Volume 71 1962 > Volume 71, No. 1 > Double descent on Yap, by David M. Schneider, p 1-24
                                                                                             Previous | Next   

- 1
DOUBLE DESCENT ON YAP

In this paper Dr. Schneider, Professor of Anthropology at the University of Chicago, explores the relationship between descent, kinship and politics on Yap, incidentally paying special attention to some of the structural problems concerning filiation and descent raised recently by Fortes, Goody, Leach and others.

The substance of this paper, in drastically abbreviated form, was delivered to the Tenth Pacific Science Congress at Honolulu, Hawaii (21st August to 6th September, 1961).

I

IN A RECENT PAPER which offers a classification of double descent, Goody 1 draws a major distinction between those systems in which both matrilineal and patrilineal descent groups are property holding corporations and those in which only one of them is a property holding corporation. This latter he excludes from the category of full double descent, though he suggests dividing this group into two sub-types: those with named, and those with unnamed complementary descent units.

There is, as every one knows, an important distinction between a kinship group and a kinship system. Nevertheless, we tend to characterize a whole system in terms of the kind of group which is either dominant or most conspicuous in it. Thus we have been inclined to consider the Tallensi as a patrilineal system even though the complementary matrilineal descent unit is clearly evident within it. Similarly, Ashanti and Fanti have often, though by no means always, been treated as having matrilineal descent systems despite the evident presence of complementary patrilineal units.

Indeed, we have known for some time now that a kinship system as a whole does not often contain one and only one kind of kin group. It very often happens that patrilineal lineages are associated with the kindred (leaving aside the question of whether the kindred is a descent group), and it may turn out that they are also often associated with ambilineal units as well. Similarly, matrilineal lineages may be associated with the kindred, ambilineal units, patrilineal units and so on.

The permutations and combinations of the different kinds of kinship units which can be found in association with one another has not been fully explored and even a rough list of the possible or conceivable combinations is not available to us at this time.

If all or almost all kinship systems contain more than one kind of kin unit, and if many of these contain more than one kind of descent unit, then what has been called “double descent” must be treated as a - 2 special case of this very general phenomenon. It would be wrong, therefore, to see double descent as a rare, unusual or in some sense “special” category of kinship system on the ground that it contains two different kinds of kin unit.

The problem of double descent is that of explaining why this particular combination of kin units occurs within the same system. It is not the problem of explaining why all, or almost all, kinship systems have combinations of different kinds of kin units.

Confusion may arise through the failure to distinguish between a descent system as a part of a kinship system, and kin groups which are not descent groups. A society with two or more types of kin group may be one in which a patrilineal lineage is found with the kindred, or it may be one in which a patrilineal lineage is found with an ambilineal unit, or it may be one in which a patrilineal lineage is found with a matrilineal unit. Since the kindred is not a descent unit in Rivers' 2 or Fortes' 3 sense, it may be that in the first two cases noted, the descent system is patrilineal, though the kinship system something else again, and only in the third case can one speak of the descent system as being double unilineal.

This is only to raise the question of the utility of designating the descent system apart from the total kinship system for problems such as this. For it seems apparent that the separate treatment of the descent system quite apart from the kinship system as a whole is not, at the present state of our enquiries, justified by the analytic results so far obtained. It would seem more sensible to treat the total kinship system as the unit for analysis and to ask how the different combinations of kinship units to be found within it are related to each other and to the whole.

Goody's segregation of full double descent from systems in which complementary groups occur without being property holding corporations therefore makes very good sense. For this segregation provides a useful distinction in terms of which whole kinship systems may be initially classified.

This paper narrows the problem to that particular kind of kinship system in which the complementary kin unit is not a property holding unit, and in which it is a unilineal descent unit—in this case a matrilineal clan. The aim is to focus the analysis of one such kinship system on the role of the complementary descent unit.

II

The island of Yap in the West Caroline Islands, Micronesia, has around 132 villages organized into eight aboriginal districts (ten municipalities now) which are in turn grouped into three shifting alliances. 4

- 3

Each village, of whatever class or caste, is divided into from two to five ranked, geographically distinct parts, and each village has functionally differentiated features within it such as three or more sacred places, a menstrual area, a place of play, a place for newly matured women, a village centre, a young men's house or houses, and at least one and often two old men's houses.

Each village has a village chief, who is the chief of the highest ranking part of the village, a chief of the young men, chiefs of each young men's house, messengers, specialists at deep-sea fishing, and so on. In turn, each sacred place has its priest, and each sacred place and priest are specialized—one for war, one for the sea, one for the clouds and fruitbearing vegetation, and so on.

A village is not a kinship unit. The land of the village is divided into a large number of plots, and a parcel of such plots—an estate—is held by a patrilineal lineage. The patrilineages are shallow and are not related by patrilineal ties to any other patrilineage either in the village or outside it. Put in other terms, each land holding patrilineal lineage is the entire lineage and not merely a minimal lineage linked to other minimal lineages to form units of larger scope. Hence the village is an aggregate of unrelated, independent patrilineal lineages, each of which holds land within the village boundary, but each of which is related to every other by ties of common village affiliation and not by ties of patrilineal kinship. (The ties between lineages established by marriage, as well as ties of matrilateral filiation are, as will be seen below, of minor structural importance.)

Patrilineages are very shallow, on Yap, and therefore patrilineal ties are very narrow in their range. When fission occurs, the separated units break all kinship ties with each other and suppress their links just as soon as possible. This is partly related to the fact that fission only occurs rarely and under conditions of extreme bitterness. Partly too, such links might constitute claims on land which the holders are quite anxious to terminate. The narrower the range of those who have claims on land, the better the Yaps like it. It follows, therefore, that fission as a process does not function so as to interrelate lineages within a village, but instead serves to maintain the shallow depth, narrow span and independence of lineages from each other within the village. This in turn tends to inhibit any tendency for the village to become a kin group.

The relationship between land holding lineages within the village is not phrased in the idiom of kinship. Rather, the relationship between lineages within a village is the relationship of common village membership, common class and common caste, as well as common interest of a more general nature, but emphatically not that of common kinship.

The highest ranking plot of land in each village is the village chief, and the eldest man of the lineage which (in the terms in which this is formulated on Yap) is held by that land is its spokesman for as long as he is alive. Hence access to the political and religious offices of the village is through membership in particular patrilineal lineages which hold particular estates. The effective internal government of the village—as distinct from its external relations—consists of the - 4 assembled heads of all patrilineages in the village. This is conceived of as a group of equals with certain members having primarily representational roles in the village's dealings with the outside world. 5

The village is thus not a kin group in any sense, and the importance of the patrilineal descent group in the political organization of the village lies in two closely interwoven facts. First, each lineage has a rank within the village which derives from the rank of the land it holds, and political (including religious) offices within the village are held by certain patrilineal lineages as a function of the rank of the land which they hold. Second, each lineage is a single indivisable unit in the governing body of the village and is represented on this body by its head. That is, the lineage is the elementary political unit of the village.

