12 Shame across cultures: the evolution, ontogeny
and function of a 'moral emotion'
Michael J. Casimir and Michael Schnegg
Our aim in this chapter is to apply a 'dual inheritance theory' and show how
'emotion universals' - in this case shame - are related to physiological pro-
cesses and linked to social and behavioural similarities across cultures on the
Olle hand and how culture-specific emotions are connected to the learning and
coding of specific social and behavioural patterns and conventions on the other.
At least since Paul Ekrnan 's major studies (e.g. 1973, 1980, 1989; Ekrnan and
Oster, 1979) on 'basic emotions' and the rise of the constructivist view in an-
thropology, there has been a controversy over whether emotions are universal or
culturally constructed (Armon-Jones, 1986; Averill, 1980; H~rre, 1986; Lutz,
1988). Many recent studies have shown that everywhere humans express emo-
tions through metaphors and/or metonyms, which are related to bodily feelings
(Kövecses, 1995, 1998), which in turn are generated by autonornic nervous
system activities. These activities seem to have a panhuman hard-wired basis
(for an overview see Levenson et al., 1992). These findings challenge the cul-
tural constructivist hypothesis, which postulates that different cultures construct
emotions in an entirely different manner. However, a comparison of the events
which lead to the feelings expressed by thesemetaphors and metonyms shows
that the antecedents and related appraisals do vary greatly between cultures.
This underlines the relevance of the criticism levelled by the constructivists,
which stresses the analytic indispensability of cognitive factors (Schachter
and Singer, 1962) inculcated in the course of childhood and related to the
variety of values, norms and behavioural prescriptions obtaining in different
The problems inherent in the theory of 'basic emotions' and the fruitless-
ness of either/or, nature/culture, nature/nurture dichotornizations can, we sug-
gest, be overcome by~~ollowing Lyon (1995, 1998; cf. also Leavitt, 1996,
p. 531; Lyon and Barbalet, 1994; Morton, 1995; Williarns, 2001) in suggesting
Michael J. Casimir would like to thank all those colleagues, students and friends, who helped collect
the language data used in this paper, and who are toD numerous to mention individually. Without
forgetting the usual disclaimer, Christine Avenarius, Monika Böck, Priya Bondre-Beil, Alexander
L. Gerlach, Thomas Helrnig, Aparna Rao, Birgitt Röttger-Rössler, Peter Tschohl and the editors of
the volume were of immense help with their perceptive and critical cornments.
Shame across cultures 271
that ideational/interpretative and materialist/positivist/analytic approaches be
surmounted by using a more comprehensive perspective:
Drawing on data on language use across 135 cultures (see appendix I), we
show that the phenomenon of blushing in shameful situations is a panhuman
Olle. This, we suggest, indicates that the emotional mechanisms and physiol-
ogy involved serve a specific function, which has a lang phylogenetic history.
The data available on the ontogeny of emotions in early childhood, and es-
pecially on the emergence of the feelings of shame) could shed light on its
panhuman function. Simultaneously, however, we must allowfor some epige-
netic modulation and flexibility in early childhood für a linkage to take place
wirb culture-specific behavioural norms. Culture prescribes and constructs the
enacting of shame-feelings - it speils out when and why these feelings can or
should be experienced and exhibited and when they should be suppressed.
Here we examine shame and embarrassment - the members of the 'family
of negative self -conscious emotions'. Thesehave been chosen because they are
most suited to reach a clearer understanding of the interrelationship between
bodily feelings and the more or less culture-specific antecedents and appraisals
leading to emotions (cf. Mesquita and Frijda, 1992; Scherer, 1993a, 1997;
Scherer and Wallbott, 1994). Finally, using the cognition/emotion b.ehaviour
complex associated in the West with the terms 'shame' and 'embarrassment'
and applying Wierzbicka's (1986, 1988, 1992, 1999; Harkins and Wierzbicka,
1997) theory of 'natural semantic metalanguage' (NSM), we demonstrate the
interdependence of metaphorical or metonymicallanguage and panhuman phys-
iological reactions and theirrelationship to universal antecedents and appraisals.
It has become increasingly clear in recent years thai only an epigenetic view
can explain the similarities and differences in emotional experiences and ex-
pressions. In some instances it is true that, as Athanasiadou and Tabakowska
(1998, p. xii) wrote, '... while the physiological background as such may well
be universal für all human beings, the actual choice of its elements für conceptu-
alisation, and subsequent expression, need not be'. I:n other instances, however,
even this choice appears to follow some universal pattern, für although 'there are
no completely universal concepts für emotions; there can nevertheless be simi-
larities in conceptualisation of emotions across different languages and different
cultures. This is due to universal human experience, especially physiological
experience' (Mikolajczuk, 1998, p. 158). This is why antecedent events, which
may lead to evolutionarily 'primitive' reactions, such as an attack or flight, are
always accompanied by similar bodily feelings and often conceptualized across
cultures in comparable metonyms and metaphors. The metonymic conceptu-
alization of the physiological reaction is an indirect description of the emo-
tion; thus, für example, the metonymic expression für anger could be 'his face
turned red wirb anger'. Metaphors are also often motivated by physiological
effects - hefe, the central metaphor für anger is 'heat' (Mikolajczuk, 1998,
272 Michael J. Casimir and Michael Schnegg
pp. 157-8; für a detai1ed discussion of the 'standard theories of metaphor' see
Johnson, 1990, pp. 67-72).
On the origins of the 'negative se]f-reflexi,'e emotions' and the
An evolutionary perspective always focuses on the function-of a given trait
or behaviour and it is in this framework that we now discuss the evolution
of shame and related emotions. Emotions - and especially the self-centred,
reflexive Olles - are linked to the sociallife of the species and must have had a
continued adaptive importance. Emde (1980,p. 3), in his biosocial approach to
emotional development, stated that:
. . . emotional states represent complex systems of organized functioning inherent in the
human person, states that are generally advantageous to the species as weIl as to the
individual in tbe course of development. Evolutionary considerations seem to highlight
both the complexity of emotions and their centrality in social adaptation.
Again, referring to Hamburg's (1964) study, he pointed out that emotions
. . . evolved because of a selective advantage in facilitating social bonds. In reviewing
the course of primate evolution, he speculated that group living operated as a powerful
adaptive mechanism and that, because of this, the formation of social bonds has been
experienced as pleasurable and their disruption as unpleasurable. (Emde, 1980. p. 3)
All members of the farnily of unpleasant and negative self-reflexive emo-
tions can then be understood as part ofthe 'social fearfulness complex', which
emerged together with the evolution of more complex social structures in the
animal kingdom. Lewis (1995) postulated anearly form ofembarrassment that
is not associated with the self's negative evaluation of its action and a second
phase when embarrassment does become associated wirb self-evaluation. 'This
[latter] form of embarrassment is related to other negative self-conscious eval-
uative emotions, such as shame and guilt' (p. 215). In evolutionary terms the
early emotional type of embarrassment may be linked to unpleasant feelings
(e.g. fear) deriving from the negative reactions experienced by the individual
when his/her behaviour wellt against the expectations of members of the social
group. If the same 'wrang' behaviour occurs again and is once again noticed
by other group members, expected 'punishrnent' may cause these feelings of
embarrassment. This type of emotion and the accompanying behaviour in the
absence of self-recognition is to be found in most species 'below' the higher
primates (Gallup, 1970, 1977); a common example is the behaviour of a dog
that hag düne something 'wrong'. Shying away and subrnissive postures '4fe the
usual signs of punishment expectation and can be interpreted as appeasement
strategies with the obvious intention of changing the opponent's behaviour
(cf. Keltner and Harker, 1998). Should, however, the opponent's strength and
Shame across cultures
dominance be questionable, aggressive
attack. In the process of social integratio
the qualities and capacities of others, m
by calculating his chances and also app
As Öhman (1986, p. 130) observes,
. . . the evolutionary origin of social fear is in systems controlling dominance/su.bmiss-
iveness in group-living animals. Biological factors are clearly involved in the det,ermining
stimulus and response factors, such as which gestures define displays of dominance and
submissiveness, as weIl as readiness to respond to them. However, learning describes
which particular individuals become associated with submissiveness and fear.
It was only later in evolution - at the latest with the emergence of the horninids
with their self-reflexive cognitive capacities - that the function of this mech-
anism was broadened and Lewis' postulated second form of embarrassment/
The exact mechanisms of the physiological process leading to a ftush/blush
are not clear. They seem to be part or epiphenomena of a general arousal
mechanism, which enables the individual to react quickly by supplying the
periphery with blood. If such a mechanism indicates aggression/rage in the
'lower' mammals (e.g. the bloodshot eyes of an attacking hüll), it probably also
existed in primates and early horninids. When the capacity für self -reftection
developed in the higher primates, the second form emerged and was linked to
the physiological mechanism already present.
Since, arguably, the early horninids bad a dark complexion, it was only when
some of them became melanin deficient that the reddening of the face was
semanticized and now functioned as a signal indicating the emotions, intentions
and possible behaviour of the other. This became a new code which, together
with other signals indicating the individual's emotions (facial expressions, body
postures and, later, language) enhanced communication between individuals,
thus leading to cohesion and consensus between group members. Leary et al.
