FAITH AND REASON 349either 32 in 98 or 33 in 98, but we don’t know which. The same goes,
relative to our knowledge, for the odds of our picking a blue marble. We
may also infer that either the odds of our picking a red marble or a green
one are better than our picking a blue one or a green one or the odds of our
picking a blue one or a green one are better than our picking a green one or
a red one; we don’t know which. In such cases, where natural laws are not
directly relevant, we can quantify both the objective odds and what the
odds are, relative to our knowledge.
Very often, we can do neither. What the odds are, if any, that my
computer is on, my window is closed, or my dog asleep are not quantifiable
in any real sense. We can say that, since the computer seems to be on, my
window appears to be closed, and my dog looks like he is asleep, the odds
are better than 5 that these things are so. But that is either metaphorical or
else a matter of forcing the notion of quantifiable probability onto a case
where it does not fit.
In cases in which quantitative probability is inapplicable, many
philosophers nonetheless use the calculus that would apply were the
probabilities in question quantifiable. The assumption is that there are
non-quantifiable probabilities and that non-quantifiable probabilities
behave as do quantifiable probabilities. We can call this the assumption of
universal probabilism. The issues that concern us in the pages ahead do not
concern quantifiable probabilities. We can remain neutral about whether
the assumption of universal probabilism is true. What requires saying can
be said using such notions as evidence and entailment.
The epistemic status of religious belief
A religious tradition that involved only unbreakable beliefs would be an
odd enterprise. Imagine a tradition that accepted mathematical truths of
the sort 1 and 1 are 2, 1 and 2 are 3, 2 and 2 are 4, 2 and 3 are 5 and so on
and propositions of the sort I exist, I am conscious and the like. Their
resources for stating, within their chosen framework, either a universal
human problem or a solution thereto would be highly limited. No delicate
propositions being permitted to stain the security of their tradition, there
could be no descriptions of institutions or practices, and so no way of
gathering or behaving included in this austere imaginary religion. No such
proposition as Kim needs salvation could be formulated, since that
proposition is neither a necessary truth nor a belief-entailed truth.^7 The
very understandable fact is that no religious tradition has ever limited
itself to such a slender doctrinal base. But this fact is no defect in religious
traditions. Nor has any political, scientific, or academic tradition ever