Positional and Personal Control Systems


Element Compared

Main means of social control

Positional Control System

Manipulation of incentives and social coercion

Personal Control System

Persuasion and manipulation of incentives

Aim of social control


Feelings, thoughts, and intentions

Psychological habits developed

Learn obedience in behavior and outward action. Emotion work less necessary

Learn to be subject to persuasion, learn to persuade self and others, learn emotion work


Stresses outward behavioral conformity

Progressive, stresses own feelings and intentions

Occupation calls for

Behavior, action and its products

The management of meanings and feelings; with role closeness or role distance

Social class

Traditional working class (both men and women) and technical sector of middle class

Upper and middle class (mainly men); new working class (mainly women)


[1] For stylistic convenience, I shall use the pronoun “she” when referring to a flight attendant, except when a specific male flight attendant is being discussed. Otherwise I shall try to avoid verbally excluding either gender.

[2] I use the term emotional labor to mean the management of feeling to create a publicly observable facial and bodily display; emotional labor is sold for a wage and therefore has exchange value. I use the synonymous terms emotion work or emotion management to refer to these same acts done in a private context where they have use value.

[3] We may misinterpret an event, feel accordingly, and then draw false conclu­sions from what we feel. (We sometimes call this neurosis.) We can handle this by applying a secondary framework that corrects habits of feeling and inference, as when we say “I know I have a tendency to interpret certain gestures as rejections.” But feeling is the essential clue that a certain viewpoint, even though it may need frequent adjustment, is alive and well.

t A black person may see the deprivations of the ghetto more accurately, more “rationally,” through indignation and anger than through obedience or resigned “realism.” He will focus clearly on the policeman’s bloodied club, the landlord’s Cadillac, the look of disapproval on the employment agent’s white face. Outside of anger, these images become like boulders on a mountainside, minuscule parts of the landscape. Likewise, a chronically morose person who falls in love may suddenly see the world as happier people do.

[4] It is not only in the world of commerce that we automatically assume insin­cerity. Political reporters regularly state not only what an officeholder or candidate wants to seem to feel but also how well he or she succeeds in the effort to convey that feeling. Readers, it is assumed, demand at least this much unveiling.

[5] In An Actor Prepares, Stanislavski points out an apparent contradiction: “We are supposed to create under inspiration; only our subconscious gives us inspira­tion; yet we apparently can use this subconscious only through our consciousness, which kills it” (1965, p. 13). The solution to this problem is the indirect method. The subconscious is induced. As Stanislavski notes: “The aim of the actor’s preparation is to cross the threshold of the subconscious…. Beforehand we have ‘true-seeming feeling,’ afterwards ‘sincerity of emotion.’ On this side of it, we have the simplicity of a limited fantasy; beyond, the simplicity of the larger imagination, [where] the cre­ative process differs each time it is repeated” (p. 267).

[6] We commonly assume that institutions are called in when individual controls fail: those who cannot control their emotions are sent to mental hospitals, homes for disturbed children, or prisons. But in looking at the matter this way, we may ignore the fact that individual failures of control often signal a prior institutional failure to shape feeling. We might ask instead what sort of church, school, or family influence was unavailable to the parents of institutionalized patients, who presum­ably tried to make their children into adequate emotion managers.

[7] Scientific writing, like scientific talk, has a function similar to that of covering the face and genitalia. It is an extension of institutional control over feeling. The overuse of passive verb forms, the avoidance of “I,” the preference for Latinate nouns, and for the abstract over the concrete, are customs that distance the reader from the topic and limit emotionality. In order to seem scientific, writers obey conven­tions that inhibit emotional involvement. There is a purpose in such “poor” writing.

[8] I heard the rationale for this company regulation discussed in class on Febru­ary 19, 1980. (It was also stated in the training manual.) Whether it has ever been enforced, and with what result, I don’t know.

[9] We may also believe that there shouldn’t be a feeling rule in a given instance. One father, for example, reported: “When Jeffrey was little, and squalled intermi­nably one morning—I felt like throwing him on the floor. I was horrified at my rage. But I told myself, it’s all right to feel the rage. It’sjust bad to acton it.”

[10] This raises the issue of display and display rules. It raises the issue of the “falseness,” as distinct from the “wrongness,” of a feeling. Wrongness refers to a discrepancy between “what I do feel and think” and “what I should feel and think.” Falseness refers to a discrepancy between “what I do feel and think” and “what I appear to feel and think.” For example, the bride may say “I’m so happy” with such a forced smile that she seems false to others. One of the display rules at weddings is that the bride should seem natural and unforced.

[11] What can be expected (at this stage, on this occasion) and what is wished for in experience deserve a certain analytic separation. But in the American middle class, there may be an “optimism norm” so that what we realistically expect and what we think is ideal are closer together than they are in other classes and other cultures.

[12] Private gender relations have a floorboard, which is the prevailing arrange­ment between the sexes in the larger society. An equalitarian couple in a society that as a whole subordinates women cannot, at the basic level of emotional ex­changes, be equal. For example, a woman lawyer who earns as much money and respect as her husband, and whose husband accepts these facts about her, may still find that she owes him gratitude for his liberal views and his equal participation in housework. Her claims are seen as unusually high, his as unusually low. The larger market in alternate partners offers him free household labor, which it does not offer her. In light of the larger social context, she is lucky to have him. And it is usually more her burden to manage indignation at having to feel grateful.

[13] When an airline commands a market monopoly, as it is likely to do when it is owned by a government, it does not need to compete for passengers by advertising friendly flight attendants. Many flight attendants told me that their counterparts on Lufthansa (the German national airlines) and even more on El AI and Aeroflot (the Israeli and Russian national airlines) were notably lacking in assertive friendliness.

t A black female flight attendant, who had been hired in the early 1970s when Delta faced an affirmative action suit, wondered aloud why blacks were not pic­tured in local Georgia advertising. She concluded: “They want that market, and that market doesn’t include blacks. They go along with that.” Although Delta’s cen­tral offices are in Atlanta, which is predominantly black, few blacks worked for Delta in any capacity.

[14] Most anger fantasies seemed to have a strong oral component, such as befoul­ing the troublemaker’s food and watching him eat it. These fantasies inverted the service motif but did not step outside it. No one, for instance, reported a fantasy about hitting a passenger.

[15] By some accounts, the company’s play on our culture’s devaluation of age in women made older female workers feel obliged to “make up” for their age by work­ing harder. There were some stories of direct harassment of older female flight at­tendants. One supervisor was reported to have asked a woman to take off her jacket and hold out her arms; he then remarked on the “unsightliness” of the flesh on the under side of her upper arms. Although the woman was personally distressed by this, another flight attendant and union official remarked: “They make us think age is a personal flaw. Actually, theyjust don’t want to pay our pensions.”

[16] Certain features of work not mentioned in job descriptions —such as incen­tive systems that join self-interest to worked-on display and feeling—may be espe­cially successful in promoting emotional labor. Salespersons working on commis­sion are a prime example. In the absence of clear self-interest, close supervision probably helps foster emotional labor most of all.

[17] A child asked to “love Aunt Hilda” might rebel by refusing to love Aunt Hilda. The child asked to feel ambitious and “love school” might rebel by hating school and disdaining success. R. D. Laing in his Politics of the Family (1971) draws attention to this middle-class “internal” mode of control by showing how parents and psychiatrists set feeling rules and how children and patients rebel against them. If authority in the middle class is more expressed through feeling rules and emotion management—if it is more through these than through rules of outer behavior that we are governed—then we would do well to examine, as Laing does, rebellion as rebellion against dictates in this realm.

t One latent message in the free-school education of the 1960s, designed al­most exclusively for middle-class students, was that personal feelings are near­sacred objects of attention and deserve frequent and detailed discussion. See Swidler(1979).

[18] Similarly, the social guardians of the positional control system are found not only in working-class families but in the traditional churches to which they go, and to some extent in the schools, where they learn to manage their behavior in ways that will be useful on the job.

sible), we might expect women to be more in touch with their feelings, which are, as a consequence, more available for conscious management. See Chodorow (1980). Men may manage feelings more by subconscious repressing, women more by con­scious suppressing. ‘

[20] Fiedler (1960) suggests that girls are trained to be “seriously” good and to be ashamed of being bad whereas boys are asked to be good in formalistic ways but covertly invited to be ashamed of being “too” good. Oversocialization into “sugar – and-spice” demeanor produces feminine skills in delivering deference.

