The study of kinship is so central to anthropology that Robin Fox has likened it to logic in philosophy, as ‘the basic discipline of the subject’ (1967: 10). This is hardly surprising, since it deals quite literally with matters of life and death, not to mention identity and personhood, honour and shame, control of property, and succession to positions of authority.
Despite the importance of these themes, anthropologists have never agreed on what kinship is, let alone how to deal with it. Their views on these matters are as varied, contested and multidimensional as human life itself, and Needham’s iconoclastic statement that there is ‘no such thing as kinship’ (1971: 5) was not meant to deny the significance of the topic, but rather to emphasize that like all such notions in comparative anthropology, kinship is not a clearly delimited ‘thing’ but an amorphous, polythetic concept.This lack of precise definition may even be liberating rather than restrictive, since it helps undermine the persistent delusion that the task of kinship studies is to isolate and analyse semi-algebraic kinship ‘systems’. For any individual, kinship does not constitute a closed system, but an open-ended set of opportunities and constraints.
Kinship and biology
It is crucial to understand that kinship relationships are quite distinct from biological relationships. Kinship systems vary greatly, but as physiological processes are the same everywhere, these variations are clearly social rather than biological. There is more to it than this, however. It had long been conventional to distinguish social parents (pater and mater) from physiological ones (genitor and genetrix), but John Barnes (1961) introduced a further distinction between genitor and genetrix, the supposed biological parents of the child, and the genetic father and mother from whose sperm and ovum the child was actually produced. Like mater and pater, genetrix and genitor are socially ascribed roles, assigned according to prevailing ideas about the biology of conception, combined with assumptions about the sexual activities of possible parents. Of these three levels — social kinship, emic views about physical kinship, and genetic relationship — social anthropology is concerned only with the first two. It deals not with biology itself, but with biological kinship as culturally defined by the society concerned.
Trobriand Islanders’ views on paternity illustrate the need for Barnes’s distinctions. Mal-inowski (1932) reported their belief that a child’s spirit entered its mother’s body through her head and was nourished in her womb.
Sexual intercourse was seen as a necessary preliminary to conception, but its sole function was the mechanical one of opening the vagina and providing space for the child to grow, so although Trobriand children presumably do have genetic parents, they have no genitors according to local dogma. Yet the role of pater is crucial, and for a woman to bear a child having no social father is highly dishonourable. Interestingly, Trobrianders were perfectly familiar with notions of biological paternity but denied their validity; in other words, this denial was ideological rather than an indication of their state of empirical knowledge. The same probably applies to many emic statements about biological kinship, even those purporting to be ‘scientific’ or factual.
For heuristic purposes, kinship can be seen as having three co-existing aspects. (1) At the categorical level it comprises forms of nomenclature and classification. These provide the conceptual framework whereby people experience and understand their environment. The relationship terminology is the most obvious example in kinship, but Aboriginal Australian section systems also qualify. (2) At the jural level, it includes the rules which affect people’s kinship behaviour, covering everything from criminal laws to ideas about good manners. Jural rules are phrased in terms of the categories just mentioned, but whereas these categories are normally taken for granted, rules are explicit, subject to disagreement, and can be broken. (3) The behavioural level consists of what people actually do. It can be subdivided into collective behaviour, expressible statistically through such notions as marriage or divorce rates; and individual practice, which depends upon the circumstances at the time. Practice is influenced by jural rules of course, but in a more complex way than is often assumed.
The three levels are far from congruent, and none fully determines the forms taken by the others. Consider choosing a spouse, for instance. In a few societies one finds prescription, whereby the relationship terminology positively defines marriageable categories of relative. In these and many more societies, there are also jural preferences for people to marry particular relatives such as first cousins. This obligation is never absolute, however, and whether individuals actually make such marriages depends on the political, economic and emotional advantages of choosing an alternative spouse rather than the ‘preferred’ partner. Marriage choice is therefore not a matter of blindly obeying a rule: obedience is only one possible strategy, and although conformity often has advantages, these may be outweighed by other considerations.
