Scheff and Retzinger – Shame, Anger, and the Social Bond: A Theory of Sexual Offenders and Treatment


Thomas J. Scheff Professor Emeritus University of California, Santa Barbara

[email protected]

Suzanne M. Retzinger Family Relations Mediator Superior Court Ventura, CA.

[email protected]


This article proposes a new approach to male sexual offenses: a theory of the relational-emotional basis of sexual compulsion, and a treatment procedure that deals directly with emotions and relationships. There are two central hypotheses: 1. Offenders have no secure social bonds. 2. The offenderís extensive unacknowledged shame takes the direction of compulsive assaults on women. To underline the meaning of social bonds and shame, we review the literature on these two related topics. Assuming for the sake of argument zero bonds and shame outside of awareness as root causes of sexual compulsion, we outline an approach to treatment that increases awareness of social bonds, uncovers hidden shame, and decreases the arousal of shame in the offenderís relationships with others.

There is a large literature on the causes of sexual assault and the treatment of sex offenders, summarized in Marshall et al (1990) and Hall et al (1993). Both summaries make it clear that there is no agreement on the causes of sexual assault, nor on its treatment: there is no theory that explains sexual assault, and no treatment that has been effective in treating sex offenders. Here we propose that the main reason for their inconclusiveness is that these studies have not dealt directly and extensively with relationships and emotion, which we take to the main elements in causation and cure. Instead, they focus on behavior, thoughts, and beliefs.

However, there is one reoccurring finding which involves the offenderís emotions in an indirect way. Studies have repeatedly shown that sex offenders seek to humiliate their victims, as summarized in Darke (1990). Typically, the authors of these studies interpret the intent to humiliate as an interest in power. The theory of shame-rage loops, to be described below, suggests a more elaborate interpretation: being ashamed of themselves in general, assaultive men are also ashamed of their sexual desires. Experiencing women as haughty and rejecting, sex offenders reject the women they see as rejecting them. Feeling humiliated and powerless (a shame state), they humiliate and dominate in return. But since their shame is unacknowledged, it leads to shame-rage loops that produce compulsive violence against women.

Since unacknowledged shame is common among human beings, it is necessary to construct a more specific theory that formulates the elements that produce sexual aggression. As already indicated, one element is the inability to deal with recurring shame. To explicate that idea:

  1. Sex offenders are hypothesized to be quick to take offense or feel insulted and humiliated.
  2. They are unable to extricate themselves from continuous loops of shame.
  3. For offenders, these loops do not take the form of being ashamed of being ashamed (shame-shame loops), which lead to withdrawal and passivity, but shame anger, which lead to continuous humiliated fury, the emotional basis for contempt and hatred.
  4. Finally, these men have no secure social bond to which they can turn to share their pent-up feelings; they are in the zero bond condition. This combination, continuous shame-anger loops and zero bonds, will produce either madness, suicide-homicide, or sexual assault. Since all of the elements in this formulation are counter-intuitive, it will be necessary to review each in turn.

A Theory of Social Bonds

Since sexual assault often takes the form of gang rape (Dark 1990), one would assume that at least offenders in this category would have strong social bonds with their gang. But we propose that gang bonds are not secure, indicating true solidarity, but based on unquestioned loyalty to the gang and to its code of behavior. We argue that a social relationship can be so close and demanding as to be suffocating, and therefore not a secure bond, just as a relationship can be distant and rejecting, also not a secure bond. Gang members, by this definition, can be just as alienated as lone individuals are from others.

We believe that a secure bond strikes a balance between being too close (engulfment) and too far (isolation). Following Eliasís (1987) idea of the “I-We” balance, the three states of the bond can be identified by the disposition of pronouns in discourse.

To give the idea of the I-We balance an empirical basis, language can be an indicator of the state of the social bond between two people, by focusing on the use of pronouns, particularly I, you, we, and it. The disposition of these pronouns within a sentence, and the relative weight accorded them, can be used as cues to three different states of the bond – solidarity, and the two opposite forms of alienation, isolation and engulfment. As discussed below, this analysis can be backed up with a study of emotion cues, showing how pride cues signal solidarity, and shame cues signal alienation.

Our approach draws upon and overlaps with Buber’s (l958) discussion of I-thou, and many other formulations. What we call solidarity language, (I-I) corresponds exactly to his I-thou. What we call the language of isolation (I-you) corresponds exactly to his I-it. We use different terms because Buber, like most philosophers and social scientists, did not consider the other form of alienation, what Bowen (l978) called fusion (me-I). The idea of engulfment is centrally important in family systems theory (sometimes called emeshment or fusion) but is absent elsewhere in the human sciences. Social scientists usually confound engulfment with solidarity. In engulfed relationships, one or both parties subordinate their own thoughts and feelings to those of the others(s). In solidarity, each party recognizes the sovereignty of the other, but balances respect for the otherís position with respect for oneís own.