The claims of patrilineal kinship are in essence claims on land, and claims on land are in turn made up of two major elements: first, they are claims on food and resources; second, they are claims on rank and privilege within the political system of the village and in turn within the district and alliance. Hence the narrow range of the patrilineage helps to maintain the elaborate and extreme rank distinctions of both the village and the national political organization by keeping at a minimum the number of persons with equal claims on a particular piece of food-giving or rank-giving land. These are, I think, the factors which have tended to maintain the Yap patrilineage as one of shallow generational depth and hence small membership.

Each patrilineage has its own pool of ancestral ghosts to which it alone has access through the head of the patrilineage. Pools of patrilineal ghosts do not, of course, overlap, for a ghost, like a person, is either in a given patrilineage or he is not. The ghosts of women who were mothers join the pools of their children (not the pools of ghosts where they were born), for it is a child, or child's child, who prays to the ghost of a woman. In Durkehim's sense, then, each pool of ancestral ghosts defines its own patrilineage as its church, in which the patrilineage itself worships, and joins in worship with no one else.

Yet the ghosts themselves only have the power to intervene, not to actually accomplish much. Supernaturals who are unrelated to each other or to the living by bonds of kinship are the prime movers on Yap. The ancestral ghosts' power lies in being able to influence and affect these supernatural figures of greater generality. Thus if a problem arises within a patrilineage that requires supernatural intervention, the head of the patrilineage divines which ancestral ghost is happy, and then prays to that happy ancestral ghost to intervene on behalf of the lineage with the appropriate supernatural figure.

Thus each patrilineage is separated from every other patrilineage by the unique composition of its pool of ancestral ghosts. So too is each village isolated from every other village by virtue of its distinct lines of access to the supernaturals. It is true that all villages have certain supernaturals in common, but it is the route to these supernaturals which is distinctive. Without access to the supernaturals through the ancestral ghosts, the supernaturals are neutral and disinterested and - 5 of no use. The link through the ancestral ghosts is therefore crucial for effective action, and it is this link which is distinctive to each lineage and to each village, and it is this unique link which in turn effectively isolates each lineage from every other and each village from every other.

Ranking, as is evident by now, is inherent in almost every aspect of Yap social organization. Each lineage within a village has certain rank with respect to the other lineages, each village has a rank in the class and caste system, and so too each lineage is internally differentiated in terms of the rank of its members.

Within the lineage, rank is largely a function of age-sex category. The basic rank categories are those of old man, old woman, young man, and young woman along with children. Each of these categories must take its food from plots of land specially reserved to it, must cook that food over fires specially reserved to it and in pots of its own. The sharing of food from the same pot, the same fire or the same plots of land across these lines is prohibited.

Nevertheless the estate as a whole, regardless of its internal differentiation, is jointly owned by the lineage and no part of it can be transferred without the consent of its entire membership. Such transfers are extremely rare.

There are no equals in a patrilineage, and all relationships are marked by modes of respect for superior from subordinate. The eldest man of the lineage is its head, and his responsibility for the members is to represent them to other patrilineages within the village, to represent them to the ancestral ghosts and the supernatural, and to provide for their needs and necessities. A father is said to take care of his children when they are young, and they obey him and pay him respect in return. Later, when he gets old, they reciprocate by providing him with food and care while he is old. Elder brother is said to be “like a father” to his younger brother. He too enjoys the respect of his younger brother, provides the younger brother with authority and guidance, and, of course, brothers should help each other out when in need, just as any members of the patrilineage would do. Brother and sister have a mild avoidance relationship, with strongly sexual overtones. Father is superior to mother, mother to daughters and sons, and brother to sister.

All relationships within the patrilineage are phrased in terms of the delayed exchange of food and care and, as will be seen in a moment, this is the essence of the definition of patrilineage relationships.

Finally, one important point must be added. Prior to the German administration of Yap, the ideology was that coitus had no bearing on conception. Conception was the reward arranged by happy ancestral ghosts, who intervened with a particular spirit to bestow pregnancy on a deserving woman. The bond between father and child, therefore, had no biological content. Even in 1947 this ideology had not been gravely altered. Despite the knowledge imparted by Germans, Japanese, and Americans, the official line on this matter had not been altered in any significant degree, partly because the Yaps themselves tended to take an attitude of indifference toward it. It was an interesting piece of information which might well be true, but it was irrelevant to any - 6 matters of significance on Yap and it was not integrated into the ideology of patrilineal relationships at the time I was there.

The formulation of patrilineage solidarity rests on the symbolic content of reciprocity in the basic necessities of life. The care which a father gives his child in the form of protection against outsiders, food from his land, fish from the sea, obliges the child to return that care when the father is older, and creates in the child an obligation to be obedient and respectful. The giving of the gift of food in the context of affectively interested care creates a father-child relationship. And this is amply demonstrated by the fact that an old man with no one to care for him can be brought food by any young man, and given care by any young man, and the young man thereby becomes his son and inherits the old man's land. The exchange of care and food are the nature of the father-child relationship.

I turn now to the matrilineal descent unit and its functions.

There are between 30 and 40 named, exogamous, totemic, matrilineal clans, called genung on Yap, each with its own founding ancestress (nik).

The matriclan is dispersed, so that members are found all over Yap regardless of class, caste or political affiliation. Despite the dispersed nature of the matriclan, certain areas are believed to have a higher concentration of clan members than others, and these areas are generally the areas associated with the believed origin of the clan. Thus the Waloi clan is said to have arisen on Map, and Map is said to have more members of this clan than other areas.

The matriclan never meets together as a whole, never acts as a unit, and with certain minor exceptions, holds no property jointly. Although the members of the same clan in any given political unit—such as a village or district—may, if they wish, turn for guidance or help to the oldest member of that clan in the area, there is no clan leadership, no clan chief, nor are there any political offices directly connected with the matrilineal clan. That certain of them have political functions will appear below.

A person has the right to ask a fellow clansman for help when such help is needed. This right is very loosely and very flexibly stated. The claim is the claim of matrilineal kinship, and constitutes the special claim that one kinsman has on another. This claim can be for sanctuary in case a man is being pursued, either because his village is at war or because he is the object of some particular person's wrath. A man can also claim hospitality when in another village, perhaps attending an exchange, or on some political business. Or a person can make claim on some clansman for labour or specialized help—perhaps the clansman is a canoe builder, or a magician whose special magic may help him. Where a man would have to reward a canoe builder, the reward to a clansman would be slight by comparison.

Where the patrilineage is in no sense a group related by biological ties, or ties believed to be biological, matriclan members are explicitly related through ties which are held to be of a biological nature. The biological relationship of clansmen is based on the belief in common descent from a particular founding ancestress, and the immediate - 7 connection with that ancestress is, of course, “demonstrably” through the mother. In fact, of course, matriclan members cannot all be biologically related. In part this is true since it is unlikely that a porpoise, rat, fungus, etc., ever had real, living, human offspring. But in part, too, adoption precludes actual biological relationship since the adopted girl transmits the food prohibitions and clan affiliation of its adoptive, not its true mother. Yet the importance of the belief that clansmen are biologically as well as socially related is illustrated by the fact that the adopted child has a dual clan affiliation and a dual set of food prohibitions, those of both the real and the adoptive mother. The totemic prohibitions and clan affiliation of the real mother terminate with the adopted child; only those of the adoptive mother are transmitted through her adopted daughters.

