Abstract
Questions about the relation between shame and embarrassment are often posed in discussion of emotion but have rarely been examined at length. In this study I assemble and examine distinctions that have been proposed in the literature with the aim of identifying the criteria that have been used to differentiate shame and embarrassment. Relevant empirical studies are also reviewed. Despite the attention paid to the question of the difference between shame and embarrassment consensus on differentiating criteria has not been reached nor has there been consideration of what kind of question is being posed. Three positions that have been adopted are identified and critically evaluated.
Are shame and embarrassment variants of the same emotion or do they represent distinct emotions? For some theorists the position is clear-cut: “Even though embarrassment and shame reflect negative self-evaluation . . . there is evidence that they are distinct and different self-conscious emotions” (M. Lewis & Ramsay, 2002, p. 1034). Keltner and Buswell (1997) reviewed the extant literature in terms of properties that are thought to differentiate emotions—antecedents, appraisal processes, subjective experience, facial expressions, and physiological responses—and concluded that evidence supported the distinctness hypothesis. A variant of this position emphasises the essentially moral nature of shame in contrast to embarrassment (Keltner & Buswell, 1997; Taylor, 1985; Wierzbicka, 1999). An alternative position is that they represent different locations along a continuum (Probyn, 2005). Embarrassment is a less intense form of shame (see Keltner & Buswell, 1997, for a critical review). Harré (1990) extends this to a two-dimensional “semantic space” of self-conscious emotion, where the axes represent, respectively, variation in the seriousness of a breach of convention and in the degree of the individual’s fault for the breach. Shame occurs when you are at fault for a serious breach, embarrassment if you are less at fault and the breach is more trivial. The aim of this study is to assemble and assess claims about the distinction between shame and embarrassment, to consider their relations to the distinctness, moral basis, and continuum positions, and to reflect on this debate in the light of concerns about the “crisis” in emotion research (Dixon, 2012).
An important issue here—as it is more generally in the study of emotion—is whether this question is essentially a matter of the language of emotion: We cannot conclude that because we have different words to name emotions there are distinct emotional states that correspond to these names (Bedford, 1957). Understanding the relations between emotion lexicon and emotional states remains hotly contested (for example, Izard, 2010; Scherer, 2005; Wierzbicka, 2009). Is the issue here one of identifying distinctive profiles of cognitive, neurophysiological, motivational, motor expressive, or subjective feeling components (Scherer, 2005) or of clarifying the circumstances in which particular words are used?
Historical and cross-cultural studies show variation across time and place in the meanings of embarrassment, implying that relations between the terms are neither constant nor universal. The meaning of shame seems to have remained consistent over time. Aristotle (2009, p. 106) defined it as a passion, “a kind of fear of dishonour” and Samuel Johnson’s Dictionary of the English Language (1799, p. 275), perhaps the first to comprehensively cover the language used in everyday discourse, defined it as “the passion felt when reputation is supposed to be lost.” Johnson would have been comfortable with current dictionary definitions, for example Chambers Dictionary (Allen & Schwarz, 1998, p. 1519) gives it as “the humiliating feeling of having appeared unfavourable in one’s own eyes, or those of others, by shortcoming, offence, or unseemly exposure, or a similar feeling on behalf of anything one associates with oneself.” On the other hand, embarrassment in its contemporary usage has a much shorter history. The Oxford Dictionary of Word Histories (Chantrell, 2004) traces the word to the early 17th century, meaning to hamper or impede. In the same century “to be embarrassed” is to encounter difficulty, for example in the context of financial problems. Johnson’s Dictionary defined it in terms of perplexity and entanglement. The Oxford English Dictionary includes definitions in terms of perplexity, confusion of thought and expression, examples of their use dating from the 18th century, and “constrained feeling or manner arising from bashfulness or timidity,” (Simpson & Weiner, 1989, p. 158) dating from the 19th century.
