Abstract

There is a certain form of philosophical argument that characteristically begins, "we have no reason to suppose that..." and goes on to deny some proposition we took to be well supported. The claim that there is no inductive argument takes the form: the sample on which such an argument would be based is taken from a special part of the population; but we have no reason to suppose that this special part of the population is not very different from the population as a whole. This claim is refuted by offering a plausible conception of probability and randomness under which the inductive arguments would be undermined only if we had reason to believe that the sample and the population differed