Kitabı oku: «Method in the Study of Totemism»

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.

For purposes of comparison with other British Columbia tribes, I give the list of Tlingit totem kins furnished by Mr. Frazer, "on the authority of Mr. F. Boas"²⁰:

Примечание 1²¹

As I found out, and proved, in many Australian tribes the name of each phratry also occurs as the name of a totem kin in the phratry; so also it is among the Tlingit —teste Mr F. Boas.²²

Thus on every point – female descent, animal-named phratries, animal-named totem kins, and each phratry containing a totem kin of its own name, the Tlingit totemism is absolutely identical with that of many South-Eastern Australian tribes of the most archaic type.

But the Tlingit, unlike the Australians, live in villages, and "the families or households may occupy one or more houses. The families actually take their names from places." (I italicise the word "families.") Mr. Frazer's authorities here are Holmberg (1856), Pauly (1862), Petroff ("the principal clans are those of the Raven, the Bear, the Wolf, and the Whale"), Krause (both here undated). Dr. Boas (1889). and Mr. Swanton (1908).

Mr Goldenweizer²³ does not mention that the "clans" of the Tlingit have animal names. Quite the reverse; he says that "the 'clans' of the Tlingit … bear, with a few exceptions, names derived from localities."²⁴ This is repeated on p. 225.

At this point, really, the evidence becomes unspeakably perplexing. Mr. Frazer, we see, follows Mr. F. Boas and Holmberg (1856) in declaring that the "clans" of the Tlingit bear animal names. Mr. Goldenweizer says that, "with few exceptions," the "clans" of the Tlingit bear "names derived from localities."²⁵ Mr. Goldenweizer's authority is "Swanton, Bur. Eth. Rep., 1904-1905 (1908), p. 398." Mr. Frazer²⁶ also quotes that page of Mr. Swanton, but does not say that Mr. Swanton here gives local, not animal, names to the clans of the Tlingit. Mr. Frazer also cites Mr. Swanton's p. 423 sq. Here we find Mr. Swanton averring that Killer Whale, Grizzly Bear, Wolf, and Halibut are in the Wolf phratry, "on the Wolf side," among the Tlingit; while Raven, Frog, Hawk, and Black Whale are on the Raven side. Here are animal names (not precisely as in Mr. Boas' list) within the phratries. But Mr. Swanton does not reckon these animal names as names of "clans"; to "clans" he gives local names in almost every case. To his mind these animal names in Tlingit society denote "crests" not "clans" and with crests we enter a region of confusion.

V

Further south than the Tsimshian dwell the Kwakiutl, of whom the most southerly are called "the Kwakiutl proper." The northern Kwakiutl are divided, says Dr. Boas, into "septs" and "clans." What a "sept" may be I am not certain. The first tribe has "clans" called Beaver, Eagle, Wolf, Salmon, Raven, Killer Whale: the usual totemic names in this region. These totemic clans are exogamous, like those of Mayne's Tsimshians. Descent is in the female line. In the next tribe we find three exogamous animal-named clans: Eagle, Raven, Killer Whale, Beaver, Wolf, and Salmon have vanished, or have never existed. In these two tribes a child is sometimes placed in the father's, not in the mother's clan, as a Dieri father sometimes "gives" his totem to his son, in addition to the inherited maternal totem.³⁹

When we reach the southern Kwakiutl ("the Kwakiutl proper") we are told by Dr. Boas that "patriarchate prevails." This appears to mean that descent is here reckoned not, as in the north, in the female, but in the male line. "We do not find a single clan that has, properly speaking, an animal for its totem; neither do the clans take their name from their crest, nor are there phratries."⁴⁰ As the northern Kwakiutl have animal-named exogamous "clans" with female descent, Dr. Boas now thinks that the northern Kwakiutl "have to a great extent adopted the maternal descent and the division into animal totems of the northern tribes."⁴¹ We do not know, elsewhere, that totemism has ever been borrowed by one tribe from another, especially by a tribe so advanced in culture as the Kwakiutl, and we have no example of a tribe in which the men have given up their social prerogatives, and transmitted them to their nephews in the female line.

Mr. Frazer writes, "The question naturally arises, Are the Kwakiutl passing from maternal institutions to paternal institutions, from mother-kin to father-kin, or in the reverse direction?.. In one passage Dr. Boas seems to incline to the former member of this alternative, that is, to the view that the Kwakiutl are passing, or have passed, from mother-kin, or (as he calls it) matriarchate to father-kin or patriarchate, for he says that "the marriage ceremonies of the Kwakiutl seem to show that originally matriarchate prevailed also among them."⁴² Yet he afterwards adopted with great decision the "contrary view." On these very intricate problems I take leave to quote the statement with which Mr. Goldenweizer has been good enough to favour me.

First, as to descent among the Kwakiutl proper.

