Читать книгу The Metamorphoses of Kinship - Maurice Godelier - Страница 9

CHAPTER THREE

Filiation and Descent

(First Component)

Filiation and descent: While both words are translated by the French filiation, Anglo-Saxon anthropologists make a distinction between the two terms. The majority of French anthropologists do not, and range themselves behind Lévi-Strauss,1 who, in a debate with Leach2 claimed that the distinction was not useful. This is not my opinion; I think that it is not only pertinent but important. For Meyer Fortes, Leach and Needham, who agree on this point at least, the term ‘filiation’ should be kept to designate the fact that every individual is at birth (or becomes by adoption) the son or daughter of a father or fathers and a mother or mothers who are themselves sons and daughters of a father or fathers and a mother or mothers, and so on. In a word, filiation is the set of ties that link children to their paternal and maternal kin. Filiation is bilateral and cognatic; it links the individual to both agnates and uterines.

DESCENT MODES

Descent is governed by other principles. The calculation of descent can be unilineal, duolineal, bilineal or non-lineal. In the case of unilineal systems, descent is traced through only one of the sexes. When descent is reckoned through men, it is patrilineal. When it is traced through women only, it is matrilineal. In the first case, a man’s sons and daughters belong to his descent group, but only sons transmit this belonging. In the second case the opposite holds true: a woman’s sons and daughters belong to her descent group, but only daughters transmit the belonging. The systems termed ambilineal or duolineal combine the preceding unilineal systems, such that each child belongs to its father’s group by virtue of a patrilineal principle and to its mother’s group by virtue of a matrilineal principle. Alternatively, bilineal systems are rare. They come in two kinds. Either descent is parallel, and the sons of a couple belong to the father’s lineage and the daughters to the mother’s. Or descent is crossed, and a couple’s daughters belong to their father’s lineage, and the sons to their mother’s.

Lastly, in non-lineal systems, the individual’s sex makes no difference, and descent is traced indifferently through the men and through the women. All descendants, in the male or the female line, of an ancestral couple can refer to this ascendance in order to claim rights to titles or land use, for instance. The sons, daughters, grandsons and granddaughters of this ancestral couple have of course married into other groups, and their descendants belong to these groups. Undifferentiated kinship systems therefore require other rules than ascendance to constitute coherent, bounded descent groups.

PATRILINEAL DESCENT

Among the societies that reckon descent grosso modo by a patrilineal principle, we find ancient Greece and Rome, ancient and contemporary China, the Nuer in Sudan, the Tallensi in Ghana, the Cyrenian Bedouins, the Iraqi Kurds, the Juang3 in central India, the Kachin in Burma, the Purum4 in India, the Melpa or the Baruya in New Guinea, the Tupinamba in Amazonia, and so on. These societies differ profoundly from one another. The Melpa prohibit the direct exchange of women and instead exchange wealth for wives. For the Baruya, direct sister exchange is the principal marriage rule, but they practise payment of bridewealth in the case of marriage with other groups. The Katchin forbid wife-takers to be wife-givers as well. When the givers are superior to the takers, the women circulate in the opposite direction of the wealth (which ‘rises’ from the takers toward the givers).

Let us spend a moment with the Juang, studied by McDougal (1964) and by Parkin in his monographic work on the Munda (1992).5 The Munda descent system is divided into three levels of segmentation. The first consists of eighteen patriclans (bok), which are exogamous, do not hold property in common, are dispersed in various villages but have different individual ‘totems’. These clans are in turn divided into thirty-eight local descent groups, each of which usually resides in its own village. The local groups are exogamous, and it is between these groups that marriages take place. The third level is formed by the lineages, which are of shallow genealogical depth, three or four generations on average, with two or three lineages living in the same village. They cooperate to gather the payment needed for their young men to marry, and redistribute the brideprice received when their daughters marry. Marriage must take place between persons with different totems, and therefore from different clans. Residence after marriage is virilocal. Lands and goods are inherited within the lineage, which also functions as a ritual unit.

MATRILINEAL DESCENT

Among the societies whose mode of descent reckoning is matrilineal, we can mention the Ashanti of Ghana, the Pende of Kasai, the Khasi of northeast India, the Nayar of South India, the Trobriand Islanders, the Maenge of New Britain, the Mnong-Gar and the Rhades of Vietnam, the Tetum of Timor, the Na of China, the Nagovisi of Bougainville, the Iroquois and the Hopi of North America. The Khasi are divided into matriclans (kur).6 The couple’s residence after marriage is uxorilocal, but in the daytime the family is composed of brothers and sisters living together with the sisters’ children. At night, the brothers join their own wives and children while the sisters’ husbands take their place. Na families, too, are composed of brothers and sisters and the sisters’ children. At night the men circulate among the houses and become the temporary lover of one or another of the women living in these houses. Marriage as a ceremony officializing a union exists only for village headmen’s families, and the kinship vocabulary, according to the Chinese ethnologist, Cai Hua, has no word for ‘husband’ or ‘father’.7

DUOLINEAL DESCENT

Duolineal systems are a rarity. We can cite the Yako of Nigeria, the Herrero of South Africa, and the Kondaiyankottai Maravar, a South Indian subcaste. The Kondaiyankottai Maravar have two kinds of clan: patrilineal (kottu) and matrilineal (kilai).8 Children belong to both clans. The patrilineal clans correspond to local, exogamous descent groups. They control inheritance of land and other forms of property, and settle successions. They are religious and ritual units that ensure the worship of family gods. The matrilineal clans are not localized but they too are exogamous. Every marriage must take into account each group member’s double clan affiliation. In duolineal systems, children belong to both clans, regardless of sex; but the continuation of the clan and the transmission of rights and statuses rest on unilineal principles: from the father to his sons for the patrilineal clan and from the mother to her daughters for the matrilineal clan.

BILINEAL DESCENT

Bilineal systems, whether parallel or cross, are even more rare. Examples of the first case are the Orokolo9 and the Omie;10 and of the second, the Mundugumor,11 three New Guinea societies, the first two of which live in the South in the Gulf of Papua, and the third in the North, in the Sepik. In the first case, an individual’s belonging to a lineage is transmitted from same sex to same sex (from father to sons and from mother to daughters), unlike unilineal or duolineal systems, in which children of both sexes belong either to their father’s or to their mother’s clan or to both at once. In cross systems, the clan to which boys and girls belong switches with each generation. Among the Mundugumor, a son belongs to his mother’s kin group, a daughter to the father’s. The son carries the name of his maternal grandfather, the daughter that of her paternal grandmother. This has been called a ‘rope’ system. Land is transmitted from father to son, but the father’s weapons and his group’s sacred flute are transmitted from father to one of his daughters, who will pass them on to her sons.

UNDIFFERENTIATED DESCENT

Examples of societies with an undifferentiated mode of descent are the Maori of New Zealand and a great many Polynesian societies – Samoa, Tonga, etc. At one end of the great axis of the Malayo-Polynesian migrations, we find the Imerina and several other societies of Madagascar. And between the two, the Penan of Borneo and other groups in Indonesia. Among the Maori,12 the descent group (hapu) stems, as its name indicates, from a group of ancestors to which the group members are connected by ties that run indifferently through the male or the female line. These ties can go back as many as ten generations or more. A person can belong to several hapu, but the fact of residing in one of them consolidates the rights to which he is entitled. Marriages very often take place within the same hapu, which is therefore strongly endogamous. Each hapu is under the political and religious authority of a chief, who descends in the male line from the eldest son of the founding ancestor. Each of the family lines that descend from the founding ancestor has a different status, determined by its distance from the family line of the eldest son. This line is a sort of aristocracy within the hapu, whereas younger sons of younger sons were treated as people of inferior status, as commoners. This example shows that in cognatic systems it is also possible to resort to unilineal principles for the purpose of transmitting certain functions and establishing various forms of hierarchy. In addition, real or potential belonging to several hapu offers individuals choices and strategies that are more open and broader than in strictly ‘lineal’ societies. In this case residence plays an important role in consolidating rights and reducing these choices.

HOUSE SYSTEMS

To the various modes of descent reckoning, long known to anthropologists, we will add a mode of constituting groups that calls upon both descent principles and marriage rules. This mode creates kinship groups that we in the West call ‘houses’; however such groups are also found in the aristocratic families of Japan and in the societies on the northwest coast of the United States and Canada characterized by ‘ranks and houses’, such as the Kwakiutl and their neighbours, or in certain politically ranked cognatic societies of Indonesia. In Europe, alongside the ‘royal’ houses and the various minor or great houses that made up the family lines of the feudal aristocracy (vestiges of which still exist) we find various peasant ‘houses’, such as the casa in Catalonia or the ostau of the Lavedan region.