(1992) rightly suggested that '... the blush reaction may have emerged für
reasons unrelated to social communication and only later came to serve an
interpersonal function among light-skinned peoples'. This would mean that a
ftush-reaction was already fully developed before self-consciousness, complex
sex- and age-specific hierarchically organized social structures - and with them
the expectations of role performance - came into being. But now its function
broadened. If violation of behavioural expectations of group members leads to
probable punishment, arousal is likely in the individual who hag transgressed
these mIes, setting the behavioural options to flight (fear) or fight (anger, rage).
Such behavioural patterns are weIl known among many mammals with complex
social structures and a 'social subrnissive system'. From the phylogenetic point
of view, it could then be suggested that shame is a specific form of social fear
that may lead to feelings of anger and/or rage, enabling the individual to react~
motivation may emerge and lead to an
n the young individuallearns to appraise
inimizing the risk of 'wrong' behaviour
raising the advantages offlight or fight.
274 Michael J. Casimir and Michael Schnegg
quickly and aggressively in order to change the behaviour of others (Gilbert
and McGuire, 1998; McGuire and Troisi, 1990).
If, however, the neurophysiological mechanisms leading to a shameful blush
are different from those leading to a flush in the context of anger/rage, the
question arises about the evolution of this new mechanism which functioned
as part of an interpersonal communication system. This can only have devel-
oped in concert with the evolution ofcomplex social systems,-the capacity für
self -consciousness and normative value-bound behaviours. If we assurne that in
early dark-skinned primate/human populations with complex social structures
and individual self -consciousness, shame and related emotions (embarrassment,
shyness) bad already developed, the function of a blush as an appeasement signal
is unlikely. Vasodilation and thus blushing in shameful situations occurs in all
humans, regardless of visibility; this suggests an early time-point in evolu.tion.
In other words, it could not have developed exclusively among those human pop-
ulations with decreased levels of melanin in the skin. It is also implausible that a
physiological'blush mechanism, independent tram that of a flush, was coupled in
'shameful situations' with lowering the gaze, shying away and other behavioural
expressions of confusion and insecurity, since these are also observable in many
other mammals living in social groups (cf. Keltner and Harker, 1998).
On the function of shame
Social consensus and the adjustment 01 the self
A major role of emotions when functioning as facial, vocal, or physiological
cues, or when expressed as gestures or body movements, consists of communi-
cating ego's state of mind as part of aseries ofindividual and culture-specific
adjustments to the social environment (Greenwald and Harder, 1998; Keltner
and Harker, 1998; Lazarus, 1991; Planalp, 1999). Mayer (1974) differentiated
between two kinds of behaviour - non-communicative and communicative.
Only the latter, he suggested, elicits responses that can be understood and re-
acted to, if Olle is to maximize chances of goal attainment. Consequently, he
subdivided the emotion 'fear' into that elicited by other species who endanger
the weIl-being of an organism and that elicited by conspecifics.
Hut we can conceive of blushing as expression of emotion as weIl as com-
munication, and to see these as dichotomies would, we suggest, be misleading
(see Ekman, 1997). It is interesting to observehow the change of colour, which
cannot normally be seen by the blusher, affects the other. The blusher only feels
the rise in temperature, and even this seems to occur only after the reddening of
the faceis visible (Shearn et al., 1990) to his/her interlocutor, who might react
even before the blusher realizes that the event hag been observed. This gives
him/her a very brief temporal advantage in which to decide between different
Shame across cultures 275
reactive strategies. The reaction will depend largely on the culture-specific code
that prescribes the correct behaviour in such asituation. Such a code obviously
varies according to class, gender and age and also depends on the specific goals
pursued in this interaction. Forinstance, mocking reactions amplify the feelings
of shame, typically expressed in the German children' s rhyme 'Pfui, pfui schäm
dich, alle Leute seh'n dich' (Fie, fie shame on rau, everyone sees rau). This
may lead to even greater self-focused attention, blushing propensity andfear of
blushing (Bögels, Alberts and de Jong, 1996; Shearn et al., 1990), and eventu-
ally tao many unpredictable reactions which the interlocutor would have td cape
wirb. But the brief period between the moment of observation and the perceived
temperature rise may be enough für an empathetic person, or für Olle who wants
to avoidfurther complications, to politely pretend not to have noticed the em-
barrassing event. It is clear then, that not only the specific cultural appraisals of
the antecedent event hut also the behavioural prescriptions and individual goals,
which are partly cu1t~ral and situation-bound, playamajor role in the perfor-
mance and outcome of (potentially) embarrassing interactions between indi-
viduals. Indeed, the 'blushing complex' is a potent mechanism to help or force
individuals to adjust to group-specific behavioural expectations, and is thus of
great significance in the development and maintenance of social conformity.
Embarrassment and shame function to rapidly integrale and reintegrate an
individual into the normative zeitgeist-specific behavioural mainstream of soci-
ety (cf. Planalp, 1999). When severe punishment (in this or the next world) can
be expected für major normative violations, persistent feelings of guilt occur;
these may lead to confession and finally to reintegration into a social group,
hut as Bastian and Hilgers (1991, p. 102) have noted, 'guilt requires shame as
The development of shame and related members of the 'family
of seil-reflexive emotions'
The results presented in the bulk of the literature dealing with the develop-
ment of gelt-reflexive emotions depend on definitions, as wen as on the types
of observations and experiments conducted. Wierzbicka (1986, 1992, 1999)
pointed to the question of generalizability of such findings. Lewis (1992, 1995)
hag shown that in Western populations the different members of the 'moral',
'social' or 'se1f-reflexive' family ofemotions, i.e. 'shyness', 'embarrassment',
'shame' and 'guilt', can be differentiated. In many languages, speakers do not
differentiate linguisticany between, für example, embarrassment and sharne,
sharne and guilt or sharne and anger (Fisher, 1985; Heron, 1992; Russen, 1989;
Schieffelin, 1983, also Larnbek and Solway, 2001); alternatively, a multitude
of related terms, all belonging to this 'family', are used. We must therefore
ask whether, and how many. such emotions which were found to emerge.
Michael J. Casimir and Michael Schnegg
tor exarnple, among children in the Uni ted States, are also distinguished or
distinguishable in other &ocieties. It is likely that all methods based on language-
perforrning experiments must fail if a specific differentiation of the 'family of
negative gelt-reflexive emotions' based on 'Western terms' is expected.
Cultural Darms and values tend to deternline which behaviours are consid-
ered shameful, and in many cultural contexts blushing can have both positive
and negative connotations. The phrase 'he/she blushes', can mean that he/she
hag trespassed a convention or role - hut it can also mean that the person is
weIl mannered. Here the event of blushing WOlle - i.e. the phenomenon without
further contextualization - cannot be correctly interpreted. A photograph pub-
lished by Izard (1992, p. 330, fig. 15.1) illustrating 'shame behaviours along
with a smile in a ten-year-old boy of northem India' is a good example of how
essential culture-specific contextualization iso Many north Indian languages and
most languages ofIran and Afghanistan distinguish between situations and feel-
ings, all of which are conducive to behaviour which could be glossed by the
single English term 'sharne'. Denotedby local terms such as 'sharm', 'laaj' and
'hayaa', these feelings can lead to blushing, lowering one's gaze, veiling, etc.,
hut their antecedents and connotations are varied and several terms in English -
such as embarrassment, modesty, respect, well-mannered, hut also pride and
honour, manliness and femininity - are required to convey the complex contexts
of their usage and meaning (cf. Cohen, Vandello and Rantilla, 1998; Lindisfame,
1998; Menon and Shweder, 1994; Wierzbicka, 1999). Fajans' (1985) discussion
of the complexity of 'sharne' in the folk model of emotions among the Baining
of New Britain and Hennann's (1995) and Jamieson's (2000) studies among
the Ngaing of Papua New Guinea and the Miskitu of Kakabila (Nicaragua),
respectively, are yet other examples of such variation.
