As regular readers may be aware, I sometimes go to the pub to discuss short-ish works in political theory. Last week, we were discussing Derek Parfit's work on intergenerational justice, in particular justice towards future generations: what we owe to those who are not yet born. Unreflectively, most of us think we have definite duties towards future generations. It is nigh-on axiomatic for example that we should not deplete the resources of the planet, for example, because we have a responsibility to ensure that our children and our children's children are not left in a situation, because of our profiligacy, of want. Parfit, however, offers something called the Non-Identity Paradox which appears to cast doubt on the conclusion that we have anything more than a fairly limited duty of beneficence towards future generations. I think Parfit's paradox involves a confusion of two types of responsibility, and the fact that it is paradoxical rests on that confusion.
What the Non-Identity paradox does is make future generations who complain about their rights-violations into future generations who complain about their own existence. Imagine there is a decision taken, and that as a result of that decision, x will be born, and live a life that is fairly miserable, but still preferable over not existing at all - a life worth living in Parfit's terms. Imagine at least of the events that makes x's life miserable is an event which, were it to be done to someone currently living, we would describe as a rights-violation. Imagine further that all this is a direct causal result of the initial decision, so that by taking that decision, we are bringing x into existence, and then doing something that would conventionally be described as a rights-violation to them. Parfit's point is that x is doing something very strange if they complain about the decision which both brought them into existence and, in conventional terms, violated one of rights, because they are saying, effectively, that they would prefer that they didn't exist, even though ex hypothesi, they do prefer existence over non-existence. Because they can't complain about the decision to bring them into existence, they can't complain about the rights-violation either, as that flows directly from the decision to bring them into existence: they are effectively given a choice between having some of their rights violated, but not being able to do anything about it and not being able to claim any compensation, and not existing at all.
Consider the application of this thought-experiment to the real world. It would suggest that Union Carbide have no duty to compensate children born with disabilities in Bhopal after the disaster there in 1984, since, plausibly, that disaster affected their parent's lives in such a way that had the disaster not occurred, different children would have been born. That, to my mind, shows immediately that Parfit must be wrong, whilst also suggesting why. Intuitively, in the Union Carbide case, we think that the negligence of Union Carbide makes them responsible for any and only harms which could be reasonably expected to flow from that negligence. Just because, for example, we could trace a causal chain from the leak at Bhopal to freak hailstorms ten years later over Kansas which destroyed a farmer's crops, or people who became more seriously injured as a result of persistent refusal to accept assistance on reasonable terms, it doesn't mean we would hold Union Carbide responsible for those harms. Causal responsibility - having caused some event - does not imply moral responsibility - being liable for its costs. Likewise, just as if I sell a knife to someone whom I have good reason to suspect will use it to cause someone else harm, even if they would have caused them similar or even identical harm without me selling them that knife, I am partially morally responsible for that harm, even though I am not causally responsible. Moral responsibility does not imply causal responsibility.
So, moral and causal responsibility come apart: neither is a sufficient condition of the other. The Non-Identity paradox, however, assumes that they are identical. In the paradox, the causal responsibility for having brought x into being becomes a moral responsbility for any benefits that x gains as a result of their existence, which then, providing x's existence is worth living, absolves those who brought x into being of the violations of x's rights, because x prefers, of the two options available, living. The paradox, I think, isn't really about identity at all - the fact that, absent the policy, x wouldn't exist at all is totally irrelevant, because if x does exist, as they will once the policy has been decided on, then x can have their rights violated in exactly the same way as any other person - it's about responsibility, and it shows, amongst other things, that even very clever philosophers can make mistakes.
(Note: much of this owes a heavy debt of inspiration to Steve Winter)