I do not think that Professor Oppenheim and I are any nearer agreement, but perhaps the various forks in the road are clearer. I will try to be as fair, as clear, and to the point as Oppenheim has been. I will consider most, but not all, of his replies and will do so in the order he has followed.

Oppenheim concedes that he is relying upon the principle that a sentence is cognitively significant if and only if (briefly) it is logically significant or empirically testable, and he claims that this principle is generally accepted by contemporary philosophers of science. I do not think it is generally accepted; indeed, on the page after the one quoted by Oppenheim, Carl Hempel states that he feels “less confident” that such a criterion can establish “sharp dividing lines” between “those sentences which do have cognitive significance and those which do not.” The proper estimate of the current situation is, I think, that whether a general criterion of “cognitive” meaning can be had, and if so, just what it is—are notoriously unsettled questions. If any estimate is correct, it is that among philosophers of language there is widespread suspicion of the neat distinctions between analytic/synthetic, a priori/a posteriori, cognitively meaningful/meaningless which prevailed prior to and during the 1950's. The work of W. V. O. Quine and Noam Chomsky has only muddied the waters further.