J ens J ohansson circularity in saying that personal identity over time should be understood in terms of (continuity of) quasi-memory As suggested by the above, for a long time memory-involving relations outshined other psychological relations on the personal identity scene This focus can perhaps be partly explained by the fact that the debate long involved a somewhat confusing mixture of metaphysical and epistemological considerations: participants often did not distinguish clearly between the question of what personal identity over time holds in virtue of and the question of what constitutes evidence for personal identity over time (This conflation is facilitated by the frequent use of the term “criterion of personal identity,” which has both a metaphysical and epistemological reading.) Memory, or quasi-memory, seems particularly relevant to the latter, epistemological question: having what appears to be a quasi-memory of an earlier experience normally provides me with an excellent reason to believe that I myself did have that experience For quasi-memory gives us a special kind of access to – a sort of “inside information” of – our own past experiences that we lack to the past experiences of others (that is, we so in fact, though not as a matter of conceptual necessity) Another partial explanation of the focus on memory relations might be that an object’s persistence through time is, plausibly, to a large extent a matter of its states at any one time being caused by its states at earlier times The appeal to memory is obviously congenial to this idea, for as just noted, nothing is a memory (or even a quasi-memory) of an experience unless it is caused by that experience Many other ordinary mental relations, such as “x believes the same things as y” and “x has the same character traits as y,” lack such a causal requirement Of course, however, we can still appeal to nearby notions that involve such a requirement: for instance, “x’s beliefs are direct effects of y’s beliefs” and “x’s character traits are direct effects of y’s character traits.” And that is precisely what participants in the debate have gradually come to do: they have gradually come to regard personal identity over time as depending on chains of these and countless other mental connections, with quasi-memory as just one relevant psychological connection among others.2 In other words, what is necessary and sufficient for personal identity over time is the holding of the more general relation of psychological continuity In addition to these fairly conservative modifications of the simple Lockean view – replacing direct memory links with indirect memory links, replacing memory with quasi-memory, and supplementing quasi-memory with other mental relations – a more drastic revision is apparently called for, in light of the so-called “fission” problem (sections 4–7) But first it will be useful to consider why we should be drawn to the psychological-continuity view in the first place 3. Reasons to accept the psychological-continuity view Historically, one important motivation behind the psychological-continuity view has been its strong empiricist credentials What makes it the case that a certain present experience belongs to the same person as a certain other present experience? 128