Variation is evident in languages other than English. Castelfranchi and Poggi (1990) identify two Italian terms: Imbarazzo refers to uncertainty or conflict about what action to take in a social situation and implies no personal shortcoming on the part of the embarrassed person; vergogna refers to embarrassment that does entail some personal shortcoming and also to shame. The Welsh language (a member of the Brythonic branch of the Celtic languages) has distinct words for shame (gwarth, also meaning reproach and disgrace) and embarrassment (chwithig) as well as words that cover both shame and embarrassment (for example, cywilydd). Cross-cultural studies identify languages that have the same word for both shame and embarrassment (Fessler, 2004; Haidt & Keltner, 1999). In short, research undertaken in different eras or societies might well make different distinctions between shame and embarrassment.
The thesis that experience can be labelled as either shame or embarrassment depending on contextual factors (Sabini, Siepmann, Stein, & Meyerowitz, 2000; Sabini & Silver, 1997, 2005) suggests a possible bridge between different positions. A psychological state is labelled shame if the evidence at hand suggests a flaw in character; it is labelled embarrassment if a flaw in the character being enacted is revealed. But how can we be sure that the nature of a flaw is the essential criterion? Might other criteria be used to differentiate shame from embarrassment? What alternative criteria are advanced in the literature and is there consensus on these? My aim in this article is to collect and review a sample of differentiating criteria that have been proposed in the literature. I review assertions about differences and similarities between them and also draw upon empirical studies reporting relevant findings. There is no dearth of studies of shame and embarrassment but there has been little systematic research into differences between them; research has tended to rely on participants’ recollections of experiences and sorting or ratings of emotion-related vignettes, with the limitation that these are inevitably influenced by folk theories or stereotypes of the meaning of shame and embarrassment. Less attention has been paid to cognitive, psychophysiological, expressive, or behavioural measures.
Claimed Differences between Shame and Embarrassment
What distinctions do emotion researchers make? I approach this question through examination of a sample of 33 publications that have made explicit claims about differences and similarities (these sources are marked with an asterisk in the references list). I cannot claim that this sample is exhaustive; the aim is to identify themes that emerge in a sample of influential sources. The publications were read carefully and any references to similarities and differences between embarrassment and shame were identified and transcribed into a table. The entries in the table were classified in an attempt to identify recurrent themes. Inspection of the texts suggests that the distinctions can be grouped as presented here.
Intensity
Shame is more intense (Buss, 1980; Ho, Fu, & Ng, 2004), devastating (Babcock, 1988), can be extraordinarily intense (Scheff, 1994), a weightier and more shattering emotion (Taylor, 1985). Embarrassment is less shattering and painful (Zahavi, 2010), a mild version of shame (Rochat, 2009; Scheff, 1994).
Duration
Shame is more persistent (Buss, 1980, 2001), protracted (Scheff, 1994) and long-lasting (Miller, 1996), producing lasting damage to self-esteem (Heywood, 2002) and social identity (Zahavi, 2010). Embarrassment is briefer (Scheff, 1994), more transient (Buss, 1980), temporary (Buss, 2001; Kristjánsson, 2010), short-lived (Zahavi, 2010) and fleeting (Ho et al., 2004); it dissipates faster (Buss, 2001) and can disappear by itself (Ho et al., 2004).
Nature of the Involvement of the Self
This has been construed in various ways, emphasising the seriousness of the flaw, the involvement of “core” aspects of the self, and the impact upon the self. Shame involves a flaw revealing of self and character (Babcock, 1988; M. Lewis, 1995, 2008; Taylor, 1985); the self is flawed, vile, revolting, disgusting (Sabini & Silver, 1997) and humiliated (Buss, 2001). In contrast, in embarrassment there is a flaw in the character being enacted (Sabini & Silver, 1997), in one’s persona (Babcock, 1988) or presented self (Tracy & Robins, 2004); an apparent rather than a real flaw is involved (Sabini, Garvey, & Hall, 2001).
