The literature on ‘sex roles’ is very large. It is also more than a little confused, especially over the differences between ‘sex roles’, ‘sex differences’ and ‘sexual character’. An example is the well – known research on androgyny by Sandra Bern, which attempts to measure psychological traits of masculinity and femininity. Her questionnaire is entitled the ‘Bern Sex Role Inventory’ though it does not include questions about roles in any strict sense of the term. In literally hundreds of other studies, information on sex differences is presented with a loose assumption that role phenom­ena explain the differences observed. So it is often quite difficult to specify exactly what theory is contained in ‘sex role’ literature.

There is, nevertheless, a definite body of social theory organized around the concept of ‘role’. While formulations of the concept (which go back to the 1930s) differ in detail, most have five points in common which form the logical core of role theory. The first two state the essential metaphor, an actor and a script:

1 an analytic distinction between the person and the social position she occupies;

2 a set of actions or role behaviours which are assigned to the position.

The other three state the means by which the social drama is set in motion and held to its script:

3 Role expectations or norms define which actions are appropriate to a given position;

4 they are held by people occupying counter-positions (role senders, reference groups),

5 who enforce them by means of sanctions – rewards, punishments, positive and negative reinforcements.

These concepts are the tools by which role theory attempts a general analysis of social interaction. Broadly, role theory is the approach to social structure which locates its basic constraints in stereotyped interpersonal expectations.

This paradigm can be applied to almost any type of human behaviour, and in either very general or very narrow terms. Textbooks make a point of this range, listing examples of roles ranging from the breadth of ‘speaker of a language’ to the narrowness of ‘astronaut’ (cases given in Bruce Biddle’s Role Theory). Accordingly the role paradigm can be applied to gender relations in various ways. In one direction the ‘roles’ can be very specific. Mirra Komarovsky’s work on the family for instance attempts detailed descriptions of role behaviour in courtship or marriage. In a more recent text called Role Structure and the Analysis of the Family a group of American sociologists list a remarkable number of roles they have discovered inside the American family, including the ‘childcare role’, the ‘kinship role’, the ‘sexual role’, the ‘recreational role’, not to mention the ‘provider and housekeeper roles’ (which are, happily, complementary). What such lists mainly demonstrate is the vagueness of the role paradigm.

Most applications of role concepts to gender are of a different kind. Their basic idea is that being a man or a woman means enacting a general role definitive of one’s sex – the ‘sex role’. There are, accordingly, always two sex roles in a given context, the ‘male role’ and the ‘female role’; less commonly but equivalently called ‘man’s role’ or ‘woman’s role’, the ‘masculine’ or ‘feminine role’, etc.

This way of talking about gender is attractive in several ways. First, it allows a shift away from biological assumptions about sex differences, emphasizing that women’s and men’s behaviours are different because they respond to different social expectations. Some of the most fruitful research stimulated by role ideas has looked at the way these expectations are defined in the mass media. The constricted character of media images of women is striking, a point noted in the 1960s by Betty Friedan as part of ‘the feminine mystique’ and repeatedly confirmed in media studies since.

Second, sex role theory connects social structure to the formation of personality, an important and difficult theoretical task. Pro­ponents of general role theory such as Ralf Dahrendorf claim that the concept ‘falls on the borderline of sociology and psychology’.

More exactly it offers a simple framework for describing the insertion of individuals into social relations. The basic idea is that this occurs by ‘role learning’, ‘socialization’ or ‘internalization’. Thus feminine character is produced by socialization into the female role, masculine character by socialization into the male role – and deviants by some kind of failure in socialization.

This argument leads to an interest in the people and institutions responsible for the learning, the so-called ‘agencies of socialization’: mother, family, teachers, peers, media. Another large body of research inspired by sex role ideas has explored the different treatment of girls and boys by these ‘agencies’, the ways models of femininity and masculinity are conveyed to children, and (in a few cases) what happens when the messages are mixed. In the most sophisticated versions of sex role theory, such as Talcott Parsons’s, the concept of socialization is linked to psychoanalytic ideas about the structuring of the unconscious. Usually, however, sex role thieory is seen as an alternative to psychodynamic explanations such as Freud’s, and the focus is firmly on overt influences and overt behaviour.