[21] Other researchers have found men to have a more “romantic” orientation to love, women a more “realistic” orientation. That is, males may find cultural support for a passive construction of love, for seeing themselves as “ falling head over heels,” or “walking on air.” According to Kephart, “the female is not pushed hither and yon by her romantic compulsions. On the contrary, she seems to have a greater measure of rational control over her romantic inclinations than the male” (1967, p. 473).

t This pattern is also socially reinforced. When women sent direct messages (persuading by logic, reason, or an onslaught of information), they were later rated as more aggressive than men who did the same thing (Johnson and Goodchilds 1976, p. 70).

[22] The use of feminine wiles (including flattery) is felt to be a psychopolitical style of the subordinate; it is therefore disapproved of by women who have gained a foot­hold in the man’s world and can afford to disparage what they do not need to use.

[23] Celebrating male humor or enhancing male status often involves the use of what Suzanne Langer has called nondiscursive symbols, “symbols which are not verifiable, do not have dictionary meanings or socially defined syntax and order” (Langer 1951, 1967).

[24] Zick Rubin’s study of young men and women in love relationships (generally middle-class persons of about the same age) found that the women tended to ad­mire their male loved ones more than they were, in turn, admired by them. The women also felt “more like” their loved ones than the men did. (See Rubin 1970; Reiss 1960.)

[25] The code of chivalry is said to require protection of the weaker fry the stronger. Yet a boss may bring flowers to his secretary or open the door for her only to make up for the fact that he gets openly angry at her more often than he does at a male equal or superior, and more often than she does at him. The flowers symbolize redress, even as they obscure the basic maldistribution of respect and its psychic cost.

[26] More women than men go to doctors, and this might seem to explain why doctors take them less seriously. But here it is hard to tell cause from effect, for if a woman’s complaints are not taken seriously, she may have to make several visits to doctors before a remedy is found (Armitage et al. 1979).

[27] The management of American Airlines objected to a union request that men be allowed to wear short-sleeved shirts on warm days, arguing that such shirts “lacked authority.” As one female union representative quipped at a union meeting, “But since only male flight attendants have authority anyway, why should it matter?”

[28] Gay males apparently did not fit this general pattern. Although they were treated by the public as males and thus commanded more respect, they did not use this fact in the same way in their relations with female co-workers. Perhaps their anticipation of company and public prejudice against homosexuality led them to adjust the value of their respect currency to that of their female co-workers. This considerably eased relations between them and female workers. One woman worker said: “The gay stewards are great. If Pan Am had any sense, it would prefer to hire them.”

[29] The other side of being called a “girl” was not being allowed, socially speak­ing, to age. Even women in their thirties were occasionally called “granny” or sub­jected to within-earshot remarks such as “Isn’t she about ready for retirement?” As one woman in her mid-thirties noted: “There is definitely a difference, oh yes. The men take it for granted that they can work until sixty or sixty-five. The women work like dogs just to prove they can still do the job. And then they have to fight the granny remarks.”

[30] People want to be their “authentic” selves. As Marshall Berman has put it: “To pursue authenticity as an ideal, as something that must be achieved, is to be self­consciously paradoxical. But those who seek authenticity insist that this paradox is built into the structure of the world they live in. This world, they say, represses, alienates, divides, denies, destroys the self. To be oneself in such a world is not a tautology but a problem’ (1970, p. xvi).

[31] “If sincerity has lost its former status, if the word itself has for us a hollow sound and seems almost to negate its meaning, that is because it does not propose being true to one’s self as an end but only as a means” (Trilling 1972, p. 9).

[32] In the early Freudian model, a lack of reflexivity implies that ego cannot much alter the character of emotion. Sometimes this is explicit: Alexander and Isaacs note, “ It seems unlikely that the ego changes the quality of the affect” (1964, p. 232). Often buttressing this view is the notion that the ego is weak, as it is for the child. For the interactionist, the prototypic ego is that of the normal adult, and it has a moderate amount of strength.

[33] Collins gets Darwin right but Durkheim wrong. Having imputed a stress on animal instincts to Durkheim (in The Elementary Forms of the Religious Life [ 1965]), he presents himself as drawing from Durkheim this heritage from Darwin. He wants to link Durkheim to Darwin via an interest in the similarity of animals and human beings (Collins 1975, p. 95). In fact, whereas Darwin stresses the similarity between humans and other animals, Durkheim stresses their differences. Animals cannot symbolize, and so Durkheim was not very interested in them.

t Thomas Scheff, in his essay “The Distancing of Emotion in Ritual,” draws on Freud’s early notion of catharsis and with it the idea of emotion as the “discharge" of one or more distressful emotions (grief, fear, embarrassment, anger). These emotions, he notes, are “physical states of tension in the body produced by stress” (Scheff 1977, p. 485). See also Hochschild (1977) and Glover (1939).

[34] We do not name feelings after physiological states, for good reason. It has long been known that physiologically, anger has much in common with fear (Schachter and Singer, 1962). Physiological differences are not pronounced enough from one feeling to another to account for the wide variety of emotion names we have in our language. Such differences can at best distinguish between general families of emotion.


Of the twelve standard occupational groups used by the U. S. Census, six contain the majority of jobs that call for emo­tional labor, as defined in Chapter Seven. These six groups, summarized in Table 1, are as follows: professional and technical workers, managers and administrators, sales work­ers, clerical workers, and service workers of two types — those who work inside and those who work outside of pri­vate households. In one way or another, probably most sales workers, managers, and administrators are called upon to do some emotion work. But among those in the professions, service work, and clerical work, only selected jobs seem to involve substantial amounts of emotional labor (see Tables 2, 3, and 4). Within these categories are some of the most rap­idly growing occupations. According to the U. S. Bureau of Labor Statistics, there will be a 30 percent growth in the 1980s for social workers, 25 percent for preschool teachers, 45 percent for health administrators, 33 percent for sales managers, 79 percent for flight attendants, and 35 percent for food-counter workers. The largest number of new jobs are expected in the retailing sector, especially in department stores and restaurants (New York Times, October 14, 1979, p. 8). Given the roughness of occupational categories, the fit of emotion work criteria to occupation and to labor produced is necessarily loose. The tables presented here are no more than a sketch, a suggestion of a pattern that deserves to be examined more closely.

Table 1 shows the number of jobs in all six occupational categories in 1970. It also shows the number of men and women in these categories. Over all, women are overrepre­sented in jobs calling for emotional labor; about half of all working women hold such jobs. Men are underrepresented; about a quarter of all working men are in emotional labor jobs. This is true for professional and technical occupations, for clerical occupations, and for service-sector jobs as well.

Table 2 examines fifteen occupations that involve sub­stantial amounts of emotional labor, selected from the twenty-seven different occupations grouped as Professional, Technical, or Kindred by the U. S. Census. It computes the proportion of all professional and technical jobs that in­volved substantial amounts of emotional labor in 1970, and shows variations by sex. Tables 3 and 4, respectively, per­form the same kind of analysis for clerical workers and for service workers outside of private households.

TABLE l. Summary Estimate of Jobs Most Calling for Emotional Labor, 1970





Professional, technical and kindred3




Managers and administratorsb




Sales workersb




Clerical and kindredc




Service workers excluding private householdd




Private household workersb




Total number of jobs calling for emotional labor




Total size of employed labor force over 14 years of age




Jobs involving




substantial emotional labor as a percentage of all jobs

aSelected occupations; see Table 2. bAII jobs.

cSeIected occupations; see Table 3. ^Selected occupations; see Table 4.