Shame reaches to the core or centre of the self (Ho et al., 2004; Keltner & Buswell, 1997) whereas embarrassment reaches only the outer self, the public self (Tracy & Robins, 2004) and not the private inner self (Ho et al., 2004). Embarrassment entails awareness of a discrepancy between public aspects of the self and other people’s evaluations (Tracy & Robins, 2004).
Shame involves negative self-evaluation (M. Lewis, 1995, 2008) or failure to live up to some ideal self-goal (Kristjánsson, 2010) whereas embarrassment is more a matter of social evaluation than of self-evaluation (Miller, 1996). Shame is associated with a decrease in self-esteem whereas embarrassment is not (Zahavi, 2010); it has a major impact on self-esteem whereas embarrassment has minor impact (Babcock, 1988; Buss, 2001). In shame, the self feels small and shrinking (Fessler, 2004). Shame is more likely to be associated with regret and depression (Buss, 1980) and to lead to sadness, regret, self-disgust, self-contempt, and separation anxiety (Babcock, 1988; Buss, 2001; Prinz, 2004).
Nature of a Breach or Violation
The notion that shame is associated with more serious breaches is conceptualised in a number of ways. In shame there has been a breach of a fundamental standard of conduct (Buss, 1980), a relatively important social standard (Fessler, 2004), more important, central standards, rules and goals (M. Lewis, 1995, 2008), a more serious breach (Harré, 1990), serious failures and moral transgressions (Tangney, 2003) and violation of moral rules (Keltner & Buswell, 1997). It responds to “fundamental personal failures and transgressions that harm others” (Tangney, Miller, Flicker, & Barlow, 1996, p. 1267). According to Bedford (1957) it is a necessary condition for shame that the individual is at fault and to blame. In contrast, embarrassment is said to follow a minor breach of codes of manners or loss of poise (Buss, 1980, 2001; Harré, 1990), faux pas and social transgressions (Tangney, 2003), violations of social conventions (Fessler, 2004; Harré, 1990; Keltner & Buswell, 1997; Keltner, Young, & Buswell, 1997; M. Lewis, 1995, 2008; Tangney et al., 1996) and minor, specific breaches of norms (Kristjánsson, 2010). Shame involves inconsistency with a universal ideal standard, embarrassment with a personal standard and a relatively trivial breach (Babcock, 1988; Babcock & Sabini, 1990).
Moral Basis
Shame is connected with personal morality (Buss, 2001; Taylor, 1985), follows acts that lead to assessment of a defective moral character (Harré, 1990), implies defects in morals or character (Miller, 1996), links social and moral concerns (Wierzbicka, 1999). In contrast, embarrassment has no moral basis (Buss, 2001), is less centrally relevant to regulation of behaviour in the moral domain (Tangney, 2003) and social and moral concerns are disassociated (Wierzbicka, 1999). Shame motivates “conformity to the most moralized cultural standards, and embarrassment motivate[s] conformity to many cultural rules that hold less moral import” (Fessler, 2007, p. 181). Indeed, embarrassment can result from public attention even when there has been no wrongdoing (Fessler, 2004). In shame, reparation is more difficult (Ho et al., 2004) or is little emphasised (Fessler, 2004). Ho et al. (2004) suggest that shame encompasses both moral and nonmoral concerns. In short, there is no consensus that the moral basis is a distinguishing criterion.
Uncertainty How to Act, Confusion, Fluster
Absence of agreement is apparent here too. Embarrassment has been characterised by uncertainty how to act when action is demanded (Taylor, 1985), is related to awkward social exposure (Zahavi, 2010) and is regarded as likely to be accompanied by feeling foolish (Buss, 1980, 2001). On the other hand there are arguments that mental confusion, disruption of behaviour, and uncertainty how to act accompany shame (Babcock, 1988; H. B. Lewis, 1971).
