Distinguish two ways to evaluate a life: welfare value (how good it is for the person living it) and contributory value (what value it adds to the world at large). Might these come apart? Not according to 'crude aggregation', i.e. the view that the value of the world is simply the sum of the welfare values of the lives in it. But we should reject crude aggregation (for one thing, it implies Parfit's repugnant conclusion). I take it this commits us to the view that welfare and contributory value can come apart. But when?
Suppose average welfare matters. Then it could make the world worse to add a new life that has low positive welfare, if this drags down the average. At the very least, the contributory value would seem less than the positive welfare value -- if not negative, then perhaps simply zero; we don't think the extra mediocre life makes the world a better place than it otherwise would have been, even though it is better for the individual to exist than not. Still, there remains some connection: holding constant the fact of their existence, making the new individual significantly better or worse off would also (respectively) improve or worsen the world.
A more radical divergence in these values might be achieved if desert matters. Suppose it makes the world worse if good things happen to bad people (who instead deserve to suffer). Then the happiness of an unrepentant murderer has negative contributory value - it makes the world worse - even though happiness is presumably good for the murderer himself, i.e. it adds to the welfare value of his life.
Can you think of any other examples? I take it the common thread here is a kind of holism about worldly value, such that the contributory value of an individual life cannot be assessed in isolation from how it affects the overall pattern or 'shape' of the world as a whole. Whether it's good to add additional lives (of certain welfare values) to the world depends on how many other people exist, and how well-off they are. As I once put it: 'Adding one to a billion does not make the world a more beautiful shape. But a world with a person is surely more beautiful than the void.'
If we start thinking about the value of the world from this more 'aesthetic' perspective, even more possibilities arise. We may find value in symbolic historical events or 'narrative arcs', independently of how they affect individual people. I recently read a news story about a 109 year old daughter of a slave voting for Obama. There's something kind of cool about that -- like a spark of light that pushes back against the shadows of history -- something that makes the world a better place. And this might still hold true even if the woman personally found the experience merely a bother, and not something that improved her life (from a self-interested perspective) at all. (No doubt communitarians could come up with more along these lines.)
Which sorts of cases do you find most plausible?