Particular transactions involving power are easy enough to observe. Mr Barrett the Victorian patriarch forbids his daughter to marry; parliament makes homosexual intercourse a crime; a bank manager refuses a loan to an unmarried woman; a group of youths rape a girl of their acquaintance. It is often difficult to see beyond individual acts of force or oppression to a structure of power, a set of social relations with some scope and permanence. Yet actions like the ones just listed are not intelligible without the structure. Rape, for instance, routinely presented in the media as individual deviance, is a form of person-to-person violence deeply embedded in power inequalities and ideologies of male supremacy. Far from being a deviation from the social order, it is in a significant sense an enforcement of it.

This connection of violence with ideology points to the multiple character of social power. Force is one important component. It is no accident that the means of organized violence – weapons and knowledge of military technique — are almost entirely in the hands of men, as shown in chapter 1. Yet ‘naked force’ is rare. Much more often violence appears as part of a complex that also involves institutions and the ways they are organized. Power may be a balance of advantage or an inequality of resources in a workplace, a household, or a larger institution. By and large the people who run the corporations, the government departments and the universities are men, who so arrange things that it is extremely difficult for women to get access to top positions. Organizational control is no more naked than force usually is. Both secrete, and depend on, ideologies. The ability to impose a definition of the situation, to set the terms in which events are understood and issues discussed, to formulate ideals and define morality, in short to assert hegemony, is also an essential part of social power. Much of the critical work of feminism and gay liberation has necessarily been devoted to contesting cultural power: for instance cultural definitions of women as weak, or of homosexuals as mentally ill.

That relations of power function as a social structure, as a pattern of constraint on social practice, is in one sense all too obvious. The constraint on practice extends to the elemental question of staying alive. Helen Ware, in Women, Demography and

Development, notes that in rich countries, where basic nutrition is not a problem, women live longer than men; in the poorest countries women die earlier than men. Here, it appears, discrimi­nation against women in forms like less food and less medical attention is operating at a level where life is at stake. Differential infant mortality, including infanticide of girls, is another case in point.

It is less obvious, but also important, that the practice of those who hold power is constrained as well. Men are empowered in gender relations, but in specific ways which produce their own limits. For instance, in a patriarchal gender order emphasizing monogamous marriage there is serious tension between men about issues of adultery; a structure that defines women as a kind of property makes men liable to reprisals for theft. Sustaining hegemonic definitions of masculinity is often an issue of importance, and homosexual men attract hostility partly because they under­mine these definitions.

As with labour, the structure of power is an object of practice as well as a condition. Many accounts of patriarchy give the impression of a simple, orderly structure like a suburban war memorial. Behind the facade is likely to be a mass of disorder and anomaly. Imposing order requires a mobilization of resources and expenditure of energy. What Donzelot calls the ‘policing of families’ is just a part of this. Research on the welfare state, such as Sheila Shaver’s study of Australian taxation and benefit payments, shows an apparatus of social policy that assumes women’s dependence on men and reinforces it.

Imposing order in and through culture is a large part of this effort. One notes, for instance, the enthusiasm with which the Catholic hierarchy – all men – emphasize an ideal of purity, meekness and obedience for women. Its effectiveness has just been demonstrated in Ireland where the Church has succeeded in defeating a referendum to permit divorce. Elsewhere in the capitalist world priests are becoming less important as ideologists of gender, journalists more so. Though ‘quality’ newspapers like The Guardian in Britain are liberal in sexual politics, most mass- circulation journalism is unrelievedly sexist and homophobic. The people doing this cultural and material ‘policing’ are not necessarily the main beneficiaries as individuals. They are, rather, participants in a collective project in which the power of men and subordination of women is sustained.

If authority is defined as legitimate power, then we can say that the main axis of the power structure of gender is the general connection of authority with masculinity. But this is immediately complicated, and partly contradicted, by a second axis: the denial of authority to some groups of men, or more generally the construction of hierarchies of authority and centrality within the major gender categories.

The authority of men is not spread in an even blanket across every department of social life. In some circumstances women have authority; in some others the power of men is diffuse, confused or contested. Research by American feminists such as Carroll Smith-Rosenberg has traced the history of institutions and practices which have been controlled by women, among them girls’ edu­cation, friendship networks, and non-market production. This insight can also be inverted. We can identify a complex of institutions and milieux where the power of men and the authority of masculinity are relatively concentrated. There is a ‘core’ in the power structure of gender, contrasted with the more diffuse or contested patterns of power in the periphery.

In the advanced capitalist countries, four components of this core are reasonably clear: (a) the hierarchies and work-forces of institutionalized violence – military and paramilitary forces, police, prison systems; (b) the hierarchy and labour force of heavy industry (for example, steel and oil companies) and the hierarchy of high – technology industry (computers, aerospace); (c) the planning and control machinery of the central state; and (d) working-class milieux that emphasize physical toughness and men’s association with machinery.

The connections between (a), (b) and (c) are familiar. President Eisenhower, not a noted feminist, warned of the power of the ‘military-industrial complex’ in the United States. There is a close analogue in the Soviet Union. In both countries, as Joel Moses observes of the latter, there is an ‘almost complete exclusion of women from the major policy-making centres’. These parts of the complex are tied together by an ideology linking masculinity, authority and technological violence, which has slowly become a focus of research. But it is their connection with (d) that is crucial in sexual politics as a whole. This connection gives a mass base to militarist beliefs and practices that might otherwise be so repellent as to destabilize the governments that rest on them. Perhaps the most striking feature of this connection is the extent

to which it is mediated by machinery, especially motor vehicles. The gradual displacement of other transport systems by this uniquely violent and environmentally destructive technology is both a means and a measure of the tacit alliance between the state and corporate elite and working-class hegemonic masculinity.

It was often pointed out by ‘men’s movement’ writers in the 1970s that most men do not really fit the image of tough, dominant and combative masculinity that the ideologists of patriarchy sell. That image is not intended to fit. The celluloid heroism of a John Wayne or Sylvester Stallone is heroic only by contrast with the mass of men who are not. The ‘justifying’ ideology for the patriarchal core complex and the overall subordination of women requires the creation of a gender-based hierarchy among men. (I stress ‘gender-based’ because discussions of power relations between men have commonly stopped after identifying divisions of class and race.) As gay liberation points out, an essential part of this process has been the creation of a negative symbol of masculinity in the form of stigmatized outgroups, especially homosexual men. In general, then, a hierarchy is created with at least three elements: hegemonic masculinity, conservative masculinities (complicit in the collective project but not its shock troops) and subordinated masculinities.

Feminist thought around 1970 commonly identified the family as the strategic site, the key to the oppression of women. If anything the pendulum has now swung too far the other way. It has become clear that household and kinship relations are not a test-tube case of pure patriarchy. The family as an institution might best be regarded now as part of the periphery rather than the core complex. Colin Bell and Howard Newby remark that the authority of husbands requires a good deal of bargaining and negotiation for it to be maintained in working order. The importance of bargaining and the tensions around marital au thority are confirmed by a considerable body of close-up research on families, from Mirra Komarovsky’s classic Blue-Collar Marriage to more recent studies like Lillian Rubin’s Worlds of Pain in the United States, Pauline Hunt’s Gender and Class Consciousness in Britain, Claire Williams’s Open Cut and Jan Harper and Lyn Richards’s Mothers and Working Mothers in Australia. These studies also suggest a recent historical shift, with husbands now finding it more difficult to impose an openly patriarchal regime in the home.

The contestation of domestic patriarchy is so widespread in

some settings that it makes sense to talk of a working-class feminism, rooted in these struggles as much as in paid employment. Marital power struggles are often won by the wives. The research that produced the case study of the Princes in chapter 1 also showed a number of families where the father’s control was eroded or simply non-existent. Komarovsky noted this in the 1950s too. I think it is important to acknowledge that there are genuine reversals of power here. It is not a question of women being conceded an apparent power which can then be revoked, but of the hard relational outcomes of domestic conflicts and negotiations over years or even decades.

It is also important to acknowledge that these local victories do not overthrow patriarchy. Komarovsky observed among her American working-class couples that where the wife was the controlling member of the marriage, this could not be acknowledged publicly; a facade of men’s authority was maintained. The general implication is that we must distinguish the global or macro­relationship of power, in which women are subordinated to men in the society as a whole, from the local or micro-situation in particular households, particular workplaces, particular settings. It is possible for the local pattern to depart from the global pattern, even to contradict it. Such departures may provoke ‘policing’, i. e., attempts to establish the global pattern locally as a norm. They may also signify structural tension that leads to large-scale change in the longer run.