Situation-Specificity and the Necessity of an Audience
Within the literature examined here these are emphasised in embarrassment which is regarded as a response to a specific situation (Bedford, 1957; Taylor, 1985), a situational problem in managing face (Katz, 1997) and is context-bound (Zahavi, 2010). It implies an embodied audience (Taylor, 1985) or actual or suspected observation by a significant other (Heywood, 2002). It is more likely with a larger audience (Keltner & Buswell, 1997) and less likely to occur while alone (Ho et al., 2004; Keltner & Buswell, 1997; Tangney et al., 1996). It is triggered by conspicuousness (Buss, 2001), public attention (Fessler, 2004), by being the object of others’ thinking (Dineen, 1990) and by what others see of the actor rather than what they know about him or her (Wierzbicka, 1999). Heywood (2002) suggests that embarrassment occurs when you know or suspect that a significant other is aware of your failure to meet obligations whereas shame occurs when you are aware of your failure to meet obligations. It is associated with accidents and mistakes (Miller, 1996) and strikes quickly (Zahavi, 2010). It can be triggered by overpraise and teasing (Buss, 2001). The reactions of others in embarrassment are acceptance and laughter whereas in shame they are scorn and rejection (Buss, 2001). Nevertheless, shame is also triggered by specific social circumstances where dignity and reputation are threatened (H. B. Lewis, 1971; Scheff, 1994), the relations among laughter, shame, and embarrassment are complex (H. B. Lewis, 1971; Retzinger, 1987), and the presence of an audience for embarrassment is thought to be unnecessary (Miller, 1996, pp. 38–41).
Physiological Changes
Embarrassment is accompanied by blushing (Buss, 2001; Fessler, 2004; Ho et al., 2004; Prinz, 2004), and greater awareness of physiological changes (Buss, 2001). Yet H. B. Lewis (1971) and Retzinger (1987) argue that blushing and sweating and other autonomic responses accompany shame; Heywood (2002) suggests that both shame and embarrassment can immediately result in physiological discomfort and Sabini and Silver (2005) propose that the physiological state is the same in both.
Display of Emotion
In addition to blushing, embarrassment is accompanied by smiling or a “silly” grin (Buss, 1980; M. Lewis, 1995, 2008), often by giggling and laughter (Buss, 2001) and by a pattern of tilted head, looking at the other and looking away, and nervous self-touching (M. Lewis, 1995, 2008). Shame is often apparent in a stricken look (Buss, 2001), by bowing the head, shrinking, lowering or averting one’s gaze, and glancing furtively at the other (H. B. Lewis, 1987; Retzinger, 1987; Scheff, 1994), by a desire for concealment, covering part of the face and false smiling (Retzinger, 1987). Laughter is not restricted to embarrassment and is also found in shame, where it serves a number of functions including reducing shame, breaking the shame–rage spiral, and enhancing social connectedness (H. B. Lewis, 1987; Retzinger, 1987).
Other
Embarrassment can be contagious (Taylor, 1985). It is often viewed by the self and/or others as amusing at the time or in retrospect (Babcock, 1988; Fessler, 2004). It is less cognition-dependent and cognitively simpler and less elaborated than shame (Tracy & Robins, 2004). The relation with self-directed hostility varies quantitatively and qualitatively between shame and embarrassment (H. B. Lewis, 1971).
Empirical Evidence
It is apparent that many differences are claimed in the literature sampled here. To what extent are these differences identified in empirical studies? One source of evidence questions participants about the meanings of shame and embarrassment by asking them to recall and rate instances of these (Keltner & Buswell, 1996; Tangney et al., 1996) or to sort descriptive statements into piles that (they believe) characterise these emotions (Miller & Tangney, 1994). These studies report quantitative differences between the two emotions on a large number of dimensions that are consistent with the criteria summarised here. For example, embarrassment has higher ratings on the unanticipated nature of the trigger event, the situation was funny, and others were amused whereas shame has higher ratings on duration and violation of moral standard (Tangney et al., 1996). Embarrassment is more likely in circumstances of physical pratfalls, loss of script, cognitive shortcomings, loss of control over body, shortcomings in physical appearance, failure of privacy, conspicuousness, and being teased (Keltner & Buswell, 1996). Antecedents of shame include hurting others emotionally, failure to meet expectations, poor performance (e.g., academic), disappointment in oneself, and role-inappropriate behaviour (Keltner & Buswell, 1996). There is more mention of humour, smiling, accidents, and short duration in the case of embarrassment; more feeling one was a bad person, immoral, long-lasting, a deep-seated flaw in shame (Miller & Tangney, 1994).