NOTE: Tables 1 to 4 enumerate the number of employed persons, 14 years or older, by occupation, from the 1970 U. S. Census.

source: U. S. Bureau of the Census, i, “Census of the Population: 1970,” Vol. 1, Characteristics of the Population, Part l, United States Summary Section I, Table 221. Detailed Occupation of the Experienced Civilian Labor Force and Employed Persons by Sex (Washington D. C.: U. S. Government Printing Office, 1973), pp. 718-724. ‘

TABLE 2. Detailed Occupational Analysis of Selected Professional, Technical, and Kindred Workers, 1970





Lawyers and judges








Personnel and




labor relations

Registered nurses








Dental hygenists




Therapy assistants




Clergymen and




religious workers Social and




recreation workers

College and




university teachers Teachers, except college




and universities Vocational and






Public relations and




publicity writers Radio and television





Physicians, dentists, and




related personnel

Total number of




persons employed in selected professional, technical, and kindred occupations (18 jobs)

TABLE 2. Continued





Total number of

persons employed in all professional, technical, and kindred occupations (34 jobs)




Percentage of all professional, technical and kindred jobs involving substantial amounts of emotional labor




TABLE 3. Detailed Occupational Analysis of Selected Clerical and Kindred Workers, 1970





Bank tellers








Clerical supervisors




Bill collectors




Counter clerks,




excluding food

Enumerators and





Insurance adjustors




and examiners

Library attendants




Postal clerks
















Teachers aides




Telegraph operators




Telephone operators




Ticket agents




Total number of




persons employed in selected clerical and kindred occupations

Total number of




persons employed in all clerical and kindred occupations

Percentage of all clerical




and kindred jobs involving substantial amounts of emotional labor

TABLE 4. Detailed Occupational Analysis of

Selected Service Workers, Except Private Household, 1970









Food counter and




fountain workers





Health service workers3




Personal service





Child care workers




Elevator operators




Hairdressers and









(excluding private household)

School monitors




Ushers, recreation and





Welfare service aides




Protective service




workers c

Total number of




persons employed in selected service

worker occupations, except private household





Total number of persons employed in all service worker occupations except private household




Percentage of all service sector occupations, except private household, involving substantial amounts of emotional labor




^Includes dental assistants; health aides, except nursing; health trainees; lay mid­wives; nursing aides, orderlies, and attendants; practical nurses.

^Includes airline stewardesses, recreation and amusement attendants, personal service attendants not elsewhere classified, baggage porters and bellhops, barbers, boarding and lodging housekeepers, bootblacks.

includes marshals and constables, policemen and detectives, sheriffs and bailiffs.


We do not have names for all the possible combinations of primary and background focuses. No one culture has a mo­nopoly on emotions, and each culture may offer its own unique feelings. As the Czech novelist Milan Kundera wrote in The Book of Laughter and Forgetting: “Litost is a Czech word with no exact translation into any other language. It desig­nates… a feeling that is the synthesis of many others: grief, sympathy, remorse, and an indefinable longing…. I have never found an equivalent in other languages for this sense of the word either, though I do not see how anyone can under­stand the human soul without it.” Only by referring to several points of focus, several inner contexts, can Kundera suggest the quality of litost, a “state of torment caused by a sudden insight into one’s own miserable self” (1981, pp. 121,122).

Why have we only the names for feelings that we have in English? Why should our set of feeling names vary from the “inventions” of feeling names in Arabic or in German? Ed­ward Sapir has noted that codified discriminations of vari­ous sights, sounds, and tastes vary culturally. The names for emotions also vary culturally.

We have names for many ways of focusing on the object of blame for a frustrating event: anger, resentment, rage, exas­peration, irritation, and indignation. A society with less ten­dency to attribute blame to people and objects outside the self might have fewer names for doing this. Consider for a moment the cultural and structural story behind our words pride, shame, and pity.

Pride. The word is opposite to shame (which implies focus on an outside audience) and also to guilt (which implies fo­cus on the self). There might be, though in English there are not, separate words for pride with and without a focus on an audience. There might be a special name for pride based on recognition actively bestowed by some known group (as in the case of honor), as opposed to recognition derived by im­personal means (Speier 1935). In Yiddish, the language of a highly familistic social group, there is a special word for pride in one’s family, and especially in one’s children — nachus. The focus to which the word refers is dual: “My chil­dren have done gloriously” and “I am tied to my children.” In English there is no word specifically designated for pride – in-my-children, or, for that matter, pride-in-my-community, or pride-in-my-political group.

Shame. There are “seeing rules” about being watched that correspond to systems of social control. Under control, the rule is to notice whether or not one is being watched. Under more impersonal social control, the rule is to notice imper­sonal rules and less attention is paid to the watchfulness of intimates. There may be a corresponding decline in an ac­tual focus on watchers, hence a decline in the experience of shame and a decline in the number of names for it (Benedict 1946-b).*

Pity. The phrase “take pity” came into common use with the establishment of the Christian church. The church, in turn, came to power in an age known for extreme differences in wealth and generally brutalizing conditions of life. There were, in addition, communal ties between people who were in dire need—widows, orphans, the elderly—and people who could “take pity” and provide for them. Now that almsgiving has been bureaucratized so that giver and receiver remain unknown to each other, the perceptual focus that corres­ponds to pity is less common (Allport and Odbert 1936).

As certain social conditions, habits of seeing, and namable feelings fade from a culture, others enter. In the last twenty years a group of new terms for emotional states has sur­faced. For example, “being on a bummer, being turned off, being turned on, being on a downer, being freaked out, hav­ing one’s mind blown, being high.” Many seem traceable to the 1960s drug culture. Whatever their origins, these new names for psychological states have been generalized and

* There is a corresponding poverty of namable feelings that have to do with empathy. We can focus not only on our own situation but on that of another person whose situation can vary in the ways our own can. We can feel empathetic sadness, empathetic frustration, empathetic anger, empathetic resentment, empathetic fear, empathetic guilt, empathetic anguish, and so on. Oddly, there are no separate terms for these potentially namable feelings.

adopted by a wider middle-class population, for whom their more impersonal focus on the dimension of tension relaxa­tion may serve some function.

Here, two social trends come together. With the wide­spread use of contraceptives and the legitimation of their use (“the sexual revolution”), more men and women sleep with each other in the early stages of acquaintanceship. Yet the re­sidual custom of getting to know each other before making love still makes its claims. “Psycho-babble,” as the genre has been called, may have emerged as a way to pay respects to the old custom of getting acquainted nonsexually while at the same time embracing more sexual permissiveness. Psycho­babble is ideally suited for bridging this contradiction: its terms for feeling states seem intended to reduce social dis­tance through personal revelation, but they are used in such vague and undifferentiated ways as to make the communica­tion a ritual formality, low in personal content. A couple ex­changing the confidences of psycho-babble may not be get­ting personally acquainted any faster than their parents or grandparents did, when they said less about themselves but in more revealing language. It is worth noting that the language of psycho-babble corresponds to the language of the airline personnel manual: one is an experiential guide to private oc­casions, the other a guide to commercial occasions. In com­paring namable feelings to named ones we can garner clues to the links between larger social arrangements and com­mon ways of seeing and feeling.

CHART 1. Emotion Name and the Individual’s Momentary Focus

Emotion Name

What I Want (Like)

What I Have (See I Have)

What I Approve

The Causal Agent of Event, Object

The Relation of Self to Causal Agent


“I loved X —I still

“What I don’t have,


love X now”

what is gone,




(Primary focus)


“I loved X or love

“It’s in the

X, a past

irretrievable past”


(Secondary focus)

thing” (Primary



“I want to maintain

“I have suddenly now

“I disapprove”

a good image

a bad one”


of myself”


“I want this now”

“It’s not there”; “It

(Primary focus)

could be there,

it’s not”


Focus on

“ You hit me”

“I feel as or more



powerful than


UV, LVV V-.V-. 11 Uttllllllg

and having (Frustration – secondary focus) “I want safety”

“What I see makes me unsafe”


(Same as anger)

(Same as anger)


“I don’t like this”

“I see I have this”


(Same as above)

(Same as above)


"■*” —….

me; I can or could attack”

“I see the

“I feel utterly

cause of the

powerless to do


anything about


this; X is more

“And I disapprove”


powerful than I”

“I disapprove”

“Here it is close to


me” and"I want to get away from it” (Primary focus)

“I disapprove”

“X is the cause

“And X is

of this bad

beneath me”






What I Have

The Causal Agent

The Relation of

Emotion Name

What I Want (Like)

(See I Have)

What I Approve

oj Event, Object

Self to Causal Agent


(Same as above)

(Same as above)

“I disapprove”

“The cause

“I am the cause”

of the unwanted event is bad”


(Same as above)

(Same as above)

“I am responsible

for the bad

thing” and"I want to undo it

and can’t”




“I see what I want”

“I don’t have it”


person has it” (Primary focus)


“I have claim to

“I might lose what I

“There is the

what I want”

have” or “I have lost

robber or

what is mine”



Love, liking

“I want XY”

“You give or represent to me XY”


“This other person wants X”

“This other person does not have X”


(Same as above)

(Same as above)


“I want to seem a

“I see the behavior or


particular way to others”

events are discrepant with how I want to seem to others” (Primary focus)


“I want to do right, good things”

“I have done wrong or bad things”


Only vaguely sensed “I want X”

“I don’t know if I will get X”



“I like this other


“This person is beneath me” (Liking variable)

“I see in detail the

audience of this




“I am the

“I see the

cause of

audience of

the event”

this: they are better than I”

“I don’t know”

“I don’t know”

I disapprove”

I don’t know”


In Appendix A, I offer a review of research on emotion and my own three-part account of emotion. In this appendix, I examine the principle according to which we name feeling.