"At first, as Mr. Frazer points out (iii. p. 329 sq.). Dr. Boas believed that the Kwakiutl were passing from maternal to paternal descent. Later investigations conducted by Dr. Farrand (cf. F. Boas, The Mythology of the Bella Coola, Jesup North Pacific Expedition, vol. i. p. 121), led to a reversal of that opinion. The main arguments for original paternal descent among the Kwakiutl are three in number. (1) The village communities, which were the original social unit of the Kwakiutl,⁴³ regarded themselves as direct descendants of a mythical ancestor, and not as descendants of the ancestor's sister, which is the case in the legends of the northern tribes, with maternal descent. (Cf. F. Boas, The Kwakiutl, etc., p. 335, where a genealogy is also given.) (2) A number of offices connected with the ceremonies of the secret societies, such as master of ceremonies, etc., are hereditary in the male line (F. Boas, Kwakiutl, etc., p. 431). The Secret Societies, with their dances, are a very ancient institution among the Kwakiutl, and the male inheritance of the above offices is a strong argument for the former prevalence of paternal descent among these people. (3) The form taken by the maternal inheritance of rank, privileges, etc., among the Kwakiutl points in the same direction. When a man marries he receives crests, privileges, etc., from his father-in-law through his wife, but he himself may not use them but must keep them for his son, who, when of proper age, may sing the songs, perform the dances, use the crest, etc., which he thus receives from his mother through the medium of his father. (Cf. F. Boas, Kwakiutl, etc., p. 334.) When the young man marries he must return his privileges to his father, who then gives them to his daughter when she marries. Thus, son-in-law No. 2 receives the privileges, but again may not use them, but keeps them for his son, etc. It appears, then, that the privileges exercised by the young man before marriage are always derived from his mother, but formally he receives them from his father, who acts as a sort of guardian of these privileges until the son is ready for them. Descent here is clearly maternal, but the form of paternal descent is preserved, a plausible condition for a people who; having become maternal, still stick at least in form to the traditional inheritance from the father. If this inference be rejected, the feature becomes quite unaccountable.

"In the sentence, 'The woman's father, on his part, has acquired his privileges in the same manner through his mother' (Frazer, vol. iii. p. 333. note i), the privileges the woman's father exercised as a young man before marriage are meant. The privileges he later acquired through his wife he, of course, could not use, but had to keep them for his son. The phrase, 'each individual inherits the crest of his maternal grandfather' (Frazer, iii. p. 331, note 2), must be similarly interpreted. The crest the individual uses before marriage is meant.

"In connection with the foregoing it must be remembered that another mode of acquiring privileges, crests, songs, etc., was common among the Kwakiutl, viz. by killing the owner (cf. F. Boas, Kwakiutl, etc., p. 424, and elsewhere).

"I also cite the actual words of Dr. Boas. He believes that the intricate law by which 'a purely female line of descent is secured, although only through the medium of the husband,' can only be explained 'as an adaptation of maternal laws by a tribe which was on a paternal stage. I cannot imagine that it is a transition of a maternal society to a paternal society, because there are no relics of the former' (maternal) 'stage beyond those which we find everywhere, and which do not prove that the transition has been recent at all. There is no trace left of an inheritance from the wife's brothers; the young people do not live with the wife's parents. But the most important argument is that the customs cannot have been prevalent in the village communities from which the present tribal system originated, as in these' (village communities) 'the tribe is always designated as the direct descendants of the mythical ancestor. If the village communities had been on the maternal stage, the tribes would have been designated as the descendants of the ancestor's sisters, as is always the case in the legends of the northern tribes.'"⁴⁴

From all this it appears that Dr. Boas believes the Kwakiutl proper to have been once, "on the maternal stage," of which the usual "relics" survive, but why should all such traces survive? Some must disappear, otherwise there could be no transition!

Apparently, in the village communities, the existence of a mythical ancestor, not ancestress, is postulated; while in the northern tribes, with female descent, mythical ancestresses are postulated. But if, among the Kwakiutl proper, male ancestry is now the recognised rule (and it dimly seems to be so), then, as usual, Kwakiutl myth will throw back into the unknown past the institutions of their present state, will say "ancestor," not "ancestress." No argument can be based on traditions which are really explanatory conjectures. There is advanced no valid reason for supposing that the Kwakiutl proper began with descent in the female line, then advanced to the male line, and then doubled back on the female line, and so evolved transmission of crests in the female line, through husbands.

The waverings of the Kwakiutl between the two lines of descent are, in fact, such as we expect to occur when a people has retained, like the Tlingit, Haida, and Tsimshians, the system of female descent after reaching a fair pitch of physical culture, and arriving at wealth, rank, and the attribution of children to the paternal stock.

VI

I now come to give my own opinion as to the ways in which Kwakiutl totemism may have attained its existing peculiarities. It is necessary first to defend my view that the essential thing in totemism – surveying the whole totemic field – is the existence of exogamous kins bearing animal and other such names. Here Mr. Goldenweizer opposes me, saying that "no particular set of features can be taken as characteristic of totemism, for the composition of the totemic complex is variable, nor can any particular feature be regarded as fundamental, for not one of the features does invariably occur in conjunction with others; nor is there any evidence to regard any other feature as primary in order of development, or as of necessity original psychologically."⁴⁵

I have already remarked that this is true; we find human associations, which are not kins or clans, bearing animal and other totemic names, while these associations are not exogamous (the Arunta nation); and we find exogamous sets, kins, or associations which do not bear animal names.