Our use of I and me is quite different than Mead’s (l934). His social psychology seems to assume perfect solidarity, without considering the possibility of alienation. In Mead’s scheme, the me is made up of the internalized representation of the roles of others. For example, the citizen utilizing a criminal court is prepared by already knowing the role of the judge, the jury, the policeman, jailer, etc.

Mead didnít consider the accuracy with which each member of a society knows the roles of the other members. By ignoring this issue, he evades the issue of imperfect relationships, of alienation. Consider the doctor-patient relationship. Obviously the patient has only a superficial knowledge of the doctor’s role, and superficiality of her knowledge can cause impediments to cooperation. For instance, since the patient understands very little of what the doctor knows of the relationship between the patient’s illness and the medication that the doctor has prescribed for treatment, the patient might fail to follow the doctor’s orders. The relationship is asymmetric in this way.

The relationship is also asymmetric the other way round. Although the patient has never learned the role of the doctor, the doctor should certainly know the role of the patient, since she has been a patient before becoming a doctor, and will continue to be a patient as a doctor. We would expect, therefore, that the doctor would understand patients. But as it turns out, there are impediments to such understanding. As part of their training, and as part of their management of their roles, many doctors seem to “forget” the patient’s experience; they do not understand their patients, not because of lack of knowledge, but because of emotional barriers which doctors erect against them.

This process of forgetting also occurs in the teacher’s role. Most teachers call very little upon their own role as students to guide their teaching. Rather their teaching seems to conform to the way other teachers teach. Once we have learned a body of knowledge, we seem to repress the difficulties we had in learning it, which erects a wall between us and our students. In the language we will use here, in our role as teachers, we are engulfed with other teachers, and isolated from our students. Bi-modal alienation, as we call it, seems to be the most common form of social relationship in the modern world (Scheff and Retzinger, 1991; Scheff, 1994). In this format, individuals or groups are engulfed within, isolated without, as in sects, cults and academic schools of thought. In the social sciences it has been customary to refer to this system of relationships in terms of ethnocentrism, but this concept is imprecise and static.

The theoretical approach to social integration most similar to the one advocated here is found in Elias’s (Introduction, l987) discussion of the “I-self” (isolation) and the “we-self” (engulfment). Elias discusses the “I-we balance” (solidarity) in a way that is quite similar to our usage, but he doesn’t apply it to actual instances. Elias proposed a three-part typology of what he called social figurations: independence (lack of cooperativeness because of too much social distance), interdependence ( a balance between self and other that allows for effective cooperation), and dependence (lack of cooperativeness because of too little social distance). Although this typology grows out Eliasís long-standing commitment to the concept of interdependence, his attention was usually limited to this concept, rather than to the other two components of the typology. In this paper, we give equal theoretical and empirical weight to all three types.

The concept of the I-We balance is oriented largely to the verbal components of interaction, ignoring the non-verbal. An approach devised by Retzinger and I (Scheff and Retzinger, 1991; Scheff, 1994) is oriented to both verbal and nonverbal elements. We argue that the emotions of pride and shame are directly related to the state of social relationships, and can serve as indicators of the nature of that relation from moment to moment. In our scheme, pride is the emotional cognate of a secure, unalienated bond, and shame signals threat to the bond.

It would appear that careful analysis of the state of social relationships might require using both the I-We and the pride-shame approaches. The first is simpler, the second more subtle and complex, requiring use of a technique developed by Retzinger (1991) to identify cues to shame. The limitation to determining the I -We balance is that human beings often use language in a less than straightforward way, disguising their own motives and goals. Word choice is largely voluntary and intentional; we can easily hide our motives and goals in our verbal expressions. Since shame signals are, for the most part, stereotyped and out of awareness, the analysis of nonverbal cues can detect strains in relationships even when they are covert and hidden. The two approaches can be used to complement each other in determining the state of the bond.

Eliasís scheme comes closest to what seems to be needed, if we are to connect the micro and macro-worlds. Many other theories are closely related to this scheme, but contain only two of the three types. Durkheim (1897), Doi (1971) and Seeman (1975) considered only the two forms of alienation that we are calling isolation and engulfment, but didnít relate them to the third, solidarity. Buberís typology contains isolation and solidarity, but not engulfment, as do Marx and Braithwaite (1989).

