Method in the Study of Totemism

Is there any human institution which can be safely called "Totemism"? Is there any possibility of defining, or even describing Totemism? Is it legitimate – is it even possible, with due regard for "methodology" and logic – to seek for the "normal" form of Totemism, and to trace it through many Protean changes, produced by various causes, social and speculative? I think it possible to discern the main type of Totemism, and to account for divergences.

Quite the opposite opinion appears to be held by Mr. H. H. Goldenweizer in his "Totemism, an Analytic Study."1 This treatise is acutely critical and very welcome, as it enables British inquirers about totemism to see themselves as they appear "in larger other eyes than ours." Our common error, we learn, is this: "A feature salient in the totemic life of some community is seized upon only to be projected into the life of the remote past, and to be made the starting-point of the totemic process. The intermediary stages and secondary features are supplied from local evidence, by analogy with other communities, or 'in accordance with recognised principles of evolution' [what are they?] and of logic. The origin and development, thus arrived at, are then used as principles of interpretation of the present conditions. Not one step in the above method of attacking the problem of totemism is logically justifiable."2

As I am the unjustifiable sinner quoted in this extract,3 I may observe that my words are cited from a harmless statement to the effect that a self-consistent "hypothesis," or "set of guesses," which colligates all the known facts in a problem, is better than a self-contradictory hypothesis which does not colligate the facts.

Now the "feature salient in the totemic life of some communities," which I "project into the life of the remote past," and "make the starting-point of the totemic process" is the totemic name, animal, vegetable, or what not, of the totem-kin.

In an attempt to construct a theory of the origin of totemism, the choice of the totemic name as a starting-point is logically justifiable, because the possession of a totemic name is, universally, the mark of a totem-kin; or, as most writers prefer to say, "clan." How can you know that a clan is totemic, if it is not called by a totemic name? The second salient feature in the totemic life of some communities which I select as even prior to the totemic name, is the exogamy of the "clans" now bearing totemic names.

To these remarks Mr. Goldenweizer would reply (I put his ideas briefly) there are (1) exogamous clans without totemic names; and there are (2) clans with totemic names, but without exogamy.

To this I answer (1) that if his exogamous clan has not a totemic name, I do not quite see why it should be discussed in connection with totemism; but that many exogamous sets, bearing not totemic names, but local names or nicknames, can be proved to have at one time borne totemic names. Such exogamous sets, therefore, no longer bearing totemic names, are often demonstrably variations from the totemic type; and are not proofs that there is no such thing as a totemic type.

Secondly, I answer, in the almost unique case of "clans" bearing totemic names without being exogamous, that these "clans" have previously been exogamous, and have, under ascertained conditions, shuffled off exogamy. They are deviations from the prevalent type of clans with totemic names plus exogamy. They are exceptions to the rule, and, as such, they prove the rule. They are divergences from the type, and, as such, they prove the existence of the type from which they have diverged.

So far I can defend my own method: it starts from features that are universal, or demonstrably have been universal in totemism. There is "an organic unity of the features of totemism," – of these two features, the essential features.

Lastly, Mr. Goldenweizer accuses us "Britishers," as he calls us, of neglecting in our speculations the effects of "borrowing and diffusion, of assimilation and secondary associations of cultural elements, in primitive societies."4

This charge I do not understand. There has been much discussion of possibilities of the borrowing and diffusion and assimilation of phratries, exogamy, and of totemic institutions; and of "ethnic influences," influences of races, in Australia. But the absence of historical information, the almost purely mythical character of tribal legends (in North-West America going back to the Flood, in Australia, to the "Dream Time"), with our ignorance of Australian philology, prevent us in this field from reaching conclusions.

(Possibly philologists may yet cast some light on "ethnic influences" in Australia. The learned editor of Anthropos, Père Schmidt, tells me that he has made a study of Australian languages and believes that he has arrived at interesting results.)