Third, sex role theory offers principles for a politics of reform. If the subordination of women is largely a result of role expectations that define them as helpmates or subordinates, their characters as passive or expressive (rather than instrumental), then the obvious path forward is to change the expectations. A great deal of energy in contemporary feminism has gone into just this kind of enterprise. It is formalized in counter-sexist curricula in schools, antidiscrimi­nation laws, equal opportunity policies in the labour market and ‘affirmative action’ campaigns. Though broadly the territory of liberal feminism, this goes beyond the classical liberal focus on the individual, as the sex role concept points to the collective dimension of social stereotyping. As Alison Ziller, the Director of Equal Employment Opportunity for the NSW government, observed: ‘an affirmative action plan… means that remedy for discrimination does not rely on grievance procedures initiated by individuals’.

These virtues are substantial. Accordingly, sex role theory must be taken seriously as an intrinsic theory of gender for reasons beyond the sheer volume of its literature. Yet these virtues are bought at the price of very serious conceptual difficulties.

The problems begin with what many role theorists see as their greatest strength, the emphasis on the social via the concept of ‘expectations’. Role theory is often seen by psychologists as a form of social determinism, stressing the^way individuals are trapped in stereotypes. Stereotyped interpersonal expectations are indeed social facts. They are made effective, in role theory, by the idea that other people reward conformity to them and punish departures from them. In role jargon, the occupants of counter-positions sanction role performance. Little boys are praised for being assertive, ridiculed for being girlish, and so on. But why do the second parties apply the sanctions? This cannot be explained by their role expectations; if so, role theory reduces to an infinite regress. It quickly comes down to a question of individual will and agency, revolving around choices to apply sanctions. The social dimension of role theory thus ironically dissolves into voluntarism, into a general assumption that people choose to maintain existing customs.

Ultimately, then, role theory is not a social theory at all. It comes right up to the problem where social theory logically begins, the relationship between personal agency and social structure; but evades it by dissolving structure into agency.

What happens in sex role theory is that the missing element of structure is covertly supplied by the biological category of sex. The very terms ‘female role’ and ‘male role’, hitching a biological term to a dramaturgical one, suggest what is going on. The underlying image is an invariant biological base and a malleable social superstructure. This is why discussion of sex roles constantly slides into discussion of sex differences. The implicit question in sex role analysis is what particular superstructure has been created in such-and-such circumstances, and how far the biological dichotomy still shows through.

The result of using the role framework, then, is an abstract view of the differences between the sexes, and between their situations, not a concrete account of the relations between them. As Suzanne Franzway and Jan Lowe observe in their critique of sex role theory in feminism, the role literature focuses on attitudes and misses the realities that the attitudes are about. The political effect is to highlight the pressures that create an artificially rigid distinction between women and men, and to play down the economic, domestic and political power that men exercise over women.

As some critics have observed, we do not speak of ‘race roles’ or ‘class roles’ because the exercise of power in these areas of social life is more obvious to sociologists. With ‘sex roles’, the underlying biological dichotomy seems to have persuaded many theorists that there is no power relationship here at all. The ‘female role’ and the ‘male role’ are tacitly treated as equal. They are of course different in content, but reciprocally dependent on each other (‘complementary’), made of the same ingredients, and (in the eyes of the mid-1970s liberals who created a critical literature on the ‘male role’) equally oppressive of the human being within.

For an account of power, role analysis substitutes a theory of norms. Anne Edwards has observed how drastically sex role theory simplifies the complexities of gender: reducing all masculinities and femininities to one dualism; sweeping all women into one feminine role, which in turn is equated to being a housewife and located in the family. Most sex role theory is not constructed around problems raised by field observation, but as analysis of a normative standard case.

This standard case is, typically, an abstract model of a nuclear family with a conventional sexual division of labour. It is ‘standard’ in the sense that most people’s lives are presumed to be like it, with a minority being deviant. It is ‘normative’ in two senses. First, it is presumed that people in general regard it as the proper way to live – so it defines actual role expectations. Second, theorists regard it as the proper (or socially functional or biologically appropriate) way to live. Much of the intuitive, common-sense appeal of sex role literature arises from the blurring of these ideas. Thus when the sexologist John Money writes of the ‘proper’ path of psychosexual development, the effect is to stigmatize any departure from it (including homosexual choice) as pathological, and to reinforce the idea that conventional heterosexuality is a good thing both for the person and the society.

