DOING GENDER

  CANDACE WEST

  DON H. ZIMMERMAN


  In the beginning, there was sex and there was gender. Those of us who
  taught courses in the area in the late 1960s and early 1970s were careful to
  distinguish one from the other. Sex, we told students, was what was
  ascribed by biology: anatomy, hormones, and physiology. Gender, we
  said, was an achieved status: that which is constructed through
  psychological, cultural, and social means. To introduce the difference
  between the two, we drew on singular case studies of hennaphrodites
  (Money 1968, 1974; Money and Ehrhardt 1972) and anthropological
  investigations of “strange and exotic tribes” (Mead 1963, 1968).

  Inevitably (and understandably), in the ensuing weeks of each term, our
  students became confused. Sex hardly seemed a “given” in the context of
  research that illustrated the sometimes ambiguous and often conflicting
  criteria for its ascription. And gender seemed much less an “achievement”
  in the context of the anthropological, psychological, and social
  imperatives we studied—the division of labor, the formation of gender
  identities, and the social subordination of women by men. Moreover, the
  received doctrine of gender socialization theories conveyed the strong
  message that while gender may be “achieved,” by about age five it was
  certainly fixed, unvarying, and static—much like sex.

  Since about 1975, the confusion has intensified and spread far beyond our
  individual classrooms. For one thing, we learned that the relationship
  between biological and cultural processes was far more complex—and
  reflexive—than we previously had supposed (Rossi 1984, especially pp.
  10-14). For another, we discovered that certain structural arrangements, for
  example, between work and family, actually produce or enable some
  capacities, such as to mother, that we formerly associated with biology
  (Chodorow 1978 versus Firestone 1970). In the midst of all this, the
  notion of gender as a recurring achievement somehow fell by the wayside.

  Our purpose in this chapter is to propose an ethnomethodologically
  informed, and therefore distinctively sociological, understanding of gender
  as a routine, methodical, and recurring accomplishment. We contend that
  the “doing” of gender is undertaken by women and men whose competence as members of society is hostage to its production.
  Doing gender involves a complex of socially guided perceptual,
  interactional, and micropolitical activities that cast particular pursuits as
  expressions of masculine and feminine “natures.”

  When we view gender as an accomplishment, an achieved property of
  situated conduct, our attention shifts from matters internal to the
  individual and focuses on interactional and, ultimately, institutional
  arenas. In one sense, of course, it is individuals who “do” gender. But it
  is a situated doing, carried out in the virtual or real presence of others who
  are presumed to be oriented to its production. Rather than as a property of
  individuals, we conceive of gender as an emergent feature of social
  situations: as both an outcome of and a rationale for various social
  arrangements and as a means of legitimating one of the most fundamental
  divisions of society.

  To advance our argument, we undertake a critical examination of what
  sociologists have meant by gender, including its treatment as a role
  enactment in the conventional sense and as a “display” in Goffman’s
  (1976) terminology. Both gender role and gender display focus on
  behavioral aspects of being a woman or a man (as opposed, for example,
  to biological differences between the two). However, we contend that the
  notion of gender as a role obscures the work that is involved in producing
  gender in everyday activities, whereas the notion of gender as a display
  relegates it to the periphery of interaction. We argue instead that
  participants in interactions organize their various and manifold activities
  to reflect or express gender, and they are disposed to perceive the behavior
  of others in a similar light.

  To elaborate our proposal, we suggest at the outset that important but
  often overlooked distinctions should be observed among sex, sex
  category, and gender. Sex is a determination made through the
  application of socially agreed upon biological criteria for classifying
  persons as females or males.’ The criteria for classification can be genitalia
  at birth or chromosomal typing before birth, and they do not necessarily
  agree with one another. Placement in a sex category is achieved through
  application of the sex criteria, but in everyday life, categorization is
  established and sustained by the socially required identificatory displays
  that proclaim one’s membership in one or the other category. In this
  sense, one’s sex category presumes one’s sex and stands as proxy for it in
  many situations, but sex and sex category can vary independently; that is,
  it is possible to claim membership in a sex category even when the sex
  criteria are lacking. Gender, in contrast, is the activity of managing
  situated conduct in light of normative conceptions of attitudes and
  activities appropriate for one’s sex category. Gender  activities emerge from and bolster claims to membership in a sex
  category.

  We contend that recognition of the analytical independence of sex, sex
  category, and gender is essential for understanding the relationships
  among these elements and the interactional work involved in “being” a
  gendered person in society. While our primary aim is theoretical, there
  will be occasion to discuss fruitful directions for empirical research that
  follow from the formulation of gender we propose.
  We begin with an assessment of the received meaning of gender,
  particularly in relation to the roots of this notion in presumed biological
  differences between women and men.
 
 

              PERSPECTIVES ON SEX AND GENDER


  In Western societies, the accepted cultural perspective on gender views
  women and men as naturally and unequivocally defined categories of
  being (Garfinkel 1967, pp. 116-8) with distinctive psychological and
  behavioral propensities that can be predicted from their reproductive
  functions. Competent adult members of these societies see differences
  between the two as fundamental and enduring, and these differences are
  seemingly supported by the division of labor into women’s and men’s
  work and an often elaborate differentiation of feminine and masculine
  attitudes and behaviors that are prominent features of social organization.
  Things are the way they are by virtue of the fact that men are men and
  women are women—a division perceived to be natural and rooted in
  biology, producing in turn profound psychological, behavioral, and social
  consequences. The structural arrangements of a society are presumed to be
  responsive to these differences.

  Analyses of sex and gender in the social sciences, although less likely to
  accept uncritically the naive biological determinism of the view just
  presented, often retain a conception of sex-linked behaviors and traits as
  essential properties of individuals (for good reviews, see Hochschild 1973;
  Thorne 1980; Tresemer 1975; Henley 1985). The “sex differences
  approach” (Thorne 1980) is more commonly attributed to psychologists
  than to sociologists, but the survey researcher who determines the gender
  of respondents on the basis of the sound of their voices over the telephone
  is also making trait-oriented assumptions. Reducing gender to a fixed set
  of psychological traits or to a unitary “variable” precludes serious
  consideration of the ways it is used to structure distinct domains of social
  experience (Stacey and Thorne 1985, pp. 307-8).