Each matrilineal clan has its founding ancestress and a place associated with the origin of that clan. Connected with each locality where the clan arose is a place which is sacred to that matriclan. In some instances, the sacred place to the clan, and the priesthood to that sacred place, are held by the patrilineage which holds the land. Such, for instance, is the case of the porpoise clan. The sacred place of this clan can be used by the priest to call porpoises, which are eaten by persons who are not members of the porpoise clan. This land is held by a particular patrilineage, and the priest for this clan is often not and need not be a clan member. It may happen that he is a clan member, but this is fortuitous and has no bearing on his functions.

Certain clans are closely related in myth and hold very important sacred places. These sacred places are held by the clan and only clan members can be priests. These are not “linked clans” in the sense that marriage restrictions obtain between them. They are linked, however, in that they form part of a reasonably well integrated system of related beliefs. For these seven clans derive from one woman and her children, who are believed to be the original founders of Yap and from whom all—except those of other clans—are descended.

These seven clans' sacred places are scattered over the Island, and of the seven, four are particularly powerful and important. One of these four, Arib, is on the border of Teb and Merur villages in the district of Tomil. This place is sacred to Margigi, the first mother, and her daughter, and is held by the Fanif clan. The priest of this place has certain powers which enable him to foretell political shifts and changes, and he has among his functions the obligation to sanction and give legitimacy to changes in political alignments at the national level. He divines the proper day on which arranged wars are to be held, and may divine the size of the various armies to be put in the field. But the welfare of all of Yap is his primary concern and this particularly powerful sacred place is the vehicle of his concerns. A second powerful and important sacred place, held by the Waloi clan, is Biluol, in the village of Gatchepar in Gagil. This office is unusual in that it is the only one which is also associated with the village, district and alliance chief, for the same person holds these offices. In addition, this sacred place is concerned with the island empire to the east, called ngek (east), and consists in the chain of islands which stretch almost as far as Truk. - 8 The ability to send typhoons as punishment for these islands, or to protect them against typhoons, is one of the most important problems to which this shrine is devoted. A third important clan sacred place, also held by the Waloi clan, is located on the island of Map, in Gagil district. A fourth is in the district of Rul.

The importance of these sacred places held by matriclans lies first in the fact that these are all part of a single cult organization which operates throughout the island. Second, two of these have very close and immediate connections with the otherwise unrelated political system. These are the sacred places in Tomil and Gagil. (The Gagil shrine is not exclusively devoted to the eastern islands by any means.) Third, the primary functions of these clan-held sacred places and of the cult of which they are a part, are integrative and beneficent. They are devoted to the welfare of the whole island irrespective of class, caste or political division, and even when they are closely linked to the political system, as in the case of the Tomil sacred place, their functions are of this integrative, beneficent order. 6

There is one further point about kinship on Yap which should be made and that concerns the nature of affinal ties. On Yap, affinal ties do not create an alliance between patrilineages, and patrilineages are not linked by marriage. There is only the network of bonds between the persons so linked. Ties between patrilineages within a village are, as I have indicated, minimized in any case, and affinal links are treated primarily as a problem for the daughter's husband and the wife's father, husband's mother and son's wife, the wife's siblings and husband's siblings, but not of lineage-to-lineage. If there is friction between lineages, the marriage breaks. The presence of offspring of a marriage links a set of consanguineals to the child, but not necessarily to each other. Such are the bonds created between mother's brother and sister's child, between father's sisters' children and mother's brothers' children, but these are not links between the corporate, property holding patrilineages; none of these relationships are defined as affinal.

Two relationships are of some interest. One is that between mother's brother and sister's child, which is supposed to be one of warmth and protection on the part of mother's brother. There are, however, no such rules as that of free access to mother's brother's property, stealing from mother's brother, seizing his things, and so on. Instead, the relationship is simply confined to that of warm support, advice, and such help and protection as the mother's brother may be called on to provide. The other relationship of interest is that of the father's sister and her children. The father's sister and her children (terminologically - 9 mother and father) have the right to dispossess ego from his land should he commit any serious offence against customary norms and they have the right to the products of his land for the year following his father's death. Ego must save the coconuts, taro, yams and so forth from his dead father's plots for the year following his father's death and give these to his father's sister. On the first anniversary of the father's death, at the drilyun a yam, a ceremonial gift of these things occurs which marks the last time this claim is honoured. The right to demand respect of ego, and the right to evict him should he commit some serious delict are not terminated. Thereafter, ego may use his father's plots as his own.

The presence of this pattern of relationships between ego and father's sister and ego and mother's brother is not surprising, for its association with patripotestality and patrilineal descent has long been noted. 7

The problematic aspects of this pattern of relationships centre on two general questions. The first is the question of the relationship between this pattern and the matrilineal clan on Yap. Is this pattern of relationships an aspect of matrilineal clanship or is it independent of matrilineal descent? The second question is the obverse of the first. Is this pattern of relationships an aspect of patrilineal lineage relationships? Do the rights of ego in his mother's brother derive from his mother's claims on her natal patrilineage? Are the rights of the father's sister to be regarded as special conditions which her patrilineal lineage membership gives rise to? Or are they distinct from relations of descent? These questions are prompted in part by the recent discussion of descent and filiation in the literature. 8

The father's sister and father's sister's children, are, of course, members of father's matrilineal clan (except as noted below), while they are not members of ego's matrilineal clan. Similarly, mother, mother's brother, and ego are all members of the same matrilineal clan. The question arises as to whether these relationships are aspects of the relations between matrilineal descent groups, or matrilineal clans, or whether these relations are independent of clanship and matrilineal descent.

These relationships are most emphatically irrelevant to matrilineal descent units or matrilineal clans in the Yap view. Informants are quite explicit on this score, saying that nik and genung (totem and clan) have nothing to do with these relationships. Instead, they explain the relationship of the father's sister and her children to ego as arising out of the fact that a woman belongs to the patrilineage of her birth and retains rights in it whether she marries or not, whether she has children or not, and that by virtue of her rights and her close connection with her children, the children are therefore “close” to her brother and the land which he uses. It is said that a woman is “born of the land” just as a man is, and that this is the essence of her right to make claim on her brother's food and land, against the claims of her brother's - 10 children. This claim is not seen as a simple, easy matter but rather as a point of extreme tension. The very explicit rules and regulations governing it, and the very severe—indeed, extreme—sanction which attaches to the breach of these rules testify to the tenseness inherent in the situation. The name for the father's sister and her children, and the name which characterizes the relationship is gidi ni m'fen, people who chase away, and quite clearly marks the relationship as one characterized by the threat of the ultimate sanction of the loss of land.