But even if we consider only Western societies, the answers to the questions
when and in which sequence the various emotions, understood as different,
develop or mature, depend on the analytic methods used - foT exarnple on the
choice of behaviour patterns andlor language capacities thought to indicate the
presence of a given emotion. Is a given behaviour or facial expression a reli-
ahle indicator tor a specific emotion? Is a parent's report of the first signs of a
child's emotional expression a reliable indicator tor the presence of a given feel-
ing/emotion, or is this a matter of interpretation? Buss, Iscoe and Buss (1979),
für instance, concluded on the basis of paTents' judgementS of their children's
behaviour (including blushing and facial expressions) that '. . . embarrassment
begins tor most children at 5 years of age'. Does a child's capacity at a given age
to name a given emotion or describe a feeling parallel an adult's understanding
of this emotion? After all, '. . . thought-feeling structures are not leamed or en-
acted in a vacuum, hut in socia1life' (Strauss, 1992, p. 15). In same languages,
such as Chinese, the emotion we gloss as 'sharne' is hypercognitive (Heelas,
1986; Levy, 1973, 1984). Shaverandcolleaguesin 1992andWangin 1994 (both
in Russell and Yik, 1996) found more than 100 sharne-related words which are
Shame across cultures
all translated in Chinese-English dictionaries as 'shame', or as combinations of
shame with other emotions. It can be supposed that when a given emotion, or
emotion-cluster, is of outstanding cultural importance, children, in accordance
with Piaget's concept of decalage, comprehend and use such terms earlier than
their counterparts in societies where the term is not hypercognitive or even
hypocognitive (Heelas, 1986; Levy, 1973). It is therefore not surprising that
about 95 per cent of Chinese children understand 'shame' and some related
terms as early as at age 2.5-3 years, in stark contrast to American children, of
whomon1y 10per cent in that age groupunderstoodthe term 'ashamed' (RusseIl
and Yik, 1996, pp. 180-1). Using aWesternsamp1eHarterandWhitesell (1989)
showed that only after the emotions 'happiness', 'sadness', 'anger' ( cf. Wellman
et al., 1995) and 'fear' matured, did chi1dren aged 6-7 years understand the feel-
ings glossed by the terms 'pride' and 'shame', and Griffin (1995, p. 224) reports
that only at the age of 7-8 years, in the stage of bi-intentional thought did the
majority of children report self -conscious emotion experience. Mascolo and Fis~)
chef (1995, p. 65) point out' . . . that shame paralleIs pride development, whereas
a positive appraisa1 develops pride, a negative Olle shame'. In their view, distress
leads to distress/shame and thus to shame (p. 84), while joy leads to pride (p. 73).
Two postulates attempt to exp1ain the emergence of 'shame' (and so,me related
emotions) in human ontogeny. The first focuses on self-recognition, the second
sees the emergence of se1f-consciousness and shame as mutually independent:
I. According to the first postulate, 'shame' can occur on1y after the infant
has deve10ped se1f-recognition. Only then can the child re1ate cultural norms,
'inculcated' by the caregiver(s) to its self, classify behaviour as 'wrong', and
recognize this as its own fault. Se1f -consciousness is aprerequisite für empa-
thy (Bischof-Köhler, 1989, 1995; cf. also Zahn-Waxler and Robinson, 1995).
During the process of socialization, caregivers 'inculcate' desired and expected
behaviour and 'punish' deviations from convention. Bischof-Köhler (1989)
rejects Izard's (1980a) and Izard and colleagues' (1980) suggestion that embar-
rassment develops as ear1y as 8 months of age, because this feeling implies the
conscious reflection ofthe se1f, which develops much1ater. Lewis (1992) states
that embarrassment and shame are basically the same emotion, hut at different
levels of intensity. Embarrassment, he points out, is of two types: the first, is
more sirni1ar to shyness than to shame, and is not related to negative evaluation.
Shyness, he states, can be understood as oscillation between fear arid interest,
or between avoidance and approach, and is not related to self -evaluation in
children in the first month of life. The second type of embarrassment is re1ated
to negative self -evaluation, and can be understood as less intense shame.
Self-recognition deve10ps not earlier than at 15 months - it is only at 21-24
months that the majority of children tested showed this capacity (Bischof-
Köhler, 1989; cf. also DiBiase and Lewis, 1997; Scherer, 1993b; Smi1ey and
Huttenlocher, 1991). Lewis (1992, 1995) also observed that the secondary or
se1f-conscious emotions are observed onlv in the middle ofthe second vear. This
278 Michael J. Casimir and Michael Schnegg
seems to be paralleled by the emergence of olle of the typical shame-related
behaviours - shying away, which, according to Mascolo and Fischer (1995),
can be observed in children about 2 years old; only at the age of 2.5 to 3 years
are clusters of such behaviours enacted in shame-related situations.
These various observations support Darwin's (1872/1965) suggestion that
very young children; who bad not yet imbibed their culture-specific Darms, did
not feel ashamed:
Children at a very early age do not blush; nor da they show those other signs of self-
consciousness which generally accompany blushing; and it is Olle of their chief charms
that they think nothing about what others think of them. (p. 326)
11. The second postulate is based on the assumption that shame and related
feelings can arise even beraTe the development of self-recognition. Ferguson
and Stegge (1995) imply that shame develops early in the infant, caused by
the experience that its behaviour does not meer the expectations of the care-
giver. Here it is thought that the withdrawal of love and/or other negative reac-
tions towards the child elicit fear or guilt, based on anxiety (cf. also Magai,
Distel and Liker, 1995; Rosenberg, 1991). Barrett (1995) also stresses the
important function of the caregivers. If they show love aso areaction to the
child's behaviour, the child experiences bis self as loveable. If the caregiver
shows negative reactions '... children [are] more likely to believe that they
have failed, and to experience the shame of failure' (p. 54). In a comparable
approach, based on Bowlby's (1969) 'Attachment theory' and developmental
neurophysiology, Schüre (1998) demonstrates the functional interaction be-
tweeD the maturation of the orbifrontal cortex, the infant-caregiver behaviour
and the development of emotions of the 'sharne farnily'. Barrett (1995, p. 35)
suggests that '. . . general cognitive acquisitions are not viewed as prerequisites
für the emergence of entire emotion "farnilies" such as shame or guilt'. Here the
crucial question arises whether 'general cognitive acquisitions' include those
necessary für self-recognition and self-reflexive emotions, and whether the self
can exist or be conceived of without self-awareness/self-consciousness?
In what can be considered an attempt to combine these two conjectures, Lewis
(1995), who differentiates 'embarrassment' flom 'sharne', postulated that an
early form of embarrassment, not associated wirb the self's negative evaluation
of the self's action, can be differentiated flom a second type of embarrassment,
occurring only between the second and third years of lire and associated wirb
self-evaluation. 'This form of embarrassment' he writes (p. 215) 'is related to
the negative self -conscious evaluative emotions, such as shame and guilt.' He
speaks of primary emotions (joy, real, anger, sadness, disgust and surprise),
which occur shortly after birth, or are seen within the first 6 to 8 months
of lire. Thereafter, wirb the growing cognitive capacity für se1f-awareness
(15-24 months), these are followed by the 'exposed emotions' (embarrass-
meßt, empathy and envy). Only then, wirb the standards, rules and goals learned,
Shame across cultures 279
by about the third year oflife, do the 'self-conscious evaluative emotions' (em-
barrassment, pride, shame and guilt) emerge (pp. 207-8 and fig. 7.1).
From the above it appears that 'specific emotions such as pride, sharne and
guilt do not emerge at a single point in development' (Mascolo and Fischer,
1995, p. 72). The lil1kS between the elicitation of specific emotions and indi-
vidual events are established in early childhood and adolescence, and triggered
by roJe models during the process of socialization; specific emotions, espe-
cially sharne, themselves influence development (Barrett, 1995; Scherf, 1990).
Later in lire, many of these emotions are mainly released in circumstances, and
following events, which generate meaning in the framework of cultural norrns
and values, because the emergence of such feelings are based on given cultural
contextualizations. However, the debate is rar from over, and wirb a look to
the future of research on emotional development, Mascolo and Fischer (1995,
p. 72) suggested that development should be considered as 'a variable web',
rather than as 'a uniform ladder of stages':
It is not fruitful to debate about the point at which an emotion first 'really' emerges in a
child. . . . Treatment of emotions as developing at one point or age will lead to unproduc-
tive debates. . . Emotions do not emerge fully formed at Olle age; instead they develop
gradually over extended periods of time, taking increasingly complex forms. One can al-
ways identify some form of a given emotion or ability at multiple points in development.l
All these observations and postulates show that many questions still need to
be asked and many problems remain to be solved before we understand those
aspects of the interrneshing of biology and culture that bring about the devel-
opment of emotions. In search of the similarities and differences of emotional
patterns in different societies we should start with the assumption that basic
human biology is the same everywhere. The analysis of the impact of diverse
cultural patterns on the elicitation of different emotions has, however, only just
begun. To develop a panhuman model of emotional development we have first
I These extremely complex issues have teen intensely debated and can not be treated hefe in
detail. As early as 1932, Bridges (in Buechler and Izard, 1983, pp. 297ff.) postulated in his
differentiation theory that generalized excitement is the only emotion present at birth and that
distress develops during the first month. The first signs of anger, he suggested, evolve from distress
between the third and sixth months of lire and clear indicators of anger were not considered to
exist earlier than in the twelfth month. Scroufe (1976, in Buechler and Izard, 1983) is of the
opinion, however, that rage emerges before anger at 3 months, and the latter on1y in the seventh
month; according to hirn anger deve1ops from rage. Charlsworth (1969, in Buechler and Izard,
1983) suggests that 'surprise' (defined as the emotion arising with the recognition of deviation
from expectancy) is present in the seventh month. It has also been suggested (Izard, 1980b; lzard
et al., 1980) that at birth, the emotions interest, disgust, discomfort and shock can be released
and that it is only after about age 6 months that happiness, anger and surprise, and after 1 year
sadness and fright, mature. Emde (1980) gave the following mean ages at which mothers feIt
that given emotions first occurred in their infants: 2.6 months tor interest, 2.9 months tor joy,
5.6 months tor surprise, für anger 6.3 months, für fear 8.1 months, für shyness and distress 9.1
months and 8.9 months, respectively. However, the capacity für the strategic control of emotional
expressions develops much later (Saarni, 1991) and is held to be shaped by culture-specific mIes
and conventions (cf. Goldschmidt, 1976; Lindholm, 1988).