Braithwaiteís theory of reintegrative shaming in crime control (1989), which he has applied to the practice of community “conferences” in Australia, excludes one of the types of alienation, engulfment. One case that we observed when consulting in Australia illustrates engulfment, rather than reintegrative shaming (solidarity) or stigmatization (isolation). The offender was an adult male in a case of driving under the influence, with three previous arrests on this charge. With the help of most of the other participants (except for the arresting officer), he successfully denied his responsibility. Even the facilitator, a male, colluded with his defense, that because of the extenuating circumstances in this instance, drinking a six-pack, apparently a norm among working class men, and driving was not morally wrong. Rather than being overshamed, often the case with juvenile offenders, this offender was undershamed, because of engulfment between him and most of the other participants in the conference. This outcome was unusual, in that the outcome of most of the conferences is closer to reintegrative shaming than to either engulfment or isolation (stigmatization).

All three types of relationships are implied in the work of Satir (1971), who contrasts defensive postures that we would call isolated (blaming, computing) or engulfed (placating, distracting ) with what she calls leveling. By leveling she seems to have meant communication which is direct but respectful. Being direct without being respectful involves isolation, just as being respectful but not direct implies engulfment, as we use these terms here. Leveling involves balancing the claims of the individual with those of the relationship, an I-We balance as we have suggested.

The scheme proposed here suggests that most earlier discussions of social alienation and solidarity and related concepts are misleading, since they ignore one of the three type of relationships described above, or confound one with another. Most classic Western studies which compare Asian and Western societies seem to have valorized Western isolation and individualism by confounding it with solidarity based on rational outcomes (Durkheimís [1895] description of organic solidarity), as Markus and Kitayama (1991) have charged. But from the point of view proposed here, it seems likely that Markus and Kitayama have made an equal and opposite error, valorizing Asian (unity-based) societies by confounding engulfment with interdependence. A similar confusion seems to exist in classical studies of autonomy, which feminist scholarship has shown to be male oriented, or in Eliasís terminology, oriented toward independence rather than interdependence. But again, most feminist scholarship may be making the equal and opposite error, confounding engulfment with interdependence or solidarity.

Shame as the Master Emotion

Because of limitations of the concept of shame in Western societies, it is difficult to convey the importance of shame dynamics for understanding both normal and pathological behavior. In earlier publications (Scheff 1990, Retzinger 1991, Scheff and Retzinger 1991, and Scheff 1994), we have argued that shame is subject to extensive repression in modern societies. This repression is both caused by and gives rise to rampant individualism.

These ideas are documented in Eliasís magisterial study (1978, 1982, 1983) of the historical development of modern civilization in the West. Elias used excerpts from European advice manuals over the last five hundred years to show the gradual but implacable repression of shame. (For a detailed discussion of one of Eliasís examples, 19th century advice to mothers on controlling their daughterís sexual knowledge through shame, and shame about shame, see Scheff and Retzinger 1991, pp. 10-12.) The idea that one can be ashamed of being ashamed leads to the concept of continuous loops of shame, which is our explanation of the mechanism of repression.

One obvious indicator of repression is the difference between the treatment of shame in the languages of modern and traditional societies. In European languages, and especially in English, the concept of shame is extremely narrow, and extremely negative. In the English language shame has the meaning of disgrace and profound emotional pain. But in other European languages, there is both a shame of disgrace (as in the French honte), and a positive, everyday shame (as in the French pudeur), which refers to modesty, shyness, and (at least in classic Greek) awe. The possibility that there was once a positive shame in English is suggested by the word humility (because of its relation to humiliation), but humility has lost its relationship to shame in modern English.

The narrowness and negativeness of the concept of shame in modern societies is still more strongly suggested by comparison with the shame lexicon in the languages of traditional societies. It has been shown that the shame lexicon in Mandarin Chinese is much larger than those of modern societies (Shaver et al 1992). The Mandarin emotion lexicon also contains a large number of shame-anger combinations unknown in English, but parallel to the Scheff-Retzinger usage (see below). The shame lexicon is rich in traditional societies because its members are sensitive to social relationships, requiring awareness of shame and embarrassment, as contrasted with the individualism of modern societies. The small and narrow lexicon of shame in Western languages, especially English, suggests that most forms of shame are being overlooked or avoided.

In order to understand shame dynamics, it is necessary to recover the positive facets of shame, and to recover the breadth of the shame concept from the maws of repression and silence. The positive aspects of shame have been explored by Lynd (1958), Tompkins (1963 V. 2) and in great detail, by Schneider (1977). Here is a representative statement (Lynd 1958 p. 66): “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 others….” The idea expressed in this passage is crucially significant for our theory of treatment: if the offender can come to the point of “sharing and communicating” his shame instead of hiding or denying it, the damaged personality of the offender can begin to be repaired.

