Creating Status Characteristics

by Murray Webster, Jr., Stuart J. Hysom
Citation
Title:
Creating Status Characteristics
Author:
Murray Webster, Jr., Stuart J. Hysom
Year: 
1998
Publication: 
American Sociological Review
Volume: 
63
Issue: 
3
Start Page: 
351
End Page: 
378
Publisher: 
Language: 
English
URL: 
Select license: 
Select License
DOI: 
PMID: 
ISSN: 
Abstract:

CREATING STATUS CHARACTERISTICS*
Murray Webster, Jr. Stuart J. Hysom

University of North Carolina, Charlotte Emory University

Status characteristics constitute significant elements of small group and so- cietal stratification systems, and understanding their creation has theoreti- cal and practical importance. We generalize Ridgeway Spathbreaking theory of status construction in three stages. First, we show that a conceptual change, using goal objects instead of exchangeable resources, permits us to explain the creation of many more status characteristics. Second, we expli- cate an interaction mechanism, behavior interchange patterns, that can trans-

form other characteristics into status characteristics, even creating status differences where no characteristic is salient. Third, we show how new char- acteristics themselves can be created and given status value from deviance and personality attribution processes. We briefly note some parallels with other theorists' work, suggest some independent tests, and consider theoreti- cal and applied implications of this work.

Status characteristics are essential con- power. Park's (1928) conception of interac- stituents of social life. People differ by tion began with individuals' classifying each gender, race, wealth, beauty, age, reading other in terms of age, sex, and race; the result- ability, dialect, and other characteristics, and ing inferences organized conduct. those distinctions may carry great social sig- In the second half of this century, theorists nificance. Not surprisingly, interest in status have analyzed status as an organizing prin- can be traced to the earliest days of sociology. ciple of social systems, using large-scale As Simmel ([I9081 1950) observed, "The sample data, laboratory techniques, compu- first condition of having to deal with tational technology, and various formal and somebody . . . is to know with whom one has mathematical tools. Dahrendorf (1959:282 to deal" (p. 307). Weber ([I9221 1968) ff.) concluded that society is defined through viewed status characteristics as connoting so- its various status categories. Garfinkel cial worth and therefore as part of a society's (1961), Goffman (1970), and others analyzed stratification system, along with wealth and ways in which individuals gain, lose, and de- fend claims to various status categories.

* Direct correspondence to Murray Webster, Collins (1981; Kemper and Collins 1990) Dept. of Sociology, University of North Carolina, analyzed how interaction chains, influenced Charlotte, Charlotte, NC 28223 (m.webster@

by impression management goals and ex- worldnet.att.net). This work developed over two change factors, combine to produce social years; many contributed. We benefitted particu- structures. Collins and collaborators (Collins

larly from discussions with Cecilia Ridgeway and Joseph Berger. Participants at a Group Process 1991; Collins et al. 1993) later elaborated roundtable at the ASA meetings in 1996, at the that approach, relating it to the status char- Iowa Theory Workshop in 1996, and at an Emory acteristic gender and showing relations be- University sociology colloquium in 1997 helped tween social power, economic production, clarify our ideas. Hamit Fisek, David Willer, Jo-

social movements, and other factors in main- seph Whitmeyer, Terry Boswell, and Cathryn taining gender stratification and transmitting Johnson provided thoughtful comments on earlier it through interaction rituals. Turner (1984, drafts. ASR reviewers, especially the deputy edi- chap 5; 1992; 1995, chap. 7) proposed sev- tor, improved theoretical context and clarity.

eral models of the dynamics of social differ- Nathan Haynes produced the figures. An award

from the National Science Foundation (SBR- entiation. Jasso (1994a, 1997) offered an ex- 951 1514) supported the manuscript preparation. planation for the rise and perpetuation of We thank all. gender inequality, tracing it to the pre-

American Sociological Review, 1998, Vol. 63 (June:351-378)

historical agricultural revolution and migra- tions across Europe. Jasso's theoretical per- spective offers a complementary explanation for some phenomena discussed in this paper; we note comparisons later.

Many recent studies focus on how status characteristics affect interaction and group structure. Such investigations include how gender creates double standards for hiring, evaluations, wages, and promotions (Biernat 1995; Eskilson and Wiley 1976; Foschi 1996; Foschi, Lai, and Sigerson 1994; Foschi, Sigerson, and Lembesis 1995; Jasso 1994); how to overcome unwanted organiz- ing effects of gender (Pugh and Wahrman 1983); how educational attainment (Moore 1968), race (Webster and Driskell 1978), and age (Freese and Cohen 1973) affect compe- tence evaluations and influence, even when they are irrelevant to the tasks at hand; how ethnicity affects careers of Israeli soccer players (Yuchtmann-Yaar and Semyonov 1979); and how beauty affects conceptions of intellectual competence (Eagly et al. 1991 ; Jackson, Hunter, and Hodge 1995; Webster and Driskell 1983). A central theme in all these studies is the existence of social advan- tages and disadvantages associated with dif- ferent status characteristics, often including estimations of overall social worth and spe- cific performance capacities.

The study of status characteristics links micro-level with macro-level theory and re- search, an approach generally followed in "structural social psychology" (Lawler, Ridgeway, and Markovsky 1993).' Status structures are features of social systems that shape and constrain face-to-face behavior; face-to-face interaction enacts, reproduces, transmits, and may even create status struc- tures. Just which characteristics actually are status characteristics-carrying certain specified social advantages and disadvan- tages-depends on cultural definitions and other features of a society. All status charac- teristics, however, whatever the society and its culture, confer advantages and disadvan- tages. Understanding how characteristics af- fect interaction, and how individuals' expe- riences and actions transmit, reproduce,

Useful discussions of micro-macro connec- tions are available in Coleman (1986), Mar- kovsky (1988), and Turner (1988).

AMERICAN SOCIOLOGICAL REVIEW

transform, and redefine a society's status characteristics, requires attention to more than one "level" of sociological theory.

A fundamental question is why a given characteristic-say, gender or skin color-is a status characteristic rather than simply a descriptor. Understanding conditions under which characteristics become status charac- teristics-that is, how they may acquire in- vidious and socially disadvantageous conno- tations-has both theoretical and practical importance. Such an understanding contrib- utes to sociological theory by explaining where status characteristics originate. Theory may be used to devise interventions to over- come some undesirable effects of status char- acteristics. At a broad level, if (for instance) gender and skin color were simply descrip- tive terms rather than status characteristics, many interaction problems would disappear.

Some years ago, Ridgeway (1991) pre- sented a theory of the social construction of status value. The theory has been extended and stated formally (Ridgeway 1997b; Ridgeway and Balkwell 1997), and parts of it have been confirmed experimentally (Ridgeway et al. 1998; Ridgeway and Glasgow 1996). Status construction theory solves some important and difficult theoreti- cal puzzles; in particular, it explains how gender can acquire status value. Ridgeway (1997a) demonstrates applied usefulness of the work, showing how status construction processes figure in organizations and else- where to reproduce gender inequality.

We hope to build on this work, taking ad- vantage of several additional lines of theory and research to extend the range of situations for which we can explain how status value develops. The ideas we adapt come from par- allel programs of theoretical and empirical investigations; they were discovered and for- mulated by other sociologists. Our intended contributions are (1) to adapt these ideas to problems of creating status value, extending the scope of the Ridgeway-Balkwell formu- lation; (2) to show how these ideas provide a theoretical mechanism for the basic status acquisition process of status construction theory; (3) to show how these ideas provide mechanisms for two other processes modeled in status construction theory; and (4) to sug- gest how new characteristics may develop in a society through changes in norms govern- ing expected behavior, and through related attribution and enforcement practices.
STATUS ORGANIZING PROCESSES

Several decades of theory building and em- pirical research have developed a well-elabo- rated understanding of how status character- istics affect interpersonal behavior. This sus- tained program has produced several desir- able outcomes: (1) We possess explicit defi- nitions of diffuse and specific status charac- teristics; (2) we know scope and interaction conditions that lead to status generalization, and how the process occurs; (3) outcomes, such as group power and prestige structures and inferences regarding individuals' abili- ties, have been predicted and assessed through a variety of research methods; (4) it is possible to predict precisely the effects of any combination of status characteristics for many situations; (5) several investigators have developed a "family" of theories apply- ing to different kinds of abstract situations, describing additional means by which power and prestige structures may develop; and (6) interventions have been designed to control aspects of status generalization in natural set- tings. We begin by summarizing features of this theoretical research program relevant to status construction.

The theory of status characteristics and ex- pectation states applies to situations where individuals are task-focused and collectively oriented. Task focus means that people's pri- mary purpose in meeting is to solve some problem or set of problems (rather than to address a primarily social-emotional con- cern). Collective orientation means that the problem belongs to the entire group. (In con- trast, in an individual task, each individual works alone, and different solutions may be produced.) Juries, work committees, and ath- letic teams are examples of such groups. For a jury, reaching a good verdict is more im- portant than anything else in the meeting (task-focus), and the verdict must be unani- mous (collective orientation).

The theory defines two types of status char- acteristics: specific, denoted by C, and diffuse, denoted by D. We focus on diffuse char- acteristics; specific characteristics meet only parts 1 and 2 of the following definition: Dif

fuse status characteristics are those having

(1) two or more states, evaluated differen- tially in the culture; (2) performance expec- tations of limited scope associated with each state; and (3) performance expectations with- out explicit limit associated with each state.

Thus gender is a diffuse status characteris- tic because it can be shown that (1) differentiation (i.e., malelfemale) is associated with differential connotations of esteem, social worth, respect, and honor; (2) members of one group (e.g., men) are thought better able, for instance, to work mathematical problems than are members of the other group; and (3) members of that group (men) also are thought to be more rational, more depend- able, stronger, more orderly, better at running machinery, etc. The etc. makes gender a dif- fuse rather than a specific status characteris- tic: Men enjoy generally higher performance expectations than women in our culture (and in most cultures), without explicit limit on areas to which those expectations apply (Wagner 1988a).

The list of diffuse status characteristics in

our society includes gender (of particular in-

terest in status construction theory) as well

as skin color and occupations (to which that

theory should apply). The list also includes

other characteristics that have been shown to

meet the definition, such as beauty; these, as

we discuss below, do not so obviously fit the

theory's explanation.

Figure 1 illustrates the status generaliza- tion process. Starting at the left, the distribu- tion of status characteristics possessed by ac- tors is an observable set of initial conditions. Unless those status characteristics are known to be irrelevant, they create aggregated per

formance expectation states, depicted in the shaded, unobserved, portion of Figure 1. Performance expectations are unobservable theoretical constructs; they translate to ob- servable components of a group's power and prestige structure, shown at the right-hand side of the figure. Thus the theory explains interaction advantages and disadvantages as- sociated with status characteristics as conse- quences of the ability inferences, or perfor- mance expectation states, created by status generali~ation.~

The theory is presented in full in Berger et al. (1977). Webster and Foschi (1988) and Wagner and Berger (1993) describe extensions and tests.

AMERICAN SOCIOLOGICAL REVIEW

Distribution of Status Elements

(e.g., gender, skin color, age, beauty, occupation, Nobel Prizes, SAT scores)
Status     Performance     Translation to
Generalization     Expectation States     Behavior
Process         
    (e.g., perceived    
    competence at    
    specific and general    
    skills such as    
    reading ability, and    
    "most tasks')    

Figure 1. Status Generalization Process

The theory has been applied to numerous natural settings and abstract situations."ot surprisingly, most of the attention has been devoted to the "big two" characteristics in our society: gender and skin color. Status generalization effects have been documented for several other characteristics as well, and some of those characteristics stimulated our initial concern to extend the theory of status construction.

To clarify these ideas and to prepare for the extensions discussed below, we present a graph-theoretic model of the theory that has been used to derive predictions for several empirical tests4 Figure 2 depicts a simple

Some people mistake theoretical explanation for justification, but we do not believe that men should ever automatically receive more power and prestige than women. Beside being unfair to individuals, status generalization can deprive a group of good suggestions from low-status people, and can permit a group to be misled by bad suggestions from high-status people. To change an undesirable situation, however, one must understand it. Status generalization provides a theoretical foundation for many effective inter- ventions (e.g., see Cohen 1993; Cohen and Catanzarite 1988).