  Taking a different tack, role theory has attended to the social construction
  of gender categories, called “sex roles” or, more recently, “gender roles” and has analyzed how these are learned and enacted.

  Beginning with Linton (1936) and continuing through the works of
  Parsons (Parsons 1951; Parsons and Bales 1955) and Komarovsky
  (1946, 1950), role theory has emphasized the social and dynamic aspect
  of role construction and enactment (Connell 1983; Thorne 1980). But at
  the level of face-to-face interaction, the application of role theory to gender
  poses problems of its own (for good reviews and critiques, see Connell
  1983, 1985; Kessler, Ashendon, Connell, and Dowsett 1985; Lopata and
  Thorne, 1978; Stacey and Thorne, 1985; Thorne 1980). Roles are
  situated identities—assumed and relinquished as the situation
  demands—rather than master identities (Hughes 1945), such as sex
  category, that cut across situations. Unlike most roles, such as “nurse,”
  “doctor”, and “patient” or “professor” and “student”, gender has no specific
  site or organizational context. Moreover, many roles are already gender
  marked, so that special qualifiers—such as “female doctor” or “male
  nurse”—must be added to exceptions to the rule. Thorne (1980) observes
  that conceptualizing gender as a role makes it difficult to assess its
  influence on other roles and reduces its explanatory usefulness in
  discussions of power and inequality. Drawing on Rubin (1975), Thorne
  calls for a reconceptualization of women and men as distinct social
  groups, constituted in “concrete, historically changing— and generally
  unequal—social relationships” (Thorne 1980, p. 11).

  We argue that gender is not a set of traits, nor a variable, nor a role, but
  the product of social doings of some sort. What then is the social doing of
  gender? It is more than the continuous creation of the meaning of gender
  through human actions (Gerson and Peiss 1985). We claim that gender
  itself is constituted through interaction.2 To develop the implications of
  our claim, we turn to Goffman’s (1976) account of “gender display.” Our
  object here is to explore how gender might be exhibited or portrayed
  through interaction, and thus be seen as “natural,” while it is being
  produced as a socially organized achievement.
 
 

                    GENDER DISPLAY


  Goffman contends that when human beings interact with others in their
  environment, they assume that each possesses an “essential nature”—a
  nature that can be discerned through the “natural signs given off or
  expressed by them” (1976, p. 75). Femininity and masculinity are
  regarded as “prototypes of essential expression—something that can be
  conveyed fleetingly in any social situation and yet something that strikes
  at the most basic characterization of the individual” (1976, p. 75). The
  means through which we provide such expressions are “perfunctory, conventionalized acts” (1976, p. 69), which convey to others
  our regard for them, indicate our alignment in an encounter, and
  tentatively establish the terms of contact for that social situation. But they
  are also regarded as expressive behavior, testimony to our “essential
  natures.”

  Goffman (1976, pp. 69-70) sees displays as highly conventionalized
  behaviors structured as two-part exchanges of the statement-reply type, in
  which the presence or absence of symmetry can establish deference or
  dominance. These rituals are viewed as distinct from but articulated with
  more consequential activities, such as performing tasks or engaging in
  discourse. Hence, we have what he terms the “scheduling” of displays at
  junctures in activities, such as the beginning or end, to avoid interfering
  with the activities themselves. Goffman (1976, p. 69) formulates gender
  display as follows:

  If gender be defined as the culturally established correlates of sex (whether
  in consequence of biology or learning), then gender display refers to
  conventionalized portrayals of these correlates.

  These gendered expressions might reveal clues to the underlying,
  fundamental dimensions of the female and male, but they are, in
  Goffman’s view, optional performances. Masculine courtesies may or may
  not be offered and, if offered, may or may not be declined (1976, p. 71).
  Moreover, human beings “themselves employ the term ‘expression’, and
  conduct themselves to fit their own notions of expressivity” (1976, p.
  75). Gender depictions are less a consequence of our “essential sexual
  natures” than interactional portrayals of what we would like to convey
  about sexual natures, using conventionalized gestures. Our human nature
  gives us the ability to learn to produce and recognize masculine and
  feminine gender displays, “a capacity [we] have by virtue of being
  persons, not males and females” (1976, p. 76).

  Upon first inspection, it would appear that Goffman’s formulation offers
  an engaging sociological corrective to existing formulations of gender. In
  his view, gender is a socially scripted dramatization of the culture’s
  idealization of feminine and masculine natures, played for an audience that
  is well schooled in the presentational idiom. To continue the metaphor,
  there are scheduled performances presented in special locations, and like
  plays, they constitute introductions to or time out from more serious
  activities.

  There are fundamental equivocations in this perspective. By segregating
  gender display from the serious business of interaction, Goffman obscures
  the effects of gender on a wide range of human activities. Gender is not merely something that happens in the nooks and crannies of
  interaction, fitted in here and there and not interfering with the serious
  business of life. Although it is plausible to contend that gender
  displays—cons trued as conventionalized expressions—are optional, it
  does not seem plausible to say that we have the option of being seen by
  others as female or male.

  It is necessary to move beyond the notion of gender display to consider
  what is involved in doing gender as an ongoing activity embedded in
  everyday interaction. Toward this end, we return to the distinctions among
  sex, sex category, and gender introduced earlier.
 