Mr. Goldenweizer represents, though unofficially, the studies of many earnest inquirers of North America, whether British subjects, like Mr. Hill Tout, or American citizens such as Dr. Boas. They vary, to be sure, among themselves, as to theories, but they vary also from British speculators. They have personally and laboriously explored and loyally reported on totemism among the tribes of the north-west Pacific coast and Hinterland; totemism among these tribes has especially occupied them; whereas British anthropologists have chiefly, though by no means solely, devoted themselves to the many varieties of totemism exhibited by the natives of Australia. These Australian tribes are certainly on perhaps the lowest known human level of physical culture, whereas the tribes of British Columbia possess wealth, "towns," a currency (in blankets), rank (noble, free, unfree), realistic art, and heraldry as a mark of rank, and of degrees of wealth.

Mr. Goldenweizer's method is to contrast the North-Western American form of totemism with that prevalent in Central Australia, and to ask, – how, among so many differences, can you discover a type, an original norm? I answer that both in North-Western America and in Central Australia, we find differences which can be proved to arise from changes in physical and "cultural" conditions and from speculative ideas. I have said that in British Columbia the tribes are in a much more advanced state of culture than any Australian peoples, and their culture has affected their society and their totemism. Wealth, distinctions of rank, realistic art, with its result in heraldry as a mark of rank, and fixed residence in groups of houses are conditions unknown to the Australian tribes, and have necessarily provided divergences in totemic institutions. Mr. Goldenweizer replies "that the American conditions are due to the fact that the tribes of British Columbia are 'advanced' cannot be admitted."5 But, admitted or not, it can be proved, as I hope to demonstrate.

II

Mr. Goldenweizer gives what he supposes some of us to regard as "essential characteristics" or "symptoms" of totemism. He numbers five of these "symptoms."

1. An exogamous clan.

2. A clan name derived from the totem.

3. A religious attitude towards the totem, as a "friend," "brother," "protector," &c.

4. Taboos or restrictions against the killing, eating (sometimes touching, seeing) of the totem.

5. A belief in descent from the totem.

Mr. Goldenweizer next, by drawing a contrast between British Columbian and Central Australian totemism, tries to prove, if I understand him, that "the various features of totemism," are, or may be "essentially independent of one another," "historically, or psychologically, or both."6

Now, looking at the five symptoms of totemism, I may repeat (speaking only for myself) that, as to 1 and 2, I think the exogamous clan, with "a clan name derived from the totem" is an institution of such very wide diffusion that I may blamelessly study it and attempt to account to myself for its existence. But this does not mean that I regard all exogamous social sets as at present totemic; or as always having borne totem names. Again, sets of people (I cannot call them "clans," for the word "clan" indicates persons claiming common descent from a male ancestor, – say Clan Gihean, Clan Diarmaid), may bear animal or vegetable or other such names, yet not be at present, as such, exogamous. Of these are the Arunta, and the Narran-ga.

3. A religious attitude towards the totem. One cannot discuss this without a definition of religion. "Totemism is not a religion," says Mr. Frazer, with whom I am here in agreement.

4. Totemic taboos. These, though extremely general, are not quite universal even in Australia.

5. A belief in descent from the totem.

This belief is post-totemic, being merely one of many aetiological myths by which men explain to themselves why they are totemists; what is the nature of the rapport between them and their totems; why they bear as a kin (or association) animal or vegetable names. One or another such myth is not an essential part of totemism, for it is, necessarily, post-totemic.

I am thus left confronting the problems, (1) why are the immense majority of exogamous kins, in societies which we call "totemic," named by animal and other such names; and (2) why are they exogamous?

As for other exogamous social sets, which bear, not animal names, but territorial, or descriptive names, or nicknames, often derisive, it is my business to show, if I can, that these sets, or some of them, have passed, in historical times, out of the stage of totem-kins, owing to circumstances which I shall describe. Next (2) I have to show, if I can, why a few sets of people, bearing, as sets or associations, animal or other such names, are now no longer exogamous.

