Keith Lehrer and Joseph Richard have cooked up four delectable examples of remembering without knowing. 1 Although the subject (S) in each example remembers that p, S either does not presently know that p (as in their first and second examples), or did not previously know that p (as in their third example), or does not and never did know that p (as in their fourth example). The examples do, in my opinion, refute the present and previous knowledge requirements in analyses of factual memory such as:
(M) S remembers that p if and only if S knows that p because S knew that p.2
In a subsequent article, Saul Traiger discusses each of the examples with a view to defending M. 3 Although I take up one part of his discussion in Section 4, where I show why Lehrer's and Richard's fourth example should not be written off as an instance of "impure" memory, determining how successful Traiger is in his treatment of the examples is not a major aim of this paper. For, quite apart from the points that Lehrer and Richard press, it seems to me that the two concepts, knowing that p from memory and remembering that p, are distinct. If, for example, I believe that p somewhat hesitantly and yet believe it from memory, I may well remember that p -and yet, for lack of confidence, I may far to know that p from memory. More importantly -whatever my degree of belief that p -there presently may be, for that belief, too little justification of the right sort for me to know that p from memory, but quite enough such justification for me to remember that p. It therefore is not with the critical part of Lehrer's and Richard's article that I shall be concerned in what follows. Instead, I want to look at the constructive proposal which emerges from their diagnosis of the examples, their proposed nonepistemic analysis of what it is to remember that p: