Mark Gardiner

Realism has traditionally been a philosophical doctrine embodying an ontological element asserting the existence of various types of entities and a meta-theoretic element asserting that the existence of those entities is independent of our knowledge of their existence. Anti-realism, on the other hand, denies that the existence of objects is independent of our knowledge. ;Recently, attempts have been made to reinterpret the basic realist/anti-realist dispute in semantic terms. Basically, realism would be the view that the truth of sentences are independent of our knowledge of their truth-values. Anti-realism, on the other hand, would hold that truth is not so independent of our knowledge. ;Michael Dummett and Hilary Putnam have presented two of the most famous extended semantic criticisms of metaphysical realism. Dummett argues that realism is committed to an unacceptable theory of meaning. Putnam argues that realism rests upon incoherent assumptions regarding truth and reference. ;Unlike many commentators, I accept basic Dummettian constraints. I argue, however, that his conclusions do not follow. Not only can the semantic realist conform to his constraints, a realist construal of truth is in fact ineliminable in such an account. Thus, I turn Dummett's framework against its own conclusions. ;Regarding Putnam, I proceed by rejecting his premises. I show that the arguments he constructs do not support the claim of incoherence levelled at metaphysical realism. Often, indeed, his arguments, if carefully understood, actually support realism. ;I thus conclude that the two most famous and formidable attempts to reject metaphysical realism on the basis of semantic considerations fail. As such, there is no reason to abandon realism traditionally understood