Two other questions might be asked to deal with the problem of whether—regardless of informants' statements—these relationships are aspects of matrilineal descent and matrilineal clan rights or not. One is to ask what happens when the father's sister is of a different matrilineal clan from the father. This situation can arise when the father and his sister have different mothers, and its clearest relevance occurs when ego's father has two or more sisters, one of whom is a member of father's matrilineal clan, and the others are not.

There is no formal distinction drawn between father's sister who is a clanswoman and father's sister who is not. Both are equally gidi ni m'fen and both have equal rights in the fruits of the father's land for the year after his death, and both have the right to dispossess ego. All father's sisters come to the first anniversary of his death to claim their share of the products of his land, and each is entitled to an equal share.

The second question is the obverse of the first. If father has no sister, or if father's sister has died, can a woman of father's clan substitute for a father's sister in the absence of one? The answer here is unequivocally in the negative. If father's sisters are all dead, and no father's sister's children are alive, or if he never had a sister, then father's father's sister's children are the next in line to the rights of the father's sister and may exercise them. Here again, the father's father's sister's children are not necessarily members of the same clan as the father, and their clan membership is deemed irrelevant to their role as gidi ni m'fen. It may happen, but purely by chance, that the father's father's sister's child is a member of the same clan as is father. But this fact is quite irrelevant.

If the actual clan membership of the father's sister is irrelevant, and if a clanswoman may not substitute for a deceased father's sister, and if the clan as a unit and matrilineal descent have no roles in the matter, and these are added to the firm assertions of informants that matrilineal descent is irrelevant, how are we to understand the roles of father's sister and mother's brother?

First, of course, as informants insist, the matter is related to the fact that she is “born of the land”, that is, on patrilineage membership. A woman born to her patrilineage retains membership in it whether or not she marries out and whether or not she bears children to other patrilineages. These rights are exercised by returning to her natal lineages on visits, to live in case of divorce or widowhood (where she chooses not to stay with her children), by consultation with her in the event of transfers of title to land, and by being the sole title holder if all other members of the lineage are dead. If she is the sole title holder she may permit her husband to act in the political or religious capacity - 11 which is part of that estate's property, but he does not hold title to it. Her children may then inherit the estate from her. But a woman's children can not inherit the estate or any part of it if there are living male members. The lineage is a perpetual corporation and a person gains membership in it through his father. He does not inherit rights in land since he gains them at birth or by adoption through his father. Only when a patrilineage consists in a woman or a group of women without living male lineage members can the children of those women “inherit” the estate.

Yet this does not account for the matter satisfactorily. Though a woman has rights in the estate as a whole, and though those rights include her right to live on and eat the products of that estate, her special claim is on the food from only that portion of the estate from which her brother took his food, and only on his death and only with respect to his children's behaviour. Her claims on her brother's son's children only become activated if her brother's son has no sisters or sister's children. Similarly, if the lineage contains men who are father's father's brother's sons' son to each other, and if these men are each members of distant nuclear families, a woman is gidi ni m‘fen to her brother's sons and “reservoir” gidi ni m‘fen to her cousin's sons. 9

A woman, therefore, has two distinct sets of claims. One set of claims concerns her rights in the estate as a whole and is based on her membership in the patrilineal lineage validated by the fact that she was “born of the land”. These claims consist at least in her right to a voice in title transfer decisions, her right to return to her natal land on divorce or widowhood, her right to hold the entire estate in her own name if there are no other living members of the patrilineage, and in her right to transfer that estate to her children if no male members of the patrilineage are alive and there is no prospect of any being born or adopted.

A second and quite distinct set of claims which a woman has are claims on her brother, and it is these claims which are summarized and symbolized in her position of gidi ni m‘fen to her brother's children. It is the brother-sister relationship which focuses the father's sister's - 12 rights narrowly with respect to her brother's children and not widely on all children of the lineage.

There is no need to go far into the special relationship between brother and sister on Yap. 10 Suffice it to say that the relationship is one which is specially close and intense, yet specially fraught with the hazards of incest and is thus one of mild avoidance and intense interest. A brother should avoid being alone with his sister, should not stand down wind of her where her odour might reach him, should never speak face to face with her. They must not share food after puberty, and a man does not share betel with his sister's husband precisely because the sister's husband shared betel, probably from the same sprig, with his wife. But a brother keeps a watchful eye on his sister and her protection is his responsibility. A brother hesitates to interfere in his sister's marital difficulties and hopes never to have to, for any interference is generally interpreted as an undue and probably incestuous interest in her. Yet a brother is a woman's last resort in trouble, and especially in marital trouble, and may very properly take violent action on behalf of his sister if this is clearly the last resort. It sometimes happens, though by no means often, that a woman lives in fear of her husband and will not dare to leave him even for the protection of her father's house. In such cases a brother must come to her protection, and properly does so with violence when forced to.

The claims of a sister on her brother's interest and protection are precisely symbolized by the foods which she could never share with her brother during his life time, but are hers for a year when he dies. And the claims of the man's own children clearly conflict with those of the sister for they are claims of precisely the same order—they are claims on food, care and protection. Children have respect for their father's sister, because her claims on their father compete, and conflict, with theirs. The year's food which the sister gets when her brother dies settles that claim and leaves to the man's children what is their essential right.

The mother's brother is of course in the opposite position with respect to his sister's children that the man's children are to their father's sister. There is no need to detail the reciprocal relationship here beyond pointing out that the claims of the sister's children on their mother's brother are without question aspects of the sister's claim on her brother—for warmth, protection, guidance and the special concern for well being. A man's relationship with his mother's brother is not beset with incestuous hazards, as a woman's relationship with her brother is, and although a woman's relationship with her mother's brother is as a child to an elder and hence not explicitly sexual in nature, it is implicitly so, and the warmth between them is especially great since it is not subject to the restraints of openly incestuous considerations.

The right of the gidi ni m'fen to dispossess ego from his land should he fail to show proper respect for them or should he commit some particularly grave breach of custom is the sanction to support and - 13 enforce the sister's claim on her brother against the claim of his children. This same sanction, it is worth noting, is available to the father if his child is disobedient or disrespectful. It does not seem possible to interpret this sanction as the father's sister's “submerged claim” on the patrilineal estate which becomes the “residual claim” of her children as Goody suggests. 11 This is a sanction, not a claim; it applies to the brother's child and not to any child of the lineage; it applies only to that portion of the land which the brother uses and not to the entire estate (unless these happen to be the same, which need not be the case). Goody's analysis of the “holder-heir” conflict applies precisely, as I have tried to show, to what Fortes calls “the domestic sphere”, 12 the relationships of filiation which on Yap are quite distinct from those of descent. The claim which is in conflict on Yap is the claim of the sister and the child for the protection and care of the same man; this is quite different from the claims of descent.

In sum, the relationship between ego, father's sister and her children, and mother's brother and his children, are quite clearly distinct from and independent of the relationships of matrilineal descent.

