Emotion Expression in Groups1

David R. Heise
John O'Brien

Excerpted from a pre-publication draft. For the final text see the printed article in:

M. Lewis and J. M. Haviland (Eds.), Handbook of Emotions. (New York: Guilford, 1993. Chapter 34, pp. 489-497.)

Face-to-face encounters are a primary arena for the display and feeling of those intense, ephemeral, and situated affective experiences we call emotions. Emotions, as much as interpersonal transactions, reflect the social structure and the encompassing culture of participants involved in interaction, and emotions help manifest that structure and culture from moment to moment. A sociology of emotions that deals with these themes emerged in the 1980s (Kemper, 1990a; Thoits, 1989).

Empirical Evidence

Earlier sociologists, beginning with Bales (1950), systematically observed groups in laboratories and found that a substantial proportion of group interaction is devoted to the socioemotional issues of expressing affect and dealing with tension (Hare, 1976, chapter 3, reviews the work). Simultaneously, field studies of liking and disliking in natural groups showed that interpersonal sentiments collate into informal social structures (Hare, 1976, chapter 7) a discovery that still is being explored in the burgeoning field of social network analysis (e.g., Burt, 1982; Knoke & Kuklinski, 1982; Bradley, 1987).

Ethnomethodologists (sociologists concerned with the methods by which people create, maintain, and heed implicit cultural rules) demonstrated emotional commitments to common understandings in social relationships by having a group participant purposely violate some mundane, shared understanding. For example, one breaching experiment required students to spend from fifteen minutes to an hour acting as boarders in their own homes (Garfinkel, 1967: 47-8). They were instructed to conduct themselves in a circumspect and polite fashion. They were to avoid getting personal, to use formal address, to speak only when spoken to. Though a fifth of the assignments were not completed because the students were afraid to carry through or because the family refused to take their performance seriously, the remaining four-fifths of the students found emotions rampant. Reports were filled with accounts of astonishment, bewilderment, shock, anxiety, embarrassment, and anger, and with charges from various family members that the student was mean, inconsiderate, selfish, nasty, or impolite. Family members demanded explanations: What's the matter? What's gotten into you? Did you get fired? Are you sick? What are you being so superior about? Why are you mad? Are you out of your mind or are you just stupid? One student acutely embarrassed his mother in front of her friends by asking if she minded if he had a snack from the refrigerator. "Mind if you have a little snack? You've been eating little snacks around here for years without asking me. What's gotten into you?" One mother, infuriated when her daughter spoke to her only when she was spoken to, began to shriek in angry denunciation of the daughter for her disrespect and insubordination and refused to be calmed by the student's sister. Some families tried to connect with the student's behavior as a joint comedy routine, but they soon gave up in irritation and exasperated anger. Other families accounted for the student's behavior in terms of a mood or a prior misfortune; they became angry when the student did not accept help in dealing with the problem. Sometimes nasty interpersonal schisms developed that could be repaired only by revealing the experiment.

Purposely breaching a norm distresses the actor at least as much as observers. Milgram (1974) sensitively related the feelings of a person asking passengers on a New York subway for their seats a demonstration of breaching that he had assigned to his students, and then decided to do himself. Waves of emotion accompanied his experience and continued until the interpersonal situation was over. To start, Milgram nearly quit because of his unexpected apprehension and intense emotional reaction. It was only through the urging of a student observer that he finally did approach a man and say, "Excuse me sir, may I have your seat?" A moment of stark anomic panic overcame me. But the man got right up and gave me the seat. A second blow was yet to come. Taking the man's seat, I was overwhelmed by the need to behave in a way that would justify my request. My head sank between my knees, and I could feel my face blanching. I was not role-playing. I actually felt as if I were going to perish. Then the third discovery: as soon as I got off the train, at the next station, all of the tension disappeared. Milgram (1974: 72)

Gregory's (1982) study of intentional breachers showed that people cope with the severe anxiety generated by rule-breaking through rationalizations that normalize their actions to themselves, and through public accounts that normalize their actions to others, including tacit accounts like Milgram's sinking head and blanched face. Garfinkel (1967: 68-70) noted changes in emotion with habitual rule breaking. For example, student breachers who had to bargain for standard priced merchandise just once were more distressed than student breachers beginning a series of six bargaining episodes. Distress declined with repetition, so most students with a sequence of trials actually looked forward to, and enjoyed, the assignment by the third episode, and some resolved to continue similar behavior in the future.