Assumption 5 of the theory translates expec- tations to behavior. According to this assumption, all components of power and prestige in the group will be direct functions of the relative perfor- mance expectations attached to actors. Many ex- perimental applications have used a linear func- tion modeling Assumption 5 to predict influence behavior. Balkwell (1991) noted some undesir- able mathematical properties of that function, and proposed another function. Balkwell's function

Elements of Interaction Power and Prestige

(e.g., chances to perform, number of performances, ageement, positive evaluations, influence, and choice for leadership positions)

status-expectations graph for task-focused intera~tion.~

In the directly observable (unshaded) part of Figure 2, status-expectation situations contain at least two actors, p (person) and o (other), two distinct outcome states T+ (suc- cess) and T- (failure), and an associated abil- ity C*, which also has high and low states (C*+ and CX-). C* is the specific status characteristic directly relevant to performing the group's task. These elements correspond to labeled points in a graph.

Connections between points are lines of three types. "Possession" may exist between an actor and a status characteristic, reflect- ing an idea such as "Actor p possesses a particular state of the gender status charac- teristic." "Relevance" may exist between non-actor elements; for example, gender status may be relevant to general ability to perform most tasks. "Dimensionality," a negative connection, exists between oppo- site states of a characteristic, such as be- tween Dl+ and Dl-; it shows that if an actor possesses one state, possessing the other state is thereby precluded.

predicts influence as well as the linear function; it also predicts other behaviors such as amount of talking in open interaction, and is more sound mathematically. In this paper we focus on how status characteristics are created. In translating combinations of status elements into behavior for experimental tests, one would use Assumption 5 and Balkwell's function.

"tatus diagrams and their analysis are de- scribed in Berger et al. 1977:114-21.

Figure 2. Graph-Theoretic Model of Status-Expectation Situations and Status Generalization

Here, p and o possess different states of a diffuse status characteristic Dl, which could be gender. Given what we know about our culture, p possesses the "male" state (Dl+), and o possesses the "female" state (Dl-). Di- mensionality links the two states of Dl. Of course C*+ is always relevant to T+, and C*- is always relevant to T-.

When status generalization occurs, (unob- served) connections form to link actors to outcome states of T according to the process described in explicit assumptions of the theory (see Balkwell 1991 or Berger et al. 1977). "General expectation states" T+ and T-are induced and linked by relevance bonds to the diffuse status characteristic D. Next, states of T are linked by relevance to states of C? The generalization process for specific status characteristics is similar to this, though not identical. For our purposes, the significant point is that both D and C characteristics function in status generaliza- tion, and they have comparable effects on ex- pectations and behavior.

"Strength" of performance expectations- roughly, how well an actor expects to per- form the group task-is a function of two factors. The first is the path length connect- ing actors to outcome states T+ and T-. The shorter the path, the greater its effe~t.~

Sec-

More precisely, strength is calculated by f-values for paths of different length; Fisek, Norman, and Nelson-Kilger (1992) provide theo- retically derived values. These are decimal pro- portions ranging from .6231 for length 2 (where an actor is connected directly to a state of C*, reflecting known ability at the group's task) to ,0211 for length 6, the shortest effective path. F-values quantify the intuitive belief that as more inferential processing is needed to connect an ac- tor to success or failure at a task, conclusions about his or her ability become weaker.

ond, the more paths connecting an actor to outcome states, the greater the effect on ex- pectations, because all possible connections function in status generalization.

Comparative expectation advantage of two actors is the difference between their aggre- gated expectations. For Figure 2, p is con- nected to outcome state T+ by two paths, one of length 4 and the other of length 5. The first path for p is

p-Dl +-r+---C*+-T+.

The second path, of length 5, is

(Because of the negative dimensionality link between two states of Dl, the path of length 5 becomes negative, and it terminates at T-, which also is negative; thus it is a positive path.)

Two paths, also of respective lengths 4 and 5, connect o to T-. The actors' relative ex- pectations may be represented as

p: +4, +5

0:-4, -5. We use paths and their lengths in some of the following discussions. Now we are prepared to address the cen- tral question of this investigation: Why are these attributes status characteristics that carry invidious distinctions and ideas about competence?
STATUS CONSTRUCTION THEORY

Status construction theory (Ridgeway 1991; Ridgeway and Balkwell 1997) identifies structural conditions sufficient to construct a status characteristic from an unevaluated nominal characteristic-that is, to transform states of an already identified nominal char- acteristic into states of a status characteris- tic. The theory can explain why gender, skin color, and (possibly) occupation are status characteristics. It does not claim to present necessary conditions, nor to deal with actual historical circumstances that mav have sur- rounded the acquisition of status value for any particular characteristic; rather, it identi- fies processes sufficient to create status char- acteristics when certain initial conditions are present (Ridgeway, 1991 :369). Indeed, our work may be regarded as identifying addi- tional circumstances sufficient to create sta- tus characteristics. Other theorists, such as Turner (1984, 1992, 1995) and Jasso (1994a, 1997), discussed below, offer historical ac- counts of inequality.

Status construction theory describes two related processes. First, a "differentiating re- source" process attaches status value to char- acteristics among certain actors through be- havioral enactment and misattribution mech- anisms. Second, a training or "altercasting" process transmits status value definitions to other actors through additional interaction mechanisms. We recast these processes in our first and second extensions below. Our third extension lacks a precedent in status construction theory.

Status construction7 begins with distinct states of a categorical or nominal character- istic, N; its states are denoted NA and NB. Ini- tially the nominal characteristic is just that: a descriptive attribute lacking connotations of social worth and performance capacities that would make it a status characteristic. Status construction theory explains how at- tributions regarding performance capacities and ideas of social desirability come to be attached to nominal characteristics.

Interaction, constrained by the distribution of individual characteristic variables, is cru- cial for acquiring, maintaining, and transmit- ting the status value of characteristics. Char- acteristics are almost never encountered in isolation: One never meets someone who is only (for instance) a rich person and who has no other socially significant characteristic.

'This follows the description in Ridgeway (1991), which gives close attention to the theory's micro-sociological processes. Ridgeway and Balkwell (1997) give more attention to macro processes.

AMERICAN SOCIOLOGICAL REVIEW

Most often we notice gender or race ("I met a rich man today"); thus gender and race are prime candidates for status construction. This association of different sorts of charac- teristics, including initially unvalued Ns with other characteristics and resources, can cause the states of N to acquire status value. Struc- tural factors (including the distributions of different characteristics) and social psycho- logical processes (including attribution, ex- pectation formation, and altercasting) are in- volved.

The process begins with a nominal (unevaluated) characteristic N-for instance, gender-that is socially salient, with an in- equality on some exchangeable resource, ER, and with regular association of a particular state of N-say NA-with a particular state of an exchangeable resource, ER+.8 An ex- changeable resource is a type of wealth in the society, something individuals value that can be transferred. Differentiating resources cre- ate differences in performance expectations associated with actors; roughly, these are an- ticipations about the likely quality of perfor- mance for the task at hand. Differences in performance expectations are enacted as dif- ferences in power and prestige behaviors in immediate interactions. Actors generalize, misattributing the behavioral differences (which initially were produced by expecta- tions from resource differences) to the corre- sponding differences on the nominal charac- teristic. Repeated experiences create status beliefs about the nominal characteristic itself.

If, for instance, people having state NA (men, in this example) regularly receive more of some exchangeable resource, such as salary, than people having the other state (NB, or women), then that resource difference will cause them to form performance expec- tations for actors that are consistent with the high and low states of ER they receive. The performance expectations affect power and prestige behaviors in immediate interaction;

As this terminology suggests, NA will come to acquire positive status value (that is, it will be- come N+) through association with ER+, and N8 will acquire negative value through association with ER-. This example shows how male and female, respectively, could become N+ and N-. Of course the theory is indifferent as to which state of N comes to be linked to the positive state of an exchangeable resource.

for instance, ER+ actors (most of whom also are men) become more assertive and talk more, and ER-actors (mostly women) defer. With time, actors come to misattribute their situational social standing to their differ- ences in gender (Ridgeway 1997b: 149). Thus actors may attach status value to an initially unevaluated characteristic N if actors pos- sessing a particular state of N regularly re- ceive a particular state of ER.9

Once differential expectations have been attached to a characteristic for some actors (those who interact with others having par- ticular states of ER and of N and who wit- ness an association of ER with N),they carry their status beliefs into future interactions in which the same characteristic is salient. This expectation transfer (see Berger, Fisek, and Norman 1989; Markovsky, Smith, and Berger 1984; Pugh and Wahrman 1983) sets the stage for altercasting and training pro- cesses by which status beliefs can diffuse throughout a society.

To review, status construction proceeds as follows:

(1)
    Individuals interact in "double dissimi- lar" situations, where a particular combina- tion of different states of N occurs in con- junction with particular states of ER. The states of ER create performance expectation differences that affect behavior. Actors with ER+ gain interaction advantages over those with ER-. This "engine" process is fueled by the (empirically given) association between states of N and states of ER.

The experience is repeated in new double dissimilar situations that also involve an association of N and ER. Other aspects of the new situations, including other character- istics of actors and goals of the interaction, may differ. The various situations seem to have nothing in common but the conjunction of a particular state of N with a particular state of ER.

In making sense of their experience, actors may infer (incorrectly) that "there is some reason" for the link of N with ER. Be-

This description is simplified; it presumes greater than usual regularity in social encounters. The theory does not require that all men receive higher salaries than all women; only that it is more likely to meet a person possessing both NA and ER+ than to meet a person with NA and ER-.

cause differences in ER have created differ- ential performance expectations, actors may misattribute differences in ability to the dif- ference in N. They come to believe, for in- stance, that actors having state NA actually can do "most things" more competently than actors having NB.At this point, these particu- lar actors have constructed independent sta- tus value for N; it now has connotations for performance and social desirability.

(4)
    Once these actors have developed dif- ferentiated expectations and behaviors in en- gine situations, they carry those expectations to new encounters. An actor possessing NA, for instance, may behave as if a stranger hav- ing NB is less competent than the actor with NA. Even if the stranger does not share the performance expectations, that stranger may align her or his behavior, showing deference in response to displays of confidence, and vice versa. Most remarkably, aligning can occur where resources are not salient, or even when actors differ in N but are the same in ER, because N by itself has become suffi- cient to create behavioral differences for at least one party to the interaction.
(5)
    Behavioral differences created during the aligning interaction will themselves cre- ate differences in performance expectations held by the new second actor. Thus the con- structed status value of a characteristic can diffuse throughout a society.

Before offering additional ideas, we pause to note the significant accomplishments of this work. The initial statement (Ridgeway 1991) offered the first explicit, micro-process theory explaining how macro-level status characteristics can be created. That theory described, for instance, how gender could become a status characteristic in a society as a result of individual-level interactions with relatively resource-rich men and resource- poor women. The explanation is not limited to our culture, however, nor to gender. It could be applied as readily to explain why female was the preferred state of gender in another society, if such a society could be found. The theory also could explain how skin color and possibly occupation could be- come status characteristics through regular interaction with individuals receiving differ- ential rewards.

In the following year, Ridgeway and Balkwell (1992) turned their attention to

AMERICAN SOCIOLOGICAL REVIEW

specifying macro-sociological processes (dif- fusion of status beliefs), and introduced quan- titative techniques to extend and improve the theory. In a recent statement, Ridgeway and Balkwell (1997) incorporate these and other improvements: (1) The range of situations is expanded from dyadic to group interactions.

(2) The theory is fully quantified, which im- proves the determinacy of its predictions and removes the need for certain simplifications in the initial statement, such as presuming dichotomous resources and nominal charac- teristics. (3) The theory takes explicit account of population distributions of resources and characteristics, which strengthens its macro- micro links. (4) The theory now explains de- cay of beliefs as well as belief acquisition (diffusion) throughout a population, through processes such as death and individual loss of belief. (5) The theory explicitly incorpo- rates additional social effects such as those of the mass media in transmission and repro- duction of beliefs. This feature extends its scope. (6) The theory permits explicit deriva- tion of consequences, thus enabling assess- ment of its correspondence to known evi- dence and also allowing predictions of novel tests. Finally, the explanation is corroborated by simulations (Ridgeway and Balkwell 1997) and by empirical tests of parts of the theory (Ridgeway et al. 1998; Ridgeway and Glasgow 1996).1°

lo Work on status construction processes has proceeded rapidly, and publication dates do not always represent accurately the intellectual growth of a theoretical research program. Ridgeway's (1991) paper was the first published statement, though some ideas of the theory were presented in talks prior to that date. That paper gives close attention to micro-sociological pro- cesses in status construction. Ridgeway and Balkwell presented their first quantitative state- ment at the ASA annual meeting in 1992. After further development, this work was published as Ridgeway and Balkwell (1997); it explicates macro-sociological diffusion processes. Experi- ments in Ridgeway et al. (1998) grew out of ideas in Ridgeway and Balkwell's (1992, 1997) state- ment. Preliminary analyses from those experi- ments were reported at ASA annual meeting in 1995, but of course the experiments do not pre- date Ridgeway and Balkwell's theoretical work. The experiments by Ridgeway and Glasgow (1996) were similarly informed by the developed theory and model.

