Shame and the Social Bond: A Sociological Theory

Thomas J. Scheff

Emotion has long been recognized in sociology as crucially important, but most references to it are generalized and vague. In this essay, I nominate shame, specifically, as the premier social emotion. First I review the individualized treatment of shame in psychoanalysis and psychology, and the absence of social context. Then I consider the contributions to the social dimensions of shame by six sociologists (Georg Simmel, Charles Cooley, Norbert Elias, Helen Lynd, Erving Goffman, Richard Sennett) and a psychologist/psychoanalyst (Helen Lewis). I show that Cooley and Lynd, particularly, made contributions to a theory of shame and the social bond. Lewis’s idea that shame arises from threats to the bond integrates the contributions of all six sociologists, and points toward future research on emotion, conflict, and alienation/ integration.

Many sociological theorists have at least implied that emotions are a powerful force in the structure and change of societies. Although Weber didn’t refer to emotions directly, his emphasis on values as the foundation of social structure implies it, since values are emotionally charged beliefs. Durkheim was more explicit about the role of emotions. Especially in his later works, he strongly implicated emotions and collective sentiments in the creation of social solidarity through moral community. G. H. Mead implied that emotion was an important ingredient in his social psychology. And Parsons promoted it to one of the components of social action in his AGIL scheme (Parsons and Shils 1955). Even Marx and Engels, who intended a purely structural theory, implicated emotions in class tensions and in the solidarity of rebellious classes.

Durkheim, especially, implied that "…what holds a society together---the "glue" of solidarity--- and [Marx implied that] what mobilizes conflict--- the energy of mobilized groups---are emotions" (Collins 1990). However, the classic formulations have not led to any theoretical or empirical pay-off. One reason is that they concerned emotions in general, a rarefied abstraction, rather than specific emotions.

WHY SPECIFIC EMOTIONS ARE NECESSARY FOR THEORY AND RESEARCH

The inclusion of emotions in classic sociology was abstract and therefore virtually meaningless. Generalized emotions have only ambiguous reference. Our knowledge of emotions is not generalized, but particular. For example, we all know a great deal about anger. No doubt some of what we think we know may not be the case. But much of what we know is probably accurate or at least accurate enough to often be able to understand each other. About anger we know or believe we know sources from which it arises, different forms and gradations it can take, and some of the outcomes that it can lead to. We also have similar kinds of knowledge and beliefs about other primary emotions, such as fear, grief, shame, contempt, disgust, love and joy.

Our knowledge about emotions held in common allows us to communicate with each other on this topic, and restrains flights of fancy. The different emotions may have several underlying similarities, but what is much more obvious is the great differences in origins, appearance, and trajectories. It is for this reason that general statements about emotions in the abstract have so little meaning. Some of what Durkheim, Mead, and Parsons said about emotions might appear plausible when applied to one emotion, say anger or fear, but not to most of the others. The sources, appearance and consequences of anger and fear are so different as to forbid lumping them together.

Treating all emotions together under a single heading amounts to a kind of dismissal. A current parallel can be found in rational choice theory, which divides behavior into the rational and the non-rational. In this theory, attention is given only to rational behavior. As in classical theory, the non-rational, the irrational, and emotional behavior is simply dismissed.

In any case, even the theorists who dealt with emotions explicitly, Durkheim, Mead, and Parsons, did not develop concepts of emotion, investigate their actual occurrence in real life, nor collect data that might bear on propositions about the role of emotions in human conduct. Their discussions of emotion, therefore, have not borne fruit.

The researchers whose work I review took the step of investigating a specific and therefore concrete, emotion. In their various studies that I will describe, they didn’t always emphasize the name of the emotion. Sennett and Cobb, for example, in the Hidden Injuries, made no move to develop a concept of shame, and named it infrequently, but their findings and many of their interpretations clearly imply it. As it turns out, the act of explicitly naming and defining is an important part of investigation. Before turning to these authors, however, I review the treatment of shame by psychoanalytic authors, in order to show the problem that the sociologists and Lewis solved.

Shame in Psychoanalytic Theory

The treatment of shame in most psychoanalytic writing is problematic because it leaves out the social matrix. Like most psychological theory, Freud’s formulations concern emotions in isolated individuals, ignoring the social context. Individualistic formulations give rise to what might be called the inside/outside problem. If one ignores the context in which emotions arise, it will inevitably be difficult to understand their place in human behavior. Freud’s solution to the inside/outside problem was to ignore the outside.

Although in his later work Freud also ignored shame, it had an important role in his first book. In Studies on Hysteria (1895), Freud and Breuer stated early on (p. 40) that hysteria is caused by hidden affects, and named the emotion of shame as one of these affects. Near the end of the book, this idea is urged more strongly:

"[The ideas that were being repressed] were all of a distressing nature, calculated to arouse the affects of shame, self-reproach and of psychical pain and the feeling of being harmed" (pp. 313).

Note that all of the affects mentioned can be considered to be shame derivatives or cognates. Self-reproach is a specific shame cognate, the feeling of being harmed (as in rejection) somewhat broader, and finally, the quite abstract phrase "psychical pain", which, like "hurt" or "emotional arousal" can be applied to any emotion. In this passage and several others, shame is given a central role in the causation of psychopathology. Freud and Breuer also proposed that shame is the inhibiting emotion that leads to repression, therefore giving it a central role in the development and maintenance of psychopathology. The idea that it is shame that causes repression would also give shame the leading role in the causation of all mental illness, not just hysteria, if Freud had stayed with it.

However, in 1905, with the publication of The Interpretation of Dreams, Freud permanently renounced his earlier formulation in favor of drive theory, especially the sexual drive. In Freud’s thinking, shame was replaced by anxiety and guilt, the appropriate emotions for responsible adults, especially male adults. By this time, Freud had become biased about shame. He thought that it was regressive emotion, seen only in children, women, and savages. His rejection of his earlier work on shame can be seen as a lapse into the ethnocentric and sexist attitudes that prevailed at the time, as well as being psychologistic.

Since 1905, shame has been largely ignored in orthodox psychoanalytic formulations. Although several psychoanalysts made crucially important contributions to shame knowledge, these contributions helped make them marginal to psychoanalysis. Shame also goes unnamed and/or undefined even in these marginal analysts. Alfred Adler, Abraham Kardiner, Karen Horney, and Erik Erickson provide examples.