This example shows that in order to comprehend kinship fully, all three levels must be taken into account. Many anthropological arguments about kinship arose because the protagonists failed to do this; having defined the problem at different levels, they proceeded to talk past each other. Nineteenth-century scholars emphasized jural aspects, partly because many — like Morgan and tJ.F. McLennan — were lawyers, but also because reports of local customs were the most accessible data for armchair theorists. Among later writers, Radcliffe-Brown also saw kinship primarily in jural terms, as a means of allocating rights and responsibilities and regulating their transmission from one generation to the next, while Levi-Strauss dealt with it mainly as a means of classifying persons. This was not always clear in their writings, however, because Radcliffe-Brown often discussed rules in language more appropriate to statistical behaviour, while Levi-Strauss conflated the classification of relatives with rules of marriageability.
Approaches to kinship: (1) descent theory
Historically, there was a clear division between those who saw kinship as based upon descent links between parents and children, and those who concentrated on alliance relationships created by marriage. Radcliffe-Brown was foremost among those who saw kinship primarily in terms of descent (1950: 4). Along with other descent theorists like "Fortes and "Jack Goody, he drew a clear distinction between kin, relatives by descent, and "affines, relatives through marriage; hence his frequent use of the phrase ‘kinship and marriage’, implying that the latter was somehow external to the former. He further distinguished "agnates, persons descended from a common ancestor through males only, from cognates, descendants of a common ancestor or ancestress counting descent through both males and females.
Radcliffe-Brown classified kinship systems according to how descent was recognized. Two particularly distinctive forms are patrilineal descent, reckoned through males only, and matrili-neal descent, reckoned only through females. These should not be confused with patriarchy and matriarchy, for in both cases official power resides primarily with men; under patriliny a child acquires social status primarily from its father, whereas under matriliny its mother’s brother is the key figure (the avunculate). Patrilineal descent is more common worldwide, perhaps because of the added complexities involved when men transmit rights to other men in the female line.
However, this kind of typology was criticized by Edmund Leach (1961: 4) as ‘butterfly-collecting’. The mere fact that two societies are labelled ‘matrilineal’ does not mean they have anything predictable in common. As no societies are wholly regulated by a single descent principle, they cannot be termed ‘matrilineal’ without specifying which social contexts are involved. Rivers (1924: 85-8) long ago stressed the distinction between: descent proper, i.e. membership by birth of a particular group such as a lineage; inheritance of property; and succession to a title or office. The latter processes are not necessarily governed by descent-based principles, and even when they are these do not always operate in a simple way. Succession to the British throne is basically patrilineal, for example, but a directly descended female takes precedence over a collateral male; by contrast, succession to noble titles is purely patrilineal, so women cannot inherit under any circumstances. The possible permutations are almost endless.
Turning to descent in Rivers’s strict sense, Fortes portrayed Tallensi society in Northern Ghana as entirely built around the "lineage system (1970: 34). Whether he is worshipping ancestors, arranging marriages, allocating work, or exerting judicial authority, a Tallensi man’s rights and responsibilities are determined by his position in his patrilineage. So although lineage membership is determined by kinship criteria, its functions are economic and political. For this reason Fortes drew a distinction between "filiation and descent. Filiation stemmed from being the legitimate child of one’s parents and was normally "bilateral, i.e. children were filiated to both parents. By contrast, jural status was determined by pedigree — descent from a particular ancestor. In patrilineal cases, a man had descent and filiation links on his father’s side, but only filiation on his mother’s side. Filiation was relevant only in domestic contexts, whereas descent was a politico-jural matter, though the fact that it was expressed in the vocabulary of kinship provided an ideological bridge between the two domains.
Unlike Radcliffe-Brown, then, Fortes saw descent as external to kinship proper, which he limited to the domestic domain. As for marriage, it was an ephemeral matter concerning only those directly involved, quite different from the enduring lineages on which Tallensi social structure was based. Yet marriages did form a ‘web of kinship’, which held society together by transcending the social barriers between lineages (1970: 82). It was in this context, too, that Fortes employed his controversial notion of complementary filiation.