Although explicating all aspects of status construction theory is beyond the scope of our work here, we do offer three theoretical extensions and elaborations. First, in the types of situations the theory considers, we generalize the theory to account for charac- teristics other than gender and skin color. Second, we explicate an interaction mecha- nism suggested in the initial statement of sta- tus construction theory, which transmits sta- tus value to actors in new situations. As a re- sult of this work, we discover that the same mechanism produces an important additional effect previously unremarked: The interac- tion process can create status value as well as transmit it, and it can do so for types of char- acteristics other than those considered in sta- tus construction theory. This mechanism even creates status value where no characteristic of any sort initially is salient. Third, we describe how status characteristics can be created in situations where initially no characteristic ex- ists. We believe that certain combinations of norms and roles, together with attribution and enforcement processes, are sufficient to cre- ate evaluated characteristics which then may become status characteristics."
FIRST EXTENSION: STATUS VALUE AND THE REWARD EXPECTATIONS MECHANISM

For the first extension of status construction theory, we have two goals. First, we provide a theoretical explication for the "differenti- ating resources" mechanism that links evalu- ations and performance expectations to ini- tially unevaluated nominal characteristics. Second, we show how this can extend the theory's scope beyond the "big two" charac- teristics to others whose link to exchangeable resources is nonexistent.

More than 30 years of research into status generalization processes has identified sev- eral status characteristics other than gender, the primary focus of status construction theory. For instance, Webster and Driskell (1983) and Jackson et al. (1995) have shown

Ridgeway (199 1 : note 1) identified our third task as appropriate for later investigation. This is actually the most advanced problem from a so- cial constructionist perspective because it begins with a nearly undefined situation.

GO-

Figure 3. Basic Status-Reward Expectations Situation Sources: Adapted from Webster (1984:79) and Berger et al. (1985:226).

that physical attractiveness-beauty-is a status characteristic in our culture; that is, its states are accorded different social value and are associated with a wide range of expected performance capacities. Also, Johnson (1995), Stulhofer and Francetic (1996), and Webster, Hysom, and Fullmer (forthcoming) show that sexual orientation-being heterosexual or homosexual-is a diffuse status characteristic. It is likely that individuals have noticed men receiving more exchange- able resources than women, but it is less plau- sible that they would notice resource differ- ences associated with beauty or with sexual orientation. That is, it is easy to show that people think men are paid more than women, but it might be less easy to show that people think attractive faces or heterosexuals are paid more than plain faces or homosexuals. l2 Thus status construction theory may not ap- ply to some other status characteristics as successfully as to gender and skin color. For- tunately, a simple conceptual change will ex- tend the theory to account for many other characteristics, including sexual orientation and beauty. For these purposes, we use vari- ants of the status-expectation diagrams intro- duced earlier.

l2 Whether income levels actually differ is ir- relevant if actors believe that they differ.

Figure 3 shows an extension of the status- expectation diagram of Figure 2, incorporat- ing reward expectations. The added elements are R+ and R-, representing high and low "reward levels" (an abstract concept), and GO+ and GO-, positively and negatively evaluated "goal objects" (specific realiza- tions of reward levels in a given situation). If relevance bonds extended from Dl+ to GO+ or to R+, and from Dl- to GO- or to R-, the figure would represent a system in which people believed that gender was rel- evant to goal objects or to reward levels: Men would be expected to receive more of that society's rewards than women. Links between GO and R show that individuals ex- pect that, in a particular pair consisting of one man and one woman, the man might re- ceive a wage of $15.00 per hour, and the woman $12.50. In fact, any wage differen- tial favoring men would be represented by paths connecting Ds (states of the diffuse sta- tus characteristic gender) with Rs (states of reward level). l3

l3 As the example shows, many specific wage levels (GOs) would fulfill the criterion of higher rewards (Rs) for men than for women. The theory explains why people expect men to be paid more than women; it does not predict any particular wage differential or wage ratio. Also, the instantiations of Dl might differ. The theory is

AMERICAN SOCIOLOGICAL REVIEW

~7

P N~ C+ T+

(-1

1
0 T

NB 8;:

GO-

Figure 4. Mechanism Linking Nominal Characteristic to Performance Expectations

Source: Drawn from a description in Ridgeway (1991:371-73).

In status construction theory, ideas about competence or performance expectations are created by differences in an exchangeable re- source. A parallel mechanism exists in a vari- ant of status generalization theory that deals with formation of reward expectations (Berger, Fisek, Norman, and Wagner 1985). That theory, supported by at least five experi- mental tests (Bobrow 1973; Cook 1975; Harrod 1980; Parcel and Cook 1977; Stewart and Moore 1992), provides a mechanism for forming performance expectations for actors consistent with their known reward alloca- tions: high expectations for those receiving high rewards, and low expectations for those receiving low rewards. Expectations form, and behavior follows in accordance with those expectations as a consequence of un- equal allocation of valued goal objects. An allocation difference is sufficient to produce differential performance expectations, which affect behavior involving actors for whom both an allocation and a nominal character- istic are salient. It is those evaluations and performance connotations that distinguish a status characteristic from a nominal characteristic. Thus gender, which may have been

indifferent to which state of D is Dreferred; thus it also applies to any society in which women en- joy status advantages over men.

only a nominal characteristic, becomes a sta- tus characteristic because of its association with differential rewards. Eventually perfor- mance expectations come to be associated with the characteristic itself, not with the in- dividuals who happen to possess it.

The structure of expectations in a small task-focused group creates a corresponding structure of power and prestige behaviors: Actors for whom higher expectations are held enjoy more deference, attention, posi- tive evaluations, and influence in the group. Later those expectations and behaviors can transfer (with attenuation) to new encoun- ters. It is worth examining the acquisition process in some detail to understand exactly how an unevaluated characteristic acquires status value from association with an ex- changeable resource or other goal object. (In Extension 2 we will explicate how interac- tion can create status characteristics and transfer their effects to new situations.)

Figure 4 represents the situation in the five reward expectation experiments men- tioned above. Here, p and o possess differ- ent states of a nominal characteristic N: Na and NB. Because NA and NB are mutually ex- clusive (though not yet evaluated), Figure 4 includes dimensionality between them. The actors, however, observe that people Passessing NA regularly receive a higher state of goal objects than do people possessing NB. "Goal objects" are socially preferred and nonpreferred outcomes in a particular situation.14 Relevance bonds connecting states of N to states of GO show that actors are connected indirectly to states of reward levels (R) and to states of the specific task ability (C*). In the allocation experiments, subjects received different amounts of money or points convertible to money as goal objects.

Figure 4 (and Figure 3) use "goal objects" instead of the "exchangeable resources" of status construction theory as instances of re- ward levels. This change may appear minor, but it is important. The concepts are differ- ent, and "exchangeable resources" is unnec- essarily limiting. Goal objects encompass both "consummatory value" (as does "ex- changeable resource") and also what Veblen ([I8991 1953) called "status valuen-as in an honorary title, a reserved parking space, or a corner office-which "exchangeable resource" does not connote.I5 Intangibles such as esteem, scorn, friendship, harassment, and acceptance or rejection are status-valued goal objects.

Although it may not be plausible to claim that people have noticed ugly people or ho- mosexuals receiving lower salaries, it is plausible that they have observed differential al- location of status-valued goal objects on such a basis. Almost anyone can recall cases in which an ugly face or gender nonconformity caused pain in elementary or high school; al- most anyone would say that heterosexual pat- terns often are admired and homosexual pat- terns often despised in adult life; almost any- one can recall how advantages seem to flow to attractive children and adults.

According to the theory, all status ele- ments, goal objects, and reward levels in a situation can become salient, and thus they can function in status generalization pro- cesses. Actors will become connected to task

l4 Whether an object has status value and whether that value is positive or negative depend on social definitions and history; for instance, po- litical titles can lose value or even change valence when governments fall.

l5 Veblen ([I8991 1953) believed that once in- dividuals surpass subsistence level, most rewards and goal objects rely more on status value than on consummatory value for their desirability.

outcome states (T+ and T-) through all pos- sible paths; the presence and distribution of goal objects caniffect performance expecta- tions associated with actors, and thus their subsequent behaviors. Among other conse- quences derivable from the theory (Berger et al. 1985:239-52), it can be shown rigorously (in Theorem 6; Berger et al. 1985:25 1-52) that actors will come to hold reward expec- tations-expectations for different reward levels-that are consistent with the relative performance expectations they hold for each other. The results of the five experiments mentioned above can be explained as a re- verse process in which actors come to hold performance expectations consistent with the distribution of valued goal objects. This is the theoretical foundation for our extension of the "differentiating resources" mechanism of status construction theory.

When status construction theory is modi- fied to rely on positively and negatively evaluated goal objects, it incorporates many other characteristics beside the "big two" of race and gender. The modification still per- mits tangible goal objects such as exchange- able resources; thus all cases within the original theory are accommodated. It also draws attention to the great importance of status value in the social construction of sta- tus characteristics; thus status construction could occur even in societies without money economies.

If we lacked the theoretical foundation for the differentiating resource mechanism and had only the results of those experiments, scientific caution would limit the discussion to ERs instead of GOs. In all experiments so far, exchangeable resources-either money or points exchangeable for money-have been used as goal objects. The theoretical foundation for that process is broader, how- ever, and we believe that status construction theory can benefit from incorporating status- valued goal objects as sources of perfor- mance expectations in the acquisition pro- cess.

Our first extension predicts that both the status value and the consummatory value of goal objects will produce the effect. Al- though the status value part of that predic- tion is untested, it would be relatively straightforward to conduct an experiment in which status-valued goal objects were allo- cated, using procedures already developed for allocation experiments. Now let us exam- ine further extensions of the theory.
SECOND EXTENSION: CREATING STATUS CHARACTERISTICS FROM BEHAVIOR PATTERNS

Our second extension describes ways in which status value may be constructed in situations involving no exchangeable resources or other goal objects. This extension must use a route other than reward expecta- tion processes because an essential element, a goal object, is absent. For this extension we use ideas from another line of theoretical work on interaction sequences, recently called "behavior interchange patterns" or BIPs (Fisek, Berger, and Norman 1991). Be- havior interchange patterns-consistent sets of unit acts displaying subordination and superordination -can produce performance expectations based on the rates and types of behavior displayed by interactants. We argue that BIPs can construct status characteristics from unevaluated nominal characteristics. Further, BIPs can create performance expec- tations independent of any salient character- istics and can attach those expectations to actors or to initially nonsalient characteris- tics that the actors may possess.

Small-group interaction is crucial in sta- tus construction theory: It conveys the dif- ferentiated performance expectations formed by resource (or goal object) inequal- ity, and through altercasting and training it can transfer those expectations to new situa- tions involving some new actors who pos- sess similar characteristics to those of actors in the original situation. As we noted above, Ridgeway et al. (1998) show that status be- liefs are created by differentiating resources combined with behavioral differences and referential information. Here we explicate some interaction processes and so provide a theoretical foundation for behavioral mechanisms that may be involved, and un- cover additional consequences of those mechanisms. We begin with two sets of em- pirical studies that show the importance of interaction in creating and modifying status- related behaviors.

Theories of status generalization explain many behaviors as consequences of perfor-

AMERICAN SOCIOLOGICAL REVIEW

mance expectations that actors come to hold; Assumption 5 of the versions cited above (in note 4) tells how expectations translate into power and prestige behaviors. More recent studies show that under certain conditions, a reverse causal ordering can take place, such that some behaviors can affect performance expectations. Much of this theoretical work is compatible with the altercasting and train- ing processes of status construction theory. We describe this work and explicate the idea of a behavior interchange pattern, which is the foundation for our second extension of status construction theory.

