[4. THE ORIGIN OF EXOGAMY.]
Though the dual organization of the tribe seems to admit of a comparatively simple and easy explanation, the totemic exogamy which is closely bound up with it offers great difficulties. As we have already seen, totemic exogamy is characterized by the fact that a member of one specific clan, or of a totem group belonging to the clan, may enter into marriage only with a member of another clan or totem group. This restriction of the marriage relationship is generally known as 'exogamy,' a term first introduced by the Scottish ethnologist and historian, McLennan. In order to distinguish this custom from later regulations of marriage, such, for example, as exist in present law, in the prohibition of the union of relatives by blood or by marriage, we may call it more specifically 'totemic exogamy.' Totemic exogamy clearly represents the earliest form of marriage restriction found in custom or law. The phenomena bound up with it may be regarded as having arisen either contemporaneously with the first division of the tribe or, at any rate, soon thereafter, for some of the Australian and Melanesian tribes practise exogamy even though they have not advanced beyond a twofold division of the tribe. On the other hand, the primitive horde of the pretotemic age remains undivided, and, of course, shows no trace of exogamy. True, marriages between parents and children seem to have been avoided as early even as in pretotemic times. But this could hardly have been due to the existence of firmly established norms of custom. Such norms never developed except under the influence of totemic tribal organization, and they are closely related to its various stages of development.
Taking as the basis of consideration the above-mentioned conditions in Australia, where an approximate regularity in the successive stages of this development is most clearly in evidence, we may distinguish particularly three main forms of exogamy. The first is the simplest. If we designate the two divisions of the tribe between which exogamic relations obtain, by A and B, and the various subgroups of A by l, m, n, o, and of B by p, q, r, s, we have, as this simplest form, unlimited exogamy. It corresponds to the following schema:—
This means: A man belonging to Class A may take in marriage a woman from any of the subgroups of Class B, and conversely. Marriage is restricted to the extent that a man may not take a wife from his own class; it is unrestricted, however, in so far as he may select her from any of the subgroups of the other class. This form of exogamy does not appear to occur except where the divisions of the tribe are not more than two in number. The marriage classes, A and B, then represent the two divisions of the tribe; the subgroups l, m, n, o, p,... are totem groups—that is to say, according to the view maintained above, cult groups. For the most part, marriage relationships between the specific cult groups meet with no further restrictions. A man of Class A may marry a woman belonging to any of the totem groups p, q, r, s, of Class B—it is only union with a woman belonging to one of the totem groups of Class A that is denied him. Nevertheless, as we shall notice later, we even here occasionally find more restricted relations between particular totem groups, and it is these exceptions that constitute the transitional steps to limited exogamy. Such transitions to the succeeding form of exogamy are to be found, for example, among the Australian Dieri, some of whose totem groups intermarry, only with some one particular group of the other tribal division.
The second form of exogamy occurs when a member of Class A is not allowed to take in marriage any woman he may choose from Class B, but only one from some specific sub-group of B. For example, a man of group n is restricted to a woman of group r.
Both forms of exogamy, the unlimited and the limited, observe the same law with respect to the group affiliation of children. If, as universally occurs in Australia, A and B are clans having exogamous relations, and l, m, n, o, p,... are totem groups within these clans, then, if maternal descent prevails, the children remain both in the clan and in the totem of the mother; in the case of paternal descent, they pass over to the clan and to the totem of the father. Of these modes of reckoning descent, the former is dominant, and was everywhere, probably, the original custom. One indication of this is the connection of paternal descent with other phenomena representing a change of conditions due to external influences—the occurrence of the same totem groups, for example, in the two clans, A and B, that enjoy exogamous relations. The latter phenomenon is not to be found under the usual conditions, represented by diagrams I and II. In the case of unlimited exogamy (I), no less than in that of limited exogamy, we find that if, for example, maternal descent prevails, and, the mother belongs to clan B and to totem group r, the children likewise belong to this group r. This condition is much simplified in the case of the American Indians. With them, totem group and clan coincide, the totem names having become the names of the clans themselves. The particular totem groups, l, m, n, o, p,... do not exist. Exogamous relations between clans A and B consist merely in the fact that a man of the one clan is restricted in marriage to women of the other clan. Wherever maternal descent prevails, as it does, for example, among the Iroquois, the children are counted to the clan of the mother; in the case of paternal descent, they belong to the clan of the father.