280 Michael J. Casimir and Michael Schnegg
to determine behavioural and terminological similarities across cultures, which
allow comparisons and enable us to develop and conduct new experiments.
Blushing and the panhuman shame complex: the 'natural
semantic metalanguage' and a cross-cultural approach
The shame-blushing complex is, we believe, eminently suitable für a cross-
cultural analysis. Darwin suggested that blushing2 occurs in all humans. and
that its meaning is always related to the same specific emotion(s). Even when
among darker-skinned persons the flush was less clearly visible than among
fairer Olles, he wrote:
. . . there can always be seen the same expression of modesty or confusion; and even in
the dark, anse oftemperature ofthe skin ofthe face can be feIt... (Darwin, 1872/1965,
Our starting point was the insight that in many European languages people
tend to describe certain emotional states by remarks such as someone's 'face
tumed red with shame because he was caught red-handed', or again 'he/she
tumed red with shame', or simply 'he/she blushed'.1n terms, of Wierzbicka's
(1986) NSM, the general structure here is: 'something happens because ofthis'.
As the basis für this analysis we shall use the structure of Wierzbicka's (1986,
p. 592; cf. also, 1999, p. 110) formulation:
X was ashamed = X feIt as ODe does when ODe thinks that other people see that ODe
hag düne something ODe should not do, when one trunks that other people may trunk
something bad of ODe because of that, and when one wants to ce ase to be seen by other
people because of that.
In OUT context the change in facial colour took place because cf, or coincided
with, a given emotion, caused by a specific event. It was now conjectured that
when, in a given language, native speakers use phrases (see appendix 11) in
colloquial speech, the physiological phenomena parallel specific events and
indicate specific emotional states. Sometimes this relationship between event
and the physiological effect is expressed in metonyms or metaphors. These
metaphors and metonyms are often related to both physiological changes and
culture-specific ideas about the world and its functioning. As Kövecses' (1995,
1998) as weIl as Athanasiadou and Tabakowska's (1998) studies demonstrate,
many such expressions are related to bodily changes or feelings (in cold blood,
his/her blood boiled,-etc.), which coincide with or parallel emotional changes.
The physiological principles involved hefe were described by Darwin (1872/1965) and it has
recently been shown (Drurnrnond, 1997; Mariauzouls, 1996) that active ß-adrenergic vasodilation
is a major process leading to a blush. Although several precise mechanisms have been postulated
(for an overview see Gerlach, 1998), the neurobiology of blushing is still poorly understood
(Stein and Bouwer, 1997; Drummond, 1994). It is also not clear whether the same and/or other
mechanisms are involved when a flush is caused by anger or rage (Drurnrnond, 1994).
Shame across cultures
The data and the analysis
On the basis of usage in several European languages, we designed a matrix -form
questionnaire, where the emotions joy/happiness, shame, anger, rage, fear, rain,
shock and envy listed in rows could be related to Olle or more colours listed
in the colurnns. The terms 'bearningishining' were added to the list of colours,
because very early it was found that the emotions 'joy' and/or 'happiness'
were orten related to the expressions 'bearning', 'shilling eyes' or 'a shilling
or bright face'. The term 'sharne' was used by us to designate a prototype tor
all emotions/feelings belonging to the farnily of more or less negative self-
conscious emotions, since in many if not most languages no terrninological
differentiation is made between shame and embarrassment (and, as we shall
see later, even between same other emotion terms).
Only native speakers were interviewed. The task was explaiiied to the infor-
mant in his/her native language/dialect or, to bilingual informants, in a European
language in which he/she was clearly fluent. Informants were asked to relate
facial colour change to emotions. In addition, they were asked to build
colloquial sentences expressing these event-relationships. Remembering that
in many languages only same 'basic colour terms' (Berlin and Kay,- 1969) are
used, descriptive terms such as 'colour of blood', 'colour of the Sky' or 'leaf
coloured' were laken as equivalents of specific colour terms, irrespective of
whether a 'basic colour term' was present in a particular language or whether
the informant was using a metaphor. All these sentences were recorded either in
the local script or transcribed and then translated literally into English, French,
German or Spanish. Whenever possible, more than Olle informant was ques-
tioned in order to counter individual and local variation. In OUT analysis hefe
we shall focus only on the data concerning the emotions shame, rage and anger.
During data collection, in most cases a semantic structure very sirnilar to
Wierzbicka's NSM (1986, 1988, 1992) emerged, consisting of the antecedent
situation (implicit the appraisal), the emotion term and the related colour. The
semantic structures corresponded mostly to Olle of the tour following:
(a) 'He/she turned colour A.'
(b) 'Colour A (in situation C) because of emotion B.'
(c) 'Colour A [occurred (on the face)] because of emotion B.'
(d) 'Because of situation C colour A [occurred (on the face) (because of emotion
Figure 12.1 shows how frequently emotions were related to facial colour
change in 98 of the 135 different languages/dialects analysed.3 The numbers
In the remaining 37 languages no relationship was found to be expressed between any of these
emotions and any colour term. However, in many of these languages such sudden emotional
transitions were expressed in metaphors indicating the same physiological changes as those
which took place in the speakers of the other 98 languages. Hence, it is plausible that bad
282 Michael J. Casimir and Michael Schnegg
t:;c:"t.~ ~~.§ e.
Figure 12.1 Frequency with which emotions were related to colour met:
phoTs in 98 of the 135 languages/dialects recorded.
on the y-axis indicate the frequency with which each of the emotions was
associated with any colour. Since an emotion can be related to more than Olle
colour, these values can exceed 98. The figure shows clearly that the differ-
ences between the emotions are neither highnor significant; nevertheless, same
measure of variation was observed: shame, fear and rage were most frequently
associated with same colour(s), whereas pain, anger and envy were found to
be less so. In a next step we switched perspectives, to ex amine the frequency
with whith specific colours were related 10 emotions. Figure 12.2 sets out
these results, whereby the frequency with which any given colour was men-
tioned is represented by the height of the bars. The rank order hefe is much
more differentiated with red, white and black dominating and red and white
figuring more than tour times as orten as other colours. Certain colours, such
as red, were related in same languages to multiple emotions, hence the fre-
quency with which these were mentioned weIl exceeded the language sampIe
Figure 12.3 shows in how many languages each of the emotions shame, anger
and rage is related to a specific colour. Shame, tor instance, is associated with
red in 78 of the 98 languages examined. In general red followed by black, and
less frequently white is associated with shame, anger and rage. The frequency
these changes been as visible among the darker complexioned speakers of these 37 languages
as they are among speakers of the remaining 98, they would have also been expressed in similar
metaphors. This logic applies equally to those languages in which no metaphors are used to
indicate the physiological changes. This assumption justifies our not treating these 37languages
as those with 'negative responses', hut rather ignoring them in our calculations.
Sharne across cultures 283
Figure 12.2 Frequency with which colours or colour metaphors were re-
lated to emotions in the various sentences recorded. The colours with a high
frequency were orten related to multiple emotions.
Figure 12.3 Frequency of associations between colours/colour change and
the emotions shame, anger and rage in 98 languages.
of all the other colours is low and roughly equal. Figures 12.4a and 12.4b set
out the geographical distribution of the languages/dialects recorded in which a
connection is made between primarily red and shame, in other words in which
the metaphor 'blushinß' is used and others in which it is absent. From these maps
Michael J. Casimir and Michael Schnegg
it can be seen that with only a few exceptions expressions indicating blushing
are used by populations with fair complexions.
The few exceptions can, we suggest, be explained as folIows. Firstly, when
the postulated blushing was hardly visible due to the speaker's relatively dark
complexion, red was sometimes substituted by black - a phenomenon that is
comprehensible, since a dark complexion becomes darker still when the blood
flow into the capillaries intensifies (Darwin, 1872/1965). We can assume that
liefe is a threshold of melanin concentration above which the effect is too small
to be noticeable.
Secondly, in many cases, especially when the speakers' complexions were
very dark, no colour term was related to shame. Sometimes, however, other
metaphors indicating physiological processes were used; many of these point
to a rise in temperature - e.g. 'he/she (his/her) (face/neck) became hot because
It should also be mentioned that in some languages in Africa shame is orten
associated with 'red eyes', 'fire', or 'blood'; in Zulu angeris associated with the
reddening ofthe sky orof an inflarned eye and Taylorand Mbense (1998, p. 203)
explain that these'. . . expressions testify to the power of the conceptualisation
of anger in terms of the body filling up with blood'. .