Breaching experiments showed that bland expressions and reserved miens, to be seen in audiences confronting heinous events in the mass media, are misleading indicators of how people emotionally respond to interpersonal deviance. In face-to-face situations, even trifling violations of mores generate intense emotions for observers and perpetrators alike. The breaching studies reveal "a definite and strong relationship between common understandings and social affects" (Garfinkel, 1967: 50).

Breaching of norms usually provokes negative emotions, but other studies show that interpersonal relationships are the primary medium for people to express positive emotions as well. Records from observational studies in laboratories (Hare, 1976, Table 4) show that tension releases (laughing, displaying satisfaction) constitute around six percent of all events in groups. Brandstatter (1983) found that housewives reported their most positive moods when others were present, especially family and friends. Kraut (1979) studied smiling in several different settings, and found that displays of happiness are correlated with the presence of people to see it, even though the basis for happiness may be a personal achievement. For example, bowlers who made a strike often did not smile until they turned to face their friends. Recordings of social interactions also reveal frequent occurrences of laughter, even in serious, task-oriented situations. For example, Grimshaw (1989: Appendix A) provided an accurate transcription of discourse by four professors during the in camera segment of a dissertation defense the part devoted to discussion of the candidate's graduate career, dissertation, and career prospects. Twelve minutes of serious discussion were interspersed with 26 instances of laughs or chuckles (often by multiple people), which amounts to a laugh about every half minute on the average.

The correlation between emoting and socializing provides the empirical basis for supposing that emotions are intertwined with the maintenance and alteration of social relationships, and with the cultural understandings that gird social events. Emotionality in groups also encourages investigating how emotions synchronize social interaction and foster group unity. We turn now to reviewing some theoretical formulations concerned with these themes.

Social Constructionism

Social constructionists approach the interpenetration of emotions and sociocultural phenomena by understanding the expression of emotions as intelligent conduct, contrived according to cultural rules so as to effect desired interpersonal outcomes. In this perspective, displays of emotion are not uncivilized eruptions coming from deep within individual psyches, but rather amount to sophisticated social discourse that is employed to influence others.

Representative of this kind of theorizing is Schieffelin's (1983) analysis of the systematic use of anger, grief, and shame among the Kaluli people in Papua New Guinea a study that expands on Goffman's (1967) analyses of the function of emotion displays in self-presentation. Among the Kaluli, anger and grief are ways of entreating others, while shaming is employed as a way of obstructing appeals. First, consider anger; it is an assertive posture used to provoke, intimidate, excite, and inspire others. A man whose expectations have been frustrated or who has suffered wrong or injury at the hands of others does not usually suppress his annoyance. Rather he is likely to orchestrate his anger into a splendid frightening rage, projecting himself with threats and recriminations against his (often equally angry) opponent ... (Schieffelin, 1983: 183).

Anger implies that the person has suffered a loss of some kind and legitimately is entitled to redress. Accordingly: an angry man is not only intimidating, he is also a figure of pathos for the Kaluli, and a display of anger is frequently meant to be a forceful plea for support. ... Thus anger attains a particular rhetorical force, a certain kind of measure and legitimacy, and a set of implications, from the way it is situated in the scenario of reciprocity. (186-7)

Influence also can be attained with Kaluli friends and relatives by posturing as needy, vulnerable, and dependent, in order to appeal to sentimentality and compassion. [G]rief is the extreme posture of vulnerability and appeal. Given meaning, like anger, within the Kaluli sense of reciprocity, grief represents a picture of a person reduced to powerlessness and vulnerability by devastating loss, a figure of great pathos, one who is in principle entitled to redress. Powerful in rage, men are reduced to a particular helplessness in grief, weeping in a hysterical and uncontrolled manner. (187)

A display of grief insinuates that a person, who is incapacitated in seeking justified compensation for a loss, awaits redress from goodhearted others who in one way or another are able to help. The appeal is largely to others' sense of justice, though among the Kaluli grief achieves some of its persuasiveness from the implicit danger that grief can turn into rage.