Moore (1985) reported experimental evi- dence showing a behavioral process by which expectations held by one actor (o), can come to create complementary expectations held by a second actor @) Roughly, this ex- plicates behaviors involved in translating p's belief "I know you think I have higher abil- ity than you," into "I think I have higher abil- ity than you do." Moore's experiments showed that in a group where actors are sub- ject to pressure to co-orient their behavior, if one actor (0) acts as if she will accept influ- ence, p may become correspondingly more resistant to influence. That interaction pat- tern, when it continues, is sufficient to cause p to form higher expectations for herself than for o-for instance to decide that she actu- ally has more ability than o and is perform- ing better at the task. In these studies, the power and prestige behaviors led to predict- able patterns of consistent performance ex- pectations; at that time, however, the mecha- nism was not clear.

Driskell (1982) showed that "valued per- sonal characteristics" (VPCs) could alter sta- tus behaviors. A VPC fulfills only Part 1 of the definition of a diffuse status characteris- tic (see p. 353); that is, it has two or more states that are evaluated differentially.16 A VPC carries no specific or general perfor- mance connotations, however. Thus someone "should" or "ought to" have one state rather than another, and one state is more desirable than the other, but no difference in perfor- mance capacities is implied. VPCs include honesty, trustworthiness, friendliness, and

l6 Berger, Wagner, and Zelditch (1985:47 ff.) use moral characteristic, apparently to denote the same concept.

Figure 5. Behavior Interchange Patterns Linking Actors to Performance Expectations

Source: Drawn from Fisek, Berger, and Norman (1991 :122).

empathy. Driskell's study showed that differ- entiating actors on empathy produced behav- ioral inequalities comparable to those pro- duced by a status characteristic. As a result of later theoretical developments, we now believe that the organizing ability of VPCs is due to an intervening behavior interchange pattern. Behavior interchange patterns, we believe, can transform VPCs into diffuse sta- tus characteristics.

Another line of work on behavioral pat- terns helps to clarify the mechanism at work in those experiments, leading to the present theoretical concept of the behavior inter- change pattern. Berger (1958), describing how cycles build and stabilize, identified an "idealized interaction process" consisting of an action opportunity, performance output, and reward action. Berger and Conner (1969, 1974) developed the idea of "full fundamental sequences" of behaviors lead- ing to and being controlled by expectation states. In that work they set forth a typology of behavior and specified how types of be- havior follow each other in unit sequences. Empirical (Willard and Strodtbeck 1972)

and theoretical (Canner 1977, 1985) studies have shown how sty1e hesitations and pauses, can lead to forma- tion of performance expectations: Interactants read those and other "status cues" (Berger et al. 1986) and infer performance expectations from them." Fisek et al. (1991) developed a significant theoretical reformulation that incorporates these ef- fects and permits graphic representation of situations in which behavior can create ex- pectations.

Balkwell (1991) and Skvoretz and Fararo (1996) developed somewhat different ver- sions of behavioral interchange patterns in theoretical models of interaction. We present one version below.

Figure 5 shows a status-expectation dia- gram incorporating behavior interchange patterns. The actors are p, a male, and o, a female. Assume that a behavior interchange pattern has developed favoring o, as repre- sented by the possession lines from o to b+ and from p to b-. Sequences of specific be- haviors, such as hesitation or speaking readily, are shown as b-and b+ respectively. Behavior sequences (b+ and b-) are connected by relevance to corresponding "status typification states," a theoretical construct represented by B+ and B-. Status typification states are behavioral classifica-
,,Ifimportant

to distinguish beha"

iors, which are of interest here, from dominating behaviors, which are not (see Ridgeway and Walker 1995:301and note 1). BIPs are behaviors consistent with high or low status. They are not attempts to threaten or to seize control.

tions that actors may make; "leaderlfol- lower," "initiatorlreactor," and "assertive1 shy" are expressions of the construct. Actor p may think "That person is a leader; she seems to have general problem-solving abil- ity, including the ability to solve our prob- lem now." Typifications are relevant to ab- stract task ability, represented in Figure 5 by Y, an induced element in status-expectation theories. States of Y are relevant to the task outcome states T+ and T-, and thus they complete the connection of actors to. task outcomes. For the case shown in Figure 5, status advantages conveyed to p by the dif- fuse characteristic D are balanced exactly by status disadvantages arising from behav- ior shown as b-. The more typical pattern of diffuse status advantage, accompanied by precedence in behavioral interchange, would augment status advantages for p and disadvantages for o.

To understand how VPCs can become sta- tus characteristics, recall that the theoretical task is to explain how the characteristic ac- quires social evaluation and performance connotations. The difference between a VPC and a status characteristic consists precisely of those connotations, and a behavior inter- change pattern is one possible way to create performance expectations. For instance, group members may believe that one mem- ber is unusually honest, friendly, or empathic (VPCs); consequently, they may give that actor additional action opportunities, posi- tive evaluations, and agreement. Those be- haviors, in time, can lead to forming high expectations for the actor so favored. In other words, we claim that VPCs can become status characteristics if they convey certain kinds of interaction advantages.

This argument also can apply to uneval- uated characteristics, such as NA and NB discussed above. For instance, cultural differ- ences in speed of speaking when offered an action opportunity might initially be unevaluated. Yet speaking slowly could bias interaction to the detriment of slow speakers. More generally, any characteristic affecting power and prestige behaviors in a group, such as chances to perform, likelihood of ac- cepting a given chance to perform, and evaluations of single problem-solving at- tempts, can create differential performance expectations for actors. The BIP process does

AMERICAN SOCIOLOGICAL REVIEW

not even require the presence of a salient characteristic. The biasing factor might be sentiment; Shelly and Webster (1997) show that liking among group members can affect interaction and thus can influence power and prestige behaviors. l8

The altercasting and training mechanisms of status construction theory can work through BIPs. If an actor holding certain per- formance expectations-because of the misattribution process described earlier- meets a second actor who does not hold ex- pectations linked to that characteristic, and if the first actor enacts his or her expecta- tions, there is some chance that the second actor will align his or her behavior to that of the first. If the first actor enacts "high-status behavior," the second may align and display "low-status behavior," and vice versa. If that happens, BIPs will cause the second actor to develop performance expectations consistent with the behaviors he or she displays. In this way, a characteristic having evaluations but no expectations (a VPC) can acquire expec- tations, and a characteristic with neither evaluations nor expectations (an N) can ac- quire both.

In general, any factor that biases behavior sequences to the advantage of one actor over another-whether that factor is a VPC, an N, a sentiment process, a structural factor such as formal authority, or individual differences in style-can lead to differential behaviors and performance expectations. That process does not require the presence of any charac- teristic. Behavior interchanges can create sta- tus value, no matter what triggers them. When performance expectations already ex- ist, behavioral interchanges will realize and maintain the pattern of performance expec- tations, and thus will maintain the power and prestige inequalities they have created. Ex- pectations originally created by a goal object that later loses status value (such as a royal title that becomes meaningless after a revo- lution) can maintain power and prestige ad-

l8 If actors were perfectly task-focused, only expectations would determine behavior, but per- fect task focus is difficult to imagine. Usually, we expect actors to allow other processes, including sentiment, to affect their behavior, even if subtly. For example, smiling while delivering a positive evaluation to a liked person, increases the impact of the evaluation.

vantages and can continue creating expecta- tion and interaction advantages in this way.

THIRD EXTENSION: DEFINING A

CHARACTERISTIC AND CREATING
STATUS VALUE FOR IT

Finally, we consider ways in which a new characteristic might be created, where ini- tially none is salient or even present. If this process is understood, then the BIP process in our second extension can explain how the new characteristic can acquire status value through interaction. This discussion is neces- sarily more speculative than the treatment of the first two extensions: It relies on less fully developed theory than the preceding. What conditions could lead to the creation of a characteristic and the bestowing of status value where previously no characteristic was evident? Related processes may cause extinc- tion as status characteristics fall into disuse.

Disparate research and theory may be adapted to the task of explaining creation and valuation of new characteristics as societies change. We adapt these ideas from the litera- tures on attribution processes, personality at- tribution, and deviance and social control, and use them to show three general tenden- cies: (1) to recognize and categorize differ- ences among individuals; (2) to impute dis- positions from behavior, using existing cat- egories; and (3) to create categories where previously they did not exist, given certain situational and interaction conditions.

Psychological research on attribution (Fiske 1992; Fiske and Neuberg 1990; Fiske and Taylor 1991) shows that "mere differ- ences," especially in appearance variables, can be sufficient to categorize individuals: People construct cultural schemas to identify self and others. Sex may be a fundamental category (Brewer 1988; Conway, Pizza- miglia, and Mount 1996; Ridgeway 1997a). Ridgeway and Balkwell (1997, note 5) as- sumed in their simulations that a tendency to prefer one's own category should have little overall effect on the status-creation process.

In status approaches to attribution, re- searchers investigate how individuals come to display behavioral consistency in interac- tion with specific others, and how actors come to think of themselves and each other as having certain traits, thus enabling them to anticipate future behavior (Gerber 1996; Johnston 1985, 1988). This work explores individuals' frequently noted tendency to de- velop complementary roles when they inter- act repeatedly. For instance, marriage part- ners may fall into roles of "worrier" and "op- timist" or "disciplinarian" and "nurturer." Some marital roles have clear evaluative connotations, but many do not. Status ap- proaches to control processes (Berger 1988; Talley and Berger 1983) describe the emer- gence of a particular set of roles, norm car- rier and norm violator, which carry clear evaluative connotations. This work describes conditions under which roles emerge and are maintained. Both branches describe situa- tions in which no status elements are present at the outset.

Johnston (1985, 1988) developed the idea that specific personality attributes-behav- ioral predispositions that individuals are be- lieved to possess-are likely to be inferred or imputed in cases of ambiguous communica- tion, where actors are not certain of each other's tendencies or their likes and dislikes. In such cases, Johnston proposes, they will turn to diffuse attributes and will "general- ize" from them in ways comparable to status generalization processes. For instance, a man who knows that his friend is an energetic, daring, exciting woman could infer that she might like to watch an adventure movie. Be- havioral consequences of this process include role behaviors: The man will treat her as if she possessed the imputed characteristic, and she, in turn, will enact a complementary role. Johnston used the idea of personality attribu- tion to analyze families in which, for in- stance, a "strict, rational" father administers discipline while a "sympathetic, emotional" mother attends to feelings. For our purposes, the significant point is that indi~id~als

structure ambiguous situations by inferring (per- sonality) traits, attaching (attributing) them to individuals, and then acting in ways that tend to perpetuate the patterns (roles).

Gerber (1996) studied police car teams and found that this process occurred regularly in that work situation. Status differences be- tween members of a team that were actually produced by gender or by seniority led to predictable differences in imputed personal- ity traits and behavior. For instance, higher- status team members were more likely to be proactive, directive, and task-focused; lower status members tended to be reactive, com- pliant, and process-oriented. Furthermore, police team members noticed those behav- iors-actually consequences of relative sta- tus positions-but attributed them to the in- dividuals' personality traits. In teams of two men, for instance, the more senior partner would describe himself as "the one who takes charge and organizes the team; I'm a doer," while his partner "takes care of people's feelings; he's more caring." Thus Gerber's work shows that behaviors which are actually status-based can lead to attribu- tions of personality traits and dispositions.

Deviance and control processes may be how entirely new characteristics come into being. In The Invention of Heterosexuality, Katz (1995) traces historical circumstances that may have created that characteristic.19 Control processes occur within social frame- works where norms and understandings ex- ist, indicating what is appropriate and what is inappropriate, and they can, we believe, lead to activation of certain roles and cre- ation of new characteristics.

In outline, the control process is described by Berger (1988), Talley and Berger (1983), and Wagner (1988b) as follows.20 In a situa- tion containing norms or understandings that define accountable (that is, evaluated) inter- action, a control process is activated by un- expected events that violate such norms and

l9 In Katz's (1995) account, heterosexuality first was identified as excessive interest in sexual activity without concomitant desire for procre- ation; that is, with proscribed recreational sexual activity. As the role "heterosexual" was distin- guished, its associated norm violation "hetero- sexuality" was named. Heterosexuality thus was formed with negative value, as norm violation was identified and attached to actors. Katz cred- its Foucault (1978, 1985, 1986) with directing at- tention to ways in which social structures and in- stitutions, such as academic social science, schools, and mental hospitals, create and enforce definitions of categories, especially those subject- ing individuals to the classifiers' control. Fou- cault (1983) contains a summary of some of his ideas, especially on power and control. Foucault apparently supposes that classification precedes rather than follows norm violation, the reverse of the causality described by Katz.