If I succeed, I think that I may regard "Totemism" as characterised by exogamous kins bearing totemic names, and as "an integral phenomenon" existing in many various forms.7

If I understand Mr. Goldenweizer this attitude and effort of mine must seem to him "methodologically" erroneous, and "logically unjustifiable." "This attitude," he says (namely the attitude of those who hold totemism to be "an integral phenomenon"), "is reflected in the way several authors deal with the so-called 'survivals' of totemism, where from the presence in some region of one or two of the 'symptoms' of totemism, or of the fragments of such symptoms, they infer the existence in the past of totemism in its 'typical form,' that is, with all its essential characteristics."8

Thus, for example, from such phenomena as standards bearing animal forms; or from animal worship, – each animal being adored in its own district, – or from myths of descent from gods in the form of animals; or from the animal names of some Roman gentes; or from animals closely associated with gods (like the Shrew Mouse with Sminthian Apollo); or from the presence of beings partly theriomorphic partly anthropomorphic, in art, many writers infer a past of totemism in Italy; Israel; Greece Hellenic and Greece Minoan; in Egypt; in Ireland; and so forth. It is not my purpose to treat of such so-called survivals. I am to deal with peoples such as the tribes of Australia, New Guinea, and North-West America, who, if not the rose, have been near the rose: if not always totemic are at least neighbours of totemists.

III

Mr. Goldenweizer tabulates the results of his comparison between the Totemism of British Columbia and that of Central Australia.9 In the latter region the totemic institutions and myths are not those of South-Eastern Australia. To the totemism of many tribes in South-Eastern Australia that of a great tribe of British Columbia, the Tlingit, bears, – if we may trust some of the evidence, – the closest possible resemblance; while, if we trust other and conflicting evidence, the resemblance is, on an important point, nearer to the institutions of certain Australian tribes of the furthest south, in Cape Yorke peninsula. The evidence for British Columbian totemism, I shall show, is so wavering as to make criticism difficult. The terminology, too, of some American students has been extremely perplexing. I am sorry to be obliged to dwell on this point, but a terminology which seems to apply five or six separate terms to the same social unit needs reform.

Dr. Boas is one of the most energetic field-anthropologists of the United States. To him we owe sixteen separate disquisitions and reports on the natives of the North-West Pacific coast and Hinterland, all of them cited by Mr. Goldenweizer in his excellent Bibliography. But Mr. Frazer observes that Dr. Boas variously denominates the kindred groups of the Kwakiutl tribe as "groups," "clans," "gentes," and "families." I must add that he also uses gentes as a synonym for phratries – "Phratries, viz. gentes."10 Now a "phratry" is not a gens; a "group" may be anything you please; a "family" is not a gens; – a "gens" is an aggregate of families, – and a "clan" is not a "family."

Mr. Goldenweizer's tabulated form of his comparisons between British Columbia and Australia contains ten categories (see the last footnote of p. 6). Of these, two at least (8) (9) indicate elements which are purely proofs that the B.C. tribes are on a much higher, or later, level of social progress than the Australians. These two are Rank and Art. Had Mr. Goldenweizer added Wealth and Towns to his ten categories he would have given four factors in B.C. culture which affect B.C. totemism, and which do not exist in Central Australia, where realistic art is all but wholly unknown: art being occupied with archaic conventional patterns. Thus, in Australia, the bewildering B.C. heraldry – the "crests" – cannot, as in B.C., confuse the statements of observers, perplex their terminology (for they often use "crests" as synonyms of "totems"), and disorganise totemism itself. But we can find, not far from Australia, a parallel to this heraldry in New Guinea. For "crests" or badges in Central British New Guinea, see Totemism and Exogamy, vol. ii. pp. 42-44. The people, like the B.C. tribes, are settled in villages. They have "a number of exogamous clans," most clans occupying several villages, and they have paternal descent. "Every clan" (as apparently in some cases in British Columbia) "has a number of badges called Oaoa, which, generally speaking, may only be worn or used by members of the clan." The "clan" names are geographical or are patronymics, they are not totemic; the badges either represent birds and mammals, or are "schematised" from some prominent feature of these. The people are not now totemists, even if they have passed through totemism.

Again (category 5), in British Columbia, "Magical Ceremonies are not associated with Totemism." In Central Australia they are "intimately associated with totemism." Yes, but in South-Eastern Australia they are not, as far as our evidence informs us. Magical ceremonies are not in Mr. Goldenweizer's list of five symptoms or characteristic peculiarities of totemism, so I leave them out of account.