To name a feeling is to name our way of seeing something, to label our perception[34] As we see in Appendix A, perception is not all there is to emotion or feeling, nor is it its sole cause, but it is the principle according to which emotion and feeling are named. This is the idea advanced by the cognitive psy­chologist Judith Katz (1980). I develop it here to show that when we do not feel emotion, or disclaim an emotion, we lose touch with how we actually link inner to outer reality.

This theory of emotion naming is an elaboration of what I have said at the end of Appendix A about the social in­fluences on the “signal function” of feeling. Feelings signal not only a newly apprehended reality (outer or inner) but what that reality impinges upon—our prior self and expec­tations. Now I want to turn this idea around and argue that the names we give emotions refer to the way we apprehend a given situation—the aspect of it we focus on—and what our prior expectations about it are. In short, feeling signals per­ception and expectation to us, and turning this around, dif­ferent patterns of perception and expectation correspond to different feeling names. Since culture directs our seeing and expecting, it directs our feeling and our naming of feeling.

Thus what feelings “signal” to us as sociologists is how culture in­fluences what we feel and how we name it.

In my attention to patterns of perception and expectation, I may seem to imply that people actively choose to focus and expect as they do. Sometimes—when people are under the directorship of Stanislavski or In-Flight Training, they do. But for the most part, we see and expect in ways we do not ac­tively direct and in ways we are often totally unconscious of.

How do we name feeling? It seems artificial and simplis­tic, often, to apply only one name to what we feel. We can feel angry, guilty, disappointed, and frustrated, all with refer­ence to the same event. This does not mean that we are mo­mentarily possessed of a certain mixture of physiological states or expressions. It means, instead, that from moment to moment we are focusing on different features of the situa­tion. Compound emotions are serial perceptions. As Katz rightly points out, when we reminisce, the mind’s eye moves from one point to another; the multiplicity of the emotions we name results from this movement of focus.

Moving our attention from one point to another in a field of details brings together one interface after another be­tween inner and outer reality. We are always wanting or ex­pecting something, but we are not always attending to all the crystallizing details of a situation simultaneously. We hold, at most, two main points of focus and thus keep two facets of a situation in mind at the same time. We focus on one facet in light of another, with other facets providing background.

Suppose, Katz suggests, an old and beloved friend of mine is killed in a car accident. My state of grief is not a condensed experience of sadness but the continual suscepti­bility to it as I reminisce. When I focus on the thought “I love him and want him” in light of the thought “he is dead now,” we call what I feel sadness. But if I focus on “I love and want him” while at the same time (through religious conviction or denial) disbelieving the evidence of his death, I do not for that moment feel sad. If through the voyages of the mind I chance upon the thought “but we had precious good times and I have those memories,” for that moment we call what I feel happy and grateful. If I see “our precious good times” in light of the thought “but they’re gone and lost now,” nostal­gia is the name for what I feel.

Prior opinions or assumptions further differentiate named feeling. For example, when I consider the other friends of the accident victim and imagine their loss, what I feel depends on how I have regarded them in the past. If I have considered them to be equals, what I feel is compassion. If I have considered them inferior in some sense, what I feel is pity.

If I dwell neither on the cause of the loss nor on the object of it (the friend) but on the intermediate fact that this trag­edy has simply happened to me, I feel frustration. I dwell on “I’m not getting what I want,” set apart from notions of why I’m not getting it. But if I focus on the cause (the driver of the car that killed him), I feel what we call anger.

Developing Katz’s idea, Chart 1 at the end of this appen­dix describes some common emotions collected from over four hundred names of emotions and sentiments found in Roget’s Thesaurus and The Random House Dictionary of the En­glish Language. Corresponding to each emotion name are five general categories of perceptual focus: (1) what I want, or like, or am attached to; (2) what I now see myself as hav­ing; (3) what I approve or disapprove; (4) the perceived causal agent of an event or object; and (5) the relation of myself to the causal agent. Each emotion has two main points of focus; about half of them have additional periph­eral points of focus. Let us consider a few examples of what Chart 1 attempts to explain.

In sadness, I am focusing on what I love, like, or want and also on the fact that it is not available to me. I do not focus on what has caused the loss or absence nor on my relation to the cause of the loss. I do focus on my relationship to the loved object. In nostalgia, the focal points are the same but the fo­cus on the love or liking is stronger than the focus on what is gone, which adds sweetness to the bitterness of plain grief. In frustration, the focus is not on what I want that I don’t have but on the self in this state of not having; the focus is on my not having rather than on the wanted thing.

Anger, resentment, indignation, contempt, guilt, and an­guish all correspond to different patterns of focus on the cause of frustration and on my relation to this cause. If I feel as powerful or more powerful than the blameworthy party on whom I focus, we say I feel anger. If I see the causal agent as very much more powerful than myself, we say I feel fear. (The brave are those who nurture the idea, or illusion, that they are as powerful as the agent of any threat to them.) In­dignation is a name for adding a focus on a thing that is dis­approved of; contempt is a name for adding a focus on one’s social or moral superiority. Guilt is a name for seeing our­selves as the author of an unwanted event. Envy is a name for noting what we do not have but want and noting further that another has it. Jealousy is a name for a focus on the threat to something that we already suppose we possess.

We call it love when we focus on the desirable qualities of a person or thing and on our closeness to him, her, or it. We call it admiration when we focus on the desirable qualities of the person in light of some attention to social distance. In awe, we take note of much greater social distance. As with all emotion, it is not that awe is a compound of entities in the sense that chemicals are compounds. What is combined are particular twists and turns of moment-to-moment noticings that lend context to seeing. As in the case of ail emotions, too, what is noted is experienced as relevant to the self. The emotion tells exactly how.


Goffman has carried the conceptual heritage of Dewey and of Gerth and Mills as far as he can without leaving his behavior­ism and his “moments and their men” perspective. But now we need a theory that allows us to see how institutions—such as corporations—control us not simply through their surveil­lance of our behavior but through surveillance of our feelings. Such a social theory of emotion must have both a social and a psychological side. It can start by extending the ques­tion Gerth and Mills ask: How do institutions influence per­sonality? But we may specify that question: How do institu­tions control how we “personally” control feeling? In pursuit of an answer to this question, I draw, as Gerth and Mills did, on Weber’s appreciation of the power of bureau­cracy and on Marx’s sense of the interests that a bureaucracy actually serves. I also draw frequently on C. Wright Mills’s focus, in White Collar, on the “sale of personality.” But I add to Mills the notion that a personality is not simply “sold”; people actively manage feelings in order to make their per­sonalities fit for public-contact work. I also add three ele­ments found in Goffman: the focus on rules, the perspective of the affective deviant (the worker who is not obeying the feeling rules of the workplace), and an awareness of the ef­fort it takes to pay our “emotional dues” to an occasion.

On the psychological side, a social theory of emotion must take into account that these emotional dues can be costly to the self. Institutional rules run deep but so does the self that struggles with and against them. To manage feeling is to ac­tively try to change a preexisting emotional state.

But then we must ask: What is emotion? Emotion, I sug­gest, is a biologically given sense, and our most important one. Like other senses—hearing, touch, and smell—it is a means by which we know about our relation to the world, and it is therefore crucial for the survival of human beings in group life. Emotion is unique among the senses, however, because it is related not only to an orientation toward action but also to an orientation toward cognition.

The connection of emotion to an orientation toward action was key for Darwin. Indeed, he defined emotion as something quite close to this: as a protoaction, as what oc­curs instead of or before an action, as an action manque. An­ger, Darwin suggests, is the preact or prelude to killing, and love is the prelude to copulation; and we may add that envy is the prelude to stealing, gratitude the prelude to giving back, and jealousy the prelude to excluding. Emotion, there­fore, is our experience of the body ready for an imaginary action. Since the body readies itself for action in physiologi­cal ways, emotion involves biological processes. Thus when we manage an emotion, we are partly managing a bodily preparation for a consciously or unconsciously anticipated deed. This is why emotion work is work, and why estrange­ment from emotion is estrangement from something of im­portance and weight.