Adler’s formulation of the core position of prestige seeking in human behavior, and his concept of the inferiority complex are clearly shame based ideas. To make the search for prestige and honor a central human motive is to focus on the pride/shame axis, as Cooley did. Similarly, the concept of an inferiority complex can be seen as a formulation about chronic low self-esteem, or the put it more bluntly, chronic shame.

Yet Adler never used the concept of shame to integrate the various dimensions of his work, as he might have. His theory of personality was that children deprived of love at key periods in their development would become adults with either a drive for power or an inferiority complex. This theory can be restated succinctly in terms of a theory of shame and the social bond: children without the requisite secure bonds will likely become adults whose affects are predominately bypassed (drive for power) or overt shame (inferiority complex) (Lewis 1971).

Like Adler, Karen Horney (1950) didn’t name the emotion of shame. But her formulations clearly implied it. Her theory of personality was based on what she called "the pride system." Most of her central propositions imply that pride and shame are the keys to understanding both neurotic and normal behavior. Her concept of the "vindictive personality" seems to imply shame/anger sequences as the emotional basis for vengeful behavior.

Abraham Kardiner was an anthropologist who applied psychoanalytic ideas to his studies of small traditional societies. In The Individual and His Society: the Psychodynamics of Primitive Social Organization (1939), he offered an extensive analysis of the role of shame in four traditional societies. Unlike Adler and Horney, he named the emotion of shame clearly, and stated directly, like Freud and Breuer, that shame is the emotion that leads to repression. Like Adler, he also gave prominence to prestige as a fundamental human motive. Going further than Adler or Freud, he named shame as the principal component of the super-ego, that is, of conscience.

Like Kardiner, Erik Erikson also named shame directly, in his analysis of the relationship between shame and guilt (1950). In his investigation of these emotions, he proposed, again contra Freud, that shame was the most fundamental emotion and that it had a vital role in the developmental stages through which all children must pass. His analysis of shame was an important source for Helen Lynd’s work on shame; reading Erikson might have been the beginning of Lynd’s interest in shame. Like most theorists who discuss shame, neither Kardiner nor Erikson tried to define it.

The work on shame by these four analysts was not recognized by the psychoanalytic establishment. Both Adler and Horney were excluded for their deviationism. Although neither Kardiner nor Erikson were excluded, there was no response to their contributions on shame, with the exception of Helen Lynd to Erikson. It is also of interest that among the disciples of Adler and of Horney that there was also no response to their work on shame.

Although there has been a re-awakening of interest in shame by current psychoanalysts, still only a small minority of analysts are involved. Even in this group, converting from drive theory to shame language is a struggle. The work of Lansky (1992; 1995) on shame preserves drive theory. Morrison (1989) has translated drive theoretic formulations into shame dynamics, trying to bridge the two worlds. Broucek (1991) has rebelled against drive theory, but doesn’t attempt a social formulation of shame. Only Lewis (1971) has succeeded in throwing off drive theory, recasting shame in social terms. I return to her work after considering sociological contributions to the study of shame.

Sociological Pioneers in the Study of Shame

Five of the six sociologists I reviewed acted somewhat independently of each other and of other researchers on emotion. In the case of Elias and Sennett, their discovery of shame seems forced upon them by their data. Certainly Elias’s work on shame thresholds in the history of four European cultures appears to arise from nowhere. Neither Simmel nor Cooley define what they mean by shame. Goffman partially developed a conceptualization of embarrassment. The exception is Helen Lynd, who was self-conscious about shame as a concept. Lynd’s book on shame was contemporaneous with Goffman’s first writings on embarrassment. She cited his article on face-work, and realized that its main point was that face-work meant avoiding embarrassment and shame. Even though Lynd was a sociologist, she was also interested in psychoanalytic theory; her book surveys the fugitive psychoanalytic literature on shame.

Helen Lewis was both a research psychologist and a psychoanalyst. In her empirical work on shame (1971), she was influenced by the work of Lynd more than by the scattered psychoanalytic literature on shame. She also was sophisticated in formulating a concept of shame, and in using systematic methods to study it.

Sennett’s work involved slight outside influence. He approvingly cited the Lynd book on shame in The Hidden Injuries of Class (1972), and he has a chapter on shame in his book Authority (1980) in which he cites the English translation of Elias’s The History of Manners (1978). As I will suggest, shame is a central thread in The Hidden Injuries, Sennett’s first approach to shame, even though there is no concept of shame in that book. The reverse is true in the case of Elias; the concept of shame is central to The Civilizing Process (1939), but goes underground in The Germans (1996), though shame tacitly forms the central core of the book. I will refer below to the parallels and connections between Elias’s and Sennett’s work.

Simmel: Shame and Fashion

Shame plays a significant part in only one of Simmel’s essays, the one on fashion (1904). In his analysis of the origins of fashion, he clearly states that the emotion of shame is it’s source. People want variation and change, he argued, but they also anticipate shame if they stray from the behavior and appearance of others. Fashion is the solution to this problem, since one can change along with others, avoiding being isolated, and therefore shame (p 553). Simmel’s idea about fashion implies conformity in thought and behavior among one group in a society, the fashionable ones, and distance from another, those who do not follow fashion, relating shame to the dynamics of alienation.

There is a quality to Simmel’s treatment of shame that is somewhat difficult to describe, but needs description, since this quality will characterize most of the other sociological treatments reviewed here. Simmel’s use of shame is casual and unselfconscious. His analysis of the shame component in fashion occurs in a single long paragraph. Shame is not mentioned before or after this paragraph. He doesn’t conceptualize shame or define it, seeming to assume that the reader will automatically know the meaning of the term. Similar problems are prominent in Cooley and Elias, and somewhat less flagrantly, in Sennett and Goffman. Lynd and Lewis are exceptions, since they both attempted to define shame and locate it with respect to other emotions.

Cooley: Shame and the Looking Glass Self

Cooley (1922), like Simmel, was direct in naming shame. For Cooley, shame and pride both arose from self-monitoring, the process that was at the center of his social psychology. To be sure, in his discussion of what he called the "self-sentiments," pride and shame are mentioned only as two of other possible emotions. But his concept of "the looking glass self," which implies the social nature of the self, refers directly and exclusively, at the level of emotions, to pride and shame. Cooley saw self-monitoring in terms of three steps (l84): "A self-idea of this sort seems to have three principal elements: the imagination of our appearance to the other person; the imagination of his judgment of that appearance, and some sort of self-feeling, such as pride or mortification."