When Lévi-Strauss made a close analysis of the workings of the numayn – the Kwakiutl ‘houses’ whose underlying structure (patrilineal and/or matrilineal principles)13 Boas, who had been the first to describe them, admitted having great difficulties in discovering – he drew anthropologists’ and historians’ attention to this type of institution. A numayn is one of a hierarchy of ‘houses’ each of which has a rank and a seat defined by a name and a crest. According to Lévi-Strauss:

The house is first of all a legal entity, in possession of a domain composed of material and immaterial goods . . . by material [I mean] the possession of a real domain, fishing sites and hunting grounds . . . The immaterial domain includes names, which are owned by houses, legends, the exclusive right to perform certain dances of rituals . . . [A house] perpetuates itself by handing down its name, its wealth and its titles in a direct or fictitious line regarded as legitimate on condition that this continuity can be expressed in the language of kinship or marriage, and most often both at once . . . There is so much freedom in this area that one can say that alliance and descent are interchangeable.14

It is thus not the house that belongs to the people but the people who belong to the house, together with material and immaterial assets which are indivisible and kept to be transmitted to their firstborn male or female descendants in the direct line. Alternatively, other parts of the domain can be temporarily detached and given to the husbands of the daughters of the house, sons-in-law, who then have the duty to transmit them in turn to their firstborn. The first rule in this system, the rule of primogeniture, implies gender equivalence for succession to certain titles and ranks. At this level, the system is cognatic, but skewed to the paternal line. The second principle – by virtue of which certain of the house’s titles are provisionally alienated by giving them to the men who marry the daughters of the house with the obligation to transmit them to their own children, who are therefore the giver’s grandchildren – posits a partial equivalence between alliance and descent. In short, instead of opposing descent through the male line and descent through the female line (or radically opposing descent and alliance), this system plays on all kinship relations in order to maintain a ‘house’ and conserve its name and its seat, in other words a rank, a crest and privileges in the political-religious system that encompasses all houses and ranks them with respect to all others within a territorial group which Boas calls a ‘tribe’. Furthermore, this system defines not only the rank and place of each house but the rank and place of each individual in the house according to birth order and sex.

Far from being merely a ‘language’, as Lévi-Strauss suggests,15 reiterating a thesis stated well before by Beattie16 and then by Leach,17 Tambiah18 and others, kinship relations actually function as relations of appropriation and transmission of the material and social conditions of the houses’ existence (fishing and hunting territories), and determine their capacity to amass the wealth that will be entered in the potlatches, the competition by means of gift-giving between numayn for titles or to legitimize their transmission. But these titles, ranks, crests and myths, the immaterial property of the ‘houses’, are not attributes of kinship. They do not exist physically as do the hunting and fishing territories that each house has appropriated. They exist socially as part of a political-religious system of titles and privileges distributed hierarchically, and therefore unequally, between the ‘houses’; and in virtue of this distribution they come to reside in the workings of the kinship relations that structure them. But, as I said, titles, ranks, crests and privileges are not attributes of kinship; they belong to a component of society that encompasses all kin groups and places them permanently at its service for its own reproduction: this part of society is the political-ritual system which enables the society to exist as a whole, and represents it as such. This political-religious system is therefore not to be confused with the worship each ‘house’ may pay its ancestors or the tutelary gods that afford support and protection. These acts of veneration of course apply to a universe of representations, myths and rites shared by all members of the society, but in themselves they are part of the specific identity of each ‘house’.

Lastly, there is one point that Claude Lévi-Strauss, Pierre Lamaison and the others who participated in the discussion on the notion of house neglected, and that is the fact that there are two major differences between the tribal aristocracy of the Kwakiutl ‘houses’ and the ‘houses’ of the feudal aristocracy of Europe and Japan. Kwakiutl society was self-governing, there was no state. In Europe and in Japan, there were various forms of state, of central power embodied in a royal or imperial house (in Japan it was the families of the emperor and the shogun). Among the Kwakiutl, commoners were kinsmen of the aristocrats, they were the younger siblings of descendants of family lines who had no seat or crest; they were freemen. In Europe and in Japan, commoners had no kin ties with the aristocratic families, and marriage was forbidden between these social classes – on pain, for the aristocrats, of forfeiting their title and their condition. Furthermore, commoners were not freemen. They were serfs or dependants, living on lands owned by the noble families, which they were allowed to use in exchange for payment of rents in labour, produce or money, thus enabling the nobles to live in keeping with their condition, surrounded by the signs of their distinction.

A certain similarity is created between the tribal and the feudal ‘houses’ (and in their successoral practices) by the fact that a large portion of the functions, and therefore of the state’s power, was shared between the families and family lines of the feudal hierarchy, who exercised these functions over their lands and their people – the right to sit in judgement, the right to raise troops and make war, the right to preside the village meeting, the right to collect taxes on the goods circulating or sold on their lands, and so forth. But, in Europe, there was yet another type of ‘house’, found more specifically in mountainous regions, where the economy was based on a combination of farming of privately owned lands and stock-raising on communal pastures, jointly owned by all members of the community. In these regions, one found village communities also organized into ‘houses’. Each had a name attached to it, whatever its owners’ patronym. Owing to the importance of their ‘house’ (i.e. the wealth composed of farming lands and livestock), these owners held a distinct rank in the organization of village power, and their voice carried more or less weight in managing community lands and business. Each ‘house’ also had its own pew in the parish church. In order to reproduce itself, the system required that each ‘house’ be transmitted as a block to a single descendant, usually the eldest son, but sometimes, if there were no sons, to the eldest daughter, whose husband would move into the ‘house’ and take its name. The system thus implied the creation of a strongly marked hierarchy between the descendants and, in general, the exclusion of all of the younger children from the inheritance; they in turn were obliged to leave the house and become priests, or soldiers, or stay at home but remain bachelors and become virtual servants to the brother or sister who had inherited the name and the property. In short, what is perpetuated in these systems is not a family line but a name and a domain that is conserved through the families which become the successive owners.

Having come this far, since we live in Europe where the feudal aristocracies have been replaced by ‘bourgeois’ dynasties – rarely very old and always turning over – and the system of peasant ‘houses’ has become residual (livestock raising no longer having the same importance in the mountain economy or being done indoors, and the French government having outlawed primogeniture), I will say a few words about another system, which was much more prevalent in Europe than ‘houses’ and continues to exist in almost all strata of our societies: this is the system of ‘kindreds’. Which brings us back to kinship.

Today’s world is populated with millions of monogamous nuclear families, often reconstituted, following a divorce, on the part of one member of the couple if not both, and including the children of this previous marriage. These individuals’ memories of their ascendants rarely go back beyond their grandparents. And the vast majority of them have forgotten the existence and the names of their grandparents’ collaterals, if they ever knew them. They may remember a few great uncles and great aunts and some of their descendants, distant cousins. Generally speaking these are descendants of their father’s and mother’s brothers and sisters, close cousins, whom one sees most often, as well as these cousins’ married or unmarried children. This is on average the extent of the genealogical memory and knowledge of most members of Western European societies and of those in the Americas who have European roots.

THE KINDRED

In this sea of nuclear families, where there is no general affiliation principle that might group them more or less mechanically into lineages and clans, there are nevertheless fuzzy collections of kin, whose existence is never guaranteed and whose longevity is always temporary. These groups are of two kinds and overlap in part.

The first is exclusively centred on an individual. This is what is called his or her ‘kindred’.19 The other kind is not centred on an individual but contains him or her. These are groups of families whose members are related to each other, and who see each other regularly and help each other; up to a certain point, they feel a bond of mutual solidarity. In so far as the family into which an individual is born (or adopted) is part of such a group, it can be said that this individual’s belonging is initially automatic. He (or she) will be socialized in childhood, at the same time as his siblings, within a group of families and kin that will overspill to a greater or lesser degree the bounds of the nuclear birth or adoptive family. In so far as these groups continue to exist, even if one or another of the individuals in them decides to break with his family or disappears, it can be said that they are decentred with respect to Ego and partially independent from Ego. They existed before Ego was born, but he determines whether they continue to exist through him. For an individual can decide to stop seeing certain relatives and their families, or vice versa certain relatives and their families can decide not to see Ego. In a word, the individual finds himself at the centre of a network of kin whom, depending on circumstances, he sees or ignores and who, for, the same reasons or others, see or ignore him.

