[227] Thus, out of 204 kinds of totems, collected by Spencer and Gillen out of a large number of tribes, 188 are animals or plants. The inanimate objects are the boomerang, cold weather, darkness, fire, lightning, the moon, red ochre, resin, salt water, the evening star, a stone, the sun, water, the whirlwind, the wind and hail-stones (Nor. Tr., p. 773. Cf. Frazer, Totemism and Exogamy, I, pp. 253-254).

[228] Frazer (Totemism, pp. 10 and 13) cites a rather large number of cases and puts them in a special group which he calls split-totems, but these are taken from tribes where totemism is greatly altered, such as in Samoa or the tribes of Bengal.

[229] Howitt, Nat. Tr., p. 107.

[230] See the tables collected by Strehlow, op. cit., II, pp. 61-72 (cf. III, pp. xiii-xvii). It is remarkable that these fragmentary totems are taken exclusively from animal totems.

[231] Strehlow, II, pp. 52 and 72.

[232] For example, one of these totems is a cave where an ancestor of the Wild Cat totem rested; another is a subterranean gallery which an ancestor of the Mouse clan dug, etc. (ibid., p. 72).

[233] Nat. Tr., pp. 561 ff. Strehlow, II, p. 71, note 2. Howitt, Nat. Tr., pp. 426 ff.; On Australian Medicine Men, J.A.I., XVI, p. 53; Further Notes on the Australian Class Systems, J.A.I., XVIII, pp. 63 ff.

[234] Thaballa means "laughing boy," according to the translation of Spencer and Gillen. The members of the clan which bear this name think they hear him laughing in the rocks which are his residence (Nor. Tr., pp. 207, 215, 226 note). According to a myth given on p. 422, there was an initial group of mythical Thaballa (cf. p. 208). The clan of the Kati, "full-grown men," as Spencer and Gillen say, seems to be of the same sort (Nor. Tr., p. 207).

[235] Nor. Tr., pp. 226 ff.

[236] Strehlow, II, pp. 71 f. He mentions a totem of the Loritja and Arunta which is very close to the serpent Wollunqua: it is the totem of a mythical water-snake.

[237] This is the case with Klaatsch, in the article already cited (see above, p. 92, n. 3).

[238] As we indicated in the preceding chapter, totemism is at the same time of interest for the question of religion and that of the family, for the clan is a family. In the lower societies, these two problems are very closely connected. But both are so complex that it is indispensable to treat them separately. Also, the primitive family organization cannot be understood before the primitive religious beliefs are known; for the latter serve as the basis of the former. This is why it is necessary to study totemism as a religion before studying the totemic clan as a family group.

[239] See Taplin, The Narrinyeri Tribe, in Curr, II, pp. 244 f.; Howitt, Nat. Tr., p. 131.

[240] Nor. Tr., pp. 163, 169, 170, 172. It is to be noted that in all these tribes, except the Mara and the Anula, the transmission of the totem in the paternal line is only a general rule, which has exceptions.

[241] According to Spencer and Gillen (Nat. Tr., pp. 123 ff.), the soul of the ancestor becomes reincarnate in the body of the mother and becomes the soul of the child; according to Strehlow (II, pp. 51 ff.), the conception, though being the work of the ancestor, does not imply any reincarnation; but in neither interpretation does the totem of the child necessarily depend upon that of the parents.

[242] Nat. Tr., p. 133; Strehlow, II, p. 53.

[243] It is in large part the locality where the mother believes that she conceived which determines the totem of the child. Each totem, as we shall see, has its centre and the ancestors preferably frequent the places serving as centres for their respective totems. The totem of the child is therefore that which belongs to the place where the mother believes that she conceived. As this should generally be in the vicinity of the place which serves as totemic centre for her husband, the child should generally follow the totem of his father. It is undoubtedly this which explains why the greater part of the inhabitants of a given locality belong to the same totem (Nat. Tr., p. 9).

[244] The Secret of the Totem, pp. 159 ff. Cf. Fison and Howitt, Kamilaroi and Kurnai, pp. 40 f.; John Mathews, Eaglehawk and Crow; Thomas, Kinship and Marriage in Australia, pp. 52 ff.

[245] Howitt, Nat. Tr., p. 124.

[246] Howitt, pp. 121, 123, 124; Curr, III, p. 461.

[247] Howitt, p. 126.

[248] Howitt, pp. 98 ff.

[249] Curr, II, p. 165; Brough Smyth, I, p. 423; Howitt, op. cit., p. 429.