Recovering the actual breadth of the shame concept is also of crucial importance for repairing the damaged bond between the offender and others, because shame is a Protean presence among all of the relevant parties. The offender will be ashamed because he has been publicly condemned for wrongdoing. The offenderís supporters will be ashamed because of their relationship to him. The victim will be ashamed in the sense of feeling betrayed, violated, and/or impotent. The victimís supporters, in so far as they identify with her, will share this kind of shame. The public at large will be inclined to shame or humiliate the offender, seeking not only justice but also, in some cases, revenge.

So long as shame is disguised and denied, it inhibits the participants from repairing the bonds between them. The vast domain of shame was discovered in modern psychology by Helen B. Lewis, a psychoanalyst and research psychologist. In her ground-breaking study (1971) she found shame to be omnipresent but almost never mentioned in psychotherapy sessions. Instead it was ignored, disguised, or denied. The sociologist Goffman (1967) implied that shame (he called it embarrassment), actual or anticipated, was a haunting present in all human contact, not just in psychotherapy. According to Goffman, social relationships are haunted not just by actual embarrassment, but much more frequently, by the anticipation of embarrassment. Retzinger (1991) has developed a systematic procedure for identifying anger, shame and embarrassment, no matter how hidden or disguised, by reference to visual, verbal, and nonverbal cues. Members of modern societies require retraining in shame language, especially the language of gesture and innuendo, in order to become aware of shame in themselves and others.

In our earlier articles and books (Scheff 1990; Retzinger 1991, Scheff and Retzinger 1991, Scheff 1994), we have treated shame as a large family of emotions and affects, drawing upon the work of Lynd (1958) Tomkins (1963), Goffman (1967), Lewis (1971) and many others. Our definition includes the positive aspects of shame, such as what Schneider called “a sense of shame”, as well as embarrassment, humiliation, shyness, modesty, and feelings of discomfort, awkwardness, inadequacy, rejection, insecurity and lack of confidence. In our usage, we give particular emphasis to shame/anger sequences, as in Mandarin Chinese. Our concept is much, much broader than the way shame is used in vernacular English. We have reason to believe that our definition of shame recovers the broadness of the concept that is in use in traditional societies.

The most detailed treatment of shame in traditional societies can be found in a discussion of whakamaa, the conception of shame in Maori society (Metge 1986). According to this study, whakamaa means shy, embarrassed, uncertain, inadequate, incapable, afraid, hurt, depressed, or ashamed (p. 28-29). Only the inclusion of afraid (fear) would seem to differ from our usage. But the examples that Metge uses for afraid suggest not the emotion of fear (danger to life or limb), but social fear, that is, anticipation of shame. As in our usage, whakamaa concerns not only feelings, but also relationships. Maori usage, like ours, also stresses the importance of acknowledging shame, and the disruptive consequences when shame is not acknowledged. Although we did not know about whakamaa when we developed our conception of shame, it is exactly parallel to Maori usage.

Most traditional societies, like the Maoris, have a subtle and wide-ranging language of emotions and relationships which we in the West have lost. The understanding of shame in its positive, broad, and relational sense might be a crucial issue in understanding the causation and treatment of sex offenders.

A theoretical issue of great importance for this article is the difference between pathological and normal shame. Like Braithwaite (1989), we think that effective crime control requires normal (re-integrative) rather than pathological shame. By paying close attention to the particular way shame is manifested, it is possible to distinguish, moment by moment, between the two forms of shame as they occur in social interaction.

According to our theory, manifestations of normal shame, although unpleasant, are BRIEF, as little as a few seconds. Shame, anger, and other related emotions which persist for many minutes are pathological. We propose that shame is a highly reflexive emotion, which can give rise to long lasting feedback loops of shame: one can be ashamed of being ashamed, and so on, around the loop, resulting in withdrawal or depression. Another loop is being angry that one is ashamed, and ashamed that one is angry, and so on around the loop. Furthermore, shame-anger loops can occur between, as well as within, participants. Indignation can be contagious, resulting in mutual and counter-indignation. Both individual and social emotional loops can last indefinitely. Persistent, relentless emotions such as continuing embarrassment, indignation, resentment and hatred are always pathological.

Finally, we argue that shame plays a crucial role in normal cooperative relationships, as well as in conflict. We show that shame signals a threat to the social bond, and is therefore vital in establishing where one stands in a relationship. Similarly, pride signals a secure bond. Shame is the emotional cognate of a threatened or damage bond, just as threatened bonds are the source of shame. This equation allows one to translate shame language into relationship language, and vice versa.

