FOM: Reply to Machover on Conclusiveness -- addendum JSHIPMAN at
Tue Dec 16 11:11:16 EST 1997

My example has the weakness that the proof that Rabin's algorithm behaves as
advertised is more complicated than Euclid's proof, so the subjective
uncertainty is indeed greater in the case of a statement "n is prime" that is
"proved" by running a random algorithm than in the case of the statement "for
any n there is a prime > n".  But Rabin's proof is quite easy, and so any
theorem whose proof is significantly more complicated than that is therefore
more uncertain than a statement of the form "n is prime" that is "proved" by
running Rabin's algorithm, unless you insist the subjective probability you are
mistaken really is less than 10^-300.   -- Joe Shipman

More information about the FOM mailing list