So rar we have shown (1) that emotions arerelated to (facial) colour change
in a broad range of languages and (2) which colour(s) is(are) associated with
the spectrum of emotions shame, anger, and rage. These findings point to a
more generalized pattern in which red (and white, though not analysed hefe)
seem crucial, a result that is strikingly corroborated by correspondence analysis
(Greenacre, 1993) of our data. The scatter-plot can be viewed as a 'statistical
map' where substantive relations between iterns are transformed into proxirnity
in the image space. Those colours clustering around a particular emotion are
typical für this emotion.4
Figure 12.5 shows the plot of the first two dimensions obtained from cor-
respondence analysis. On the horizontal axis that captures 51.9 per cent of
the variance in the data, there is a clear split between shame/anger/rage and
pain/fear/shock corresponding respectively to red and white. The predorni-
nant dimension underlying the relationship between colours and emotions is,
however, Olle that separates white from red and simultaneously profi, fear and
shock from sharne, anger and rage. On the vertical axis, which explains an ad-
ditional 30.8 per cent of the variance of the data, a split between envy, mostly
associated with green (for which no physiological explanation is available), and
all the rest of the emotions and colours is visible.
The results presented from the cross-cultural sampIe (see appendix I) of
98 out of the 135 languages in which colour metaphors were used to signal
4 Shining/beaming was excluded tram this analysis because it occurs only in combination with
the emotion joy/happiness. This would introduce a new, partly artificial dimension.
Shame across cultures 287
I. I. I I
Figure 12.5 Correspondence analysis of the use of colour terms or metaphors
and the emotions examined.
emotional change indicate a high degree of universality or intercultural con-
sensus. Yet they leave room für variation, and none of the relationships holds
true für all individuals, languages or other speech forrns. To determine more
precisely to what degree these exceptions to the overall pattern are random,
or hide substantial differences across languages or cultures, a second sampIe
of informants was interviewed. This intracultural sampIe consisted of forty-
rune Gerrnan students, socialized in different parts of the country and hence
exposed to different Gerrnan dialects. Though tlle units of analysis are differ-
ent (languages in Olle sampIe and individuals in the other) this second sampIe
can be used as a control group to determine (1) how sirnilar or different the
picture of the emotion-colour relationship can be within Olle language and
(2) the degree of difference between the internal variations within this sin-
gle language sampIe on the Olle hand and the languages in the cross-cultural
288 Michael J. Casimir and Michael Schnegg
sampIe laken together on the other. A sirnilar or even lower variation within a
single language than that within a cross-cultural sampIe would strongly sup-
port our contention that we are looking at a universal phenomenon. We shall
ex amine both hypotheses separately: first, the similarity of the two colour-
emotion spaces and second the relative amount ofinternal variation in the two
The simplest approach is to compare the relative judgements on each pair
of items and to compute from these the absolute differences between the two
samples.5 The index of similarity (Sab) varies between 0 and 1, where 1 indicates
total equality between two populations and 0 total difference. Computing this
index of similarities für our two sampIes we retain a similarity value (Sab) of
0.91 between the two.6 Thus, within Olle population speaking Olle language there
is roughly the same degree of variation as in a sampIe composed of members
speaking 98 languages.
Still the theoretically more challenging question remains. In spite of sound
evidence that Iorty-nine speakers of Olle language are as good (if not bet-
ter) predictors of cross-cultural content as individuals speaking ninety-eight
languages/dialects, we do not know the extent of intern al variations. Only
if the internal variations within the two sampIe populations. are similar can
we be certain that there is a universal understanding of the colour-emotion
Variations can be defined as deviation from consensus on a specific judge-
meßt. If all members of Olle population agree that there is a relationship between
a specific emotion and a specific colour (say shame and red), there is no vari-
ation; there exists 100 per cent consensus within this group. The same extent
of consensus can also result from an agreed negative judgement: if all mem-
berg of a group share the notion that there is no relationship between any two
items (e.g. shame and green) we would also find 100 per cent consensus or zero
variation für that pair of items. The amount of variation within Olle population
can be calculated if we sum up all absolute departures from perfect consensus:
The index of variation (V) varies between 0 and the number of judgements
divided by 2, where 0 indicates no variation and the maximal value total dis-
agreement; this value of maximum variation would be reached if, on each pair
of items, 50 per cent of the population agrees with the judgement and 50 per cent
5 The similarity (Sab) between two sampIe populations (a and b) can be expressed as Sab =
1- I: IXij/n - Yij/ml / k where Sab is the index of similarity, Xij is the judgement of the
first population about a given pair of i.tems, Y ij the judgement of the second population about a
given pair of items, n the size of the first population, m the size of the second population and k
the number of pairs. See note 2 for the reasons für taking 98 as the basis of our calculations in
the cross-cultural sampIe.
6 The validity of the measure is also supported by a Pearson correlation of r = 0.84 between the
7 Va = I: Xij/n for Xij/n ~ 0.5 + I: l-Xij/n für Xij/n > 0.5. All indices are the same as above.
Shame across cultures
disagrees. The coefficients Va and Vb für the populations considered hefe are 5
für the intercultural sampie and 6 für the German sampie. Given the restriction
that the index varies between 0 and 24.5 für the size of the matrices analysed,
both values are very close to each other. Even more surprising is the fact that
the variations within the German sampie are greater than those in the sampie
of ninety-eight languages from all over the world. This points clearly to the
existence of a panhuman relationship between experiencing specific physiolo-
gical effects and the perception of facial colour change. The contexts/situations
(antecedents) in which the emotions are feIt are, however, culturally constructed,
especially in the case of 'shame'.
Shame, fea1; anger and rage
The metaphor linked in most languages to the emotions 'fear', 'shock' and 'profi'
was either 'white' o! terms close to it, such as 'grey', 'ash coloured' or 'pale'.
Here, obviously, the principal physiological process is the reverse ofblushing or
flushing - hefe blood withdraws from the capillaries in frightening situations. In
cases ofboth anger and fear there is arise in diastolic bloodpressure (Ax, 1953;
Roberts and Weerts, 1982), hut it is significantly higher in cases of intense anger
than of extreme fear. Anger and rage (defined as 'anger out of control', Lewis,
1992, p. 153) are accompanied by a higher heart rate, rising blood pressure and
higher breathing frequency, and in many cultures/languages the related feelings
are expressed with what Kövecses (1995,1998; see also Lakoffand Kövecses,
1987) termed the 'container under pressure' metaphor. In a shameful situation,
however, in the short moment of blushing, neither a rise in blood pressure flor
a higher heart rate can be observed (Gerlach, 1998).
An analysis of the subsampIe of seventy-eight languages in which red was
associated with shame reveals that in sixty-two of these (79 per cent) this colour
was also associated either with anger or with rage; in fifty-one cases (65 per cent)
it was related to anger, and in forty-eight (62 per cent) to rage. In twenty-seven
(35 per cent) of these seventy-eight languages red was associated with all the
three emotions - shame, anger and rage.
The connotations of the term 'shame' exhibit remarkable fuzziness in many
languages and the close relationship between shame, fear and anger (Fisher,
1985; Heron, 1992; Schieffelin, 1983) has orten been described. In some soci-
eties feelings of anger/rage are feared and should be suppressed; if they arise
they lead to the feeling of shame (Dentan, 1978; Levy, 1978; Rosaldo, 1983)
in the angry person. In other societies shameful emotions are related to feel-
ings of anger (Fisher, 1985; Heron, 1992; Rosaldo, 1984; Schieffelin, 1983)-
a famous example is that of 'running amok' which occurs in Malay society
and leads to anger and enraged homicidal attacks (Winzeler, 1990) - which
are projected either on the seIt that has failed socially, or on those who have
Michael J. Casimir and Michael Schnegg
noticed this failure and thus constitute a danger to the self (cf. also EI-Islam,
1994; Okano, 1994; Sachdev, 1990; Sharpe, 1987). This danger may, in turn,
lead to anger/rage in the ashamed person. It can be assumed that after the first
intense moments of shameful insecurity, confusion, fear, anger or rage may
overwhelm olle, leading to a secondary flush, and once again to shame. This
is in accordance with Lewis' (1992, pp. 149-53) findings that 'the shame-rage
connection can lead to a spiral in which the rage itself becomes a new source
ofshame' (cf. Retzinger, 1987; Scheffand Retzinger, 1991; Schüre, 1998).
In the foregoing we saw that the negative gelt-reflexive emotions belonging to
the shame/embarrassment/shyness complex and terrninologically differentiated
in different languages and societies are orten expressed locally by using either
the 'shying a"Yay' or 'bltishing' metonyms, or both. These emotions play an
important role in the psychological and behavioural shaping of the individual in
childhood and early youth, thus allowing für his/her integration and functioning
in a given society. The physiological principle operating on all the occasions
glossed by the English term 'sharne' also activates the various facial motor
patterns. As Ekman (1972, p. 279 in Ekman, 1980, p. 138) noted:
. . . the elicitors, the particular events which activate the affect program, are in largest
part socially learned and culturally variable, and.. . many of the consequences of an
aroused emotion also are culturallv variable. . .