[250] Howitt, pp. 101, 102.

[251] J. Mathews, Two Representative Tribes of Queensland, p. 139.

[252] Still other reasons could be given in support of this hypothesis, but it would be necessary to bring in considerations relative to the organization of the family, and we wish to keep these two studies separate. Also this question is only of secondary interest to our subject.

[253] For example, Mukwara, which is the name of a phratry among the Barkinji, the Paruinji and the Milpulko, designates the eagle-hawk, according to Brough Smyth; now one of the clans of this phratry has the eagle-hawk as totem. But here the animal is designated by the word Bilyara. Many cases of the same thing are cited by Lang, op. cit., p. 162.

[254] Spencer and Gillen, Nat. Tr., p. 115. According to Howitt (op. cit., pp. 121 and 454), among the Wotjobaluk, the clan of the pelican is found in the two phratries equally. This fact seems doubtful to us. It is very possible that the two clans may have two varieties of pelicans as totems. Information given by Mathews on the same tribe seems to point to this (Aboriginal Tribes of N.S. Wales and Victoria, in Journal and Proceedings of the Royal Society of N.S. Wales, 1904, pp. 287 f.).

[255] In connection with this question, see our memoir on Le Totémisme, in the Année Sociologique, Vol. V, pp. 82 ff.

[256] On the question of Australian matrimonial classes in general, see our memoir on La Prohibition de l'inceste, in the Année Soc., I, pp. 9 ff., and especially for the tribes with eight classes, L'Organisation matrimoniale des societés Australiennes, in Année Soc., VIII, pp. 118-147.

[257] This principle is not maintained everywhere with an equal strictness. In the central tribes of eight classes notably, beside the class with which marriage is regularly permitted, there is another with which a sort of secondary concubinage is allowed (Spencer and Gillen, Nor. Tr., p. 106). It is the same with certain tribes of four classes. Each class has a choice between the two classes of the other phratry. This is the case with the Kabi (see Mathews, in Curr, III, 162).

[258] See Roth, Ethnological Studies among the North-West-Central Queensland Aborigines, pp. 56 ff.; Palmer, Notes on some Australian Tribes, J.A.I., XIII (1884), pp. 302 ff.

[259] Nevertheless, some tribes are cited where the matrimonial classes bear the names of animals or plants: this is the case with the Kabi (Mathew, Two Representative Tribes, p. 150), the tribes observed by Mrs. Bates (The Marriage Laws and Customs of the West Australian Aborigines, in Victorian Geographical Journal, XXIII-XXIV, p. 47), and perhaps in two tribes observed by Palmer. But these facts are very rare and their significance badly established. Also, it is not surprising that the classes, as well as the sexual groups, should sometimes adopt the names of animals. This exceptional extension of the totemic denominations in no way modifies our conception of totemism.

[260] Perhaps the same explanation is applicable to certain other tribes of the South-East and the East where, if we are to believe the informers of Howitt, totems specially attached to each matrimonial class are to be found. This is the case among the Wiradjuri, the Wakelbura and the Bunta-Murra on the Bulloo River (Howitt, Nat. Tr., pp. 210, 221, 226). However, the evidence collected is suspect, according to his own admission. In fact, it appears from the lists which he has drawn up, that many totems are found equally in the two classes of the same phratry.

The explanation which we propose, after Frazer (Totemism and Exogamy, pp. 531 ff.), raises one difficulty. In principle, each clan and consequently each totem, is represented equally in the two classes of a single phratry, since one of the classes is that of the children and the other that of the parents from whom the former get their totems. So when the clans disappeared, the totemic interdictions which survived should have remained in both matrimonial classes, while in the actual cases cited, each class has its own. Whence comes this differentiation? The example of the Kaiabara (a tribe of southern Queensland) allows us to see how it may have come about. In this tribe, the children have the totem of their mother, but it is particularized by some distinctive mark. If the mother has the black eagle-hawk as totem, the child has the white eagle-hawk (Howitt, Nat. Tr., p. 229). This appears to be the beginning of a tendency for the totems to differentiate themselves according to the matrimonial classes.

[261] A tribe of only a few hundred members frequently has fifty or sixty clans, or even many more. On this point, see Durkheim and Mauss, De quelques formes primitives de classification, in the Année Sociologique, Vol. VI, p. 28, n. 1.