Groups and categories
Is ‘descent’ best seen as an empirically observable characteristic of real groups out there in society, or an emic ideology for making sense of social life? Barnes (1962) argues that despite their superficial similarities to Tallensi-style seg-mentary lineages, local groupings in the New Guinea Highlands cannot be understood using theoretical models developed in Africa. Whereas lineage membership is irrevocably determined by birth for a Tallensi man, New Guinea communities are ad hoc groupings round charismatic Big Men, based on individual choice. Many men do join their fathers’ groups, but the all-embracing descent ideology so prevalent in Africa is absent.
Harold Scheffler (1966) draws an important distinction between descent groups composed of people related by unilineal descent, and descent constructs, which are classificatory concepts used to help visualize one’s society. It is quite common to find one in the absence of the other. Wherever sons inherit land from their fathers local males are likely to be related by patrilineal descent, yet as in New Guinea people may not think of their society in that way. Conversely, communities may be spoken of in lineage terms even when most people are not actually members (or spouses of members) of that lineage. This happens among the Nuer of the Sudan, whose villages contain many persons linked to the local lineage in other ways (Evans-Pritchard 1951: 28). Local group composition may be very similar in societies with strong descent ideologies like the Nuer, and societies lacking such ideologies, as in New Guinea. The difference lies not in the residential pattern itself, but in how people think about it.
Approaches to kinship: (2) alliance theory
Descent theory made sense in some lineage-based African societies, but proved inappropriate for much of Australasia and the Americas, either for the reasons just discussed or because marriage alliance, far from being ephemeral, forms enduring patterns which persist over time. For example, marriage among the Kachin of Burma is hypogamous, i.e. the bride’s family is higher status than the groom’s family (Leach 1961). It follows that marriage can only occur in one direction, and that any two lineages are in a wife-giver/wife-taker relationship which is just as persistent over time as the lineages themselves.
Alliance theory, which stresses marriage as a structural principle, is especially associated with Dumont, Leach and Needham, but its pioneer was Levi-Strauss (1969 ). He saw marriage as the other side of the coin from incest prohibitions: both helped prevent local groups from becoming sexually self-sufficient, and so encouraged wider social cohesion. Although this conflation of incest rules and exogamy is open to criticism, it is certainly true that marriage in some societies follows regular, enduring patterns. These ‘elementary structures of kinship’ therefore involve both positive and negative marriage rules, whereas complex structures only have negative rules. He identified three elementary structures, though others have argued that these can be reduced to two, symmetric and asymmetric exchange.
In the symmetric case, women from group A marry men from B, and conversely B’s women marry A’s men. Similar reciprocal marriages occur between C and D. Levi-Strauss views this as a fairly rudimentary form of social cohesion, and terms it restricted exchange because it seems to link only a few people. This need not actually be so, however, as B and C could also be linked symmetrically.
In the simplest case (Figure 2), where A and B are nuclear families, all marriages involve first cross-cousins. Cross-cousins are the children of a brother and sister, whereas parallel cousins are the children of two brothers or two sisters. This particular form is called bilateral cross-cousin marriage, because the spouses are related on both their fathers’ sides (patrilaterally) and mothers’ sides (matrilaterally).
Do not take such diagrams too literally, however! It is more realistic to envisage A and B as two large descent groups, containing hundreds or even thousands of people. First-cousin marriage may be rare, and most marriages will involve second, third, or Nth cross-cousins. In the extreme case, that of dual organization, A and B together make up the entire society. Finally, these units need not even be descent groups at all, and could just as easily be local groups such as villages.
Figure 2 Bilateral cross-cousin marriage.
Suppose A’s women marry B’s men, but B’s women cannot marry A’s men and must marry men from C instead. Similarly, C’s women marry D’s men, and so on. This entails a clear distinction between wife-givers and wife-takers, as with the Kachin.
