Identity and Extrinsicness

In 'Essence, Origin and Bare Identity' (Mind, I987) Penelope Mackie analyses possible responses to a problematic set of possible worlds, and argues that our reaction to this set must be either: (i) to abandon talk of transworld identity in favour of some counterpart relation; (2) to embrace some non-trivial form of essentialism, such as the necessity of origin; (3) to reject the principle that there can be no 'bare' transworld identities or non-identities; or (4) to reject the principle that transworld identities and non-identities cannot be 'extrinsically determined'. Mackie argues for the third solution to the problem. In this note I want to discuss one of her arguments in connection with the fourth response, a response which I have endorsed elsewhere,1 and to suggest that it may not be as unattractive as she makes out. The problem, in rough outline, is this. Suppose that in the actual world wi, an oak-tree, Oi, developed from acorn Ai at place pI. Suppose, further, that there are two other possible worlds, w2 and w3, such that in w2, OI develops from a different acorn A2 at pI, and, in w3, Oi develops from Ai, as in the actual world, but at a different place, p2. There appears to be no objection to the existence of a fourth world, w4, in which an oak-tree, 02, develops from acorn A2 at pi and another oak-tree, 03, develops from Ai at p2. (See Fig. i below.) Since 02 and 03 are clearly distinct trees, it must be the case that either Oi in w2 # 02 in w4 or OI in w3 # 03 in w4 (or both). If we want to avoid commitment to non-trivial essential properties such as the necessity of origin (which would imply, in effect, that w2 is not a possible world relative to wi) and retain talk of transworld identity then, whichever we choose, we are forced to reject either the No Bare Identities principle or the No Extrinsic Determination principle. Either Oi in w2 # 02 in w4 or OI in w3 # 03 in w4 (or both). Suppose, for example, that Oi in w2 # 02 in w4. This non-identity is either bare (i.e. it does not supervene upon facts of any sort) or else it holds in virtue of the existence of another, equally good, candidate for identity with Oi in w4 (namely, 03). The latter explanation of the non-identity implies '. . . that although the oak tree 01 could have grown from the acorn A2 at pi and have had the material composition, appearance, and so on exemplified in w2, and although there could have been an oak tree that had all those characteristics and also had a companion growing some distance away at p2, none the less Oi could not have been like that and also have had such a companion.'2 Mackie finds this consequence 'extraordinary', and advocates rejection of the No Bare Identities principle.