----- Original Message ----- From: Brian Scurfield <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, April 16, 2002 6:47 AM Subject: RE: Holodeck guy tries to prove 'Bruno theory'
> In this paper Olum defends the self-indicating assumption which says that > given the fact you exist you should favour hypotheses according to which > many observers exist over hypotheses according to which few observers exist. > He sees the following possible problem: > > "However, one should at least give some consideration to the idea that a > theory which involves a very large number is less likely to be correct that > one which does not. For example, suppose I have a crazy theory that each > planet has 10^10^100 copies of itself on "other planes". Suppose that I (as > cranks often do) believe this theory in spite of the fact that every > reputable scientist thinks it is garbage. I could argue that my theory is > very likely to be correct, because the chance that every reputable scientist > is independently wrong is clearly more than 1 in 10^10^100. To avoid this > conclusion, one must say that the a priori chance that my theory was right > was less that 1 in 10^10^100. It seems hard to have such fantastic > confidence that a theory is wrong, but if we don't allow that we will be > prey to the argument above." (p. 15) > > A case of having your cake and eating it too? Maybe, but the idea that any theory with such large numbers must be given low a priory probability also follows from independent reasoning. E.g. a natural measurure on the set of all possible universes would show this behaviour if it is consistent with Ocham's Razor. Saibal
