Conclusively Repugnant

I am not actually a utilitarian.  (I hew to a multi-value consequentialist theory that cares about hedons, preferences, and also a whole bunch of other stuff.)  But I spend a lot of my time trying to model utilitarians, because there are a lot of them around, and I would like to be able to engage them in productive moral discussions without constantly falling into the pit of “But that’s bad!” “No, it’s good!

I had thought that I was pretty good at this.  Recent evidence, however, suggests that maybe my models were all super flawed.  It’s possible that I just don’t understand utilitarians at all.

So let me just ask flat-out: does anyone actually take Derek Parfit’s Repugnant Conclusion seriously? 


That is to say —

As part of the formal philosophical discourse, Parfit’s work is extremely valuable.  I get that.  The Repugnant Conclusion is a way of illustrating the obvious flaws with one easily-accessible, intuitively-compelling, unsophisticated way of making utility calculations.  And that’s a very good thing to have!  It shows you that you have to build a better calculator.

But I’ve encountered a number of people who accept the Repugnant Conclusion (or something very much like it) as an act of bullet-biting, on the theory that all its constitutent steps are unassailable according to their moral logic.  And even the people who firmly reject the Conclusion often do so in a way that is way more respectful to its premises than I would have expected.  This is totally baffling to me.  I’d been under the impression that it was obvious and well-understood why no real utilitarian would have to accept the Conclusion; I just thought that, for various reasons, it was something people didn’t like to talk about explicitly.


There are two separate paths that get you to the place you need to go, at least within my own model of how this works.

The first involves ditching one of Parfit’s implicit premises: that it is always possible to add a positive-utility life into your system, or at least that you can count on being able to do so for enough iterations of the expand-and-redistribute game that you wind up with a huge low-but-positive-utility populace.  If your marginal added life is a negative-utility life, or even if it’s substantially likely to be a negative-utility life, then you can’t assume A+ > A and the whole thing comes crashing down.

And, like…don’t we secretly know that there are a lot of negative-utility lives out there?  Like, maybe a majority of all the human lives currently in existence?  That, at this point in our techno-cultural development, the marginal life is almost definitely negative-utility?

Poor Folk Do Smile, it’s true, but they also frown and cry.  And (of course) using the “Poor Folk Do Smile” tagline for this issue is in fact grossly misleading, because only a few of the relevant factors have anything to do with material deprivation.  There is a lot of sadness in the world!  A lot of boredom!  A lot of angst, a lot of fear, a lot of physical hurt!  And, let’s be clear, there’s also a lot of pleasure…but our moments of actual experienced happiness, the moments when positive utility is registered, are not so vastly more prevalent than our moments of disutility as to obviously outweigh them by any substantial margin.  No one’s measured the joy against the pain to see which one predominates, no one possibly could, but it seems obvious that the default estimate should be something like “they’re pretty close.”  Meaning that in many cases the pain wins out.  Lives that aren’t worth living: they turn out to be pretty common, when you step back and look.  That’s not a Repugnant Conclusion, it’s a Repugnant Fact.

Which is really scary, and kind of embarrassing if you’re trying to make utilitarianism sound appealing.  Especially since the seemingly-logical next step involves killing all those negative-utility people, and that’s obviously terrible, and you have to come up with some sort of sophisticated patch to explain why that is.  (This is not actually that difficult.  You just need to point out that most people have a really strong preference for not being killed, even if their lives are net negative-utility, and the fact that this preference may be irrational doesn’t make it any less real.  If you’re a preference utilitarian, bam, you’re done.  If you’re a more traditional type, you need to go one step further and say “pretty much everyone would take a giant utility hit from living in a world where the Department of Serenity will murder you if you start suffering too much.”  Easy-peasy lemon squeezy.  But by the time you’ve spent two minutes explaining why your ethics don’t obligate you to kill half the people in the world for Happiness Failure, your audience has probably left to go find some friends who aren’t so creepy.)

So I understand why this train of reasoning would be something that just doesn’t get discussed very often.  But I thought it was something that all the utilitarians out there pretty much understood, or at least that many of them bought into it enough that they didn’t feel they had to worry about the Mere Addition Paradox.

Maybe not?  Maybe that was just me?


And beyond that…

The Repugnant Conclusion is only a thing if you’re genuinely a total utilitarian.  If you believe that, in the normal circumstance, A+ > A.  If you believe that every positive-utility life necessarily makes the universe better overall.