From the interactional theorists, then, we learn what gets done to emotion and feeling and also how feelings are a pre­amble to what gets done to them. From Darwin, as from other organismic theorists, we gain a sense of what, beneath the acts of emotion management, is there to be managed, with institutional guidance or in spite of it. Yet this is not the whole story. It is not simply true that the malleable aspect of emotion is “social” (the focus of the interactional theorists) and that the unmalleable aspect of emotion is its biological link to action (the focus of the organismic theorists). Rather, the unmalleable aspect of emotion (which is what we try to manage) is also social. This point could be analytically sepa­rated from the rest of the thesis with no harm done, but I add it because I think it introduces still another avenue through which to develop a social theory of emotion. And for this account of the social influence on the unmalleable aspect of emotion I move to Freud’s notion of the signal function of emotion, and from there to the influence of our prior expectations about how signals “signal.”

I have said that one reason emotion is unique among the senses is because it is related to cognition. Broadly inter­preted, cognition is involved in the process by which emo­tions “signal” messages to the individual. Freud wrote about the “signal function” of anxiety; anxiety, according to Freud, signaled the presence of a danger from within or outside the individual. It was a means by which the individual told of an apprehended danger. Similarly, other emotional states— such as joy, sadness, and jealousy—can be seen as the send­ers of signals about our way of apprehending the inner and outer environment. Thus to Darwin’s idea of emotion as an action manque, we may add Freud’s idea of the signal func­tion; they are two elaborations on how emotion, as a sense, differs from our other senses.

But signaling is complex —it is not the simple conveying of information about the outside world. It is not a telling. It is a comparing. When an emotion signals a message of danger or safety to us, it involves a reality newly grasped on the tem­plate of prior expectations. A signal involves a juxtaposition of what we see with what we expect to see—the two sides of surprise. The message “danger” takes on its meaning of “danger” only in relation to what we expect. (Sartre develops this point further, 1948.)

In this regard, expectation enters into the signal function of feeling even as it enters into the signaling of other senses—sight, for example. What we see is known to be me­diated through our notions of what we expect to see. As the classic experiments of Solomon Asch have shown, a person who expects to see a long rod on a screen because others around him say they see a long rod reports that he “sees” a long rod even when the rod is short and what the person “sees” is short (1952).

Prior expectations are part and parcel of what we see, and in the same way they are part of what we feel. The idea of prior expectation implies the existence of a prior self that does the expecting. For example, when we feel afraid, the fear signals danger. The realization of danger impinges on our sense of a self that is there to be endangered, a self we expect to persist in a relatively continuous way. Without this prior expectation of a continuous self, information about danger would be signaled in fundamentally different ways. Most of us maintain a prior expectation of a continuous self, but the character of the self we expect to maintain is subject to profoundly social influence. Insofar as our self and all we expect is social—as by the time of adulthood it inevitably is—the way emotion signals messages to us is also influenced by social factors.

Mechanisms of defense are ways of altering the relation of expectation to grasped fact as well as ways of altering each in itself in order to avoid pain. For example, if a woman sud­denly learns that her life partner has been killed, she may alter the character of her understanding of this event so as to keep it in line with what she expects—that he will still be living. She may defend against the self-relevance of the event: “This isn’t happening to me’.’ Or she may defend against the event itself: “He’s still alive. I know he is. I don’t believe he’s dead.” In these ways she holds prior expectation and current perception in a relation to one another that avoids pain.

When we finally go on to make inferences from our feel­ings to “how I must be interpreting this event” or “what must be happening,” we seem to presume that our emotion signals not simply our apprehension of the world but our prior ex­pectations about it. It signals the relation between the two. As practical actors in the world, if not as theorists, we seem to read feeling as a tell-tale sign of “what we must have ex­pected or wanted” as well as a sign of “what was going on.”

To sum up, I am joining three theoretical currents. Draw­ing from Dewey, Gerth and Mills, and Goffman within the interactional tradition, I explore what gets “done to” emo­tion and how feelings are permeable to what gets done to them. From Darwin, in the organismic tradition, I posit a sense of what is there, impermeable, to be “done to,” namely, a biologically given sense related to an orientation to action. Finally, through Freud, I circle back from the organismic to the interactional tradition, tracing through an analysis of the signal function of feeling how social factors influence what we expect and thus what feelings “signal.”


The organismic view reduces us to an elicitation-expression model. The interactional model presupposes biology but adds more points to social entry: social factors enter not simply before and after but interactively during the experi­ence of emotion. Let us say that a man becomes violently angry when insulted. What, in his cultural milieu, consti­tutes an insult? As his anger rises, does he recodify the real­ity to which he responds? Does some feature of the social context aid or inhibit him in this? Simultaneous to his out­

burst, does he react with shame or with pride at the anger? Does he express the anger in ways that work it up or ways that bind it? These are the questions of the interactionist. If we conceptualize emotion as instinct, we never pose ques­tions about these points of social entry in the first place. By virtue of its greater complexity, the interactional model poses a choice between models of how socialfactors work. *

Dewey, Gerth and Mills. Impulse, Dewey argued in 1922, is organized in interaction on the spot. “There are an in­definite number of original or instinctive activities which are organized into interests and dispositions according to the situation to which they respond” (Dewey 1922, p. 147). Thus, fear or anger have no common origin in a constitutional dis­position. Rather, each feeling takes its shape, and in a sense becomes itself only in social context. Dewey talks of how the self, in the process of charting a course of action, actively recharts and alters that course while interacting with the sit­uation. He does not apply these ideas of emergence and vari­ability to emotion, but he prepared the way for Gerth and Mills to do so.

In the same way, George Herbert Mead did not talk about emotion, but he further cleared a path for doing so from an interactional perspective. In Mead’s schema, the self is di­vided into the spontaneous uncontrolled “I” and the reflec­tive, directing, monitoring “me.” Had Mead developed a the­ory of emotion, he would have begun by elaborating his idea of the “I.” To Mead, one person’s “I” was as “spontaneous” as another’s. He looked for no social differences in this aspect of self. But his own notion of the importance of interaction in formulating the “me” that interacts could also be applied to

* The task of integrating social patterns with “basic emotionality” was early recognized by Marvin Opler: “If, for example, there is no latency period, as is well known, in the Trobriands; if Zuni women feel little social sense of deprivation, Okinawans no great sexual shame or guilt, or Samoans little spontaneity and per­sonal freedom in contrast to Navajos; then not only do the mechanisms of adjust­ment vary, but the basic emotionality involved in a type of adjustment will vary as well” (1956, p. 28; emphasis mine).

the “I”; there may well be differences between the “I” in com­parable interactions of, say, an Englishman and an Italian.

Gerth and Mills combine a theory of interaction from Mead, a notion of motivation from Freud, and structural ideas from Weber and Marx in their effort to discover ways in which social structure shapes character (1964, p. xiii). In essence, they do this by linking creeds and symbols to the motivations required for the enactment of institutional roles. Their ideas about emotion are their own; as they say, “George Mead had no adequate notion of emotions and mo­tives, no dynamic theory of the affective life of man” (p. xvii). They distinguish three aspects of emotion: gesture (or be­havioral sign), conscious experience, and physiological pro­cess. Of these three aspects they focus most on gesture—not as Darwin did, outside an interactional context, but as we see below, within an interactional context. Here, in their words, is how interaction enters into the process of defining feeling:

When our feelings are vague and inchoate, the reactions of oth­ers to our gestures may help define what we really come to feel. For example, if a girl has been jilted at the altar and is generally upset about it, the responses of her mother may define the girl’s feelings of sadness and great grief, or of indignation and anger. In such cases, our gestures do not necessarily “express” our prior feelings. They make available to others a sign. But what it is a sign of may be influenced by their reactions to it. We, in turn, may internalize their imputation and thus define our in­choate feeling. The social interaction of gestures may thus not only express our feelings but define them as well. (p. 55)

The girl cries. The mother defines the crying as a sign of anger. The girl responds to her mother’s interpretation of her tears. “Yes, anger more than sadness.” And what the cry­ing “is a sign of” is in this way swayed in interaction with the mother. How do other people influence our understanding of what we feel and, more deeply, even change the “object” of our understanding? How does this influence work differ­ently in different cultural contexts? Gerth and Mills pose these questions, but they pursue them no further.

