us argue from the nature of effects to the nature
of the agency. The first impulse would be to ascribe every intelligent
effect to some human agency, but other circumstances would subsequently
incline the savage reluctantly to divest the agent of one or more of the
limitations of humanity, and to clothe him with preter-human attributes.
Nearly all the supernormal phenomena believed in by primitive man--so
far as we can judge of him from contemporary savagery--would suggest the
agency of an invisible man; clairvoyance, and other manifestations of
preternatural knowledge, would suggest independence of the senses in the
acquisition of knowledge; every kind of "miracle" would bespeak an
extension of power over physical nature beyond human wont; while all
these together would point to that freedom from the trammels of space
and time, which is of the very essence of immaterial or spiritual
subsistence. Thus, by a gradual process of dehumanization, the mind
would be instinctively led from the notion of a man magnified in all
excellences and refined from all limitations, to the conception of
spirit. But coexistently with this progress of the reason, the
imagination would ever strain to clothe the thought in bodily form as
far as possible, and would cling to the notions suggested by dreams and
waking hallucinations, while language, after its wont, would speak of
the spirit as the _umbra_, the _imago_, the shadow, the breath, the
attenuated replica of the body. Thus we find among all men, savage and
civilized, a certain unsteadiness in their notion of spirit, whether
created or divine--a continual tendency to corruption and
anthropomorphism, due to the conflict between reason and imagination,
resulting so often in the domination of the latter.

For this view of the subject it is not necessary that we should admit
the preternatural character of the phenomena which form the
subject-matter of psychical research, but only that we should
acknowledge the hardly disputable fact that belief in such marvels is
universal and persistent among savages--a fact which science is bound by
its own principles to explain, and not to ignore. Whether, as Mr. Lang
seems inclined to think, among much illusion, chicanery, and ignorance,
there may not be truth enough to make the inference of an X-world
legitimate, whether the said universality, persistence, and
recrudescence of this seeming credulity can be accounted for in any
other satisfactory way, i