Sortals and Trans-world Identity
Allan Gibbard claims that while same-world identity need not be relative to a sortal, trans-world identities do need to be like that, that is, two things in different worlds cannot be the same simply but only the same “F.” “We speak and think of ‘the same person’ but not of ‘the same entity’.” (Metaphysics: An Anthology, 106) Are these claims true? Suppose I tell you that Qwerty is identical to Poiuyt, though Qwerty is a Ruritarian term, and Poiuyt is a Waldavian term. You say: “Interesting, I’ll keep that in mind; but to what single entity do both Qwerty and Poiuyt refer?” And I reply, “A computer keyboard.” The talk of entities leaves us wanting more information, but it is hardly meaningless.
Now consider the following questions: Could Richard Nixon have been a different
(1) entity
(2) man
(3) president
(4) Nixon
than what he in fact was?
Answers:
(1) Since every property is accidental to Nixon qua entity, yes. There is no reason why he could not be, say, a bottle of wine.
(2) Yes, if, e.g., he had been, contrary to fact, defeated by Humphrey in the 1968 presidential election.
(3) Yes, if he had not been implicated in Watergate, or if he had disliked poker, etc.
(4) No, because every property of Nixon is now essential to him.
So, sortals do change the outcome in questions of trans-world identity.