Erving Goffman. Gerth and Mills address the link between institutions and personalities. Yet the evanescent situations that make up what we call institutions, the situations in which we show our personality, are far more clearly por­trayed in the works of Erving Goffman.

The work of Erving Goffman adds two useful ideas—or more precisely, vantage points, to Gerth and Mills: that of the affective deviant, the person with the wrong feeling for the situation and for whom the right feeling would be a con­scious burden; and that of the fly on the wall, for whom each second of human action is a long, long tale.

The vantage point of the affective deviant allows Goff­man to demonstrate how the social solidarity we take for granted must be continually recreated in daily life. He seems to say, in portrait after portrait, it takes this much work for a group to laugh in simultaneous spontaneity, that much work to achieve engrossment in a game. The nature of the work varies marvelously, but the fact of it remains quietly constant. Beneath this constant is an implicit comparison with what it might be like for the actor to express what he or she feels regardless of social constraint or to what it might be like if conformity came naturally. For unlike Erich Fromm, Goffman does not assume that the individual is ef­fortlessly, pliantly social. On the other hand, the individual’s social feelings are not repressed and made unconscious, as they are for Freud, but consciously suppressed or con­trolled. The social uses of emotion are clearly stated, but it is not so clear how the individual, apart from the group, can use them.

As fly-on-wall, Goffman focuses on the scene, the situa­tion. Each situation, in his view, has a social logic of its own that people unconsciously sustain. Each situation “taxes” the individual, who in return gets protection from unpredict­ability and membership in something larger. The affective deviant is one who tries to avoid paying these social taxes. Taxes, in turn, come in emotive currency. For example, em­barrassment is an individual’s contribution to the group in the singular sense that embarrassment indicates that the in­dividual cares how he seems in company. Not to feel embar­rassed in certain situations is to violate the latent rule that one should care about how the group handles or mishandles one’s identity.

The problem with this rendition of reality is that there is no structural bridge between all the situations. There are “taxes” here and “taxes” there but no notion of an overarching pat­tern that would connect the “collections.” Social structure, to Erving Goffman, is only our idea of what many situations of a certain sort add up to. One moves, as Harvey Farberman puts it, “from one fractured island of reality” to the next, and all the work of making a situation seem real must begin afresh each time. To solve this problem, we should take what Goff­man has developed and link it to institutions on the one hand and to personality on the other. This would enable us to ac­count for what we predicate from one situation to the next, in both institutions and individuals.

Goffman sharpens his focus by identifying the rules and microacts that are conceptual elements of any situation. Rules establish a sense of obligation and license as they apply to the microacts of seeing, thinking about, remembering, recognizing, feeling, or displaying. Consider, for example, the relation of obligation to act: “He will be obliged to prevent himself from becoming so swollen with feelings and a readi­ness to act that he threatens the bounds regarding affect that have been established for him in the interaction” (1967, pp. 122-123). Or, a gamesman “has a right to deeply involve him­self’ (1974, p. 225).

A rule can be distinguished by the micro-act it addresses. Some rules apply to paying attention (1967, p. 115) and thus govern feeling indirectly by governing what might evoke feeling. Other rules apply to feeling directly. For example: “Participants will hold in check certain psychological states and attitudes, for, after all, the very general rule that one should enter into the prevailing mood in the encounter car­ries the understanding that contradictory feelings will be held in abeyance” (1961, p. 23). For the most part, however, rules apply only to what the individual thinks and displays, and the link to emotion is left unspecified.

These rules are, in the main, not consciously recognized, “the questioned actor saying he performs for no reason or because he feels like doing so” (1967, p. 49). They are known indirectly, by the reaction that occurs when a rule is broken. They are also assumed to be generally agreed upon and un­changing. (Goffman does pose conflict, but it is less between one set of rules and another than between individual inter­ests and those of the group.)

Just as Freud specialized in analyzing anxiety, so Goffman specializes in studying embarrassment and shame. Goffman shows us the self coming alive only in a social situation where display to other people is an issue. We are invited to ignore all moments in which the individual introspects or dwells on outer reality without a sense of watchers. Thus guilt, the sign of a broken internalized rule, is seldom if ever discussed. To discuss it would be to put the rule “inside” the actor, inside a sort of self that Goffman does not deal with.

In discussing rules, micro-acts, and shame-prone actors, Goffman applies the overarching metaphor of acting. His rules are generally rules that apply when we are “on stage.” We play characters and interact with other played characters. But for Goffman, acting is surface acting (see Chapter Three). The actor’s mental focus is on the slope of a shoulder, the angle of a glance, or the tightness of a smile, not on any inner feeling to which such gestures might correspond. Deep acting is not as empirically alive in Goffman’s work, and the theoreti­cal statement about it is correspondingly weaker.

To develop the idea of deep acting, we need a prior notion of a self with a developed inner life. This, in Goffman’s ac­tors, is generally missing. From no other author do we get such an appreciation of the imperialism of rules and such a hazy glimpse of an internally developed self. Goffman him­self describes his work as a study of “moments and their men, not men and their moments” (1967). This theoretical choice has its virtues, but also its limitations.

At this point, a brief discussion of those limitations will be a convenient way to introduce my own approach to the study of emotion. Goffman’s theory of rules and his theory of self do not correspond. He posits a relation between rule and feeling. Yet the actor he proposes has little inner voice, no active capacity for emotion management that might enable him or her to respond to such rules. Even as rules and micro – acts come alive in Goffman’s work, the self that might per­form such acts, the self that might acknowledge, obey, or struggle against such rules, is correspondingly unreal. Where is the self as subject of emotive experience? What is the relation of act to self? Goffman speaks as if his actors can induce, or prevent, or suppress feeling—as if they had a ca­pacity to shape emotion. But what is the relation between this capacity to act and the self? Whatever other problems they posed, William James and Sigmund Freud proposed a self that could feel and manage feeling. Goffman does not.

Goffman defines the self as a repository of inner “psycho­logical contributions.” As he puts it: “The self… as a per­formed character, is not an organic thing that has specific location. .. [the performer and] his body merely provide the peg on which something of collaborative manufacture will be hung for a time… and the means for producing and maintaining self do not reside inside the peg” (1959, pp. 252­253). Actions happen to the self; but the self does not do them. Hence Goffman’s language is riddled with passives. For “a person becomes engrossed,” he writes that “a visual and cognitive engrossment occurs” (1961, p. 38). In addition, nouns do the work of verbs. For “people get involved,” he writes that “focused gatherings do have. . . significant prop­erties … [and] the most crucial of these properties… is the organismic psychological nature of spontaneous involve­ment” (1961, p. 38). Conversely, “frames” are said to act; they organize cognitive and visual attention, as if autonomously. In order to divest himself of the concept of self, Goffman must reify the concepts just adjacent to the self. Thus frames, or on rare occasions even feeling states, are given the thickness and weight and reality that is denied to the self. Both Herbert Blumer’s critique of the implied passivity of the actor in sociological writing (1969) and Roy Schafer’s cri­tique of psychoanalytic writing (1976) must be addressed to Goffman. When the self is theoretically dissolved into “psy­chological materials,” no relation between social rules and private experience can be developed.

In Goffman’s theory, the capacity to act on feeling derives only from the occasion, not from the individual. The self may actively choose to display feelings in order to give out­ward impressions to others. But it is passive to the point of invisibility when it comes to the private act of managing emotion. The “I” is there, of course, in the many stories from the San Francisco Chronicle, in the passages from novels, in hangmen’s accounts, in Ionesco plays, in Lillian Gish’s auto­biography. But the private “I” is simply not there in theory. Feelings are contributions to interactions via the passive me­dium of a bodily self. We act behaviorally, not affectively. The system affects our behavior, not our feelings.


Charles Darwin. Darwin’s The Expression of the Emotions in Man and Animals (1872) has offered a model of emotion for vari­ous other theorists and researchers. Darwin focuses on emo­tive expressions—that is, on visible gestures—and not on the subjective meanings associated with them. These gestures, he posits, were acquired during a prehistoric period and have survived as “serviceable associated habits.” Originally linked to actions, these emotive gestures become actions manque. The emotion of love, for example, is the vestige of what was once a direct act of copulation. The baring of teeth in rage is a vestige of the once immediate act of biting. The expression of disgust is the vestige of what was once the im­mediate act of regurgitating a noxious thing. For Darwin, there is no emotion without gesture although there may be gesture without action.