If, as Goffman (1967) and others have argued, normal shame and embarrassment are an almost continuous part of all human contact, we can see why the visible expression of shame by the offender looms so large in symbolic reparation. When we see signs of shame and embarrassment in others, we are able to recognize them as human beings like ourselves, no matter the language, cultural setting, or context. The central role of shame in human contact has long been recognized in the scientific-humanist tradition, as expressed by Darwin, Nietzsche, Sartre, and many others. We think that the difficulty in overcoming the Western views of shame is the principle impediment to success in understanding and treating sex offenders.

The Theory of Shame-Rage Loops

In her study of shame in psychotherapy, Lewis (1971) uncovered a mechanism linking unacknowledged shame to anger and verbal aggression. Using hundreds of transcripts of psychotherapy sessions, she applied the Gottschalk-Glaser scale (1969) to the words in each transcript for emotions. Although her study yielded other emotions such as grief and fear, the only pattern she found to repeat itself in all sessions was shame followed by anger and aggression. When a patient felt distant from, rejected or criticized by the therapist, or inadequate, rather than mentioning the feeling (shame or embarrassment), the patient would either withdraw or show anger toward the therapist. Since the shame that was elicited by the coding procedure was seldom explicitly mentioned by the patient or the therapist, Lewis called it unacknowledged shame.

Lewis proposed that the patientís shame could take three principal routes. It could be discharged harmlessly through discussion or laughter, it could lead to withdrawal or termination of therapy, or it could lead to verbal aggression (Lewis 1987) . Because of this last route, Lewis suggested that there is a strong affinity between shame and anger.

In her study of marital quarrels, Retzinger (1991) investigated the affinity between shame and anger reported by Lewis. Using the Gottman (1979) method for eliciting quarrels, Retzinger analyzed the emotions in videotapes of four martial quarrels. She found that unacknowledged shame always preceded anger in the 16 escalations of the quarrels she studied, confirming Lewisís earlier study.

On the basis of the Lewis, Retzinger, and other studies, Scheff and Retzinger (1991) proposed a theory of shame/rage loops leading to verbal or physical violence. We suggested that when shame is not acknowledged and discharged, it refuses to subside; one can be ashamed of being ashamed, and so on, a shame-shame loop, which leads to withdrawal. But an alternative route is a shame/anger loop: one can be angry one is ashamed, and ashamed that one is angry. This is the loop which occurred most frequently in Lewisís (1971) and in Retzingerís (1991) studies. Tracing back from anger or angry escalation, these studies invariably found an incident of unacknowledged shame.

The idea of emotion loops provides new meaning to the familiar idea that pathological emotions are reactive, rather than being primary emotions (Greenberg and Safran 1987). Most theories of emotions in psychotherapy have long suggested that anger is usually not the primary emotion in quarrels, but is a reaction to some other emotion. Such theories, however, have not been specific about the name of the underlying primary emotion, referring to it by a generic name like “hurt.” In our theory we propose that the major component in “hurt” is shame, that is, feelings of rejection and/or inadequacy. Nor have earlier theories specified the mechanism relating primary and reactive emotions, or pointed to empirical indicators. This article seeks to fill in these missing elements.

Drawing upon her findings, Lewis (1971, 1976, 1987) devised a type of shame therapy. When she detected anger toward herself, she encouraged the patient to first acknowledge their anger, then trace back from it to see if they could find a moment of shame or humiliation. Her patients discovered that instead of revealing their feelings of inadequacy or humiliation, they were instead masking them by verbally attacking the therapist or withdrawing. Apparently feelings of humiliation, even in therapy, are so painful that they are usually not revealed, even to self. Instead the humiliated person often seeks to humiliate the person who they feel humiliated them: humiliation leads to counter-humiliation.

The humiliation-counter-humiliation sequence is well-known in the world of everyday life. Most sports coaches, for example, seem aware of it. When oneís team is winning by a large margin, the knowing coach will substitute heavily, not only to give the second and third teams practice, but also to avoid, if possible, “rubbing in” the defeat to the point that the other coach plots revenge. The defeated coach may experience an overwhelming defeat as an insult, and vow to insult the other coach in return, giving rise to shame/anger sequence within and between the coaches. Shame/anger loops may also be the emotional basis for institutionalized conflict between individuals and between groups, as in duels, feuds, vendettas, and wars (Scheff and Retzinger 1991; Scheff 1994).