20 This follows the description in Wagner and Berger 199359-60.

AMERICAN SOCIOLOGICAL REVIEW

understandings. As deviance theorists have noted, the typical initial response to those events is to try to ignore or to normalize them (e.g., Becker 1973; Scheff 1966). Where such efforts fail, actors initiate behav- iors aimed at managing the events and defin- ing or redefining actors' moral relations to each other. An outcome of those redefinitions is the creation of norm carrier and norm vio- lator roles. Along with role definitions, ac- tors may create characteristics, inferred dis- positions that are part of the roles. These roles then affect future behaviors (such as apologies and outrage), as well as the inter- pretation of future behaviors. (Greater detail is available in Berger 1988:464-69.)

Norm violation thus can produce the complementary roles of norm carrier and norm violator. A norm carrier might say, "You did that and it violated our understand- ings." The violator could argue, offer ac- counts, or accept the role and apologize. Whether or not the role of norm violator is accepted, and whether or not the control pro- cess is resolved, it will have cognitive and behavioral consequences including assign- ment of personality traits, activated affect and sentiment states, behaviors, and some- times even the creation of new norms spe- cific to the pair relationship.

In a simple case, when an event occurs, a norm carrier emerges to display disappoint- ment or outrage, and another actor accepts the role of violator. Before roles have been attached to actors, their acts are not problem- atic, even if those acts are what-is often called "primary deviance" (Lemert 1951). That is, whatever an outside observer might think of the acts, they have not been labeled deviant by those involved. After roles are at- tached, "secondary deviance" may occur: The norm carrier becomes more vigilant and more likely to impute deviance to acts and deviant motivations to the norm violator. and the norm violator may decide to live up to the new (deviant) role. Interaction patterns change, structured by the roles that have emerged.

TWO types of norms, which we call adequacy and excellence, function in the devi- ance process. Norms of adequacy specify minimal or expected levels for people to meet. They define roles such as "healthy1 sick," "sanelinsane," "mature/immature," or "normal/weird." In most cases, norms of ad- equacy are categorical rather than continu- ous. That is, as long as an act or a person is within a particular range, the exact location in that range is unimportant. Whether an act is normal or weird matters more than how weird it may be. Similarly, several different levels of physical well-being are classified as "healthy," from being able to conduct a sed- entary life using few medications to being an Olympic athlete.21 Norms of excellence, however, are graduated: They define perfor- mance capacities such as "expert," "skillful," "beginner," "incompetent." One lives up to these norms to varying degrees, and degree is significant. The difference in spelling abil- ity between a slow fourth-grade child and a national spelling champion in high school is enormous and socially significant-many intermediate levels exist.

Norm violation is clearly meaningful for norms of adequacy; a person belongs to one category or the other. In such a case, every- one "ought to" fall into one category. What it means to violate a norm of excellence dif- fers somewhat, because not everyone is ex- pected to attain the top of the range. Rather, individual performance may be compared with the standard and judged "normal" if someone is average, "above average" or "ex- cellent" if above, and "below average" or "failing" if lower. In these cases, it is "fortu- nate," "good," or "pleasing" when an act is toward one end, and "unfortunate," "bad," or "disappointing" when toward the other end.22 The third stage begins when an actor cast into the role of deviant (whether or not

21 Brekhus (1996) describes "binary" and "trinary" models of markedness, which appar- ently have similar properties. Binary models dis- tinguish zones of "perverse" and "generic" (or typical) behavior; trinary models distinguish "perverse" and "exceptional" at the ends of a con- tinuum and "generic" behavior in the middle range.

22 Insights on norm violation and the assign- ment of roles are available in the work of social constructionists, including Tannenbaum (1938), Lemert (1951), Erikson (1966), Goffman (1970), Becker (1973), and Scheff (1974). At first, a so- cial system contains widely shared definitions of acceptable and unacceptable behavior, often called "moral boundaries." Boundaries may shift, depending on circumstances including ethnic and

the casting was resisted) commits new simi- lar acts. This is "secondary deviance," enact- ment of the deviant role that was adopted willingly or resisted unsuccessfully. Second- ary deviance is more likely than primary de- viance to be noticed and classified as devi- ant and less likely to be "normalized," and it confirms moral boundaries, deviant roles, and imputed traits and characteristics.

In a situation where norms define a moral boundary, and actions therefore may be de- fined as normal or deviant, the stage is set for a new characteristic to emerge and (un- der certain conditions) to acquire status value. One way of looking at norm-following and deviance is to say that they are treated similarly to performance outputs in task-fo- cused interaction. That is, norm conformity and deviance are subject to unit evaluations whereby normative acts receive positive evaluations and deviant acts receive nega- tive. Thus, whenever cultural norms exist to define limits on acceptable behavior, acts be- come subject to positive and negative evalu- ations as they conform to or deviate from the moral boundary.

In studies of cognition (e.g., Brewer 1988; Fiske 1992; Heider 1958; and Neuberg and Newsome 1993) researchers note that viola- tion of expected behavior, including viola- tion of the moral order, provides contrast suf- ficient for perception. For norms of ad- equacy, only violation may be noticed, and it is evaluated negatively. For norms of excel- lence, an action at either extreme may be no-

cultural diversity, but they always define limits to acceptability. For convenience, we call accept- able acts "normal," and unacceptable acts "devi- ant." Acts tend to be viewed as normal even when they are problematic, to be "normalized" through excuses, accounts, and disclaimers.

The second stage begins when the acts are no- ticed by observers (or by the actor herself) and are classified according to moral boundaries. Some gray area always exists, and whether an act is made deviant depends partly on history, being more likely when the act is repeated and when extreme violations occur. If an act is classified (consensually) as deviant, it is an instance of "pri- mary deviance." An actor responsible for deviant acts may be cast into the role "deviant," though the likelihood of that occurrence also depends on history, on whether the actor resists the role, and on her or his resources for contesting it.

AMERICAN SOCIOLOGICAL REVIEW

ticed and may receive either negative or positive evaluation, depending on whether it fails to meet the standard, or surpasses it. In either case, evaluation appears simulta- neously with salience: An act becomes sa- lient only when it is evaluated. This is the transition from "primary" to "secondary" in the social creation of deviance.

If an evaluated act is viewed as a person's disposition, it becomes a characteristic-in fact, a valued personal characteristic. The at- tribution process here results in outcomes described by Johnston (1988), Conway et al. (1996), and Gerber (1996). Little is known of the conditions that must be present so that attribution of a characteristic will occur, but the great number of cases of attribution ar- gues that those conditions occur frequently. Certainly attribution becomes more likely when norm conformity is especially impor- tant to a group, when an act is repeated, and when departure from the norm is extreme.

Because we are concerned with creation of status characteristics, we focus on situations of task orientation and collective orienta- tion-the scope conditions for status gener- alization processes. In those conditions, VPCs may be created from any case of norm violation. As discussed above, VPCs can lead to formation of performance expectations by affecting the behavior interchange process. Thus we posit a two-stage process by which norm violation can lead to creation of status characteri~tics.~~

First, as discussed here, violation can produce a salient characteris- tic. Furthermore, because it arises in com- parison with either a normative standard of how things "ought to be" (for a norm of ad- equacy) or a desirable standard (for a norm of excellence), a characteristic so created is evaluated from the moment it is created; it is a VPC. Next, through the process described

23 It is an open question whether a moral boundary presumes existence of a characteristic (as Foucault [1985, 19861 seemed to believe) or whether a role-creating control process, as de- scribed here, also is needed. Perhaps a way to ap- proach the question is to ask under what conditions a moral boundary can exist without a named characteristic. For our purposes, the characteris- tic could either preexist or be created by a con- trol process. In either case, it affects BIPs and thus creates performance connotations that can create a status characteristic.

in our second extension, a VPC can become a status characteristic because it affects be- havior interchange patterns.
TESTING THE THREE EXTENSIONS

Experimental tests of the three extensions can provide information on their accuracy and determinacy, as well as suggesting av- enues for further theoretical elaboration. Here we outline three possible tests, all us- ing variants of a basic experimental design often employed for studies of status gener- alization (described in Berger et al. 1977, chap. 3). In this design, information on sta- tus or performance is presented to pairs of subjects in Phase 1. A standard measure of expectation states-influence over team de- cisions-is collected in Phase 2. The influ- ence measure, one component of a group's power and prestige structure, has proved to be a reliable measure of relative performance expectations of actors.

Extension 1

In the basic test, the designs used in the five reward-expectation experiments would be varied by allocating different states of goal objects possessing status value and virtually no consummatory value. Such goal objects might include honorific titles, distinctive clothing or medals, or obviously high-qual- ity or low-quality furnishings associated with actors. If pairs of individuals in the basic ex- periment were allocated different states of status-valued goal objects, that allocation should (if our extension is correct) produce a consistent difference in performance expec- tations. A further test might add comparable conditions in which actors received different consummatory-valued goal objects ("ex- changeable resources") such as money. That step would allow comparisons for refining the theory, for instance, for questions such as whether status-valued and consummatory- valued goal objects have equal power to structure performance expectation^.^^

24Thi~raises an interesting possibility. Perhaps evaluations based on goal objects transfer to nominal characteristics only when Ns become rel- evant to GOs in a particular situation.

Extension 2

According to the second extension, charac- teristics lacking performance expectations can acquire such expectations through their biasing effects on interaction. Different set- tings might test this extension. For instance, a discussion group could be used in which one individual was distinguished by a VPC such as friendliness or unfriendliness. The setting employed by Ridgeway et al. (1998), in which the discussion group includes a trained confederate, might be used for this test. Confirmation would require that VPC advantage and disadvantage lead to a corre- sponding advantage and disadvantage in the power and prestige structure, perhaps as measured by participation rates and by a q~estionnaire.~~

For further exploration, one might examine how much time was required for an association to form between Ns or VPCs and performance expectations. It seems unlikely that a single behavioral en- counter would be sufficient, and it is of in- terest to specify how many would be needed.

A laboratory test of this extension would have abstract features similar to the discus- sion group design, but might produce clearer results. We would use variants of the design employed by Driskell (1982): Differentiating actors by a VPC in some conditions and by a D in others would permit comparison of their efficacies in creating performance expecta- tions through behavioral interchanges. Un- like Driskell (1982), however, we predict that differentiation would take some time to become apparent, because of the need for in- teraction to translate the friendliness or em- pathy into behavioral advantages and disad- vantages.

25 Ridgeway et al. (1998) present experimental data showing that differences in a nominal char- acteristic linked to ER differences, along with be- havior consistent with the resource difference, produced status construction. Ridgeway and Glasgow (1996) show that behavioral differences alone are sufficient to create status value. These experiments are consistent with status construc- tion theory; additional confirmation, however, is desirable. The Ridgeway et al. (1998) experiment incorporates behavior patterns and status cues in addition to exchangeable resources. A simpler design using only differentiation on a status-val- ued goal object would test our first extension.

Extension 3

This extension might be tested in an experi- mental setting that includes a well-known set of rules, such as a game. Actors playing the game might include a confederate of the ex- perimenter, who had been instructed previ- ously to commit a clear violation of a central norm of the game. This first experiment would be concerned primarily with demon- strating that a characteristic could be created in this way rather than with details of the cre- ation process; thus the norm violation could be repeated several times to ensure that it was noticed and to reduce the effectiveness of any attempts to normalize it. At this point, measures could be introduced to ascertain that individuals indeed had formed an idea of a characteristic.

Further work might investigate processes by which a characteristic comes to exist in- dependently of the actor who embodied it. For instance, if one actor created a VPC by displaying a tendency to cheat at a game in some distinctive way, that behavior might lead to changes in interaction patterns, such as surveillance and taking precautions. In later encounters with others having the same characteristic as had been "discovered" in the first encounter, the same interaction pat- terns might occur.
RELATIONSHIPS TO OTHER THEORIES

Social inequality and its origins and mecha- nisms have long been central concerns of our discipline. We consider some alternative theoretical treatments of certain issues in the Ridgeway (1991) and the Ridgeway-Balkwell (1997) theories and our extensions of those theories, for potentially valuable cross fertilization of ideas. Related ideas ap- pear in the work of Collins and associates (Collins 1971; Collins et al. 1993), Turner (1984, 1992, 1995), and Jasso (1994b, 1997); we discuss some of them here.

This presentation is suggestive rather than definitive. A thorough analysis of theoretical relations requires an explicit statement of ini- tial conditions and theoretical assumptions, logical calculus for derivations, and coordi- nating definitions of concepts for comparing hypotheses. Such tasks are beyond our scope here,26 but we do describe some apparently related topics addressed by these theorists.