 

              SEX, SEX CATEGORY, AND GENDER


  Garfinkel’s (1967, pp. 118-40) case study of Agnes, a transsexual raised as
  a boy who adopted a female identity at age 17 and underwent a sex
  reassignment operation several years later, demonstrates how gender is
  created through interaction and at the same time structures interaction.
  Agnes, whom Garfinkel characterized as a “practical methodologist,”
  developed a number of procedures for passing as a “normal, natural
  female” both prior to and after her surgery. She had the practical task of
  managing the facts that she possessed male genitalia and that she lacked
  the social resources a girl’s biography would presumably provide in
  everyday interaction. In short, she needed to display herself as a woman,
  simultaneously learning what it was to be a woman. Of necessity, this
  full-time pursuit took place at a time in her life when most people’s
  gender would be well-accredited and routinized. Agnes had to consciously
  contrive what the vast majority of women do without thinking. She was
  not faking what real women do naturally. She was obliged to analyze and
  figure out how to act within socially structured circumstances and
  conceptions of femininity that women born with appropriate biological
  credentials take for granted early on. As in the case of others who must
  “pass,” such as tranvestites, Kabuki actors, or Dustin Hoffman’s
  “Tootsie,” Agnes’s case makes visible what culture has made
  invisible—the accomplishment of gender.

  Garfinkel’s (1967) discussion of Agnes does not explicitly separate three
  analytically distinct, although empirically overlapping, concepts— sex,
  sex category, and gender.
 

  Sex

  Agnes did not possess the socially agreed upon biological criteria for
  classification as a member of the female sex. Still, Agnes regarded  herself as a female, albeit a female with a penis, which a woman ought not
  to possess. The penis, she insisted, was a “mistake” in need of remedy
  (Garfinkel 1967, pp. 126-7, 131-2). Like other competent members of our
  culture, Agnes honored the notion that there are essential biological
  criteria that unequivocally distinguish females from males. However, if we
  move away from the common-sense viewpoint, we discover that the
  reliability of these criteria is not beyond question (Money and Brennan
  1968; Money and Ehrhardt 1972; Money and Ogunro 1974; Money and
  Tucker 1975). Moreover, other cultures have acknowledged the existence
  of “cross-genders” (Blackwood 1984; Williams 1986) and the possibility
  of more than two sexes (Hill1935; Martin and Voorheis 1975, pp. 84-107; but see also Cucchiari ,1981, pp. 32-5).

  More central to our argument is Kessler and McKenna’s (1978, pp. 1-6)
  point that genitalia are conventionally hidden from public inspection in
  everyday life; yet we continue through our social rounds to “observe” a
  world of two naturally, normally sexed persons. It is the presumption
  that essential criteria exist, and would or should be there if looked for,
  that provides the basis for sex categorization. Drawing on Garfinkel,
  Kessler and McKenna argued that “female” and “male” are cultural
  events—products of what they term the “gender attribution
  process”—rather than some collection of traits, behaviors, or even physical
  attributes. Illustratively, they cite the child who, viewing a picture of
  someone clad in a suit and a tie, contends, “It’s a man, because he has a
  pee-pee” (Kessler and McKenna 1978, p. 154). Translation:
  “He must have a pee-pee [an essential characteristic] because I see the
  insignia of a suit and tie.” Neither initial sex assignment
  (pronouncement at birth as a female or male) nor the actual existence of
  essential criteria for that assignment (possession of a clitoris and vagina or
  penis and testicles) has much—if anything—to do with the identification
  of sex category in everyday life. There, Kessler and McKenna note, we
  operate with a moral certainty of a world of two sexes. We do not think,
  “Most persons with penises are men, but some may not be” or “Most
  persons who dress as men have penises.” Rather, we take it for granted
  that sex and sex category are congruent—that knowing the latter, we can
  deduce the rest.
 
 

  Sex Categorization

  Agnes’s claim to the categorical status of female, which she sustained by
  appropriate identificatory displays and other characteristics, could be
  discredited before her transsexual operation if her possession of a penis became known and after by her surgically constructed genitalia (see
  Raymond 1979, pp. 37, 138). In this regard, Agnes had to be continually
  alert to actual or potential threats to the security of her sex category. Her
  problem was not so much living up to some prototype of essential
  femininity but preserving her categorization as female. This task was
  made easy for her by a very powerful resource, namely, the process of
  common-sense categorization in everyday life.

  The categorization of members of society into indigenous categories, such
  as girl or boy, or woman or man, operates in a distinctively social way.
  The act of categorization does not involve a positive test, in the sense of a
  well-defined set of criteria that must be explicitly satisfied prior to
  making an identification. Rather, the application of membership
  categories relies on an “if-can” test in everyday interaction (Sacks 1972,
  pp. 332-35). This test stipulates that if people can be seen as members
  of relevant categories, then categorize them that way. That is, use the
  category that seems appropriate, except in the presence of discrepant
  information or obvious features that would rule out its use. This
  procedure is quite in keeping with the attitude of everyday life, in which
  we take appearances at face value unless we have special reason to doubt
  them (Bernstein 1986; Garfinkel 1967, pp. 272-7; Schutz 1943). It should
  be added that it is precisely when we have special reason to doubt
  appearances that the issue of applying rigorous criteria arises, but it is
  rare, outside legal or bureaucratic contexts, to encounter insistence on
  positive tests (Garfinkel 1967, pp. 262-83; Wilson 1970).

  Agnes’s initial resource was the predisposition of those she encountered to
  take her appearance (her figure, clothing, hair style, and so on) as the
  undoubted appearance of a normal female. Her further resource was our
  cultural perspective on the properties of “natural, normally sexed persons.”
  Garfinkel (1967, pp. 122-8) notes that in everyday life, we live in a world
  of two—and only two—sexes. This arrangement has a moral status in that
  we include ourselves and others in it as “essentially, originally, in the
  first place, always have been, always will be once and for all, in the final
  analysis, either ‘male’ or ‘female’ “ (Garfinkel 1967, p. 122).
  Consider the following case:

  This issue reminds me of a visit I made to a computer store a couple of
  years ago. The person who answered my questions was truly a
  salesperson. I could not categorize him/her as a woman or a man. What
  did I look for? (1) Facial hair: She/he was smooth skinned, but some men
  have little or no facial hair. (This varies by race, Native Americans and
  Blacks often have none.) (2) Breasts: She/he was wearing a loose shirt
  that hung from his/her  shoulders. And, as many women who suffered through a 1950s’ adolescence
  know to their shame, women are often flat-chested. (3) Shoulders:
  His/hers were small and round for a man, broad for a woman. (4) Hands:
  Long and slender fingers, knuckles a bit large for a woman, small for a man.
  (5) Voice: Middle range, unexpressive for a woman, not at all the
  exaggerated tones some gay males affect. (6) His/her treatment of me: Gave off
  no signs that would let me know if I were of the same or different sex as this
  person. There were not even any signs that he/she knew his/her sex would be
  difficult to categorize and I wondered about that even as I did my best to hide
  these questions so I would not embarrass him/her while we talked of computer
  paper. I left still not knowing the sex of my salesperson, and was disturbed by
  that unanswered question (child of my culture that I am). (Margolis 1985)

  What can this case tell us about situations such as Agnes’s (cf. Morris 1974;
  Richards 1983) or the process of sex categorization in general? First, we infer
  from this description that the computer salesclerk’s identificatory display was
  ambiguous, since she or he was not dressed or adorned in an unequivocally
  female or male fashion. It is when such a display fails to provide grounds for
  categorization that factors such as facial hair or tone of voice are assessed to
  determine membership in a sex category. Second, beyond the fact that this
  incident could be recalled after “a couple of years,” the customer was not only
  “disturbed” by the ambiguity of the salesclerk’s category but also assumed
  that to acknowledge this ambiguity would be embarrassing to the salesclerk.
  Not only do we want to know the sex category of those around us (to see it at a
  glance, perhaps), but we presume that others are displaying it for us in as
  decisive a fashion as they can.
 

  Gender


  Agnes attempted to be “120 percent female” (Garfinkel 1967, p. 129), that
  is, unquestionably in all ways and at all times feminine. She thought she
  could protect herself from disclosure before and after surgical intervention
  by comporting herself in a feminine manner, but she also could have
  given herself away by overdoing her performance. Sex categorization and
  the accomplishment of gender are not the same. Agnes’s categorization
  could be secure or suspect, but did not depend on whether or not she lived
  up to some ideal conception of femininity. Women can be seen as
  unfeminine, but that does not make them “unfemale.” Agnes faced an
  ongoing task of being a woman—something beyond style of dress (an
  identificatory display) or allowing men to light her cigarette (a gender
  display). Her problem was to produce  gender behavior.

  Agnes’s strategy of “secret apprenticeship,” through which she learned
  expected feminine decorum by carefully attending to her fiancé’s
  criticisms of other women, was one means of masking incompetencies
  and simultaneously acquiring the needed skills (Garfinkel 1967, pp.
  146-7). It was through her fiancé that Agnes learned that sunbathing on
  the lawn in front of her apartment was “offensive” (because it put her on
  display to other men). She also learned from his critiques of other women
  that she should not insist on having things her way and that she should
  not offer her opinions or claim equality with men (Garfinkel 1967, pp.
  147-8). (Like other women in our society Agnes learned something about
  power in the course of her “education.”)

  Popular culture abounds with books and magazines that compile idealized
  depictions of relations between women and men. Those focused on the
  etiquette of dating or prevailing standards of feminine comportment are
  meant to be of practical help in these matters. However, the use of any
  such source as a manual of procedure requires the assumption that doing
  gender merely involves making use of discrete, well-defined bundles of
  behavior that can simply be plugged into interactional situations to
  produce recognizable enactments of masculinity and femininity. The man
  “does” being masculine by, for example, taking the woman’s arm to guide
  her across a street, and she “does” being feminine by consenting to be
  guided and not initiating such behavior with a man.

  Agnes could perhaps have used such sources as manuals, but, we contend,
  doing gender is not so easily regimented (Mithers 1982; Morris 1974).
  Such sources may list and describe the sorts of behaviors that mark or
  display gender, but they are necessarily incomplete (Garfinkel 1967, pp.
  66-75; Wieder 1974, pp. 183-214; Zimmerman and Wieder 1970, pp.
  285-98). To be successful, marking or displaying gender must be finely
  fitted to situations and modified or transformed as the occasion demands.
  Doing gender consists of managing such occasions so that, whatever the
  particulars, the outcome is seen and seeable in context as
  gender-appropriate or purposefully genderinappropriate, that is,
  accountable.
 
 

               GENDER AND ACCOUNTABILITY


  As Heritage (1984, pp. 136-7) notes, members of society regularly engage
  in “descriptive accountings of states of affairs to one another,” and such
  accounts are both serious and consequential. These  descriptions name, characterize, formulate, explain, excuse, excoriate, or
  merely take notice of some circumstance or activity and thus place it
  within some social framework (locating it relative to other activities, like
  and unlike).

  Such descriptions are themselves accountable, and societal members orient
  to the fact that their activities are subject to comment. Actions are often
  designed with an eye to their accountability, that is, how they might look
  and how they might be characterized. The notion of accountability also
  encompasses those actions undertaken so that they are specifically
  unremarkable and thus not worthy of more than a passing remark, because
  they are seen to be in accord with culturally approved standards.
  Heritage (1984, p. 179) observes that the process of rendering something
  accountable is interactional in character:

 
  [This] permits actors to design their actions in relation to their
  circumstances so as to permit others, by methodically taking account of
  circumstances, to recognize the action for what it is.

