Some of my best friends are women, but I would not want my sister to marry one of them. Modern-minded persons criticize me for manifesting such out-dated prejudices, and would like to send me to Coventry for a compulsory course of reindoctrination. They may be right. It could conceivably be the case that in due course the Sex Discrimination Act will be tightened up, even to the extent of our recognizing that there are no `good reasons why the State should not recognize contracts which are in all respects like marriage, except for the sex of the parties concerned'. 1 We can envisage a society so enlightened that the relations between men and women will be purely platonic and it will be a matter of no concern whether two people are members of the opposite sex or not: or, alternatively, our feelings could be so completely homogenized, that it will make no difference to an emotional relationship whether Leslie and Julian have only one Y-chromosome between them, or two, or none at all. In that event I shall be shown to be wrong, and my critics entitled to erect a monument to female equality on my grave. But I have doubts, and suspect that long after I am dead men will g,.? on falling in love with women and women will continue to find their hearts wooed and won by men. And this fact, if it is a fact, makes a profound difference to our social institutions. I want to follow out the logic of social differentiation, relying as little as I can on putative facts about the differences between the sexes. Scientifically attested knowledge is scarce in this field. Many academics take refuge in a safe suspension of judgment, but this is a craven dereliction of duty. We have yet to develop an adequate theory of knowledge in social matters, and only by thrashing out the apparently telling arguments on this and similar issues can we clear our minds about the nature of social knowledge. Moreover, if those who might argue rationally draw back from doing so, they leave the field clear for others who suffer from fewer inhibitions and fewer scruples. And finally, serious social consequences can follow from public confusion about the logic of a situation, and what things are possible and what desirable. There is a danger that in our attempts to remedy the real wrongs done to women we shall only succeed, as with much modern legislation, in making a bad case worse. Two paradigm forms of social organization can be distinguished, based on the twin facts of human nature, that Men are all alike, and that Men are all different. We are all alike in being featherless bipeds, language-using  animals, sentient beings, centres of consciousness, and, granted certain conditions of age and health, rational agents. Each of us is different in spatio-temporal location, and, identical twins apart, in his genetic inheritance and the detailed biochemistry of his body; and, at a more conceptual level, in as much as each has a mind of his own, and can make up his mind for himself, and can make it up differently from anyone else. The fact that men are in some respects the same engenders fellow-feeling and gives rise to the `peer-group': the realization that each man is different is the basis of his sense of individuality and uniqueness, and sometimes of isolation, and is, I shall argue, catered for by the `pair-bond' giving rise to the `pair-group'. The peer-group is based on the principle of similarity. All the members of a peer-group have something in common, on the basis of which they all belong, and often this engenders a strong sense of social solidarity. Members of the same age-group, of the same college, of the same regiment, of the same trade or profession, of the same country, or of the same religion, all feel bound together by strong ties of togetherness. We are familiar with these forms of organization and the feelings they engender, but it is worth following out certain logical features of this type. It is, first, inclusive. Peer-groups are what the logician calls `equivalence classes'; that is, they are based on some equivalence relation, which is symmetrical and transitive. If I am the same age as you, you are the same age as me; and if I am the same age as you, and you are the same age as Peter, then I am the same age as Peter. And so it is that if we are associating together on the basis of our being the same age, or having any other property in common, and if someone else turns up who forms a friendship on the same basis with any one of us, he will be readily accepted by us all. `Any friend of yours is a friend of mine' is the logical basis of peer-group feeling. Peer-groups operate at a somewhat low emotional temperature, and the larger the group the lower the temperature usually is. This is because the members are not alike in all respects, and, in general, if there are to be many members, then the things they have in common must be fewer. Hence, although we may emphasize very much the things we have in common, they will not completely characterize any particular member, and each man will be more or less aware of a somewhat loose fit between his position as a member of the group and his own real individuality. Although I am an Englishman, I am not just an Englishman, and should feel insulted if I were described as being merely a typical Englishman; there is, I like to feel, much more to me than just that. Hence, although there can be some surges of peer-group feeling, they are unlikely to be intense for a long time. There are short-lived groups which generate intense, but transient, feelings; and there are groups which continue over the years to attract the loyalty and affection of men, but without getting them steamed up or keeping them awake at night.  