The role of the self in shame

The role of the self in shame

Michael Lewis

SHAME, as one of the self-conscious emotions, differs from what have been called the primary emotions because it comes about through self-reflection. While primary emotions require a self to experience the state (Lewis and Michalson, 1983), self-conscious emotions require a self both to produce the state and then to experience it. For example, a loud noise may put me in a state of fright. But to experience this state, I need to be aware of my state of fright. To be in a state of shame I must compare my action against some standard, either my own or someone else’s. My failure, relative to the standard, results in a state of shame. Once in the state, I may or may not experience it. To experience it depends on whether I focus my attention on my state. Again, this requires consciousness. Others have suggested that the state of shame can be produced in a more automatic fashion, one that does not require objective self-awareness.

Consider the case of a 30-month-old child who has a bowel movement in his pants. We could assume that this event automatically causes a state of shame; that is, that there is some connection between the bowel movement and shame. Alternatively, we could think of shame as the consequence of what the child was thinking, either about the accident or about his parent’s response to it. In my work, I have focused on the latter explanation, because I believe that most examples of shame-eliciting events cannot readily be explained by an automatic process. Consider another example, this time involving adults. A man asks an acquaintance who claims to be well read if he has read Joyce Carol Oates’s latest novel. The well-read friend has not read the Oates book, but answers that he has read it in order to maintain his reputation for being cultured. In other words, he lies so as not to be shamed by his lack of knowledge. Clearly, here the event leading to a state of shame has to do with how the second man thinks rather than with some automatic process.

The confusion may reside not in the nature of the process but in the likelihood that some events lead to the state of shame more readily than others, perhaps because they are more likely to elicit specific thoughts than are others. Toilet accidents are more likely than knowledge gaps to lead to disapproval from others. The suggestion of prototypicality with regard to shame-eliciting events must rest on the assumption that certain events lead to a shame state because they are more likely than others to lead to shame-producing thoughts. But in most discussions these factors have remained somewhat confused. For example, Darwin’s argument illustrates the confusion. He believed that blushing could be produced in some automatic fashion by drawing attention to a particular part of the person: attention to the face would cause the face to blush. But Darwin also believed that blushing was caused by how we appear to others–as he put it, “the thinking about others thinking of us … excites a blush” (Darwin, 1965 [1872]: 325).

His observation about blushing and shame indicates his concern with two issues: the issue of appearance and the issue of consciousness. In his model of blushing, the personal appearance of the individual and the consciousness that others were attending to it were the critical elements. Darwin repeatedly made the point that shame depends on sensitivity to the opinion of others, whether good or bad. Thus, self-conscious emotions require the organism’s own sensitivity. Even so, Darwin repeatedly returned to the point that shame is specifically related to external appearance. Darwin did not distinguish between embarrassment and shame, but he did point out that pride, admiration, and even disapproval can cause blushing! He was quite specific in this regard. Somehow, personal appearance, and not moral conduct, was what produced blushing.

Darwin’s description of shame and guilt indicates that he saw them as distinguished by their eliciting events, although he could not find the specific behaviors that would mark the difference. He saw guilt as regret over some fault committed; a person’s relationship to others could then turn that guilt into shame. So, for example, he pointed out that one can feel guilty in solitude but would not blush because “it is not the guilt over an action, but what others think or know about our own guilt which leads to the blushing [shame]” (327). It was the social significance of the act, the eye of the other, which produced shame. He did note that a blush might occur in solitude, but only because we might be thinking about what others might be thinking of us. Again, the opinions of others about our appearance, especially the appearance of our faces, or our conduct, are Darwin’s elicitors of shame (345).

Although Darwin would not distinguish behaviorally between shyness, shame, guilt, and embarrassment, he did offer a variety of possible stimuli likely to elicit these emotions. He was clear in specifying that action per se is not the cause of the shame state: shame arises from how others see us. Understand here that the metaphorical “see” means “evaluate.” Once again, Darwin’s discussion presages others a century or more later.

Sylvan Tomkins elaborated on and further developed the Darwinian view (Tomkins, 1963). For Tomkins, the self-conscious emotions are not readily distinguished from each other at the level of affect; however, they are distinguished at the level of conscious awareness and by the stimuli that elicit them. Having said this, Tomkins spends most of his discussion considering shame, and directs relatively little attention toward shyness, disgust, contempt, embarrassment, or guilt. All of these emotions are incorporated into an overview of interconnected constructs. It is therefore difficult to separate one from another and claim that one pattern represents shame as opposed to embarrassment or guilt.