Faced with forceful interpersonal rhetorics, like anger and grief, as means of commandeering one's energies and resources, Kalulis depend on shame as a tool to help preserve their own interests. A Kaluli may question another's rights to something lost, attempting to evoke shame over illegitimate claims. The phrasing of the challenge as a rhetorical question aims to avoid a confrontation or clash of wills and risk of anger, and throws the one to whom it is directed on the defensive. Moreover, for an assertive move, the question `Is it yours?' implies that the request for the object amounts to a kind of potential theft that suggests the threat of retaliation in line with the Kaluli's sense of reciprocity. (188) , Shame is revealed here as a situation as much as a private emotion: both its existence and meaning are products of interaction, not loss of reputation or a sense of right and wrong. ... The legitimacy of one's basic posture of assertion or of appeal has been removed. (189)

Thus, Kaluli emotions form a nonverbal system of interpersonal communication, used in negotiating culturally normative rights and obligations. [They] are socially located and have a social aim. To this degree they are located not only in the person, but in the social situation and interaction which, indeed, they help construct. (190-1) Being declarations of mind for social consumption, public displays of emotion may not reveal what a person really feels. Asked what other persons feel about an event, a Kaluli informant typically replies, "`I don't know. How is one to know what another man feels?' ... [or that they] `acted as if they were angry (happy, dismayed, etc.).'" (Schieffelin, 1983: 184)

The Kalulis' use of anger, grief, and shame has parallels in industrial societies, and Westerners engage in the calculated use of emotional displays no less than New Guineans. Clark (1990) identified a number of strategies by which emotion presentations keep others in their place and allow an individual to gain or recover status; Heise (1989) presented an alternative approach to these issues. Moreover, because emotion displays in everyday social affairs articulate underlying ideologies that some people want legitimated, sophisticated expressions of emotion have turned into a form of labor that individuals market in capitalist societies (Hochschild, 1979, 1983, 1990); we return to the issue of commercial emotion work in a later section.

Social Determinism

Social determinists approach the interpenetration of emotions and sociocultural phenomena by understanding emotions as authentic, involuntary responses that occur during social interactions. Emotions emerge from the operative social structure in a situation, and emotions allow people to sense that structure, as well as the social consequences of actions. Moreover, because displays of emotion broadcast a person's subjective appraisals to others, emotions contribute tacitly to sharing views about social structure and to synchronization of action and feeling within a group.

Kemper's deterministic framework (1978, 1981, 1990b; Kemper & Collins, 1990) posits that people in social interaction are arrayed along two relational dimensions. First, a person, Ego, has a degree of status that is determined by Alter's esteem for Ego and that determines the extent to which Alter is likely to engage in unconditional acts of regard for Ego. A person with high status realizes other people's ideals so that the others voluntarily and spontaneously function on that person's behalf. Second, Ego has a degree of power in a relationship that is a net outcome of Alter's dependency on Ego and that determines the likelihood that Alter will perform as Ego wills, conditional on Ego's surveillance of Alter and on Ego foiling Alter's power plays. A person with power is seen as resourceful by others, so others function on that person's behalf strategically as a way of sustaining their own objectives, trying to consolidate control over the resources held by the powerful person.

Emotions arise as interpersonal events affirm and change individuals' levels of status and power. Events affirming social status and power yield "structural" emotions that characterize the nature of the relationship. For example, acting in such a way as to affirm another's exalted status yields emotions related to love, and Kemper (1972) identified different kinds of love-related emotions. The variations are related to the power of the loved person and the status and power of the lover (e.g., a feeling of "romantic" love arises when both have high status and high power, "brotherly" love when both have high status and low power).

"Consequent" emotions occur during social interaction as events change social structure, as people distribute responsibility for these changes, and as they compare structural changes to their anticipations. Increases or decreases in status or power of self or of other lead to specific emotions, as discussed by Kemper in his contribution to this volume. According to Kemper (1981: 338), "particular social stimulus keys fit particular locks to produce particular emotions." This explains why people may feel a different emotion than the one they socially display: they interpret the personal implications of interaction and feel the appropriate emotion even if constructing a culturally or ideologically suitable emotional display that advances their rational interests.