  The key word here is circumstances. One circumstance that attends
  virtually all actions is the sex category of the actor. As Garfinkel (1967,
  p. 118) comments:

  [T]he work and socially structured occasions of sexual passing were
  obstinately unyielding to [Agnes’s] attempts to routinize the grounds of
  daily activities. This obstinacy points to the omnirelevance of sexual
  status to affairs of daily life as an invariant but unnoticed background in
  the texture of relevances that compose the changing actual scenes of
  everyday life. (emphasis added)


  If sex category is omnirelevant (or even approaches being so), then a
  person engaged in virtually any activity may be held accountable for
  performance of that activity as a woman or a man, and their incumbency
  in one or the other sex category can be used to legitimate or discredit their
  other activities (Berger, Cohen, and Zelditch 1972; Berger, Conner, and
  Fisek 1974; Berger, Fisek, Norman, and Zelditch 1977; Humphreys and
  Berger 1981). Accordingly, virtually any activity can be assessed as to its
  womanly or manly nature. And note, to “do” gender is not always to live
  up to normative conceptions of femininity or masculinity; it is to engage
  in behavior at the risk of gender assessment. Although it is individuals
  who do gender, the enterprise is fundamentally interactional and
  institutional in character, because accountability  s a feature of social relationships and its idiom is drawn from the
  institutional arena in which those relationships are enacted. If this is the
  case, can we ever not do gender? Insofar as a society is partitioned by
  “essential” differences between women and men and placement in a sex
  category is both relevant and enforced, doing gender is unavoidable.
 
 

              RESOURCES FOR DOING GENDER


  Doing gender means creating differences between girls and boys and
  women and men, differences that are not natural, essential, or biological.
  Once the differences have been constructed, they are used to reinforce the
  “essentialness” of gender. In a delightful account of the “arrangement
  between the sexes,” Goffman (1977) observes the creation of a variety of
  institutionalized frameworks through which our “natural, normal
  sexedness” can be enacted. The physical features of social settings provide
  one obvious resource for the expression of our “essential” differences. For
  example, the sex segregation of North American public bathrooms
  distinguishes “ladies” from “gentlemen” in matters held to be
  fundamentally biological, even though both “are somewhat similar in the
  question of waste products and their elimination” (Goffman 1977, p.
  315). These settings are furnished with dimorphic equipment (such as
  urinals for men or elaborate grooming facilities for women), even though
  both sexes may achieve the same ends through the same means (and
  apparently do so in the privacy of their own homes). To be stressed here is
  the fact that:

  The functioning of sex-differentiated organs is involved, but there is
  nothing in this functioning that biologically recommends segregation;
  that arrangement is a totally cultural matter. . . toilet segregation is
  presented as a natural consequence of the difference between the sex-classes
  when in fact it is a means of honoring, if not producing, this difference.
  (Goffman
  1977, p. 316)

  Standardized social occasions also provide stages for evocations of the
  “essential female and male natures.” Goffman cites organized sports as one
  such institutionalized framework for the expression of manliness. There,
  those qualities that ought “properly” to be associated with masculinity,
  such as endurance, strength, and competitive spirit, are celebrated by all
  parties concerned—participants, who may be seen to demonstrate such
  traits, and spectators, who applaud their demonstrations from the safety of
  the sidelines (1977, p. 322).

Assortative mating practices among heterosexual couples afford still
  further means to create and maintain differences between women and men.
  For example, even though size, strength, and age tend to be normally
  distributed among females and males (with considerable overlap between
  them), selective pairing ensures couples in which boys and men are
  visibly bigger, stronger, and older (if not “wiser”) than the girls and
  women with whom they are paired. So, should situations emerge in which
  greater size, strength, or experience is called for, boys and men will be
  ever ready to display it and girls and women to appreciate its display
  (Goffman 1977, p. 321; West and Iritani 1985).

  Gender may be routinely fashioned in a variety of situations that seem
  conventionally expressive to begin with, such as those that present
  “helpless” women next to heavy objects or flat tires. But, as Goffman
  notes, heavy, messy, and precarious concerns can be constructed from any
  social situation, “even though by standards set in other settings, this may
  involve something that is light, clean, and safe” (Goffman 1977, p. 324).
  Given these resources, it is clear that any interactional situation sets the
  stage for depictions of “essential” sexual natures. In sum, these situations
  “do not so much allow for the expression of natural differences as for the
  production of that difference itself” (Goffman 1977, p. 324).

  Many situations are not clearly sex categorized, nor is what transpires in
  them obviously gender relevant. Yet any social encounter can be pressed
  into service in the interests of doing gender. Thus, Fishman’s (1978)
  research on casual conversations found an asymmetrical “division of labor”
  in talk between heterosexual intimates. Women had to ask more
  questions, fill more silences, and use more attention-getting beginnings in
  order to be heard. Her conclusions are particularly pertinent here:

  Since interactional work is related to what constitutes being a woman,
  with what a woman is, the idea that it is work is obscured. The work is
  not seen as what women do, but as part of what they are. (Fishman 1978,
  p. 405)

  We would argue that it is precisely such labor that helps to constitute the
  essential nature of women as women in interactional contexts (West and
  Zimmerman 1983, pp. 109-11; but see also Kollock, Blumstein, and
  Schwartz 1985).

  Individuals have many social identities that may be donned or shed,
  muted, or made more salient, depending on the situation. One may be a
  friend, professional, citizen, and many other things to many different
  people or to the same person at different times. But we are always  women or men—unless we shift into another sex category. What this
  means is that our identificatory displays will provide an ever- available
  resource for doing gender under an infinitely diverse set of circumstances.
  Some occasions are organized to routinely display and celebrate behaviors
  that are conventionally linked to one or the other sex category. On such
  occasions, everyone knows his or her place in the interactional scheme of
  things. If an individual identified as a member of one sex category
  engages in behavior usually associated with the other category, this
  routinization is challenged. Hughes (1945, p. 356) provides an illustration
  of such a dilemma:
 

  [A] young woman . . . became part of that virile profession, engineering.

  The designer of an airplane is expected to go up on the maiden flight of
  the first plane built according to the design. He [sic] then gives a dinner to
  the engineers and workmen who worked on the new plane. The dinner is
  naturally a stag party. The young woman in question designed a plane.
  Her co-workers urged her not to take the risk—for which, presumably,
  men only are fit—of the maiden voyage. They were, in effect, asking her
  to be a lady instead of an engineer. She chose to be an engineer. She then
  gave the party and paid for it like a man. After food and the first round of
  toasts, she left like a lady.


