bewilderingly various matter; while in an Appendix he at last inclines to a totally different theory about him which at once renders nearly all these two volumes irrelevant. Elsewhere (VII, 2, p. 218) he embodies in the text 'a suggested theory of Totemism' in which he no longer fully believes, 'because it serves as a convenient peg on which to hang a collection of facts which are much more valuable than any theory of mine.' This is not quite worthy of the dignity of a magnum opus; and he does himself an injustice. Yet his words recognise the necessity of some theory, without which our perception, memory and sense of values would be drowned and lost beneath the flood of accumulated facts. A fact by itself is not yet of any value. past and present history of mankind contains potentially billions of facts,' most of which are useless, and none of which are useful until interpreted and illumined by some theory or hypothesis. Even 'The Golden Bough,' by far the richest storehouse of anthropological facts that has ever been accumulated, is only a selection; and every selection implies a theory.

The

We cannot, then, be indifferent to theories; but at the same time it is well to realise their inadequacy. Sir James Frazer is modest enough about his own; but it is not clear that he recognises with sufficient acuteness how inadequate are his own theories, as well as others that have been put forward, to explain nine-tenths of the facts in certain domains of anthropology, especially mythology and ritual. Let us accept and apply to the phenomena of ritual and myth all the most prevalent hypotheses, theories of sun-worship, weather-magic, fear of ghosts and witchcraft, ideas of purification and taboo, totemism and the worship of animals; then we are pulled up short by a question of detail-'Why do savages knock out their two front teeth?' or, 'Why must the priestess of Athena at Athens abstain from native cheese?' or by any other of the thousands as embarrassing as these in Dr Frazer's collection; and all our theories fall helpless. Anthropology can show the leading ideas in such a priestly code as Leviticus, but no theory can rationalise more than a fraction of the amazing details. Nothing is more eccentric and baffling than the ceremony of a savage funeral. Nothing in

our writer's discussion on the worship and reverence paid to animals explains why men should hunt the wren with such curious rites (v, 2, p. 317). Mythology is, even more than ritual, elusive of the net of the theorist, especially if he is narrow-minded enough to work on the sole theory that all myth comes from ritual. Dr Frazer is too much of a veteran to cherish that illusion; but it may be that inadvertently he encourages it in the tyro.* He puts forth, for instance, a theory of Balder and thinks that, if accepted, it explains the myth. We feel that at best it only explains a fraction of it; the myth of Balder is too rich and original to be explained as the 'text-book' of any conceivable 'sacred drama.' It is the fault of much modern research into the origins of Greek mythology, that, when it has discovered and proclaimed the origin of a particular myth, it leaves all that is really interesting in the myth unexplained and more mysterious than before. A capricious fancy, at once exuberant and meticulous, appears to have prevailed in primitive ritual and myth-making. It is our duty to endeavour to systematise their products by reference to general principles that regulate them; but it is also our present duty to confess how slight our success has been and how evasive the detail is.

Our author's theories are mainly concerned with magic, religion and sociology in the broadest sense of these terms. Now, it is a vice of much modern anthropology that it blurs the distinction between magic and religion and fails to arrive at any precise definition of them. This cannot be said of our writer, who gives a clear and workable account of the difference between the two spheres (Part 1, 1, p. 222). By religion he understands a propitiation or conciliation of powers superior to man which are believed to direct and control the course of nature and of human life.' Magic is not defined with such precision in this treatise; but he uses the term to denote a mode of control or compulsion exercised upon the outer world (including men and gods) according to a law of sympathy between phenomena; which sympathy he analyses into the ideas of homoeopathy and contiguity. His analysis is a reasonable

* Cf. vII, 2, p. 88, where he gives a too narrow definition of myth.'

interpretation of the facts from the point of view of the modern observer; but it may be thought to fail as a vivid representation of what is going on in the mind of the savage magician. His theory might have been more complete and more real, if he had engrafted upon it Dr Marett's exposition of magic in his 'Threshold of Religion' as an ebullition of will-power or 'mana,' using such means as mesmeric words, gestures, and actions. Such a view will protect us against the temptation which is strong in Sir James Frazer to affiliate the modern man of science to the primitive magician, as if the point of view of both in respect of the outer world were the same, each assuming the invariability of nature and its obedience to certain fixed laws, while the votary of religion regards it as variable and depending on the capricious or incalculable will of a personal power. might be truer to say that the magician is distinguished from both the other characters by his egoism, his selfconfident assertion of force; and that both he and the typical religious person are distinct from the man of science in that they consider the world of nature as an elastic and sensitive medium, indefinitely responsive at any point to an immediate discharge of will-power.