The cross-cultural analysis and the related correspondence analysis have
shown that the metonyms indicating the reddening, mainly of the face, are used
to indicate states of rind - emotions which are part of this shame complex
and/or of the anger/rage complex. If the physiology of the shame-blushing
complex occurs in all human beings, whether visible or not, it is worthwhile to
explore its phylogenetic roots and examine whether it serves similar functions
in the social context.
We suggest that there is a panhuman experience deeply rooted in evolution,
which first links specific events and appraisals, then leads to the elicitation of
these self-retlexive emotions and finally to behaviour which serves the function
of enhancing the chances of individual survival. We posit the following: indi-
vidual emotions in different sets of 'emotion families' are released by events
of specific sets of 'event farnilies'. The fuzziness of both sets, between cultures
and within 'olle' culture over time, is the result of different culture-specific
concepts about the world. These emotions are experienced and paralleled by
physiological reactions that cause bodily symptoms or syndromes; again, the
cultural setting greatly intluences which bodilyfeeling(s) are taken into account
when speaking of emotions.
Shame across culture: ')Ql
We know that the junction of specific extern al events with physiological
reactions leading to bodily feelings developed in the course of evolution. This
development took place in the differential process of optirnizing the proximate
means of reaching the ultimate goal of maxirnizing Darwinian fitness. With
the evolution of complex social and cultural structures the number of events
included in such' external event farnili"es' grew. Feelings of fear and anger were
now released, not only by predators, hut also by conspecifics and members of
the same social group, when hierarchically ordered sex- and age-related Darms
began to constrain the degrees offreedom. The 'self-conscious emotions' now
enabled the individual to react quickly in self-defence. A second function:
however, developed in all groups where the flush or blush can be observed: it
serves as a signal für the 'other' about ego's state of mind and gives the former
the opportunity to react according to his/her own purposes - für the benefit or
detriment of the blusher.
The data used in this charter indicate that whenever a behavioural norm has
been violated and this violation has been observed by others, feelings belong-
ing to the farnily of self -conscious emotions arise in the violator. These are
often accompanied by a blush. When a more or less fair complexion fenders the
blush visible, this reddening of the face expresses and comrnunicates (Ekrnan,
1997) specific feelings. In these cultures theblushing phenomenon is expressed
metonyrnically or metaphorically in everyday speech, to indicate shame, embar-
rassment or sirnilar feelings that belang to this 'emotion farnily'. Such speech is
encountered in the hypercognitive and hypocognitive emotion vocabularies of
different cultures (Heelas, 1986; Levy, 1973, 1984) with their varied behavioural
Darms. This points to epigenetic d~velopment processes in early childhood and
to the flexibility ofhuman 'emotional intelligence'; it enables humans to cape
optimally, though differentially, with their specific social environments.
Arrnon-Jones, C. (1986). The thesis of constructionism. In R. M. Harre (ed.), The social
construction 0/ emotions. Oxford: Basil Blackwell, pp. 32-56.
Athanasiadou, A. and Tabakowska, E. (1998). Speaking ofemotions. conceptualization
and expression. Berlin: Mouton de Gruyter.
Averil1, J. R. (1980). A constructivist view of emotion. In R.Plutchik and H. Kellermann
(eds.), Emotion: theory, research, and experience. San Diego: Academic Press,
Ax, A. (1953). The physiological differentiation between fear and anger in humans,
Psychosomatic Medicine, 15,433-42.
Barrett, K. C. (1995). A functionalist approach to shame and guilt. In J. P. Tangney
and K. Fischer (eds.), Self-conscious emotions. The psychology 0/ shame, guilt,
embarrassment, and pride. New York: Guilford Press, pp. 25-63.
Bastian, T. and Hilgers, M. (1991). Scham als Teil des Minderwertigkeitsgefühls - und
die fehlende Theorie der Affekte (Shame as part of feelings of inferiority - and
292 Michael J. Casimir and Michael Schnegg
the missing theory of affect). Zeitschrift fiir Individualpsychologie, 16, 102-
Berlin, B. and Kay, P. (1969). Basic color reims: their universality and evolution.
Berkeley: University of CaIifomia Press.
Bischof-Köhler, D. (1989). Spiegelbild und Empathie. Die Anfänge der sozialen Kog-
nition (Mirror image and empathy. The beginnings of sociaI cognition). Stuttgart:
(1995). The development ofempathy in infants. In M. E. Lamb and H. Keller (eds.),
Infant development: perspectives from Gelman-speaking countries. HilsdaIe, NJ:
Lawrence Erlbaum, pp. 1-33.
Bögels, S. M., Alberts, M. and De Jong, P. (1996). Self-consciousness, self-focused
attention, blushing propensity and fear of blushing. Personality and Individual
Differences, 21(4), 573-81.
Bowlby, J. (1969). Attachment and lass, vol. 1: Attachment. New York: Basic Books.
Buechler, S. andIzard, C. E. (1983). On theemergence, functions, andregulationofsome
emotion expressions. In R. Plutchik and H. Kellerrnann (eds.), Emotion: theory,
research;\a!ld experience, 2nd edn. San Diego: Academic Press, pp. 305-40.
Buss, A. H., Iscoe, I. and Buss, E. (1979). The development of embarr~ssment. Journal
of Psychology, 103,227-30.
Cohen, D., Vandello, J. and Rantilla, A. K. (1998). The sacred and the sociaI: cul-
tures of honor and violence. In P. Gilbert and B. Andrews (eds.)~ Shame: interper-
sonal behaviol; psychopathology, and culture. New York : Oxford University Press,
Darwin, C. (1872/1965). The expression of emotions in man and animals. Chicago:
University of Chicago Press (Originally published by D. Appleton and Company,
(1872/1998). The expression of the emotions in man and animals. Definitive edi-
tion. Introduction, Afterword and Commentaries by P. Ekman. New York: Oxford
Dentan, R. K. (1978). Notes on childhood in a nonviolent context: the Semai case. In
A. Montagu (ed.), Learning non-aggression: the experience of non-literate soci-
eties: Oxford: Oxford University Press, pp. 94-143.
DiBiase, R. and Lewis, M. (1997). The relation between temperament and embarrass-
meßt. Cognition and Emotion, 11 (3), 259-71.
Drummond, P. D. (1994). The effectof anger and pleasure onfacial blood flow.Australian
Journal of Psychology, 46(2), 95-9.
(1997). The effect of adrenergic blockade on blushing and facial flushing. Psychophys-
iology, 34, 163-8.
Ekman, P. (ed.) (1973). Darwin andfacial expressions: a century ofresearch in review.
New York: Academic Press.
(1980). Theface ofman: expressions ofuniversalemotions in aNew Guinea Village.
New York: Garland STPM Press.
(1989). The argument and evidence about universals in faciaI expressions of emotions.
In L. H. Wagner and A. Manstead (eds.), Handbook of social psychophysiology.
New York: John Wiley, pp. 143-64.
(1997). Should we caIl it expression or. cornmunication? Innovation, 10, 333-
293Shame across cultures
Ekman, P. and Oster, H. (1979). FaGial expressions of emotion. Annual Review oi
Psychology, 30, 527-54.
EI-Islam, M. F. (1994). Cultural aspects of morbid fears in Qatari women. Social
Psychiatry and Psychiatric Epidemiology, 29, 137-40.
Emde, R. N. (1980). Levels of meaning für infant emotions: a biosocial view. In W. A.
Collins (ed.), Development oi cognition, affect, and social relations, The Minnesota
Symposiaon ChildPsychology,vol.13.Hillsdale,NJ: Lawrence Erlbaum, pp. 1-37.
Fajans, J. (1985). The person in social context: the social character of Baining
'Psychology'. In G. M. White and J. Kirkpatrick (eds.),Person, seit and experience:
exploring Pacific ethnopsychologies. Berkeley: University of California Press,
Ferguson, T. J. and Stegge, H. (1995). Emotional states and traits in children: the case of
guilt and shame. In J. P. Tangney and K. W. Fischer (eds.), Self-conscious emotions:
the psychology oi shame, guilt, embarrassment and pride. New York: Guilford
Press, pp. 174-97.
Fisher, S. F. (1985). Identity of two: the phenomenology of shame in borderline devel-
opment and treatment. Psychotherapy, 22(1), 101-9.
Gallup, G. G. (1970). Chimpanzees: selfrecognition. Science, 157,86-7.
(1977). Self-recognition in primates. American Psychologist, 32, 329-38.
Gerlach, A. L. (1998). Blushing, embarrassment, and social phobia: physiological,
behavioral, and self-report assessment. Marburg: Tectum Verlag.
Gilbert, P. and McGuire, M. T. (1998). Shame, status, and social roles: psychobiology
and evolution. In P. Gilbert and B. Andrews (eds.), Shame: interpersonal behavio1;
psychopathology, and culture. New York: Oxford University Press, pp. 99-125.
Goldschmidt, W. (1976). Absent eyes and idle bands: socialization für low affect among
the Sebei'. In T. Schwartz (ed.), Socialization as Guttural communication. Berkeley:
University of California Press, pp. 65-71.
Greenacre, M. J. (1993). Correspondence analysis in practice. London: Academic Press.