In this passage he restricts self-feelings to the two which he seems to think are the most significant, pride and shame (considering "mortification" to be a shame variant). To make sure we understand this point, he mentions shame three more times in the passage that follows (l84-85, emphasis added):

"The comparison with a looking-glass hardly suggests the second element, the imagined judgment, which is quite essential. The thing that moves us to pride or shame is not the mere mechanical reflection of ourselves, but an imputed sentiment, the imagined effect of this reflection upon another's mind. This is evident from the fact that the character and weight of that other, in whose mind we see ourselves, makes all the difference with our feeling. We are ashamed to seem evasive in the presence of a straightforward man, cowardly in the presence of a brave one, gross in the eyes of a refined one and so on. We always imagine, and in imagining share, the judgments of the other mind. A man will boast to one person of an action—say some sharp transaction in trade—which he would be ashamed to own to another."

Although Cooley is explicit in suggesting that pride and shame are social emotions, he made no attempt to define either emotion. Instead he used the vernacular words as if they were self-explanatory.

To give just one example of the ensuing confusion over not defining emotions such as pride: In English and other European languages, the word pride used without qualification usually has an inflection of arrogant pride or hubris (Pride goeth before the fall.) In current usage, in order to refer to the kind of pride implied in Cooley’s analysis, the opposite of shame, one must add a qualifier like justified or genuine. Using undefined emotion words is an invitation to the Tower of Babel.

However, Cooley's analysis of self-monitoring suggests that pride and shame are the basic social emotions. At this point intellectual history takes a somewhat surprising turn. Elaborating on Cooley’s idea of self-monitoring, G.H. Mead and John Dewey based their entire social psychology upon the process of role taking, the ability of humans to continuously monitor themselves from the point of view of others. Yet neither Mead nor Dewey mention what was so obvious to Cooley, that social monitoring usually gives rise to feelings of pride or shame. Mead and Dewey treat role taking, their basic building block of human behavior, as entirely a cognitive process. Neither has anything to say about pride and shame, as if Cooley never existed.

Furthermore, until quite recently, all other readers of Cooley have also ignored the central role in human experience he gave to pride and shame. The "looking-glass self" is by far the best known of Cooley’s formulations. Particularly puzzling is the obliviousness of current symbolic interaction to Cooley’s presentation of shame. How could the blatant attention he gave to pride and shame be ignored?

Cooley’s formulation of the social basis of shame in self-monitoring can be used to amend Mead’s social psychology. Perhaps the combined Mead/Cooley formulation can solve the inside-outside problem that plagues psychoanalytic and other psychological approaches to shame, as I suggest below.

Elias: Shame in the Civilizing Process

In the first book he wrote, Elias undertook a historical analysis of what he calls the "civilizing process" (1994). He traced changes in the development of personality and social norms from the onset of modern civilization to the present. Like Weber, he gave great prominence to the development of rationality. Unlike Weber, however, he gave equal prominence to emotional change, particularly to changes in the threshold of shame: "No less characteristic of a civilizing process than "rationalization" is the peculiar molding of the drive economy that we call "shame" and "repugnance" or "embarrassment" (292, l982)."

Using excerpts from advice manuals over a very long historical span, the last five centuries, he outlined a theory of modernity. By examining advice concerning etiquette, especially table manners, body functions, sexuality, and anger, he suggests that a key aspect of modernity involved a veritable explosion of shame. Although he uses somewhat different terminology, Elias' central thesis concerning modernity is closely related to my own (Scheff, l990). In my analysis of modernity, I point to what I consider the alienating consequences of modernity: decreasing shame thresholds at the time of the break-up of rural communities, and decreasing acknowledgment of shame, which also may have had powerful consequences on levels of awareness and self-control.

I will cite only one of many advice excerpts used by Elias. He first presents a lengthy excerpt from a l9th century advice book, The Education of Girls (von Raumer, l857), that advises mothers how to answer sexual questions. In response to the question "Where do babies come from? Von Raumer suggests "Children should be left as long as possible in the belief that an angel brings the mother her little children..." If the issue comes up again, the child is to be sternly warned: ‘ It is not good for you to know such a thing, and you should take care not to listen to anything said about it.’ " Von Raumer concludes this passage with advice that both shames the mother and advises her to shame the daughter: "A truly well-brought-up girl will from then on feel shame at hearing things of this kind spoken of."

This advice suggests three different puzzles:

l. Why is the author, von Raumer, offering the mother such absurd advice?

2. Why does the mother follow his advice (as most did, and still do)?

3. Why do the daughters follow their mothers’ advice (as most did, and still do)?

Modern feminist theory might respond to the first question that von Raumer's advice arises from his position of power: he sought to continue male supremacy, by advising the mother to act in a way that is consonant with the role of women as subordinate to that of men. That is, he was promulgating the woman's role as Kirche, Kueche, Kinder (church, kitchen, children). Keeping women ignorant of sexuality and reproduction would help to continue this system.

This formulation is probably part of a complete answer, but it does not attend to the other two questions. Why do mothers and daughters submit to ignorance and shame? Elias's formulation provides an answer to all three questions, without contradicting the feminist answer. Each of these persons, the man and the two hypothetical readers, the mother and the daughter, is too embarrassed about sexuality to think clearly about it. It could be true that von Raumer's advice is part of his male chauvinist position, and also true that he is too embarrassed to think about the meaning of his advice. Thoughts and emotions are both parts of a causal chain.

Elias's study suggests a way of understanding the social transmission of a taboo on shame. The adult, the author von Raumer in this case, is not only ashamed of sex, but he is ashamed of being ashamed, and probably ashamed of the shame that he will arouse in his reader. The mother responding to von Raumer's text, in turn, will probably react in a similar way, being ashamed, and being ashamed of being ashamed, and being ashamed of causing further shame in the daughter. Von Raumer's advice is part of a social system in which attempts at civilized delicacy result in an endless chain reaction of unacknowledged shame. The chain reaction is both within persons and between them, three spirals (one spiral within each party, and one between them. The spiral idea integrates social and psychological processes, and suggests a solution to the usual separation of inside and outside, as I suggest at the end of this article.