The pair-group is by contrast exclusive, intimate, intensive, and, ideally, permanent. Whereas a peer-group is indefinitely large, a pair-group is necessarily as small as any group can be. Its motto is `two is company, three is none'. There is a one-one relationship, which emphasizes each person's uniqueness; he is the sole partner of his partner. Anyone else is seen as a threat to the relationship and his unique position under it, and because he is unique, he sees himself as being also indispensable. In a peer-group it is easy come and easy go; I am readily included in, but if I were to go away or disappear, the gap would be soon filled-there are many other Englishmen, many other Oxford men, many other philosophers, many other members of the committees on which I serve; I am not indispensable, and when I die, I shall be only briefly missed. It is quite different inside my family, where I stand in a unique relation and could not be replaced. The one-one relationship also makes for intimacy. There is no need to restrict the respects in virtue of which the relation obtains, in order that it may be more extensive. And therefore, because it is intimate, it is also intense. Everything I do is significant to you, and it matters very much to me not only what you do but what you think and feel and would like to do. From this certain other features follow. There is a tendency towards further differentiation of roles. In all groups each member is anxious to contribute towards the common good; but, whereas in peer-groups there is a competitive spirit in which each tries to out-do the others in achieving what are generally regarded as desirable goals, in a pair-group the contributions are complementary rather than competitive. Each partner wants to do things for the other, but will tend to do those things that the other cannot do so well herself, and will readily accept from the other services which were not in the first place indispensable, but rapidly become so. Many married men lose their bachelor abilities to darn socks and boil eggs, and few wives retain their facility at filling in forms or cleaning sparking plugs. We do not need to posit a natural differentiation between the sexes in respect of these abilities, but only a social pressure towards differentiation that will arise in any partnership. Exactly the same division of labour can be observed where two sisters, or two brothers, or a mother and daughter, or a father and son, or just two friends, set up house together. A division of labour is not only often economically more efficient but emotionally more enjoyable: each is doing something for the other, each is more needed by the other. Even in the days of our innocence, Milton assigned Adam and Eve different parts of the Garden of Eden to tend. Because the pair-bond is exclusive, intimate and intense, it needs to be permanent. Nobody can envisage its dissolution with equanimity. A peer-group relationship depends on relatively few respects, and I can envisage my being altered in those respects while still remaining essentially the same person. I shall stop being a teenager or a young person, I can leave  my school or college, I might even emigrate and become a Canadian or a New Zealander, and still remain essentially me. But the ties of family are so intimate, that they constitute a large part of my identity. I cannot see myself being replaced by any one else, nor can I imagine that a pair-bond with any other partner would mean the same as my existing partnership does. Hence the tendency in every such partnership to require an intention of permanence. There can, of course, be partnerships that are explicitly intended to be only temporary-two friends going off for a holiday in Greece, or two authors working together on a joint book. But as the partnership deepens, and becomes more intimate and complementary, the question of the future will insistently arise, and the partners will be faced with the dilemma of either making it clear to each other that their partnership is only for a limited time and limited purposes, or entering into an unlimited commitment to each other. They must either draw back or else go on. And if they go on, they come to mean that they shall go on for good. As the bond between the partners increases in intimacy and intensity, the pair become more and more of a unity. It becomes less and less meaningful to enquire what each is doing in isolation from the other, and more and more natural to enquire, rather, what they are doing together. At the same time, the focus of each becomes less concentrated on the other, and more merged with that of the other in joint attention on some external concerns. At this stage the complementary differences between the two partners come to be taken for granted, and the background similarities become more important. The conventional wisdom that marriages are most likely to succeed where the husband and wife have a good deal in common is based on this fact, which seems to run counter to the doctrine I have put forward that pair-groups are based on difference, and similarities give rise, rather, to peer-groups. But neither difference nor similarity can be absolute. Although we may differ in many respects, we must resemble each other in some: and although we may all have some things in common, we must differ from one another in some ways, and each man must in some respects be different from everyone else. Even in the small, cohesive peer-group there is, and has to be, some recognition of differences between different members; and even in marriage, based as it is