Anti-realism, truth-conditions and verificationism

The article begins by distinguishing a number of theses which, in the past, have sometimes been lumped together under the heading of 'anti-realism'. One of the theses is that there is something wrong with truth-conditional theories of meaning (what a truth-conditional theory of meaning is a matter discussed at some length), another is what I take to be the central thesis of anti-realism, that all truths are knowable. Several writers on the subject, such as Wright and Prawitz, have defended the latter thesis while jettisoning the former. I argue that this position is exactly the wrong way around. Given the 'meaning is use' principle, which is also called the 'manifestation requirement', a very powerful case can be made that true theory of meaning cannot be truth-conditional. But I argue that, given the current state of our logical knowledge, there is no good reason for concluding from this that a true theory of meaning must be of the 'verificationist' type, as Dummett seems to think, and still less for thinking that anti-realism follows. I end by examining theories of meaning against Dummett's criticisms. © Oxford University Press 1997.