The definitions of shame that Tomkins gives vary from the prototypical, which is automatic in its ability to produce a shame state, to more specific situations and events. Tomkins’s model uses the idea of an automatic elicitor, one that does not require thought. Thus, he has no difficulty assuming that infants experience a shame state. His automatic eliciting event is any event that inhibits interest and enjoyment: “Shame is an innate auxiliary affect and a specific inhibitor of continuing interest and enjoyment. The innate activator of shame is the incomplete reduction of interest or joy” (Tomkins, 1963: 123). Tomkins’s association between shame and this reaction is correct; however, it is more likely we could transpose the elements and state that shame results in the interruption and termination of excitement and enjoyment. The automatic causal connection that Tomkins makes is predicated on an attempt to understand the utilitarian nature of shame in its evolutionary niche. Such a definition of shame makes its elicitation much more mechanistic and controlled by particular inhibitory mechanisms, rather than by a conscious or self-conscious attribution or process.

Tomkins’s prototypical and automatic elicitors of shame are barriers to enjoyment and excitement. I think the term “blocking of desire” would approximate Tomkins’s meaning of shame. On the other hand, Tomkins described shame: failure at work, loss of friendship, loss of close relationships through death, and feelings about the body. In particular, “unattractiveness is seen as producing shame because it produces a loss of interest in pride in our own bodies” (Tomkins, 1963: 194). Tomkins does not consider shame from the point of view of violating social norms. Because of his idea of interruption, and because of its causal nature, Tomkins’s theory of the elicitation of shame is unique. For him, the prototypical shame event is any event that leads to the reduction of interest. Here, then, is prototypicality at the level of function, not in terms of specific events. It is hard to know the direction of the causal chain between interrupt and shame, because careful sequential analysis of facial patterns prior to or during shame has not been seriously undertaken. It seems reasonable to think of shame as causing an interrupt of excitement and enjoyment rather than being caused by the interrupt of excitement and enjoyment. I suspect the former causal chain is closer to the truth, especially given the results of our studies on the interrupt of learning and the resultant emotions of anger and sadness but not shame (see Alessandri, Sullivan, and Lewis, 1990; Lewis, Alessandri, and Sullivan, 1990).

In considering what causes the shame state, Cal Izard vacillates between the notion of an automatic elicitor in the manner of Tomkins and a phenomenological position developed more fully by H. B. Lewis (1971). Shame can occur from an unrequited smile: excitement is interrupted, and positive emotions and enjoyment are reduced, just as Tomkins believes. But shame, for Izard, also has another dimension. Shame becomes a heightened consciousness of the self, an unusual and distinct form of self-perception. The self is seen as small, helpless, frozen, emotionally hurt. Although Izard does not explore the question himself, we should ask, By whom is the self seen as small, helpless, frozen, emotionally hurt? It appears that Izard is prepared to see shame as elicited by some belief or thought of the person. It is more than an interruption, it is an idealization. Izard defines shame phenomenologically “as a heightened degree of self-conscious self-awareness, or self-attention: our consciousness is filled with self and we are aware of some aspect of self we consider innocuous or inadequate” (Izard, 1977: 389; emphasis added). Here, then, is the precipitating event, not only an interruption or reduction of positive emotions, but also a heightened awareness of some aspect of self. Thus, the elicitation of the shame state requires objective self-awareness: the precipitating stimulus is that which we consider in ourselves to be innocuous or inadequate. It is what we could think of as contempt for the self.

Central to H. B. Lewis’s notion of shame-eliciting events is the belief that shame is a state of self-devaluation that can, but does not have to, emanate from “out there” (H. B. Lewis, 1971). Shame for her involves self-consciousness and self-imagery; that is, the idea of the other’s feelings. She distinguishes shame, which is about the self, from guilt, which is about action related to another. For her, shame and guilt are confused because of their common origins as modes of correcting lost affective bonds. They are fused under the heading of guilt but do not represent the same phenomena. One of H. B. Lewis’s major contributions was her stress on the belief that shame is produced by events located in the head of the person experiencing it. While it is true that she suggests that shame arises out of, and in large part is caused by, the loss of approval of a significant other, the source of the shame is our thoughts about our selves. The stimulus eliciting the state is self-thought about the self. For example, the disapproval of a significant other leads to thoughts about our selves. The stimulus eliciting the state is self-thought about the self. The disapproval of a significant other leads to thoughts about self-degradation, and these thoughts, in turn, lead the person to feel shame. This model is very different from one in which the shame state is the natural consequence of the specific action directly eliciting the state.

In general, we can conclude from this that the elicitors of shame appear to reside in one’s evaluation of the negative evaluation of others or of one’s self (Lewis, 1992). The particular reasons are varied. They can include failure to adhere to standards, such as failure to meet the requirement needs of neatness and cleanliness as described by Erikson; physical appearance issues as described by Darwin; or the loss of a significant other, as pointed to by Klein and the object relation theorists (Erikson, 1950; Klein, 1975). In all cases, however, it is the focus of the self on the self’s failure, and an evaluation of that failure, that leads to shame, not some automatic elicitor (Lewis and Michalson, 1983).

