And now it may seem high time to pass on from the sociological and external view that has hitherto been taken of primitive religion to a psychological view of it--one that should endeavour to disclose the hidden motives, the spiritual sources, of the beliefs that underlie and sustain the customary practices. But precisely at this point the anthropological treatment of religion is apt to prove unsatisfactory. History can record that such and such is done with far more certainty than that such and such a state of mind accompanies and inspires the doing. Besides, the savage is no authority on the why and wherefore of his customs. "However else would a reasonable being think of acting?" is his sufficient reason, as we have already seen. Not but what the higher minds amongst savages reflect in their own way upon the meaning of their customs and rites. But most of this reflection is no more than an elaborate "justification after the event." The mind invents what Mr. Kipling would call a "Just-so story" to account for something already there. How it might have come about, not how it did come about, is all that the professed explanation amounts to. And when it comes to choosing amongst mere possibilities, the anthropologist, instead of consulting the savage, may just as well endeavour to do it for himself. Now anthropological theories of the origin of religion seem to me to go wrong mainly because they seek to simplify too much. Having got down to what they take to be a root-idea, they straightway proclaim it _the_ root-idea. I believe that religion has just as few, or as many, roots as human life and mind.