I've previously argued that consequentialist moral theories respect the "separateness of persons" when they recognize individual persons as being of distinct intrinsic value, rather than seeing them as mere means to the single token value of aggregate welfare. (This entails more fine-grained non-instrumental desires, and associated emotions like regret, but doesn't ultimately affect what actions are the right ones to perform.) So I was interested to come across a different conception of the separateness of persons in Michael Otsuka's 'Prioritarianism and the Separateness of Persons'. According to Otsuka, a theory respects the separateness of persons when it is sensitive to "competing claims" and so treats "non-identity" cases differently:
It is morally relevant that there are distinct persons with competing claims to receive benefits. Such competing claims ground moral complaints on the part of those who would be worse off, relative to others, and the case for giving benefits to people with such complaints is stronger than it otherwise would be in analogous [intra-personal or non-identity] cases in which the prioritarian value of distributing goods in one way rather than another is equally great, yet such complaints are lacking. (371-2)
To illustrate, suppose we must choose between offering a larger benefit to a mildly impaired person or a lesser benefit to a more severely impaired person. Otsuka claims that it makes a moral difference whether the lesser benefit would be identity-affecting. If it would cause a different, slightly-less-severely impaired person to come into existence, distinct from the severely-impaired person who would otherwise come into existence, then there is no "competing claims" justification for choosing this lesser benefit. If you instead help the mildly impaired person (yielding a greater benefit), the severely impaired person has no personal grounds for complaint against you, since there is nothing you could have done to help them more -- treating the more severe impairment would (ex hypothesi) have instead brought someone else into existence. So, Otsuka claims, it is easier to justify helping the mildly impaired person in this identity-affecting case than it would be in the simpler case where such a choice would amount to failing to help the worse-off person.
Suppose we fill in the details so that this difference in justification actually tips the scales: a morally-motivated stranger ought (according to the "competing claims" view) to choose the greater benefit in the identity-affecting case, but the lesser benefit in the basic (constant identities) case.
I'm suspicious of the idea that mere identity facts can, in this way, affect what we (as morally-motivated strangers, bracketing any partiality towards our "nearest and dearest") ought to do. I'd initially thought that such suspicions might be buttressed by appeal to the "Veil of Ignorance", which forces fair and impartial moral choices by hiding all identity facts and forcing us to choose between outcomes on purely qualitative grounds. But now that I think of it, I guess there's not any reason why the Veil couldn't allow through facts about sameness or difference of identity, just so long as it blocks the chooser behind the veil from knowing which identity is hers. So, never mind that then.
Let me instead raise a different challenge: Why think that the only way for a moral theory to be sensitive to "competing claims" is via its verdicts for right action? Couldn't a theory instead respect the separateness of persons by seeing these cases as calling for different emotional responses (as I previously argued)? If we reject the idea that one's baseline welfare makes any difference to the strength of one's claim, then this seems like the way to go. We should (according to the separateness-respecting utilitarian) always bestow the greater benefit to one rather than a lesser benefit to another. But sometimes this decision will call for more "inner turmoil" than others. In particular, we can continue to see it as regrettable for Bob's sake if Bob could have benefited from an alternative decision (even though the maximizing decision was the right one), whereas in an identity-affecting tradeoff there will be no such grounds for pro tanto regret -- the decision will be, in this sense, a psychologically "easier" one, though the various reasons for action present in the case are unaffected.
Using my previous account as a model, then, it seems that a separateness-respecting prioritarian could say something similar. They will hold that we have more reason to benefit people the worse off they are (regardless of whether or not the benefit would be "identity-affecting"). We have no additional reasons for action given by "competing claims" in the constant-identity cases. But we may have additional reasons for regret, whenever a particular person is less well-off than they otherwise could have been. So we can, in this way, respect the separateness of persons -- even where this is understood to entail "sensitivity to competing claims" -- without requiring this sensitivity to take the particular form of affecting our actions.