A person’s kindred is first of all a network of individuals who are either directly attached to him, and he is at the centre; or are related by ties that start or end with him. To the ascendants and descendants we must add Ego’s brothers and sisters, who share the same ascendants, Ego’s half-brothers and half-sisters, who share some of them, their spouses, brothers-in-law and Ego’s nieces, etc. A person’s kindred includes not only paternal and maternal kin in the direct and collateral lines, but also their close affines – and of course those of Ego if he or she is married.

A kindred is thus not a descent group, in the sense of a lineage or a clan; it is a collection of people related by various close or distant ties with whom Ego maintains relations and whom, for example, Ego will invite to his wedding or whose funeral he will attend; whereas he will ask only certain ones – a brother-in-law rather than a brother, a maternal uncle rather than his father – for help or advice in certain circumstances. Quite often, the ties with certain kinsmen fade even before these relatives die because they have moved away or their position or social class has changed thereby creating too much distance with respect to Ego, not because of their kin ties, which do not change, but because of their social success or comedown – or even decline (in the eyes of society and/or Ego).

But we must not forget that, in addition to Ego’s kindred, there is another network, formed by his friends, some of whom are neighbours, others, more numerous, having become friends in other contexts – school, workplace, sports, etc. With his friends, the network of Ego’s personal relations with a portion of society’s members – people he will listen to or who will listen to him, people he will help or who will help him – expands.

Nevertheless, in Europe, as in America and Australia, with their myriads of families with short memories, there are a few who conserve (and even cultivate) the memory of ancestors going back several generations, of kin in the direct line but also collaterals. Without belonging to the European aristocracy, now of marginal social importance, these families carry a name that raises them above other families at the local, regional, national or even international level. This can be the name of an industrial dynasty, like the Schneiders in France, or a political dynasty, like the Kennedys or the Bushes in the United States. Because certain members of these families have made a name for themselves through various deeds whereby the family has gained exceptional prestige, their descendants are invited to preserve the memory of their ties with these famous ancestors.

The name becomes an immaterial shared asset that is usually unaccompanied by any other form of shared property – material or other – but which may be used by those who carry it as social capital. Moreover, it is in their interest to do so. Some even feel obliged to show themselves worthy of their name, either by doing as well as their illustrious ancestor in the same area or by excelling in others.

We thus see the spontaneous creation of cognatic pseudo-descent groups strongly skewed to the male line, which bring together all those who have inherited a name and pass it on to their children (men) or who carried the name in childhood but do not transmit it to their children (women) although they teach them that they have a stake in this name through their mother, and so on. Much less visible, and as though dotted, female lines thus grow up alongside the immediately visible male lines. Sometimes these groups, who keep their family tree up to date – including direct descendants and affines – hold family reunions, which everyone or almost everyone attends to get to know each other personally and count their members (like the Monods, whose reunions can number up to 700 persons from several countries in Europe). All that would need to happen is for these groups of families to own in common material wealth – land, factories, banks – or inherit political (or other) functions in order for genuine descent groups to come into being that would adopt rules for distributing the use of these resources among the families and for transmitting the task and honour of carrying out these functions to one or another of their number. It is clear, in this case, that the generalization of private, individual and family-held property, of the means of production and of money, make it difficult for descent groups to come into being, for they would tend to close ranks through the systematic application of a criterion of kinship (perhaps completed by other criteria allowing the inclusion or exclusion of certain types of kin).

Proof that this possibility once existed in Europe can be found by observing the customs of the inhabitants of Karpathos Island in the Aegean Sea. Until sometime in the 1920s, in order to avoid dividing the family assets, the rare farming lands and the houses, only first-born children married: the eldest son of one family with the eldest daughter of another.20 The eldest children bore a name, Kanacares, which distinguished them from their younger siblings. The latter were obliged to emigrate or, if they stayed on, to serve their elder sibling after he or she married. The assets that formed the dowry of each of the spouses were again passed on to the next generation, to the eldest sons and daughters. These marriage rules ultimately ensured the circulation of land, houses and statuses in two parallel lines of descent and inheritance, one male and the other female. The system disappeared sometime between the two World Wars, when the value of land fell and money became the principal form of wealth: the younger siblings who had emigrated to the United States or Australia and had made money came home and were able to marry an eldest daughter. The old logic was broken and did not survive.21 Today Karpathos has roughly the same matrimonial regime as the rest of Greece.

It must be stressed that kindred systems are not restricted to Europe. They are found in Borneo, among the Iban, where the kindred underpins the organization of the ‘long houses’, the Bilek, described by Derek Freeman,22 and in New Guinea among the Garia,23 and so on.

And if we extend a person’s kindred to all those individuals linked to him by cognatic ties, we will find ourselves in the Polynesian or Malagasy systems, where the kindred is a principle that works in tandem with the existence of cognatic descent groups. Taking yet another step, we will point out that the extended kindred, as the collection of Ego’s cognates and affines, exists in all descent systems – unilineal, bilineal and non-lineal. But it is usually pushed into the background, placed in sleep mode, or completely masked by the interplay of rules that assigns an individual to a descent group that is not Ego-centered. The Nuer,24 a society whose patrilineal system stresses agnatic kinship and lineages descending from a common ancestor, a relationship they designate by the term buth, use another word, jimarida, for Ego’s kindred, that is to say not only paternal and maternal kin but also the affines of Ego’s lineage.

The elegant table below, borrowed from Needham,25 shows the various descent modes that have been discussed. The letters m and f stand for ‘male’ and ‘female’.262728

m → m	patrilineal	Nuer, Tallensi,26 Turk,27 Baruya, Juang, etc.
f → f	matrilineal	Trobriand, Khasi, Iroquois, Ashanti, Nagovisi, Hopi, etc.
[m → m] + [f → f]	duolineal	Yako, Herrero, Kondaiyankottai Maravar
[m → f] + [f → m]	cross bilineal	Mundugumor
[m → m] + [f → f]	parallel bilineal	Orokolo, Omie, Apinaye
mf → mf	non-lineal	Maori, Imerina,28 Penan

The formula mf → mf, when it does not function as a rule for constructing cognatic descent groups, as in Polynesia, also represents (European, Euro-American) kindred-based systems, which are cognatic as well but do not give rise to genuine long-lasting descent groups.

DISTRIBUTION OF DESCENT MODES

Although we are far from having even a relatively complete inventory of the descent modes characteristic of the some 10,000 societies or local groups that currently exist on the face of the earth, taking our cue from the figures advanced by Roger Keesing,29 we can suggest the following distribution:

Patrilineal	Matrilineal	Duo-bilineal	Cognatic
45%	12%	4%	39%

1. Patrilineal systems predominate in the Chinese zone, in India, in the expansion zone of Islam, in a portion of sub-Saharan Africa, in New Guinea and in numerous North and South American Indian groups.

2. Matrilineal systems are numerous in Central Africa (the ‘matrilineal’ belt of Africa), in part of North America, in eastern New Guinea and in the Solomon Islands, and in a few minorities of Southeast Asia or China.30

3. Duo- and bilineal systems are few and far between; we find them here and there in Africa, Asia and Oceania.31

4. Cognatic descent systems are plentiful in the expansion zone of Malayo-Polynesian-speaking societies, which stretches from Madagascar to Easter Island, by way of Taiwan, which is close to this expansion zone. Large numbers are also found in Amazonia.

5. Cognatic systems based on kindred prevail in Europe, in populations of European origin in North and South America, in Japan and in certain areas of Malaysia and Indonesia.

In the absence of a map of the world-distribution of these systems, it is evident that we cannot go very far toward finding correlations between the nature of these systems and other aspects of the culture and structures of their associated societies.

I will complete this inventory by examining several theoretical problems raised by the analysis of these systems over the last decades.

Application of a descent rule, whether lineal or non-lineal, gives rise to groups of individuals of both sexes, in several successive generations, that have been dubbed ‘lineages’ and ‘clans, in the case of uni-, duo- and bilineal systems, and ‘branches’ or ‘demes’ – terms unfamiliar to the public – sometimes used to designate cognatic descent groups. Whereas applying the descent rule in lineal and bilineal groups suffices to define and close the boundaries of the groups, in the case of undifferentiated, non-lineal groups, other principles are needed to define a descent group and give it boundaries (individuals having resided on an ancestral site for a long time, for instance).

