Non-identity problem

Why is it so hard to name anyone wronged by a choice that decided which person would exist at all?

A country chooses to burn through its resources rather than conserve them. Three centuries later people are living in a depleted, harsher world. Most of us want to say those people were wronged. So ask the obvious follow-up: wronged compared with what?

The tempting answer is "compared with the lives they would have had if we had conserved." But hold that up to the light. Under the conserving policy people take different jobs, move to different towns, meet different partners, conceive on different evenings. Within a few generations not one of the same individuals is born. The people in that other future are not better-off versions of the depletion generation. They are other people entirely. So who exactly is worse off?

Derek Parfit set this out in Reasons and Persons in 1984, and the sharpest version is smaller than a policy. A fifteen-year-old decides to have a child now. Everyone agrees she should wait: a child raised by a teenager who is not ready will have a harder start. She has the child anyway, and he has a life that is difficult but clearly worth living.

Now try to say what she did to him. Had she waited, he would not have had a better childhood — he would not have existed, because a different conception at a different time is a different person. His options were this life or none. Unless we are prepared to say his existence is worse for him than never being, she did not make him worse off. And that runs straight into a principle almost everyone holds without noticing: that an act wrongs someone only if it makes that someone worse off than they would otherwise have been.

Notice the shape of the trap: the very act we want to condemn is the act that fixes who the person is. Change the act and you do not improve the victim, you replace him. That is why it is called non-identity. It is not a puzzle about uncertainty or distance in time; it survives perfect foresight.

Two responses appear immediately, and both cost something.

The first is to stop asking who benefits and ask only how much good there is. On such an impersonal view the conserving future contains more wellbeing, so depletion is worse and no victim is needed. That works — and it opens Parfit's other door. If what matters is the total, a vast population of lives barely worth living can beat a small flourishing one by sheer numbers. Parfit called that the Repugnant Conclusion, and the theory he wanted, which would condemn depletion without licensing it, he called Theory X. He never claimed to have found it.

The second response keeps persons in the picture but widens the comparison: instead of asking whether this individual was made worse off, ask whether a better-off person could have occupied that place. This family of wide person-affecting views condemns the depletion policy without impersonal machinery. The cost is that it treats people as occupants of slots — precisely the thought person-affecting morality was meant to rule out.

A third line drops comparison from the idea of harm altogether: to harm is to cause someone to fall below a decent threshold, whether or not an alternative existed. That handles a child conceived into serious deprivation, but goes quiet where the depletion case lives, since a diminished future can still sit above any plausible threshold.

There is no settled answer here. Formal work in population ethics, notably Gustaf Arrhenius's impossibility results, suggests no theory can satisfy all the adequacy conditions we independently find plausible — which would make the price structural rather than a failure of cleverness.

Three refinements hold the pieces together.

First, this is not the imperceptible-harms puzzle in different clothes. There, real people are made real worse off and the difficulty is that each contribution is too small to detect. Here detection is not the issue: the contribution is enormous and perfectly visible, and there is still no worse-off individual. That is the fixed point of difference — one puzzle loses the victim in the noise, the other has no victim to lose.

Second, "his life is worth living" is doing exact work, not rhetorical work. If the life were genuinely worse than non-existence, the ordinary comparative account applies and the problem dissolves. It lives only in the band where a life is worth living but avoidably worse than it might have been for whoever occupied that position.

Third, this is a conceptual disagreement and should be labelled as such. No experiment adjudicates between an impersonal principle and a wide person-affecting one; progress would look like a theory that meets the adequacy conditions, a proof that none can, or a principled argument for giving one condition up. One thing here is empirical, though, and worth stating as a test: whether a given choice is actually identity-affecting. If a policy demonstrably leaves the same individuals conceived and merely alters their circumstances, the structure does not apply and ordinary comparative harm returns. That check disqualifies most everyday uses of the argument.

• The Repugnant Conclusion, and why "just cap the population's size" is harder to defend than it sounds.
• Whether wronging without harming can be built on contractualist grounds.
• How the same structure appears in reproductive medicine, where selecting one embryo over another is identity-determining by construction.

Every term the collection defines is gathered in the glossary.