Again, as to Taboo (category 3), in British Columbia, "non-totemic taboo is common; totemic, absent."

As to this "absence," Mr. Frazer has a great deal to say. For example, we have Commander Mayne's book, Four Years in British Columbia, a work of 1862, in which is given information from Mr. William Duncan, a missionary among the Tsimshian tribe. All such evidence given prior to controversies about totemism is valuable. According to this account, the Indians used, as "crests," representations of Whale, Porpoise, Eagle, Raven, Wolf, Frog, etc. Every person was obliged to marry out of the name of the animal represented by his crest, and each "clan" tabooed its animal, "will never kill the animal which he has adopted for his crest, or which belongs to him as his birthright," that is, apparently, his "familiar," and his inherited totem. This is original totemism in North-West America.

Mr. Frazer says, "So far as I remember, no other writer on these North-Western Indians has mentioned their reluctance to kill their totemic animals. In the course of this work I have repeatedly called attention to the paucity of information on this important side of totemism in the writings of American ethnologists."11 Mr. Frazer also finds the usual totemic taboo among the Yuchi, a tribe of the Gulf nations.12

In Central Australia are "numerous totemic and non-totemic taboos." But in other parts of Australia there are also tribes where people even kill and eat their totems. The totemic taboo is an extremely common institution, but not a note stantis vel cadentis ecclesiae.

Another category is (4), "Descent from the Totem." As I have said, the belief in this descent is a mere explanatory myth to account for totemism; and, like all other such myths, could only arise after men were not only totemic, but wondered why they were totemic. Consequently such myths are not of the essence of totemism, and their varieties are of no importance.

The belief, or myth, of totemic descent is absent in British Columbia, says Mr. Goldenweizer, in the Tlingit, Haida, and Tsimshian tribes, and present "among the Kwakiutl and further south." In Central Australia descent from the totem is "universal."

But it is a queer kind of "descent," is not, in the usual sense, descent at all, and, notoriously, is not descent by physical generation.

Then we have the category (7), "Guardian Spirits, intimately associated with Totemism" in British Columbia, "not associated with it in Central Australia." Yet, in Central Australia, a man's spirit is a totemic spirit. Again (10), "Number of Totems." In British Columbia "small," in Central Australia "large." But it is "small" in such central regions of Australia as those of the Dieri and Urabunna, and in South-Eastern Australia; and why it is so large among the Arunta no man knows. It is an unexplained peculiarity, and not essential.

"Reincarnation" (6) is, in British Columbia, "not associated with Totemism," in Central Australia "intimately associated with Totemism." Here, Mr. Strehlow, for the Southern Arunta, reports otherwise; while for the Northern Arunta and other tribes, this "reincarnation" is part of a speculative explanatory myth. The myth, as I can show, explains, at one stroke, how men come to have souls, and why men are totemic We know the kind of savage philosophy which accounts for this category.

I have now remarked on eight out of Mr. Goldenweizer's ten categories of differences between British Columbian and South Australian totemism; all of them, I think, are separable accidents of totemism; and most of them are easily to be accounted for by actual differences of culture, of social conditions, and by variety of savage taste and fancy in making guesses as to why totemists are totemistic.

IV

We next arrive at the two first of Mr. Goldenweizer's categories. These are concerned with points of such very wide diffusion in the totemic world that I, under correction, take leave to regard them as "normal," while I hold that such variations from the norm as exist can be explained – as aberrations.

The first of these two categories is announced as:

BRITISH COLUMBIA.

1. Exogamy

Totemic phratries (Tlingit).

Totemic clans (Haida, Tsimshian, Northern Kwakiutl).

CENTRAL AUSTRALIA.

2. Exogamy

Phratries.

Classes.

Totem clans (generally not independent exogamous units).

This needs explanation! By "totemic phratries" in the case of the Tlingits, Mr. Goldenweizer means the two main exogamous divisions of the tribe, Wolf and Raven. By "totemic clans," in the case of the Haida, he also means the two main exogamous divisions, Raven and Eagle, which, really, are phratries. But it is also clear that Mr. Goldenweizer is here using the word "clans" as it exists in the peculiar terminology of Dr. Swanton. Mr. Goldenweizer informs us that "Dr. Swanton now fully recognises the strict parallelism of the social units of the Tlingit and Haida, and sanctions the use of 'phratry' and clan in both cases." This terminological source of confusion happily disappears.