Second, this pattern of relationships on Yap, although consistent with patrilineal lineage relationships, is nevertheless equally distinct from and independent of these relationships as well. The obligations of the brother-sister relationship are clearly distinguished from the obligations of one lineage member to another. The rights of the father's sister to the products of the land used by her dead brother for the year after his death are symbols of the claims of a sister on a brother, they are not the claims of common lineage membership. A man's claim to the warm, beneficent support of his mother's brother does not give him rights to the land or property of the mother's brother's lineage. The right of the gidi ni m'fen to dispossess ego of his land should he show disrespect or commit some grave delict is the sanction which supports the sister's claim on her brother against the claims of her brother's children, and which supports the father's sister's children's claim on their mother's brother against the mother's brother's children. It cannot be interpreted as a “residual claim” arising out of relationships of descent, but is instead a claim embodied in relationships of filiation.

Third, this pattern of relations is distinct from relations of affinity which are, on Yap, of only minor structural relevance.

Finally, the bonds of filiation, like the bonds of affinity, do not function on Yap to unite lineages nor do they create alliances between lineages. 13

III

Thus far it is apparent that the Yap patrilineage and the matrilineal clan are two very different kinds of kin groups. The patrilineage is the property holding unit, of very narrow range and confined within a village where it is but one of a larger number of such units. The - 14 matrilineal clan has very wide range and is dispersed throughout the island, and holds no productive property unless the sacred places are so regarded. Both have important but different religious functions; the patrilineage is isolated from other lineages by the unique composition of its ancestral ghosts and the help of that group is limited to the lineage, but the clan's religious functions are national in scope and relevance. The benefits which the patrilineage ancestral ghosts try to achieve are limited to the lineage itself, or, if the lineage holds a village shrine, to the village as a whole. The benefits of the cult which binds the seven matrilineal clan shrines together are for the whole island, regardless of village, class, caste or district affiliation. Both have important but different political functions. The patrilineage is the basic unit of the village, holds all political offices of village, district and alliance, including the religious offices of the village. The matrilineal clan's political functions are different. Only in the case of the chieftainship of the district of Gagil are the offices of district chief and priest of the most important sacred place combined. The office of the priest of the sacred place in Tomil, on the other hand, co-ordinates political activities and integrates them by divining auspicious times and ways of settling political differences. Further, the claims of matrilineal clanship which transcend district or village lines are primarily claims to help, to sanctuary, hospitality and material help in the form of money or work. Where it is patrilineal lineages which fight one another, villages which war against one another, it is the ties of matriliny which provide some safety for the individual from these conflicts.

Finally, patrilineage and matrilineal clan differ in both the form and essential content of their structure. The patrilineal lineage is internally differentiated in terms of rank, and relationships are only those of superordinate to subordinate, relationships of rank and respect. The content of these relationships is stated in terms of interdependent reciprocity; the father feeds the child first and creates a debt; the child pays the debt by feeding the father when the child is grown. There is a gift and a counter gift, and it is the debt which is created by the first gift, which, in Yap, makes the return gift obligatory. This theme of gifts and return of equal magnitude is played out in a number of different ways on Yap; in the ceremonial exchanges between villages, the exchanges between patrilineages, the relations between districts, the relations between castes.

But the matrilineal clan is not ranked either among other clans or against any other unit. It is totally outside the rank system, and in a world which is so systematically devoted to the detailed elaboration of rank differences, this non-ranking feature of the matrilineal clans stands out in sharp relief. Neither are the members of a matrilineal clan ranked, and whatever their rank may be apart from common clanship, the claims of one member of a clan on another member are seen as just and legitimate and equal because they are both equal—as are all members—as members of the matrilineal clan. Neither does a gift in these terms create a debt.

The idiom of kinship, the symbols in terms of which the different kin groups are structured, is consistent with this difference. The Essence - 15 of patrilineal ties is the gift and the return, the care and the reciprocation. The essence of matrilineal clanship is biological equivalence; members are biologically related, and they are all equally biologically related to the ancestress as well as to each other. The one then is based on the concrete ties of dependence and interdependence, the other on the mystical tie of biological relatedness through the mother; the one is based on the sanctions which the father can employ to bring an errant son into line—by dispossessing him, by not giving him the gifts of food and protection—the other has no sanctions to enforce conformity. Clansmen help one another “because they are of the same belly”.

IV

The mere inventory of the differences between patrilineal lineage and matrilineal clan does not tell us much about why these two kinds of kin group are present in a society, nor why one is a complementary kin group and is not structurally similar to the other, which is a property holding corporation.

There is one more difference in function between patrilineal and matrilineal clan, and it is this difference which is probably crucial in accounting for the presence of the matrilineal clan and its organization as a complementary kin group. This is the role of matrilineal clanship in the total social structure and particularly with regard to the political system.

The political system on Yap is predicated on rank. Although rank is in part conceived of as a permanent, unalterable and inalienable attribute of the ranked element, it is in fact none of these and it is toward the problem of the alteration of unalterable rank that the whole political system is oriented.

As I have indicated, each village as a whole occupies a particular class position in a system of nine classes and seven rank strata. The first three ranks constitute the endogamous upper caste (pilung) while the lower four ranks contain the endogamous lower caste (pimilingai). The land of the lower caste villages is owned by patrilineages in the villages of first rank.

The very highest rank contains two different classes, and so does the second rank. This division, which has been called a division of war parties, operates largely within each district and works roughly as follows: The two first classes are called ulun and bilche, and these constitute the leaders of the baan e pagel and baan e pilung, the side of the young men and the side of the chiefs respectively. The second rank strata is also divided into two, metheban and tetheban, with one traditionally allied to the side of the young men, the other traditionally allied to the side of the chiefs. The third ranked class in the upper caste is dourchig, and a village of this class may be allied with the side of the chiefs or the side of the young men, depending on the last re-alignment of forces, or the last outcome of some dispute or war. The four classes of the lower caste are allied with the ulun or bilche village which contains the patrilineage which owns their land.

This division into “sides” may split a district, or may operate at the national level. It is said, for instance, that the job of dealing with - 16 foreigners, which first presented itself as the opportunity to seize a Spanish ship and divide its spoils, was given to the side of the young men, and all villages from that side participated most prominently in the earliest difficulties which the Spanish encountered in trying to gain a foothold on Yap.

Irrespective of the “war party” or “sides” division, however, groups of districts form into three major alliances, and along this axis there is again constant manoeuvring where the aim of each alliance is to dominate the other two.

The ultimate goal of political activity is to raise the rank of the concerned unit. The alliance's goal is to dominate the other two alliances, and this of course is the necessary condition to victory, for when any one alliance shows signs of gaining ascendency, the other two form a coalition against the first. Wars between alliances were arranged wars, and were on the whole bloodless and lacking in destructiveness. When a particular alliance felt that it had lined up sufficient strength within its own unit and had gained sufficient commitment from districts which were nominally committed to another alliance, it would call for an arranged war. The date and time were set in advance by the priest of the matriclan sacred place between the villages of Teb and Merur in Tomil, and on the appointed day the armed forces of each would show up. This date was the showdown, for at this time the districts which one alliance might count on would, finally, have to take sides, and an anticipated district might in the end not shift its allegiance, or an anticipated force from certain villages might not come through, and the mere show of the forces on the field of battle—sometimes on water, sometimes on land—would make clear what the line-up actually was. This show of force was the war, and the alliance which felt it was dominant might find itself insufficient to gain control, or it might actually gain control for a time. Then, on the settlement of that engagement, the whole process would begin all over.