How large are the differences? Tangney et al. (1996) report that 22 of 31 comparisons on the rating scales were significant but if we adopt a more stringent criterion, considering for example only mean ratings where the means for shame and embarrassment fall on opposite sides of the mid-point of the 5-point scale, this yields five items referring to “situation serious/funny,” “hurts now/can laugh about it now,” “emotion lasted a long time,” “felt violation of moral standard,” “others angry/indifferent.” On the other hand, Keltner and Buswell (1996) report only a modest degree of overlap (9%) between shame and embarrassment (computed as the number of times the same antecedent is listed for each emotion). Other research suggests that ratings responses are susceptible to changes in the nature of the task. Harris (2003) applied factor analysis to ratings of items relating to shame, guilt, and embarrassment made by motorists who recalled an actual event within a specific context—appearing in court or attending a conference to discuss their offence during legal proceedings ensuing from a drink-driving offence—and obtained results at variance with studies that draw upon more abstract or hypothetical vignettes.
Another strand of research examines the recognition of facial displays of emotion (Haidt & Keltner, 1999; Keltner, 1995; Keltner, Young, & Buswell, 1997). There is evidence for a prototypical embarrassment display (Keltner, 1995), a sequential pattern of gaze aversion, smiling, and face touching that is apparent in young children from about 18 months of age. Keltner et al. (1997) found that following brief exposure to faces expressing embarrassment or shame, participants tended to infer that the expression of embarrassment was in response to violations of conventions such as tripping in public whereas the shame display was elicited by more serious transgressions. Keltner et al. (1997) found that in responses to still photographs the embarrassment display evoked more amusement in others than did the shame display after less serious transgression while the shame display evoked higher levels of sympathy after both less serious and more serious transgressions (the study did not control for a face displaying sadness). Haidt and Keltner (1999) compared recognition of photographs of facial displays of embarrassment and shame in American and Indian samples (the language of the Indian participants has a single word, lajya, that covers both emotions). Recognition rates for the embarrassment display were high in both samples whereas rates for shame were substantially lower in both (consistent with findings generally when participants assign their own emotion labels as opposed to choosing among experimenter-supplied labels; Crozier, 1981; Widen, Christy, Hewett, & Russell, 2011). Among the Indian sample, lajya was the most common response to the embarrassed face whereas the word for sadness was the most common response to the shame display. Analysis of the kinds of antecedents of displays nominated by participants showed that both shame and embarrassment were perceived to be triggered by self-conscious situations (defined as “events in which the self’s behaviour or interactions are improper or awkward,”; Haidt & Keltner, 1999, p. 247), although shame was more likely to be triggered by “events with negative implications for self’s goals, attachments, hedonics” (p. 247); this pattern held for both samples. Shame was also associated with negative emotions—sadness, worry, and fear—whereas embarrassment had associations with the positive emotions happiness, love, and amusement (perhaps influenced by the non-Duchene smile portrayed in the embarrassment display). Thus shame and embarrassment are perceived as distinct in terms of antecedents and displays even though the Indian participants’ language had a single name for both. The kinds of antecedents nominated for the display of embarrassment are consistent with those in the English-speaking sample and in other empirical research.
