The institution of marriage, which has been defined as ‘the union of man and woman such that the children born from the woman are recognized as legitimate by the parents’ (RAI, .Notes and Queries on Anthropology 1951) has constituted a central area of anthropological research, usually in the context of studies of the family (Morgan 1871; Westermarck 1921). Claude Levi-Strauss has made the most spectacular contribution to the development of this field; by placing marriage alliance at the very heart of kinship, he has shown how marriage is a structure of exchange resulting from the incest prohibition. The prohibition of incest, which is universal and requires the avoidance of union between close relations, has as its positive counterpart the institution of exogamy, the obligation to choose a marriage partner outside the close family group. Alliances are, nevertheless, not made randomly: exogamy has its own counterpart in endogamy, which demands or recommends marriage within a prescribed group or locale. Modern societies, for example, combine strict prohibitions on marriage between close or distant kin and affines, with what might be termed a less restrictive class and ethnic ‘homogamy’ (marriage with the ‘same kind’). Most so-called primitive societies have both strict exogamy between certain close relations and an equally strict endogamy between other relations or prescribed groups.
Thus it can be seen that this ‘universe of rules’ (Levi-Strauss 1969 [1949]) governing the organization of all human groups, has many implications. The prohibition of incest leads to exogamy, which implies marriage with others, and this in turn produces exchange and reciprocity. These marriage exchanges, in which women are exchanged by men for other women, do not take place between individuals but between groups of ‘givers’ and ‘takers’ who thus circulate women between them. In certain societies, these groups are always the same, with the result that they find themselves linked to each other by an endless series of matrimonial exchanges, each of which reinstates the first alliance and in so doing reaffirms their solidarity. As it is women who are at the heart of this exchange and as they are ‘a priceless good’, they transform it, adding a sort of extra value. It is this which justifies the giving of a woman as a wife to seal political alliances and to cancel debts of blood, honour or money.
But marriage is not only an opportunity to create new social bonds, it is often accompanied by "prestations, of greater or lesser importance, and as much symbolic as material, which firmly establish the marriage and set up long-lasting bonds between the two groups party to the exchange. Within the broad category of marriage payments, one can distinguish societies that practise bridewealth, involving prestations, in nature or kind, which the family of the groom must give to the family of the bride, from those which practise dowry, whereby a collection of goods and services are offered by the bride’s family to that of her future husband, either with or without the dower (the part of the goods to be returned to the wife if her husband dies before her). The structure of marriage prestations can have important political, economic and ritual consequences for the society as a whole (Comaroff 1980), and there have been a number of attempts to construct grand comparisons. For Jack Goody (1973), the choice between bride-wealth and dowry derives from a more general opposition between societies that have unilineal kinship systems, emphasizing descent, and societies that have bilateral systems, valuing alliance. For Collier and Rosaldo (1981), important symbolic and political consequences of gender relations can be correlated with a slightly different distinction: between bride-wealth societies, in which objects are transferred from the groom’s side to the wife’s side, and bride-service societies (often, but not always, relatively egalitarian hunting and gathering societies) in which the groom’s labour is given to the wife’s side.
Although emphasis is often put on the economic aspects of marriage or on marriage as a form of exchange, it should not be reduced to these functions alone, nor can we justify its nec-cessity in terms of these functions. The variation in forms of conjugal union between different societies clearly indicates its relative nature. Everyone knows that there are societies in which men or women have more than one marriage partner at the same time. These societies are referred to as "polygamous. The term "’polygyny’ is used when a man can have more than one wife at the same time, and "’polyandry’ when a woman can have more than one husband. However, there are also societies where marriage may occur between people of the same sex: among the Nuer for instance, studied by "Evans-Pritchard (1951), there is the institution of "woman-marriage. In this case a woman can give bridewealth to the relatives of another women and marry her. She then has absolute control over this woman and her children, delegating to a male "genitor the duties of procreation. In addition, the death of the spouse does not bring to an end the rights and obligations brought about by marriage. The Nuer also practise ghost-marriage, contracted by a widow on behalf of her dead husband if he has no heirs, or by a sister on behalf of her deceased brother if he has no progeny. We should also mention the levirate, which is the institution obliging a brother to marry his widowed sister-in-law; and the sororate which obliges a widower to marry his deceased wife’s sister. In the 1950s, Edmund Leach used examples like these, especially the marriage practices of the matrilineal Nayar of South India (see family), to query the universality of the Notes and Queries definition with which we started (Leach 1955).
These varied forms of conjugal union pose several questions. They may appear to us as aberrant, even absurd, but they cannot be dismissed as archaic or primitive elements of societies yet to reach civilization. On the contrary, such institutions occur in demographically important populations which have highly developed cultures and can in no way be regarded as at the dawn of humanity. Besides, do not our own changing practices, for example as a consequence of new reproductive technologies — allowing a mother to be impregnated by her dead husband, or a woman to rent her womb to beget children for another mother — suggest that we might be discovering equally unexpected arrangements ourselves?
It is also clear that all of these marriage institutions cannot be reduced to rational causes. In each case the particular form of the institution is an essential element of the social system to which it belongs. And, taken in their entirety, they are ‘the illustration of a series of logical possibilities derived from the application of very simple conceptual principles that appear to us as strange because we have not been able to imagine them for ourselves’ (Muller 1976). Finally these polygamous marriages between women, between the living and the dead, provide little evidence of the stability of the marriage institution. The divorces and remarriages, the common law marriages and free unions which are on the increase in Western societies, also point to the fragility of the institution and show above all that the functions that have been ascribed to marriage — the transfer of goods, the solidarity between the relations, the sexual division of labour, rearing of children — cannot be the consequence of any natural imperative. Therefore we can only call for a rethinking of the universality and durability of marriage.
