Rationality and Disconfirmation

Abstract

In a recent article in Social Studies of Science,' Michael Mahoney argues that scientists are not the paragons of intellectual and moral virtue that some have claimed them to be, but share to a considerable extent the vices of more ordinary folk. Readers of this journal will be familiar with the general range of considerations which Professor Mahoney adduces, so I shall not rehash them here. I wish, rather, to draw attention to some features of that section of his article which deals with the rationality of scientists. His discussion of this topic is flawed, it seems to me, by inaccuracies in the use of logical terminology, by a too-simple philosophical account of rationality, and by his appeal to an experiment of dubious design. The inaccuracy of terminology, which is even more apparent in NMahoney's earlier book, Scientist as Subject,2 is essentially a surface problem, but it needs to be set straight. For example, Mahoney asserts that 'Modus ponens and modus tollens are the only valid forms of inference, and the latter appears to be the only real option for the practising scientist.'3 Strictly speaking, modusponens and modus tollens are only two of infinitely many valid forms of deductive inference, and those other valid forms are not reducible to modus ponens and modus tollens even if we confine ourselves to the relatively simple forms of inference found in propositional logic. Far from its being the case that rnodus tollens is the 'only real option for the practising scientist,' a scientist (or anyone else) who confined himself to modus tollens would scarcely be able to engage in any reasoning whatever.4 Mahoney's confusion seems to have its root in a mistaken identification of the truth-functional connective '-_.,commonly used to represent the weakest construal of the English 'If...then _', with the much stronger 'implies' or 'entails'. Thus, in Scientist as Subject, he says: