explanations only. For example, if one con-
strues a prohibited action as manifesting a disposition to act freely against important values, shame will prob-
ably ensue. If construed as a mistake, guilt will probably ensue. This is so because mistakes are less easily
construed as undermining ones values than dispositions to act against them. But this is not always the right
explanation. For instance, shame of being dragged naked in front of an audience does not usually result from
construing the event as manifesting a disposition to behave as to be so seen. The fact that values such as pri-
vacy or honor are perceived as undermined when others see us in these circumstances are powerful motives for
construing ourselves as worthless or degraded.

References

Ausubel, D. P. (1955). Relationships between shame and guilt in the socializing process. Psychological Review, 62, 378390.
Barrett, K. C., & Campos, J. J. (1987). Perspectives on emotional development: II. a functionalist approach to emotions. In J. Osovski
(Ed.), Handbook of infant development (2nd ed., pp. 555578). New York: Wiley.
Barrett, K. C., Cole, P. M., & Zahn-Waxler, C. (1993). Avoiders versus amenders: Implications for the investigation of guilt and shame
during toddlerhood?. Cognition and Emotion 7, 481505.
Baumeister, R. F., Stillwell, A. M., & Heatherton, T. F. (1995). Interpersonal aspects of guilt: Evidence from narrative studies. In J. P.
Tangney, D. Fischer, & W. Kurt (Eds.), Self-conscious emotions (pp. 255273). New York: The Guilford Press.
Benedict, R. (1947). The Chrysanthemum and the sword. Reprint New American Library.
Boonin, L. (1983). Guilt, shame and morality. Journal of Value Inquiry, 17(4), 295304.
Calhoun, C. (2004). An apology for moral shame. The Journal of Political Philosophy, 12(2), 127146.
Caplovitz Barrett, K. (1995). A functionalist approach to shame and guilt. In J. P. Tangney, Fischer, & W. Kurt (Eds.), Self-conscious
emotions (pp. 2563). New York: The Guilford Press.
Crozier, W. R. (1998). Self-consciousness in shame: The role of the other. Journal for the Theory of Social Behaviour, 28(3), 273286.
De Sousa, R. (1987). The rationality of emotions. Cambridge, Mass: MIT Press.
Deigh, J. (1983). Shame and self-esteem: A critique. Ethics, 93, 225245.
Deigh, J. (1999). All kinds of guilt. Law and Philosophy, 18, 313325.
Deonna, J., & Teroni, F. (2008). Is shame a social emotion? Under Review.
Ferguson, T. J. (2005). Mapping shame and its functions in relationships. Child Maltreatment, 10(4), 377386.
Ferguson, T. J., Stegge, H., & Damhuis, I. (1990). Guilt and shame experiences in elementary school-age children. In R. J. Takens (Ed.).
European perspectives in psychology (Vol. 1, pp. 195218). Wiley.
Fessler, D. (2001). Emotions and cost/benet assessment: The role of shame and self-esteem in risk-taking. In R. Selten & G. Gigerenzer
(Eds.), Bounded rationality: The adaptive toolbox (pp. 191214). Cambridge MA: MIT University Press.
Fessler, D. M. T. (2004). Shame in two cultures: Implications for evolutionary approaches. Journal of Cognition and Culture, 4(2),
207262.
Fischer, K. W., & Tangney, J. P. (1995). Self-conscious emotions and the aect revolution: Framework and overview. In J. P. Tangney &
K. W. Fischer (Eds.), Self-conscious emotions (pp. 322). New-York: The Guilford Press.
Fontaine, J. R. J., Luyten, P., de Boeck, P., & Corveleyn, J. M. (2006). Untying the gordian knot of guilt and shame: The structure of guilt
and shame reactions. Journal of Cross-cultural Psychology, 37, 273292.
Frank, R. (1988). Passions within reason: The strategic role of the emotions. New York: W.W. Norton & Co.
Frankfurt, H. (1988). The importance of what we care about. Cambridge: Cambridge University Press.
Gibbard, A. (1990). Wise choices, apt feelings: A theory of normative judgement. Oxford: Oxford University Press.
Gilbert, P. (2003). Evolution, social roles, and the dierences in shame and guilt. Social Research, 70.4.
Helm, B. (2001). Emotional reason. Cambridge: Cambridge University Press.
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and aect. Psychological Review, 94, 319340.
Higgins, E. T., Klein, R., & Strauman, T. (1985). Self-concept discrepancy theory. Social Cognition, 3, 5176.
Jager, C., & Bartsch, A. (2006). Meta-emotions. Grazer Philosophische Studien, 73, 179204.
Konrad, M. (2000). Werte versus Normen als Handlungsgrunde, Peter Lang Verlag.
Lamb, R. E. (1983). Guilt, shame and morality. Philosophy and Phenomenological Research, 43(3), 329346.
Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.
Leary, M. R. (2007). Motivatioanl and emotional aspects of the self. Annual Review of Psychology, 58, 317344.
Lewis, H. B. (1971). Shame and guilt in neurosis. New York: International Universities Press.
Lewis, M. (1992). Shame: The exposed self. New York: Free Press.
Lewis, M. (2000). Self-conscious emotions: embarrassment, pride, shame, and guilt. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook
of emotions (2nd ed., pp. 623636). New York: The Guilford Press.
Lindsay-Hartz, J. (1984). Contrasting experiences of shame and guilt. American Behavioral Scientist, 27, 689704.
740 F. Teroni, J.A. Deonna / Consciousness and Cognition 17 (2008) 725740

