The Blackwell Companion to Sociology (97 page)

exchange principles, is Cecilia Ridgeway's theory of status value (Ridgeway,

1991). Ridgeway asked the question: how do nominal characteristics (categories

like male/female, young/old, black/white) acquire consensual status value in

society? Using expectation states logic, she argued that if the nominal characteristic is correlated initially with some material resource, interactions between

people who differ on both resource levels and the nominal characteristic will

lead to inferences about performance expectation that will eventually diffuse

through the society, creating status value for the nominal categories. New

simulation work by Mark et al. (1999) indicates that in small societies, a

correlation between resources and the nominal characteristic is not needed;

uncorrelated differentiation on both resources and the nominal characteristic is enough to create status value in some instances.

The research traditions above ± social exchange and expectation states ± both

look at how people attempt to gain valued rewards, the first through exchange

with others and the second in the context of a group performing a collective task.

We now turn to research traditions that are based on a different conception of

the actor.

Meaning and Emotion in Social Interaction

The symbolic interactionist tradition in sociology has at its root three central principles: (a) people act toward things, including each other, on the basis of the meanings that those things hold for them; (b) the primary source of these

meanings is social interaction; and (c) meanings are managed and transformed

through an interpretative process (Snow, 2000). The fundamental conception of

the actor is that of a meaning-creator, who actively works to interpret what is

happening around him or her in terms of meanings accrued from past interac-

tions and who actively generates new lines of action to maintain a coherent,

meaningful view of self and others. Historically a heavily cognitive perspective, symbolic interaction has turned its eye toward affective meaning and emotion in

the past 20 years (MacKinnon, 1994). There are two dominant perspectives in

this new work. The first views emotion as a signal about how well events are

maintaining identities. The second takes a more cultural approach, concentrat-

ing on the norms associated with emotion.

Control Theories of Identity

Sociologists use the term `ìdentity'' to refer to the many meanings attached to a person, both by the self and by others. The concept embraces both (a) structural features like group affiliations, role occupancy, and category memberships, and

(b) the character traits that the individual displays or that others attribute to him 414

Lynn Smith-Lovin

or her. There has historically been some tension between symbolic interactionists who focus on the creative, actively negotiated process through which people

make their identities within social interaction and the symbolic interactionists who emphasize the extent to which our identities are shaped by the social

structures in which we live. (See Stryker (1980) for a nice overview of the history of interactionist thought and Snow (2000) for a short summary of differences

and communalities between the two schools.) In recent years, theoretical think-

ing in the latter camp (often called structural symbolic interactionism) has been dominated by a control system model that makes it much more dynamic,

creative and processual.

David Heise (1979) initially developed the control system view, borrowing

control models from engineering and measurement technology from psychology.

He argued that people acquire general meanings for role-identities (like mother

and daughter) and social actions (like compliment or criticize) through interac-

tion. The meanings are affective in character and universal in their form. People respond to all concepts in terms of three basic dimensions: goodness, powerfulness, and liveliness. After people acquire these meanings, they are quite stable.

Therefore, when people recognize an interaction as an instance of a certain type of situation (a mother and a daughter), they perceive events and engage in

actions that maintain the meanings that those role-identities evoke. So a

woman who sees herself as a mother interacting with a daughter will do things

that maintain the good, powerful, and lively meanings associated with that

identity, and will expect her daughter to do things to maintain that somewhat

less potent but more lively identity. Thus the role-identity meanings and action meanings act as reference signals that are maintained by social interaction. When events do not maintain identity meanings, emotions signal the extent and direction of the deflection away from the reference value (Smith-Lovin and Heise,

1989; Smith-Lovin, 1990).

There are several important ways in which this affect control theory repres-

ented a major advance over earlier formulations. The control model focuses

explicitly on the relational context of interaction. The unit of analysis here is the event (an actor±behavior±object combination), not the individual. The use of

universal dimensions of meaning (goodness, powerfulness, and liveliness) allows

all types of social entities (actors, behaviors, emotions, traits, settings) to be characterized in the same system. Using an impression formation paradigm from

psychology, the theory specifies exactly how the many elements of a situation

combine to form a situation-specific meaning that is then compared to the

reference value to ascertain how well events are maintaining meanings. Most

importantly, the control view of the relationship between role-identity and social action accurately represented the extraordinary flexibility of actors' application of cultural information as they moved through varying situations. Role occupants (for example, mothers) did not follow a simple, static script of role

expectations vis-aÁ-vis a particular alter (a daughter). What they did was powerfully shaped by events that occurred and their interpretation of them. A mother

who has just been appreciated by her daughter is expected to behave very

differently than a mother who has been ignored. A final insight from affect

Social Psychology

415

control theory is the sense in which cognition and affect are intertwined

inexorably. One cannot process a cognitive understanding of an interpersonal

situation without responding to it emotionally. Nor is emotion likely to occur (in any but the simplest stimulus±response modes) without the active work of

defining the situation.