Another branch of social determinism focuses on simultaneous emotional arousal within a collectivity, as a result of rituals involving cultural symbols (like a flag, the word "democracy", presence of a baby, etc.), or as a result of emotional contagion (Hatfield, Cacioppo, & Rapson, 1992). Early social analysts like LeBon (1895/1960) thought that emotional contagion in crowds swept away individual rationality and substituted emotion-based goal seeking and synchronization of action. However, McPhail's (1991) extensive observational studies of crowds suggest that contagion phenomena are rare in those crowd situations where they are supposed to thrive; emotion contagion may occur mainly in collective situations where conditions encourage altered states of consciousness (Prince, 1982).

Durkheim (1912/1954) proposed that rituals hold society together by producing sacred objects and moral constraints, and Collins (1975, 1981, 1990) expanded on Durkheim's theme by proposing that a common emotional mood generated in rituals creates social solidarity and diffuses charismatic emotional energies that preserve and disseminate normative group patterns. Profaning a symbol usually will elicit anger and conflict between groups or between group factions, but reaffirming symbols generates positive emotion and synchronization within the group. Collins holds that the inherent emotionality of even commonplace interaction rituals is the glue that holds society together and the driving force that mobilizes social change.

Social Interactionism

Social interactionists approach the interpenetration of emotions and sociocultural phenomena by understanding that emotions are both constructed and determined. The interactionist model points to a certain paradox; a feeling is what happens to us. Yet it is also what we do to make it happen. (Hochschild 1990: 120) Emotions erupt during social interaction. Then they are judged for suitability according to cultural and ideological standards, and managed to effect culturally acceptable displays that yield social accord.

Affect control theory (Heise, 1977, 1979, 1986; Smith-Lovin & Heise, 1979; Smith-Lovin, 1990; Mackinnon & Heise, in press) proposes that people construct and understand social action so as to have important cultural meanings affirmed by impressions generated in manifest behavior. People credit themselves and others with specific identities during social encounters. They then engage in physical and mental work so that events create impressions that maintain sentiments attached to their identities, as well as to other categories of action (i.e., behaviors, settings, and personal conditions emotions included).

Sentiments for all categories of action can be measured quantitatively on three culturally universal dimensions of affective meaning (Osgood, May, & Miron, 1975):

evaluation (the extent to which things seem good versus bad);
potency (impressions of powerfulness versus powerlessness); and
activity (impressions of activation versus tranquillity).
The same dimensions serve for measuring transient impressions of people, behaviors, settings, and personal conditions, as generated by particular events. The theory proposes that people seek experiences in which transient Evaluation-Potency-Activity impressions, created by an event, match their pre-existing sentiments as much as possible. Since research indicates that the impressions generated by an event are transformations of feelings that exist before the event (Smith-Lovin, 1987), people seek experiences that transform current feelings into new, sentiment-confirming feelings.

In affect control theory, emotions are momentary personal states that reflect how events affect people. The emotion depends on the current impression of the person, and on how that impression compares to the sentiment attached to the person's identity. For example, a person might feel overwhelmed or anxious if made to seem bad and weak, but a more extreme response feeling ashamed, desperate, or depressed is to be expected if the person has a particularly good and powerful role in a group.

Affect control theory has an empirical base and is implemented mathematically in a computer program (Heise & Lewis, 1988) that allows prediction of the probable emotions resulting from events. For example, in an analysis where one person is identified as a father and the other as a daughter, we can specify that the father educates the daughter (a behavior predicted by the affect control model). The theory predicts that the father then will feel generous, secure, or forgiving while the daughter feels humble, relaxed, touched. Forcing the daughter to disobey the father leads to predictions that the daughter must be feeling irate, angry, anxious, and that the father feels melancholy, apprehensive, shocked. These particular analytic outcomes are based on Southern U.S.A. sentiments, and predictions would be different with sentiment measurements from other cultures. The affect control simulation system currently allows analyses to be conducted for U.S.A., Canadian, Irish, German, and Japanese interactants.

People protect their pre-existing sentiments both through their own conduct and through their interpretations of others' conduct. For example, continuing the father-daughter example, the father could exonerate the daughter who disobeyed him, and his action would begin to re-affirm pre-existing sentiments about fathers and daughters. Alternatively, according to analyses, he could begin to see his daughter as greedy, manipulative, or mean reconceptualizations that match pre-existing sentiments to the girl's behavior. Thus, in affect control theory, a socially undesirable emotion can be changed by implementing a new event that will replace the unwanted emotion with a different emotion, or by reinterpretation of a past event so that the emotion produced by the original interpretation is replaced by another emotion.

