Against the Repugnant Conclusion

Another wonkish philosophy post

In my previous post on utilitarianism, I started with two crucial observations.

First, utilitarianism is a political philosophy, dealing with the question of how the resources in a community should be distributed. It’s not a system of individual ethics

Second, (this shouldn't be necessary to state, but it is), <em>there is no such thing as utility</em>. It's a theoretical construct which can be used to compare different allocations of resources, not a number in people's heads that can be measured and added up.

Failure to accept these points is at the heart of the kind of ‘longtermism’ advocated by William McAskill and, earlier, Parfit’s Repugnant conclusion. The claim here is that the objective of utilitarians should be to maximise total utility, including people who are brought into existence as a result of our decisions. In particular, that means that it is desirable to bring children into existence who will have a miserable life, provided that no one else is made worse off, and the life is not so bad that the children in question regret being born.

As well as being intuitively unappealing, this idea makes no sense in the two main contexts in which it is relevant: families deciding how many children to have, and polities deciding whether to promote pro-natalist policies[1]

The members of a family, and of a polity, have to allocate resources among themselves. Utilitarianism says that the welfare of each member should be given equ2 weight. In deciding whether to bring an additional child into existence, it’s necessary to compare two situations

(i) the child is born, and has an equal weight with everyone else; or

(ii) the child isn’t born, and all the current members of the group are weighted equally

It’s nonsensical in case (ii) to add in some extra weight to the hypothetical child who doesn’t exist. And it’s clear that if everyone in case (ii) is better off than everyone in case (i), the correct utilitarian decision is to go with (ii).

This leads to the conclusion that the social order we want is one where average utility is maximized (remembering that utility is a way of comparing allocations, not a real thing).

Another way to reach this conclusion is from behind a Harsanyi/Rawls veil of ignorance where we choose a social order of which we will be a member, without knowing where we will be situated. There’s no way to make this work if we are also supposed to consider the infinite set of possible people who won’t come into existence at all.

The counterarguments I’ve seen don’t impress me. Many of them start with some version of the utility monster, an individual who can have massively more utility than anyone else. But, as I showed in my last post, utilitarianism as a political philosophy doesn’t work that way. The utility of a trillionaire is outweighed by small improvements for a hundred people.

Parfit’s arguments, as quoted in the Stanford Encyclopedia of Philosophy rest on appeals to intuition derived from situations that can’t possibly exist in an actual polity.

“For instance, the principle implies that for any population consisting of very good lives there is a better population consisting of just one person leading a life at a slightly higher level of well-being”

This is both untrue and analytically faulty.

Untrue because humans are social animals, and human societies require a minimum number of people to deliver anything beyond bare existence. Solitary individuals don’t live well, so the supposed “better population” can’t exist. But if it were possible, we wouldn’t need polities or utilitarianism, any more than bears or skunks do

It’s analytically faulty because the point isn’t to compare different populations in the abstract but for families and polities to make choices about population. Starting from an existing population, it’s entirely possible (and is now the case) that people might choose below-replacement fertility so that they and their children can have even better lives.

So, we could easily see the population of the world gradually decline from billions to hundreds of millions, with steadily rising living standards. But below some point (Charlie Stross estimates 100 million) it would become impossible to sustain a modern civilisation. So, at this point (many generations away) it might be necessary to encourage people to have more kids.

Until then, the choice can be put as one between

(i) Letting families make their own choices, leading to a world with a shrinking population living better lives; or

(ii) Adopting pro-natalist policies [2} to deliver a growing population, living worse lives

Parfit called (ii) the repugnant conclusion, and he was right to do so.

[1] Migration raises a whole new set of issues about who counts. My position is essentially cosmopolitan (everyone counts, wherever they live), but this needs a whole new post, or maybe a book.

[2] To get fertility rates above replacement under current conditions, such policies would have to be very intrusive.