In both academic social science and the popular literature on gender in the last two decades the commonest approach has been through concepts of social moulding or ‘socialization’.

Schematically, the main argument runs like this. The new-born child has a biological sex but no social gender. As it grows older society provides a string of prescriptions.,templates, or models of behaviour appropriate tr’ sex-or the other. Certain agencies

of socimizatimT-notablv the family, the jned-i-a thp-pppf grn^ip and the schooL— makethese expectations and models concrete and provide the settings in which they are appropriated by the

child. The sequence of agencies may be important and distinctions between primary and secondary socialization are often made. Various mechanisms of learning come into play: conditioning, instruction, modelling, identification, rule learning. There is much debate in the gender-socialization literature about the relative importance of these mechanisms. Whatever the details, the social models or prescriptions are internalized to a greater or lesser degree. The result is a gender identity that in the usual case corresponds to the social expert a i-ffin^tor that sex. Some cases will ‘deviate’ because of the abnormal functioning of an agency of socialization (for example, a father-absent family), or because of exceptional biology. The products of these deviations are homosexuals, transsexuals, intersexes and others whose gender identity fails to correspond in the usual way to their sex.

A close affinity between socialization concepts and sex role theory is obvious. The social prescriptions are often called ‘norms’ and the process of спгіяі Ірятігцг iq often ПІ1Н ‘role learning’, ‘role acquisition’, nrfspv rnjp qnriali^atinn’ T n this extent the critique of role theory^stSfeHin chapter 3 applies in full force. Role-learning theory is internally incoherent, and incapable of providing a genuinely social analysis of social process. But though they are commonly phrased in sex role terms, strictly speaking the main elements of socialization analyses can be separated from the role framework and need to be considered in their own right.

The notion of a school, family, peer group or television network as an ‘agency of socialization’ implies a definite script, a charter under which the agency acts on behalf of the society, and a degree of consensus about what it is to do and how to do it. Such smooth pictures of socializing agencies are given by most of the literature on sex-role socialization, including the writings of liberal feminism.

Both historical evidence about the institutions involved, and close-up studies of their contemporary working, oblige us to reject this picture of consensual agencies acting on behalf of the social order as a whole. Social history shows schools and families often in conflict with each other and with larger social structures. Examples are Richard Sennett’s Chicago study Families Against the City, and Pavla Miller’s Long Division which shows how the mass school system in South Australia was imposed as an intervention into working-class families and in the teeth of considerable resistance from them.

Nor are these institutions internally homogeneous, consensual, or even roughly consistent in their dealings yith the people being ‘socialized’. Psychoanalytic studies of families such as Laing and Esterson’s Sanity, Madness and the Family ahd Frangoise Dolto’s Dominique show intolerable contradictions in the pressures and demands placed on children. These case^ are, by definition, exceptional in the intensity of the pressures they generate. But other evidence goes far to show that internal conflict and cross­pressure is quite normal. Examples are American research on families like Mirra Komarovsky’s Blue-Collar marriage and Lillian Rubin’s Worlds of Pain, and the British researcnv on working-class schools cited earlier. The classic of liberal feminism, Betty Friedan’s The Feminine Mystique, was precisely an exploration of contradiction within the American bourgeois family. When institutions appear consensual and consistent, closer examination is likely to show this is because a great deal of energy has been put into overriding conflicts of interest to create a front of harmony and good order. This is shown for ruling-class schools in Teachers’ Work, and is evident in case studies of families like the Princes in chapter 1.

These data have considerable implications for our understanding of the way institutions shape people. The notion of ‘socialization’ rests on the idea of internalization through one or other of the metdTahisms described bv the psychology of learning. What is produced inside the person is a psychological structure that reproduces 0i>-reftert5The^characteristics of the~socializing agency. The potioja-Qf-^nodclling1, taking on the attributes ol the soctalizer, encapsulates this feature of socialization theory.

This becomes untenable as a general conception of gender formation once we recognize the fact of contradiction within the process. This eliminates the possibility of mechanical causation: the only mechanical outcome of contradiction would be random­ness, and the evidence will not support that. Instead it introduces two possibilities, both of which lead away from ‘socialization’ notions to something radically different. One is the production within the person of psychological structures qualitatively different from the properties of the institution or milieu. This is the basis of classical psychoanalysis. The other is the possibility of constitutive choice in a field opened up by contradiction. This is the basis of existential psychoanalysis.

The popularity of socialization notions in academic research has been supported by two occupational blindnesses, the inability of sociologists to recognize the complexities of the person, and the

unwillingness of psychologists to recognize the dimension of social power. Both groups have been willing to settle for a consensual model of intergenerational transfer – playing down conflict and ignoring violence – and for a consensual model of the psychological f * structure produced.

