er these
relatively Supreme Beings, so conceived of by men in very rudimentary
social conditions, can be, as anthropology declares, mere developments
from the belief in ghosts of the dead.

We shall end by venturing to suggest that the savage theory of the soul
may be based, at least in part, on experiences which cannot, at present,
be made to fit into any purely materialistic system of the universe. We
shall also bring evidence tending to prove that the idea of God, in its
earliest known shape, need not logically be derived from the idea of
spirit, however that idea itself may have been attained or evolved. The
conception of God, then, need not be evolved out of reflections on dreams
and 'ghosts.'

If these two positions can be defended with any success, it is obvious
that the whole theory of the Science of Religion will need to be
reconsidered. But it is no less evident that our two positions do not
depend on each other. The first may be regarded as fantastic, or
improbable, or may be 'masked' and left on one side. But the strength of
the second position, derived from evidence of a different character, will
not, therefore, be in any way impaired. Our first position can only be
argued for by dint of evidence highly unpopular in character, and, as a
general rule, condemned by modern science. The evidence is obtained by
what is, at all events, a legitimate anthropological proceeding. We may
follow Mr. Tylor's example, and collect savage _beliefs_ about visions,
hallucinations, 'clairvoyance,' and the acquisition of knowledge
apparently not attainable through the normal channels of sense. We may
then compare these savage beliefs with attested records of similar
_experiences_ among living and educated civilised men. Even if we attain
to no conclusion, or a negative conclusion, as to the actuality and
supernormal character of the alleged experiences, still to compare data of
savage and civilised psychology, or even of savage and civilised illusions
and fables, is decidedly part, though a neglected part, of the function of
anthropological science. The results, whether they do or do not strengthen
our first position, must be curious and instructive, if only as a chapter
in the history of human error. That chapter, too, is concerned with no
mean topic, but with what we may call the X region of our nature. Out of
that region, out of miracle, prophecy, vision, have certainly come forth
the great religions, Christianity and Isla