Semantic Naturalism and the Liar

Abstract

Few metaphysicians or philosophers of language have seriously studied the literature on the Liar paradox, let alone attempted to work out its implications for more general issues in their fields. It has generally been supposed that one can first develop a general theory of the metaphysics of truth, for example, and then later adjust one's theory only slightly, if at all, to deal with supposed 'pathologies' like the semantical paradoxes. The Liar sentence has usually been thought to be somehow meaningless, or at any rate to say nothing about the physical world. Thus, philosophers have often assumed that, at the very least, theories of truth for meaningful sentences about the physical world can be developed without taking the Liar paradox into account. In this paper I will argue that the Liar in fact gives us reason to reject a metaphysical theory of truth that I will call 'semantic naturalism' the view that there are natural facts about which tokens of sentences about the natural world are true. More precisely, in this paper semantic naturalism will be the view that it is possible to consistently introduce into any language adequate for natural science a predicate 'has N' of sentence tokens that satisfies the following two conditions: