Theism and Explanation - Appeared-to-Blogly
Theism and Explanation - Appeared-to-Blogly
Theism and Explanation - Appeared-to-Blogly
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Successful Theistic <strong>Explanation</strong>s 121<br />
But does this mean that just any known fact could corroborate a theory,<br />
given that it is an observed consequence of that theory? This view—the idea<br />
that any positive case of a hypothesis would corroborate it 14 —is sometimes<br />
referred <strong>to</strong> as “Nicod’s criterion” (named after the logician Jean George<br />
Pierre Nicod [1893–1924]). But if Popper’s criterion is <strong>to</strong>o restrictive,<br />
Nicod’s criterion, without further qualifi cation, is <strong>to</strong>o permissive. Firstly, it<br />
would allow corroboration by observed facts that differ from one another<br />
in ways that are irrelevant <strong>to</strong> the hypothesis being tested. 15 Secondly, it<br />
would allow a hypothesis <strong>to</strong> be corroborated by citing further instances of<br />
the same phenomenon we are seeking <strong>to</strong> explain. And thirdly, it leads <strong>to</strong><br />
Carl Hempel’s no<strong>to</strong>rious “raven paradox.” 16 So this criterion needs <strong>to</strong> be<br />
tightened up a little.<br />
A heuristic account. There are a number of ways in which one might<br />
tighten up Nicod’s criterion. The fi rst is John Worrall’s “heuristic view” of<br />
corroboration: the idea that the facts that are considered <strong>to</strong> corroborate a<br />
theory should not be those that were used <strong>to</strong> construct it. 17 There is much<br />
<strong>to</strong> be said for this view, which would at least help <strong>to</strong> eliminate the accumulation<br />
of irrelevant instances. A theory cannot be corroborated by citing<br />
further instances of the phenomena of the type it was devised <strong>to</strong> explain;<br />
it can be corroborated only by citing other kinds of facts. But this view,<br />
<strong>to</strong>o, faces diffi culties. The fi rst is that of deciding which facts were used<br />
in the construction of a theory, which is by no means as simple a task as it<br />
might appear. 18 And the heuristic view does not avoid the raven paradox, 19<br />
which Popper’s original view has the virtue of solving. (If background<br />
knowledge already “predicts” that my yellow cup is a non-black non-raven,<br />
then observing the cup cannot count as a severe test of the theory that “all<br />
ravens are black.” 20 )<br />
An his<strong>to</strong>rical account. There is, however, a second way of tightening up<br />
Nicod’s criterion, namely by taking in<strong>to</strong> account competing theories. I<br />
will call this an “his<strong>to</strong>rical” (as opposed <strong>to</strong> purely logical) account of corroboration.<br />
When our background knowledge includes another potential<br />
explanation of the fact in question, then our hypothesis can be corroborated<br />
if it is able <strong>to</strong> “predict” some fact that cannot be explained by, or<br />
that apparently falsifi es, its predecessor. 21 Whether that fact is known or<br />
unknown is irrelevant.<br />
What if the theory we wish <strong>to</strong> test has no predecessor? In particular, what<br />
if it is the only hypothesis on offer, with regard <strong>to</strong> a particular fact? As we<br />
have seen (4.3.2), proposed theistic explanations are often like this. When<br />
Richard Swinburne argues that the theistic hypothesis would explain the<br />
very existence of laws of nature, 22 he is claiming <strong>to</strong> explain a fact which—<br />
as Elliot Sober notes—“no other hypothesis really engages as a problem.<br />
This is an instance of inference <strong>to</strong> the ‘best’ explanation only in the Pickwickian<br />
sense that just one explanation has been suggested.” 23 Assuming