Putnam's Paradox: Metaphysical Realism Revamped and Evaded*

Hilary Putnam's argument against metaphysical realism (commonly referred to as the "model theoretic argument") has now enjoyed two decades of discussion. 1 The text is rich and contains variously construable arguments against variously construed philosophical positions. David Lewis isolated one argument and called it "Putnam's Paradox".2 That argument is clear and concise; so is the paradoxical conclusion it purports to demonstrate; and so is Lewis' paradoxavoiding solution. His solution involves a position I call "anti-nominalism": not only are classes real, but they are divided into arbitrary and 'natural' classes. The natural classes 'carve nature at the joints', being (as other philosophers might say) the extensions of 'real' properties, universals, or Forms.3 Thus the argument was turned, in effect, into support for a metaphysical realism stronger than Putnam envisaged. I will offer a different way to look at Putnam's model theoretic argument. If we insist on discussing language solely in terms of a relation between words and things, we may well be forced into a metaphysical realist point of view, on pain of paradox. But on the level of pragmatics, in a discussion of language that also addresses the roles of user and use, the air of paradox dissolves all by itself. I shall also try to show how we can resist Lewis' argument, derived from Putnam's, for anti-nominalism. No metaphysical postulates will be needed to avoid the threatened disasters of self-understanding.