A particular village could raise its rank through such inter-alliance operations, inter-district operations, and through inter-village rivalries. Although generally the district is the unit of allegiance in alliance politics, individual villages might sometimes find themselves aligned against their own district in such alliance politics. And some villages have raised their rank in such political engagements.

The village could raise its rank in a number of ways. In exchanges with other villages of equal rank, one village might take the lead and so become generally recognized as being superior to another village. Or the district chief could elevate a particular village because of its particularly valiant or crucial part in some inter-district rivalry or inter-alliance rivalry. Or a particular village might gain or lose rank by order of the district chief for either an infringement on the rights of another village—an offence against that village punished by losing rank—or by being the object of another village's offensive acts.

In general, the rank of a village is largely dependent on the decision of the district chief, who in turn may or may not—given the particular circumstances—be subordinate to the alliance chief. If the particular village strongly supports a particular district chief, and that district - 17 chief wins out in some political engagement, then the village might be rewarded for its loyalty. If, on the other hand, that district chief loses, the succeeding alliance chief or political group in the district might reduce a village's rank.

Low caste villages owe allegiance to the patrilineal lineage of the first ranking village which owns their land. The patrilineal lineages which own a low caste village put the services of that low caste village at the disposal of their village's political endeavours, particularly in times of war. Thus, through its strong support and bravery in defence of its high caste, owner village, a low caste village might be elevated in rank within its own caste. More rare, but certainly possible, is the elevation of a low caste village to the lowest rank of the high caste for deeds of great valour and the loyalty of its support.

Wars between districts and within districts occur as well. These might be quite violent, with the loss of many lives and the burning of many houses, and in such wars villages might gain or lose rank either as a direct result of the war or as a result of the loss of manpower it suffers in the war, which affect its strength and its resources.

Still another kind of war is the limited warfare which occurs when two villages fight in anger over the theft of a woman or some insult to their rank and standing. Such wars might be viewed with neutrality by other villages as none of their concern and without political relevance.

But there is usually a stale-mate implied in any balance of power at any given moment, so that the resort to warfare is by no means the usual course which politics takes. Any given village or district would usually have such an alignment of allies that any other village or district hesitates to test the strength of that alliance, for it might well turn out to be stronger than anticipated. It is by no means a simple process of mustering arms and mounting an attack to see who wins. Caution prevails partly at least because the political leaders can not depend on their followers to obey them, and if the body politic should feel that it does not stand a chance, it would resist the political leader's efforts to muster arms.

It is the village as a single unit which is the basic element in the political system at the national level. A district may be divided between the party of the chiefs or the party of the young men; a village may give its loyalty temporarily to another district which is in turn pitted against the village's own district; a village may throw its support to one or the other alliance under certain circumstances. But it is the village as a whole which operates as a single, irreducible element in the national political system.

A village's relations with the district, alliance, or any other village are narrowly and rigorously controlled and channeled through two and only two representational offices of the village; those of the village chief and the messengers. Since the messengers only act on direct instructions from the village chief, the representational functions are essentially in the hands of the village chief. No other person may formally represent the village in its relations with the outside world. In fact, of course, any village member is treated as if he represented his village when he is outside it, so that if a man is outside his own village and - 18 insults another village, or steals a woman, or attempts to, his whole village is held responsible for his action and his whole village takes responsibility for his action. But his actions are not, of course, formally aimed at specific political consequences, and he becomes a representational figure more by accident than political design. Yet the very fact that his action has political consequences—a war may ensue, he may be killed and his village have to retaliate, the name of his village is demeaned by others and so on—means that strict and firm control over the actions of unauthorized persons outside their villages must be exercised. The village chief and the village as a whole prefer that no person leave the village and precipitate difficulties outside it and so tend to enforce the confinement of the villagers within the village boundaries. On the other hand, relations between villages are such as to tend to discourage venturing into a village other than one's own. For one is simply not safe outside one's own village. It is said that prior to the German period no man in his right senses would spend the night outside his own village unless he were actively invading another village with a body of warriors. Villages are and were constantly on the alert to seize whatever advantage they could, and so intensely hostile to one another. Even in 1947 many of the people in the area in which I worked had never been to the southern part of the island, and some had never been outside their own district.

The isolation of each village as a single unit, related to others primarily through its representational figures, is (and was formerly more so than it is today), further isolated by another fact of some importance. And that is the prevalence of what might be called “dirty politics”. The use of fraud and deceit, false promises, slander, character assassination, treachery, lying, poisoning, sorcery, ambush, beating, and political action of this sort was not unusual and is not uncommon today. The prevalence of these as possibilities tends to discourage much trust between villages, tends to reduce the contact of the members of one village with the members of another, and tends, even, to make the formal, politically proper representational activities of even chiefs a matter of more than ordinary risk.

The isolation of each village from all other villages does not prevail without break or without relief. The system of ceremonial exchanges between villages creates a situation of general truce when there is considerable freedom of movement and considerable safety outside one's own village. These events bring large numbers of people into one village from all surrounding villages, and thus integrate diverse units partly through common participation in an event aimed at maintaining and perhaps elevating rank of the villages involved, and partly through the physical fact of bringing people into contact with one another. Yet even such events are not lacking in inter-village rivalry, in tenseness and in feelings of cautious distrust.

The total picture of the political system, then, is one in which sharply isolated units—villages—are organized into districts, alliances and sides and in any of these arrangements each village as an isolated, indivisable unit has a particular rank. Two countervailing pressures are constantly at work. One is the pressure of each village or district - 19 or alliance or side to improve its rank position. The other is to maintain the given rank system in status quo. Since it is ultimately the village's rank which is at issue, district and alliance and “side” alignments are alignments of larger units each element of which is oriented primarily to the end of raising its own rank. Hence these alignments are relatively unstable. Districts do not disband or re-align at a moment's notice, but the degree of solidarity within the district or the alliance is very variable.

The tendency to press for rank improvement and the tendency to press for status quo for all other units than one's own result in constant pulling and hauling, but generally a state of stalemate. And the state of stalemate in turn tends to help maintain each village as a unit whose commitment to any given alignment is shallow and uncertain at best. This in turn helps to maintain the general situation of distrust and uncertainty which is a facet of the isolation of each village in the total system.

The political system of Yap is such, then, that divisive and isolating factors are dominant. The hostility between villages and districts is a dominant feature of their relations. An index of this state of division and the fissive tendencies inherent in this structure is given by the very clear language dialect differences in different parts of the island and of the differences in minor customs and vocabulary. Such differences tend to be a function of the conditions of isolation that I have described.

V

The structure of the patrilineage and the political system are markedly similar; that of matrilineal clanship radically different.