^ / ^The notion of a ‘gender identity’ at the core of femininity and ^ ^masculinity is the psychological "counterpart of the notion of a ‘sex / role’ into wMclTone is s^oa^tzedrlndeed its basis seems tcTBeTthe act ofTecognizingmieself as the kind of person that conventional images of femininity and masculinity define. Accordingly investi­gators like Robert Stoller trace the foundations of gender identity back to the first years of life, wTienthe child is introduced to definitions of femaleness and maleness. Xikewise the origins of aberrant gender identity, in StolIeFs case studies of transsexuals, are traced to aberrant family constellations at that stage of life.

The notion of identity as a coherent core to personality or sexual character was popularized in the 1940s and 1950s by Erik Erikson in Childhood and Society and by the culture-and-personality school of anthropology. It is not accidental that this happened at the same time as the rise of role theory in sociology. Identity theory traded on the psychoanalytic idea of different levels in personality, but gutted it by removing the idea of fundamental conflict between the levels. Surface levels of personality became the more or less straightforward expression of the core.

A full critique of such models of personality would involve a long detour. I will state the main points briefly. A homogeneous or consensual model of gender identity loses the ability to account for creativity and resistance. It recognizes the production of different gender practices only as deviance resulting from inad­equate or aberrant socialization. Its homogeneous picture of the core misses or marginalizes that mixing of elements of gender which classical psychoanalysis referred to as normal bisexuality and which even scalar conceptions of personality have now acknowledged as important (androgyny etc.). And it makes a cognitive act – recognizing oneself as male or female – central to the structuring of emotion without acknowledging either the social structuring of such cognition (gender involves social categories presented to the child in the practices of adults) or its dependence on emotional processes that are not cognitively organized (what the Freudians call ‘primary process’).

Socialization theory, supposing a mechanism of transmission

and a consensual model of what is produced, has been credible only to the extent that social scientists have been willing to ignore both choice and force in social life. I would argue, with Sartre and Laing, for seeing them as constitutive. ‘Agencies of socializ­ation’ cannot produce mechanical effects in a growing person. What they do is invite the child to participate in social practice on given terms. The invitation may be, and often is, coercive – accompanied by heavy pressure to accept and no mention of an alternative. The emotional pressure involved has been well brought out in autobiographical writings from the American antisexist men’s movement, such as those by Seymour M. Miller and Michael Silverstein. Yet children do decline, or more exactly start making their own moves on the terrain of gender. They may refuse heterosexuality, though gay autobiographies show it often takes a long time to construct a positive alternative as homosexual. They may set^about blending masculine and feminine elements, for example girls insisting on cornpetitive~sport at~school. They may start a S^pirnTtheir own’Ty^S’, ‘for example boys d nTsImfTnTlTag when by themselves. They may TmTstruct a fantas-y~IIfe~at odds with their actual practice, which is perhaps the ccunmonest move of all.—————– —

If socialization were as smooth and successful as socialization theory presupposes, it would be difficult to account for the degree of violence that has existed historically in relations between adults and children. It seems that its level has declined a great deal in capitalist core countries over the last hundred years – one of the few senses in which this really has turned out to be, as Ellen Key predicted, ‘the century of the child’. Yet even here the decline has been very selective. In a Sydney survey of 1969, over 70 per cent of adolescent boys aged eleven to fourteen reported being caned or strapped at school in the current year. Observation in any supermarket will indicate that hitting young children is still a commonplace of daily life.

The imbalance of power makes it possible for adults to put very heavy pressure on children if they set out to. Adults too respond to contradiction in the ‘socialization’ process, and their response may take the form of attempting to impose order and direction on it, creating vehement regimes of character formation and indoctrination. A classic description of such a regime in an Irish Jesuit school was provided by James Joyce in Portrait of the Artist as a Young Man. In many cases vehement regimes do work as

intended: the novitiate produces the nun, the school produces scholars, the boy grows up to be the image of his father. Suddenly the socialization paradigm seems apt. Yet notions of‘role learning’ are all too mild for the fierce pressures generated in such settings. And even here the response may be rather different from what was intended; then we get not a good Catholic Civil Servant but a James Joyce.

An adequate account of gender formation must be able to understand such events as something more than random exceptions or social deviance. That is only possible if the analysis of the successful case of ‘socialization’ is in terms that also allow us to understand the unsuccessful.

Several criteria for a theory of gender formation are implicit in what has just been said. It must be able to reckon with social contradiction and contradiction within personality. It must be able to reckon with power and its effects without turning people into automata (for instance recognizing motives for compliance like the child’s need for love). It must be able to recognize different levels in personality without depending on the notion that one level simply expresses the characteristics of another. Most important – and in a sense subsuming the others – it must be historical, both in the sense of seeing the person in terms of a trajectory through time and situations, and in recognizing the constant historical reconfiguration of the social forces impinging on personal growth.

Those are reasonably strong criteria. There are, I think, only two approaches that come close to meeting them, classical psychoanalysis and existential psychoanalysis. Their accounts of gender formation are the subject of the rest of this chapter. For different reasons, neither fully meets the criteria just outlined. Considering these reasons points towards a reformation of person­ality concepts, which is attempted in the following chapter.