Collins analyzes gender from a power per- spective, viewing subordination as a conse- quence of either interpersonal violence (Collins 1971) or of political change that fa- vors some occupations at the expense of oth- ers (Collins 1991). Thus his theory links mi- cro-sociological behavior (interpersonal ad- vantages) to macro phenomena such as sub- sistence production and control of political institutions. For Collins, the primary moti- vating principle is power, not exchangeable resources or goal objects. He hardly speci- fies the mechanisms by which large-scale phenomena translate into interpersonal rela- tions. The primary consequence of gender stratification is economic position, though Collins also examines how disinterested ac- tors come to support universalistic (egalitar- ian) or particularistic (gender-differentiated) social systems. Collins et al. (1993) appar- ently rely on an implicit desire of individu- als to construct hierarchies (and to be at the top); gender is only one such hierarchy, and various other types (economic, erotic, repro- ductive) are somewhat interchangeable.

26 Comparing theories in great enough detail to be informative is an ambitious task. In this sec- tion we identify only some apparent relationships between status construction and other theories of inequality. For an example showing the steps whereby hypotheses are derived from three ver- sions of a single theory, see Berger et al. (1992). An excellent tool for tracing the growth of a theory is what Jasso (1993a, 1993b, 1993c) calls a "Merton Chart." To construct a Merton Chart, a theorist lists a theory's assumptions in columns and its derivations in rows. Cells then may be checked to show which assumptions function in each derivation, whether some assumptions bear most of the burden of the theory's explanatory power, and the like. Beside assessing the state of a theory at any given point, a Merton Chart could be used to encourage further development of a theory by adding observations that appear to deal with the same concepts as are contained in the theory, but for which the present set of assump- tions is insufficient explanation. Although this would be a worthwhile task for any developing theory, a Merton Chart including all the theories used here would be far too long to include-it would contain 26 columns, for instance, repre- senting all the unique assumptions of the theories cited-and hence it is best left for a future paper.

AMERICAN SOCIOLOGICAL REVIEW

Turner (1984, 1992, 1995) has constructed models of the development and perpetuation of several sorts of stratification systems; he views gender as one of the most important. Turner's models incorporate a great many variables, and thus consider factors and pro- cesses far more elaborate than those on which we focus, such as stratification of ma- terial wealth. Like Collins, Turner generally favors an economic interest mechanism for stratification systems; he regards actors as engaged in conflictual interaction at both the individual and the group level. In contrast, status theories presume cooperative status al- locations. In fact, one of the striking findings that these theories seek to explain is that in- dividuals encountering status disadvantages generally do not contest the status relations. At least interpersonally, people accept low- status positions-perhaps regretfully, but not combatively-when social conditions create them (Ridgeway and Walker 1995).

Jasso (1993a, 1993b, 1994a, 1994b, 1997) has consistently applied and extended a gen- eral theory of comparison processes, of which feelings of justice (who should get what) is the most visible consequence. This, however, is far from the only consequence of the processes in her theory; happiness, self- esteem, willingness to gamble, and other concretely distinct phenomena all are deriv- able from the theory of comparisons.

Jasso's theory distinguishes between quantity goods (also called "cardinal goods") and quality goods (also "ordinal"): The former, such as dollars of income or numbers of cattle owned, are additive and transferrable: the latter, such as beauty, wittiness, or ath- letic skill, are neither transferrable nor sus- ceptible to cardinal measurement. These types appear to be what we have called re- spectively "exchangeable resources" or "consummatory valued characteristics" and "status-valued characteristics." Jasso. in this theory, uses different functions for calculat- ing justice for the two types of goods, and certain justice parameters consequently dif- fer. One consequence that invites exploration in status construction theory isthat, in Jasso's theory, the upper extreme value of justice for a quality good-roughly, the maximum experienced over-reward-is limited and quite low. By contrast, the upper limit for a quantity good is infinite. Thus the wittiest or most beautiful person experiences only modest feelings of over-reward, while the richest person could feel over-rewarded much more strongly.

Status construction theory, like other sta- tus generalization theories, treats all status characteristics alike without attending to possible differences in their weights in pre- dicting their effects. That is certainly an over-simplification (Driskell and Mullen 1990; Foschi 1971). Probably it has not been addressed before because the theory, as stated, makes some very precise predictions that have been confirmed (see, e.g., those in Berger et al. 1992). Yet if the distinction in Jasso's theory applies also to status generali- zation theories, we might expect quality goods (or status-valued characteristics) to have less "strength" of effect than quantity goods. For instance, wealth might be a more powerful engine for creating-status value than beauty or sexual orientation. Although such parallels are worth exploring, the task must be approached cautiously, because race also is a quality good (purely status-valued), and we know that race exerts powerful ef- fects (e.g., Webster and Driskell 1978).

Another connection between status con- struction theory and Jasso's comparison theory is the difference in the explanations they offer for status differences associated with gender. As noted earlier, status con- struction theory, including all our proposed extensions, seeks to identify sufficient con-ditions-to describe a way in which gender or any characteristic could acquire status value; it does not claim that those conditions are the only possible ones-that is, that they are necessary. Given historical assumptions about the relative wealth of immigrant young men and native women in Europe during the agricultural revolution, Jasso (1994a, 1997)27 proves that wealth inequality is a necessary precondition for gender inequality (though, by itself, it is not sufficient). (This work con- stitutes elegant proof of a suggestion Rousseau made nearly 250 years ago.) Jasso's approach is consistent with certain elements of status construction theory: in

27 We thank Professor Jasso for providing an unpublished manuscript (1994a), which gives de- tails of this argument presently unavailable else- where.

particular, that gender inequality can arise and be perpetuated without deliberate at- tempts by advantaged individuals to oppress others, and also without the power differ- ences that are regarded as mechanisms in theories relying on a Marxist approach (such as Collins's [1971], outlined above). The re- lationships between Jasso's explanation for gender inequality and the approach of status construction theory provides a promising av- enue for further theoretical and empirical in- vestigation.
DISCUSSION

We now briefly consider some implications of the Ridgeway-Balkwell (1997) model of diffusion of status beliefs, including the rela- tive effectiveness of different mechanisms for generating status differences, the persistence and decay of status beliefs, and some practi- cal applications of this theoretical work.

Ridgeway and Balkwell (1997) present and analyze a model of the diffusion process by which individuals who already associate status value with a characteristic transmit that set of beliefs to others through interac- ti~n.~~

Our second extension provides a theo- retical mechanism for one part of the pro- cess, namely behavior control through be- havior interchange patterns (BIPs). The Ridgeway-Balkwell model also portrays pat- terns of spread of status beliefs that create consensus and dissensus regarding status characteristics as well as describing the de- cay of status value: how characteristics may lose their status connotations. Here we con- sider some implications of models of consen- sus, and some further ideas on gain and loss of status value.29

28 Other diffusion models, such as Carley's (1990) model of information diffusion through scientific communities, and Burt's (1980; Lin and Burt 1975) analyses of how network structure af- fects adoption of innovations, may address simi- lar issues. Many diffusion models, when applied to organizational behavior, presume self-interest on the part of actors acquiring information; this idea does not appear in the Ridgeway-Balkwell model. Assessing relationships between the Ridgeway-Balkwell model and other models, al- though a worthwhile task, is beyond the scope of this paper.

29 Brekhus (1996) names several characteristics

AMERICAN SOCIOLOGICAL REVIEW

On consensus, the Ridgeway-Balkwell model identifies several factors affecting the spread of status beliefs through a society. These include (1) the proportion of the popu- lation receiving high and low states of a goal object, GO+ and GO-; (2) the proportion of the population possessing each state of a nominal characteristic, NA and NB; (3) the structural correlations between GO+ and NA, and between GO-and NB;and (4) homophily biases regarding both GO and N. Where the correlation in Factor 3 is high, of course, sta- tus beliefs are likely to be acquired and to become consensual throughout the society. Skewing on goal objects or on N, by itself, does not have much effect, though diffusion is markedly impeded when the proportions are "reversed" (e.g., a 20-80 split on GO and an 80-20 split on N) (Ridgeway and Balkwell 1997, note 15). On the basis of these factors, we would expect race and gen- der to produce virtually unanimous agree- ment as to their status value. Factors 1, 2, and 3 promote consensus; only 4, homophily, retards it in the case of race.30

That may not be true, however, for beauty or sexual orientation. Although a correlation may well exist between these characteristics and states of goal objects such as esteem and friendship (Factor 3), neither characteristic is discerned so unambiguously as race and gen- der. Thus Factor 2 may not contribute strongly to status construction here. Also, the differential allocation of goal objects (Factor 1) may be less regular than for race and gen- der, and homophily (Factor 4) probably is stronger, at least for sexual orientation. Therefore the asymptotic consensus on beauty and orientation may be far from uni- versal. In keeping with this conjecture, Webster et al. (forthcoming) found different degrees of expectation formation based on sexual orientation: About one-third of their

(not all of them status characteristics) that have either gained or lost social significance. For in- stance, "premature ejaculators" and "spouse rap- ists" have been identified only recently, while "Catholic," "married," "miscegenation," "virgin- ity," and "living in sin" may be decreasing in sig- nificance.

30 Even a strong preference for same-race in- teraction cannot prevent mixed-race interaction in a heterogeneous society such as ours. See Ridgeway and Balkwell (1997, note 5).

respondents reported little status effect from that characteristic. The idea needs indepen- dent testing, however.

On decay of status value, the model notes two main ways in which decay may occur: mortality (death and emigration of individu- als holding the status beliefs) and conversion (individuals coming to realize that the per- formance connotations of the characteristic are not accurate). Authors of several empiri- cal studies (Ridgeway, Johnson, and Diekema 1994; Smith-Lovin and Brody 1991; Stewart 1988) have reported that col- lege-age women may no longer recognize gender as a status characteristic. It would be valuable to know whether the mechanism at work here is individual conversion or mor- tality. One way to study that question would be to compare individuals of different ages on their status beliefs associated with gender. An association of age with status beliefs (stronger belief among older respondents) would support a mortality mechanism; simi- larity across ages would be consistent with individual conversion, perhaps encouraged by changes in the culture, the legal system, and television depictions.

Many studies exploring how status affects interaction are motivated, at least in part, by a concern for reducing or overcoming unde- sirable effects of status generalization (e.g., see Cohen 1993; Entwisle and Webster 1974; Pugh and Wahrman 1983). In related fashion, we may ask, "What are the prospects that ex- isting status characteristics such as gender or race can lose status value?" More generally, we may ask what conditions cause character- istics to increase or decrease in status value, and whether the processes are similar for sta- tus gains and status losses. We might expect that exposure to inconsistent status charac- teristics and goal objects could diminish re- ward expectations based on those status char- acteristics. The sight of women receiving high salaries or enjoying status-valued ob- jects such as managerial positions can reduce advantages in performance expectations that otherwise might accrue to men in mixed-gen- der groups. At present, however, we do not know whether that process would erode the status value of the gender characteristic itself.

Some suggestions for intervention are im- plied above in our extensions. For instance, status construction theory and our first exten- sion suggest that deliberately controlling in- formation about the associations of goal ob- jects or exchangeable resources with charac- teristics might weaken an existing status link. Demonstrations of individuals or groups who possess the (currently) negatively valued state of some status characteristic but who are linked reliably to the high state of a goal ob- ject might overcome the association of per- formance expectations with states of that characteristic. Women earning high salaries, or unattractive people receiving esteem, when portrayed on television and in movies might be expected to work toward overcom- ing the status aspects of those respective char- acteristics. "Overcoming" here means chang- ing not only the differential estimations of social worth, but also the performance expec- tations associated with the characteristic. The visible presence of high-earning women might overcome ideas that women lack skills that men are thought to possess.

The second extension also may carry im- plications for overcoming status value. For instance, if BIPs are inconsistent, for some reason, with the relative evaluations of sta- tus characteristics (as in Figure 5), or if the behavior patterns are completely equal, the structuring power of a status characteristic might be weakened. Sometimes it is sug- gested that status generalization effects are less important in mixed-gender interaction among intimates than among strangers. If sentiments such as love or friendship tend to produce behavior patterns consistent with equality among intimates, that situation would help to overcome status generalization effects based on gender.