CONCERNING THE IMAGINARY CHARACTER OF DESCENT PRINCIPLES

An initial remark is needed here to underscore the abstract, social, and in part imaginary character of these rules and therefore of the kinship relations they engender. To favour the relations that pass exclusively through the male or exclusively through the female line is to attribute a different social status to the two sexes by overdetermining one to the detriment of the other in the production of social ties between the generations and, in each generation, between the sexes. Even in cognatic systems, where the rule of gender equivalence ensures the transmission of descent and the continuity of the group, in certain contexts only the men or only the women, the eldest or the youngest, etc. are selected to transmit certain components of social life – titles, land, ritual functions – which are, at the same time, components of the kin group’s identity and local conditions of the global reproduction of the society as a whole.

Later, we will discover another proof of the abstract, social character of kinship when we look at marriage systems and kinship terminologies that distinguish between parallel and cross cousins, two categories of individuals of both sexes whom we know (something we have learned recently) to stand at the same genetic and biological distance from Ego but whose social status is basically different and even opposed, since theoretically one may marry one’s cross cousins but not one’s parallel cousins, because the latter are assimilated to siblings. This remark poses the problem of the relationship between kinship and biology. But to tackle the question, we must start from a universal fact anyone, anthropologist or not, can see when travelling the world: there is no language that does not have a so-called kinship vocabulary. In all languages, whatever the society’s descent and marriage rules, there are specific words to designate the positions and relationships of individuals of both sexes in at least five generations – two ascending and two descending generations starting with an individual (or a class of individuals in positions equivalent to that of Ego) characterized universally by his or her sex and by the place he or she occupies in his or her generation with respect to others of both sexes born before or after and connected to either the same ‘relatives’ or to the relatives of these relatives.

In short, all societies are concerned with organizing, on the one hand, the succession of generations – an obvious condition of their physical continuity – and, on the other hand, the relationships between persons of both sexes belonging to a certain number of successive generations (usually five). All cultures attribute meaning to these facts and to these relationships. All languages talk about them. Individuals of both sexes belonging to different generations, which succeed each other and are linked to each other in the process of reproducing human life, as well as being linked by the place their society attributes to them in this process – these are, to my way of thinking, both the biological and the social underpinnings of kinship relations. Kinship relations are therefore not just any kind of social relations. They are not, for example, relations springing from the desire on the part of a certain number of individuals or groups to create a hometown football club and who get together to bring it about, finance it, train local players or recruit outside players, and who fight to see their club rise higher and higher in the national and international tables.

Some anthropologists, like Mary Bouquet or Janet Carsten,32 concerned with ridding the definition of kinship of any reference to the biological process of reproduction, devote considerable time to scrutinizing the word ‘relative’, which designates any kind of kin and comes from the verb ‘to relate’, in an attempt to discover the real meaning cleansed of any hint of biology and therefore of kinship. Of course all they found at the heart of the word was the formal notion of a ‘tie’ or ‘link’ between two components, henceforth connected by this tie, a notion that can apply to all manner of ties without permitting their distinction. For, as Holy reminds us,33 one can be connected in a thousand social ways – people can be close friends, come from the same town or the same country, etc. – and to distinguish between these different types of relatedness we need to reintroduce contexts, specific contents that will allow us to distinguish, for example, the relatedness of friendship from the relatedness of ethnicity from the relatedness of citizenship. Eliminating all reference to the reproduction of life in analyzing the domain of kinship, and glossing the etymology and the semantic field of the word ‘relative’, do not lead to any positive deconstruction of kinship theories, but only to their dissolution in a sea of formal discourses that provide no hold on the realities.

In reality it is impossible to grasp the domain of kinship relations if one completely separates kinship relations between the sexes from the reproduction of life. The recent appearance in Western Europe and North America of same-sex families, and the legal and ethical debates this continues to arouse, will not contradict us here. For with the appearance of gay or lesbian couples the question is still one of gender, and, with the adoption of children or insemination, it is very much that of the transformation of a couple into a family. Sex, gender, couple and family: we are once more immersed in the universe of kinship.

It is impossible deliberately to disregard, even more so to deny, that – whatever role a society assigns to the man or the woman in making babies, whatever the rule followed by this society to determine who the child will belong to after birth, what groups of adults will have rights and duties with regard to the child (which are liable not to be identical for the father’s or fathers’ and the mother’s or mothers’ sides) – in all known societies there are rules (associated with representations and value judgements) that define the conditions under which unions between individuals of the opposite sex (and today of the same sex) will be socially recognized and which set out in advance the social identity of any children that may be attributed to these couples either by birth, by adoption (with or without insemination) or by other means. And it is by putting these rules into practice – or into action if one prefers – that people produce, between themselves and between the social groups to which they belong by birth (families, houses, clans, lineages, etc.), the specific social relations that are precisely kinship relations. This is true of all known kinship systems, including that of the Na of Yunnan, where the unions between a woman and the men who visit her at night (or a man and the women he visits at night) almost never result in a ‘marriage’ and therefore in an ‘alliance’ between two kin groups, two houses. For in this case, too, the union of the sexes and the status of the children are defined by society, since the sexual permissiveness of the adults is countered by a total taboo on sexual relations within the houses; and within the resident matriline, there is a taboo, on pain of death, on sexual relations between a brother and his sisters, an uncle and his nieces, an aunt and her nephews, and of course between a mother and her sons. The children born of the men’s nocturnal visits to other houses belong exclusively to their mother’s house and are raised by all of the members – male and female – of their matriline.

It must also be recalled that all societies, even though they may believe that certain births are not the result of a union between a man and a woman but stem from the union of a woman with a spirit or a god, recognize that there is a connection between sexual intercourse between two people of opposite sex and the birth of a new human being. Nevertheless, recognizing this does not necessarily mean that the man and the woman who unite are perceived as the genitors of the children born to them. In many matrilineal societies, for example, the ‘father’, in other words the mother’s ‘husband’, is not recognized as being the ‘genitor’ of the children his wife bears. In the Trobriand Islands, as we will see, a woman is not pregnant because she has been fertilized by her husband’s sperm. She becomes pregnant when the spirit of a male or a female ancestor of her clan wishes to come back to live with his or her people and leaves Tama, the island of the dead, in the form of a spirit-child who floats over the water to Kiriwina and enters the woman’s body. The spirit-child then mingles with the woman’s menstrual blood, which coagulates and becomes a foetus. When the woman tells her husband she is pregnant, he multiplies his sexual relations with her to nourish the foetus in her womb and shape it to resemble him. The patrilineal Baruya do not consider the woman to be the genetrix of her children. It is the man who makes the foetus in her womb from his sperm, and it is the Sun that makes the foetus into a human child. The woman’s womb is a ‘netbag’ where the actions of the husband and the Sun, the child’s only genitors, work together.

In short, in the Trobriand Islands as among the Baruya, people are perfectly capable of detailing their genealogy over several generations – moreover they have numerous reasons to keep it in mind. But reconstructing their genealogy with them by no means implies, as Schneider claims, that the anthropologist projects Western cultural presuppositions onto the ties being described, namely: that through sexual intercourse a man and a woman become the genitors of the children that will be born to their couple, that children are of the ‘same blood’ as their parents, that ‘blood is thicker than water’ and so on. If a married Trobriander is not the genitor of his children, who in all events will not belong to his clan but to that of their mother and their mother’s brother, he is therefore not a ‘father’ in the Western sense of the term; and nothing obliges an anthropologist from the West to project his own cultural representations when he asks people he knows to tell them their genealogy. In fact he is prevented from doing so by what he knows of their culture. The Trobriand example is a particularly convincing demonstration. How can one project the Western notion of consanguinity onto a society where, as Annette Weiner showed much more clearly than Malinowski had a half century earlier, the blood (dala) that flows in the veins of all members of a clan is always the ‘same’ blood, which comes from the blood of the founding ancestress of the clan who one day emerged, alone or with a brother, from a hole in the ground? And the very word that designates that blood also means ‘clan’, since every child that comes out of a woman’s womb is made from its mother’s menstrual blood, which the spirit of a clan ancestor in search of reincarnation came to coagulate and inhabit.