Levi-Strauss calls this generalized exchange, arguing (controversially) that it provides greater social cohesion than restricted exchange, and so permits more complex social systems to develop. If the exchange units are single families, as in Figure 3, every man marries his matrilateral cross-cousin (MBD). The same caveat applies as before, however; in practice they are much larger groups, so most people do not marry actual first cousins.
Patrilateral cross-cousin marriage
Levi-Strauss regards the converse case, where men marry their patrilateral cross-cousins (FZD), as a third distinct type in which exchanges reverse in each generation. But is this just an illusion fostered by Figure 4? As the real exchange units are much larger than single families, there can be no sharp distinction between generations. Even as the youngest people in one generation are marrying, so are the oldest of the next; at any given moment exchange is occurring in both directions, making the situation indistinguishable from symmetric exchange. Figure 4 cannot therefore be taken literally, though it is always possible that people may visualize their society in this way — as with descent, we should perhaps distinguish exchange groups from exchange constructs.
Figure 3 Matrilateral cross-cousin marriage.
Figure 4 Patrilateral cross-cousin marriage.
Approaches to kinship: (3) the cultural approach
Although some kin relationships, like those between ‘in-laws’, are seen as purely social, most are felt to have a biological basis. "David Schneider found that Euro-Americans define relatives both in terms of common ‘blood’, and by the fact that they behave like relatives. In his memorable phrase, kinship involves both natural substance and code for conduct (1968: 29) -though ‘natural’ is itself a cultural notion, of course.
Fieldworkers adopting this approach are often led to examine emic ideas about the person. As Schneider found, personhood is commonly seen as constituted partly by kinship practices, and partly by genetics as locally understood. For example, Marriott and Inden (1977) argue that South Asia is characterized by monistic thinking quite different from the mind-body dualisms of much Western thought. Every kin group, from caste to family, has its own specific ‘code for conduct’, yet this moral code is thought to take physical form, such as shared blood or substance. Group members must adhere to the codes regarding marriage and sexual intercourse, so that their personal identity will remain secure and their distinctive substance will be passed on to their offspring.
All group members broadly share this common substance, but there are also individual differences and fluctuations. Daniel (1984) shows that Tamils see the timing, location, and style of sexual intercourse as having auspicious or inauspicious implications for the well-being of a couple, and the gender, health, and fortune of their offspring. Moreover, mere mixing of sexual fluids is not enough. To mix properly they must be compatible, and this is investigated by studying horoscopes before marriage. As social and physiological factors are dialectically related, a horoscope is simultaneously an account of one’s social destiny, and a biopsy revealing the character of one’s bodily substance. Incompatible sexual fluids produce unhealthy children, or none at all in extreme cases.
Such accounts are vital because they indicate the means whereby people represent their own actions to themselves and others, but they also have potential pitfalls. First, ethnographic evidence is rarely clear-cut: some South Asians say that substance derives from semen and ‘spirit’ from uterine blood, but others say substance comes from both parents and spirit derives from the ether. Moreover, the sociology of such theories is complex; rival views are often current, and many people have no coherent theory at all. Second, somewhat ironically given the insistence on working through indigenous rather than Western concepts, there is a tendency to write individuals out of the script in favour of abstractions like ‘Indian thought’. Third, it is often implicitly assumed that indigenous explanations have the same purpose as anthropological theories, namely, to provide abstract explanations of social phenomena. As the next section shows, this is far from the case.
Approaches to kinship: (4) kinship practice
Classic kinship writing emphasized jural rules, but conformity to such rules was usually seen as unproblematic and explanations were sought only when they were ‘broken’. Yet obedience and disobedience are both matters of choice, so it is necessary to explain why some people adhere to the rules as well as why others ignore them. Even this takes too naive a view of the connection between rules and behaviour, however. In brief, rules do not direct (or fail to direct) behaviour; rather, they are used to interpret, explain or justify it.
The atemporal character of earlier approaches led to a downplaying of the strategic aspects of kinship behaviour. By contrast, Bourdieu (1977) sees kinship as an open-ended set of practices employed by individuals seeking to satisfy their material and symbolic interests. Thus, marriage choice is made in the light of one’s social situation at the time, including the options available in the form of marriageable persons and the material and symbolic capital to be gained by choosing each of them.
One factor in this complex calculation is that by ‘obeying the rules’ one gains respect. However, such rules refer to official, publicly acknowledged kinship entities — lineages and the like — rather than the practical kinship units called into existence for specific purposes. Whereas official kin generally come to the fore in celebrating marriages, many other kinds of practical, ad hoc kinship links may be involved in setting them up.
The major determinants of kinship behaviour, therefore, are not the explicit rules themselves but people’s largely implicit knowledge about ‘how things are done’, i.e. about social practice. Explicit ideologies are manifestations rather than explanations of this practical knowledge, which Bourdieu terms habitus. For these reasons, the study of rules alone yields a picture of kinship which is not only incomplete but also seriously misleading. For example, because anthropologists use genealogies to depict ‘real’ relationships, they tend to forget that informants often use genealogical discourse in other ways, to support or justify particular activities and concerns.
Analyses of this kind can be criticized for reducing human motives to cynical attempts at maximizing personal, material advantage. But just as it is necessary to steer a middle course between descent theory and alliance theory, using insights from either as the situation demands it, so too this approach based on praxis provides a necessary dynamic and inherently sociological counterweight to the somewhat asocial, atemporal insights provided by the study of emic theories of personhood.
Kinship and gender
The central place occupied by kinship in anthropological history has gone hand-in-hand with an emphasis on its socio-structural rather than domestic aspects. The locus classicus for the kinship ethnographer has been the public arena of lineage politics and ostentatious weddings, rather than the private one of the cooking hearth or vacuum cleaner. In small-scale societies, kinship — or at least the kinship metaphor — often encompasses politics, economics and religion too, making the resulting theoretical debates central to the discipline as a whole. This is in marked contrast to the status of the sociology of the family as an empirical backwater. Yet the downplaying of household and family relationships has undesirable consequences too; for example, kinship theorists are ill-equipped to say much of value about those intimate contexts to which kinship in urban, large-scale societies is increasingly confined. Above all, it leads to the major paradox that kinship studies have paid almost no attention to issues of gender.
Most kinship writing has been thoroughly chauvinistic — a feature not wholly eradicated above because the terminology itself usually presumes the male viewpoint. To some degree this can be empirically justified by the fact that in many societies groups of men exchange women in marriage rather than the other way round, but this argument again privileges the public domain over the private. In any case, there is a world of difference between consciously assuming the male perspective in order to depict events as male participants see them, and doing so unthinkingly, which was what normally happened in the past. The problem was not that ethnographers misrepresented indigenous views, but that they represented only one such viewpoint, that of men.
Howell and Melhuus (1993) argue that such distortions were inevitably exposed once the focus of interest shifted from the first two approaches outlined above to the second two. There are self-evidently at least two kinds of person in every society, male persons and female persons. Similarly, when ethnographers look at kinship practices rather than social structure, they frequently discover that whereas ‘official kin’ are indeed predominantly male, ‘practical kin’ are far more likely to be female. It may seem bizarre that it has taken so long for kinship studies to take gender relations fully into account, but at least there is no longer any excuse for not doing so.
Reproductive technology and future kinship
This article began with the premise that biological facts are universal and unchangeable, yet the reproductive technologies now available have the potential to bring about previously unimaginable changes in kinship. Two particular areas of interest are the changing patterns of rights and obligations involved; and the possible redefinition of kinship relations themselves. Even that (as culturally defined) most natural of all notions, motherhood, is called into question by the emergence of previously unknown statuses like ‘egg-donor’ or ‘surrogate mother’, while genetic fingerprinting, which allows paternity to be unambiguously established, may lead to conflation of the roles of ‘genitor’ and ‘genetic father’. A new kinship entity has also made its appearance, with distinctive rights and a legal personality — the embryo. Even more basically, ideas about ‘nature’ itself, which have always been fundamental to cultural notions of kinship, seem bound to change (Strathern 1992).