  On this occasion, the parties reached an accommodation that allowed a
  woman to engage in presumptively masculine behaviors. However, in the
  end, this compromise permitted demonstration of her “essential”
  femininity, through accountably “ladylike” behavior.

  Hughes (1945, p. 357) suggests that such contradictions may be
  countered by managing interactions on a very narrow basis, for example,
  by “keeping the reiationship formal and specific.” But the heart of the
  matter is that even—perhaps, especially—if the relationship is a formal
  one, gender is still something one is accountable for. Thus, a woman
  physician (notice the special qualifier in her case) may be accorded respect
  for her skill and even addressed by an appropriate title. Nonetheless, she is
  subject to evaluation in terms of normative conceptions of appropriate
  attitudes and activities for her sex category and under pressure to prove
  that she is an “essentially” feminine being, despite appearances to the
  contrary (West 1984, pp. 97-101). Her sex category is used to discredit her
  participation in important clinical activities (Lorber 1984, pp. 52-4), while
  her involvement in medicine is used to discredit her commitment to her
  responsibilities as a wife and mother (Bourne and Wikier 1978, pp.
  435-7). Simultaneously, her  exclusion from the physician colleague community is maintained and her
  accountability as a woman is ensured.

  In this context, “role conflict” can be viewed as a dynamic aspect of our
  current “arrangement between the sexes” (Goffman 1977), an arrangement
  that provides for occasions on which persons of a particular sex category
  can “see” quite clearly that they are out of place and that if they were not
  there, their current troubles would not exist. From the standpoint of
  interaction, what is at stake is the management of our “essential” natures
  and, from the standpoint of the individual, the continuing
  accomplishment of gender. If, as we have argued, sex category is
  omnirelevant, then any occasion, conflicted or not, offers the resources for
  doing gender.

  We have sought to show that sex category and gender are managed
  properties of conduct that are contrived with respect to the fact that others
  will judge and respond to us in particular ways. We have claimed that a
  person’s gender is not simply an aspect of what one is, but, more
  fundamentally, it is something that one does, and does recurrently, in
  interaction with others.
  What are the consequences of this theoretical formulation? If, for example,
  individuals strive to achieve gender in encounters with others, how does a
  culture instill the need to achieve it? What is the relationship between the
  production of gender at the level of interaction and such institutional
  arrangements as the division of labor in society? And, perhaps most
  important, how does doing gender contribute to the subordination of
  women by men?
 
 

                   RESEARCH AGENDAS


  To bring the social production of gender under empirical scrutiny, we
  might begin at the beginning, with a reconsideration of the process
  through which societal members acquire the requisite categorical apparatus
  and other skills to become gendered human beings.
 

  Recruitment to Gender Identities


  The conventional approach to the process of becoming girls and boys has
  been sex-role socialization. In recent years, recurring problems arising
  from this approach have been linked to inadequacies inherent in role
  theory per Se: its emphasis on “consensus, stability and continuity”
  (Stacey and Thorne 1985, p. 307), its ahistorical and depoliticizing focus
  (Stacey and Thorne 1985, p. 307; Thorne 1980, p. 9), and the fact  that its “social” dimension relies on “a general assumption that people choose
  to maintain existing customs” (Connell 1985, p. 263).

  In contrast, Cahill (1982, 1986a, 1986b) analyzes the experiences of preschool
  children using a social model of recruitment into normally gendered
  identities. Cahill argues that categorization practices are fundamental to
  learning and displaying feminine and masculine behavior. Initially, he
  observes, children are primarily concerned with distinguishing between
  themselves and others on the basis of social competence. Categorically, their
  concern resolves itself into the opposition of “girl/boy” versus “baby”
  classification (the latter designating children whose social behavior is
  problematic and who must be closely supervised). It is children’s concern with
  being seen as socially competent that evokes their initial claims to gender
  identities:

  During the exploratory stage of children’s socialization. . . they learn that only
  two social identities are routinely available to them, the identity of “baby,” or,
  depending on the configuration of their external genitalia, either “big boy” or
  “big girl.” Moreover, others subtly inform them that the identity of “baby” is a
  discrediting one. When, for example, children engage in disapproved
  behavior, they are often told “You’re a baby” or “Be a big boy.” In effect, these
  typical verbal responses to young children’s behavior convey to them that
  they must behaviorally choose between the discrediting identity of “baby”
  and their anatomically determined sex identity. (Cahill l986a, p. 175)

  Subsequently, little boys appropriate the gender ideal of “efficaciousness,”
  that is, being able to affect the physical and social environment through the
  exercise of physical strength or appropriate skills. In contrast, little girls learn
  to value “appearance,” that is, managing themselves as ornamental objects.
  Both classes of children learn that the recognition and use of sex
  categorization in interaction are not optional, but mandatory (see also Bem
  1983).

  Being a “girl” or a “boy,” then, is not only being more competent than a
  “baby,” but also being competently female or male, that is, learning to produce
  behavioral displays of one’s “essential” female or male identity. In this
  respect, the task of four- to five-year-old children is very similar to Agnes’s:

  For example, the following interaction occurred on a preschool playground. A
  55-month-old boy (D) was attempting to unfasten the clasp of a necklace when
  a preschool aide walked over to him.

  A: Do you want to put that on?
  D:   No. It’s for girls.
  A: You don’t have to be a girl to wear things around your neck. Kings wear things around their necks. You could pretend you’re a king.
  D: I’m not a king. I’m a boy. (Cahill 1986a, p. 176)


  As Cahill notes of this example, although D may have been unclear as to the
  sex status of a king’s identity, he was obviously aware that necklaces are
  used to announce the identity “girl.” Having claimed the identity “boy” and
  having developed a behavioral commitment to it, he was leery of any
  display that might furnish grounds for questioning his claim.
  In this way, new members of society come to be involved in a
  self-regulating process as they begin to monitor their own and others’
  conduct with regard to its gender implications. The recruitment process
  involves not only the appropriation of gender ideals (by the valuation of
  those ideals as proper ways of being and behaving) but also gender
  identities that are important to individuals and that they strive to maintain.
  Thus gender differences, or the sociocultural shaping of “essential female and
  male natures,” achieve the status of objective facts. They are rendered
  normal, natural features of persons and provide the tacit rationale for
  differing fates of women and men within the social order.