[262] Except among the Pueblo Indians of the South-West, where they are more numerous. See Hodge, Pueblo Indian Clans, in American Anthropologist, 1st series, Vol. IX, pp. 345 ff. It may always be asked whether the groups which have these totems are clans or sub-clans.

[263] See the tables arranged by Morgan, Ancient Society, pp. 153-185.

[264] Krause, Die Tlinkit-Indianer, p. 112; Swanton, Social Condition, Beliefs and Linguistic Relationship of the Tlingit Indians, in XXVIth Rep., p. 308.

[265] Swanton, Contributions to the Ethnology of the Haida, p. 62.

[266] "The distinction between the two clans is absolute in every respect," says Swanton, p. 68; he gives the name clan to what we call phratries. The two phratries, he says elsewhere, are like two foreign nations in their relations to each other.

[267] Among the Haida at least, the totem of the real clans is altered more than that of the phratries. In fact, usage permits a clan to sell or give away the right of bearing its totem, as a result of which each clan has a number of totems, some of which it has in common with other clans (see Swanton, pp. 107 and 268). Since Swanton calls the phratries clans, he is obliged to give the name of family to the real clans, and of household to the regular families. But the real sense of his terminology is not to be doubted.

[268] Journals of two Expeditions in N.W. and W. Australia, II, p. 228.

[269] Kamilaroi and Kurnai, p. 165.

[270] Indian Tribes, I, p. 420; cf. I, p. 52. This etymology is very doubtful. Cf. Handbook of American Indians North of Mexico (Smithsonian Inst. Bur. of Ethnol., Pt. II, s.v., Totem, p. 787).

[271] Schoolcraft, Indian Tribes, III, 184; Garrick Mallery, Picture Writing of the American Indians, in Tenth Report, 1893, p. 377.

[272] Hearne, Journey to the Northern Ocean, p. 148 (quoted from Frazer, Totemism, p. 30).

[273] Charlevoix, Histoire et description de la Nouvelle France, V, p. 329.

[274] Krause, Tlinkit-Indianer, p. 248.

[275] Erminnie A. Smith, Myths of the Iroquois, in Sec. Rep. of the Bur. of Ethnol., p. 78.

[276] Dodge, Our Wild Indians, p. 225.

[277] Powell, Wyandot Government, in First Rep. of the Bur. of Ethnol., 1881, p. 64.

[278] Dorsey, Omaha Sociology, in Third Rep., pp. 229, 240, 248.

[279] Krause, op. cit., pp. 130 f.

[280] Krause, p. 308.

[281] See a photograph of a Haida village in Swanton, op. cit., Pl. IX. Cf. Tylor, Totem Post of the Haida Village of Masset, J.A.I., New Series I, p. 133.

[282] Hill Tout, Report on the Ethnology of the Statlumh of British Columbia, J.A.I., XXXV, p. 155.

[283] Krause, op. cit., p. 230; Swanton, Haida, pp. 129, 135 ff.; Schoolcraft, op. cit., I, pp. 52-53, 337, 356. In the latter case the totem is represented upside down, in sign of mourning. Similar usages are found among the Creek (C. Swan, in Schoolcraft, V, p. 265) and the Delaware (Heckewelder, An Account of the History, Manners and Customs of the Indian Nations who once inhabited Pennsylvania, pp. 246-247).

[284] Spencer and Gillen, Nor. Tr., pp. 168, 537, 540.

[285] Ibid., p. 174.

[286] Brough Smyth, The Aborigines of Victoria, I, p. 99 n.

[287] Brough Smyth, I, p. 284. Strehlow cites a fact of the same sort among the Arunta (III, p. 68).

[288] An Account of the English Colony in N.S. Wales, II, p. 381.

[289] Krause, p. 237.

[290] Swanton, Social Condition, Beliefs and Linguistic Relationship of the Tlingit Indians, in XXVIth Rep., pp. 435 ff.; Boas, The Social Organization and Secret Societies of the Kwakiutl Indians, p. 358.

[291] Frazer, Totemism, p. 26.

[292] Bourke, The Snake Dance of the Moquis of Arizona, p. 229; J. W. Fewkes, The Group of Tusayan Ceremonials called Katcinas, in XVth Rep., 1897, pp. 151-263.

[293] Müller, Geschichte der Amerikanischen Urreligionen, p. 327.

[294] Schoolcraft, op. cit., III, p. 269.

[295] Dorsey, Omaha Sociol., Third Rep., pp. 229, 238, 240, 245.

[296] Spencer and Gillen, Nat. Tr., p. 451.

