“nation”), in no tribe does the same totem ever occur in both phratries. Totems and totem names are inherited by the children from the mother, in this primitive type. Thus a man, Eagle Hawk by phratry, Snipe by totem, marries a woman Crow by phratry, Black Duck by totem. His children by her are of phratry Crow, of totem Black Duck. Obviously no person can marry another of his or her own totem, because, in the phratry into which he or she must marry, no man or woman of his or her totem exists. The prohibition extends to members of alien and remote tribes, if of the same totem name.
The same rules exist in the more primitive North American tribes, but as the phratry there has generally, though not always, decayed, the rule, where this has occurred, merely forbids marriage within the totem kin.
(2) We find this type of organization, where the child inherits phratry and totem from the father, not from the mother.
(3) We find tribes in which phratry and totem are inherited from the mother, but an additional rule prevails: the rule of “Matrimonial Classes.” By this device, in phratry “Dilbi,” there are two classes, “Muri” and “Kubi.” In phratry “Kupathin” are two classes, “Ipai” and “Kumbo” (all these names are of unknown meaning). Each child inherits its mother’s phratry name and totem name, and also the name of that class of the two in the mother’s phratry to which the mother does not belong. No person may marry into his or her own class—practically into his or her own generation: the rule makes parental and filial marriages impossible,—but these never occur even among more primitive tribes which have not the institution of classes. Suppose that the class names are really names of animals and other objects in nature—as in a few cases they actually are. Then the rules, where classes exist, would amount to this: no person may marry another who, by phratry, totem or generation, owns the same hereditary animal name as himself or herself. In practice, where phratries exist, a man who knows a woman’s phratry name knows whether or not he may marry her. Where class names exist (even though the phratry name be lost), a man who knows a woman’s class name knows whether or not he may marry her. Nothing can be simpler in practice.
(4) The same rules as under (3) exist, but the phratry, totem and class are inherited through the father: the class of the child of course not being the father’s, but the linked class in his phratry.
(5) In the fifth category (Central North Australia), while phratry name (if not lost) and totem name are inherited from the father, by a refinement of law which is spreading southwards there are four classes in each phratry (or main exogamous division unnamed), and the choice of a partner in life is thus more restricted than in more primitive tribes.
(6) Finally we reach the institutions of the group of tribes called, from the name of the most powerful tribe in the set, “the Arunta nation.” They occupy the Macdonnell Ranges and other territory in the very centre of Australia. The Arunta reckon kinship in the male line: Arunta customs. their phratry names they have forgotten, in place of phratries eight matrimonial classes regulate marriage. In these respects they resemble most of the central and northern tribes, but present this unique peculiarity, that the same totems may and do exist in both of the opposed intermarrying exogamous divisions consisting of four classes each. It thus results that a man, in the Arunta tribe, may marry a woman of his own totem, if she be in the class with which he may intermarry. This licence is unknown in every other part of the totemic world, and even in the Kaitish tribe of the Arunta nation intertotemic marriages, in practice, almost never occur.
Among the Arunta the totems are only prominent in magical ceremonies, unknown in South-Eastern Australia. At these ceremonies (Intichiuma) the men of the totem do co-operative magic for the benefit of their plant or animal, as part of the tribal food-supply. The members of the totem taste it sparingly on these occasions, apparently under the belief that to do so increases their magical power: the rest of the tribe eat freely. But, as far as denoting kinship or regulating marriage is concerned, the totems, among the Arunta, have no legally important existence. Men and women of the same totem may intermarry, their children need not belong to the totem of either father or mother.
The process by which Arunta totems came thus to differ from those of all other savages is easily understood. Like the other tribes from the centre to the north (including the Urabunna nation, which reckons descent through women), the Arunta believe that the souls of the primal semi-bestial ancestors of the Alcheringa or “dream time” are perpetually reincarnated. This opinion does not affect by itself the usual exogamous character of totemism among the other tribes. The Arunta nation, however, cultivates an additional myth, namely that the primal ancestors, when they sank into the ground, left behind them certain oval stone slabs, with archaic markings, called churinga nanja, or “sacred things of the nanja.” The nanja, again, is a tree or rock, fabled to have risen up to mark the spot where a group of primal ancestors, all of one and the same totem in each case (Cats here, Grubs there, Ducks elsewhere), “went into the ground.” The souls of these ancestors haunt such spots, especially they haunt the nanja tree or rock, and the stone churinga nanja. Each district, therefore, has its own oknanikilla (or local totem centre of the ghosts), Cat ghosts, Grub ghosts, Hakea flower ghosts and so on. These spirits enter into women and are reborn as children. When a child comes to birth, the mother names the oknanikilla in which she conceived it, and, whatever the ghost totem of that place may be, it is the child’s totem. Its mother may be a Grub, its father may be a Crow, but if the child was conceived in a Duck, or Cat, or Opossum or Kangaroo locality, it is, by totem, a Cat, Opossum, Duck or Kangaroo. The churinga nanja of its primal ancestor is sought for at the place of the child’s conception, and is put into the sacred repository of such objects.
Thus the child does not inherit its totem from father, or from mother, as everywhere else, but does inherit the right to do ceremonies for the paternal totem: a proof that, of old, totems were inherited, as elsewhere, and that in the male line. If totems among the Arunta, as everywhere else, were once arranged on the plan that the same totem never occurs in both exogamous moieties, that arrangement has been destroyed, as was inevitable, by the existing method of allotting totems to children,—not by inheritance,—but at haphazard. By this means (a consequence of the unique Arunta belief about churinga nanja) the same totems have got into both exogamous moieties, so that persons of the same totem, but of appropriate matrimonial classes, may marry. This licence is absolutely confined to the limited region in which stone churinga nanja occur.
The whole system is impossible except where descent is reckoned in the male line, for there alone is local totemism possible, and the Arunta system is based on local totemism, plus the churinga nanja and reincarnation beliefs. With reckoning of descent in the female line, no locality can possibly have its local totem: all the totems indiscriminately distributed everywhere: and thus no woman can say in what totemic locality her child was conceived, for there is not and cannot be, with female descent, any totemic locality. Now it is admitted that reckoning by female descent is the earlier method, and it is granted that in rites and ceremonies the Arunta are of a relatively advanced and highly organized pattern. Their social organization is local, and they have a kind of local magistracies, hereditary in the male line.
In spite of these facts, Spencer and Gillen conceive that the peculiar totemism of the Arunta is the most primitive type extant (cp. Spencer, J.A.I. (N.S.), vol. i. 275-281; and Frazer, ibid. 281-288). It is not easy to understand this position, as, without male kinship and consequent local totemism (which are not primitive), and without the churinga nanja (which exist only in a strictly limited area), the Arunta system of non-exogamous totems cannot possibly exist. Again, the other tribes cannot have passed through the Arunta stage, for, if they had, their totems would have existed, as among the Arunta, in both exogamous moieties, and would there remain when they came to be inherited;