  Additional studies of children’s play activities as routine occasions for the
  expression of gender-appropriate behavior can yield new insights into how
  our “essential natures” are constructed. In particular, the transition from what
  Cahill (1986a) terms “apprentice participation” in the sex-segregated world,
  which is common among elementary school children, to “bona fide
  participation” in the heterosocial world, which is so frightening to
  adolescents, is likely to be a keystone in our understanding of the
  recruitment process (Thorne 1986; Thorne and Luria 1986).
 

  Gender and the Division of Labor


  Whenever people face issues of allocation—who is to do what, get what,
  plan or execute action, direct or be directed, incumbency in significant social
  categories such as “female” and “male” seems to become pointedly relevant.
  How such issues are resolved conditions the exhibition, dramatization, or
  celebration of one’s ~‘essential nature” as a woman or man.
  Fenstermarker Berk (1985) offers an elegant demonstration of this point in
  her investigation of the allocation of household labor and the attitudes of
  married couples toward the division of household tasks.  Berk found little variation in either the actual distribution of tasks or
  perceptions of equity in regard to that distribution. Wives, even when
  employed outside the home, do the vast majority of household and
  child-care tasks. Moreover, both wives and husbands tend to perceive this
  as a “fair” arrangement. Noting the failure of conventional sociological and
  economic theories to explain this seeming contradiction, Berk contends
  that something more complex than rational arrangements for the
  production of household goods and services is involved:

  Hardly a question simply of who has more time, or whose time is worth
  more, who has more skill or more power, it is clear that a complicated
  relationship between the structure of work imperative and the structure of
  normative expectations attached to work as gendered determines the
  ultimate allocation of members’ time to work and home. (Berk 1985, pp.
  195-6)

  She notes, for example, that the most important factor influencing wives’
  contribution of labor is the total amount of work demanded or expected by
  the household; such demands had no bearing on husbands’ contributions.
  Wives reported various rationales (their own and their husbands’) that
  justified their level of contribution and, as a general matter, underscored
  the presumption that wives are essentially responsible for household
  production.

  Fenstermarker Berk (1985, p. 201) contends that it is difficult to see how
  people “could rationally establish the arrangements that they do solely for
  the production of household goods and services”—much less how people
  could consider them “fair.” She argues that our current arrangements for
  the domestic division of labor support two production processes:
  household goods and services (meals, clean children, and so on) and, at
  the same time, gender. As she puts it:

  Simultaneously, members “do” gender, as they “do” housework and child
  care, and what [has] been called the division of labor provides for the joint
  production of household labor and gender; it is the mechanism by which
  both the material and symbolic products of the household are realized.
  (1985, p. 201)

  It is not simply that household labor is designated as “women’s work,”
  but that for a woman to engage in it and a man not to engage in it is to
  draw on and exhibit the “essential nature” of each. What is produced and
  reproduced is not merely the activity and artifact of domestic life, but the
  material embodiment of wifely and husbandly roles and, denyatively, of
  womanly and manly conduct (see Beer 1983, pp. 70-89). What are also frequently produced and reproduced are the dominant and
  subordinate statuses of the sex categories.
  How does gender get done in work settings outside the home, where
  dominance and subordination are themes of overarching importance?
  Hochschild’s (1983) analysis of the work of flight attendants offers some
  promising insights. She found that the occupation of flight attendant
  consisted of something altogether different for women than for men:

  As the company’s main shock absorbers against “mishandled” passengers,
  their own feelings are more frequently subjected to rough treatment. In
  addition, a day’s exposure to people who resist authority in a woman is a
  different experience than it is for a man. . . . In this respect, it is a disadvantage
  to be a woman. And in this case, they are not simply women in the biological
  sense. They are also a highly visible distillation of middle-class American
  notions of femininity. They symbolize Woman. Insofar as the category
  “female” is mentally associated with having less status and authority, female
  flight attendants are more readily classified as “really” females than other
  females are. (Hochschild 1983, p. 175)

  In performing what Hochschild terms the “emotional labor” necessary to
  maintain airline profits, women flight attendants simultaneously produce
  enactments of their “essential” femininity.
 

  Sex and Sexuality


  What is the relationship between doing gender and a culture’s prescription of
  “obligatory heterosexuality” (Rich 1980; Rubin 1975)? As Frye (1983, p.
  22) observes, the monitoring of sexual feelings in relation to other
  appropriately sexed persons requires the ready recognition of such persons
  “before one can allow one’s heart to beat or one’s blood to flow in erotic
  enjoyment of that person.” The appearance of heterosexuality is produced
  through emphatic and unambiguous indicators of one’s sex, layered on in ever
  more conclusive fashion (Frye 1983, p. 24). Thus, lesbians and gay men
  concerned with passing as heterosexuals can rely on these indicators for
  camouflage; in contrast, those who would avoid the assumption of
  heterosexuality may foster ambiguous indicators of their categorical status
  through their dress, behaviors, and style. But “ambiguous” sex indicators are
  sex indicators nonetheless. If one wishes to be recognized as a lesbian (or
  heterosexual woman), one must first establish a categorical status as female.
  Even as popular images portray lesbians as “females who are not feminine”  (Frye 1983, P. 129), the accountability of persons for their “normal,
  natural sexedness” is preserved.

  Nor is accountability threatened by the existence of sex-change
  operations—presumably, the most radical challenge to our cultural
  perspective on sex and gender. Although no one coerces transsexuals into
  hormone therapy, electrolysis, or surgery, the alternatives available to
  them are undeniably constrained:

  When the transsexual experts maintain that they use transsexual
  procedures only with people who ask for them, and who prove that they
  can “pass,” they obscure the social reality. Given patriarchy’s prescription
  that one must be either masculine or feminine, free choice is conditioned.
  (Raymond 1979, p. 135, italics added)

  The physical reconstruction of sex criteria pays ultimate tribute to the
  “essentialness” of our sexual natures—as women or as men.
 