It

As regards the relations between magic and religion in the history of man, our writer is well aware of their deep interfusion; but he maintains the thesis, which is scarcely susceptible of historic proof, that in our human history an age of magic preceded an age of religion. At the same time he strongly and clearly refuses to regard religion as in any sense evolved from magic; for the rise of religion involved a self-abnegation, a breach with the past, a new point of view which was adopted by the higher minds as they discovered the futility of the old. That the human race has passed through a wholly nonreligious era is a quite defensible hypothesis; but our author's proofs of it in the fourth chapter of Part I (Vol. 1) will hardly be considered adequate. For instance, such a statement as that on p. 233-'obviously the conception of personal agents is more complex than a simple recognition of the similarity and contiguity of ideas'—is hardly obvious or convincing as phrased.

As regards his general theories on religion and its progress through the various stages of lower and higher

life, there is some characteristic vagueness arising from lack of clear definition, He has no perception or imagination of a possible preanimistic mode of religious feeling; he is vague concerning the meaning of fetichism, confusing it with pure animism. For instance, he defines the former as 'the view that the fruits of the earth and things in general are divine or animated by powerful spirits' (IV, 2, p. 24); and he is occasionally loose in his application of the term 'god.' When he deals with the interesting question concerning the genesis of the idea of 'gods,' he tries to preserve an open mind; and in an excellent preface to his treatise on 'Spirits of the Corn and of the Wild' (Part v), he warns his readers against attributing too much weight to the quantity of the evidence that he there accumulates, and against the narrow view that each and every deity was in germ and origin a vegetation-spirit. He rightly proclaims his belief in many different origins of deities. Yet here and there, we may think, his early and life-long devotion to the theory of Mannhardt, the discoverer and the champion of the Corn-dæmon, still leads him astray; for instance, it misleads him into a jejune and narrow-viewed account of Demeter and Persephone (v, 1), who in origin and throughout their career are vaster potencies than the Corn-Mother or the Corn-Maiden, and whose careers cannot be wholly explained, as he would explain them, by reference to that trifling peasant-fetich.

Though he strives to be open-minded, studies so wide and copious as Sir James Frazer's naturally entitle him to a predilection for certain views. There are two concerning the genesis of gods that are salient in these volumes. The first, to which, I think, he has only been inclined in recent years, we may call the Euhemeristic. At the end of the two volumes, somewhat daringly entitled 'Balder the Beautiful,' he adds an Appendix on 'African Balders,' in which he finally gives his adhesion to the opinion that Balder was a real man. Again, in Part IV, after sketching the religion of Osiris, he comes to favour the suggestion that Osiris also was once a real personage, in fact to be identified with Khent, an Egyptian king of the first dynasty, who was buried at Abydos. Now, the Greek who gave his name to this theory of the human and historical genesis of gods may have been

a very foolish theorist; his followers in antiquity were undoubtedly among the most fatuous of the later writers. But this ought not to prejudice us against the tenability of the theory as an abstract proposition. Like other theories, it has its turn to be true; in fact, within historic times, as Sir Alfred Lyall in his 'Asiatic Studies' has shown in respect of India and China, and recent anthropologists have proved in respect of certain African communities, it gives us an undoubted vera causa of the growth of gods. Therefore we have the right to posit it as an operative cause in the prehistoric period.

It is only a slight personal misfortune for our author that, if his final theory about Balder is right, the title he gives to the two volumes of Part VII becomes even more irrelevant than before. But what alone concerns us here is to ask whether he gives us any criteria, or whether any such can be found, for determining, in any particular case, whether a god was once a real man. I believe such criteria exist and can be stated. But criticism and sharp definition of criteria are not our writer's strong points; the vaguely synthetic and lax habit of the roving comparative method has led him to discover Balders in Africa, who are not Balders except at one point, and who do not help our question at all; when he ought to have spent more intensive study on Scandinavian religion and the Balder-myth. The reader will find the theory of a real man-Balder more critically set forth and the evidence for it seriously considered in Golther's 'Handbuch der Germanischen Mythologie,' published in 1895.

A further reflection bearing on religion may occur to the reader of these volumes on Balder; and it is strange that it did not occur to the writer. Whether a god was once man or not man, in popular mythology many human traits and attributes are sure to be attached to him, which are the same as those that belong in the real world to men and heroes, for instance, a coat of mail, a spear, a wife, a horse, a ship. It was an actual custom for a great Norse hero to drink beer, and on his death to be burnt on his viking-ship. Therefore Odin's gods drink beer, and Balder is burnt on his ship. It is the real heroic life of the period that suggested this trait, not the peasants' midsummer fires, with which Balder is not known to have had any concern. Sir James Frazer's