Elias showed that there was much less shame about manners and emotions in the first part of the period he studied, than there was in the 19th century, and therefore, I infer, far fewer shame loops. In the late l7th and early l8th century, a change began occurring in advice on manners. What was said openly and directly earlier begins only to be hinted at, or left unsaid entirely. Moreover, open justifications are offered less and less. One is mannerly because it is the right thing to do. Any decent person will be courteous; the intimation is that bad manners are not only wrong but also unspeakable, the beginning of repression.

The change that Elias documents is gradual but relentless; by a continuing succession of small decrements, etiquette books fall silent about the reliance of manners, style, and identity on respect, honor, and pride, and avoidance of shame and embarrassment. By the end of the l8th century, the social basis of decorum and decency had become virtually unspeakable. Unlike Freud or anyone else, Elias documents, step by step, the sequence of events that led to the repression of emotions in modern civilization.

By the l9th century, Elias proposes, manners are inculcated no longer by way of adult to adult verbal discourse, in which justifications are offered. Socialization shifts from slow and conscious changes by adults over centuries to swift and silent indoctrination of children in their earliest years. No justification is offered to most children; courtesy has become absolute. Moreover, any really decent person would not have to be told, as suggested in the text interpreted above. In modern societies, socialization of most children automatically inculcates and represses shame.

Although Elias’s analysis of the changing shame threshold in European history now seems extraordinarily important, it is the last time that he explicitly referred to this process. None of his many subsequent works return to an explicit analysis of shame.

Richard Sennett: Is Shame the Hidden Injury of Class?

Although The Hidden Injuries of Class (1973) carries a powerful message, it is not an easy book to summarize. The narrative is developed from quoted excerpts from the interviews, and the authors’ brief, low-level interpretations of the meanings of these quotes. They do not devise a conceptual scheme and a systematic method for analyzing their interviews and observations. For this reason, readers are required to devise their own conceptual scheme, as I will do below.

The book is based on participant-observation in communities, schools, clubs and bars, and 150 interviews with white working class males, mostly of Italian or Jewish background, in Boston for one year beginning in July of 1969 (p. 40-41).

The hidden injuries that Sennett and Cobb discovered might be paraphrased in this way: their working class men felt that first, because of their class and occupational position, they were not accorded the respect that they should have gotten from others, particularly from their teachers, bosses, and even from their own children. That is, these men have many complaints about their relationships. Secondly, a more subtle injury: these men also felt, in some ways, that their class and occupational position was at least partly their own fault. Sennett and Cobb imply that social class is responsible for both injuries. They believe that their working men did not get the respect they deserved because of their social class, and that the second injury, lack of self-respect, is also the fault of class, rather than the men’s own fault, as most of them thought.

Sennett and Cobb argue that in American society, the respect one receives is largely based on one’s individual achievement, the extent that one’s accomplishments give one a unique identity that stands out from the mass of others. The role of public schools in the development of abilities forms a central part of Sennett and Cobb’s argument. Their informants lacked self-respect, the authors thought, because the schooling of working class boys did not develop their individual talents in a way that would allow them to stand out from the mass as adults. In the language of the sociology of emotions, they carry a burden of feelings of rejection and inadequacy, which is to say chronic low self-esteem (shame).

Sennett, who did the participant-observation part of the study, reported most fully on a particular grammar school, "Watson School," that he observed. He suggests that "teachers act on their expectations of the children in such a way as to make the expectations become reality" (p. 81). One of his observations concerns a second-grade class:

(Indent) In this class there were two children, Fred and Vincent, whose "...clothes were pressed and seemed better kept" than the other children’s’ clothes. "In a class of mostly dark Italian children, these were the fairest skinned. From the outset the teacher singled out these two children...To them he spoke with a special warmth in his voice. He never praised them openly... but a message that they were different, and better, was spontaneously conveyed" (p. 81). (End indent)

Sennett and Cobb argue that teachers single out for attention and praise only a very small percentage of the students, usually students who are either talented or middle class or closest in actions and appearance to middle-class. This praise and attention allows the singled-out students to develop their potential for achievement. The large majority of the boys, however, are ignored and, in subtle ways, rejected.

"... by the time the children are ten or eleven the split between the many and the few who are expected ‘to make something of themselves’ is out in the open... [The mass of] boys in class act as though they were serving time, as though schoolwork and classes had become something to wait out, a blank space in their lives they hope to survive..." (pp. 82-83).

This statement is a damning indictment of public schools. There are a few working class boys who achieve their potential by virtue of their superior academic or athletic talents. But the large mass does not. For them, rather than opening up the world of culture and accomplishment, public schools close this vision off. Education, rather than becoming a source of personal and cultural growth, provides only shame and rejection. For the majority of students in public schools, surviving the days and years of large classes means running a gauntlet of shame and embarrassment every day. These students learn by the second or third grade that is better to be silent in class rather than risk ridicule or humiliation of a wrong answer. Even students with the right answers must deal with having the wrong accent, clothing or physical appearance. For most students, schooling is a vale of shame.

Helen Lynd: Shame and Identity

During her lifetime, Helen Lynd was a widely known sociologist. With her sociologist husband, Robert, she published the first American community studies, Middletown and Middletown in Transition. But Lynd was also profoundly interested in developing an interdisciplinary approach to social science. In her study On Shame and the Search for Identify (1958), she dealt equally with both the social and psychological sides of shame. She also clearly named the emotion of shame and its cognates, and located her study within previous scholarship, especially psychoanalytic studies. But Lynd also modified and extended the study of shame by developing a concept, and by integrating its social and psychological components.

In the first two chapters, Lynd introduces the concept of shame, using extended examples from literature, often discourse, to clarify every point she makes. In the next section, she critiques mainstream approaches to emotion and behavior in psychology and the social sciences. She is especially critical of the orthodox psychoanalytic and the conventional quantitative approaches. She then introduces ideas from lesser-known approaches, showing how they might resolve some of the difficulties. Finally, she has an extended discussion of the concept of identity, suggesting that it might serve to unify the approach to persons in the social sciences by integrating the concepts of self, ego, and social role under the larger idea of identity. This latter idea strikes me as particularly insightful.

