David Jensen’s paper in the latest JME considers a possible Kantian argument against the use of genetic enhancement for the sake of comparative advantage in one’s children.  Essentially, the argument rests on the idea that the maxim describing such a course of action would not be universalisable; universalised, it would be self-defeating, since the very idea of comparative advantage relies on exceptionalism.  Thus

1. If a consideration w is a reason for person x to do y in circumstances z, then w is a reason for everyone in circumstances z to do y (universal character of reasons).

2. If a consideration w is a reason for person x to do y in circumstances z, then x can do y in circumstances z where consideration w obtains (ought implies can).

3. If a consideration w is a reason for person x to do y in circumstances z, then everyone can do y in circumstances z where consideration w obtains (from 1 and 2).

4. If comparative advantage w is a reason for person x to do y in circumstances z, then some persons cannot do y in circumstances z where comparative advantage obtains (definition of comparative advantage).

5. If comparative advantage w is a reason for person x to do y in circumstances z, then everyone can do y in circumstances z where consideration w obtains (instance of 3). Therefore,

6. It is not the case that comparative advantage is a reason for person x to do y in circumstances z (from 4 and 5).

There’s something appealing about the argument, although I don’t think it quite works.  But one of the reasons why it doesn’t work is interesting in its own right.

My quibbles start with premise 2. I accept that ought implies can, but I have a suspicion that Jensen is mixing up “ought to…” with “has a reason to…”.  Though to have an obligation is to have a kind of reason, it doesn’t follow that to have a reason is to have an obligation; the two things aren’t identical.  (And, thinking about it, even the supposition that an obligation is a reason might be disputable: it doesn’t seem to be analytically the case.)  Granted that, it’s not obvious that having a reason to act is dependent on being able to.  And so the role of the ought/ can rule seems to have been knocked out (and the gap from 2. to 3. has, at best, become a lot wider); an obligation might evaporate in the face of possibility, but it isn’t obvious that reasons do at the same time – which is why it is possible (and makes sense) to talk about one’s regrets when one has not been able to do what one would have wanted to do.  “I regret having missed your wedding: the accident made attendance impossible” is perfectly coherent, and recognises that the speaker had a reason to attend, even though he could not.  What we’re left with is a question of whether it is possible to universalise w in its own terms.

Can we do that in this case?  I don’t see why not.  While it is, of course, impossible for everyone to give their child the same comparative advantage, it seems to me all the same to be possible for everyone to will such a thing, and to take steps to realise it: there’s no internal contradiction here, either in the pure practical reason or the desire.  Jensen himself seems to touch on this point when he writes about basketball:

There may be a variety of good reasons for undertaking an athletic competition: one desires health benefits of exercise by playing basketball.  However, in the case of basketball, the goal of each team is to score the most points, and yet both teams cannot score the most points.  But we normally think that we have a reason to score the most points, since that is the object of the game.  Thus, though competitive basketball conflicts with the universal character of reasons, in the larger context of the game and its purposes, we can set aside this conflict as meaningful.

If we can put aside a conflict with the universal character of reasons in this case, it’s not at all clear to me why we can’t put it aside in all cases, too – and, if we can’t in other cases, moral rules being universal, it would seem that we can’t here.  But, on closer examination, Jensen has admitted why we don’t have to anyway (and why the last two sentences of that quotation are mistaken): though it is impossible for both teams to have more points than the other at the end of the contest, it is not impossible for either to set out to have the more points.  Both teams has a reason to (try to) score as many points as possible; they might even have an obligation to do so.  Both teams also have a reason to score more points than their opponent – though the obligation has clearly petered out in this case.  And that’s precisely where the competition arises.  There is, though, no contradiction.

And the same applies to parents who want to give their children a comparative advantage.  Allowing (arguendo) that there is no other moral reason not to adopt such a maxim, the impossibility universalising a maxim of of ensuring that one’s child is the best does not imply impossibility in the universalisation of a maxim of trying to ensure that one’s child is the best.

What’s the lesson here?  Well, it points to an old problem in Kantian ethics about how you know what your maxim is, and defining it aright.  If you’re not careful, you can outlaw all kinds of things – you can show that naming your child Isabella Rose is wrong, for example.  After all, implicit in the practice of naming someone is the idea of giving them an identifier – a name is not really a name if it’s the same for everyone.  But “One should act so as to name one’s child Isabella Rose for the sake of identification” cannot, on this basis, be universalised – because if everyone named their child that at all times, it would cease to identify anyone at all, and so the maxim would be self-defeating.  And since a self-defeating maxim is one that’s incompatible with the moral law, giving your child that name is also incompatible with the moral law.  And since to say that something is incompatible with the moral law is, to all intents and purposes, to say that it’s wrong… well, that looks to mean that naming your child Isabella Rose is wrong.

It’ll come as a relief to my brother and his wife (who gave that name to their sprog) to learn that this is balls, though.  And the reason why it’s balls is that the maxim is badly expressed.  “One should act so as to name one’s child for the sake of identification” can be universalised; and once you’ve accepted that, the question of which identifier you choose is a whole different game.  Jensen has, it seems to me, chosen a maxim that is too specific – and so defined it into impermissibility.  (By contrast, “One should act so as to give one’s child the best possible start in life” seems universalisable.  It can then have “… including genetic modification if that seems right to you” appended without difficulty.)  Moreover, the impossibility universalising a maxim of of ensuring that one’s child is the best doesn’t generate an ought-not – only a not-ought.  But that’s a different matter.

Read Jensen’s paper.  It’s quite fun – and there’s a lot to it that’s much stronger than the impression here gives.  (The stuff that seems right simply isn’t as entertaining to blog about.)  And even where it does go a bit squiffy, I think it does so for interesting reasons.