3.1 If we accept the Vorlons’ offer to create a stroke-free duplicate, that does not mean that we are claiming that the duplicate is the same (numerically identical) person. We could instead say that some things are more important than numerical identity, and that this is such a case. I think that’s what we should say.
3.2 The reversal test for status quo bias is no good here, because which status is quo is relevant. The best way I can think of to explain this is an analogy. Take gold: a valuable substance which, we all know from our science classes, is atomic number 79. That is, we thought we knew. But the Vorlons appear in a hologram and explain that, as a result of their ancient war, the atomic numbers have been fluctuating and all that beautiful yellow jewelry is actually now atomic number 75 - we just failed to keep up with events. Atomic number 79 is now that stuff we call “bismuth”. They explain a whole new physics of vorlonicles (as we name the new particles) showing that while the beautiful yellow jewelry was undergoing a change of atomic number, it still maintained characteristic invariant vorlonicle-structures, which allowed it to stay yellow, beautiful, heavy, and so on.
Does this mean you should go back to the jeweler’s and yell at them for selling you fakes? Should you return your rings and demand a bottle of Pepto-Bismol instead? No, it means that the essential feature of gold never was its atomic number, despite appearances that it was so. Your jewelry is still golden, and that bottle of Pepto still contains bismuth.
Similarly, in Mark Walker’s scenario, nobody was born yesterday, their children are still theirs, and somatic identity never had anything to do with personal identity, because the status was quo. The status quo in one’s world, or more precisely the normal workings of things, determines the reference of one’s terms. In the Vorlon-world “gold” refers to a certain invariant of vorlonicles—but in the real world, “gold” still refers to atomic number 79.
3.3 The Hatfield-McCoy argument fails. To see why, we just need to look at the obverse side of the coin. Thus: Hatfield hates McCoy so badly that he is willing to undergo radical changes in his personality and memory in exchange for killing McCoy. So he does. The surviving human being claims—with perfect sincerity! - not to remember hating McCoy, and to be a nonviolent person. If we accept psychological continuity as the public norm, Hatfield gets away—from his viewpoint, at least—with murder. The practical-ethical argument is a wash.
6.0. If P2‴ were presented without argument:
Preservation of personal identity requires preservation of numerical [token] identity.
then it would indeed beg the question. But I can think of two good arguments for it right off the bat. First, types are abstract entities, and while we may be quite happy with abstract entities in logic, math and semantics, to posit that you are an abstract entity is quite a surprise. Second, people, like animals and insentient objects, are routinely identified and re-identified, and individuated, by bodily characteristics. I’ve never heard of two human bodies being considered one person based on psychological similarity. That just isn’t how we normally think about people.
Which is not to deny that type-identity is important. But it can be important without its being equivalent to personal identity. See my comments on 3.1 above.
Some comments, organized by section.
Nice analysis, gwern.
3.1 If we accept the Vorlons’ offer to create a stroke-free duplicate, that does not mean that we are claiming that the duplicate is the same (numerically identical) person. We could instead say that some things are more important than numerical identity, and that this is such a case. I think that’s what we should say.
3.2 The reversal test for status quo bias is no good here, because which status is quo is relevant. The best way I can think of to explain this is an analogy. Take gold: a valuable substance which, we all know from our science classes, is atomic number 79. That is, we thought we knew. But the Vorlons appear in a hologram and explain that, as a result of their ancient war, the atomic numbers have been fluctuating and all that beautiful yellow jewelry is actually now atomic number 75 - we just failed to keep up with events. Atomic number 79 is now that stuff we call “bismuth”. They explain a whole new physics of vorlonicles (as we name the new particles) showing that while the beautiful yellow jewelry was undergoing a change of atomic number, it still maintained characteristic invariant vorlonicle-structures, which allowed it to stay yellow, beautiful, heavy, and so on.
Does this mean you should go back to the jeweler’s and yell at them for selling you fakes? Should you return your rings and demand a bottle of Pepto-Bismol instead? No, it means that the essential feature of gold never was its atomic number, despite appearances that it was so. Your jewelry is still golden, and that bottle of Pepto still contains bismuth.
Similarly, in Mark Walker’s scenario, nobody was born yesterday, their children are still theirs, and somatic identity never had anything to do with personal identity, because the status was quo. The status quo in one’s world, or more precisely the normal workings of things, determines the reference of one’s terms. In the Vorlon-world “gold” refers to a certain invariant of vorlonicles—but in the real world, “gold” still refers to atomic number 79.
3.3 The Hatfield-McCoy argument fails. To see why, we just need to look at the obverse side of the coin. Thus: Hatfield hates McCoy so badly that he is willing to undergo radical changes in his personality and memory in exchange for killing McCoy. So he does. The surviving human being claims—with perfect sincerity! - not to remember hating McCoy, and to be a nonviolent person. If we accept psychological continuity as the public norm, Hatfield gets away—from his viewpoint, at least—with murder. The practical-ethical argument is a wash.
6.0. If P2‴ were presented without argument:
then it would indeed beg the question. But I can think of two good arguments for it right off the bat. First, types are abstract entities, and while we may be quite happy with abstract entities in logic, math and semantics, to posit that you are an abstract entity is quite a surprise. Second, people, like animals and insentient objects, are routinely identified and re-identified, and individuated, by bodily characteristics. I’ve never heard of two human bodies being considered one person based on psychological similarity. That just isn’t how we normally think about people.
Which is not to deny that type-identity is important. But it can be important without its being equivalent to personal identity. See my comments on 3.1 above.