order. This would yield a probability like at least 0.00000033 that an intelligent designer
of
end p.133
the universe would produce a universe exhibiting order, which is so much higher than the
astronomically small probability on Humean design-free intuitions that it significantly
increases the ratio of the probability of the design hypothesis to the Humean hypothesis.
An alternative reply to the Swinburne argument is to invoke MUAP. Recall that MUAP
posits infinitely many universes but notes that there is a selection effect: we can observe
only a universe that has observers in it. Now, a universe that for the most part displays
causal regularity is a necessary prerequisite for there to exist finite knowers and agents,
since empirical knowledge depends on identifying persisting objects. If so, then we have
no right to be surprised at the order in the universe given a many-universes theory.
Swinburne (1968) attacks the MUAP reply to his argument by noting that it is at most
order in the past, and even then only in our local neighborhood, that is required for
knowers and observers. Thus, even if there are many universes and we preselect for those
that contain observers, nonetheless on Humean grounds we should still find future order,
and order outside our local neighborhood, to be quite improbable. To see this more
clearly, suppose in our toy model above we preselect first for those universes where the
minimal condition for regularity is satisfied for the first fifty time steps. Nonetheless,
only fewer than one in 10^50 of these universes continues being regular for the next fifty
steps. Indeed, on a Humean MUAP account, we would expect future disorder to be
highly probable, and hence as order continues to be observed, the Humean MUAP reply
becomes more and more disconfirmed. Likewise, order outside our galaxy disconfirms
the Humean MUAP reply.
Observe that in a number of the nondeductive teleological arguments, issues of
probability theory require further investigation. We intuitively feel that it is highly
improbable prima facie that there be a nondesigned universe that exhibits regular
lawlikeness. But making this intuition precise is a nightmare. There are infinitely many
possible universes that exhibit lawlike regularity and infinitely many that do not. The
infinite numbers here may even be beyond cardinality (for instance, it has been shown
that the collection of all possible worlds is not a set and hence lacks cardinality; see Pruss
2001). Perhaps the argument can be made only on an intuitive level, on the same intuitive
level at which we say that it is highly unlikely that a given integer about which nothing
more is known is in fact prime even though the cardinality of the set of prime numbers is
the same as that of the integers.
Likewise, the thorny issue of how initially plausible the hypothesis of the existence of a
designer is to someone needs to be discussed. If one thinks that the existence of a
designer has astronomically small epistemic probability, then one will not be impressed
by arguments showing that some form of complexity has a similarly small probability of
arising by chance. However, few reasonable people think that the existence of God has a
probability as low as 10−^100.
end p.134