A Sense of Shame

In any theory of shame, the stimuli that elicit this state can best be understood if we consider shame from a phenomenological point of view. It would ease our task if we could define the state of shame by compiling a list either of a set of unique behaviors, or of a unique set of stimuli likely to elicit the particular feeling, or of some combination of the two. But this is impossible. A combination of behaviors and situations offers us a very powerful matrix in which to define, observe, and study individual differences in shame.

Consider the case in which two students each received the same grade on a paper. Elizabeth is overjoyed at an 85, but John is disturbed. Clearly, one could not define students’ reactions to their performance on an exam without knowing what their expectations were and how they would evaluate the grade they received. People’s responses to events and situations are, obviously, specific to their unique histories of experiences, expectations, desires, and needs.

Shame is like a subatomic particle. One’s knowledge of shame is often limited to the trace it leaves. We have proposed the desire to hide or to disappear as one very important feature of the phenomenology of shame; that is, that desire is an overpowering component of the experience. A second feature in descriptions of shame is intense pain, discomfort, and anger. In fact, these distinguish shame from embarrassment and shyness. A third feature is the feeling that one is no good, inadequate, unworthy. It is a global statement by the serf in relation to the self. And a fourth feature is the fusion of subject and object. In shame, we become the object as well as the subject of shame. The self system is caught in a bind in which the ability to act or to continue acting becomes extremely difficult. Shame disrupts ongoing activity as the serf focuses completely on itself, and the result is confusion: inability to think clearly, inability to talk, and inability to act.

This fourth phenomenological feature enables us to differentiate shame from guilt. Shame is the complete closure of the self-object circle. In guilt, although the self is the subject, the object is external to the self. The focus of the self is upon the behavior that caused the interruption, namely the inadequacy to meet certain standards, and upon the object that suffers from that failure. Many have used terms like concern or regret as synonyms for guilt, suggesting a focus on something external to the self rather than on the self itself. This four-feature phenomenological definition will be used throughout the remainder of this essay. Nonetheless, I shall not forget nor neglect the behaviors and stimulation likely to be associated with shame and its companion emotions of guilt, embarrassment, and pride.

Before we leave the phenomenology of shame, I must ask an important question: What is the function of these particular negative-feeling states? While functional analysis is always difficult, it is reasonable to ask what adaptive significations these emotions have. The function of guilt and shame is to interrupt any action that violates either internally or externally derived standards or rules. The internal command, which I call bringing into consciousness, says, “Stop. What you are doing violates a rule or standard.” This command serves to inhibit that action. The difference between shame and guilt resides in the nature of the interruption. In guilt, the command is essentially “Stop. What you are doing violates the standard or rule. Pay attention to what you did and alter your behavior.” Guilt is designed to alert the organism that the behavior violates some rule or standard and to alter that behavior. Its function is to alert or to provoke anxiety. In addition, it directs behavior toward alternative action patterns that repair the inappropriate behavior that has been called into question.

In shame the command is much more severe: “Stop. You are no good.” More important, it is about self, not about action; thus, rather than resetting the machine toward action, it stops the machine. Any action becomes impossible since the machine itself is wrong. The shame interruption is more intense given the identity of the subject-object. That the violation involves the machine itself means, functionally, that all behavior ceases. Its aversiveness functions to ensure conformity to the standards and rules. While shame, more than guilt, should be likely to change behavior, thought, or feeling, its aversiveness may be so extreme that shame is bypassed. If bypassing occurs, a shame state may be ineffective in producing a change in behavior. There is some evidence from the work of Janis that too intense a message will be disregarded (Janis, 1965).

In reviewing ideas about shame, we have come to see that shame involves specific behaviors associated with certain phenomenological experiences. These behaviors are elicited by two classes of events: those related to specific physical events, like exposure of the genitals, and those related to thoughts about the self. While both classes have been recognized, they have not been separated, in part because no careful analyses of the cognitive aspects of shame have been undertaken. Without introducing the mental state of awareness and thoughts about standards, our understanding of shame will remain inadequate. To understand the development of shame, we need to understand the development of this mental state of awareness (Lewis and Brooks-Gunn, 1979).

The Mental State of Me

Although human adults possess the capacity of consciousness, or the mental state of the idea of “me,” for most of our organizing and regulation functions we do not avail ourselves of this capacity. Moreover, there are times when its use can actually lead to difficulty. For example, when we try to listen to what we are saying while we are talking, we block our ability to continue to talk.

Much has been written about this complex capacity, only some of which will be referred to here. I have already indicated the importance of self for emotions in general, and for shame in particular, and I will return to this topic when I focus more narrowly on shame. Here I am concerned simply with the capacity for self-reflection. As we will see, this aspect of the self goes by many names: consciousness, the mental state of the idea of “me,” explicit as opposed to the implicit consciousness, and the I.

The notion of a self and its development is vitally important for understanding shame. A mechanistic model of human behavior can ignore the self, but human beings are not simple mechanisms that operate like switchboards: information coming in, information going out. All of us consider ourselves. We not only know ourselves, we know ourselves in many different ways. We know our names, we know what we look like, we know what it feels to be us.

For the last 25 years, I have been studying the problem of self-development and emotional life. In Social Cognition and the Acquisition of Self (1979), I undertook a series of studies in which we observed children’s self-recognition as an attempt to study serfdevelopment. In Children’s Emotions and Moods (Lewis and Michalson, 1983), we focused on emotional development and the role of self in its development. From the series of studies and work we have done in this area, I have articulated a theory of self-development. My interest in empirically exploring the origins of self started with our interest in self-recognition. Given that infants and very young children do not have the capacity for language, it is difficult to study the acquisition of these concepts of the serf, so I decided to look at children’s behavior in the presence of reflective surfaces. Our results indicate that true consciousness as defined by self-referential behavior does not emerge until the second half of the second year of life. Moreover, our work shows that personal pronoun usage and pretend play all emerge at around the same time (Lewis and Ramsay, 1999). Thus, we have reason to believe that consciousness emerges in the human infant at this point of time.

In the Genesis story of creation, shame is the only emotion that is discussed at any length. The biblical version of the origin of shame has considerable significance for the Western mind. More over, the issue of shame as presented in the myth is particularly cogent for my view of the developmental process. In the Genesis creation story, Adam and Eve recognize their nakedness, their sense of exposure before God, and they hide their nakedness because of their shame. There are three critical features to this story. Adam and Eve disobey God because they are curious. Their curiosity leads them to sample the fruit of the tree of knowledge: in other words, their curiosity leads to knowledge. This knowledge in turn leads to their sense of shame. Here, then, is the core of the myth. Curiosity leads to knowledge, which leads to shame. The structure of this early creation story fits the ontogenetic development of shame for the child and describes the process of shame production for the human adult. In discussing shame and the other self-conscious emotions, I have already suggested that one has to have specific cognitions in order for them to occur. One has to know something about 1) standards, rules, and goals; 2) one’s own behavior with regard to these standards; and 3) oneself. Only when these cognitions develop can shame occur.

The Development of Shame

Figure 1 presents a schematic of the model of the development of shame. The model has five major features that account for the change over the first three years of life. The development of shame involves the transformation and integration of structures and functions. It does not exist at birth, but instead develops over age. The first emotions to emerge are the primary emotions. Not all of these emotions appear at the same time. Bridges has described how the earliest two classes of emotion, positive emotion (joy/happiness) and distress, differentiate into the other emotions (Bridges, 1932). Disgust emerges from distress and is followed by anger, appearing somewhere between ages 2 and 4 months; fear emerges sometime later, at about 8 months. Likewise, surprise emerges early, perhaps developing from the interest/joy axis. Somewhere around age 8 months these differentiated primary emotions are already evident. The socialization of these primary emotions within the interpersonal life of the child as well as maturation contributes to the next phase of development, the cognitive capacity of objective self-awareness.


Because I have already spent considerable time discussing consciousness, I will only mention here that near the middle of the second year of life the child gains this capacity. At this point, when the child becomes conscious of himself/herself, several emotions emerge that are related to this cognitive ability. These emotions are not evaluative in nature; they do not arise because of correct or incorrect thoughts, actions, or feelings. Three such emotions are embarrassment, empathy, and envy, but there may be others. These emotions, the consequence of consciousness, make up the first set of self-conscious emotions (see Lewis, 1993 for further discussion of this first set).

The emergence of shame and the other self-conscious emotions requires something more than consciousness: the child must have certain additional cognitive capacities. Standards, rules, and goals must exist that, when accompanied by consciousness, give rise to this new set of self-conscious emotions, the self-conscious evaluative emotions, shame being one of them.

Self-Cognitions and Attributions and the Development of Shame

Following the onset of these self-conscious emotions and the development of standards, rules, and goals, shame emerges. Figure 2 presents the structural model that I will use to define shame and the rest of the self-conscious emotions. Three subscripts–A, B, and C–represent cognitive processes that serve as stimuli for these emotions. I will use terms that are commonly used both in the literature and in everyday discussion. It may seem to some, at least at first, an idiosyncratic usage of the terms. However, as I have already noted (Lewis, 1992), one problem associated with the study of the self-conscious emotions is the mixing of terms, for example, shame and guilt.