The final strand of research reviewed by Keltner and Buswell (1997) involves physiological responses and focuses on blushing, which is often regarded as the “hallmark of embarrassment” (Buss, 1980) and is considered to be specific to embarrassment (Miller, 1996). Yet blushing is associated with shame in theoretical accounts (Castelfranchi & Poggi, 1990; H. B. Lewis, 1971; Tomkins, 1963); historical analysis of descriptions of blushing, shame, and embarrassment (Crozier, 2012); and cross-cultural studies of associations between colour words and emotion (Casimir & Schnegg, 2002). Current research emphasises its value as a signal of appeasement or nonverbal form of apology and shows that people feel better disposed towards someone who has committed a transgression if that person is seen to blush (de Jong & Dijk, 2013). Yet communicating acknowledgement of wrongdoing and appeasement is surely valuable whether we are ashamed or embarrassed.
What Shame and Embarrassment Share
It is useful to provide a brief summary of the themes that emerge when the sample of texts is analysed for what is common to shame and embarrassment. Unsurprisingly, given that both are regarded as self-conscious emotions, there are frequent references to self-conscious emotion, self-consciousness, or self-awareness (mentioned in 15 of the 33 sources). There are also frequent references to evaluation, standards, norms, and flaws (20 sources); to feelings of distress, inadequacy, pain, and inhibition (14); to being conspicuous, the object of public attention, the loss of reputation or dignity, or being viewed or evaluated by others (15); and to being exposed and experiencing feelings of exposure (7). In summary, a consensual view might be that both shame and embarrassment are painful states, where the self is focal in attention, the individual believes she or he has failed to meet appropriate standards of conduct, and is seen to have done so in the eyes of others.
Discussion
There is no dearth of claims about the nature of differences between shame and embarrassment but little relevant empirical literature is available and no consensus has been reached on the issue. I have proposed one organisation of recurrent themes, acknowledging that there is considerable variation within these themes. Does this analysis throw light on the three positions identified in the introduction: that they occupy different locations on a continuum, they are distinct emotions, and the distinction relates specifically to the essentially moral nature of shame? Each of the positions has its advocates but all present difficulties.
Variation on a Continuum
This thesis is most frequently asserted, not only in terms of intensity and duration but also in terms of the seriousness of faults and flaws and the nature of failures to meet standards: Moral dereliction is more serious than a breach of manners, disgrace more serious than forgetting your lines. Yet a convincing case has not been made for intensity or seriousness. There exists little empirical evidence for variation in intensity as a defining characteristic; Tangney et al. (1996) reported that significant differences in ratings of shame and embarrassment remain after variation in intensity ratings is statistically taken into account. Variation in seriousness has problems too. Sabini and his colleagues (Sabini et al., 2000; Sabini & Silver, 1997, 2005) focus on the relative seriousness of flaws but do not explain why you might be ashamed of yourself because, say, the state of your appearance is due to the poverty of your family rather than a flaw in character or why you should be embarrassed about failings and accidents that are suffered by others or by praise and positive attention. Finally, it offers no justification for assuming that exposure of flaws should be the defining characteristic.
One problem with the dimensional thesis is where to draw the line: How intense, persistent, or serious does an experience have to be in order to be described as shame rather than embarrassment? Moreover, each emotion varies along these dimensions. Embarrassment need not always be mild or short-lived; it can be intense in terms of its feelings, its effects on behaviour, and the seriousness of its consequences. Nevertheless despite difficulties in settling upon specific dimensions, this explanation cannot be ruled out. It is consistent with lay accounts of the distinction and with findings from rating studies notwithstanding their limitations. Where it falls short is in its focus on specific, discrete, additive dimensions. Research to date suggests that this has proved unproductive.