Darwin’s theory of emotion, then, is a theory of gesture. The question for later students thus became: are emotive gestures universal or are they culturally specific? Darwin’s own general conclusion was that they were universal.* The

* Darwin distinguished between facial expressions of emotion that are innate and universal and facial gestures (not necessarily of emotion) that are learned and thus culturally variable. He devised a sixteen-item questionnaire and sent it to thirty-six missionaries and others who had lived in non-Western societies. One question was: “Can a dogged or obstinate expression be recognized, which ischiefly shown by the mouth being firmly closed, a lowering brow, and a slight frown?" Based on his returned questionnaires, Darwin concluded that “the chief expressive actions” of human beings were innate and therefore universal. Despite his gener­ally universalist interpretations, however, Darwin concluded that some nonverbal behaviors (such as weeping, kissing, nodding, and shaking the head in affirmation and negation) were not universal but culture-specific and “learned like the words of a language” (quoted by Dane Archer in Rosenthal et al. (1979, p. 352).

debate has been carried forward by those who argue that emotional expressions are probably innate (Ekman 1971, 1983; Ekman et al. 1972) and those who argue that they are modeled on language and therefore culturally variable (Klineberg 1938; Birdwhistell 1970; La Barre 1964; Hall 1973; Rosenthal 1979, p. 201). What is missing from both sides of this debate is what was missing in Darwin’s theory from the beginning: a conception of emotion as subjective experience and a more subtle and complex notion of how social factors impinge.

Taking another tack, but subject to the same critique, Randall Collins unites a Darwinian concept of emotion with a Durkheimian notion of ritual as a means of arousing emo­tion (1975, p. 95).[33] He then argues (drawing on his conflict model) that men compete with each other for control of the ritual apparatus, which is a powerful tool for commanding people by controlling their emotions (pp. 59,102). Yet in this interesting development of Darwin, the same push-button model of emotion remains unquestioned.

Sigmund Freud. Freud’s thinking on emotion, or affect, went through three major developments. In his early writ­ings he thought affect to be dammed-up libido indicating itself as tension and anxiety; affect was the manifestation of instinct. f At the turn of the century, he came to think of af­fect as a concomitant of drive. Then in 1923, in The Ego and the Id, he came to stress the role of the ego as a mediator between the id (drive) and conscious expression. Affects were now seen as signals of impending danger (from inside or outside) and as an impetus to action. The ego was as­signed the capacity to postpone id drives, to neutralize or bind them (see Brenner 1974, p. 537).

Unlike Darwin, Freud singled out one emotion—anxi­ety—as the model for all others, reasoning that it was more important because the unpleasantness of anxiety led to the development of various ego defenses against that unpleas­antness. As Brenner notes, “As analysts we recognize that anxiety occupies a special position in mental life. It is the motive for defense. Defenses serve the purpose of minimiz­ing, or, if possible, preventing the development of anxiety” (1974, p. 542). Anxiety was initially defined in a way that by­passes the ego: anxiety was “the reaction to an influx of stim­uli which is too great for the mental apparatus to master or discharge” (p. 533). Rejecting this model, Brenner suggests:

Anxiety is an emotion… which the anticipation of danger evokes in the ego. It is not present as such from birth or very early infancy. In such very early periods the infant is aware only of pleasure or displeasure…. As experience increases, and other ego functions develop (e. g., memory and sensory percep­tion), the child becomes able to predict or anticipate that a state of displeasure (a “traumatic situation”) will develop. This dawn­ing ability of the child to react to danger in advance is the begin­ning of the specific emotion of anxiety, which in the course of further development we may suppose to become increasingly sharply differentiated from other unpleasant emotions. (Bren­ner 1953, p. 22)

Freud’s focus on anxiety was part of his concern with mas­sive, incapacitating, “pathological” emotions that exagger­ate the normal case. Furthermore, important as it is to un­derstand it, anxiety is not typical of all other emotions in several ways. We do not try to avoid joy or love in the way that we typically try to avoid anxiety. Anxiety is also atypical in that it is an emotion without a defined object; one is not anx­ious at someone in the same way that one is furious at or in love with someone.

For Freud, unlike Darwin, the meaning of a feeling (the ideational representations associated with affect) is crucial but often unconscious. As Freud explained, “To begin with it may happen that an affect or an emotion is perceived but misconstrued. By the repression of its proper presentation it is forced to become connected with another idea, and is now interpreted by consciousness as the expression of this other idea. If we restore the true connection, we call the original affect ‘unconscious’ although the affect was never uncon­scious but its ideational presentation had undergone repres­sion” (Freud, 1915b, p. 110).* Thus the focus in Freud’s early writing on instinctual givens, on anxiety as the main connec­tion the individual has with them, and on the unconscious as a mediator between individual understanding and instinct led him to conceive of social influences mediated through the ego and superego as relatively unimportant. Like Darwin, he had little to say about how cultural rules might (through the superego) apply to the ego’s operations (emo­tion work) on id (feeling).

William James. If for Darwin emotion is instinctual gesture and if for the early Freud emotion (affect) is the manifesta­tion of dammed-up libido, for James emotion is the brain’s conscious reaction to instinctual visceral changes. As James noted in his Principles of Psychology (1890): “My theory… is that bodily changes follow directly the perception of the ex­citing fact and that our feeling of the same changes as they occur is the emotion” (cited in Hillman 1964, p. 50).

This theory has been at the heart of much controversy between the centralists (such as Cannon and Schachter) and the peripheralists (such as James and Lange), t James

* There is a lively debate on the question of whether, apart from ideas, feelings can be unconscious (see Pulver 1971). Motive and wish, as aspects of affect, are certainly assumed to be potentially unconscious. Fenichel (1954) and Greenson (1953), for example, hypothesize that boredom involves an unconscious attempt to convince oneself that one does not want to gratify an instinctual wish that is fright­ening, and therefore one has no wish to do anything.

t As Hillman points out, there was a difference between James and Lange. For James, emotion is conscious feeling and bodily change together at the same time. For Lange emotion is bodily change, the feeling of which is secondary in conse­quence (Hillman 1964, p. 50). For a careful exegesis of James, see Hillman (1964), pp. 49-60.

equates emotion with bodily change and visceral feeling. From this it follows that different emotions will be accompa­nied by different, not similar, bodily states. Manipulation of bodily states, by drugs or surgery, will also manipulate emo­tional states. Cannon’s 1927 experimental work (1929) re­futed the James-Lange theory. He found that the total sepa­ration of the viscera from the central nervous system (which gives us our sensations) does not alter emotional behavior. The dog operated on could still, it was presumed, feel emo­tion. Further, the viscera are relatively insensitive and change slowly, unlike emotions (see Schachter and Singer 1962, 1974; Kemper 1978; and Chapters Seven and Eight). After Cannon’s work, psychologists sought to discriminate between emotional states according to cognitive factors. Thus, the Cannon research set the stage for future social psychology. Gerth and Mills note: “There do not, for exam­ple, seem to be noteworthy differences in the visceral accom­paniments of fear and anger…. We must go beyond the or­ganism and the physical environment to account for human emotions” (Gerth and Mills 1964, pp. 52-53). While “going beyond” does not mean ignoring the importance of physiol­ogy in emotion, it does mean working with a more intricate model than organismic theorists propose of how social and cognitive influences join physiological ones.


Two basic models of emotion have emerged in the last cen­tury. From the work of Charles Darwin, William James, and the early Sigmund Freud, an organismic model appears.*

* McDougall (1937, 1948) and Tomkins (1962) have also contributed to the or­ganismic model of emotion. Although Tomkins’s theory covers a broad range of

From the works of John Dewey, Hans Gerth and C. Wright Mills, and Erving Goffman, versions of an interactional model appear. The two models differ in several fundamen­tal respects.

First, the organismic model defines emotion as mainly a biological process. For the early Freud, emotion (affect) is libidinal discharge, for Darwin it is instinct, and for James it is the perception of a psychological process. By virtue of the stress on instinct and energy, the organismic theorists postu­late a basic fixity of emotion and a basic similarity of emotion across categories of people. For the interactionists, on the other hand, it is enough to say that emotion always involves some biological component. Whether the biological pro­cesses involved in fear, for example, actually differ from those involved in anger (James thought they do; Cannon proved they do not) is a matter of little theoretical interest to the interactionist, whose main concern is the meaning that psychological processes take on.

Second, in the organismic model, the manner in which we label, assess, manage, or express an emotion is seen as extrin­sic to emotion and is therefore of less interest than how the emotion is “motored by instinct.”