Causation of Sexual Assault: A Theory

To understand the role of shame in sexual assault, it will be necessary to briefly review relevant aspects of the shame literature. One major finding of gender differences in emotionality involves studies of what has come to be called “field dependence and independence” (For a summary of some of the findings selected from a very large corpus, see Wapner and Demick 1991). It was Lewis (1971; 1976) who called attention to emotional elements in what until then had been construed as a perceptual phenomenon.

The concept of field dependence grew out of the work on conformity by Asch and others. Persons who yielded to group opinion in making perceptual judgments (such as lengths of lines) were said to be field dependent. Those who did not yield, field independent. Lewis (1971; 1976) showed a correlation between field dependence and the experience of shame. Those who were field dependent also showed an inclination toward a certain type of shame, which Lewis called overt, undifferentiated shame. This type of shame involves the experience of emotional pain and unwanted bodily symptoms, such as sweating, blushing and rapid heartbeat. This type corresponds to the vernacular meaning of a shame experience, which is painful, visible, and conscious.

But in the same study (Lewis 1971) also uncovered a different, less overt type of shame, which she called bypassed. This type is completely outside of awareness, but is indicated by repetitive, obsessive thoughts and rapid, slightly off-key speech. Lewis identified this emotion as shame because it occurred in the kind of context which sometimes resulted in overt shame, the patient experiencing the therapist as distant, critical, or rejecting.

Lewis showed that bypassed shame was characteristic of persons who were field independent. The relevance to this article is that she and others also showed a strong relationship between gender and field dependence: women strongly tend toward field dependence, men toward field independence. If, as Lewisís studies suggest, men tend toward bypassed shame, women toward overt shame, these findings would have implications for differences in male and female sexual shame.

The concepts of overt and bypassed shame correspond closely to central aspects of Adlerís (1956) theory of personality. Adler thought that if children were denied love during critical stages of maturation, they would develop either one of two types of adult personality. The first type he called the inferiority complex, which seems to be an exact parallel to what we would call chronic overt shame. The second direction he called the drive for power, which parallels what we are calling bypassed shame. His whole theory implies a shame dynamic, since we expect that children to whom love was denied would suffer from feelings of rejection (shame).

In the language of our theory, overt shame is correlated with shame-shame loops, of being ashamed of being ashamed. This idea is exemplified by shy people who blush: when they become aware they are blushing, they experienced a self-perpetuating self-consciousness: they are ashamed of being ashamed, and so on. Since women tend more than men to experience shame in the overt mode, we would expect to find women tending toward shame-shame loops more than men. This loop is associated with passivity, shyness or withdrawal. Real or imagined distance, criticism or rejection by men may be expected, therefore, to lead to passivity or withdrawal in shame-prone women.

Men are more likely than women to experience shame in the bypassed mode. This mode is much less available to consciousness than the overt mode, and associated with shame-anger loops. Men may bypass the painful experience of shame by becoming angry. But they are ashamed of their anger (in the bypassed mode), and angry that they are ashamed, and so on. Unacknowledged shame in this mode gives rise to what Lewis (1971) called “humiliated fury.”

Male and Female Sexual Shame

Although there are few empirical studies, we will assume that shame about sexuality is quite prevalent in modern societies, and that it is particularly obvious in adolescents and young adults. In his study of the civilizing process, Elias (1978, 1982) used excerpts from European advice and etiquette manuals from the 12th through the 19th century to show that there has been a rapidly accelerating rise in shame in many areas of life, especially shame about the body, emotions, and violence. He specifically points to sexuality as one of the areas of increasing shame (1978, 169-190). It seems clear that virtually all children in modern societies learn to be embarrassed and ashamed about their bodies and body functions, including their sexuality. What happens to their sexual shame as they grow to adults?

Again following Eliasís lead, we propose that sexual shame does not disappear for most adults, but goes underground. Elias (1978, 180-181) suggests that women are especially shamed into silence about their sexuality. Contrary to Elias, who did not have the concept of bypassed shame available to him at the time he wrote (the original text was written in 1939), we argue that most adults continue to be ashamed of their sexuality, but that men and women manage their shame differently. The theory of shame loops, in conjunction with the studies of gender differences described above, suggests that there should be a substantial difference between the way in which men and women experience shame about sexuality.

Our theory would help to explain differences in male and female sexuality. If women tend toward shame-shame loops in reaction to their shame about sexuality, this process would explain the direction that sexuality often takes with women: lack of sexual interest, withdrawal, passivity, or late-blooming interest, since shame-shame loops lead to silence and withdrawal.