Lynd’s approach to shame is much more analytical and self-conscious than the other sociologists reviewed here. They treated shame as a vernacular word, rather than as a concept. For them, shame sprung out of their data, unavoidable. But Lynd encounters shame deliberately, as part of her exploration of identity. First she explains her discomfort with mainstream approaches and domains in the social sciences:

...techniques for the study of human nature were never so abundant; there were never so many people engaged in using them. If understanding of identity... could be discovered by such means, [the problem of identity] would be assured of solution. But since every way of seeing is also a way of not seeing, the very multiplication of categories and the very precision of techniques may ... act as barriers...

Certain pervasive experiences, not easily labeled, may slip through the categories altogether, or if given a location and a name, may be circumscribed in such a way that their essential character is lost...Among such experiences are...shame, anxiety, joy, love, sense of honor, wonder, curiosity, longing, sense of pride, self-respect. Of these, only anxiety has been the subject of extensive specialized study (1958 p. 16).

Note that Lynd’s list contains three other topics besides shame that are closely related to shame: sense of honor, sense of pride, and self-respect. The large-scale cross-cultural study of emotions by Paul Ekman (1972) is a good example of what she is saying about shame slipping through the technique net altogether. Although shame is obviously a central emotion in all cultures and historical eras, Ekman and his co-workers have left it out completely, as have many other students of emotion.

Lynd explains that shame and its cognates and related emotions get left out because it is more deeply hidden, but at the same time so pervasive in human affairs as to be like water for fish. She makes this point in many ways, particularly in the way she carefully distinguishes shame from guilt. She notes that guilt is usually extremely specific and therefore close to the surface; it involves specific acts done or not done. Guilt is about what one did, shame is about the self, what one is. Guilt also involves feeling that the ego is strong and intact: one is powerful enough to injure another, and one is also powerful enough to make amends. By contrast, shame feels like weakness and dissolution of the self, even for the wish that the self would disappear. Guilt is a highly individualist emotion, reaffirming the centrality of the isolated person; shame is a social emotion, reaffirming the emotional interdependency of persons.

One point that Lynd makes is profoundly important for a social theory of shame and the bond, that sharing one’s shame with another can strengthen the relationship:

The very fact that shame is an isolating experience also means that if one can find ways of sharing and communicating it this communication can bring about particular closeness with other persons" (Lynd 1958 p. 66).

In another place, Lynd went on to connect the process of risking the communication of shame with the kind of role-taking that Cooley and Mead had described: "communicating shame can be an experience of …entering into the mid and feelings of another person" (p. 249).

Lynd’s idea about the effects of communicating and not communicating shame was pivotal for Lewis’s (1971)concepts of acknowledged and unacknowledged shame, and their relationship to the state of the social bond, as outlined below.

Goffman: Embarrassment and Shame in Everyday Life

Since his Interaction Ritual (1967) and other books that I will mention are the most widely appreciated of the studies reviewed here, Goffman advanced the idea of the centrality of shame in social relations more than any of the other sociologists. Although shame goes largely unnamed in his early work on presentation of self, embarrassment and avoidance of embarrassment is the central thread. Goffman’s Everyperson is always desperately worried about his image in the eyes of the other, trying to present herself with her best foot forward to avoid shame. This work elaborates, and indeed, fleshes out, Cooley’s abstract idea of the way in which the looking glass self leads directly to pride or shame, giving the idea roots in the reader’s imagination.

Interaction Ritual made two further contributions to shame studies. In his study of face-work, Goffman states what may be seen as a crucial model of "face" as the avoidance of embarrassment, and losing face as suffering embarrassment. This is an advance, because it offers readily observable markers for empirical studies of face; otherwise the idea of face is only an ambiguous metaphor. The importance of this idea is recognized, all too briefly, at the beginning of Brown and Levinson’s (1987) study of politeness behavior. In this work, the authors define politeness as the maintenance of face, and, following Goffman, state that maintaining face involves avoiding embarrassment (p. 61). Oddly, however, following this introduction, embarrassment is seldom referred to in the rest of the book; none of the empirical studies use it as a concept, and it is mentioned infrequently and in passing. Embarrassment, since it is one type of shame, has a slippery quality, even for those authors who don’t completely ignore it, its usual fate.

Goffman’s second contribution to the study of shame was made in a concise essay on the role of embarrassment in social interaction (1967). Unlike any of the other shame pioneers in sociology, he begins the essay with an attempt at definition:

"An individual may recognize extreme embarrassment in others and even in himself by the objective signs of emotional disturbance: blushing, fumbling, stuttering, an unusually low- or high-pitched voice, quavering speech or breaking of the voice, sweating, blanching, blinking, tremor of the hand, hesitating or vacillating movement, absentmindedness, and malapropisms. As Mark Baldwin remarked about shyness, there may be "a lowering of the eyes, bowing of the head, putting of hands behind the back nervous fingering of the clothing or twisting of the fingers together and stammering, with some incoherence of idea as expressed. in speech." There are also symptoms af a subjective kind: constriction of the diaphragm, a feeling of wobbliness, consciousness of strained and unnatural gestures, a dazed sensation, dryness of the mouth, and, tenseness of the muscles."

This definition is a definite advance, since it sets the stage for an operational definition of shame that can be used in systematic research. But it also foretells a limitation of the whole essay, since the defnition is behavioral and physiological, rather than concerning inner experience. The part that Goffman refers to as "subjective" is entirely physiological. He makes no attempt to include the ideation of embarrassment. Framing his analysis in what he thought of as a Durkheimian, purely sociological mode, Goffman therefore omitted most of the feelings and thoughts experienced in embarrassment. His solution to the inside/outside problem was to ignore most of inner experience, just as Freud ignored most of outside events.

However, Goffman affirms Cooley’s point on the centrality of the emotions of shame and pride in normal, everyday social relationships. In Goffman’s language: "One assumes that embarrassment is a normal part of normal social life, the individual becoming uneasy not because he is personally maladjusted but rather because he is not… embarrassment is not an irrational impulse breaking through social prescribed behavior, but part of this orderly behavior itself" (1967, p. 109 and 111.).

Even Goffman’s partial definition of the state of embarrassment represents an advance. One of the most serious limitations of current contributions to the sociology of emotions is the lack of definitions of the emotions under discussion. Much like Cooley, Elias, and Sennett, Kemper (1978) offers no definitions of emotions, assuming that they go without saying. Hochschild (1983) attempts to conceptualize various emotions in an Appendix, but doesn’t go as far as to give concrete definitions of emotional states. Only in Retzinger (1991; 1995) can conceptual and operational definitions of the emotions of shame and anger be found.