The lineage is the land holding corporation; so is the village. The lineage is a unit which is treated by outsiders as internally undifferentiated; so is the village. The lineage is represented to outsiders by its head; so is the village, the district and the alliance. The head of the lineage is its highest ranking member, the oldest competent male; the head of the village is its highest ranking lineage for whom the eldest male is spokesman. The lineage can only act through its representational figure and his actions are only legitimate insofar as he has the consent and concurrence of his lineage, though he has ways and means at his disposal to enforce consent and concurrence by devices which are not always strictly proper; the same is true for the village and village chief, district and district chief, alliance and alliance chief. Internally, the lineage consists of units which are differentially ranked so that no two are equal; so are the parts of the village, villages within districts. And districts within alliances. The relationship between the differentially ranked units within the lineage is that of superordinate to subordinate, marked by respect, and hedged about with systematic symbols which maintain the difference between units (the food taboos, separate eating, etc.); so, too, villages within districts, districts within alliances. The role of the superordinate in the lineage is explicitly phrased as one in which the superordinate gives protection and care and managerial guidance to the subordinate, the subordinate is obliged by those gifts to - 20 be obedient, loyal, respectful and to return those same gifts when they are requested by the superordinate; so, too, the relationship between village chief and villagers, district chief and villages of the district, alliance chief and districts in the alliance. The exchange of food, ritually and ceremonially regulated, between superordinate and subordinate, marks the tie between adjacent ranks of the lineage; so, too, between villages, districts and alliances. Finally, the isolating and divisive tendencies between the units of the lineage are paralleled by the same tendencies in the village, district and alliance.

Yet the structure of the patrilineage and the political system have one significant difference. What, then, is the major difference between the lineage and the rest of the political structure? It cannot be the symbolic value of consanguinity, for biological relatedness as a sign of that value is missing from Yap. Patrilineal kinsmen are not biologically related since coitus is believed to be irrelevant to conception. Lineages are not different from villages or alliances because one is presumed to be a biologically related unit, the other not. None of these are presumed to be biologically related units. Instead, all are units in which the essential bond which maintains their unity is phrased as one of the gift and its return, and the gift is that of care and protection, in return for loyalty and obedience and the delayed return of care and protection from the subordinate.

The difference between the lineage and the village and all other political units stems from the fact that the lineage holds land which has two values; the land as food and the land as rank. For the lineage takes its livelihood from its land, and it takes its rank in the political unit of the village from the same land. But the village and district are less concerned with land as livelihood than they are with rank per se, and the rank of the village or district or alliance is less closely tied to the inherent and unchangeable rank of land than is the rank of the lineage.

The difference between the lineage and the rest of the political system, then, rests on the fact that the lineage is the corporation which holds land as livelihood and which exploits and manipulates that, while the political system holds rank as its scarce commodity and manipulates and exploits that. These two are articulated through the fact that lineage land also has rank.

It is notable that essentially the same kind of structure organizes these two very different scare commodities. And it is also notable how the one unit which is organized as a kin group can be structured almost identically with the other units which are not organized in terms of kinship principles.

Since the patrilineal lineage and the political system are structured in almost identical ways it is not surprising that they both exhibit the same kinds of strains and the same tendencies toward instability.

The nature of the instability lies essentially in the tendency of each unit to become isolated from all other units, to fail to relate to other units in terms of stable ties. This tendency arises from the heavy pressures put on each unit to keep to its ranked place, which coincides with the pressures from within each unit to improve its position in the rank system at the expense of other units; for the young man to get - 21 control of his land, to increase his independence of his father and minimize his subordination to him, and for the village to improve its rank in the district, etc.

The normative regulation of the means for improvement of rank is more difficult to maintain at the level of political action than it is with respect to the internal differentiation of the lineage. The maintainance of the norms governing proper political action is most difficult and the presence of so much “dirty politics” is a statement of this condition. For the Yap view is that “dirty politics” is expected, but not proper; understandable, but not ideal; “the way things are”, not “the way they should be”.

VI

The essence of matrilineal kinship on Yap is that of solidarity and loyalty unqualified by considerations of rank. The symbols of such solidarity and loyalty are those of common descent, expressed in the phrase “we are of the same belly”, the myths of the founding ancestresses, and the cult which ties the seven sacred places of all Yap together into a single unit.

Both in symbolic and in practical form the seven sacred places have as their primary responsibility the well-being of all of Yap without regard to rank. The sacred places associated with matrilineal kinship cannot help one at the expense of another, cannot advance one claim at the expense of another, but by their power produce undifferentiated solidarity by doing good. The major sacred place is set apart from, though related to, the political system.

The major values stated in and implicitly formulated by matrilineal kinship are those of undifferentiated solidarity. These values directly counterbalance the extreme fissive tendencies of the political system. On the other hand, they constitute a set of limiting conditions, though flexible limits to be sure, to the values around which the political system is organized. When the fissive tendencies go too far, they are checked by the unifying tendencies which derive from the values of matrilineal clanship.

On the other hand, at the level of concrete action, they add one very important element to the function of merely setting limiting values. When the concrete separation between groups proceeds so far that the basis for relationship between them diminishes to the vanishing point as far as the framework of political relationships are concerned, there always remain matrilineal bonds between chiefs, between villagers, between the concrete units thus split apart. These bonds then constitute the avenues along which reintegrative efforts can proceed guided by the values, and explicitly in terms of those values, of matrilineal kinship. There remains, that is, a socially workable and culturally sanctioned means of re-establishing relations between villages and districts no matter how far a particular political schism may have gone.

It is this element which accounts in part for the national unity of Yap, for the fact that it is not an island containing two or more relatively independent, warring societies, but rather an island with a single political system.

- 22

For the given actor, for the person in need, the claims of matrilineal clanship provide avenues toward supportive ends, but provide no means by which another can be injured directly. A man can go to a matrilineal clansman for help with a loan of valuables, for help in building a canoe, for sanctuary in time of war, for refuge away from home or hospitality if he is in need of it. For a political figure in need, matrilineal ties provide the same claims and the same avenues but for different purposes. Matrilineal ties are money in the bank for the politician who has no other way of approaching a possible ally, or for successfully concluding an alliance when other things might not quite permit it to go.

The crucial functions of matrilineal kinship on Yap are, therefore, essentially integrative. I believe that it could be argued that the political system, including the way in which lineages are organized, could not exist in their described form were it not for the operation of integrative mechanisms such as those of matrilineal kinship.

It should be noted that I have not argued that matrilineal kinship is the only such integrative device, although I do argue that it plays a crucial role. The system of ceremonial exchanges of food and valuables between villages has a similar function, and functions which are at the same time closely articulated with the prestige and rank systems.

The problem which I posed at the outset was that of accounting for these two kinds of kinship units as part of the same kinship system, and particularly the question of why one of them is a complementary unit rather than a property holding corporation.

The answer which I propose is simply that a complementary unilineal descent unit has primarily integrative functions; that when the political system—be it the kinship system itself or be it differentiated from kinship—is so organized that isolating and devisive tendencies are of sufficient magnitude to constitute a threat to the maintenance of the system, then complementary unilineal descent units can provide the integrative functions necessary to the maintenance of the system and that when they are present they do have such functions.