A crucial difficulty is that what is normative, i. e., expected or approved, is not necessarily standard, i. e., actually the way things usually happen. Perhaps especially not in the case of sexuality. Research has produced a series of upsets. The Kinsey studies are the most celebrated, finding frequencies of homosexual behaviour in the American population that normative sex role theory has never come to terms with. Data on premarital and extramarital heterosexuality make further trouble. Data on intra-family violence are also difficult to absorb: sober estimates by Straus and others that more than 50 per cent of American families have experienced domestic violence suggest a very large gap between what is morally approved and what actually happens. Statistics on household composition show that the mother-father-two-kids-cat-and-dog nuclear family routinely invoked by priests, presidents and advertis­ing copy-writers, is not the majority form of household now and perhaps never was. The normative pattern of husband-as – breadwinner and wife-as-home-maker, still powerful in ideology, has been undermined in fact by economics: as shown in chapter 1, around a third of the world’s paid workers are women, while a good many men are not in the labour force. Role theorists who do close-up field research, like Komarovsky, have to push the data very hard to get anything like the theoretical model of ‘roles’ out of it.

If we distinguish what is normative from what is common, instead of blending them together, new and important questions emerge. It becomes possible to see what is ‘normative’ not as a definition of normality but as a definition of what the holders of social power wish to have accepted. This raises questions about whose interests are embodied in the ‘norms’; how far the daily life of other people represents resistance to those interests; and what potentially normative principles might emerge from currently non – normative but widespread practices.

Sex role theory has a way of accommodating departures from the normative standard case, through the concept of deviance. This term is out of favour as a result of sharp criticisms of ‘labelling’ in welfare practice, but it persists in role theory because it is logically required by the normative concept of ‘role’. Euphemisms are available – inadequate self-concept, non­conformity – but it is no surprise when a role theorist like Biddle roundly talks of‘deviant behaviour’, ‘deviant identities’, ‘causes of maladjustment’, and contrasts them with ‘successful role learning’.

The dominance of the normative standard case in sex-role literature, plus the concept of deviance, have a distinct effect. They create the impression that the conventional sex role is the majority case, and that departures from it are socially marginal and likely to be the result of some personal eccentricity, produced by imperfect or inappropriate socialization. Lesbianism, men’s homosexuality, chastity, prostitution, marital violence and transves­tism are all liable to this treatment. Again the role framework eliminates the element of power from gender relations. It also eliminates the element of resistance to power and social pressure, the fact of social struggles – open or covert – going on around definitions of sexuality and gender.

Here, in thinking about the conflict of interests, the use of sex – role theory in feminism is most limiting. The conceptualization of sexual politics as role reform – as updating, liberalizing, or expanding the ‘female role’ (and in the men’s movement version the ‘male role’ also) – means that there is no social theory of the reforming movement as such, no conception of the constitution of collective interests within gender relations. The motive of role reform is individual discomfort in the existing version of the sex role. When comfort is achieved, there is nothing to carry either the politics or the analysis forward.

The lack of a theory of movement and social struggle reflects, at a deeper level, the lack of a way of grasping social contradiction and formulating a social dynamic. The sex role framework is fundamentally static as social theory. This is not to say that role analysts ignore change. Far from it. Modernization of the female role was a leading theme in one of the first important statements of sex role theory, by Talcott Parsons in 1942. Shifting expectations have been a leading issue in North American discussions of the male sex role in recent decades. Changing – definitions of the female role are the central theme in academic social science’s response to feminism.

The point is that sex role theory cannot grasp change as history, as transformation generated in the interplay of social practice and social structure. Structure is given to sex role theory in the form of biological dichotomy. And the ultimate voluntarism of the practical side of role concepts prevents the formulation of a concept of social determination. As a result, change is always something that happens to sex roles, that impinges on them. It comes from outside, from society at large, as in discussions of how technological and economic changes demand a shift to a ‘modern’ male role. Or it comes from inside the person, from the ‘real self’ demanding a relaxation of the constricting sex role. The role itself is always being pushed. Sex role theory has no way of grasping change as a dialectic arising within gender relations themselves. As an intrinsic theory of gender it is therefore limited in a fundamental way.

To sum up, four basic considerations oblige us to abandon sex role theory as a framework for the social analysis of gender: its voluntarism and inability to theorize power and social interest; its dependence on biological dichotomy and its consequently non­social conception of structure; its dependence on a normative standard case and systematic misrepresentation of resistance; and the absence of a way of theorizing the historicity of gender.

To recognize these weaknesses, as Edwards notes, does not prevent fruitful research on stereotypes of femininity and mascu­linity, that is, on sex roles as social constructs, cultural ideals, media contents and so forth. (I will take up these questions in chapter 11.) But as Edwards also argues, it is necessary to look for ways of theorizing the domain of sex and gender that give more attention to social institutions and social structures.

The critique of sex roles has nevertheless given some useful pointers towards what such a theory should be. Plainly one of the issues it must be able to grasp is the formation and conflict of social interests that occur in gender relations. I will now turn to approaches that make this their central concern.