theory of origins. The first form of his theory distinguishes (i.) honorific, (ii.) piacular and (iii.) mystical or sacramental sacrifices; but the latter type is traced back to the same cycle of ideas as that in which the piacular sacrifice originated. (i.) The essential feature of this type was that the god and his worshippers shared the sacrifice and might thus be regarded as commensals, or table companions. The human commensals were the totem-kin, whom Robertson Smith conceived to have been in the habit of sharing a common meal in daily life, or at least of not mixing with other kins. The object of sharing the meal with the god was to renew the blood bond. The victim was the animal of a hostile totem-kin or an animal commonly oliered to the god. The god was originally a stranger, taken into the kin by a rite of blood brotherhood, and this constitutes the dark point of the theory; for Robertson Smith regards the blood bond as relatively late; hence we do not see how the god became associated with the kin. (ii.) The piacular sacrifice arose from the need of atoning for bloodshed within the kinship group; properly speaking, the culprit himself should suffer: should he be unknown or substitute had to be found.

non-human member of the

a sense, therefore, the god

mystical sacrifice the god is

beyond the reach of vengeance, a

This was naturally found in the

totem-kin-the totem animal; in

died for his people. (iii.) In the

himself slain and eaten by his worshippers; In the Religion of the Semites (and ed., 1894) the theory was remodelled so as to overcome the difficulty pointed out above. The god, the victim and the human group are regarded as of the same kin; the animal (totem) is the earlier form of the god; the deity was originally female, for under matrilineal rules the mother alone is of kin to her children, but, with the rise of descent in the male line, the god was transformed into a male. The sacrifice is in its origin a communion; god and worshippers have a bond of kinship between them; but it is liable to be interrupted or its strength diminished. Ceremonies of initiation are the means by which the alliance is established between the deity and the young man, when the latter enters upon the rights of manhood; and the supposed bond of kinship is thus regarded as an artificial union from the outset, so far as the individual is concerned, although Robertson Smith still maintains the theory of the fatherhood' of the god, where it is a question of the origin of the totem-kin. From the communion sacrifice sprang the piaculum, which here becomes a subsidiary form and finds its full explanation in the ideas connected with the mystic union of god and worshippers. For the object of the piaculurn is the re>establishment' of the broken alliance, which was precisely that of the communion sacrifice. With the decline of totemism arose the need for human sacrifice-the only means' of' re-establishing the broken tie of kinship when the animal species was no longer akin to man; "t-This

theory of Robertson Smith's has been attacked from two sides. In the first place, L. Marillier (Rev. de l'h1}st. des religions, xxxvi. 243) argues- that if there was -an original bond of kinship between the god and the kin, there is no need to maintain it by sacrificial rites, 'and cites against Smith's view the practice of totemic groups. To this it might be replied that the real significance of 'initiation ceremonies is still obscure; it is a plausible argument 'that the child does not form part' of the kin till after initiation, but this'argumerit'seems inconclusive, for in West Australia there is solidarity, according to Grey (Journals, 23Q)', between the whole -of the kinship group; whether adult or not; and, moreover, nowhere are rites found which are intended to strengthen the union between a man and his totem by means of the blood bond, unless we include the aberrant totemism of the Arunta (Spencer and Gillen, Native Tribes of Central Australia, 167), who eat their totems in order togain magical powers of increasing the stock of the totem animal. Marillier further argues that if, *on' the other hand, there was no bond between god and people but that of the common meal, it does not appear that the god is a totem god; there is no reason why the animal should have been ia* totem; and in any case this idea; of sacrifice can hardly have been anything but a slow growth and consequently not the origin of the practice. In the second place, |.

I

MM. Hubert and Mauss point out that °Robertson'Smith is far from having established either the historical or the logical connexion between the common meal and the other types of sacrifice; the simplest Semitic forms known to us are the most recently recorded; further their simplicity may mean no more than documentary insufficiency, and in any case does not imply any priority; the piaculum is found side by sidewith the communion. at all times; Moreover, under piaculurn are confused purification, propitiation and expiation's; Smith's contention that purification, whose magical character he recognizes but interprets as late, are not sacrificial, 'is far from conclusive; (0) Building in part on the foundation laid by Robertson Smith, Dr J. G. Frazer hasput forward the view that while the sacrifice of the god mayhave been piacular, it was also intended to preserve his divine life against the inroads of.old»'age.' This theory he exemplifies by two orders of cases, (i.) the putting to death of the man-god, who is often also the king, onwhose health is held' to depend the safety of his people, of the world, »or even of the universe; and (ii.) the annual killing of 'the representative of the spirit of vegetation or of the Corn-spirit (see DEMONOLOG Y). (d) For L. Marillier sacrifice was, at its origin, essentially a magical rite-the liberation by the effusion of a victim's bloodi of a magical force which was to bend the gods'to the will of man; from this arose, underthe influence of cult of theidead, the gift theory of sacrifice. Adopting the theory of W. R.Smith, Marillier'also maintained, but without clearly explaining' the relation of this part of his theory to the preceding, that a human kinship group conceived the idea of allying itself with one god in particular. ' This they did by sacrificing a victim and effecting communion with the god by the application of its blood to the altar; or, more directly, 'by the-sacrifice of thefanirnal-god-and the contact of the sacriflcer 'with its blood. ' A

(e) Dr Westermarck takes the view that human sacrifice is as af rule an act of substitution, 'in that men offer a victimiin the hope of saving themselves; but he 'also recognizes funeral sacrifices of various kinds. ~ Certain sacrifices of animals he explains as intended to transfer a conditional curse. (f) The preceding theories are attempts; in the main, to derive from one source all the forms 'of sacrifice. MM. Hubert' and Mauss, while 'admitting that in all sacrifices is found some idea of purchase for substitution, decline to admit* thatall have issued from one primitive form. *In their view, based on an analysis of Hebrew and Hindu forms of sacrifice, the unity of sacrifice consists in the- immediate 'aim of the ritual, not in the ultimate end to be attained; for we rarely find a rite other than complexi and by the3same sacrifice more than one result may be soughtor attained. The unity 'of procedure consists in the fact that every sacrifice involves 'putting the divine in communication with the profane by an intermediary-*the victim-which may be piacularor honorific, a messenger or a means of divination, a means of alimenting the eternal life of the species or a source of magical energy which the rite diffuses over objects in its neighbourhood., ~ i ~ - V

(g) Our knowledge of primitive forms of sacrifice is meagre; even were it more extensive, it would probably be impossible to determine the origin or origins of sacrifice; for no ritual =has5 necessarily survived unchanged in form and meaning sinceits inception, and even permanence of form cannot be takenlto imply a corresponding permanence of meaning for the worshippers. If, however, we turn to Australia, where sacrifice is iunknown, we find more than"one class of rites in which we 'can trace an idea akin to some forms of sacrifice. just as the German reaper leaves the last ears of corn as an offering to Wodan, so the Australian black offers a portion of a find of honey; in New South ~Wales a pebble is said to have been offered or a number of spears, in'Queensland' the skin removed in forming the body-scars. Thus 'it appears that the gift theory may after all be primitive; the worship of, or care for, the dead' may have supplied' inother areas the motive for the transition from offering to sacrifice or the evolution may have been clue to the

spiritualization of the gods. In" Australia, among the Hottentots,