upon a fundamental differentiation between the sexes, there is, and has to be, some reliance on similarities between the spouses. These become more evident, as the couple become more firmly united, and act together more and more as a single person. We regard them as a unit, and characterize them, therefore, in terms of the features and dispositions they both possess. There is, moreover, some emotional necessity that each should cease to be exclusively occupied with the other, and that they should engage in joint concerns external to them both. Dante portrays Paolo and Francesca as condemned to be for ever making love to each other in hell, their hen being constituted by the sweet nothings that they were continually addressing  to each other. Mutual self-absorption becomes self-absorption simpliciter, and turning in on itself becomes empty and futile. If a pair-group is to grow, it must develop a common concern for something other than itself. Mutual absorption must give way to joint creativity. Normally and naturally in marriage this finds expression in the procreation of children. It does not have to be so: there are marriages in which both husband and wife find fulfilment in a vocation, or some crusade or other joint enterprise they are both engaged on; and other partnerships must find fulfilment in the contribution they can together make to society or some good cause. But relatively few people are called on to make a very definite contribution of their own, and it is correspondingly difficult for a couple to find a joint identity in their work or other vocation. For most people, therefore, the creative urge will find its expression most naturally in the bearing and rearing of children. Thus far, the argument has not depended essentially on the facts of sexual differentiation. Any pair-bond has an inherent tendency towards exclusiveness, permanence, intimacy, intensity and creativity, and will therefore manifest both a differentiation of roles and a sharing of interests and activities. Quite apart from biological facts, there would be some advantage in there being some antecedent differentiation of partners. This is largely on account of our acting under conditions of imperfect information. Partners cannot know each other all that well at first, yet much depends on their concerting their activities and their mutual aspirations and responses. If we had a unisex pair-bond in a unisex society, although it might be possible for the members to get themselves satisfactorily paired off, there would be many problems in discovering by trial and error who was to play what role in their mutual relationship, and quite a few pairs would be likely to come unstuck in the process. Homosexual relationships seem to be much less stable than heterosexual ones. It may, of course, be argued that this is due to social disapproval and the absence of any institutional framework: but it could also be explained as due to the absence of antecedent complementarity. If each party starts with some idea, either innate or the result of social convention, of what to do, then there are likely to be far fewer misunderstandings and mutual disappointments. Each will know what to expect of the other, and neither will be likely to let the other down simply through not knowing what part to play. A pair-bond involves complementarity, and pair-bonding is likely to be successful if many of the respects in which partners will be able to complement each other are immediately evident. Fewer feet get trodden on if men always dance with women than if men dance with men or women with women, because at the beginning of each dance the newly coupled partners each know who is to give the lead and who to follow it. These games-theoretical pressures towards differentiation of roles have clearly been manifested in social mores and may have led to genetic differences in the innate behaviour  patterns of the sexes. Certainly, once granted the fundamental biological difference between the sexes, there will be a tendency to further differentiation in a wide variety of respects between one sex and another. It does not follow that the differences have to be exactly those that have customarily obtained. Cooking has traditionally been regarded in England as a feminine accomplishment, but in Ancient Greece it could be said oudeis mageirainan eide popote, `Nobody ever saw a woman cook'. 2 Until this century it was a man's job to be knowledgeable about horses, whereas now horsiness is typically a girlish condition. The agricultural labourers in Victorian Oxfordshire regarded the carrying of water as women's work, but towards the end of the century began to think of it as a burden better borne by themselves. 3 Whatever the social conventions are, different men and women have different attitudes and aptitudes. Masculine men do needlework, feminine women play trains. If there are innate differences between the sexes, they are not uniform or invariable: nor should social differences be insisted on too rigidly. Although there are great advantages in having clear and unambiguous guidelines when we have to act under conditions of imperfect information, we should not automatically extrapolate to conditions of fuller information. Stereotypes are valuable, but we should not be completely bound by them. We may base social institutions upon definite assumptions about what people can be expected to want or do, but should allow for the possibility that some may reasonably have other aspirations, and should build some flexibility into whatever structures we set up. Nevertheless, the need for flexibility should not be taken as requiring flabbiness. The different roles assumed by different partners do not have to be the same in every age, nor in any one age the same in every case. But if they are to be varied, they need to be varied explicitly, with a fairly full understanding of the implications of the change. Else, many people will be confused, and some may be badly hurt. The Liberation movements object to there being any social recognition of differences between the sexes. It is partly on general egalitarian grounds. Women caught up in radical movements in the early 1960s noticed the discrepancy between the egalitarian professions of the, mostly male, leaders and their actual behaviour, and instead of seeing this as another instance of the incoherence of egalitarianism, concluded that `the last equality-that between men and women' was the one now needing to be achieved. 