The first feature of the model concerns the standards, rules, and goals that govern our behavior. All of us have beliefs about what is acceptable for ourselves and for others with regard to these standards concerning actions, thoughts, and feelings. This set of standards, rules, and goals, or beliefs is derived from the information one acquires through acculturation in a particular society. Standards (for simplicity, this term will serve as shorthand for standards, rules, and goals) differ across societies, across groups within societies, across time periods, and between individuals of different ages. The standards of our culture are varied and complex, yet each of us knows at least some of them. Moreover, each of us possesses a set of standards unique to ourselves. To become a member of any group requires that we learn the group’s standards. I can think of no group that does not have such standards, nor of any group that does not employ negative sanctions to enforce their standards. These standards are acquired through a variety of processes. They always are associated with human behavior, including thinking, action, and feelings. They are prescribed by the culture, which extends from the family group, through other social groups (peers, work groups, etc.), to the culture at large. The knowledge can be acquired through a passive or an active process. In the passive view, we are acted upon by cultural forces; in the active view, humans act upon these forces (Overton, 1984). In the passive view, the causes of behavior, action, or thought are forces that act on the organism, causing it to acquire information. These may be internal biological features of our species or the external social control of conspecifics, determined by the shaping effect of the differential rewards. In contrast to this passive view is the constructionist paradigm, which is based on the worldview that the organism acts on its environment and participates in it (see Lewis, 1979, 1990). The organism has desires and plans. These desires and goals are constructed, as are most of the actions that enable the organism to behave adaptively within its culture.

One particular aspect of the active process is the child’s observation of how people behave toward one another. One of the most curious facts about the study of this problem has been what I have called the error of the didactic method. The didactic method leads us to conclude that children or adults learn through direct interaction between the “teacher”–the person extending the information–and the “learner”–the person receiving the information. Most of our learning is not through this didactic method, but through other, less direct methods.

For example, consider the case of the five-year-old girl who has a two-year-old sibling. The five year old receives some crayons as a present and uses those crayons and colors on the walls of their house. The two year old watches as her older sister does this. Mother appears and starts to yell at the five year old, exclaiming, “Don’t you know you are not supposed to write on the walls? For this I’m going to punish you; give me your crayons and go to your room,” at which point the five-year-old is reduced to tears.

In fact, she may not have known the rule, and the violation of it comes as a surprise. She now has learned the rule and, under most circumstances, is unlikely to color on the walls again. But what about the two year old? Here we see something quite different. The two year old has not marked with crayons on the wall, nor has she been yelled at; yet, I think it is safe to say that this child will not use crayon on the walls since she has learned indirectly about a standard.

However we might consider how these standards are acquired, they are necessary for the emergence of these self-conscious evaluative emotions. The evaluation of one’s actions, thoughts, and feelings in terms of standards is the second cognitive evaluative process that serves as a stimulus for serf-conscious emotions. Two major aspects mark this process. The first concerns the internal and external aspects of evaluation. For my model to work in describing the process of eliciting emotions, internal evaluation, as opposed to no evaluation or external evaluation, is necessary. Obviously, individuals differ in their characteristic evaluative response. Moreover, situations differ in the likelihood that they are internally caused. The second aspect of the evaluative process is concerned with how young children make a determination about success or failure in regard to any specific standard. The work of Hans Heckhausen and Deborah Stipek, as well as our own, seems to indicate that by the beginning of the third year of life, children already have their own sets of standards, rules, and goals and seem to show distress when they violate them (Heckhausen, 1984; Stipek, 1983). These results indicate that by the beginning of the third year of life, children become upset when they violate their own standards. The importance of these observations for the model is that the evaluation of our behavior in terms of our standards is a natural process independent of the nature of the standards themselves.

Some standards, rules, and goals are more valuable than others. For me, the goal of driving well is less valuable than the goal of helping a student or solving a problem. Thus, the evaluation of my behavior in relation to these two different standards should result in different emotions. The violation of those more central to the definition of the self are more likely to lead to shame. What constitutes a more central or peripheral standard is defined by the individual as well as by the family, various intermediate groups, and the culture at large. Moreover, as new standards, rules, and goals are added, their positions relative to the centrality of self differ. Thus standards, rules, and goals and their importance change over the life course.

The field of attributional studies has investigated another problem regarding evaluation, the problem of internal versus external attribution (Weiner, 1986). People violate standards, rules, and goals, but often do not attribute the failure to themselves. Instead, they may explain their failure in terms of chance or the actions of others (Seligman, 1975; Seligman et al., 1984).