Let us travel back, to the Classical China of the Chu. At this time, society was ruled by a warrior nobility, by kings, princes and dukes organized into big patrilineal clans (tsu). These clans encompassed all descendants of a common ancestor, with whom they shared not the same blood but the same breath (ch’i). This breath bound them together into a community of feeling (kan-tung), which they shared with the ancestor from whom they were never separated and whom they regularly worshipped at his tomb (tsung). Even if their bodies were separate, the breath that gave life to a father and his sons, and to the sons together, was always the same breath. The Chinese notion of agnation thus has nothing to do with the idea of ‘one blood’, which the West inherited from the Romans. We also see why the terms used to designate the elder lineages (ta-tsung) or the younger lineages (hsiao-tsung) that make up a clan are constructed with reference to the descendants’ obligation to make offerings on the ancestor’s tomb (tsung) and to venerate him down through the generations.34 This viewpoint ignores the woman’s reproductive capacities. Life here can be perpetuated and extended only through the breath (ch’i) that dwells in men.

The examples of the Trobriand Islanders and of the ancient Chinese nobility point up another aspect of kinship relations. If, in the Trobriand Islands, it is the same blood that unites a mother and her children, brothers and sisters, uncles and aunts on the mother’s side to their nephews and nieces, and connects them all to the founding clan ancestress, and if, in Chu China, it is the same breath that unites a father and his sons and brothers, and connects them to the founding clan ancestor who has a separate body but the same breath, it means that the birth of a child in each of these societies is the result of a gift of essential vital principles stemming from invisible but ever-present and ever-acting ancestors who are transferred to the new human being taking shape through the agency of either men (the fathers) or women (the mothers). We must therefore take careful note that these vital principles are not made by these fathers or mothers, who are merely the chosen, favoured vehicles of their transmission. What is essential is that individuals think of themselves and experience themselves as not separated by part of themselves from certain ancestors.

But there is nothing exceptional about these examples. They simply place us in a particularly clear manner in the presence of the fact that, in all known societies, the making and birth of a baby are the outcome of a series of gifts that human beings, but also ancestors, spirits or gods, have made in order to assemble and unify the components of a new human being. Whether these components appear in the form of what we in the West call bodily substances (blood, bones, flesh, etc.) or of less visible realities (breath, soul, spirit – which, moreover, many societies do not see as totally immaterial realities, since souls are liable to reappear after death), for the child all of these are gifts that beings, which can be human and/or non-human, close or remote, have made so that it can be born. For a child (and for its parents) this has two consequences, which take various forms depending on the society. The child will be led to conceive of himself (or herself) (and even to feel) as being identical to or resembling all those of whom he (or she) will learn that he shares some component of his being, by virtue of which he is ‘the same’ as they. Hence the community of breath and feeling invoked by ancient Chinese authors. Second, as soon as adults have taught the child that he owes his existence and the elements of his being to a certain number of visible or invisible human and non-human donors who have made him what he is, the child will find himself indebted to them for his life and therefore under obligation.

But this feeling of obligation can die out or be called into question if the individual discovers or imagines that the human or non-human beings to whom he believes he owes his birth and being do not act as they should toward him; furthermore there often exists the idea that a child’s ‘real’ parents are not the people who gave birth to him but those who raised him and behaved toward him as parents should. One must also be careful about thinking that in all societies a person’s identity is constructed from what they received from others at birth. Most societies consider that a person’s identity is constructed over their lifetime, both by what they receive from others and which becomes incorporated in them and by what detaches itself from them and becomes incorporated in others. Some societies even consider that what one receives from others after birth counts as much, if not more, in creating kinship ties between oneself and others as what one received from one’s parents at birth. This is the case for the Baining of New Britain.35 As the Anglo-Saxons say, nurture is often stronger than nature.36

In reality, the ancestors’ strength and identity are not always transmitted solely through blood, breath, sperm or the soul, which borrow the medium of sexual relations in order to mingle and act. In numerous Melanesian societies, the ancestors and their powers are present in the ground they cleared. It was their flesh that fattened the land, their bones placed in trees that attract game and keep away evil spirits. These beliefs and rites explain why an adopted lost child or an outsider taken in after an unfortunate war can gradually become a relative who can legitimately claim the same ancestors. By working the same land year after year, by sharing and eating the same products of this land impregnated with the ancestors’ subsistence and presence, the outsider acquires a new body and with it a new social identity. He becomes one more descendant of the ancestors of the clan that took him in, a descendant whose identity was not transmitted to him through sexual intercourse, a descendant without genealogical ties, direct or indirect, with the other members of his clan.

And yet his age, his sex, the position he will occupy in the kin network of those who took him in and fed him will make the outsider a kinsman – a son for some, an uncle or a younger brother for others, a brother-in-law, and so on. In a word, the kin terms used by others to designate or address him, and to which he responds using the appropriate terms, act in his case as categories devoid of genealogical content but which nevertheless assign him a symbolic genealogical position in the pre-existing network of kin ties between the members of the society that is now his. What is exceptional in this case – the fact that an individual finds himself related to others without having any genealogical connection with them – is frequent in societies divided into sections (or matrimonial classes), as are a large number of Australian and a few Amazonian societies (the Pano, for example). Belonging by birth or adoption to a particular section gives an individual specific kin ties with all other members of his society, whether they belong to his section or to the three others and even though he has no genealogical ties with most of them.

Let us take a simple example. Imagine a society divided into four sections, of the so-called Kareira type. A man born to section A must marry a woman from section B, and his sister will marry a man from section B. The children of the couple AB belong to section C. The children will in turn marry members of section D, and their children will belong to section A, like their grandfather. For a man from A, all women in B are potential wives, whereas only a few of them will actually be. All children born to women in section B will be A’s sons and daughters, whereas he is not their father. We thus see that, in this system, kinship relations coincide only partially with real genealogical ties. But we also see that this does not imply that kinship relations are purely abstract social relationships that have nothing to do with sexuality and the reproduction of life. In reality, kinship relations, in so far as their abstract categorial content is concerned, are structured by a double reference: descent rules and alliance rules, which back onto the prohibition of incest between members of the same section, and the roles played by the men and women of each section in the initiation ceremonies and the rites ensuring the reproduction of plants and animals, and human beings.37

It is because kinship relations refer to ties between individuals who provide or have provided members of the new generation with the components of their life, in other words of their physical existence and/or their social identity as new human beings, and because the individuals have neither negotiated nor produced these components themselves but have received them as a gift, that these ties constitute a separate domain. This domain is initially dominated by the values of gift and debt, by the existence of rights that have not been acquired by the individual himself but with which he finds himself vested by virtue of relations with certain other individuals who appear as his father(s) and mother(s), his brothers or his sisters, his cousins or his affines, and so on. For, in most societies, there is no such thing as descent without alliance, and the exercise of kinship consists in transforming affines into consanguines and ensuring that, after a certain lapse of time, distant consanguines become potential affines.

In short, kinship ties, whether with descendants or with affines, are inseparable from the relations humans must produce with each other so that new generations of men and women may come into being and human life go on, which by no means implies that humans think they are the only ones involved. But I already hear the objections of those who claim that kinship is a purely social relationship, that there is nothing in its content that refers to sex or gender and to the biological process of reproducing life. Every anthropologist faces these objections and must address them.

PARENTS BY ADOPTION

Let us therefore take the case of adoption, that is, of the creation of a non-engendered line of descent. First of all we must distinguish between adoption and fosterage.

Adoption implies the definitive replacement of the ascendants by the adopting parents. Fosterage is merely the momentary replacement of the ascendants by designated guardians. Adoption entails a change of identity for the child, whether this is deep-seated or not. Fosterage offers a means of preserving the child’s identity and social status.38

It is readily understandable that not just anyone can give his or her child up for adoption, and not just anyone can adopt a child. In ancient Rome, a woman could not adopt a child. Only a male Roman citizen could. But a man was not allowed to adopt a man older than he. There had to be a sufficient age difference, according to ‘nature’, that is to say, which corresponded to the average age at which a man is able to engender children. And yet in Rome eunuchs and men recognized as being impotent could adopt. In the Middle Ages, eunuchs could not legally adopt a child, while the adopter had to be older than the adoptee by a margin that imitated nature. Even in the legal fictions of Antiquity and the Middle Ages, then, where adoption is conceived of as an abstract and not a biological filiation, the reference to engendering and biological filiation is present.39

Adoption,40 when not practised between two sisters, as in Polynesia where two sisters give each other children who were therefore already related to their new parents, makes children (and sometimes even adults, as in Rome) from outside the family into descendants who theoretically enjoy the same attention, the same rights as the other children (if they exist) engendered by the adoptive parents. More generally, adopted children are supposed to enjoy the same status as non-adopted children. So we see that the standard status accorded adopted children does not result from adoption itself but from the rules of conduct set out by the society in question for dealing with non-adopted children.

