Monday, August 01, 2005

Foreknowledge, Memory, Personal Identity

According to a memory-based accounts of personal identity, that in virtue of which I count as the same person as the person named Alex who lived in my apartment a year and a half ago is my memory of his activities. The bond between me and Past-Alex is founded on memory.

[DIGRESSION: This is a strange view, since, of course, I remember other people in addition to Past-Alex and it is not clear why remembering Past-Alex's activities should be thought to ground identity, unless one is sneaking in some identity-presupposing conception of memory. That is a worry anyway. END DIGRESSION]

What I would like to talk about is a different kind of problem: that posed by our knowledge of the future. At its most general, a memory-based theory of ID, claims that identity is explained by reference to knowledge: the reason Past-Alex was me is because I have access -- via memory -- to the facts about, for example, what it was like to be him. To the extent my knowledge is defeated -- by defective memory -- my claim to be identical to Past-Alex is undermined. If I have forgotten a great deal about Past-Alex, then I no longer count as him... or something.

But note that, on certain views, one's access to Future-Self is much more fraught than one's access to Past-Self. It is a mainstream position in the Divine Foreknowledge/Freedom Debate to claim that there are no truths about the future. That there is nothing to know about the future (excepting necessary truths?). If God is cut off from the knowledge of the future, I can hardly do better. So, while I presumably know many things about Past-Alex -- and in virtue of that knowledge count as him -- I cannot know about Future-Alex. If knowledge grounds identity, then no Future-Alex will be me.

But this seems crazy. No doubt, the guy who may well finish typing this sentence will be the same guy who started typing it. But consistency with an abstracted version of memory theory ought to rule that out. No one person ever completes a sentence!

Some options:

1) Show that the abstraction from memory to knowledge is illegitimate. Memory is as deep as identity goes.

2) Explain how there is some principled reason why epistemic access from now-to-future is different from epistemic access from now-to-past. That is, explain why identity should be grounded in a temporally asymmetrical manner. I mean: weirdly, I may become Future-Alex, but only because he will remember me and not because I now know about him. Isn't that a little strange. It is as if: though I am not now identical to that future person, I will become identical to him once he gets around to remembering me. Sigh.

3) Believe that sentences are sometimes finished by people, and ditch the claim that ID is based on memory.

Think about: the funny differences in our attitudes about future and past suffering -- also: prenatal non-existence vs posthumous non-existence.

Think about: temporal asymmetry generally.

Think about: the weirdness of tense... maybe there is a kind of error in reasoning here, which is brought on by the funniness of time and tense.