The third extension invites investigation to assess its applied usefulness. The first ques- tion might be, "How can this process be blocked?" That is, what actions can an actor take to avoid being cast in the violator role? Under what conditions do those actions pre- vent creation of a characteristic, or at least prevent attaching status evaluations to it?
CONCLUSIONS

In this paper we hope to contribute to the important work begun by Ridgeway and her colleagues, which first provided a micro- level theoretical explanation of a way in which status characteristics can be created. We have taken advantage of others' empiri- cal results and theoretical work, adapting them where possible to extend ideas of cre- ating status value. By employing findings and ideas from a wide range of investigators, we have been able to suggest a considerable widening of the range of characteristics and situations in which status construction can take place. Our extensions are consistent with existing evidence, and we have identi- fied leads for independent assessment.

We have offered three extensions to widen the scope of understanding the creation and maintenance of status beliefs. First, substi- tuting "goal objects" for "exchangeable re- sources" permits extension to status charac- teristics in addition to gender and race, par- ticularly to beauty and sexual orientation. Goal objects subsume exchangeable resources by including status-valued goods as well as goods possessing exchange or con- summatory value.

In our second extension, we describe how status characteristics can be created by be- havior interchanges even in the absence of goal objects (or exchangeable resources). Here we use ideas from several theoretical and empirical investigations of ways in which interaction processes can lead to formation of performance expectations. Behavior inter- changes, we believe, can create status char- acteristics from valued personal characteris- tics, from nominal characteristics, and even in the absence of any salient characteristic.

Our third extension integrates ideas on norm-violation processes with ideas on per- sonality attribution to describe a way in which new characteristics can be created. The process we describe creates valued personal characteristics. Our second extension de- scribes how those can become status charac- teristics through behavior processes.

From this work, one can predict to a far wider set of situations than considered until now in status construction theory, such as (1) creating status characteristics from purely status-valued goal objects such as expres- sions of esteem or honorific titles, (2) transforming either Ns or VPCs to status charac- teristics through behavior interchange pat- terns, and (3) creating VPCs from norm en- forcement and personality attribution pro- cesses. The next steps are to test these exten-

AMERICAN SOCIOLOGICAL REVIEW

sions, and then to apply the work for socially desirable ends.

Beyond testing the extensions, it would be worthwhile to develop further the theoretical foundations of this work. Works by Ridgeway and Balkwell, and by Jasso, are more highly developed quantitatively than are the new ideas we present here, and ben- efits will accrue from stating our ideas more precisely using mathematical tools. In addi- tion to permitting rigorous comparison of predictions from different theories, formal statement of our ideas will establish differ- ent and overlapping scope conditions. Per- haps most important, we will be able to adopt developments from other theories, such as Jasso's work on distributions of quantity and quality goods, to understanding how status characteristics are created.

A considerable program of research lies ahead. In acknowledging that fact, it is good to recall that the issues treated are signifi- cant, and the claims for theoretical extension are ambitious. Sociologists long have won- dered how status characteristics came to ex- ist and to persist in cultures. Status construc- tion theory first provided ways to think about those questions, and furnished some answers to them. We have offered ways to ask addi- tional theoretical questions and have pro- posed some new answers as well.

Murray Webster, Jr. is Professor of Sociology at the University of North Carolina, Charlotte. With John Skowetz and Joseph Whitmeyer (and sup- ported by the National Science Foundation), he is conducting research to extend a dynamic model of task group differentiation. With Joseph Whitmeyer he is developing a theory of second- order expectations and behavior, and with Guillermina Jasso he is testing and extending a theory of comparison processes and fair earnings

for women and men.

Stuart J. Hysom is a Ph.D. candidate in the De- partment of Sociology at Emory University. His current work investigates leadership differentia- tion in task-oriented groups, the status effects of sexual orientation and gender, and links between stratification systems at various levels of analysis.
REFERENCES

Balkwell, James W. 1991. "From Expectations to Behavior: A General Translation Function." American Sociological Review 56:355-69.

Becker, Howard S. 1973. Outsiders: Studies in the Sociology of Deviance. 2d ed. New York: Free Press.

Berger, Joseph. 1958. "Relations Between Perfor- mance, Rewards, and Action-Opportunities in Small Groups." Ph.D. dissertation, Department of Social Relations, Harvard University, Cam- bridge, MA. . 1988. "Directions in Expectation States Theory." Pp. 450-74 in Status Generalization: New Theory and Research, edited by M. Webster, Jr. and M. Foschi. Stanford, CA: Stanford University Press.

Berger, Joseph and Thomas L. Conner. 1969. "Performance Expectations and Behavior in Small Groups." Acta Sociologica 12: 186-98. . 1974. "Performance Expectations and Behavior in Small Groups: A Revised Formu- lation." Pp. 85-109 in Expectation States Theory: A Theoretical Research Program, edited by J. Berger, T. L. Conner, and M. Hamit Fisek. Cambridge, MA: Winthrop Publishers.

Berger, Joseph, M. Hamit Fisek, and Robert Z. Norman. 1989. "The Evolution of Status Ex- pectations: A Theoretical Extension." Pp. 100- 30 in Sociological Theories in Progress: New Formulations, edited by J. Berger and M. Zelditch, Jr. Newbury Park, CA: Sage.

Berger, Joseph, M. Hamit Fisek, Robert Z. Nor- man, and David G. Wagner. 1985. "Formation of Reward Expectations in Status Situations." Pp. 215-61 in Status, Rewards, and Influence: How Expectations Organize Behavior, edited by J. Berger and M. Zelditch, Jr. San Francisco, CA: Jossey-Bass.

Berger, Joseph, M. Hamit Fisek, Robert Z. Nor- man, and Morris Zelditch, Jr. 1977. Status Characteristics and Social Interaction: An Ex- pectation States Approach. New York: Elsevier.

Berger, Joseph, Robert Z. Norman, James Balk- well, and Roy F. Smith. 1992. "Status Incon- sistency in Task Situations: A Test of Four Sta- tus Processing Principles." American Socio- logical Review 57:843-55.

Berger, Joseph, David G. Wagner, and Morris Zelditch, Jr. 1985. "Introduction: Expectation States Theory: Review and Assessment." Pp. 1-72 in Status, Rewards, and Influence: How Expectations Organize Behavior, edited by J. Berger and M. Zelditch, Jr. San Francisco, CA: Jossey-Bass.

Berger, Joseph, Murray Webster, Jr., Cecilia L. Ridgeway, and Susan J. Rosenholtz. 1986. "Status Cues, Expectations, and Behavior." Pp. 1-22 in Advances in Group Processes, vol. 3, edited by E. J. Lawler. Greenwich, CT: JAI Press.

Biernat, Monica. 1995. "The Shifting Standards Model: Implications of Stereotype Accuracy for Social Judgment." Pp. 87-1 14 in Stereotype Accuracy: Toward Appreciating Group Differ- ences, edited by Y. Lee, L. J. Jussim, and C. R. McCauley. Washington, DC: American Psy- chological Association.

Bobrow, Sue Berryman. 1973. A Balance Theory of Distributive Justice and Experimental Tests of Consequences. Ph.D. dissertation, Depart- ment of Social Relations, Johns Hopkins Uni- versity, Baltimore, MD.

Brekhus, Wayne. 1996. "Social Marking and the Mental Coloring of Identity: Sexual Identity Construction and Maintenance in the United States." Sociological Forum 11 :497-522.

Brewer, Marilyn B. 1988. "A Dual Process Model of Impression Formation." Pp. 1-36 in Advances in Social Cognition, vol. 1, edited by T.

K. Srull and R. S. Wyer, Jr. Hillsdale, NJ: Erlbaum.

Burt, Ronald S. 1980. "Innovation as a Structural Interest: Rethinking the Impact of Network Po- sition on Innovation Adoption." Social Net- works 4:337-55.

Carley, Kathleen. 1990. "Structural Constraints on Communication: The Diffusion of the Ho- momorphic Signal Analysis Technique through Scientific Fields." Journal of Mathematical So- ciology 15:207-46.

Cohen, Elizabeth G. 1993. "From Theory to Prac- tice: The Development of an Applied Research Program." Pp. 385-415 in Theoretical Re- search Programs: Studies in the Growth of Theory, edited by J. Berger and M. Zelditch, Jr. Stanford, CA: Stanford University Press.

Cohen, Elizabeth G. and Lisa Catanzarite. 1988. "Can Expectations for Competence Be Altered in the Classroom?" Pp. 27-54 in Status Gener- alization: New Theory and Research, edited by

M. Webster, Jr. and M. Foschi. Stanford, CA: Stanford University Press.

Coleman, James S. 1986. "Social Theory, Social Research, and a Theory of Action." American Journal of Sociology 91: 1309-35.

Collins, Randall. 1971. "A Conflict Theory of Sexual Stratification." Social Problems 19:3

21.

. 1981. "On the Microfoundations of Macrosociology." American Journal of Sociol- ogy 86:984-1014.

. 1991. "Historical Change and the Ritual Production of Gender." Pp. 109-20 in Macro-Micro Linkages in Sociology, edited by J. Huber. Newbury Park, CA: Sage.

Collins, Randall, Janet Saltzman Chafetz, Rae Lesser Blumberg, Scott Coltrane, and Jonathan

H. Turner. 1993. "Toward an Integrated Theory of Gender Stratification." Sociological Per- spectives 36: 185-216.

Conner, Thomas L. 1977. "Performance Expecta- tions and the Initiation of Problem Solving At-

tempts." Journal of Mathematical Sociology 5:187-98.

. 1985. "Response Latencies, Performance Expectations, and Interaction Patterns." Pp. 189-2 14 in Status, Rewards, and Influence: How Expectations Organize Behavior, edited by Joseph Berger and Morris Zelditch, Jr. San Francisco, CA: Jossey-Bass.

Conway, Michael, M. Teresa Pizzamiglio, and Lauren Mount. 1996. "Status, Communality, and Agency: Implications for Stereotypes of Gender and Other Factors." Journal of Person- ality and Social Psychology 7 1 :25-38.

Cook, Karen S. 1975. "Expectations, Evaluations, and Equity." American Sociological Review 40:372-88. Dahrendorf, Ralf. 1959. Class and Class Conflict in Industrial Society. Stanford, CA: Stanford University Press.

Driskell, James E. 1982. "Personal characteristics and performance expectations." Social Psy- chology Quarterly 45:229-37.

Driskell, James E. and Brian T. Mullen. 1990. "Status, Expectations, and Behavior: A Meta- Analytic Review and Test of the Theory." Personality and Social Psychology Bulletin 16: 541-53.

Eagly, Alice H., Richard D. Ashmore, Mona G. Makhijani, and Laura C. Longo. 1991. "What Is Beautiful Is Good, but . . . : A Meta-Analytic Review of Research on the Physical Attractive- ness Stereotype." Psychological Bulletin 110: 109-28.

Entwisle, Doris R. and Murray Webster, Jr. 1974. "Expectations in Mixed Racial Groups." Sociology of Education 47:301-18.

Erikson, Kai T. 1966. Wayward Puritans: A Study in the Sociology of Deviance. New York: Wiley.

Eskilson, Arlene and Mary Glenn Wiley. 1976. "Sex Composition and Leadership in Small Groups." Sociometly 39: 183-94.

Fisek, M. Hamit, Joseph Berger, and Robert Z. Norman. 1991. "Participation in Heteroge- neous and Homogeneous Groups: A Theoreti- cal Integration." American Journal of Sociol- ogy 97:114-42.

Fisek, M. Hamit, Robert Z. Norman, and Max Nelson-Kilger. 1992. "Status Characteristics and Expectations States Theory: A Priori Model Parameters and Test." Journal of Math- ematical Sociology 16:285-303.

Fiske, Susan T. 1992. "Thinking Is for Doing: Portraits of Social Cognition from Daguerreo- type to Laserphoto." Journal of Personality and Social Psychology 63:877-89.

Fiske, Susan T. and Steven L. Neuberg. 1990. "A Continuum of Impression Formation, from Cat- egory-Based to Individuating Processes: Influ- ences of Information and Motivation on Atten- tion and Interpretation." Pp. 1-74 in Advances in Experimental Social Psychology, vol. 23, edited by M. P. Zanna. San Diego, CA: Aca- demic Press.

Fiske, Susan T. and Shelley E. Taylor. 1991. Social Cognition. 2d ed. New York: McGraw- Hill.

Foschi, Martha. 197 1. "Contradiction and Change of Performance Expectations." Canadian Re- view of Anthropology and Sociology 8:205-22. , 1996. "Double Standards in the Evalua- tion of Men and Women." Social Psychology Quarterly 59: 237-54.