Carol Dweck and her associates have conducted studies in which young boys and girls were asked about the causes of their success and failure, especially within academic fields (for example, Dweck and Leggett, 1988). Many children blame their success or failure on external forces (some children, of course, also evaluate success and failure in terms of their own action). Interestingly, strong sex differences emerged. In academic achievement, boys are more apt to hold themselves responsible for their success and to blame others for their failure, whereas girls are apt to attribute their success to others and to blame themselves for failure.

Another feature of evaluation concerns the socialization of what constitutes success or failure. Having assumed responsibility (internal evaluation), how do we go about evaluating ourselves as succeeding or as failing? To begin with, the individual’s evaluation of the success or failure relative to standards can reflect an “accurate” interpretation; that is, his evaluation might be judged by an unbiased observer as normal, or as approximating the kind of evaluation a typical self would make in like situations. Alternatively, the individual’s evaluation can reflect a unique interpretation. The unique interpretation could be a function of the consequences of establishing too high a standard. The student who sets his standards for school performance at a perfect score is likely to think of himself as a failure for receiving a grade of 90. Likewise, the student who sets too low a standard is likely to think himself a success just for achieving a barely passing grade. I should note that the inaccuracy of these students’ interpretation of success and failure has to do with what other selves would conclude given the same situation.

An alternative reason for the inaccuracy of the interpretation concerns misunderstanding. Take, for example, a student who thinks she knows the answer to a question presented in a test but, in fact, does not. From her point of view she did well, but from the point of view of the professor who created and grades the test, she did poorly. Here we have a case of misjudgment of success or failure relative to a standard.

Many factors are involved in producing inaccurate or unique evaluations of success or failure. These include early failures in the serf-system that lead to narcissistic disorders, harsh socialization experience, and high levels of reward for success or high levels of punishment for failure. The evaluation of one’s goal-directed behavior in terms of success and failure is an important aspect of the organization of plans and the determination of new goals and new plans. Its importance in relation to the notion of interrupt is central. If, as I believe, the self-conscious emotions of shame and guilt serve as interrupt signals to inform us that the actions we have taken have failed, the interrupt clearly serves the biological function of enabling the organism to reconsider and alter its strategy. This process must be borne in mind when we start to consider what happens when we succeed. Success results in positive affect such as joy, interest, excitement, and pride. The consequence here is to reward the self in terms of the action, thought, or feeling. In the negative as well as positive evaluation of one’s action, the consequence of evaluation is an affective discharge.

We now come to the final aspect of the cognitions needed to create the self-conscious emotions. Recall that in almost all of the models of self that describe the phenomenological experience of shame, there appears to be some agreement as to the role of the self as object. H. B. Lewis, for example, speaks of shame as being more about the self, whereas guilt is more about the other. The phenomenology of these emotions suggests that the object of the self’s orientation is quite different when we feel shame than when we feel guilt.

I want to suggest that shame is elicited when the serf orients toward the self as a whole and involves an evaluation of the total self, whereas in guilt it is orientation of the self toward the actions of the self, either in terms of the actions of the serf alone or in terms of the actions of the serf as they have affected another.

Personality theory has drawn attention to an important feature of self-regard and evaluation of the self. Studies on depression (notably that of Aaron Beck) also have focused on individual differences in making serf-attributions (Beck, 1967, 1979). Global and specific are used to specify the tendency of individuals to make evaluations about the serf. Global refers to an individual’s propensity to focus on the total serf. Thus, for any particular behavior violation, some individuals, some of the time, are likely to focus on the totality of the self. These individuals use such evaluative phrases as, “Because I did this, I am bad (or good).” Janoff-Bulman’s distinction between characterological and behavioral self-blame is particularly relevant here (Janoff-Bulman, 1979).

We can see how closely this idea agrees with the phenomenological experience of shame. In shame situations, the interrupt and focus is upon the serf, both as object and as subject. The self becomes embroiled in the serf because the evaluation of the self by the serf is total. There is no way out. The focus is not on the individual’s behavior, but on the total serf. The individual who makes global attributions focuses on herself, not on her action. Focusing inward, such a person is unable to act, and is driven from the field of action into hiding or disappearing.

By specific attribution, I mean that in some situations, some of the time, some individuals have the propensity to focus on specific actions of the self. That is, their evaluation of their serf is not global, but specific. It is not the total self that has done something wrong or right, bad or good; instead, specific behaviors are examined and judged. At such times as these, individuals will use such evaluative phrases as, “What I did was wrong. I mustn’t do it again.” Notice that the individual’s focus here is on the behavior of the self in interaction with objects or persons, and also on the individual’s effect on other selves.

