This definition is scientifically drawn so as to include what is common to all Totems and not to include any character that is not universally connected with them. It is true of the Totems of individuals (or guardian genii), very common among the Northern Amerinds; as well as of the Sex-Totems (say Bat for men and Owl for women) which occur in Australia; and of the Totems of clans found in many parts of the world, of which the Australian may be supposed most nearly to represent the original institution. When one speaks of Totems without qualification, one means these Clan-Totems, being incomparably the most important: others probably derive from them.
“The Clan-Totem” (to quote the same source) “is reverenced by a body of men and women who call themselves by the name of the Totem, believe themselves to be of one blood, descendants of a common ancestor, and are bound together by common obligations to each other and by a common faith in the Totem.”[511] This definition is also strict: whereas it is not uncommon for writers on Totemism to include in their notion of the Totem other characters which are accidents more or less often associated with it: as that the clan believes itself descended from the Totem; though in many cases the clan traces its origin to something that was neither Totem nor man, or regards the Totem animal as descended from their own human ancestress, or tells some story of the transformation of a man into an animal, or of an animal into a man, or of an ancestor’s friendship with the Totem animal; for the institution is a copious source of explanatory myths. Again, it may be written that no member of a clan may kill or injure the Totem, or eat, or utilise it (if an animal or plant); and this is often, but not always true. If the Totem is water, or the most important food in the district—dugong, turtle, coconut, sago—abstention is impossible or intolerably inconvenient. Hence among the Kaitish a Water clansman when alone may drink; but if others are with him he may drink only when water is given him by a man of another Totem. And to eat a little of the Totem sacramentally, in order to confirm one’s identity with it, may be allowed or (rather) required. Among the Warramunga the Snake clan may kill snakes. It may even be the clan’s duty to destroy or drive away their Totem; as happens with the Mosquito clan of the Kakadus in North Australia, and with the Reptile clan amongst the Omahas. Similarly, the clan sometimes, but by no means always, imitate the Totem in character, in costume, in dances, or bear in tattoo or otherwise its badge to manifest the community of nature. They may be thought to have special magical powers to influence its fertility or to control its actions (actions of game or wind or rain), and may use these powers for the benefit of the tribe. At death the soul of a clansman may pass into, or may appear as, the Totem animal. The Totem may be the protector of its clansmen, and in some cases it seems to have become a god. These accidents of Totemism, however, are very irregularly diffused and some of them are rare.
Most important socially is the connexion between Totemism and Marriage-customs. In the great majority of cases, men and women of the same Totem may not intermarry; the Totem clans are said to be exogamous; but there are exceptions even to this rule. In many parts of the world—America, Africa, India, Indonesia, as well as Australia—tribes are divided into non-intermarrying or exogamous classes (moieties or phratries); sometimes the classes are subdivided, and even the sub-classes again (in parts of Australia), so that in all the exogamous classes are eight; and the rules concerning these classes are such as to prevent the marriage of brothers and sisters, parents and children and first cousins. The Totem clans are usually distributed amongst the Marriage-classes; being probably the older institution (whether or not originally exogamous), they may be said to have been subordinated by the formation of the classes; and in the matter of marriage they follow the class rules. But the Arunta with some neighbouring tribes of Central Australia, though divided into eight exogamous classes, have not subordinated the Totem clans; which, accordingly, are not exogamous. If Exogamy was originally the vital utility of Totemism, the institution is undermined by the Marriage-classes, whether the clans are subordinated to the classes and merely repeat their rules, or are not subordinated and lose their Exogamy. This may explain why the classes prevail more widely than Totems, and may be found where Totemism seems to have become extinct.
§ 2. Of the Origin of Totemism
It has already been said that Totemism does not prevail universally amongst mankind. There is very little evidence of its having existed amongst the ancestors of the Indo-European peoples; and it seems to be unknown to many of the most primitive of surviving tribes—the Boschmans, Veddas, Puranas (Borneo), Andamanese, Yaghans (Tierra del Fuego) and Californian Indians. It is not, therefore, something founded in human nature itself, but must have originated (whether in one or in several places) under particular conditions.
There is some probability that the earliest Totems were animals. Wherever Totems are found most of them are animals. A list of Totems recognised in North Australia, given by Spencer and Gillen,[512] comprises 202 altogether: Animals 164,[513] Plants 22, Inanimate 16. The animals (except insects) and plants are nearly all edible. To hunters animals must have been of all things the most interesting; next, certain plants, especially to women. Animals lend the greatest plausibility to any notion of blood-relationship. The inanimate Totems include some which it is very desirable to control by Magic, such as Water, Wind, Fire, Sun, Moon and the Boomerang.
Generally, no doubt, Totemism is ancient; but in some cases the existing Totems must be recent, as among the Bahima of Uganda.[514] Nearly all their Totems are different coloured cows, or some part of a cow, and must have been adopted since the advance to pastoral life—at a stage of culture, therefore, much above the Australian. “Split Totems,” arising on the division of a clan by dividing its Totem (as the Omahas comprised Buffalo-heads, Buffalo-tails, etc.[515]), were probably formed by deliberate agreement, and had the incidental advantage that a clansman’s food-taboo did not extend to the whole buffalo, the staple food of the tribe, but only to the head or tail.
Amongst the Australian aborigines, the group of beliefs and practices included in, or connected with, Totemism are of such intense and widespread social importance, that it must have prevailed amongst them for many generations. The northern Amerinds are so much in advance of the Australians in social organisation, culture and mentality, that if their Totemism has descended to them from a time when they lived at the Australian level, its history must go back not merely for many generations but for thousands of years. And if Egyptian gods, such as Hathor and Anubis, were formerly Totems, the retrospect becomes still longer. I am not aware of any direct evidence of the prehistoric existence of Totemism, such as we have, in some ancient burials, of the existence of Animism; but some palæolithic carvings or paintings may have had totemic significance.
It follows that any account of the origin of Totemism can only be hypothetical. Were this all, indeed, it would be on the same foot with Magic and Animism; but it is not all. That Magic should arise from belief in mysterious forces and from confusing coincidence with causation, or that Animism should result from a confusion between dreams and objective experience (with the help of Magic-ideas) is highly probable; because such errors are active to this day, and they seem to spring up in primitive minds by psychological necessity; though it may not be intrinsically necessary that the errors should be perpetuated and systematised as, in general, they have been. But the case of Totemism is different; for we do not see—at least, no one has yet shown—any sort of necessity why in certain cases clans should have borne the names of certain animals: and this is the root of the whole matter. Indeed, this practice, not being universal, cannot be necessary; and it may have had several different origins. Any relevant hypothesis, therefore, can claim no more than to agree with the known facts better than rival hypotheses; we cannot expect to deduce it from laws of human nature; whilst still another hypothesis just as good may any day be put forward by some speculative genius; and the doubt must always remain whether some important facts of Totemism have not been lost which, could they be recovered, would prove all our guesses to have been made in vain.
In spite of these discouraging considerations, several hypotheses have been proposed: this also is for us a sort of psychological necessity. They may be grouped into two classes, according as they assume the totemic names to have been originally names of individuals or of whole clans. All seem to assume that one explanation must hold good for all the cases.