If the question is raised as to why the complementary descent unit is organized as a clan rather than as localized lineages, the answer is simply that it must be a clan if the devisive or fissive tendencies are “national” or society-wide in scope. Otherwise the complementary unit would not function at the points of relevant strain. On Yap the matrilineal descent system (the clans, their shrines, their myths and their activities) operate for Yap as a whole and operate at the same time at the level of the lineage, the village, and the district. Localized matrilineal units would only operate at the level of the lineage, yet some of the sharpest fissive tendencies occur at the national level.

Granting, then, that the matrilineal descent system has integrative functions of the sort I have described, and granting that the fissive tendencies in the total political system (including lineages as well as villages, districts and alliances) are such as have been indicated, the question arises as to whether certain of these integrative functions could only be performed by a kinship or descent system, or whether other kinds of mechanisms would work quite as well. To put the question - 23 in another way, one might ask whether the integrative functions of the ceremonial exchange system do not sufficiently counteract the fissive tendencies, and what special conditions can be shown to make a kinship based mechanism necessary in addition to a politically based system (the units in the ceremonial exchanges are political units, namely, villages).

I believe that the answer to this question lies in the symbolic value of kinship, and the affective implications of biologically defined kinship, as well as the very special position which the mother-child relationship can (but does not necessarily have to) assume as a symbol of solidarity.

The integrative functions of matrilineal kinship stem from the fact that these relationships are conceived as relationships of unqualified solidarity without regard to rank. A central element in that definition of the relationship, which contrasts sharply with its patrilineal counterpart, is that the claim on a matrilineal kinsman, or the matrilineal sacred place, or any matrilineal relationship is a claim which explicitly does not entail a stipulated kind of reciprocity. The man who claims sanctuary from a matrilineal clansman does not have to pay him back; the man who asks a clansman for help on his canoe does not have to formally pay him back with a stipulated return; the people who benefit from the intercession of a matrilineal sacred place do not have the obligation of returning a concrete gift. Yet a son becomes obliged by the father's care and must return that care; a village becomes obligated to return another village's ceremonial gift in equal quality and amount; a village must return the protection of the district chief by respect, loyalty, arms and valuables.

The essence of this relationship is structurally identical with that which is defined as proper for a Yap mother and her child. A child respects and obeys his father because of what the father does for him; a child loves his mother, on Yap, and a mother loves her child and what she does for the child she does without generating the formally defined and explicit obligation to reciprocate.

This relationship is symbolized by the biological bond which is believed to exist between mother and child, and of course, between all those who are related by the presumption of uterine ties. It is the unity of biological identity which requires no specifically stipulated kind of reciprocity.

The kind of solidarity which ties so conceived entail is thus of a different nature than those which entail stipulated forms and stipulated measures of obligatary return. Affectively, such solidarity is referred to the child's view of his mother's relationship to him and to such unconscious yearnings as every child, grown up, carries along with him—for complete, unqualified, interdependent warmth and love and care. And it is precisely the need for this kind of relationship which is most acutely generated by the hazards and the risks of the fundamentally untrustworthy bonds of political alliance.

Indeed, it might be suggested that on the one hand the whole political system is seen by the actors to consist in an attempt to control by force—force of prestige, force of alliance, the symbol “strength”—what cannot be obtained in any other way. On the other hand, the very - 24 attempt to control by force exacerbates the need which it attempts to satisfy.

It is, therefore, precisely kinship bonds so phrased and kinship bonds of such affective meaning to the actors, that are required as integrative mechanisms in a situation of the sort described. And it is for this reason, I suggest, that the integrative functions are performed by a descent unit and not by a further elaboration of contractually phrased reciprocities.

REFERENCES
  • FORTES, M., 1953. “The Structure of Unilineal Descent Groups.” American Anthropologist, 55:17-41.
  • — — 1959. “Descent Filiation and Affinity.” Man,
  • GOODY, J., 1959. “The Mother's Brother and the Sister's Son in West Africa.” Journal of the Royal Anthropological Institute, 89:61-88.
  • — — 1961. “The Classification of Double Descent Systems.” Current Anthropology, 2, 1:3-25.
  • LEACH, E. R., 1961. Rethinking Anthropology. London School of Economics Monographs in Social Anthropology, No. 22. London, The Athlone Press.
  • RADCLIFFE-BROWN, A. R., 1952. Structure and Function in Primitive Society. London, Cohen and West.
  • RIVERS, W. H. R., 1924. Social Organization.
  • SCHNEIDER, D. M., 1953. “Yap Kinship Terminology and Kin Groups.” American Anthropologist, 55:215-236.
  • — — 1957. “Political Organization, Supernatural Sanctions and the Punishment for Incest on Yap.” American Anthropologist, 59:791-800.
  • SCHNEIDER, David M. and GOUGH, Kathleen, 1961. Matrilineal Kinship. Berkeley and Los Angeles, University of California Press.
1   Goody 1961.
2   Rivers 1924.
3   Fortes 1959.
4   Field work on which this paper is based took place in 1947-8 and was sponsored by the Peabody Museum of Harvard University and the Co-ordinated Investigation of Micronesian Anthropology of the Pacific Science Board of the National Research Council.
5   Schneider 1957.
6   This account of the matrilineal clans is somewhat more abbreviated than I would wish. In part the details of the religious cult are of less significance than the fact that their functions are primarily beneficent in nature. In part, too, the very nature of matrilineal clanship is highly generalized and functionally diffused, centering as it does on help and aid and assistance, and the inventory of specific ways in which such help may be given lends unwarranted specificity to the general rule that clansmen help each other out, that clan shrines are for the public good, in whatever way they can be.
7   Radcliffe-Brown 1952:15-31.
8   Fortes 1959; Goody 1959; Leach 1961.
9   This is another instance of the way in which kinship relationships on Yap are divided into active and inactive categories so that a reservoir of inactive statuses is always at hand to keep a particular kinship status occupied with persons in an active relationship. Elsewhere (Schneider 1953) I have shown how, for example, Fa and FaBr are both classified terminologically as “father”, but that only if Fa is dead is FaBr referred to as “father” and only then does he play the role of “father”. Just as there is a reservoir of fathers, mothers, and so on, there is also a reservoir of gidi ni m‘fen. I must call attention here to an error in the 1953 publication which stated (Schneider 1953, table I, p. 225) that FaSiCh is in active relationship with ego as gidi ni m‘fen, while FaSi is in active relationship with ego but not as “mother”. The correct facts are that FaSi, FaSiCh, FaFaSiCh are all classed as gidi ni m‘fen, but it is FaSi, if she is alive, who is in active relationship with ego in that capacity. Lacking FaSi, FaSiCh is in active relationship with ego, and lacking FaSiCh, FaFaSiCh assumes the active role.
10   Schneider 1953 and 1957.
11   Goody 1959:81 ff.
12   Fortes 1959.
13   Fortes 1959; see also Leach 1961.