Possibly the institution grew up among people who did not look so far forward, who 'took short views.' It is certain that, if the object of the classes was to stop marriages between young and old, it is a failure. 'The old men marry young wives at present,' says Mr. Mathews. If so, Herr Cunow may be right. Dr. Durkheim offers a theory. But his theory takes for granted, as we saw, that the two 'phratries,' originally, were only two totem groups, containing within them no members of other totem kins. 'They were not yet subdivided' into other totem kins. But I have tried to show that there was no such 'subdivision' into 'secondary clans' or totem kins. Dr. Durkheim regards these totem kins as colonies split off from the two original totem groups which became phratries.[13] My reasons against accepting this position have already been given. This being the case, it is unnecessary to unfold Dr. Durkheim's theory of the origin of the classes. Probably that of Herr Cunow comes nearest to the truth.
Mr. Mathews offers another solution of the problem. 'Phratry' Dilbi, for example, has 'classes' Murri and Kubbi, while the linked phratry, Kupathin, has classes Ippai and Kumbo. 'It is possible that the group Dilbi was divided into (female) Matha and Kubbitha to distinguish the mothers from the daughters, and that the terms Murri and Kubbi were adopted to provide names for the uncles and nephews of their respective generations.' Thus we return to distinction of generations. In any case the 'classes' 'have the effect of preventing consanguineous marriages, by furnishing an easy test of relationship when the tribe has become so numerous or widespread that kinship could not otherwise be well determined.[14] Later (p. 168) Mr. Matthews writes, 'The mother of a man's wife, and also his daughters, belong to the same section' ('class'), 'and therefore his marriage with that section is prohibited.' That is, he cannot marry out of his generation above or below, as indicated by 'class' names. 'Neither can he marry into the section to which his mother belongs, although a woman might be found in either case, who was in no way connected with him.' In short, as far as the names rudely indicate the generation above, and the generation below a man, he cannot marry into these classes. But, as old men do marry young wives, the apparent intention of the rules is to some extent frustrated. We can say no more, till we are told what the class names mean in a literal sense. Does nobody inquire into this essential question?
As if to accentuate the problems raised by the change of 'class' names in each generation, Mr. Matthews has discovered that when a man may marry a woman of his own 'phratry,' but out of a set of totems not his own, the totems of his children by her alter as the class names do. 'The children take the totem name,' not of their mother, but of their maternal grandmother. 'One totem is the mother of another totem.'[15] This is an unusual phenomenon, and looks like the effort of a desperate ingenuity.
The class system exists among the Arunta, with male descent. One moiety of the southern part of the tribe consists of Panunga and Bulthara, linked classes, calling themselves Nakrakia; the other moiety is of Purula and Kumara, calling themselves Mulganuka. A Bulthara man of the first moiety can only marry a Kumara woman, of the second moiety: a Purula man marries a Panunga woman only. The children of a Bulthara man's union with a Kumara woman take neither the Bulthara nor Kumara name, but are called Panunga, while the children of a Purula man and a Panunga woman are Kumara: of a Panunga man and a Purula woman, Bulthara; of a Kumara man and a Bulthara woman, Purula.
That is to say, the Arunta reckoning in the male line, a man's children do not take his 'class' name but the name of the 'class' linked to his, and forming, with his, one division of the tribe. Further each of these four divisions consists of two moieties, and a Panunga man, though he can marry a Purula woman, must choose her out of the proper moiety of the Purula division. These moieties of each division, among the Northern Arunta, have names; Uknaria, Appungerta, Umbitchana, Ungalla, and the children of each marriage fall under these names.
This restricts a man to only an eighth of the women of his generation, but, on the other hand, among the Arunta, the totem prohibition no longer exists: the totems are not restricted to one or another class, but skip among them, as we have shown in the section on the Arunta. The eight class system, perhaps the four class system, may be regarded as later and conscious modifications of the old phratry and totem rules, which, on my hypothesis, had no conscious moral origin.
[1] Die Verwandtschafts-Organisationen der Australneger. Diek, Stuttgart, 1894.
[2] This can hardly be, as the most backward tribes have phratries and totems, but no 'classes.'