Goffman, like Sennett and Elias, named the emotion of shame infrequently. In his early work on presentation and self and on facework, embarrassment and avoidance of embarrassment is the central thread, but goes largely unnamed. When he does clearly name embarrassment in his essay on this affect, he carefully avoids a complete analysis, which would include both its social and its psychological side. Although only implicitly, he seems to separate embarrassment from shame, as if they did not belong too the same family of emotions. In his studies of stigma and asylums, shame again is a key element, but he mentions it only in passing.

Lewis’s Discovery of Unacknowledged Shame and its Social Implications

Helen Lewis was a research psychologist and a psychoanalyst who published many systematic studies in psychology. Like Elias and Sennett, her discovery of shame seemed forced upon her by her data. Her book on shame (1971) was based on an analysis of verbatim transcripts of hundreds of psychotherapy sessions.

She encountered shame because she used a systematic method for identifying emotions in verbal transcripts, the Gottschalk-Gleser method (1969; 1995). This method involves use of long lists of key words that are correlated with specific emotions, such as anger, grief, fear, anxiety, and shame.

Lewis found that anger, fear, grief, anxiety cues showed up from time to time in some of the transcripts. What she was unprepared for was the massive frequency of shame cues. The findings from her study most relevant to this article are:

1. Prevalence: Lewis found a high frequency of shame markers in all the sessions, far outranking markers of all other emotions combined. This finding alone suggests that shame was a dominant force in the sessions she analyzed.

2. Lack of awareness; two forms: Lewis noted that although shame markers were frequent in all of the sessions, patient or therapist almost never referred to shame or its near cognates. Even the relatively mild word embarrassment was seldom used. In analyzing the context in which shame markers occurred, Lewis identified a specific context: situations in which the patient seemed to feel distant from, rejected, criticized, or exposed by the therapist, generated a cloud of shame markers. This context fits the Darwin/Cooley proposition that shame arises from seeing one’s self negatively from the point of view of the other.

However, the patient’s showed two different, seemingly opposite responses in the shame context. In one, the patient seemed to be suffering psychological pain, but failed to identify it as shame. Lewis called this form overt, undifferentiated shame. In a second kind of response, the patient seemed not to be in pain, revealing an emotional response only by rapid, obsessional speech on topics that seemed somewhat removed from the dialogue. Lewis called this second response bypassed shame. Identifying or calling shame by its right name seems to be an important aspect of understanding and managing it.

3. Shame, Anger and Conflict

In her transcripts, Lewis found many episodes of shame that extended over long periods of time. Since emotions are commonly understood to be brief signals (a few seconds) that alert us for action, the existence of long-lasting emotions is something of a puzzle. Lewis’s solution to this puzzle may be of great interest in the social sciences, since it provides an emotional basis for longstanding hostility, withdrawal or alienation.

She argued that her subjects often seemed to have emotional reactions to their emotions, and that this loop may be extended indefinitely. She called these reactions "feeling traps." The trap that arose most frequently in her data involved shame and anger. A patient interprets an expression by the therapist as hostile, rejecting, or critical, and responds with shame or embarrassment. However, the patient instantaneously masks the shame with anger, then is ashamed of being angry. Apparently each emotion in the sequence is brief, but the loop can go on forever. This proposal suggests a new source of protracted conflict and alienation, one hinted at in Simmel’s treatment of conflict, but stated explicitly by Lewis.

Although Lewis didn’t discuss other kinds of spirals, there is one that may be as important as the shame/anger loop. If one is ashamed of being ashamed, it is possible to enter into a shame/shame loop that leads to silence and withdrawal, as already suggested in the discussion of Elias’s work above.

4. Shame and the Social Bond: Finally, Lewis interpreted her findings in explicitly social terms. She proposed that shame arises when there is a threat to the social bond, as was the case in all of the shame episodes she discovered in the transcripts. Every person, she argued, fears social disconnection, of being adrift from understanding and being understood by the other.

Lewis’s solution to the outside/inside problem parallels and advances the Darwin/Cooley definition of the social context of shame. She proposed that shame is a bodily and/or mental response to the threat of disconnection from the other. Shame, she argued, can occur in response to threats to the bond from the other, but in can also occur in response to actions in the "inner theatre", in the interior monologue in which we see ourselves from the point of view of others. Her reasoning fits Cooley’s formulation of shame dynamics, and also Mead’s (1934) more general framework: the self is a social construction, a process constructed from both external and internal social interaction, in role-playing and role-taking.

Lewis’s formulations also suggest a correction to the sociologists’ analysis of shame. Perhaps Goffman’s analysis of the harried Everyperson desperately seeking to maintain face may be somewhat overgeneralized. Most of Goffman’s examples of persons presenting self in everyday life seem to be either bypassing shame, or experiencing it only in its overt, undifferentiated form, to use Lewis’s terminology. This formulation suggests that Goffman’s people have insecure bonds, to the extent that they are unable to acknowledge, and therefore dispel shame. His analysis might be most appropriate to persons alienated from self and/or others. Similarly, Sennett and Cobb’s workmen seem to experience shame only in its overt/undifferentiated form, again suggesting alienation. Similar reasoning may also apply to Simmel’s fashion seekers. Like Goffman and Sennett, Simmel may be considering only the alienated.

Naming and Unnaming Shame

Lewis’s analysis of unacknowledged shame has implications for this review of the sociological exploration of shame. Although Elias made it clear in The Civilizing Process that shame analysis is a key element in his argument, it is much less clear in his study (with Scotson), The Established and the Outsiders (1965). Shame was not the central concern, but it is indirectly present in the idea that outsiders are stigmatized. In his study of the Germans (1996), though again not made explicit, shame plays a much larger role. Although the word shame and its variants (embarrassment, humiliation, low self-esteem, lack of self-confidence, etc) occur repeatedly, Elias does not make explicit that shame is a key concept, as he did in TCP. (In a word count from a file of the ms. of The Germans I found more than five hundred references to shame and its cognates). For reasons that are not immediately obvious, in the later two books shame was demoted from a central issue to its common cognates.

