What has science to do with truth?

Abstract

Recent interest in the problem of verisimilitude stemmed originally from Popper's desire to provide a non-inductive criterion of merit that will select between two false theories.1 But the problem has also been taken up by others2 who are not committed to Popper's anti-in ductivism. Indeed Ilkka Niiniluoto has argued3 that the estimated degree of truthlikeness of a generalisation g which is compatible with evidence e can be equated with the inductive probability of g on e, wherever g is a constituent in Hintikka's sense. It might therefore be worth while to approach the problem afresh, in order to determine quite generally whether a measure of verisimilitude is what is prin cipally needed for the evaluation of scientific progress. I shall argue that it is not. The attractiveness of the verisimilitudinist research-programme springs from the inherent plausibility of the thesis that science has the pursuit of truth as its characteristic objective. "The aim of science", Popperians tell us, "is simply truth; not some epistemically dis tinguished variety of truth, but truth alone. Or, more enterprisingly, more truth, as much truth as can be achieved."4 A concordant thesis is asserted by certain inductivists. "Science at its best", they say, "makes progress towards more and more truthlike theories within conceptual systems with a great unifying power."5 To attack such a doctrine which I shall hereafter refer to as 'the truth doctrine' has the appearance of intellectual sacrilege, almost of academic Satanism. It is as if, instead of Iago's "Evil, be thou my good", one were proclaiming "Falsehood, be thou my truth". Surely those who are committed to the advancement of learning, or to the useful application of intellectual effort, should uphold the sanctity of truth and the ideal of maximising its discovery? Yes, indeed; but it is nevertheless misleading to characterise the aim of science as the pursuit of truth or verisimilitude. Nor is the case for objecting to the truth doctrine based on some anti-realist argument for rejecting any conception of scientific achievements in propositional or linguistic