This collection of essays and lectures consists mainly of pieces which have been published before. What is new is the lecture on G.E. Moore and the second and third lectures on memory originally delivered at Princeton in 1962, « Three Forms of Memory » and « A Definition of Factual Memory ». The other essays appear substantially as they did in philosophical publications. The longest essay is the one on the Verification Argument, which consists of a vigorous criticism of the analysis Carnap and C.I. Lewis make of the notions of « theoretical certainty » and « absolute certainty » as applied to empirical statements. Two chief errors are alleged of Carnap and Lewis. First, a failure to see that on their view of « theoretical certainty », it would require an infinite number of acts of verification to be performed before we could say that any statement was theoretically certain. Since it is a contradiction to assert that an infinite number of actions have been performed, they are guilty of misusing the expression « theoretically certain ». « What they call ' theoretically certain ' cannot be attained even in theory » (p. 56). The second, and more serious, error is to equate the expression « absolutely certain » with « theoretically certain ». This would have as a consequence that the perfectly straightforward statements, whether true, false or problematic, like « It is absolutely certain that Socrates had a wife », are self-contradictory. — Some of the assumptions and implications of the position are discussed in other essays. M.-M.V.