 

            GENDER, POWER, AND SOCIAL CHANGE


  Let us return to the question: Can we avoid doing gender? Earlier, we
  proposed that, insofar as sex category is used as a fundamental criterion
  for differentiation, doing gender is unavoidable. It is unavoidable because
  of the social consequences of sex-category membership: the allocation of
  power and resources not only in the domestic, economic, and political
  domains but also in the broad arena of interpersonal relations. In virtually
  any situation, one’s sex category can be relevant, and one’s performance as
  an incumbent of that category (i.e., gender) can be subjected to evaluation.
  Maintaining such pervasive and faithful assignment of lifetime status
  requires legitimation.

  But doing gender also renders the social arrangements based on sex
  category accountable as normal and natural, that is, legitimate ways of
  organizing social life. Differences between women and men that are
  created by this process can then be portrayed as fundamental and enduring
  dispositions. In this light, the institutional arrangements of a society can
  be seen as responsive to the differences, the social order being merely an
  accommodation to the natural order. Thus if, in doing gender, men are
  also doing dominance and women are doing deference (cf. Goffman 1967,
  pp. 47-95), the resultant social order, which supposedly reflects “natural
  differences,” is a powerful reinforcer and legitimator of hierarchical
  arrangements.

Frye observes:  For efficient subordination, what’s wanted is that the structure not appear
  to be a cultural artifact kept in place by human decision or custom, but
  that it appear natural—that it appear to be quite a direct consequence of
  facts about the beast which are beyond the scope of human manipulation.
  That we are trained to behave so differently as women and men, and to
  behave so differently toward women and men, itself contributes mightily
  to the appearance of extreme dimorphism, but also, the ways we act as
  women and men, and the ways we act toward women and men, mold our
  bodies and our minds to the shape of subordination and dominance. We
  do become what we practice being. (Frye 1983, p. 34)

  If we do gender appropriately, we simultaneously sustain, reproduce, and
  render legitimate the institutional arrangements that are based on sex
  category. If we fail to do gender appropriately, we as individuals— not the
  institutional arrangements—may be called to account (for our character,
  motives, and predispositions).

  Social movements such as femininism can provide the ideology and
  impetus to question existing arrangements and the social support for
  individuals to explore alternatives to them. Legislative changes, such as
  those proposed by the Equal Rights Amendment, can also weaken the
  accountability of conduct to sex category, thereby affording the possibility
  of more widespread loosening of accountability in general. To be sure,
  equality under the law does not guarantee equality in other arenas. As
  Lorber points out, assurance of “scrupulous equality of categories of
  people considered essentially different needs constant monitoring.” What
  such proposed changes can do is provide the warrant for asking why, if
  we wish to treat women and men as equals, there needs to be two sex
  categories at all (see Chapter 18).

  The sex category/gender relationship links the institutional and
  interactional levels, a coupling that legitimates social arrangements based
  on sex category and reproduces their asymmetry in face-to-face interaction.
  Doing gender furnishes the interactional scaffolding of social structure,
  along with a built-in mechanism of social control. In appreciating the
  institutional forces that maintain distinctions between women and men,
  we must not lose sight of the interactional validation of those distinctions
  that confers upon them their sense of “naturalness” and “rightness.”
  Social change, then, must be pursued at the institutional and cultural
  levels of sex category and at the interactional level of gender. Such a
  conclusion is hardly novel. Nevertheless, we suggest that it is important
  to recognize that the analytical distinction between institutional and
  interactional sphere does not pose an either/or choice when it comes to  he question of effecting social change.

Reconceptualizing gender not as a
  simple property of individuals but as an integral dynamic of social orders
  implies a new perspective on the entire network of gender relations: the social subordination of women, and the cultural practices which help
  sustain it; the politics of sexual object-choice, and particularly the
  oppression of homosexual people; the sexual division of labor, the formation
  of character and motive, so far as they are organized as femininity and
  masculinity; the role of the body in social relations, especially the politics of
  childbirth; and the nature of strategies of sexual liberation movements.
  (Connell 1985, p. 261)

  Gender is a powerful ideological device, which produces, reproduces, and
  legitimates the choices and limits that are predicated on sex category. An
  understanding of how gender is produced in social situations will afford
  clarification of the interactional scaffolding of social structure and the social
  control processes that sustain it.
 
 

                         NOTES

  1.
    This definition understates many complexities involved in the relationship between biology and culture (Jaggar 1983, pp. 106-13). However, our point is that the determination of an individual’s sex classification is a social process through and through.
  2.  This is not to say that gender is a singular “thing,” omnipresent in the same form historically or in every situation. Because normative conceptions of appropriate attitudes and activities for sex categories can vary across cultures and historical moments, the management of situated conduct in light of those expectations can take many different forms.
  3.  Bernstein (1986) reports an unusual case of espionage in which a man passing as a  woman convinced a lover that he/she had given birth to “their” child, who, the lover, thought, “looked like” him.
 
 

                       REFERENCES


  Beer, W. R. 1983. Househusbands: Men and Housework in American Families. New York: Praeger.
   Bem, S. L. 1983. “Gender Schema Theory and Its Implications for Child Development:
  Raising Gender-Aschematic Children in a Gender-Schematic Society.” Signs: Journal of Women in Culture and Society 8:598-616.

  Berger, J., B. P. Cohen, and M. Zelditch, Jr. 1972. “Status Characteristics and Social
  Interaction.” American Sociological Review 37:24 1-55.

Berger, J., T. L. Conner, and M. Hamit Fisek, eds. 1974. Expectation States Theory: A
  Theoretical Research Program. Cambridge, MA: Winthrop.

Berger, J., M. Hamit Fisek, R. Z. Norman, and M. Zelditch, Jr. 1977. Status
  Characteristics and Social Interaction: An Expectation States Approach. New York:
  Etsevier.