Elias's argument was that the Germans, both as persons and as nation, historically have been unable to respond to humiliation in any other way than fighting. His argument is quite similar to my analysis of the humiliated fury that arose during the period of the three Franco-German wars, 1870-1945 (Scheff, 1994). I was unaware of this similarity until after I had published my book. I proposed that because of the French defeat in 1871, unacknowledged shame was a key element on the French side leading to the First World War, and following their defeat in 1918, on the German side leading to the Second World War. Like Elias, I propose unacknowledged shame as one source of protracted conflict.

Following his success in analyzing shame in The Civilizing Process, why did Elias not develop a technical conception of shame, just as he moved toward a technical conception of interdependence? There is no sure way of answering this question, but one possibility concerns the response of the audience to that book. Unless I am mistaken, there was virtually no response to his shame analysis, even though it plays a central role in his overall thesis. The only mention I have been able to find is by Sennett (1980), who recognized the applicability of Elias's shame analysis to the problem of social control. Sennett argued that shame and social-economic dependence are intertwined (pp. 45-49), and that shame plays a central role as a tool of discipline of workers by management (pp. 92-97).

Although shame was the key injury in Sennett and Cobb’s earlier study, The Hidden Injuries of Class (1971), the authors had not conceptualized it as such. Although they refer approvingly to Lynd’s (1958) book on shame (p. 127), their study of working class men is couched largely in the men’s own vernacular. These men spoke of feeling that it was their fault that they had not gotten ahead, that there must be something wrong with them, that they lacked self-respect. The phrase "lack of self-respect" occurs frequently, if not universally in the men’s litany of complaints about themselves. Sennett and Cobb did not classify these various responses as shame clues (Retzinger 1990; 1995). Instead they used the same codewords the men did, to continue to hide the already hidden injury.

Apparently the publication of The Civilizing Process in English (1978), with its open conceptualization of shame as a key component of the civilizing process, encouraged Sennett to point to shame directly and openly in Authority (1980). However, just as there was virtually no response to Elias's shame analysis, there was also none to Sennett's. Just as Elias failed to develop a technical concept of shame, this emotion disappeared from Sennett's later work. It might not be stretching a point to conclude that both Elias and Sennett were shamed into silence by the silence of their audiences. Just as shame goes unacknowledged in most social interaction, it is also unacknowledged, and for similar reasons, in social science.

Shame as the Social Emotion

Drawing upon the work of the pioneers reviewed here, it is possible to take further steps toward defining shame. By shame I mean a large family of emotions that includes many cognates and variants, most notably embarrassment, humiliation and related feelings such as shyness that involve reactions to rejection or feelings of failure or inadequacy. What unites all these cognates is that they involve the feeling of a threat to the social bond. That is, I use a sociological definition of shame, rather than the more common psychological one (perception of a discrepancy between ideal and actual self). If one postulates that shame is generated by a threat to the bond, no matter how slight, then a wide rage of cognates and variants follow: not only embarrassment, shyness, and modesty, but also feelings of rejection or failure, and heightened self-consciousness of any kind. Note that my definition will usually subsume the psychological one, since most ideals are social, rather than individual.

If, as proposed here, shame is a result of threat to the bond, shame would be the most social of the basic emotions. Fear is a signal of danger to the body, anger a signal of frustration, and so on. The sources of fear and anger, unlike shame, are not uniquely social. Grief also has a social origin, since it signals the loss of a bond. But bond loss is not a frequent event. Shame on the other hand, following Goffman, since it involves even a slight threat to the bond, is pervasive in virtually all social interaction. I propose that shame is the emotion that Durkheim should have named as the social emotion. As Goffman’s work suggests, all human beings are extremely sensitive to the exact amount of deference they are accorded. Even slight discrepancies generate shame or embarrassment. As Darwin (1872) noted, the discrepancy can even be in the positive direction; too much deference can generate the embarrassment of heightened self-consciousness.

Especially important for social control is a positive variant, a sense of shame. That is, shame figures in most social interaction because members may only occasionally feel shame, but they are constantly anticipating it, as Goffman implied. Goffman’s treatment continually, but only implicitly, points to the slightness of threats to the bond that lead to anticipation of shame or embarrassment.

My use of the term shame is much broader than its vernacular use. In common parlance, shame is a negative, crisis emotion closely connected with disgrace. But this is much too narrow if we expect shame to be generated by even the slightest threat to the bond. Because Simmel, Cooley, Elias, Goffman and Sennett used only the vernacular word, when they named the emotion of shame at all, there is considerable confusion in interpreting their work today.

Conclusion

An obvious question arises from my description of the zigzag progress of shame studies described above. What gives rise to the slipperiness of the concept of shame? Why did Elias, Sennett, Goffman and others make fundamental contributions to shame knowledge, yet fail to explicitly name and define the emotion they studied as shame, or ignore it in their later work? Why did Mead and Dewey ignore the obvious importance of shame in Cooley? Why did Brown and Levinson recognize the importance of Goffman’s concept of face as the avoidance of embarrassment, but fail to utilize it in their empirical studies? My description of the history of shame studies by psychoanalysts suggests many similar questions, particularly Freud’s early discovery of shame and his later disavowal.

My explanation derives from Elias’s idea of the advance of the shame threshold, and Lewis’s work on unacknowledged shame. Elias's response to his data led him to an analysis of the underlying process in our civilization that was too advanced for his audience. In Western societies, as Elias pointed out, the threshold for shame has been decreasing for hundreds of years, but at the same time awareness of this emotion has been declining. As his own analysis could have predicted, in our era the level of awareness of shame is so low that only those trained to detect unacknowledged shame could understand the point that Elias was making. Because Retzinger and I were guided by Lewis's (1971) work, we were responsive to Elias's shame analysis. Within psychology and psychonalysis, Lewis’s work is widely acclaimed but seldom used.

The development of a concept of shame, which includes both analytical and operational definitions of shame, is crucially important for the scientific study of shame. It would appear that subjects’ testimony about shame states, and indeed the presence or absence of any other emotion, may not be valid. Perhaps most emotional states are disavowed or exaggerated. Following, Lewis, it would appear that most shame states are not experienced in consciousness, but are either unconscious or misnamed (bypassed or overt, undifferentiated shame, in Lewis’s (1971) terminology). For this reason studies that rely on testimony of subjects, rather than analysis of their behavior and their discourse, are apt to leave out most shame. It is also not clear that subject’s reports of their own shame and that of others accurate. Studies are needed to test the validity of subjective reports of shame. In my view, such a test would mean validating standardized shame measures against analysis of discourse.