Greenwald, D. F. and Harder, D. W. (1998). Domains of shame: evolutionary, cultural,
and psychotherapeutic aspects. In P. Gilbert and B. Andrews (eds.), Shame: inter-
personal behavio1; psychopathology, and culture. New York: Oxford University
Press, pp. 225-45.
Griffin, S. (1995). A cognitive-developmental analysis of pride, shame, and embarrass-
ment in middle childhood. In J. P. Tangney and K. W. Fischer (eds.), Self-conscious
emotions: the psychology oi shame, guilt, embarrassment and pride. New York:
Guilford Press, pp. 219-36.
Hamburg, D. A. (1964). Emotions in the perspective ofhuman evolution. In P. H. Knapp
(ed.), Expression oi emotions in man. New York: International University Press,
Harkins, J. and Wierzbicka, A. (1997). Language: a key issue in emotion research.
Harre, R. M. (ed.) (1986). The social construction 01 emotions. Oxford: Oxford
Harter, S. and Whitesell, N. R. (1989). Developmental changes in children's understand-
ing of single, multiple, and blended emotion concepts. In C. Saami and P. L. Harris
(eds.), Children's understanding oi emotion. Cambridge: Cambridge University
Press, pp. 81-116.
294 Michael J. Casimir and Michael Schnegg
Heelas, P. (1986). Emotion talk across cultures. In R. M. Harre (ed.), The social con-
struction of emotions. Oxford: Oxford University Press, pp. 234-66.
Hermann, E. (1995). Emotionen und Historizität. Der emotionale Diskurs über die Yali-
Bewegung in einer Doifgemeinschaft der Ngaing, Papua New Guinea (Emotions
and historicity. The emotional discourse on the Yali movement in a village com-
munity of Ngaing, Papua, New Guinea). Berlin: D. Reimer Verlag.
Heron, J. (1992). Feeling and personhood: psychology in another key. London: Sage.
Izard, C. E. (1980a). Cross-cultural perspectives on emotion and emotion communica-
tion. In H. C. Triandis and W. Lonner (eds.), Handbookofcross-cultural psychology,
val. 3: Basic processes. London: Allyn and Bacon, pp. 185-221.
(1980b). The emergence of emotions and the development of consciousness in infancy.
In J. M. Davidson and R. J. Davidson (eds.), The psychobiology of consciousness.
New York: Plenum Press, pp. 193-216.
(1992). The psychology of emotions. New York: Plenum Press.
Izard, C. E. and Buechler, S. (1980). Aspects of consciousness and personality in terms
of differential emotions theory. In R. Plutchik and H. Kellerman (eds.), Emotion:
! theory and.research, val. 5: Emotion, psychopalhology and psychotherapy. London:
Acadernic Press, pp. 165~87.
Izard, C. E., Huebner, R. R., Risser, D., McGuinnes, G. C. and Doughen, L. M. (1980).
The young infant's ability to produce discrete emotion expressions. Developmental
Psychology, 16, 132-40. .
Jarnieson, M. (2000). It's shame that makes men and warnen enernies: the politics of
intimacy among the Miskitu of Kakabila. Journal of the Royal Anthropological
Institute (n.s.), 6, 311-24.
Johnson, M. (1990). The body in the mind: the bodily basis of meaning, imagination,
and reason. Chicago: University of Chicago Press.
KeImer, D. and Harker, L. (1998). The forms and functions of nonverbal signals of
shame. In P. Gilbert and B. Andrews (eds.), Shame: inter personal behavio1; psy-
chopathology, and culture. New York: Oxford University Press, pp. 78-98.
Kövecses, Z. (1995). Metaphor and folk understanding of anger. In J. A. Russel,
J.-M. Femandez-Dols, A. S. R. Manstead andJ. C. Wellenkamp (eds.), Everyday
conceptions of emotions: an introduction to the psychology, anthropology and
linguistics of emotion, Series D: Behavioral and social sciences 81. London: Kluwer
Acadernic, pp. 49-71.
(1998). Are there any emotion-specific metaphors? In A. Athanasiadou and
E. Tabakowska (eds.), Speaking of emotions: conceptualization and expression.
Berlin: Mouton de Gruyter, pp. 127-51.
Lakoff, G. and Kövecses, Z. (1987). The cognitive model of anger inherent in American
English. In D. Holland and N. Quinn (eds.), Cultural models in language and
thought. Cambridge: Cambridge University Press, pp. 196-221.
Lambek, M. and Solway, J. S. (2001). Just anger: scenarios of indignation in Botswana
and Madagascar. Ethnos, 66(1), 49-72.
Lazarus, R. S. (1991). Emotion and adaptation. Oxford: Oxford University Press.
Leary, M. R., Britt, T. W., Culip 11, W. D. and Tempelton, J. L. (1992). Social blushing.
Psychological Bulletin, 112(3), 446-60.
Leavitt, J. (1996). Meaning and feeling in the anthropology of emotions. American
Ethnologist, 23(3), 514-39.
Shame across cultures 295
Levenson, R. W., Ekman, P., Heider, K. and Friesen, W. V. (1992). Emotion and au-
tonomic nervous system activity in the Minankabau of West Sumatra. Journal of
Personality and Social Psychology, 62(6), 972-88.
Levy, R. I., (1973). Tahitians: mind and experience in the Society Islands. Chicago:
University of Chicago Press.
(1978). Tahitian gentleness and redundant controls. In A. Montagu (ed.), Learning
non-aggression. the experience ofnon-literate societies. Oxford: Oxford University
Press, pp. 222-35.
(1984). Emotion, knowing, and culture. In R. A. Shweder and R. A. LeVine (eds.),
Culture theory: essays on mind, self, and emotion. Cambridge: Cambridge Univer-
sity Press, pp. 214-37.
Lewis, M. (1992). Shame: the exposedself New York: Maxwell Macmillan Press.
(1995). Embarrassment: the emotion of self-exposure and evaluation. In J. P.
Tangney and K. W. Fischer (eds.), Self-conscious emotions: the psychology of
shame, guilt, embarrassment, and pride. New York: Guilford Press, pp. 198-218.
Lindholm, C. '( 1988). The social structure of emotional constraints: the court of Louis
XIV and the Puk9tun ofNorth Palcistan. Ethos, 16,227-46.
Lindisfarne, N. (1998). Gender, shame, and culture: an anthropological perspective. In
P. Gilbert and B. Andrews (eds.), Shame: interpersonal behaviol; psychopathology,
and culture. New York: Oxford University Press, pp. 246-60.
Lutz, C. (1988). Unnatural emotions: everyday sentiments on a Micronesian atoll and
their challenge to Western theory. Chicago: University of Chicago Press.
Lyon, M. L. (1995). Missing emotions: the limitations of cultural constructionism in the
study of emotions. Cultural Anthropology, 10(2), 244-63.
(1998). The limitations of cultural constructionism in the study of emotions. In
G. Bendelow and S. J. Williams (eds.), Emotions in sociallife. London: Rout-
ledge, pp. 39-59.
Lyon, M. and Barbalet, J. M. (1994). Society's body: emotion and the 'somatization'
of social theory. In T. J. Csardas (ed.), Embodiment and experience: the existen-
tial ground ofculture and self Cambridge: Cambridge University Press, pp. 48-
Magai, C., Distel, N. and Liker, R. (1995). Emotion socialization, attachment, and pat-
terns of adult emotional traits. Cognition and Emotion, 9(5),461-81.
Mariauzouls, C. (1996). Psychophysiologie von Scham und Erröten (The psychology
of shame and blushing), Ph.D. thesis, University of Munich.
Mascolo, M. F. and Fischer, K. W. (1995). Developmental transformations in appraisals
forpride, shame and guilt. In J. P. Tangney and K. W. Fischer (eds.), Self-conscious
emotions: the psychology of shame, guilt, embarrassment and pride. New York:
Guilford Press, pp. 64-113.
Mayer, E. (1974). Behavior programs and evolutionary strategies. American Scientist,
McGuire, M. T. and Troisi, A. (1990). Anger: an evolutionary view. In R. Plutchik and
H. Kellermann (eds.), Emotion: theory, research, and experience. San Diego:
Academic Press, pp. 43-57.
Menon, U. and Shweder, R. A. (1994). Kali's tongue: cultural psychology and the
power of shame in Orissa, India. In S. Kitayama and H. Markus (eds.), Emotion
and culture. Washington, DC: American Psychological Association, pp. 241-82.
296 Michael J. Casimjr and Michael Schnegg
Mesquita, B. and Frijda, N. H. (1992). Cu1tural variations in emotions: a review.
Psychological Bulletin, 112(2), 179-204.
Mikolajczuk, A. (1998). The metonymic and metaphorica1 conceptualization of anger
in Po1ish. A. In Athanasiadou and E. Tabakowska (eds.), Speaking of emo-
tions: conceptualization and expression. Berlin: Mouton de Gruyter, pp. 153-
Morton, J. (1995). The organic remains: remarks on the construction and development
of people. Social Analysis, 37, 101-18.