But the co-existence of the exogamous kin with the totemic name of that kin is found in such an immense and overwhelming majority over every other arrangement; the exogamous "totem clan" is so hugely out of proportion in numbers and width of diffusion over the Arunta animal-named non-exogamous associations and other rare exceptions, that we have a right to ask – Are not the exceptions aberrant variations? Have not the Arunta, with non-exogamous sets bearing totemic names, and other peoples with exogamous sets not of totemic names, passed through and out of the usual stage of animal-named exogamous kins? A mere guess that this is so, that the now non-exogamous human sets with totem names have once been exogamous, would be of no value. I must prove, and fortunately I can prove, that it was so.

It is certain, historically, that some exogamous units which now bear non-totemic names, in the past were ordinary totem kins with totemic names. As we can also demonstrate to a certainty that the Arunta have been in, and, for definite reasons, have passed out of, the ordinary stage of exogamous totem kins, we have a right, I think, to say that, normally, the feature of the totemic name is associated with the feature of exogamy, and that the exceptions really prove the rule, for we can show how the exceptions came to vary from the rule.

Mr. Goldenweizer, in a very brief criticism of my own theory of Totemism, given by me in Social Origins (1903), and in The Secret of the Totem (1905), writes "Why is the question, How did the early groups come to be named after the plants and animals? – the real problem? Would not Lang admit that other features may also have been the starting point?" (I not only admit but insist that "other features" were among the starting-points of exogamous totemism.) Among "the other features" Mr. Goldenweizer gives "animal taboos, or a belief in descent from an animal, or primitive hunting regulations, or what not? I am sure that Lang, who is such an adept in following the logos, could without much effort construct a theory of totemism with any one of these elements to start with – a theory as consistent with fact, logic, and the mind of primitive man, as is the theory of names accepted from without."

Now as to the last point, I have written "unessential to my system is the question how the groups got animal names, as long as they got them and did not remember how they got them" (et seq.)⁴⁶ I did show how European and other village groups obtained animal names, namely as sobriquets given from without; and I proved the same origin of the modern names of Siouan "gentes," of two Highland clans; of political parties, religious sects; and so forth.

This mode of obtaining names is a vera causa: that is all: and nobody had remarked on it, in connection with totemism.

Next I cannot "without much effort" (or with any effort) construct a theory of totemism out of (1) "animal taboos." They are imposed for many known and some unknown reasons, and not all totem kins taboo the totem object. Next (2) as I must repeat that "belief in descent from an animal," is only one out of many post-totemic myths explanatory of totemism; I cannot possibly use it as the starting-point of totemism. If Mr. Goldenweizer has read the book which he is criticising, he forgets that I wrote⁴⁷ "it is an error to look for origins in myths about origins," and that I refused to accept as corroboration of my theory an African myth which agrees with my own view.

As to (3) "primitive hunting regulations," Mr. Goldenweizer does not tell us what they were. It is a very common "regulation" that no totem kin may hunt its own totem animal, but to suggest that the totem kin was created by the regulation is to mistake effect for cause.

Finally (4), who can take "or what not" for the starting-point of an investigation? But every totem kin has a totemic name: if there is no totemic name how can we know that we have before us a totem kin? If the Tlingit "clans" be exogamous but not named by totemic names (as Mr. Swanton tells us), then the Tlingit clans are not totemic, now, whatever they may have been in the past: and we are not concerned with them.

Of every totem "clan" the totem name is a universal feature; and therefore I must begin my study from what is universal – the names. Here (though we must not appeal to authority), I have the private satisfaction of being in agreement with Mr. Howitt. The assumption by men of the names of objects "in fact must have been the commencement of totemism," says Mr. Howitt.⁴⁸

I start then, from the totemic names because, – no totemic name, no totemic "clan"! With the totemic name of a social unit in the tribe, I couple exogamy, (though exogamy may exist apart from totemism), because exogamy is always associated with a "clan" of totemic name, except in a very few cases of which the Arunta "nation" is much the most prominent. But it is not to the point, for the Arunta have no totemic clans. Mr. Frazer's latest definition of totemism is "an intimate relation which is supposed to exist between a group of kindred people on the one side and a species of natural or artificial objects on the other…"⁴⁹ Now the Arunta associations of animal names are not (I must keep repeating) kindreds, are not "clans," are not composed of persons who are, "humanly speaking," akin. The totem is not inherited from either parent or through any kinsman or kinswoman. The Arunta bearers of the same totem name, in each case, do not constitute a "clan." This puts the so-called Arunta "totem clans," non-exogamous, out of action as proofs that "totem clans" may be non-exogamous.

Moreover, the non-exogamous Arunta associations bearing totemic names have once been exogamous totem clans. The usages of the Arunta, and their traditions, and the actual facts of their society, prove that their totems were originally hereditary and exogamous.⁵⁰