Moreover, unless they appeared by magic, adopted children were themselves engendered and, even if their social parents are absent, they still had genitors. All societies have various reasons for allowing or forbidding the transfer of children between adults and between social groups. And, in most cases, kinship by adoption is a complementary, second kinship. Nevertheless there are extreme cases in which adoption appears as a plus with respect to kinship by filiation, as among the Mbaya-Guaycuru, cited by Lévi-Strauss in Tristes Tropiques.

It was a society remarkably adverse to feelings that we consider as being natural. For instance, there was a strong dislike for procreation. Abortion and infanticide were almost the normal practice, so much so that perpetuation of the group was ensured by adoption rather than by breeding, and one of the chief aims of the warriors’ expeditions was the obtaining of children. It was estimated at the beginning of the nineteenth century, that barely 10 percent of the members of a certain Guaycuru group belonged to the original stock.41

When children were born and survived, they were not brought up by their parents but entrusted to another family, and visited only at rare intervals.42

Without going to this extreme, other societies testify to the importance that may be given to the adoption of captives. Such is the case of the Txicao of central Brazil studied by Patrick Menget.43 The Txicao are slash-and-burn agriculturalists, hunters and fishers who today live on the left bank of the Xingu River. Their social organization is based on a system of cognatic kindreds organized around uxorilocal residence. The Txicao practise two forms of adoption, one within the group and the other resulting from the capture of enemy children. The term for internal adoption is anumtxi, which means to ‘lift up’. The usual reason for adopting a child is the death of its mother. The child is then entrusted to one of its classificatory mothers to be raised. This form of adoption never gives rise to the definitive rupture of kinship relations with the paternal and maternal kin. The child enjoys all of these relationships concurrently. Another reason for adoption is the woman’s sterility. In this case, her husband’s brother or one of his sisters gives a child who will be adopted when it is weaned. This kin tie is not a substitute for filiation but rather a complement.

The adoption of a captured child is entirely another matter. To understand this practice, it must be seen in its cultural context. For the Txicao, death is not a natural occurrence but the effect of another’s malice. Following a series of deaths in a village, for instance, the villagers would organize a raid to punish the enemy group they suspected of causing these deaths by means of sorcery. The aim was both to punish these enemies by killing some of their numbers and to capture children to replace the dead Txicao. The adult prisoners were killed on the spot and their bodies used to make a series of trophies, which the warriors displayed upon returning to celebrate their victory.

The status of captured children was complex and they held both a central and an ambiguous position in Txicao social organization. The captured child is called an egu, a term that designates at the same time trophies taken from enemy corpses, house pets, the bamboo horns played at initiations or any member of an animal species as it relates to the spirit master of its species. The child’s education follows two rules: to root out the child’s ethnic origins by making fun of them and to reward the acquisition of Txicao qualities. Gradually, as the alien child acquired more and more Txicao features, he becomes a source of pride for his adoptive kin. As Patrick Menget writes, it is as though the adopted child became a superlative child, adding something to the group’s identity. When the adopted child grows up he or she becomes a favoured sexual partner for the Txicao, and as such the object of rivalries. Nevertheless, the union between a Txicao man and an adopted women is never a marriage in the strict sense of the term. The word used to designate this woman is still that used for family pets. But at the same time, captives are considered to be the principal source of proper names for the Txicao, and they thus contribute something of their alien identity to making new Txicao. This practice corresponds to the logic shared by many Amazonian societies, which are constantly on the lookout for elements in outsiders that could be used to constitute themselves, whether in the form of another’s flesh (cannibalism), symbolic components (trophies), disembodied entities (spirits of the dead) or finally identities connected with names. But the ambiguous status of the adopted prisoner, who is both more and less than a Txicao, shows that the ultimate reference is still that of the native-born Txicao.

Another, this time more serious, objection: there are a number of kinship systems, the Australian section systems, for instance, in which individuals from different sections are, with respect to each other, as fathers, sons, husbands or wives, without being linked, for the most part, by any real genealogical tie or alliance. Yet an individual will call the child of a woman, of a ‘wife’ whom he has never married, ‘son’ or ‘daughter’. This is an example of the ‘categorial’ character of many kin terms and explains the fact that individuals who do not stand in the same genealogical relation to Ego are classified under the same term (the term ‘cousin’, for example, designates both the son of a paternal uncle or a maternal aunt) and who may even not have any genealogical tie whatsoever with Ego.

We will deal later with this problem, which has continually enflamed debates between the partisans of the two opposing theses. For one side, from Malinowski to Scheffler by way of Murdock, these categorial terms result from the extension of the kin terms used to address the father, the mother, the brothers and the sisters who live alongside the child or to refer to them beyond the nuclear family circle (where children worldwide are supposed to be born and live their early childhood). According to this theory, there is a shift from the real father to the metaphorical fathers (the father’s brothers, who also are ‘fathers’) and from the primary relationships within the nuclear family to the secondary relationships created by the extension of the genealogical links between individuals and between their generations. For the other side, from Hocart to Leach and Dumont, this categorial classification is neither secondary nor derived from a genealogical classification but exists in and of itself, the class of ‘fathers’ being made up of a number of men in an equivalent relation to Ego one of whom is the mother’s husband. It is therefore not by extending the notion of father but by reducing the class to a single individual that one obtains the genealogical father. Extension? Reduction? We favour the reduction thesis, and this is because of the very argument Lounsbury and Scheffler use to combat it.

For what matters are the social values and statuses a society attaches to the fact of being a father, a son, etc., whether we are dealing with a classificatory or with a ‘genealogical’ father or mother. What matters are the attitudes, obligations and rights that the recognition of these relationships entails for the individual. And the more essential or vital components of their ‘flesh-and-blood’ existence or their social identity these relations appear to contribute or to have contributed, the stronger the bonds between the individuals.

A FAUX DÉBAT

We would like to show here that the quarrel between partisans of an exclusively genealogical (and in the last analysis biological) theory of kinship and those of a purely social theory (which in the last analysis excludes any reference to genealogical and biological ties) fuels a false debate based – on both sides – in part on real facts but used to ask bad questions. Whatever they may think, both sides have to explain what a father is and what is expected of fathers (or mothers, etc.) in a given culture and in a given kinship system. In each case they are confronted with the various ways societies have come up with to think and regulate the process of reproducing life, which means organizing the succession of the generations and the appropriation of the children born in each through the individuals of both sexes who claim to be their parents.

In short, the quarrel between extensionists and reductionists in no way calls the definition of kinship into question. Both sides miss the point. For however you become a relative – by birth or by adoption, by eating the same food or by living on the same piece of land, etc. – once you are, you have to behave like a member of the family, that is, treat the other members as though they had given you certain pieces of themselves – their blood or their sperm, their breath or their name, their spirit – elements that helped give you life and a social identity and which became part of you and which you can give in turn, can detach from yourself and give to others, who will in turn exist partly thanks to you.

Of course, as we will see later, these representations of the components (sperm, blood, breath, etc.) that are believed to pass from one individual to another and from one generation to the next are partly fictitious, and a society’s ethnobiological conceptions are not a matter of biology but of ideology. But beforehand, to show those features of kinship that cannot be explained by biology or ethnobiology, let us recall that it is not the biological ties that explain that two related individuals will (or will not) share the same residence, own (or not own) land together, intervene on the same side or on the other side of a political conflict: all these social behaviours are connected with kinship relations and mean that these are also social relations which can in no way be reduced to biological facts (even though kinship relations are largely underpinned by biology).

Whatever the role each culture ascribes to the man, the woman, the ancestors and the gods in making children, kinship relations attach divers social realities to the relationships that grow up between individuals owing to their (real or fictive) place in the reproduction process according to their sex and generation. Sex and sexuality are at the heart of kinship. And the fact that the taboo on incest, which concerns the exercise of sexuality, is a universal condition44 for the reproduction of kinship relations is a constant reminder of their existence and importance.