Burke (1991) later developed an alternative identity control model. He used a

more intricate, role-specific method for measuring meaning (so that meanings

cannot be compared across role domains). Burke's model focused only on the

maintenance of the focal actor's identity meanings; he assumed that people tried to maintain their own identities, but are concerned with the identities of others only insofar as they impact self-meanings. In contrast, affect control theory

attempts to maintain meanings associated with all elements of a situation.

The most marked difference between the two control models is in their view of

emotion, however. In Heise's theory, people may feel exceptionally good, potent

or lively if their immediate experiences deflect them upward on one or more of

the affective dimensions. An employee who receives an unexpectedly good

performance evaluation might feel ``high as a kite.'' While people are predicted to act in ways that bring situated meanings back into line with reference signals, emotions can signal deflections above or below the reference standard. In

Burke's identity control formulation, failure to confirm meanings is always

assumed to produce a negative emotion, stress (Burke, 1991).

Researchers have tested the control theories successfully using a wide array of

experimental, survey and qualitative methods (see reviews in chapters 2 and 5 of Cook et al., 1995). Experiments showed that affect control theory does a good

job of predicting the emotions of both actors and the recipients of actions.

Experiments also have confirmed the counterintuitive prediction that people

will choose to interact with others who confirm a negative self-identity (as a

poor public speaker) even when the other option is someone who evaluates them

more positively than they evaluate themselves. Qualitative data from two church

congregations showed how gay Christians created new identities and rituals to

generate positive emotions in a religious context (while more traditional identity meanings for homosexual and religious identities generated negative emotion).

Burke and his colleagues have tested their control theory extensively in a data set of newly married couples in Washington state who were followed for three years.

Gender identities, personal identities about control, and parental/spousal iden-

tities all appear to be maintained by the respondents in that survey.

New research focuses on three areas. First, both control theories have devel-

oped such substantial research traditions at this point that scholars are beginning to consider how to judge their relative strengths. The key points of disagreement

± the measurement of meaning, the location of the reference standard, and the

valence of emotion ± will focus research in the future. Second, both research

traditions have begun to explore how the multiple identities that constitute the self are processed during interaction. These studies deal with the salience of

identities within a relatively stable self-structure (Stryker, 1980) and the parallel processing of multiple identities that may be evoked within a single situation

(Smith-Lovin, 2001). Third, researchers are applying the control theories in the 416

Lynn Smith-Lovin

context of instrumental interaction to view the interaction of affective and

instrumental processes. This work is discussed more extensively in the final

section of the chapter.

Feeling Rules, Emotion Work, and Emotional Labor

While the control models use identity dynamics to generate emotions, most work

in the sociology of emotion emphasizes cultural norms. Arlie Hochschild (1979)

drew upon Erving Goffman's insights about the management of self-presentation

to discuss how people manage their emotions to make them appropriate to a

situation. Hochschild developed the concept of emotion work to describe the act

of trying to change the type or intensity of emotion that one was experiencing.

People use cognitive thoughts, physical manipulation, or expressive display to

try to change how they feel. Often people manage emotions in order to conform

to feeling rules ± norms about what type or intensity of emotion we should feel.

Peggy Thoits (1985) has linked emotion work to mental health. She argued that

when emotions routinely fail to match our cultural expectations and manage-

ment efforts fail, the emotional deviance could be labeled as mental illness. Such deviance could be created by inadequate socialization or by structurally induced stress.

When emotion work is done for pay, and feeling rules are job requirements that

represent à`commoditization of feeling,'' Hochschild argued that workers could

lose touch with their real, unmanaged emotional responses. She speculated that

the consequences would be a disturbing sense of inauthenticity and burnout on

the job. Her study of the emotional labor that flight attendants were required to perform to put up a pleasant, friendly face in interaction with sometimes annoy-ing, abusive passengers (Hochschild, 1983) illustrates the process. Emotional

labor in many other occupations ± bill collectors, supermarket checkers, medical students, nurses, and library workers ± has been observed using

similar ethnographic methods. Researchers have also studied emotion norms

in non-work contexts using undergraduate students, advice books, women's

magazine columns, and ethnographic work in other cultures. This work convin-

cingly demonstrates that norms for both feelings and emotion display are an

important component of many roles. These norms can change over time and vary

from culture to culture. Research on the psychological effects of managing one's emotions for pay is less convincing; this is a promising area for future work.

Social Structural Positions and Emotional Experience

Survey researchers often study mental well-being while drawing on both the

symbolic interactionist and cultural perspectives on emotion. This research

tradition looked at how social structural positions (and the role-identities that they imply) affect emotional experience. In general, this research showed that

those who are disadvantaged or lower status report more emotional distress

than those who have greater control and more resources to cope with trouble

(see review in Mirowsky and Ross, 1989). Women report more emotional

Social Psychology

417

distress than men. The uneducated poor report more than the educated rich.

Poor racial minorities experienced more distress than poor whites (although the

effect is non-existent or occasionally reversed among the middle class). People

who experience undesirable life events (for example, loss of a job, death of a

spouse, accidents) report more distress than those who do not. The young and

the very old are more unhappy than the middle-aged. Those who have few

network ties and few social roles have less happiness than those with more