Memory sometimes yields knowledge and sometimes does not. It is, however, natural to suppose that i f a man remembers that p, then he knows that p and formerly knew that p. Remembering something is plausibly construed as a f o rm of knowing something which one has not forgotten and which one knew previously. We argue, to the contrary, that this thesis is false. We present four counterexamples to the thesis that support a different analysis of remembering. We propose (...) that a person remembers that p (at t) if and only if the thought or conviction that p comes from memory (at t) when, in fact, it is true that p. (shrink)