[297] Spencer and Gillen, Nat. Tr., p. 257.

[298] The meaning of these relations will be seen below (Bk. II, ch. iv).

[299] Spencer and Gillen, Nor. Tr., p. 296.

[300] Howitt, Nat. Tr., pp. 744-746; cf. p. 129.

[301] Kamilaroi and Kurnai, p. 66 n. It is true that other informers contest this fact.

[302] Howitt, Nat. Tr., p. 744.

[303] Swanton, Contributions to the Ethnology of the Haida, pp. 41 ff., Pl. XX and XXI; Boas, The Social Organization of the Kwakiutl, p. 318; Swanton, Tlingit, Pl. XVI ff.—In one place, outside the two ethnographic regions which we are specially studying, these tattooings are put on the animals which belong to the clan. The Bechuana of South Africa are divided into a certain number of clans; there are the people of the crocodile, the buffalo, the monkey, etc. Now the crocodile people, for example, make an incision in the ears of their cattle whose form is like the jaws of this animal (Casalis, Les Basoutos, p. 221). According to Robertson Smith, the same custom existed among the ancient Arabs (Kinship and Marriage in Early Arabia, pp. 212-214).

[304] However, according to Spencer and Gillen, there are some which have no religious sense (see Nat. Tr., pp. 41 f.; Nor. Tr., pp. 45, 54-56).

[305] Among the Arunta, this rule has exceptions which will be explained below.

[306] Spencer and Gillen, Nat. Tr., p. 162; Nor. Tr., pp. 179, 259, 292, 295 f.; Schulze, loc. cit., p. 221. The thing thus represented is not always the totem itself, but one of those things which, being associated to this totem, are regarded as being in the same family of things.

[307] This is the case, for example, among the Warramunga, the Walpari, the Wulmala, the Tjingilli, the Umbaia and the Unmatjera (Nor. Tr., 339, 348). Among the Warramunga, at the moment when the design is executed, the performers address the initiated with the following words: "That mark belongs to your place; do not look out along another place." "This means," say Spencer and Gillen, "that the young man must not interfere with ceremonies belonging to other totems than his own: it also indicates the very close association which is supposed to exist between a man and his totem and any spot especially connected with the totem" (Nor. Tr., p. 584 and n.). Among the Warramunga, the totem is transmitted from father to child, so each locality has its own.

[308] Spencer and Gillen, Nat. Tr., pp. 215, 241, 376.

[309] It will be remembered (see above, p. 107) that in this tribe, the child may have a different totem than his father, his mother, or his relatives in general. Now the relatives on both sides are the performers designated for the ceremonies of initiation. Consequently, since in principle a man can have the quality of performer or officiant only for the ceremonies of his own totem, it follows that in certain cases the rites by which the young man is initiated must be in connection with a totem that is not his own. That is why the paintings made on the body of the novice do not necessarily represent his own totem: cases of this sort will be found in Spencer and Gillen, Nat. Tr., p. 229. That there is an anomaly here is well shown by the fact that the circumcision falls to the totem which predominates in the local group of the initiate, that is to say, to the one which would be the totem of the initiate himself, if the totemic organization were not disturbed, if among the Arunta it were what it is among the Warramunga (see Spencer and Gillen, ibid., p. 219).

The same disturbance has had another consequence. In a general way, its effect is to extend a little the bonds attaching each totem to a special group, since each totem may have members in all the local groups possible, and even in the two phratries. The idea that these ceremonies of a totem might be celebrated by an individual of another totem—an idea which is contrary to the very principles of totemism, as we shall see better after a while—has thus been accepted without too much resistance. It has been admitted that a man to whom a spirit revealed the formula for a ceremony had the right of presiding over it, even when he was not of the totem in question himself (Nat. Tr., p. 519). But that this is an exception to the rule and the product of a sort of toleration is proved by the fact that the beneficiary of the formula does not have the free disposition of it; if he transmits it—and these transmissions are frequent—it can be only to a member of the totem which the rite concerns (Nat. Tr., ibid.).

[310] Nat. Tr., p. 140. In this case, the novice keeps the decoration with which he has thus been adorned until it disappears of itself by the effect of time.

[311] Boas, General Report on the Indians of British Columbia in British Association for the Advancement of Science, Fifth Rep. of the Committee on the N.W. Tribes of the Dominion of Canada, p. 41.

[312] There are also some among the Warramunga, but in smaller numbers than among the Arunta; they do not figure in the totemic ceremonies, though they do have a place in the myths (Nor. Tr., p. 163).