Third, in the organismic model, emotion is assumed to have a prior existence apart from introspection, and intro­spection is thought to be passive, lacking in evocative power. As one psychoanalytic theorist reasoned:

Introspection provides abundant examples, one of which the reader, if he is so inclined, may notice in himself at this very moment. We know that a “feeling tone,” an affective quality, is always present as a part of our stream of experience, conscious or unconscious. Yet if this paper has captured your interest, it is probable that you have not been aware of your feelings during the past few minutes in which you have been reading it. If you

phenomena, it focuses on the relationship between drive and emotion. He distin­guishes eight innate affects, which are said to be evoked by “innate activators” that serve as “drive signals.”

now set it aside for a moment and introspect, you will notice your own immediate feeling. You may be comfortable, slightly irritated, mildly depressed, etc., but some feeling will be there. The affect, until you noticed it, had been present but not in awareness: it was preconscious. (Pulver 1971, p. 351; my emphasis)

For the interactionist, it is highly questionable that the feeling had been present all along. How do we know, they ask, that the very focusing of attention and use of cognitive power does not in itself evoke the feeling? And if the act of attending to feeling helps shape the feeling itself, that feel­ing cannot be referred to independently of these acts. Simi­larly, for the interactionist, the act of management is insepa­rable from the experience that is managed; it is in part the creation of that emerging experience. Just as knowing affects what is known, so managing affects what is “there” to be managed. This reflexivity of expression is generally doubted by organismic theorists (see Lofgren 1968). In the organic, “discharge” theory of affects, the manifestation of an emo­tion is almost epiphenomenal because emotion is presumed to be linked to impervious organic givens.[32] In sum, for the interactional theorists, emotion is open-ended whereas for the organismic theorists it is fixed.

Fourth, the organismic stress on instinctual fixity reflects an interest in the origins of emotion, a subject of little con­cern to interactionalists. Darwin, for example, traces emo­tion back to its phylogenetic origin and points to evidence of similarities between emotions in animals and in human be­ings. Freud traces emotion experienced in the present back to ideas whose origin often lies in childhood (Brenner 1974, p. 542). The interactional model, on the other hand, points attention away from origins and focuses instead on aspects of emotion that uniquely differentiate social groups of nor­mal adult humans.

Each difference between the two models implies different links between social factors and emotion. In the organismic model, social factors merely “trigger” biological reactions and help steer the expression of these reactions into custom­ary channels. In the interactional model, social factors enter into the very formulation of emotions, through codification, management, and expression.

MODELS OF EMOTION From Darwin to Goffman

Most of the arguments about specific aspects of emotion can be traced to a more fundamental difference between what may be called the organismic and the interactional view­points. Before I summarize these viewpoints and state my own position, it will be useful to acknowledge two barriers to any serious inquiry on this matter: first, the practice among social scientists of ignoring emotion or subsuming it under other categories; and second, the acceptance of several ideas about emotion that confuse any discussion of it.

Some theorists have gone so far as to deny that emotion is a tenable concept. Thus the psychologist Elizabeth Duffy, after distinguishing between longitudinal concepts (which describe phenomena that occur sequentially) and cross-sectional con­cepts (which describe phenomena such as perception, thought, and emotion, which occur simultaneously) argued for dispensing with cross-sectional concepts altogether. She was correct to point out that they represent loose and over­lapping categories of phenomena (1941, p. 184). Unfortu­nately, her alternative simply eliminates the complexity we ought to be trying to describe. The same objection applies to social psychologists who believe that the exquisite care they take to avoid discussing feeling, in order to focus ever more intently and narrowly on cognition, increases the scientific character of their work. A content analysis of their own per­sonal speech habits over an average week would certainly

show that emotion is more central to life as they live it than to life as they study it.

Many social psychologists give emotion short shrift by subsuming it under some conceptual umbrella. For example, in an otherwise informative study of soldiers’ attitudes to­ward the Women’s Army Corps in 1950, Suchman and col­leagues subsume emotion under the concept of affect: “Af­fect toward an object can be very generally classified as either positive or negative. For our purposes, however, an­noyance, anger, distrust, and fright are all shadings of nega­tive affect, and these shadings we shall ignore” (cited in New­comb et al. 1965, p. 48). When emotion is subsumed in this way, the interesting dimension of emotion becomes the “how much.” What precisely there is “a lot” of or “a little” of is unclear. We lose the distinction between a fearful dislike of the Women’s Army Corps and an angry dislike of it. We lose a wealth of clues about the various definitions of reality that people apply when adopting an attitude. We lose the idea that emotions reflect the individual’s sense of the self-rele­vance of a perceived situation. We lose an appreciation of what the language of emotion can tell us.*

For those who do not deny or subsume emotion, two other ideas sometimes obstruct our clear understanding of emotion. These are: (1) The idea that an emotion, like anger or jeal­ousy, can have an independent presence or identity within a person through time. (2) The idea that when possessed by emotion we are led to act irrationally and see distortedly. Be­cause these notions are sometimes applied by writers in both the organismic and the interactional camps, we should exam­* There is loss when emotion is conceptually ripped away from the situation to which it is attached. When Aristotle discusses his fifteen emotions, Descartes his six, Hobbes his seven, Spinoza his three (with forty-eight derivatives), McDougall his seven, and Tomkins his eight, the immediate relation of emotion to viewpoint or frame is lost. This is also a problem with Joel Davitz’s otherwise interesting attempt to formulate a dictionary of emotions (1969). Just as modern linguists now examine language as it is used in social context, so emotion, another sort of language, is best understood in relation to its social context.

ine their content before turning to the assumptions that di­vide the organismic and interactional theorists.

Does emotion have a presence or identity independent of the person it is “in”? We talk as if it did. We commonly speak of “expressing,” “storing,” “getting in touch with,” or even “spreading” an emotion. We speak of guilt as something that “haunts” us, or fear as something that “grips,” “strikes,” “be­trays,” “paralyzes,” or “overwhelms” us. Fear, as we talk about it, is something that can lurk, hide, creep, look up, or attack. Love is something we fall into or out of. Anger is something that overtakes or overwhelms us. In this way of talking we use the fiction of some independent, outside agency in order to describe a contrasting inner state.

As Roy Schafer points out in A New Language for Psycho­analysis (1976), the very way we normally talk about an emo­tion, our very use of nouns such as “anxiety” or “love” or “anger” suggests entity. Even verbals such as “fearing” or “dreading” (we can’t speak of “anxiousing”) are themselves abstractions and carry the same implications as the nouns they replace. Schafer proposes a new action language as a substitute for common parlance. He would remove expres­sions like “to fear” or “fearing” because they refer abstractly to a number of separate actions and modes of action; thus “to fear may subsume to flee, to avoid, to act timidly, or pla­catingly” (p. 275). Though Schafer is perceptive in identify­ing common expressions that embody problematic assump­tions, his action language seems to me too simple an apparatus for coping with the complexity of everyday emo­tional life.

Commonly we find ourselves speaking of emotion as if it had a location or residence. When we speak of love as residing in the heart and envy in the bile, the heart and the bile are put in place of the person. The speaker personifies an organ or portrays emotion as “a substance or quantity of energy of a certain kind.” We also speak of emotions as having some sort

of continuous identity, as when we say an emotion is “stored” or “accumulated,” or when we refer to an “old” emotion.

Metaphors that suggest agency, residence, and continuity through time often convey with uncanny precision just what it feels like to experience an emotion; they enjoy a poetic ac­curacy. But they can get in the way of understanding how emotion works.

A second idea that impedes our understanding of emotion is that the inner state of emotion is always associated with outer action that is irrational. This is sometimes the case and sometimes not. A man who feels fear at the sight of a rattle­snake moving toward him may run to safety. He may act ra­tionally. Were he not afraid, he might not run, which in the absence of other forms of protection, would be irrational. Again, a mother may, with the feeling of love, reach out to hug her child. Here, too, the feeling and act seem consonant and “rational” in the sense that what a person does, under the in­fluence of feeling, gets people where they want to go as much if not more than what a person would do if not under the influence of feeling. They only reason I pose these obvious examples is that when people talk about “acting emotionally,” it is often not these examples they cite. That is, we tend to associate the idea of emotion more with irrational or unwise actions than with rational or wise actions. This tendency results more from our cultural policy toward emotional life (“watch out for it, manage it”) than it results from observing the relation between feeling and action in all the common but inconspicuous instances in which they are related.