I sincerely believed that this was a fringe viewpoint, held only by a few crazy people.

I mean, pretty much nobody in utilitarian circles acts like A+ > A.  We don’t talk about bringing new lives into the world as a moral obligation — to the contrary, most of the utilitarians I know are very clear on the idea that no one is under any kind of ethical injunction to reproduce.  People’s expressed visions of utopia never seem very heavily populated, and contra Scott Alexander’s version of Job’s dialogue with God, the people who formulate those utopias rarely seem concerned about the all the potential people living less-than-utopian lives that are still positive-utility.  Total utilitarianism, to me, seems like it’s just a bad operational model of the values that actual utilitarians hold — like evaluating the tastiness of a piece of food by counting how many times the eater smiles in the hour after he eats it.

And, again, it makes sense that this would fall into the category of “embarrassing thing that we try not to talk about.”  Utilitarianism doesn’t work very well theoretically if you don’t have some kind of utility-aggregating system, and the only other high-profile option (average utilitarianism) is even worse for most purposes; it generates its own share of “repugnant conclusions,” and the averaging methodology feels weird and artificial and not very much like the thing that our internal moral machinery is trying to do.  So it wouldn’t be surprising for utilitarians to throw a blanket over their aggregation failure by sticking to conversations about the utility of the actual people that we actually have to work with, where the unsophisticated pragmatic ways of calculating utility actually work just fine, and by avoiding population ethics.

But more and more, I’m seeing reason to believe that people are actually total utilitarians.  Or, at least, that they’ve talked themselves into thinking that they have to be, for consistency’s sake.

This seems super sad, to me, because total utilitarianism is ghastly on its face; even the regular conclusions are repugnant.  (Why would you ever want A+ over A?  Why would you prefer a world of 20 billion happy people to a world of 1 billion happy people?)  The real answer, I think, is that we need to come up with a better aggregator.  I actually have one of which I’m pretty fond; I call it “holistic consequentialism,” and the “holistic utilitarianism” variant works just fine.  I’ll be talking about it in depth in a later post.



I am aware that the claims I make about many lives being negative-utility are empirical claims.  I’m also aware that I provide no real empirical evidence for them, and indeed I have none; I’m arriving at my conclusions solely through introspection and experiential knowledge.

If anyone has real data on this point, I would be very interested in seeing it.

That said, I should point out a couple of things:

  • Anything that comes from happiness surveys is going to be viewed with extreme skepticism.  I have no reason to believe that such surveys provide information that is more reliable than randomness.  People lie a lot…especially to themselves.
  • In particular: the fact that someone claims straight-out that his life is worth living is not actually evidence that he’s right.  Humans are equipped with powerful wired-in survival instincts, which are very hard to overcome.  These are not guided by any overall utility-calculating module, and there’s no reason to think that most negative-utility people aren’t simply in thrall to overpowering urges telling them to live on through the suffering.
Conclusively Repugnant

2 thoughts on “Conclusively Repugnant

  1. Idomeneus says:

    Well given the audience of this blog, you may not be getting many responses to your question. But here are my own thoughts at least.

    It’s called “the Repugnant Conclusion” not the “Bullet Biting Conclusion”, so obviously there’s some sense that this whole thing isn’t reasonable. And I think most people do take this as a sign that utilitarianism based on lives not yet existing is unworkable and counter-intuitive.

    I think your intuition about net-negative value lives is not widely shared. I take it you were at least partly inspired by SSC’s post about chickentarianism lately, which also made assumptions like this. I don’t know whether you are right or wrong, but you are at least broaching territory that most people haven’t thought seriously about and don’t already agree with you.

    Besides rationalists worrying about X-risk and other futuristas, I think the repugnant conclusion comes up most often when considering whether it’s acceptable to create and kill a life when the likely option otherwise is not to create it at all (ie, whether to create the demand for both the birth and death of a cow.) It does seem pretty weird to say someone is responsible for murdering 100 cows when the alternative is that those cows wouldn’t have existed. And as much as we don’t feel people are *obligated* to have many many kids, we do celebrate people who choose to have kids in part with “the world will be better off for that new person existing”. That may be a very squishy example of this logic going on, but well we’re talking about people’s intuitions, and it seems a pretty relevant example.


  2. tomatlee says:

    I am VERY interested in your concept of “holistic utilitarianism”, having just this morning come up with the phrase myself and finding very few people using it. You say you will cover it in a later post. Where is that later post? I, for one, eagerly await your thoughts.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s