Thoits' (1990) typology of emotion management techniques includes implementation of new events and reinterpretation of past events as ways of managing emotions through reconstruction of the situations that evoke the emotions. Thoits notes that emotions also can be managed directly by manipulating accompanying physiology (as with drugs), by performing or fantasizing expressive gestures for desired emotions, or by re-categorizing one's affective sensations in terms of a desired emotion.

Spontaneous eruptions of emotion arise from appraising security, control, and stimulation with reference to a personal identity, and the consequent emotion may not accord with emotion norms appropriate to a publicly negotiated identity (Heise & MacKinnon, in press). The personal and the normative systems unite when group members are deeply committed to their group identities; in that case, people spontaneously emote and act according to group norms in order to experience affirmation of self through the reflected appraisals of others (Burke & Reitzes, 1991). The two systems diverge when a person maintains multiple definitions of a situation simultaneously, and the actor's deepest commitment is to an identity other than the public identity. In that case, emotion management is required to prevent the display of emotions appropriate to the private identity, and to authenticate one's supposed commitment to the public identity.

Hochschild's (1983) study of emotion work among flight attendants illustrates how people have to manage emotions in order to hide tacit definitions of a situation and their alienation from the public definitions of the situation. In the heyday of friendly airline service, attendants were instructed to treat passengers as if they were family and friends in their own living rooms, and the attendants deeply committed to their prestigious roles were able to do so and gain satisfaction from their work. "When feelings are successfully commercialized, the worker does not feel phony or alien; she feels somehow satisfied in how personal her service actually was" (Hochschild, 1983: 136). However, as airlines began encountering financial problems in the early 1970s and demanded more cost-efficient flying, attendants had to work longer shifts and cope with more people. Then deregulation of airlines dropped the price of tickets and brought more difficult passengers on board (e.g., children, frightened elderly, people unacquainted with airplane travel) and further increased the work load for attendants. Hochschild (1983: 121) summarizes the consequences as follows. When an industry speed-up drastically shortens the time available for contact between flight attendants and passengers, it can become virtually impossible to deliver emotional labor. In that event, the transmutation of emotion work, feeling rules, and social exchange will fail. Company claims about offering a smile 'from the inside out' (Delta) will become untenable. The living room analogy will collapse into a flat slogan. The mosaic of 'as if' techniques will fall to pieces, and deep acting will be replaced by surface displays that lack conviction. Nowadays airlines demand overwhelming housework and serving from flight attendants, leaving little time for pleasantries. The ideal of the friendly attendant still is advertised by airlines, but attendants have little commitment to this role, which they express with superficial smiles. Their commitment is to a job with lowered status, in which they see passengers as crowds of strangers, a job that yields private feelings of estrangement, even hostility, as they do their work.

A Challenge for the Future

Sociologists accumulated extensive evidence that interpersonal dynamics are laden with emotional processes during decades of research based on laboratory studies, participant observation, and ethnographic projects. Attempts to organize the data has produced a rich body of sociological theory that relates affect to social processes and social organization.

Currently sociologists are returning to empirical study to test and refine their theories. Kemper (1991) found that his power-status model could account successfully for more than two-thirds of emotion episodes self-reported by some of the respondents in an eight-nation study of emotions (Scherer, Wallbott, & Summerfield, 1986). At the time of this writing, Lynn Smith-Lovin and her students are demonstrating high levels of success for affect control theory's predictions of emotions in social encounters described in vignettes.

The challenge now is to develop sensitive measures of emotion for studying rapid changes in affect that occur during social interaction in natural groups. One approach is to employ the latest in sound-image recording technology and code non-verbal emotional expressions (Scherer & Ekman, 1982), drawing especially on Ekman and Friesen's (1978) explication of facial expressions. Meaningful sociological analyses of emotion also require assessing people's definitions of situations and their appraisals of events: such meanings might be tapped by replaying recordings for participants and eliciting their interpretations of what was happening during social interaction. It remains to be seen whether social psychologists can develop less costly, labor-intensive, and intrusive techniques for assessing emotions and operative meanings in group encounters.



  1. We gratefully acknowledge help from Peggy Thoits in structuring this essay and from Neil MacKinnon in articulating some ideas.