At 8:11:51 PM on Friday, July 9, 2004,
enlil@... wrote:
> Brian:
>> but you can't exclude it on methodological grounds.
> Statistics is statistics.
This tautology doesn't change the fact that it's a
methodological error to refuse a priori to reconstruct
something simply because it's rare. Unlikely != impossible.
The legitimate points of argument are whether the evidence
actually justifies reconstructing something that is
empirically of low probability, and whether there actually
is evidence that the thing is possible.
The empirical facts -- descriptive statistics -- are
relevant to such questions, but statistics as a discipline
has little if anything to say here.
Brian