Foschi, Martha, Larissa Lai, and Kirsten Sigerson. 1994. "Gender and Double Standards in the Assessment of Job Applicants." Social Psychology Quarterly 57:326-33.

Foschi, Martha, Kirsten Sigerson, and Marie Lembesis. 1995. "Assessing Job Applicants: the Relative Effects of Gender, Academic Record, and Decision Type." Small Group Re- search 26:328-52.

Foucault, Michel. 1978, 1985, 1986. The History of Sexuality. Translated by R. Hurley. New York: Pantheon. . 1983. "Afterword: The Subject and Power." Pp. 208-26 in Michel Foucault: Be- yond Structuralism and Hermeneutics, 2d ed., edited by H. L. Dreyfus and P. Rabinow. Chi- cago, IL: University of Chicago Press.

Freese, Lee and Bernard P. Cohen. 1973. "Elimi- nating Status Generalization." Sociometry 36: 177-93.

Garfinkel, Harold. 196 1. "Conditions of Success- ful Degradation Ceremonies." American Jour- nal of Sociology 61 :420-24.

Gerber, Gwendolyn L. 1996. "Status in Same- Gender and Mixed-Gender Police Dyads: Ef- fects on Personality Attributes." Social Psy- chology Quarterly 59:350-63.

Goffman, Erving. 1970. Strategic Interaction. Oxford, England: Blackwell. Harrod, Wendy. 1980. "Expectations from Un- equal Rewards." Social Psychology Quarterly

43: 126-30: Heider, Fritz. 1958. The Psychology of Interper- sonal Relations. New York: Wiley.

Jackson, Linda A,, John E. Hunter, and Carole N. Hodge. 1995. "Physical Attractiveness and In- tellectual competence: A Meta-Analytic Re- view." Social Psychology Quarterly 58:108

+q

LL.

Jasso, Guillermina. 1993a. "Choice and Emotion in Comparison Theory." Rationality and Soci- ety 5:231-74. . 1993b. "Analyzing Conflict Severity: Predictions of Distributive Justice Theory for the Two-Subgroup Case." Social Justice Re- search 6:357-82. . 1993c. "Building a Theory of Comparison Processes: Construction of Postulates and Deri-

AMERICAN SOCIOLOGICAL REVIEW

vation of Predictions." Pp. 213-64 in Theoretical Research Programs: Studies in the Grown of Theory, edited by J. Berger and M. Zelditch, Jr. Stanford, CA: Stanford University Press.

. 1994a. "Exploring the Origins of Gender Inequality." Unpublished manuscript available from the author, Department of Sociology, New York University, New York.

, 1994b. "Assessing Individual and Group Differences in the Sense of Justice: Framework and Application to Gender Differences in Judg- ments of the Justice of Earnings." Social Sci- ence Research 23:368-406.

. 1997. "The Common Mathematical Structure of Disparate Sociological Questions." Sociological Forum 12:37-5 1.

Johnson, Cathryn. 1995. "Sexual Orientation as a Diffuse Status Characteristic: Implications for Small Group Interaction." Pp. 115-37 in Advances in Group Processes, vol. 12, edited by

B. Markovsky, M. Lovaglia, and K. Heimer. Greenwich, CT: JAI Press.

Johnston, Janet. 1985. "How Personality At- tributes Structure Interpersonal Relations." Pp. 3 17-49 in Status, Rewards, and Influence: How Expectations Organize Behavior, edited by J. Berger and M. Zelditch, Jr. San Francisco, CA: Jossey-Bass.

. 1988. "The Structure of Ex-Spousal Re- lations: an Exercise in Theoretical Integration and Application." Pp. 309-26 in Status Gener- alization: New Theory and Research, edited by

M. Webster, Jr. and M. Foschi. Stanford, CA: Stanford University Press. Katz, Jonathan. 1995. The Invention of Hetero- sexuality. New York: Dutton.

Kemper, Theodore D. and Randall Collins. 1990. "Dimensions of Microinteraction." American Journal of Sociology 96:32-68.

Lawler, Edward J., Cecilia Ridgeway, and Barry Markovsky. 1993. "Structural Social Psychol- ogy and the Micro-Macro Problem." Sociological Theory 11:268-90.

Lemert, Edwin M. 1951. Social Pathology: A Sys- tematic Approach to the Study of Sociopathic Behavior. New York: McGraw-Hill.

Lin, Nan and Ronald S. Burt. 1975. "Differential Effects of Information Channels in the Process of Innovation Diffusion." Social Forces 54: 256-74.

Markovsky, Barry. 1988. "From Expectation States to Macro Processes." Pp. 35 1-65 in Status Generalization: New Theory and Research,

edited by M. Webster, Jr, and M. Foschi. Stanford, CA: Stanford University Press.

Markovsky, Barry, Roy F. Smith, and Joseph Berger. 1984. "Do Status Interventions Per- sist?" American Sociological Review 49:373

82. Moore, James C. 1968. "Status and Influence in Small Group Interactions." Sociometry 3 1 :47-

63.

. 1985. "Role Enactment and Self-Identity: An Expectation States Approach." Pp. 262-316 in Status, Rewards, and Influence: How Expec- tations Organize Behavior, edited by J. Berger and M. Zelditch, Jr. San Francisco, CA: Jossey- Bass.

Neuberg, Steven L. and Jason T. Newsom. 1993. "Personal Need for Structure: Individual Dif- ferences in the Desire for Simple Structure."

Journal of Personality and Social Psychology

65:113-31.

Parcel, Toby and Karen S. Cook. 1977. "Status Characteristics, Reward Allocation and Eq- uity." Sociometry 40:311-24.

Park, Robert E. 1928. "Bases of Race Prejudice." The Annals 140: 11-20.

Pugh, Meredith and Ralph Wahrman. 1983. "Neutralizing Sexism in Mixed-Sex Groups: Do Women Have to Be Better than Men?" American Journal of Sociology 88:746-62.

Ridgeway, Cecilia L. 1991. "The Social Con- struction of Status Value: Gender and Other Nominal Characteristics." Social Forces 70: 367-86. . 1997a. "Interaction and the Conservation of Gender Inequality: Considering Employ- ment." American Sociological Review 62:218

35.

, 1997b. "Where Do Status-Value Beliefs Come From? New Developments." Pp. 137-58 in Status, Network, and Structure, edited by J. Szmatka, J. Skvoretz, and J. Berger. Stanford, CA: Stanford University Press.

Ridgeway, Cecilia L. and James W. Balkwell. 1992. "Group Processes and the Diffusion of Status Beliefs." Presented at the annual meet- ing of the American Sociological Association, August 20-24, Pittsburgh, PA.

. 1997. "Group Processes and the Diffu- sion of Status-Beliefs." Social Psychology Quarterly 60: 14-3 1.

Ridgeway, Cecilia L., Elizabeth Heger Boyle, Kathy J. Kuipers, and Dawn Robinson. 1998. "How Do Status Beliefs Develop? The Role of Resources and Interactional Experience." American Sociological Review 63:33 1-78.

Ridgeway, Cecilia L. and Kristan Glasgow. 1996. "Acquiring Status Beliefs from Behavior in In- teraction." Paper presented at the annual meet- ing of the American Sociological Association, August 20, New York.

Ridgeway, Cecilia L., Cathryn Johnson, and David Diekema. 1994. "External Status, Legiti- macy, and Compliance in Male and Female Groups." Social Forces 72: 105 1-77.

Ridgeway, Cecilia L. and Henry A. Walker. 1995. "Status Structures." Pp. 281-310 in Social Psy- chology: Sociological Perspectives, edited by

K. S. Cook, G. A. Fine, and J. S. House. Needham Heights, MA: Allyn and Bacon. Scheff, Thomas J. 1966. Being Mentally Ill: A So- ciological Theory. Chicago, IL: Aldine.

, 1974. "The Labeling Theory of Mental Illness." American Sociological Review 39: 444-52.

Shelly, Robert K. and Murray Webster, Jr. 1997. "How Formal Status, Liking, and Ability Sta- tus Structure Interaction: Three Theoretical Principles and a Test." Sociological Perspec- tives 40:81-107.

Simmel, Georg. [I9081 1950. "The Secret and the Secret Society." Pp. 305-76 in The Sociology of Georg Simmel, translated by K. H. Wolff. Glencoe, IL: The Free Press.

Skvoretz, John and Thomas Fararo. 1996. "Status and Participation in Task Groups: A Dynamic Model." American Journal of Sociology 101: 1366-1414.

Smith-Lovin, Lynn and C. Brody. 1991. "Inter- ruptions in Group Discussions: The Effects of Gender and Group Composition." American Sociological Review 54:424-35.

Stewart, Penni. 1988. "Women and Men in Groups: A Status Characteristics Approach to Interaction." Pp. 69-85 in Status Generaliza- tion: New Theory and Research, edited by M. Webster, Jr, and M. Foschi. Stanford, CA: Stanford University Press.

Stewart, Penni and James C. Moore. 1992. "Wage Disparities and Performance Expectations." Social Psychology Quarterly 55:78-85.

Stulhofer, Aleksander and Hana Frencetic. 1996. "Sexual Orientation as a Diffuse Status Char- acteristic: A Bi-Cultural Comparison." Revija za Sociologiju 27: 193-204.

Talley, Jerry and Joseph Berger. 1983. "Social Control as a State Organizing Process." Unpublished manuscript available from Berger, Department of Sociology, Stanford University, Stanford, CA.

Tannenbaum, Frank. 1938. Crime and the Com- munity. New York: Ginn.

Turner, Jonathan H. 1984. Societal Stratification: A Theoretical Analysis. New York: Columbia University Press. . 1992. "The Production and Reproduction of Social Solidarity: A Synthesis of Two Ra- tional Choice Theories." Journal for the Theory of Social Behaviour 22:3 11-28.

. 1995. Macrodynamics: Toward a Theory on the Organization of Human Populations.

ASA Rose Book Series. New Brunswick, NJ: Rutgers University Press.

Veblen, Thorstein. [I8991 1953. The Theory of the Leisure Class: An Economic Study of Insti- tution. Rev. ed. Reprint, New York: New American Library.

Wagner, David G. 1988a. "Gender Inequalities in Groups: A Situational Approach." Pp. 55-68 in

Status Generalization: New Theory and Re- search, edited by M. Webster, Jr. and M. Foschi. Stanford, CA: Stanford University Press.

. 1988b. "Status Violations: Toward an Expectation States Theory of the Social Con- trol of Deviance." Pp. 261-74 in Status Gener- alization: New Theory and Research, edited by

M. Webster, Jr. and M. Foschi. Stanford, CA: Stanford University Press.

Wagner, David and Joseph Berger. 1993. "Status Characteristics Theory: The Growth of a Pro- gram." Pp. 23-63 in Theoretical Research Pro- grams: Studies in the Growth of Theory, edited by J. Berger and M. Zelditch, Jr. Stanford, CA: Stanford University Press.

Weber, Max. [I9221 1978. Economy and Society. Edited by G. Roth and C. Wittich. Reprint, Berkeley, CA: University of California Press.

Webster, Murray, Jr. 1984. "Social Structures and the Sense of Justice." Pp. 59-94 in Research in the Sociology of Organizations, vol. 3, edited by S. P. Bachrach and E. J. Lawler. Greenwich,

AMERICAN SOCIOLOGICAL REVIEW

CT: JAI Press.

Webster, Murray, Jr. and James E. Driskell. 1978. "Status Generalization: A Review and Some New Data." American Sociological Review 43:220-36.

, 1983. "Beauty as Status." American Journal of Sociology. 89: 140-65. Webster, Murray, Jr. and Martha Foschi. 1988.

Status Generalization: New Theory and Re- search. Stanford, CA: Stanford University Press.

Webster, Murray, Jr., Stuart J. Hysom, and Elise

M. Fullmer. Forthcoming. "Sexual Orientation and Occupation as Status." In Advances in Group Process, vol. 15, edited by J. Szmatka and J. Skvoretz. Greenwich, CT: JAI Press.

Willard, Don and Fred L. Strodtbeck. 1972. "La- tency of Verbal Response and Participation in Task Groups." Sociometry 35: 161-75.

Yuchtmann-Yaar, Ephraim and Moshe Semyonov. 1979. "Ethnic Inequality in Israeli Schools and Sports: An Expectation-States Ap- proach." American Journal of Sociology 85: 576-90.

Comments
  • Recommend Us