dividual man, and again such expressions as Pietas Legionis,
by analogy with which von Domaszewski, wrongly as I think, would explain
those we have been discussing.[314]

Before leaving this complicated and cloudy system of divine
nomenclature, it is as well to ask the question once more, even if we
cannot answer it, whether if left to itself it might have developed into
a polytheistic system of personal deities. I will give my own opinion
for what it is worth. I do not think that such a result could have been
reached without the magic touch of the Greek poet and artist, or the
arrival of Greek deities and their images in Latium. Professor Sayce, in
his Gifford lectures on the religion of Babylonia, has shown how the
non-Semitic Sumerians knew only of spirits and demons until the Semite
arrived in the Persian Gulf with his personal gods of both sexes;[315]
and I gather that he does not suppose that without such immigration the
Sumerian ideas of divinity could have become personalised. The question
is not exactly the same at Rome; for there the spirit world had passed
into the hands of an organised priesthood occupied with ritual, and
especially with its terminological aspect; and the chance of
personalisation, if it were there at all, lay in the importance of the
functional name. But the question is after all beside the mark; we shall
see what happened when the Greeks arrived. We may be content at present
to note the fact that they found the functional terminology sufficiently
advanced to take advantage of it, and to revolutionise the whole Roman
conception of the divine.

Dr. Frazer gives me an opportunity of adverting to another point bearing
on the question we are discussing,--the way in which the old Roman
thought of his deities. "It is difficult," he says,[316] "to deny that
the epithets Pater and Mater, which the Romans bestow on so many of
their gods, do really imply paternity and maternity; if this implication
be admitted, the inference appears to be inevitable that these divine
beings were supposed to exercise sexual functions, etc." In a footnote
he adds a number of formidable-looking references, meant, I suppose, to
prove this point. I have closely examined these passages; what they do
prove is simply that many deities were called Pater and Mater. Not one
even suggests that paternity and maternity were in such cases to be
understood literally and, so to speak, physically. The two that come
nearest to what he