Lindsay-Hartz, J., De Rivera, J., & Mascolo, M. (1995). Dierentiating shame and guilt and their eects on motivation. In J. P. Tangney,
Fischer, & W. Kurt (Eds.), Self-conscious emotions (pp. 255273). New York: The Guilford Press.
Livet, P. (2006). Les Normes. Paris: Armand Colin.
Lynd, H. (1956). On shame and the search for identity. New York: Harcourt Brace Jovanovich.
Mead, M. (1937). Cooperation and competition among primitive peoples (1st ed..). New york: McGraw-Hill Book Company.
Morretti, M., & Higgins, E. T. (1990). Relating self-discrepancy to self-esteem: The contribution of discrepancy beyond actual self-rating.
Journal of Experimental Social Psychology, 26, 108123.
Morris, H. (1988). The decline of guilt. Ethics, 99:1, 6276.
Mulligan, K. (1998). From appropriate emotions to values. The Monist, 81:11, 161188.
Niedenthal, P. M., Tangney, J. P., & Gavanski, I. (1994). If only i werent vs If only I hadnt: Distinguishing shame and guilt in
counterfactual thinking. Journal of Personality and Social Psychology, 67, 585595.
Olthof, T., Ferguson, T. J., Bloemers, E., & Deij, M. (2004). Morality- and identity-related antecedents of childrens guilt and shame
attributions in events involving physical illness. Cognition and Emotion, 18(3), 383404.
Ozgul, S., Heubeck, B., Ward, J., & Wilkinson, R. B. (2003). Self-discrepancies: Measurement and relation to various negative aective
states. Australian Journal of Psychology, 55, 5662.
Piers, G., & Singer, M. B. (1953). Shame and guilt. A psychoanalytic and a cultural study. Springeld.
Prinz, J. (2004). Gut reactions: A perceptual theory of emotion. New York: Oxford University Press.
Rawls, J. (1971). A theory of justice. Harvard: Harvard University Press.
Roberts, R. (2003). Emotions: An essay in aid of moral psychology. Cambridge: Cambridge University Press.
Sche, T. (1988). Shame and conformity: The deference emotion system. American Sociological Review, 53, 395406.
Tangney, J. P. (1992). Situational determinants of shame and guilt in young adulthood. Personality and Social Psychology Bulletin, 18,
199206.
Tangney, J. P. (1993). Shame and guilt. In C. G. Costello (Ed.), Symptoms of depression (pp. 161180). New York: John Wiley.
Tangney, J. P., & Dearing, R. (2002). Shame and guilt. New York: The Guilford Press.
Tangney, J. P., Miller, R. S., Flicker, L., & Barlow, D. H. (1996). Are Shame, guilt and embarrassment distinct emotions? Journal of
Personality and Social Psychology, 70, 12561269.
Tangney, J. P., Niedenthal, P. M., Covert, M. V., & Barlow, D. H. (1998). Are shame and guilt related to distinct self-discrepancies? A test
of Higgins (1987) hypotheses. Journal of Personality and Social Psychology, 75(1), 256268.
Taylor, G. (1985). Pride, shame and guilt. Oxford: Oxford University Press.
Teroni, F. (2007). Emotions and formal objects. Dialectica, 61.3, 395415.
Velleman, J. D. (1999). A rational superego. The Philosophical Review, 108(4), 529558.
Velleman, J.D. (2001). Identication and identity. Reprinted in Self to self: Selected essays. Cambridge: Cambridge University Press
(2006), pp. 330360..
Velleman, J. D. (2003). Dont worry, feel guilty. In A. Hatzimoysis (Ed.), Philosophy and the emotions (pp. 235248). Cambridge:
Cambridge University Press.
Wallbott, H. G., & Scherer, K. R. (1995). Cultural determinants in experiencing shame and guilt. In J. P. Tangney, Fischer, & W. Kurt
(Eds.), Self-conscious emotions (pp. 465487). New York: The Guilford Press.
Weiner, B. (2000). Intrapersonal and interpersonal theories of motivation from an attributional perspective. Educational Psychology
Review, 12.1, 114.
Williams, B. (1993). Shame and necessity. Berkeley: University of California Press.
Wollheim, R. (1999). On the emotions. New Haven: Yale University Press.