Some recent attempts to refute contemporary antirealism are puzzling in at least two important respects: Firstly, it is puzzling how philoso? phers who represent opposite sides of the realist-antirealist debate as this pertains to recent semantic theory, can both be attacked as antirealists and moreover, be attacked on fairly analogous grounds. Thus, we find Michael Devitt, in his 'Dummett's Anti-Realism' offering a refutation of Dummett; and we find Crawford Elder, in his 'The Case Against Irrealism' attacking philosophers like Quine and Davidson as irrealists or antirealists.1 This first puzzle can be cleared up by pointing to an often ignored ambiguity in the recent use of terms like 'realism' and 'antirealism'. But secondly, there is another puzzle having to do with the fact that some of these recent refutations of antirealism seem to succeed, if they succeed, only in virtue of another use or meaning for these terms-a use that seems quite irrelevant to the issues at hand. I think that this puzzle is more serious. Accordingly, I shall first discuss the current (and understandable) ambiguity in the use of terms like 'realism' and 'antirealism', and then address what I consider to be the illegitimate introduction of another meaning for these terms as this occurs, tacitly or otherwise, in some of the recent refutations of antirealism. The contemporary use of the term 'antirealism' is ambiguous-at least in the context of recent semantic theory. On the one hand, the term is associated with philosophers in the camp of Michael Dummett who apply the term primarily to their epistemic notion of truth and derivatively to a theory of meaning that explicates sentence meaning in terms of warranted assertibility or verificationist conditions. This is done in order to satisfy the constraint that any adequate theory must account for a speaker's acquisition and understanding of sentence meaning as this is manifested in publicly discernible speech dis? positions and linguistic usage. In this sense, antirealism is presumably opposed to 'realist' theories like Davidson's that explain sentence meaning in terms of unconstrained truth conditions and construe truth value as a non-epistemic feature of sentences for which the principle