This detour having been completed, we can now resume our study of descent groups. To simplify the task, we will limit ourselves to groups, lineages and clans resulting from the application of a unilineal descent rule. These social groups are constructed with reference to ancestors, and therefore to a memory, and are based on a criterion that selects from among all the descendants of these male and female ancestors who share an identity those men and women who transmit it. This criterion can be the fact of coming from the same sperm (Baruya), from the same blood (Trobrianders) or from the same breath (Chinese). In most cases memory of the founding ancestors does not go back more than four or five generations. And if the name of the lineage founder is usually known by his descendants, the name of the most remote ancestor, from whom all of a clan’s lineages are believed to descend, is either unknown, legendary or mythic (a supernatural being). But even if the name is unknown, the existence of real ties between the lineage ancestor and the clan ancestor is posited. In some societies – Mandarin China as well as the aristocratic families of Tonga – a written record (ancestor tablets) or an oral account (recitation of genealogies) traces the ancestors back fifteen or more generations. Because they share the same ancestors and the same bodily and social identity, together the members of a lineage make up a sort of collective individual, a ‘corporate group’, as Meyer Fortes called it,45 following Maine46 and Max Weber,47 that is supposed to act as a single individual, which never dies, not only because its members are replaced by other members but because they own land, titles and rights, all of which must be kept and transmitted intact to the following generations.

We came across analogous phenomena when describing the ‘house’ system. Let us pause for a moment and examine the fact that all members of a lineage are seen as coming from the same sperm (the patrilineal Baruya) or from the blood of the same ancestress (the matrilineal Ashanti or Trobrianders). This blood, which is the same from one generation to the next and is transmitted solely by the women, is not a biological reality. It is a fiction. The concept does not stem from empirical knowledge about the body. It is a representation that makes it possible to exclude certain individuals (who are kin) from the formation and reproduction of the kin groups – lineages, clans – that as collective rights-holders play a role in reproducing the society as a whole. Blood is not merely a concept that makes it possible to determine the internal composition of a lineage and its boundaries, it is also a criterion that legitimizes this exclusion by referring to the presence of a vital component of a person’s identity, which some possess and others do not. Furthermore, it is a component that only some of those who possess it (women in the case of matrilineal systems) have the capacity to transmit to the next generation. A lineage, a matrilineal clan in the Trobriand Islands, thus appears as a kin group built on a twofold fiction – namely that blood is constitutive of an individual’s identity, it is his or her essential substance which only women can transmit, and therefore that the blood of all of the members of the matrilineage, men and women alike, is female blood.

‘Blood’ is therefore not simply the mental representation of an imaginary identity projected onto concrete individuals, it is a concept that gives social meaning to a vital, concrete component of these individuals, an element of their body which connects them simultaneously with their ancestors who made them this gift and to all those who received it in common. At the same time as their life and their identity, members of a lineage also receive the use, to be transmitted to the following generations, of lands, titles and functions, all of which are also gifts from the ancestors and remain the joint property of the group for as long as it exists.

KINSHIP BISECTED

It was on the basis of such facts, and using almost exclusively examples from Africa, that Meyer Fortes48 proposed to distinguish between two components of unilineal descent-group structure, one juridical-political and the other domestic. The first would correspond to the lineage and the clan, as the collective owner of the land and other means of production as well as of the political-ritual functions that give this corporate group a rank and an influence in reproducing the society (cf. the eight matrilineal clans of the Ashanti ‘kingdom’); the second would correspond to each of the families of the married members of the different lineages. In the matrilineal Ashanti society, married men live with their wives during the day, but in the evening they go home to their sisters and their mother, who live together under the authority of the oldest women of the matriline. The family is the unit in charge of bringing up the children. But it is also a unit of production and consumption. The relationships between individuals within the family are relations of bilateral filiation, but they are cross cut and marked by the matrilineal descent principle, which stipulates that a man does not transmit his goods or his functions to his own children but to those of his sister, and that his own children will inherit from their maternal uncle, their mother’s brother.

To account for the close ties between children in this matrilineal system and their father and the members of his lineage, Fortes proposed the notion of ‘complementary filiation’.49 He was strongly criticized by Leach, who felt that the notion concealed or reduced to purely personal and domestic ties between children, on the one hand, their father’s clan and family and, on the other hand, the relations of affinity between two lineages and two clans. As proof of the importance of the father and his lineage in the constitution of the Ashanti person, he cited the fact that, while all members of a matrilineage share the same maternal ancestral blood, called abusua, every child receives from its father and his lineage the spirit (atore) that animates his or her body.50 Leach’s criticism seems to me altogether founded.

But coming back to the distinction between the two spheres – juridical and domestic – that Fortes saw as being combined in kinship, Fortes has also been criticized for having projected a Western perspective onto the Ashanti and the Tallensi in so far as, until recently in the West, the political-juridical sphere was considered to be the preserve of men and the domestic sphere the preserve of women. In many societies however, there is no such clear-cut separation between these domains. For example, women, in their capacity as sisters, may actively participate in managing the patrimony shared by their lineage. Annette Weiner, discussing gender relations in Trobriand society, showed the essential role played by women as sisters of the deceased in the funeral rites, which are extremely important in this part of New Guinea. During these ceremonies, the sisters redistribute large quantities of female wealth in order to restore to their lineage’s patrimony the elements their brother had dispersed over his lifetime by making gifts, for example, to his sons, who in virtue of the matrilineal descent role do not belong to the same dala, the same ‘blood’, the same lineage as their father. The gifts presented by his sisters also make it possible to consolidate alliances threatened with extinction by this death. These facts had been ignored or perhaps deemed to be of little interest by Malinowski, which explains the charges of androcentrism Annette Weiner addressed to his work.

But it is not only in matrilineal societies that women in their role as sisters play an important part in managing lineage or clan resources and wealth. Among the Kako of Gabon, an agriculturalist–hunter-gatherer society with an Omaha-type patrilineal kinship system, the eldest sister of the head of the lineage, who is married in another lineage, intervenes throughout her life in her brothers’ management of lineage goods. She performs the rites that will ensure him success in hunting and war, but also abundant harvests. She has authority over his wife and can, by magical means, make her fertile or barren, thus depriving her brother of children. In short, although she left her original lineage when she married, she continues throughout her life to play a key role enjoyed by neither her brothers’ nor her nephew’s wives.

In Polynesia – Tonga but also Samoa – women as sisters rank socially higher than all their brothers, including their older brothers. They are in fact regarded as being closer to the ancestors and the gods, and play an essential role in the funeral rites. In Tonga, the most sacred, highest ranking person is not the Tu’i Tonga, but his sister, the Tu’i Tonga Fafine, to whom her brother, after having received the first fruits of the harvests from all the kainga (descent groups) in his kingdom, presents the best fruits together with other gifts, always of the finest quality. In short, we must not generalize a simplistic conception of gender relations and the possible forms and contents of male domination.

Another aspect of Fortes’ distinction calls for discussion as well: what did he mean by the political-juridical sphere within kinship relations? If one means the existence of relations of authority and responsibility within the kin group, lineage or clan, and the fact that the same persons represent the entire group to the rest of society, that they affirm and defend its rights in pieces of land, persons and functions, then we have only part of the political-juridical – or better, political-religious – relations that bring a society to exist as a whole, even if there is no central power capable of submitting the reproduction of the society as a whole to its will and power. We have seen this in Baruya society. Each lineage has collective rights in hunting territories, rivers and arable land. The lineage elders manage their distribution and use, and all male members of the lineage take up arms and call on their affines to defend their resources against those who, for instance, have cleared a garden in their forest without permission. But political-religious relations always extend beyond kinship relations and kin groups.

For example, each Baruya clan plays a specific role in the initiation rituals, and it takes the ritual and material cooperation of all clans to make a new generation of warriors that will defend not only their clan lands but the entire tribal territory. Or to take another example from ancient Rome, the pater familias at the head of his gens (a patrilineal descent group) had the right of life and death over his children and over the other members of the gens. He managed its resources. But he and his gens occupied particular positions in Rome’s political system. They belonged to the senatorial families or to the equites (knights). They were able to accede to or were excluded from given functions in the city. But all sons born to freemen, once accepted by their father and presented to the city magistrates, became Roman citizens (civis romanus) and enjoyed life-long privileges attached to this status wherever Rome exercised its power and its dominion.

