While the family contains essentially the two principles, male and female, present in procreation, in the physiological division of functio.:s and in sociological protection, the clan is based upon the elimination of either the paternal or the maternal element from relevant kinship. It is rather the clan of the relevant parent, plus the clan of the irrelevant parent, plus the other clans related to Ego by marriage or other forms of affinity, which together embrace the classificatory body of relatives. In fact the classificatory nomenclature always refers to the tribe or the community or a wider portion of it, and never to one clan only. It is the tribe, therefore, as a correlated system of clans, or such portion of it as is embraced by the classificatory nomenclature, which corre sponds to the widest circle of kinship extensions.
It is an easy but dangerous mistake to maintain that "the classi ficatory system and our own are the outcome of the social insti tutions of the clan and the family respectively," and to say that as "among ourselves this (the essential) social unit is the family" so "amongst most peoples of rude culture the clan or other exog amous group is the essential unit of social organization" (Rivers, Kinship and Social Organization, pp. 74, 75). This view carries on Morgan's mistaken opinion that the clan is a domestic insti tution, made ad hoc for purposes of group-marriage, a mistake which has recently been reaffirmed in the phrase that "the clan, like the family, is a reproductive group" (Briffault, 1927). All this is a continuous source of error in that it construes the clan into an independent, self-sufficient kinship unit, whereas the clan is essentially a group correlated to other groups of a similar nature, and dependent upon their existence. In its simplest form the correlated system is reduced to two clans, but never to one. It is this compound system which corresponds to the family, which itself is a self-sufficient independent kinship unit. The clan in fact never bears the imprint of extended full family kinship, but only of one side of it.
It is a curious mistake to take savage fiction and linguistic simile at their face value, and to regard, with Morgan, the clan as a "domestic institution," made ad hoc for purposes of group marriage ; or, with Rivers, to imagine that the clan has been the foundation of classificatory nomenclature in the same sense as the family is the basis of our own terminology; or to affirm that "the clan, like the family, is a reproductive group." The function of the clan system is neither generative nor domes.
tic ; exogamy is not primarily an injunction to marry a woman of another clan, but the prohibition of sexual intercourse within the clan. Again the relations between the older and younger genera tion within the clan, or between age-grades, are neither an equiva lent nor a copy of the parent to child relations—above all, not as regards reproductive functions! The relation of the members of a clan is a modified and ex» tended kinship solidarity; it implies co-operation in most com munal undertakings and the exclusion of sexual interests. Thus some elements of the later parent to child and brother to sister relationship are carried over into clanship, but two elements never enter it : the matrimonial relation and early parent to child rela tion. The first of these is extended, in a modified form, into the relationship between different clans, members of which may pur sue amusements and sexual interests in common, as between males and females ; and between individuals of the same sex, render each other reciprocal services from group to group, and join in enter prises on a tribal scale.
XXIV. Summary and Conclusions.—We can now define Xxiv. Summary and Conclusions.—We can now define kinship, in the first place, as the personal bonds based upon pro creation, socially interpreted ; and, in the second place, as the wider bonds derived from the primary ones by the process of gradual extensions which occur in all communities during the lif e history of the individual. On the level of savagery and lower bar barism, the powerful persistence of family bonds is given freer play, hence the extensions are more numerous and more definitely systematized ; they are backed up by legal fictions of totemic descent; by ideas of one-sided procreation or mystic identity; and they lead to the formation of wider groups such as the clan, moiety or exogamous division.
Kinship is thus a class of social relations, which must be sub divided into several varieties : primary kinship always founded on marriage and family; and the derived forms, correlated with the group of cognate households, the village-community and the clan. The terms of kinship, which are but linguistic expressions of all these relationships, have obviously also a manifold meaning, which corresponds to the social reality. Thus is explained the existence, side by side of individual and classificatory terms, of the family and the clan, of the individual and communal aspects of kinship. The enigmatic and apparently anomalous character of primitive kinship vanishes with a closer scrutiny of the facts.