[313] Other names are used by other tribes. We give a generic sense to the Arunta term because it is in this tribe that the churinga have the most important place and have been studied the best.

[314] Strehlow, II, p. 81.

[315] There are a few which have no apparent design (see Spencer and Gillen, Nat. Tr., p. 144).

[316] Nat. Tr., pp. 139 and 648; Strehlow, II, p. 75.

[317] Strehlow, who writes tjurunga, gives a slightly different translation to the word. "This word," he says, "means that which is secret and personal (der eigene geheime). Tju is an old word which means hidden or secret, and runga means that which is my own." But Kempe, who has more authority than Strehlow in this matter, translates tju by great, powerful, sacred (Kempe, Vocabulary of the Tribes inhabiting Macdonell Ranges, s.v. Tju, in Transactions of the R. Society of Victoria, Vol. XIII). At bottom, the translation of Strehlow is not so different from the other as might appear at first glance, for what is secret is hidden from the knowledge of the profane, that is, it is sacred. As for the meaning given to runga, it appears to us very doubtful. The ceremonies of the emu belong to all the members of that clan; all may participate in them; therefore they are not personal to any one of them.

[318] Nat. Tr., pp. 130-132; Strehlow, II, p. 78. A woman who has seen a churinga or a man who has shown one to her are both put to death.

[319] Strehlow calls this place, defined in exactly the same terms as by Spencer and Gillen, arknanaua instead of ertnatulunga (Strehlow, II, p. 78).

[320] Nor. Tr., p. 270; Nat. Tr., p. 140.

[321] Nat. Tr., p. 135.

[322] Strehlow, II, p. 78. However, Strehlow says that if a murderer takes refuge near an ertnatulunga, he is unpityingly pursued there and put to death. We find some difficulty in conciliating this fact with the privilege enjoyed by animals, and ask ourselves if the rigour with which a criminal is treated is not something recent and should not be attributed to a weakening of the taboo which originally protected the ertnatulunga.

[323] Nat. Tr., p. 248.

[324] Ibid., pp. 545 f. Strehlow, II, p. 79. For example, the dust detached by rubbing a churinga with a stone, when dissolved in water, forms a potion which restores health to sick persons.

[325] Nat. Tr., pp. 545 f. Strehlow (II, p. 79) contests this fact.

[326] For example, the churinga of the yam totem, if placed in the soil, make the yams grow (Nor. Tr., p. 275). It has the same power over animals (Strehlow, II, pp. 76, 78; III, pp. 3, 7).

[327] Nat. Tr., p. 135; Strehlow, II, p. 79.

[328] Nor. Tr., p. 278.

[329] Ibid., p. 180.

[330] Nor. Tr., pp. 272 f.

[331] Nat. Tr., p. 135.

[332] One group borrows the churinga of another with the idea that these latter will communicate some of the virtues which are in them and that their presence will quicken the vitality of the individuals and of the group (Nat. Tr., pp. 158 ff.).

[333] Ibid., p. 136.

[334] Each individual is united by a particular bond to a special churinga which assures him his life, and also to those which he has received as a heritage from his parents.

[335] Nat. Tr., p. 154; Nor. Tr., p. 193. The churinga are so thoroughly collective that they take the place of the "message-sticks" with which the messengers of other tribes are provided, when they are sent to summon foreign groups to a ceremony (Nat. Tr., pp. 141 f.).

[336] Ibid., p. 326. It should be remarked that the bull-roarers are used in the same way (Mathews, Aboriginal Tribes of N.S. Wales and Victoria, in Jour. of Roy. Soc. of N.S. Wales, XXXVIII, pp. 307 f.).

[337] Nat. Tr., pp. 161, 259 ff.

[338] Ibid., p. 138.

[339] Strehlow, I, Vorwort. in fine; II, pp. 76, 77 and 82. For the Arunta, it is the body of the ancestor itself; for the Loritja, it is only an image.

[340] When a child has just been born, the mother shows the father the spot where she believes that the soul of the ancestor entered her. The father, accompanied by a few relatives, goes to this spot and looks for the churinga which the ancestor is believed to have left at the moment that he reincarnated himself. If it is found there, some old man of the group undoubtedly put it there (this is the hypothesis of Spencer and Gillen). If they do not find it, a new churinga is made in a determined manner (Nat. Tr., p. 132. Cf. Strehlow, II, p. 80).