In short, even in ‘segmentary’ tribal societies like the Nuer and the Bedouins, political relations are not to be confused with the relations of power or solidarity that spring up between clans or lineages in times of conflict. The well-known Arab saying, ‘Me against my brother; my brother and I against my cousins; I, my brother and my cousins against the world’, is not the last word on political relations in these societies. For even in the event of a scission between the kin groups that make up a tribe or a tribal confederation, the seceding groups associate with others, and the broadest resulting social group once again takes the overall form of a tribe. Furthermore, the Arab proverb is not universally valid. When a conflict arises between two Baruya lineages both of which have given wives to a third lineage (and have received wives in return), the third lineage splits into two groups, each of which goes to the aid of one of the feuding lineages. This practice can be seen as revealing of a concern to balance forces and avoid one of the lineages undergoing such a defeat that it is for all practical purposes doomed to disappear or to abandon its lands and seek asylum in a neighbouring tribe. But it is also the outcome of a marriage rule that obliges brothers to marry into different lineages and is the concrete proof that solidarity between members of a same lineage, of a same blood, comes second to each man’s solidarity with the lineage that gave him a wife. In this case marriage alliance plays a direct role in the way the descent reckoning operates. This is perhaps because marriage is based on the direct exchange between two men and two lineages of one of their sisters (or for fathers, of one of their daughters). The marriage alliance does not have the same importance when there is no direct exchange of women and a lineage exchanges wives for bridewealth, which used to be the most prevalent practice in Africa, but also in Asia and certain parts of Oceania.

These facts hardly concord with Fortes’ theoretical position on kinship. For above and beyond the question of the division of kinship into two domains, or the distinction between descent and complementary filiation, Fortes is probably the most eminent representative of those theoreticians of kinship for whom the essence of kinship resides in descent and not in alliance. He repeatedly declared that he was drawn to the study of institutions that were indispensable for ensuring the temporal continuity without which there can be no enduring society, and for this process to continue, he stressed, institutional forms of alliance are not essential.51 This places him in direct opposition to Lévi-Strauss, for whom kinship is first and foremost alliance. For Fortes, however, ‘the ways in which the reproductive cohabitation of men and women is regulated are of secondary concern’.52

Let us return to the problem of descent groups that persist over many generations. We have remarked on the very small number of rules used to generate these groups and which operate by manipulating the difference between the sexes (unilineal, duo- and bilineal systems) or by cancelling this difference (undifferentiated systems). Why have some societies chosen this principle rather than another? Meyer Fortes always refused to ask this kind of question, claiming that societies are not like individuals, who can over their lifetime ‘choose’, for example, the languages they want to speak. He is right on this last point. But one day we will have to come up with satisfactory answers to this type of question. Murdock53 and in his wake Goodenough54 in effect attempted to explain the appearance of matrilineal systems by change of residence, which in certain circumstances (importance of women’s gathering activities), they suggested, switched from virilocal to uxorilocal and entailed substitution of a matrilineal descent rule for the patrilineal rule that had previously applied. But these hypotheses proved unconvincing.

Jack Goody took another tack. Leaving to one side the unresolved and unresolvable question of why certain societies adopted a patri- or matri- or nonlineal descent rule for forming kin groups, and adopting Fortes’ idea that descent groups (whatever their principle of organization) are corporate groups, ‘moral persons’ who act as a ‘collective individual’, Goody turned to the raisons d’être of these collective subjects that traverse generations and transcend the life and death of their members. He thus advanced the hypothesis that the appearance of these groups had its basis in different forms of common ownership of the resources indispensible for the survival, reproduction and development of tribal societies – hunting and fishing grounds, lands cultivated using extensive agricultural or horticultural techniques.

However Fortes had already rejected Goody’s hypothesis in his own explanation of the formation of corporate descent groups in tribal society. ‘I maintain that it is a mistake to interpret the model of corporate descent groups to imply that productive or durable or any other form of property is the formative basis of corporate group structure in tribal society.’55 His reason for rejecting hypotheses based on the social importance of various forms of property was that ‘filiation and descent would probably be accepted as endogenous variables that are predominantly if not entirely independent of exogenous forces’.56 For him, the relations and institutions entailed in kinship have a sui generis, autonomous status in all societies.

RELATIONS OF DESCENT AS A SOURCE OF DUTY OF ALTRUISM

The formation of these corporate descent groups that transcend the existence of their physical members from one generation to the next would be explained by the fact that, wherever and whenever they are found, kinship relations link people personally, whether or not they love or even like each other. Meyer Fortes called this principle that sums up the obligations engendered by these ties, by their inherent strength, ‘the rule of prescriptive altruism’, ‘the axiom of Amity’. In short, he believed this duty of altruism between kin to be the psychological, moral and religious force that transformed a group of related individuals into a collective individual, a corporate group, which encompassed them all, present in each member while remaining distinct.

I do not deny that the universe of kinship is a place where solidarity, cooperation and sharing often prevail over competition, refusal to help others and egotism, if only because it is into the universe of kinship relations that children are usually born and that they survive only because they are cared for by adults who are related to them and for this reason feel and believe themselves to be obliged to do this. But the ‘duty-of-altruism’ explanation is too general and proves to be inadequate if one seeks to explain the formation of social groups that exist and develop by placing (and keeping) their material and immaterial resources in common and ensuring that each generation considers itself less as the owner of these resources than as a ‘steward’ whose first duty is to transmit them in turn to the following generations. One has only to remember that for hundreds of years, in both the East and the West, in both the Old and the New World, land, whether for growing or for hunting, was excluded from those things a person could buy or sell. There were fundamental reasons for this, which were not only moral or religious but also pragmatic and material. For by holding their material and immaterial resources in common, individuals and families gave themselves the means to survive and grow together, even in hard times. In this common holding and refusal to divide up resources (which by no means implies that they were exploited collectively) a material and social force was added to the moral force of the obligations that bound kin together.

I will conclude this point by voicing my agreement with Fortes that kinship relations are specific and that their evolution and their transformations lead to the formation of other types of kinship relations and not to something else, for example class relations. But transformations and evolution are not brought about by internal factors alone. While they are specific, in no case are kinship relations completely autonomous, but neither do they depend mechanically on the transformations and evolutions occurring elsewhere in society. Nevertheless these relations always entail stakes that do not stem from kinship. Before going into this second component of kinship – alliance relations – I would like to give what seems to me a particularly striking example of the material and social stakes entailed in the functioning and reproduction of kin groups, and at the same time an example of the way these groups manipulate the imaginary and symbolic stakes involved in kinship relations to serve their own interests. The example is that of child sacrifice in the Mandak and Barok societies of the South Pacific island of New Ireland.57 This sacrifice used to enable members of the child’s lineage to integrate the lineage of its father and therefore to change clans.

The Mandak and the Barok are matrilineal societies divided into two exogamous moieties each of which has a number of matriclans and matrilineages which own lands, rivers, sacred emergence sites and rights in the making of the carved cult objects known as malanggan. These goods were regarded as ‘food’ that each clan gave in abundance to each of its members and which were an extension of the food received from the mother while in the uterus. Each matriclan has an enclosure where the men of the clan sleep and a burial site for all of the clan members, both men and women. The clan, the male enclosure and the burial site are, in the Mandaks’ minds, a sort of vast maternal womb, which contains and nourishes its members and then receives them when they die.

In certain circumstances, when a matrilineage that is short of land wanted to acquire some of the clan territory of a man who had married one of the women of the lineage or when a woman cut off from her clan and living on her husband’s lands was widowed and sought to ensure herself and her descendants the right to go on living in the village she should have left after her husband’s death, in both cases the man’s youngest child was beaten to death at the funeral by his maternal uncle, ‘who then took the child’s body to the male enclosure of the paternal lineage and placed it in the tomb where the father’s corpse had been laid’.58 The sacrificed child was supposed to live a second imaginary and symbolic gestation in the same ‘maternal container’ that gave birth to its father. The child was thereby ‘reborn’ in death as a member of its paternal matrilineage and automatically became the owner of a usage right in the lands of this lineage. At the same time, this right extended to the child’s mother and to her descendants. Paradoxically, it was the son who came to belong to a new lineage after death and transmitted the belonging to his mother and brothers, thus re-engendering them, as it were, in a different lineage from that of their birth. One can measure the imaginary character of all of these transformations, but also the fundamental social importance in this society of a reassignment of kinship acquired in this instance at the cost of child sacrifice.

In other cases, instead of sacrificing one of its children, a lineage desirous of acquiring permanent rights in pieces of another clan’s territory would pay a very large compensation in pigs and other valuables. Wealth was substituted for a life, for life, which is the principle of all systems in which women are not exchanged for women or (which also happens) men for men.