Given these three sets of activities–1) standards, rules, and goals; 2) the evaluation of success or failure of one’s action with regard to these; and 3) the attribution of the serf–it now is possible to observe the full set of self-conscious emotional states. Significantly, this model is symmetrical relative to positive and negative self-conscious emotions. Because of this symmetry, I will focus not only upon shame and guilt, but also upon the other side of the axis, hubris and pride. Throughout the discussion of these cognitive-evaluative processes, I have indicated that particular events may have unique eliciting properties. Nevertheless, it is the cognitive-evaluative process of the organism itself that elicits these states. The immediate elicitors of these serf-conscious emotions are cognitive in nature.

In the model, I distinguish between four emotional states. Notice that shame is a consequence of a failure evaluation relative to the standards when the person makes a global evaluation of the self. Guilt also is the consequence of a failure; however, with guilt, the focus is on the serf’s action. A parallel exists as a consequence of success. When success is evaluated and the person makes a global attribution, hubris (pridefulness) is the resulting emotion; when success is evaluated and the person makes a specific attribution, pride is the resulting emotion. I use here the Greek term hubris to differentiate this emotion from pride, with which it is too often confused. As I have warned before, word usage issues associated with these emotions render careful analysis most difficult. Not only do we have the problems of differentiating shame from guilt, embarrassment, and shyness, but we also have the difficulty of distinguishing between different kinds of pride. One can think of pride in terms of its use in two ways. On the one hand, I think of pride in terms of achievement, the feeling one has when being successful in fulfilling a particular goal and activity. In my discussion of achievement motivation, I noted how children learn to feel proud about their achievements in terms of a particular standard, rule, or goal. On the other hand, I also use the term pride to indicate a negative emotional state. I speak of the proud man or the proud woman with some disdain. The Bible speaks of pride coming before the fall; throughout the story of the Old and New Testaments, we find examples of false pride and how this pride brings down the man. It is clear that the term pride carries a surplus of meaning; if we are to understand the term at all, we need to distinguish between specific pride and global pride. I have done so by the use of the term “hubris” to represent global pride and “pride” to represent specific achievement. I see hubris as the counterpoint to shame. The pride-shame axis has been described by others who have noticed the similarity between these two emotional states (Morrison, 1986, 1989).

Measurement of Shame and Early Effects

While the primary emotions can be measured from the face, the self-conscious evaluative emotions require both facial and bodily observation. Following the coding system of Geppert (1986), shame is defined as bodily collapse, shoulders falling in, downward/lower lip, eyes lowered with gaze downward or askance, withdrawal from a task or person, and negative self-evaluation, such as “I’m no good at this.”

To measure shame, we construct an achievement task where easy and difficult problems are presented to children. Their success or failure at these tasks can be controlled by varying the time to complete the task. Unknown to the children, we can shorten or lengthen the time we give them to solve the problem. Thus, if we want them to fail, we sound the buzzer marking the end of their time before they are finished, and if we want them to succeed, we sound the buzzer after they complete the task. In this way, we can control their success or failure in easy or difficult tasks. The procedure works well and children are unable to detect the time manipulation. Lewis, Alessandri, and Sullivan (1992) explored such procedures with three-year-old children. We found that children showed more shame faces when they failed an easy task than when they failed a difficult task. We also measured pride and found that three year olds showed more pride when they succeeded in a difficult task than when they succeeded in an easy task. These differences in emotion by success and failure and as a function of task difficulty indicated that children by the age of three years have developed the cognitive capacities that we have proposed underlie these self-conscious emotions. Using the same procedure, we have been able to show that children from three to eight years show similar shame responses when placed in these achievement tasks.

Large individual differences exist early, reflecting both temperament and socialization affects (Lewis, 1992). Most interesting sex differences are evident even at three years of age, with girls showing more shame when they fail than do boys. In particular, girls show more shame when they fail a difficult task relative to boys while there are relatively no sex differences when children fail an easy task. These early sex differences appear to be enduring and are found across the three- to eight-year-old period. Interestingly, there is reason to believe that these sex differences extend into adulthood with women showing more shame than men (see Lewis, 1995). Other individual differences in shame proneness have been demonstrated as a function of maltreatment and sex abuse, with abused children showing more shame. We have postulated that shame is the important source of subsequent pathology (Lewis and Haviland-Jones, 2000).

Thus, the child’s emotional development takes place over the first three years of life. Starting with the simplest pre-wired responses seen at birth, by three years the child is able to evaluate his or her own actions and with these new cognitive capacities shows self-conscious evaluative emotions. The emergence of shame as part of this complex of new emotions acts to inhibit children’s actions, thoughts, and feelings that do not conform to the internal standards learned through socialization.