Shame about sexuality would lead men in a different direction, however. Since shame-anger loops are associated with the way in which men tend to manage shame, through bypassing (denial, which, in psychoanalytic terms, leads to a reaction formation), we would expect male sexual shame to lead to boldness, anger and aggression. The shame-anger loop can be seen as the emotional basis of machismo, the hypermasculinity that is explicitly named in Spanish-speaking societies, but is prevalent in most current and past cultures. To the extent that male shame is bypassed in any society, we would expect a generalized shame-proneness, manifested by sensitivity to insults, real or imagined, hostility and aggressiveness. Although counter-intuitive, the theory of shame loops provides a testable theory of male aggressiveness, which posits that machismo rests upon personal and social insecurity, which translated into our language, means shame-proneness.

There is some indirect evidence that links sex offenses with insecurity (shame) about oneís masculinity. This evidence comes from a series of studies of the self-esteem of sex offenders. Before discussing it, some comments are needed concerning the measurement of self esteem.

The measure that comes nearest to uncovering emotions are self-esteem scales. It seems likely that high-self esteem is based on persistent pride in self, and low self-esteem, on persistent shame. But these scales do not take into account the subjectís defenses against their emotions (Scheff et al, 1989). All self-esteem scales confound exaggerated confidence in self as a defense against shame, with true self-respect.

There is one series of studies of rapists which strongly support this inference (Marshall and Marshall 1981, Marshall and Turner 1985, and Lawson et al, 1979). In their investigation of the self-esteem level of rapists, these studies found that most of them scored either very low or very high. But these investigator then took a step that is quite unusual in self-esteem studies, they sought external corroboration of self-esteem levels from persons who knew the subjects.

For those rapists who scored well above the mean of non-offenders, we had prison staff who were acquainted with them provide their evaluations of the offenders using the same scale. These staff reported that high-scoring offenders were low in self-esteem, but compensated for that with a bravado style that reflected hyperidentification with the traditional male role (Lawson, et al 1979)

This study suggests support for out argument, that the most powerful component of self-esteem is shame and pride, and that existing self-esteem scales confound exaggerated self confidence with pride. The exaggerated masculinity and aggressiveness of sex offenders may be produced by unacknowledged shame.

A Theory of the Cause and Treatment of Sex Offenders

The discussion above leads to the following propositions:

1. The compulsion to assault women is a product of unacknowledged shame about sexuality, which has taken the path of shame-rage loops. Men, much more than women, tend to bypass their shame, making it less accessible to consciousness. Bypassed shame about sexuality is the hidden cause of sexual compulsions.

2. A treatment procedure based on this theory would seek to interrupt hidden shame cycles, bring them into consciousness, and to teach social skills which would avert the generation of new shame.

a) Uncovering hidden shame. As already indicated above, the technique that Lewis (1971; 1977; 1987) used to uncover shame in psychotherapy was to first note the occurrence of anger and hostility, no matter how subtle, toward herself, the therapist. She then encouraged the client to first acknowledge the hostility, then trace it back to earlier events. Invariably she found that the patient had experienced feelings of rejection or inadequacy earlier, which were expressed however, as hostility. (This mechanism may lead to male hostility to women, discussed above: the man feels rejected by women, does not acknowledge this feeling, but rejects them in retaliation). Lewis then helped the client to explore the feeling of shame or embarrassment verbally, leading to the uncovering and discharge of shame.

The latter step, the actual experiencing of shame during verbal discussions of it, requires the most effort on the part of the client and therapist. Detection of hostility by the therapist is straightforward. The client usually needs to learn to trace back from anger to shame, requiring patience from both participants, but most clients learn the skill. Finding and experiencing the clientís actual feeling of shame through discussion may require considerable cooperation between patient and therapist. (For other approaches to shame therapy, see Broucek 1991; Goldberg 1991; Kaufman 1980 and Wurmser 1981).

To apply the type of shame therapy developed by Lewis to male sex offenders, it would probably be focused on their unacknowledged shame in general, as well as the emotions aroused by the clientís sexuality, and by women. If a man feels ashamed of his sexuality and/or rejected by, or inadequate with women, but does not acknowledge these feelings even to himself, a likely outcome is the kind of hostility toward women that can lead to assault. If the buried feelings of sexual rejection or inadequacy can be uncovered in therapy, they can be experienced and discharged.

b) Communication skills. If the theory outlined here is correct, sex offenders need to learn new skills in communicating with others, particularly women. Friction in close relationships is unavoidable, including feelings of inadequacy and/or rejection. Learning to level (Satir 1972), to be direct but respectful, with others, particularly with women, can forestall the creation of new shame and hostility, by teaching offenders to acknowledge shame, rather than denial.


