Atheism and Theism 113hypothesis that it is. Accordingly, unless other factors exclude the hypothesis
of design it is to be preferred over that of chance.
Suppose, however, someone argues that there are infinitely many other
universes, ordered either in parallel or in temporal sequence, and hence that
it is inevitable that one with the fundamental configuration of ours should
exist. To begin with, this needs correction. Even if there were an infinity of
universes it would not be inevitable that this or any other one should be
among them. All one can say is that as the number of universes proceeds
towards infinity the probability of a difference between the actual distribution
and the probable one diminishes to zero. Taking the earlier example of the
pages, if one shuffles and piles them over and over again infinitely many times
then the chances of not getting 1 to 100 diminish. However it is not guar-
anteed that 1 to 100 will eventually result. Infinitely many operations may
never yield the significant ordering. Nonetheless they will make its occurrence
very much less unlikely than if there were only one operation.
The logic of the many worlds response involves postulating an infinity of
actual universes, because while this does not determine that the significant
one will occur it diminishes its improbability. There is another reason for
postulating an infinity rather than just finitely many universes and this is
connected with the next argument I shall be considering, viz. the cosmological
one. For if one envisages an infinity of possibilities, but stipulates that only
some (however many) shall be realized, this invites the question of what
debars the others, or equivalently of what occasions the occurrence of those
that are realized. This then introduces the idea of contingency and of the
need of a source of selection from among possibilities. In order to avoid
this issue, and to eliminate any element of improbability in the occurrence of
this universe, one might suggest that the set of worlds (this included), is the
totality of all possible universes; or one might claim that there is and could
only be one world – the actual one. In either event since it could not fail to
exist no question arises as to the fact of its existence.
Deferring consideration of the cosmological argument, how effective is
the many universes response? Unless it claims that all possibilities are or
must be actualized, it concedes that a finely tuned universe might not have
existed and thereby allows scope for a probability argument for design. Rather
than try to build on that reduced base, however, the theist may query the
coherence of the many universes hypothesis itself. Can it be excluded? The
question is ambiguous. If it asks whether there is any argument to show
that it is contradictory or otherwise impossible then I suspect that there
is not; at any rate I do not have one. However, one might mean less than
that, for we often exclude suggestions on the grounds that they are obscure or
inadequately supported, and here I think there is a significant weakness in
the hypothesis.