Alessandri, S., M. W. Sullivan, and M. Lewis. “Violation of Expectancy and Frustration in Early Infancy.” Developmental Psychology 26 (1990): 738-744.

Beck, A. T. Depression: Clinical, Experimental and Theoretical Aspects. New York: Harper and Row, 1967.

–. Cognitive Therapy and the Emotional Disorders. New York: Times Mirror, 1979.

Bridges, K. M. B. “Emotional Development in Early Infancy.” Child Development 3 (1932): 324-334.

Darwin, C. R. The Expression of Emotions in Man and Animals. Chicago: University of Chicago Press, 1965 (1872).

Dweck, C. S., and E. L. Leggett. “A Social-Cognitive Approach to Motivation and Personality.” Psychological Review 95 (1988): 256-273.

Erikson, E. H. Childhood and Society. New York: Norton, 1950.

Geppert, U. A Coding System for Analyzing Behavioral Expressions of Self-Evaluative Emotions. Munich: Max-Planck Institute for Psychological Research, 1986.

Heckhausen, H. “Emergent Achievement Behavior: Some Early Developments.” The Development of Achievement Motivation. Ed. J. Nicholls. Greenwich, Conn.: JAI Press, 1984.

Izard, C. E. Human Emotions. New York: Plenum Press, 1977.

Janis, I. L. “Psychodynamic Aspects of Stress Tolerance.” The Quest for Self-Control. Ed. S. Z. Klausner. New York: Free Press, 1965.

Janoff-Bulman, R. “Characterological versus Behavioral Self-Blame: Inquiries into Depression and Rape.” Journal of Personality and Social Psychology 37 (1979): 1798-1809.

Klein, M. “Envy and Gratitude” and Other Works, 1946-1963: On the Theory of Anxiety and Guilt. New York: Delacorte Press, 1975 (1948).

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

Lewis, M. “The Self as a Developmental Concept.” Human Development 22 (1979): 416-419.

–. “The Development of Intentionality and the Role of Consciousness.” Psychological Inquiry 1 (1990): 231-248.

–. Shame: The Exposed Self. New York: The Free Press, 1992. Paperback ed. 1995.

–. “Self-Conscious Emotions: Embarrassment, Pride, Shame and Guilt.” Handbook of Emotions. Eds. M. Lewis and J. Haviland. New York: Guilford Press, 1993.

Lewis, M., S. Alessandri, and M. W. Sullivan. “Expectancy, Loss of Control, and Anger in Young Infants.” Developmental Psychology 26 (1990): 745-751.

–. “Differences in Shame and Pride as a Function of Children’s Gender and Task Difficulty.” Child Development 63 (1992): 630-638.

Lewis, M., and J. Brooks-Gunn. Social Cognition and the Acquisition of Self. New York: Plenum Press, 1979.

Lewis, M., and J. Haviland-Jones, eds. Handbook of Emotions. 2d ed. New York: Guilford Press, 2000.

Lewis, M., and L. Michalson. Children’s Emotions and Moods: Developmental Theory and Measurement. New York: Plenum, 1983.

Lewis, M., and D. Ramsay. “Intentions, Consciousness, and Pretend Play.” Developing Theories of Intention. Eds. P. D. Zelazo, J. W. Astington, and D. R. Olson. Mahwah, N.J.: Lawrence Erlbaum Associates, 1999.

Morrison, A. P. Shame: The Underside of Narcissism. Hillsdale, N.J.: Analytic Press, 1989.

Morrison, A. P., ed. Essential Papers on Narcissism. New York: New York University Press, 1986.

Overton, W. “World Views and Their Influence on Psychological Theory and Research: Kuhn-Lakatos-Lauden.” Advances in Child Development and Behavior. Vol. 18. Ed. H. W. Reese. New York: Academic Press, 1984.

Seligman, M. E. P. Helplessness: On Depression, Development, and Death. San Francisco: Freeman, 1975.

Seligman, M. E. P., et al. “Attributional Style and Depressible Symptoms among Children.” Journal of Abnormal Psychology 39 (1984): 235-238.

Stipek, D.J. “A Developmental Analysis of Pride and Shame.” Human Development 26 (1983): 42-54.

Tomkins, S. S. Affect, Imagery, and Consciousness. Vol. 2. The Negative Affects. New York: Springer, 1963.

Weiner, B. An Attributional Theory of Motivation and Emotion. New York: Springer-Verlag, 1986.

Michael Lewis is University Distinguished Professor of Pediatrics and Psychiatry, Director of the Institute for the Study of Child Development at Robert Wood Johnson Medical School, and Professor of Psychology at Rutgers University.

COPYRIGHT 2003 New School for Social Research

COPYRIGHT 2004 Gale Group