This article proposes a new theory and treatment of male sexual violence, which focuses on a new idea, that hyperidentification with masculinity is based on carefully hidden shame and insecurity, rather than on true self-respect. The theory leads to a type of treatment for sex offenders leading to the uncovering of repressed shame, and learning of communication techniques which decrease the arousal of shame.


Adler, A. 1956. The Individual Psychology of Alfred Adler. New York: Basic Books.

Bowen, M. 1978. Family Therapy in Clinical Practice. New York: J. Aronson.

Bowen, M. and M. Kerr. 1988. Family Evaluation. New York: Norton.

Braithwaite, J. 1989. Crime, Shame and Reintegration. Cambridge: Cambridge U. Press.

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

Buber, M. 1958. I-Thou. New York: Scribners.

Darke, J. 1990. Sexual Aggression: Achieving Power through Humiliation: 55-72. in Marshal, et al (Eds.), Handbook of Sexual Assault. New York: Plenum

Doi. T. 1971. The Anatomy of Dependence. Tokyo: Kodansha International.

Durkheim, E. 1895. Rules of Sociological Method. London: Macmillan. (1982).

__________ 1897. Suicide. London: Routledge (1952).

Elias, N. 1978. The History of Manners. New York: Vintage.

__________ 1982. Power and Civility. New York: Vintage.

__________ 1987. Involvement and Detachment. Oxford: Blackwell.

Goldberg, C. 1991. Understanding Shame: Theoretical and Clinical Implications. New York: Jason Aronson.

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

Gottman, J. 1979. Martial Interaction. New York: Academic Press.

Gottschalk, L. and G. Gleser. 1969. Manual for the Gottschalk-Gleser Scales. Berkeley: U. of California Press.

Hall, G. C., R. Hirschman, J. Graham, and M. Zaragoza. 1993. Sexual Aggression. Washington, D.C.: Taylor and Francis.

Kaufman, G. 1989. The Psychology of Shame: Theory and Treatment of Shame-based Syndromes. New York: Springer.

Lawson, J. S., Marshall, W. L., and McGrath, P. 1979. The Social Self-Esteem Inventory. Educational and Psychological Measurement, 39: 803-81l.

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

__________ 1976. Psychic War in Men and Women. New York: NYU Press.

__________ 1987. The Role of Shame in Symptom Formation. Hillsdale: LEA

Lynd, Helen. 1958. On Shame and the Search for Identity. New York: Harcourt. NewYork: Plenum

Markus, H. and S. Kitayama. 1991. Culture and the self: implications for cognition, emotion, and motivation. Psychological Review. 98: 224-253.

Marshall, W. L., D. Laws, and H. Barbabee. 1989. Handbook of Sexual Assault. New York: Plenum.

Marshall, W. L. and Marshall, P. 1981. Social Functioning in Penitentiary Inmates. Unpublished MS.

Marshal, W. L. and Turner, B. 1985. Life Skills Training in Penitentiary Inmates in Canada. Report to the Solicitor General of Canada, Ottawa.

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

Metge, Joan. 1986. In and Out of Touch: Whakamaa in Cross Cultural Perspective. Wellington: Victoria U. Press.

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

Satir, V. 1971. Peoplemaking. Palo Alto: Science and Behavior Books.

Scheff, T. 1990. Microsociology . Chicago: U. of Chicago Press.

__________ 1994. Bloody Revenge. Boulder: Westview Press.

Scheff, T. and S. Retzinger. 1991. Emotions and Violence. Lexington: Lexington Books

Scheff, T., S. Retzinger, and M. Ryan. 1989. Crime, Violence and Self-Esteem. in A. Mecca, N. Smelser, and J. Vasconcellos (Eds.). The Social Importance of Self-Esteem. Berkeley: U. of California Press.

Schneider, Carl. 1977. Shame, Exposure, and Privacy. Boston: Beacon Press.

Seeman, M. 1975. Alienation Studies. Annual Review of Sociology. 1: 91-124.

Shaver, Philip, et al. 1992. Cross-cultural similarities and differences in emotions. In M. Clark, (Ed.), Review of Personality and Social Psychology 13: 175-212. Newbury Park: Sage.

Tomkins, Silvan. 1963. Affect, Imagery, Consciousness, V. 2. New York: Springer.

Wapner, S. and Demick. 1991. Field Dependence. Hillsdale: LEA.

Wurmser, L. 1981. The Mask of Shame. Baltimore: Johns Hopkins U. Press.

* This article was earlier turned down by Social Problems as being too psychological, and by the Archives of Sexual Research as being too sociological. Does that mean that these two journals define their fields so narrowly that no overlap with the other will be tolerated?

© 1997 Copyright Electronic Journal of Sociology