To continue to develop, the sociology and psychology of emotions should follow up the leads offered by the authors reviewed here. If we could agree upon a method for studying shame that would be reliable and valid, we might start by testing the key hypotheses on collective shame Elias stated: shame is increasing in modern societies, but at the same time awareness of shame is decreasing. Another hypothesis, following Sennett and Cobb, is that members of the working and lower classes are shamed by their status.

One direction that a sociology of emotions might take concerns the dynamics of racial, gender, ethnic, and class relationships. In her chapter "Honor and Shame", Howard (1995) proposes that women and blacks are likely to be ashamed of themselves. She suggests that they are dishonored, that their status is consistently derogated. To coordinate their actions in a white male dominated society, women and blacks must take the role of white males, which leads to seeing themselves as they are seen. She supports this idea by pointing to the amount of "self-mutilation" that women and blacks undergo in attempting to fit themselves into the male or white ideal. She argues that women’s sustained attempts to be slender and have small waists and feet, to the point of self-starvation, suggest shame in these women. Similarly, she proposes that hair straightening and the high status of light skin among blacks has the same implication.

Howard’s analysis of shame and honor in race and gender relations is suggestive, but is only a first step. If her formulation is accurate, it would mean that there is an emotional/relational structure that sustains the domination of white males, in addition to legal, political and economic causes. In order to test this idea, however, shame would need to be investigated so that its presence or absence in women and blacks could be documented. Retzinger’s theory of conflict (Retzinger 1991) and my application of it to collective conflict (Scheff 1994) suggest that protracted and intense hatred, resentment, and envy are all products of unacknowledged shame. Research on gender, race, ethnic and class emotional tensions and alienation could be inspired by this idea.

As indicated at the beginning of this article, the classic sociologists believed that emotions are crucially involved in the structure and change of whole societies. The authors reviewed here suggest that shame is the premier social emotion. Lynd’s work, particularly, suggests how acknowledgement of shame can strengthen bonds, and by implication, lack of acknowledgment can create alienation. Lewis’s work further and in much more detail suggests how shame/anger loops can create perpetual hostility and alienation. Acknowledged shame, it seems, could be the glue that holds relationships and societies together, and unacknowledged shame the force that blows them apart. Since we are now in a position to clearly define shame as a working concept, perhaps the time has come to begin systematic empirical studies of the effect of individual and collective shame on social solidarity and alienation.

References

Brown, Penelope, and Stephen Levinson. 1987. Politeness: Some Universals in Language Usage. Cambridge: Cambridge U. Press.

Broucek, Francis. 1991. Shame and the Self. New York: Guilford.

Collins, Randall. 1990. Stratification, emotional energy, and the transient emotions. In T. D. Kemper (Editor), Research Agendas in the Sociology of Emotions, pp. 27-57. Albany, NY: State University of New York Press.

Cooley, Charles H. 1922. Human Nature and the Social Order. New York: Scribner’s.

Darwin, Charles. 1872. The Expression of Emotion in Men and Animals. London: John Murray.

Ekman, Paul, P. Ellsworth, and W. Friesen. 1972. Emotion in the Human Face. New York: Pergamon

Elias, Norbert. 1978, 1982, 1983. The Civilizing Process: V. 1-3. New York: Pantheon.

1994 The Civilizing Process. Oxford: Blackwell

1996. The Germans. Cambridge: Polity

Elias, N. and John Scotson. 1965. The Established and the Outsiders. London: Frank Cass.

Erikson, Erik. 1950. Childhood and Society. New York: Norton.

Freud, Sigmund and J. Breuer. 1895. Studies on Hysteria. New York: Avon (1966).

Goffman, Erving. 1967. Interaction Ritual. New York: Anchor.

Gottschalk, Louis. 1995. Content Analysis of Verbal Behavior. Hillsdale, NJ: Lawrence Erlbaum Associates.

Gottschalk, Louis, Winget, C. and G. Gleser. 1969. Manual of Instruction for Using the Gottschalk-Gleser Content Analysis Scales. Berkeley: UC Press.

Hochshild, Arlie. 1983. The Managed Heart. Berkeley: U. of California Press.

Horney, Karen. 1950. Neurosis and Human Growth. New York: Norton.

Howard, Rhoda E. 1995. Human Rights and the Search for Community. Boulder: Westview

Kardiner, Abraham. 1939. The Individual and His Society. New York: Columbia U. Press.

Kemper, Theodore. 1978. A Social Interactional Theory of Emotions. New York: Wiley.

Lansky, Melvin. 1992. Fathers Who Fail: Shame and Psychopathology in the Family System. Hillsdale, NJ: Analytic Press.

1995. Posttraumatic Nightmares. Hillsdale, NJ: Analytic Press.

Lewis, Helen B. 1971. Shame and Guilt in Neurosis. New York: International Universities Press

Lynd, Helen M. 1961. On Shame and the Search for Identity. New York: Science Editions.

Mead, G. H. 1934. Mind, Self, and Society. Chicago: U. of Chicago Press.

Morrison, Andrew. 1989. Shame: the Underside of Narcissism. Hillsdale, NJ: Analytic Press.

Parsons, T. and Shils, E. l951. Toward a General Theory of Action. Cambridge: Harvard U. Press.

Retzinger, Suzanne. 1991. Violent Emotions. Newbury Park: Sage.

1995. Identifying Shame and Anger in Discourse. American Behavioral Scientist 38: 1104-1113.

Scheff, T. l990. Microsociology. Chicago: U. of Chicago Press.

1994. Bloody Revenge: Nationalism, War, and Emotion.

Scheff, T., and S. Retzinger. l991. Violence and Emotions. Lexington, Mass: Lexington Books.

Sennett, Richard, and Jonathan Cobb. 1972. The Hidden Injuries of Class. New York: Vintage Books.

Sennett, Richard. 1980. Authority. New York: Alfred Knopf.

Simmel, Georg. 1904. Fashion. International Quarterly X: 130-55. (Reprinted in the American Journal of Sociology 62: 541-559)

Von Raumer, W. l857. Education of Girls. (cited in Elias, l978). 10, 040 words