We are now, alas, entering a region where the variations of evidence, the confusions of terminology, and the influence of wealth and rank in the creation of heraldry, cause extreme perplexity. Meanwhile, as the Haida "clans" of the category are, in fact, phratries; on the other hand the "totemic clans" of the Tsimshians and Northern Kwakiutl (Raven, Eagle, Hawk, Wolf), and six "totemic clans" of the Northern Kwakiutl seem destitute of phratries, which, among the Arunta of Central Australia, have also died out Mr. Goldenweizer, however, assigns phratries to Central Australia, the Arunta have none;13 also "totem clans," where there are none, for the totemically named associations of the Arunta are not "clans," in the normal and usual sense of that word; they are not kins but associations.

Mr. Goldenweizer, in his first category, speaks of Central Australia as possessing totemic "clans" ("generally not independent exogamous units"). If by "Central Australia" he means the Arunta group of tribes, they have, I repeat, no "totemic clans"; they have only clubs with totemic names, and these associations are not "exogamous units." Where phratries with totem kins in them exist, no totem kin is or can be "an independent exogamous unit," except where one totem to one totem marriage prevails, as among certain Australian tribes. But if the phratry rule be dropped, as Morgan says it was among the Iroquois, then people may marry into any totem kin except their own, and each totem kin becomes an "independent exogamous unit."14

Thus the first category in Mr. Goldenweizer's list needs a good deal of explanation and criticism.

The second category is Totemic Names. Under these, in British Columbia, are:

"Phratries (Tlingit)."

"Clans (Haida)." (But these are phratries.)

"Two of four clan Tsimshian."

"Clans (Northern Kwakiutl)."

In place of two animal-named clans out of four, Mr. Frazer assigns four animal-named clans to the Tsimshians;15 Raven, Eagle, Wolf, and Bear. (T. and E., vol. iii. pp. 307-308.) Mr. Goldenweizer himself16 also assigns these four animal-named clans to the Tsimshians. But, in his table,17 he docks two Tsimshian clans of their totem names. He does so also in his p. 190. Thus (p. 187) all of the four Tsimshian "clans" have animal names. But (p. 190), and also in the tabular arrangement, only two of the Tsimshian clans have animal names. Mr. Frazer gives to all four Tsimshian clans the names of animals. Whom are we to believe18 Method is here a little to seek.

A much more serious puzzle meets us when, in his second category (totemic names), Mr. Goldenweizer assigns no totemic names to the "clans" of the Tlingit, while Mr. F. Boas (whose list is quoted by Mr. Frazer) and Holmberg (1856) do assign totemic names to the Tlingit clans.

Let us examine this situation.

If we take a South-East Australian tribe of the Barkinji pattern, we find it divided into two animal-named intermarrying phratries (or exogamous intermarrying "classes" or "moieties," I call them "phratries"). In each phratry are totem kins, that is, kins named after animals, vegetables, or other things in nature. The names of phratries and totem kins (I know no other word for them but totem kins or totem clans) descend in the female line. No such totem kin occurs in both exogamous phratries, therefore all these units are necessarily exogamous.

Two-thirds of the Australian phratry names are untranslated, like those of the Dieri; the other third, with a single exception (the Euahlayi), are names of animals.19

Now turn to the disputable case of the Tlingits of British Columbia. I first examine Mr. Frazer's account of them in Totemism and Exogamy (vol. iii. pp. 264-278). The Tlingits are divided into two exogamous phratries, or "classes," of animal names, Raven and Wolf. (In the north the Wolf "class" is also known as the Eagle.) Phratry exogamy is the rule; descent is in the female line. Each phratry is subdivided into a number of "clans," which are named after various animals. As no "clan" is represented in both phratries, and as all folk are obliged to marry out of their own phratry, the "clans" are, inevitably, exogamous.