Öhman, A. (1986). Face the beast and fearthe face: anima1 and social fears as prototypes
für evolutionary analyses of emotions. Psychophysiology, 23(2), 123-45.
Okano, K. (1994). Shame and social phobia: a transcultural viewpoint, Bulletin ofthe
Manninger Clinic, 85(3) 323-38.
Planalp, S. (1999). Communicating emotion: social, moral, and cultural processes.
Cambridge: Cambridge University Press.
Retzinger, S. R. (1987). Resentments of1aughter: video studies ofthe shame-rage spiral.
In L. Lewis (ed.), The role of shame in symptom formation. Hillsdale, NJ: Lawrence
Er1baum, pp. 151-81.
Roberts, J. R. and Weerts, T. C. (1982). Cardiovascular responding during anger and
fear imagery. Psychological Reports, 50, 219-30.
Rosaldo, M. Z. (1983). The shame of head hunters and the autonomy of gelt. Ethos,
(1984). Towards an anthropology of self and feeling. In R. A. Shweder and R. A.
LeVine (eds.), Culture theory: essays on mind, seit and emotion. Cambridge:
Cambridge University Press, pp. 135-57.
Rosenberg, C. (1991). Understanding shame. London: Jason Aronson.
RusselI, J. (1989). Culture, scripts, and children's understanding of emotions. In
C. Saarni and P. L. Harris (eds.), Childrcn 's understanding of emotion. Cambridge:
Cambridge University Press, pp. 293-318.
Russen, J. and Yik, S. M. (1996). Emotion among the Chinese. In M. H. Bond (ed.),
The handbook of Chinesc psychology. Hong Kong: Oxford University Press,
Saarni, C. (1991). Children's understanding of strategic control of emotional expression
in social transactions. In C. Saarni and P. L. Harris (eds.), Children 's understanding
of emotion. Cambridge: Cambridge University Press, pp. 181-208.
Sachdev, P. S. (1990). Whakama: culturally determined behavior in the New Zealand
Maori. Psychological Medicine, 20, 433-44.
Schachter, S. and Singer, J. E. (1962). Cognitive, social and physiological determinants
of emotional state. Psychological Review, 69(5), 379-99.
Scherf, T. J. (1990). Socialisation of emotions: pride and shame as causal agents. In
T. D. Kemper (ed.), Research agendas in the sociology of emotions. Albany: State
University ofNew York Press, pp. 281-304.
Scherf, T. J. and Retzinger, S. M. (1991). Emotions and violence: shame and rage in
destructive conflicts. Lexington, MA: DC Heath.
Scherer, K. R. (1993a). Studying the emotion-antecedent appraisal process: an expert
system approach. Cognition and Emotion, 7,325-55.
(1993b). Les emotions: fonctions et composantes. In B. Rime and K. Scherer (eds.),
Les emotions. Neuchate1: Delachaux et Nie~tl~ nn q7-111
. (1997). Profiles of emotion-antecedent appraisa1: testing theoretical predictions across
cultures. Cognition and Emotion, 11 (2), 113-50.
Scherer, K. R. and Wallbott, H. G. (1994). Evidence für universality and cultural varia-
tion of differential emotion response patteming. Journal of Personality and Social
Psychology, 66, 310-28.
Schieffelin, E. L. (1983). Anger and shame in the tropical forest: on affect as a cultural
system in Papua New Guinea. Ethos, 11(3), 181-91.
Schüre, A. N. (1998). Early shame experiences and infant brain development. In
P. Gilbert and B. Andrews (eds.), Shame: Interpersonal behavior; psychopathology,
and culture. New York: Oxford University Press, pp. 57-77.
Sharpe, J. (1987). 'Shame' in Papua New Guinea. GroupAnalysis, 20,43-8.
Sheam, D. (1992). Blushing as a function of audience size. Psychophysiology, 29(4),
Sheam, D., Bergman, E., HilI, K., Abel, A. and Hindis, L. (1990). Facial coloration and
temperature responses in blushing. Psychophysiology, 27(6), 687-93.
Shweder, R. A. (1996). Thinking through cultures. expeditions in culturalpsychology.
Cambridge, MA: Harvard University Press.
Smiley,P. and Huttenlocher, J. (1991). Young children's acquisition ofemotion concepts.
In C. Saami and P. L. Harns (eds.), Children 's understanding of emotion.
Cambridge: Cambridge University Press, pp. 27-49.
Stein, D. J. and Bouwer, C. (1997). Blushing and social phobia: a neuroethological
speculation. Medioal Hypotheses, 49(}), 101-8.
Strauss, C. (1992). Models and motives. In R. D' Andrade and C. Strauss (eds.),
Human motives and cultural models. Cambridge: Cambridge University Press,
Taylor, J. R. and Mbense, T. G. (1998). Red dogs and rotten mealies: how.Zulus talk
about anger. In A. Athanasiadou and E. Tabakowska (eds.), Speaking of emotions:
conceptualization and expression. Berlin: Mouton de Gruyter, pp. 191-226.
Wellman, H. M., Harris, P. L., Baberjee, M. and Sinclair, A. (1995). Early understanding
of emotions: evidence from language. Cognition and Emotion, 9(2), 117-49.
Wierzbicka, A. (1986). Human emotions: universal or culture-specific? American
Anthropologist, 88(3), 584-94.
(1988). The semantics of emotions:fear and its relatives in English. Australian Journal
of Linguistics, 10(3),359-75.
(1992). Talking about emotions: semantics, culture, and cognition. Cognition and
Emotion, 6(3/4), 285-319.
(1999). Emotions across languages and cultures: diversityand universals. Cambridge:
Cambridge University Press.
Williams, S. (2001). Emotion andsocial theory: corporeal refiections on the (ir)rational.
Winzeler, R. (1990). Amok: historical, psychological and cultural perspectives. In
W. J. Karim (ed.), Emotion of culture: a Malay perspective. Oxford: Oxford
University Press, pp. 97-122.
Zahn- Waxler, C. and Robinson, J. (1995). Empathy and guilt: early origins of feelings of
responsibility. In J. P. Tangney and K. Fischer (eds.), Self-conscious emotions: the
psychology of shame, guilt, embarrassment, and pride. New York: Guilford Press,
298 Michael J. Casimir and Michael Schnegg
These follows a list of languages referred to in this paper and the countries
in which they are spaken. Bold type indicates the absence of rnetaphors and
rnetonyrns linking facial colours (or colour change) and emotions, while the
numbers give the sampIe of informants für the given language.
Guabibo I Sukuani
Gur I Kabre
Italian / Venetian
Lamutic I Ewenic
Italy / Venice
Philippines I Luzon
China I Peking
Yugoslavia I Croatia
Arabic I Egyptian
Arabic I Palestinian
Arabic / Tunisian
Ashanti I Twi
Boran I GaUa
Chinese I Taiwan
Eipo I Mek
English I British
English I American
W. India (Barrner)
W. India (Jodhpur)
N .E. India
Lamutic I Ewenic
Naga I Angami
Persian I Dari
Persian I Farsi
Tibet I China
lndonesia I Romeo
Russia I Yacutia
Examples of metaphors and metonyms indicating shameful feelings in same
Armenian: amantchouchounits karmrav
Literal from shame become red
Translation he/she became red with shame
Bahasa lndonesia: karena dia malu mukanya
Literal: because he is ashamed his face
Translation: his face is red because he is ashamed
Balinese: gobane barak ulian lek
Literal: face red because shame
Translation: his face tumed red because of shame
300 Michael J. Casimir and Michael Schnegg
Bengali: notun barer shaathe chokh miliye bauer mukh lojjaae laal hoe gaelo
Litera!: new groom's with eyes having met bride's face shame red became
Translation: the bride blushed when her eyes met those of her new groom
Bulgarian: potscherwenjawam ot sram
Literal: becoming red with shame
Translation: he/she tumed red with shame
Chinese: ta bei shangsi dang zhong piping jiu lian hang se
Literal: he is head/boss in public criticized therefore face red colour
Translation: because the head/boss was criticized in public bis face tumed red
Croatian: pocrvenelo mu je lice od stida
Literal: becomes red hirn is face with shame
Translation: his face became red with shame
Greenlandic: kanngutsikkami aapillerpoq
Literal: shame red
Translation: he/she tumed red because of shame
Hungarian: v(Jrös, mint a paprika
Literal: red as the paprika
Translation: he/she is red as a paprika (from shame or rage)
Kashmiri: washlyok sa vits watse tember
Literal: shame he with became copper-coloured
Translation: his face became copper-coloured with shame
Toba: napagaaqtec cha' aye ncoq
Literal: he/she blushes because he/she is ashamed
Translation: he blushes because he is ashamed
Toba-Batak: marrara bobina ala tarboto ibana manangko
Literal: becoming red his face because recognized he a thief
Translation: he blushed when he was recognized as a thief
Thrkish: utanmaktan kipkirmizi oldu
Litera!: shame from very red has became
Translation: he/she tumed red because of shame
Yakutic: kybystan sireje kytarda
Literal: having been ashamed face becarne red
Translation: his face became red because he has been ashamed