It has been common to regard Wittgenstein as being, in his later years, an ordinary language philosopher in some fairly obvious sense of that phrase. The reason for this is that in his later writings he frequently admonished us to consider how words are actually used. And he declared that the philosopher's task “is to bring words back from their metaphysical to their everyday use” (PI, §116). All of us, I suppose, have noticed that Wittgenstein persisted in the metaphysical use of various words. But these failures have commonly been regarded as nothing more than isolated errors he would have been happy to correct. No one, I think, saw in these a pattern indicative of a pervasive metaphysical theory. And yet this, I will argue, is precisely what these seeming lapses actually were. Wittgenstein, if I am right, never rejected the empiricist metaphysics that forms the basis of the Tractatus. By 1916 he had embraced that version of empiricism that William James called “radical empiricism” and Bertrand Russell later called “neutral monism.” From that date until his death his fundamental views changed very little. In his later writings he did revise the Tractatus account of language, but beyond that he merely tinkered with empiricism, adjusting both it and ordinary language until he could bring them to a conformity that suited him.