Scepticism and Anti-Realism

Abstract

My aim in this paper is to counteract the belief that Michael Dummett's anti-realist argument provides a refutation of scepticism, or at least a very deep insight into the nature of scepticism. I argue here that it provides neither. I should say at the outset that I believe Dummett's anti-realism, properly understood, is probably true; I am disputing not its truth but its alleged implications. I am also unwilling to present my negative conclusions as criticisms of Dummett himself, who has refrained from endorsing the claims of epistemological interest made for his argument by others. But he has, I must admit, made some remarks that encourage those claims. Dummett describes analytic philosophy 'as that which follows Frege in accepting that the philosophy of language is the foundation for the rest of the subject' ('Can Analytical Philosophy be Systematic, and Ought it to Be?', p. 441). 1 Before Frege, he explains, epistemology was the foundation of philosophy: 'From the time of Descartes until very recently the first question for philosophy was what we can know and how we can justify our claim to this knowledge, and the fundamental philosophical problem was how far scepticism can be refuted and how far it must be admitted' ('Frege's Philosophy', p. 89). Analytic philosophy, he believes, has liberated us from the Cartesian preoccupation: 'We can get on with ... whatever interests us', he writes, 'without first having undertaken any epistemological inquiry at all' (ibid.). These remarks have led some of his readers to suppose that antirealism offers a roundabout but improved of doing anti-sceptical epistemology. Richard Rorty thinks that 'Dummett sees philosophy of language as foundational because he sees epistemological issues now, at last, being formulated correctly as issues within the theory of meaning. '2 Dummett 'agrees with Descartes about the importance of the issues which emerged out of the way of ideas', Rorty explains, 'but he thinks that we have only recently been able to state them properly'.3 Dummett, as Rorty interprets him, believes that once the old questions are expressed in the right way-as questions in the philosophy of language-analytic philosophers will succeed where the old epistemologists failed. Yet I suspect that Dummett's attitude toward the old questions is actually very close to Rorty's own; the point of Dummett's remarks seems to be that when