One standard criticism of the memory (or virtually any psychological continuity) account of personal identity is that it is vulnerable to duplication. If person A is somehow duplicated, so that the resulting persons are B and C, and both have A’s memories and are otherwise psychologically continuous with A, this shows that psychological continuity cannot be the bearer of personal identity. After all, in this case, B would be identical with A, and C would be identical with A, so by the transitivity of identity, B would be identical with C. But since B and C are, ex hypothesi, two distinct persons, they cannot both be the same person as A. I have never found this argument convincing or relevant—it seems to me to miss what personal identity is about, because “same person” doesn’t mean “same variable” and personal identity involves temporal considerations that the duplication argument simply ignores. If we want to insist on using personal identity for a formal, atemporal, relation, my sense is that Parfit is right—personal identity isn’t a real property to begin with and we should switch to a different word that will be less confusing to metaphysicians. But I was just reading Lynne Rudder Baker’s summary of Gareth Matthews’ religiously motivated attempt to save the memory criterion and it strikes me as completely off track; let’s hope psychological continuity theorists don’t need to appeal to intuitions this vague!
I haven’t read Mathews’ argument, but only Baker’s short summary of it. Presumably he defends some of his premises, though I can’t imagine how. Here I will just quote Baker’s summary in full (from her “Death and the Afterlife” (p. 380); there is no citation for Mathews, so I’m not sure where he discusses this):
The reason it would be metaphysically impossible for B and C to have A’s memories is this: A deserves punishment. God is essentially just and judges everyone. Suppose that B and C both had A’s memories (caused in the right way). Whom does God punish? If God punished B but not C, or C but not B, then God would not be essentially just: B and C are related to A in exactly the same way; it is impossible to be just and to judge B and C differently. On the other hand, if God punished both B and C, then there would be twice the punishment that A deserved, and again God would not be essentially just. Either way, supposing that B and C both had A’s memories (caused in the right way) violates God’s essential justice in judgment. Because God is essentially just, if A deserves punishment, it is metaphysically impossible for God to bring it about that B and C both have A’s memories.
Now there are all sorts of problems—the reliance on theological premises, for example, or on the idea that despite being a “loving” God, He is preoccupied with making sure everyone gets to suffer for their sins (I’ve always wondered how much of the psychological pull towards such views is about love, and how much—pace Nietzsche—is just about sticking it to the people you don’t like). But here it’s the conception of justice I don’t buy. I get that, if God is all about punishing, it would be unjust of him to punish B but not C. Sure, ok, both of them deserve it equally (though—pace Anselm, this time—we could insist that God doesn’t necessarily punish everyone based exclusively on their desert, since God must also be just to Himself, and we don’t have any idea how that works; I’m pretty sure Anselm’s view of divine justice would undermine this prong of the dilemma).
But I’m willing to accept this. It’s the second prong that smells fishy. “If God punishes both B and C, then there would be twice the punishment that A deserved, and again God would not be essentially just.” Wait. Come again? This seems to me obviously, and trivially, wrong at best. First, there seems to be something very weird going on: the argument assumes that there is a fixed ratio between the amount of guilt and the amount of appropriate punishment. So if, say, someone commits crime X, which deserves Y amount of punishment, then it would be unjust to meet out more (or less, I suppose) than Y amount of punishment. But this can’t be the whole story. For surely if A commits crime X, it would be unjust to meet out Y amount of punishment to Z, a completely different person. So it matters not simply how much punishment is meted out, but to whom it is meted out—the right person has to be punished. And that person is A. So on the argument as given, it seems like there are two criteria in play: (1) Punishment must be meted out to the person who deserves it, and (2) the amount of punishment for crime X in the universe must be proportional to the severity of crime X. Now, (2) may be a modification of a reasonable assumption, (3) the amount of punishment must be proportional to the severity of the crime. But (3) is perfectly compatible with duplication—if both B and C are psychologically continuous with A, then both B and C deserve the amount of punishment proportional to A’s crime.
The difference between (2) and (3) should be clear enough. (3) insists that everyone get what they deserve, but no more. But (2) insist that in the universe as a whole, there not be meted out more than the number of people who initially deserved it now deserve. In other words: the assumption of (2) is that, if only one person committed crime X deserving Y punishment, then at any time after X is committed, only Y and no more may be justly doled out in the universe. But I haven’t got a clue why we should believe that. It makes sense, of course, to say—with (3)—that if A committed crime X, which deserves Y punishment, then it would be unjust to punish A with more than Y. But if, as we are assuming, A is split into B and C, where both remember committing the crime, both remember thinking beforehand about the consequences, etc., I can see no reason why Y punishment would not be appropriate to each person who committed the crime. Why should the universe demand—if one person committed a crime, but now two people stand in that one person’s place—that only one of them may be punished? Whatever the idea behind this, it doesn’t seem to me to be related to any conception of justice.
Perhaps the idea is this: since only one person committed the crime, only one person may be punished. This is still dubious, but it’s also irrelevant: if, at the time of crime X, one person committed the crime, but now there are two people responsible, to insist that only one of them deserves punishment would be question begging. The argument must assume, it seems to me, that two people cannot both be the same as one person that used to be. Perhaps that isn’t question-begging: the argument isn’t supposed to show that if A is duplicated into B and C, then B and C are not identical with A. The argument simply assumes this. It is supposed to show only that God could not allow both B and C to be duplicates of A. But that isn’t right—if B and C were both duplicates of A but were not the same person, then there would be no problem here at all, because neither B nor C would deserve punishment. So in that case, God would have no reason to prevent the duplication. He would have reason to prevent the duplication only if B and C in fact were the same person as A. But then it seems perfectly reasonable to think that both B and C deserve the punishment for A’s crime, since it is also B’s crime and C’s crime. So Mathews’ argument is either question-begging, or involves saddling justice with a weird assumption that is foreign to the idea of justice, since it isn’t germane to the issue of what punishment anyone who committed a crime deserves.