Distinct and Different
Assertions of distinct emotions have appealed to the display of embarrassment, particularly the dynamic pattern involving smile control and smiling, the association between embarrassment and humour, and the presence of dramaturgical uncertainty. Yet here too there are counterclaims that these characteristics are not unique to embarrassment, and smiling, humour, mental confusion, and failure of effective functioning are also associated with shame (H. B. Lewis, 1971; Retzinger, 1987). Empirical evidence has identified a patterned display associated with embarrassment but is less convincing on whether the display is unique to it. Similarly arguments that the blush is peculiar to embarrassment are met with counterarguments and do not seem compatible with findings from cross-cultural research (Casimir & Schnegg, 2002; Strathern, 1977). In the absence of explicit criteria for distinguishing among emotions it is difficult to know what weight to attach to these claims.
Moral Basis
The thesis that shame is connected with morality whereas embarrassment has no necessary moral basis also faces difficulties. You can be ashamed of things about yourself that are not necessarily to do with morality and which may be beyond your control: your name or nickname; your affiliations; or the actions of your affiliates, forebears or compatriots. You can be ashamed of your social status, family background, accent, clothes, or distinctive physical characteristics; of being hospitalised. As Fessler (2007) has argued, you can be ashamed simply by occupying a subordinate position in a social hierarchy. You can also surely be embarrassed about all these things. Goffman (1956) considered that embarrassment and morality are connected, that embarrassment results from failure to meet moral expectations, specifically those confronting the individual as a participant in social encounters. Embarrassment can be about moral issues, shame not necessarily so.
Conclusions
In summary, no consensus has been reached on how shame and embarrassment differ. One might be disappointed that these efforts have not produced a more consistent and coherent picture of relations between two emotions that play important roles in scientific accounts of social regulation and self-regulation and of psychological conditions such as depression and social anxiety disorder. One cannot envisage a comprehensive English language account of emotional life which neglects the concepts of shame and embarrassment and it is worthwhile asking whether this is also the case for other languages and cultures. Intuitively, for English speakers at least, the two seem to be neighbours or members of the same family, one of which is somehow psychologically more challenging than the other, and they are both associated with the blush. We feel that the nature of the differences, although elusive, ought to be within our grasp. Will more research along similar lines finally bring order to this diversity of positions or do the difficulties reside in the approaches themselves, particularly those starting with English language labels for vignettes or emotion displays or which provide participants with these labels or translations as names of response categories?
One alternative approach draws upon the methodology of metalanguage (Wierzbicka, 1999). Dineen (1990) has applied natural semantic metalanguage (NSM) to shame and embarrassment, the difference she emphasises being that shame refers to “the kind of bad thing that happens to people/when they know that they have done something bad/and they do not want other people to know it” (p. 221) whereas embarrassment is characterised as “the kind of bad thing that happens to people/when they know that other people are thinking of them/and they do not want this to go on after now” (p. 220). These are not proposed definitions of the two emotion words but are characterisations of states that are called by these names in English and Danish but which might be labelled differently or not at all in other languages. Nevertheless, as we have noted, other differential criteria have been proposed, which might also be couched in metalanguage, and it can be asked what the basis is for singling out these particular criteria. Knowing you have done something bad can be labelled as guilt rather than shame; there are arguments that shame is experienced when we are aware that others may reasonably conclude that we have done something bad even though we do not share this judgment (Castelfranchi & Poggi, 1990). NSM accounts need to be tested.
More generally, the sources reviewed here make strong claims about the nature of the relations between shame and embarrassment. There “really” is either one emotion or two distinct ones and determining which is the case is a matter of establishing which conditions give rise to which emotion. Recurrent themes are identifiable yet it has proved difficult for the research community to converge on the criteria for distinguishing them. This conclusion may seem familiar in the context of current debates about “the crisis” in the study of emotion (Dixon, 2012) and criticisms that the field has failed to provide a convincing definition of emotion, confuses emotions with emotion concepts, is guilty of reification of concepts derived from the English language, and is misled by unacknowledged reliance on English vocabulary (Wierzbicka, 2009). How can shame be distinguished from embarrassment when there are no established criteria for distinguishing any pair of emotions? In the absence of such criteria surely seizing upon any feature or combination of features and regarding these as decisive is premature.
Progress will require developments in both theory and research. There are theories of shame (H. B. Lewis, 1971; Scheff, 1994; Tomkins, 1963) that have explored the phenomenology of shame in detail and have considered differences between shame and guilt but there have been no equivalent analyses of relations between shame and embarrassment. Empirical research designs to date share the limitation that they conceive of emotions as static entities whereas they might more usefully be construed as dynamic psychological states. Thus, H. B. Lewis (1971) regards shame as a family of feeling states with embarrassment as one of several variants, and in her account feeling states alternate, fuse, spiral, are discharged or not, and one can underlie or absorb another. Shame can be recognised by the ashamed person but need not be and might be identified as shyness, self-consciousness, or embarrassment or other terms in the emotion lexicon at the individual’s disposal.
Empirical enquiries should proceed on several fronts. For example, within English language populations it would be instructive to explore what people have in mind when they say they are both embarrassed and ashamed (this is meaningful and not a redundancy since people can and do say this; to take one example from a quick Internet search, Ruth Madoff, the wife of the convicted financier, was reported by CBS News to say that she was “embarrassed and ashamed” (CBS News, 2009) at her husband’s actions) and to investigate what is added to a psychological description when we say that someone is not only ashamed about something but is embarrassed as well, or what is added when we say that someone is ashamed as well as embarrassed; it would be essential to eschew prescribed categories and rating scales in eliciting responses. Other studies could investigate the circumstances that give rise to spontaneous descriptions in terms of embarrassment and go beyond vignettes or posed still photographs to draw upon participants’ and observers’ understandings of emotion-eliciting events; such naturally occurring events are now more accessible given the frequency of instances of loss of dignity, humiliation, self-consciousness, difficulties in functioning, and so on available in television reality programmes and Internet movie clips. This research should acknowledge that speakers of other languages may describe these kinds of circumstances differently and be alert to the issues this raises. Examination of the literature suggests that relevant factors to explore might include how serious are the consequences of failure to meet standards for the actor and for the social order in terms of social acceptance or rejection, the management of social identity in the short and long term, and whether immediate remedial action is required and is achievable.
The studies reviewed here suggest that identifying discriminatory criteria will not be a simple matter, which will result in a list of some kind, and approaches should be sufficiently sensitive to detect complexity. It may be more fruitful to think in terms of patterns of antecedents, experiences, and potential consequences, some of which are more likely to be described by English speakers as embarrassment, some more likely to be identified as shame, and others are labelled with less confidence, perhaps identified as sadness or amusement. These patterns and their descriptions may not be consistent across time and place, and this issue would be a matter for empirical enquiry. This approach is consistent with the inductive approach advocated by Barrett (2006). Rather than assuming that shame and embarrassment are “natural kinds” to be identified we should begin by investigating how these concepts are applied to circumstances that give rise to descriptions in terms of these emotions.
In conclusion, this study finds little consistency in the literature on the question of relations between shame and embarrassment and no convincing evidence that they are “natural kinds.” Recurrent themes are identifiable and all seem to capture something of common sense notions of the differences between them. What proves difficult to achieve is the establishment of explicit differential criteria whenever comparative claims are made that one emotion is more serious than the other, impacts on more central aspects of the self, or breaches more fundamental norms or rules. Failure to reach agreement together with acknowledgement that other languages make different distinctions, as historically English seems to have done in the case of embarrassment, implies that an approach that starts from the assumption that there are two emotion states and it is a matter of pinning them down by identifying their correlates is unlikely to succeed. The study of emotion more generally provides few guidelines for deciding upon criteria to distinguish any pair of emotions. Insufficient attention has been paid to precisely what is entailed when the question of the relations between shame and embarrassment is raised, whether it is about emotions as “natural kinds” or about emotion concepts whether scientific or couched in ordinary language. The differences between the two as captured in ordinary language are subtle and this implies that research should be more nuanced than is evident in current practice.