Abstract

That there is an edge at all is, of course, philosophically controversial; it would be denied by anti-realists of a verificationist stripe. However, we accept, since G¨odel, that there are true propositions of elementary arithmetic that are unprovable in arithmetic; just so, we should accept—by analogy—that there are true statements that are unknowable. An argument called the Fitch Argument tells us that it is so. Williamson has long argued that the Fitch Argument cannot by itself refute antirealism—because the anti-realist is already committed to the denial of some of the principles of classical logic required to derive the anti-realist conclusion. The point is well made.1 In Knowledge and its Limits, however, Williamson looks at what the Fitch argument tells us if we adhere to classical logic: and that is that there are unknowable truths