4 More important is the objection that difference has traditionally  involved an element of dominance. According to the traditional patterns, the man takes the initiative and has the final say. An inevitable corollary is that he also bears the burden of responsibility, and if things go wrong, he carries the can. Often men are expected to recognize a further obligation, also correlative to their having the decisive role, although the correlation in this case is not a conceptually necessary one, to pay particular regard to the needs and wishes of women. In the code of chivalry, although decisions are to be taken by men, much greater consideration is to be given to women. So, too, in the Book of Common Prayer, the woman promises to obey, the man to `worship' (ascribe worth to) and to endow. What other relationships are possible within a partnership? Clearly, the roles could be reversed, with the women wearing the trousers, and meek submissive menfolk doing their chores at their bidding. Unkind critics of America say that such a sex-reversal has taken place there, and that the destructiveness of American children is due to their moms, needing to be bossed but no longer having their husbands to take firm decisions for them, allowing junior, therefore, to rule the roost. But maybe only male chauvinists say this, and certainly many people would controvert both the alleged facts and the suggested interpretation. Matriarchy is a possible form of family organization. It may have its disadvantages, but then it may have advantages too, and cannot be ruled out of court a priori. More common, however, than proposals for reversing the position of the sexes is the claim that there should be no difference between them as far as dominance is concerned. Neither should have either the first or the last say, but each should have an equal say with the other. It is an attractive proposal in as much as it appears to express in a phrase the requirement of mutual consideration: even if the husband (or the wife) had the last word, he (or she) ought to pay great regard to his wife's (or her husband's) wishes, and therefore often defer to her (or his) actual words: but, although attractive, it fails to meet the essential criterion of a decision-procedure, of being able always, even in disputed cases, to yield a clear decision. If there is a dispute---say, whether to live in Slough, England, or Trenton, New Jersey---and neither is persuaded of the merits of the other's choice, then in default of some background understanding-for example, that the wife goes to live where the husband wants to work-they can only part and go their separate ways. Faced with such a possibility, one of them may give way, and go where the other wants rather than break up their marriage. But then there will be a premium on being the last to give way, and, instead of being resolved against a background of shared understanding, disputes will be fought out by threat and counter-threat, bluff and calling of bluff. The whole internal structure of the partnership will be altered. Moreover, if both are determined to be the last to give way, neither will give way in time, and they will break up. The price of each really having an equal say is that they will be more external to each other  and more likely to part. It would be different in a threesome. Then each could be bound to the others on the basis of a majority vote's being ultimately decisive of what they should all do. But a pair-bond cannot be based on a majority vote, and therefore must, if it is to maintain its characteristics of intensity and permanence, be based on some other, non-egalitarian, principle of dominance. The structure of threat and bluff employed by a warring couple can be analysed in terms of the theory of games, where it is known as `The Battle of the Sexes'. Each would rather live with the other than live apart, but each would rather the other give way rather than go and live where the other wanted. He prefers Slough to Trenton, but Trenton to breaking up, she prefers Trenton to Slough, but still would put up with Slough rather than live on her own. We can express this in the form of a 2 X 2 matrix, with his choices given by columns, hers by rows, and the value of each outcome tabulated for him and then for her:
|He lives at|
|She lives at Slough||(20,10)||(0,0)|
|She lives at Trenton||(0,0)||(10,20)|
Susan Haack, `On the Moral Relevance of Sex', Philosophy 49 (1974), 93.
2. Pherecrates, fr. 64.
3. Flora Thompson, From Larkrise to Candleford (Oxford, University Press, 1954), ch. 1, pp. 7-9.
4. Betty Roszak, `Women's Liberation', in Betty Roszak and Theodore Roszak, Masculine/Feminine (New York, Harper and Row, 1969), 297-298; reprinted in James Rachels (ed.), Moral Problems, 2nd edn (New York, Harper and Row, 1975), 123-I24.