Abstract

In this paper I shall attempt to outline a nominalistic theory of mathematical truth. I call my theory nominalistic because it avoids a real (see [4]) ontological commitment to abstract entities. Traditionally, nominalists have found it difficult to justify any reference to infinite collections in mathematics. Even those who have tried to do so have typically restricted themselves to predicative and, thus, denumerable realms. I Indeed, many have linked impredicative definitions to platonism; nominalists have tended to agree with Weyl that impredicative analysis is "a house built on sand" in a "logician's paradise."2 As a result, they have either worried about how much of mathematics empirical science requires (e.g., [34]) or renounced mathematical truth altogether (e.g., [12], [17]). My theory, in contrast, seeks a nominalistic interpretation of the entire body of classical mathematics. It tries to secure for the nominalist not merely the natural numbers but the reals, the ordinals, and inaccessible cardinals as well. If I am right, then nominalists can declare, with Goodman and Quine, that "any system that countenances abstract entities we deem unsatisfactory as a final philosophy" ([17],105), and, at the same time, with Hilbert, that "no one shall drive us out of the paradise which Cantor has created for us" ([20], 141). I cannot hope to convince you, in this small space, that my goal is fully achievable. Here I shall try to justify a smaller contention: that if we can explain our knowledge of logic (and of language generally), we can also explain our knowledge of mathematics. Given a nominalistic theory of logic, therefore, we can construct a nominalistic theory of mathematical truth. I shall thus argue for an epistemological thesis of logicism: any epistemology that suffices for our knowledge of logic also suffices for mathematics