138 From Spell to Prayer.
" proceeds upon " such and such " assumptions " ; and so on.^ Now on the face of them these appear to be glaring instances of what is known as "the psychologist's fallacy." The standpoint of the observer seems to be confused with the standpoint of the mind under observation. But there are indications that Dr. Frazer expects us to make the necessary allowance for his metaphorical diction. Thus one of the "assumptions " of magic is said to consist in a " faith" that whilst "real and firm" is nevertheless "im- plicit." ^ Meanwhile, from the point of view of the psycho- logical purist, implicit, that is, unconscious, inferences, assumptions, and so on, are little better than hybrids. Now doubtless a considerable amount of real inference may be operative at certain stages in the development of magic. Nay, certain forms of magic may even be found to have originated in a theorising about causes that did not arise out of practice save indirectly, and was the immediate fruit of reflection. I refer more especially to divination, if divination is to be classed under magic, as Dr. Tylor thinks that it should. ^"^ But, speaking generally, the working principle we had better adopt as inquirers into the origin of magic is, I suggest, the following : to expect the theory to grow out of the practice, rather than the other way about ; to try to start from a savage Monsieur Jourdain who talks prose whilst yet unaware that he is doing so.
In what follows I shall seek to observe this working principle. Meanwhile, I cannot pretend to a systematic and all-inclusive treatment of a subject which, for me, I confess, has at present no well-marked limits. Dr. Frazer's division of magic into two kinds, imitative and sympathetic,^^ is highly convenient for analysis, but I am not so sure that it directly subserves genesis. Not to speak of the question already touched on whether divination falls under magic,
8 G. B.,- 49. 3 G'. B.- Cf. 62 wiih 61.
"" Sec his article " Magic " in Emycl. Brit, (iiinlli edit.) " (7. ^.,-'i.,9.
