One of the common critiques of total utilitarianism is Parfit’s The Repugnant Conclusion. It was originally formulated:
For any possible population of at least ten billion people, all with a very high quality of life, there must be some much larger imaginable population whose existence, if other things are equal, would be better even though its members have lives that are barely worth living.
It is my contention that this is a lot less interesting than people think it is.
This is a chart I made in Paint. Essentially, what the repugnant conclusion means is that total utilitarians are indifferent between any point on the red line. We are equally okay with a world of trillions of people with lives barely worth living, a world of billions of people with pretty good lives, or a world of millions of people with ecstatically happy lives.
This means that the repugnant conclusion could just as well be phrased “for any population of ten billion people, all with a very high quality of life, there is some imaginable smaller population with a much much higher quality of life which, if other things were equal, is preferable.”
First, there is nothing in total utilitarianism that says that you have to prefer the high-population-lower-utility to the low-population-higher-utility. You could very well say that, if the population * utility of two states of the world are equal, people should prefer the happier one.
Second, the Repugnant Conclusion depends on the assumption that lives cost less than happiness does. Think about it this way: one implication of the Repugnant Conclusion is that if you could devote all the resources of the earth to make a single, ecstatically happy human, trillions of times happier than any human has ever been before, then it would be ethical to do so. This is not particularly interesting to total utilitarians, because obviously we cannot devote all the resources of the world to create a single ecstatically happy human. They would be tremendously lonely and having all of Europe doesn’t do you much good once you already have all of America.
Similarly, the actual Repugnant Conclusion depends on this unstated premise: if you have X amount of resources, you could use that to either make two people who have 1 utility each, or one person who has 1.9 utility, and therefore it is always ethical to devote those resources to creating more people and no one gets to have more than 1 utility. Without this assumption, the Repugnant Conclusion is as trivial as the conclusion that we ought to create a single human with trillions of utility. If you can use your X resources to create two people who have 1 utility each, or one person with 4 utility, it is obviously better to create the single person, and the Repugnant Conclusion is nothing but an amusing parlor trick.
However, this premise seems dubious. It is possible to increase both population and utility per person (for instance, by curing diseases or preventing famines): quality of life has improved tremendously over the past couple centuries, even as the population has soared.
Furthermore, one may look into less resource-intensive ways of making a life worth living. Having more resources does not necessarily make people happier. New Yorkers use fewer resources than most Americans; their carbon footprint is 30% of the average American’s. The claim that a New Yorker’s life has 30% of the utility of an average American’s seems, at best, dubious.
A total utilitarian utopia may be one where everyone lives in a dense urban environment, takes the subway to work, only eats meat on special occasions, and takes vacations a few days away from where they live. However, they can still have interesting work, deep and meaningful relationships, access to the world’s great art and music, and a sense of contributing to the world. That is an unusual definition of “life barely worth living”.
Furthermore, at current margins, it costs fewer resources to make a person happier than to create a person (although not by much). Therefore, we can guess that– whatever the ideal point of trading off putting-resources-into-lives and putting-resources-into-utility is– it is probably not too far below where the median person on Earth is now. Total utilitarianism may require a population of ten or twelve billion people; but it is unlikely to call for a population of thirty or forty billion.
Thing of Things 
This seems more like an argument for a form of Ulitarianism that only cares about the population who already exist or who are likely to exist and not people who only exist because making them exist creates additonal utility.
LikeLiked by 2 people
My problem with this line of attack is that it could possibly imply that it’s okay to take actions now that will result in future generations being stuck in a RC scenario, if doing so benefits the currently existing population. Adding the qualifier “likely to exist” like you have defends against that. But I think “likely to exist” needs to be defined more explicitly.
LikeLike
Utilitarianism is supposed to cover the fairly simple ideas that the well being of entities capable of subjectivity is what matters, and to give useful guidance when addressing trade offs.
It’s not going to give clear answers at tolerances beneath our ability to reckon, nor will it generate results over domains for which it is not defined. Once you recognize this, the vast majority of utilitarianism related shenanigans evaporate.
LikeLike
I suspect that you are overestimating how much it costs to create a person because you’re implicitly assuming that you have to create the child with a comparable standard of living to your own.
There are a lot of differences between death and birth, at least according to my ethical intuitions, but still – in this case we’re considering them as QALYs or hedons, right?They’re analogous.
When you look for ways to save the marginal life, you look to the third world. You don’t add up the average American’sfood and housing bills and call it The Cost Of A Life.Poor Folks Do Smile, so each one you add is gonna be adding a lot of the things.
i
LikeLiked by 2 people
I disagree with part of the first point.
“First, there is nothing in total utilitarianism that says that you have to prefer the high-population-lower-utility to the low-population-higher-utility. You could very well say that, if the population * utility of two states of the world are equal, people should prefer the happier one.”
That just means you’d either be indifferent between a low-population high-utility and a high-population with a slightly higher average utility than you did before. This would change the slope of the indifference line, but not the end values, which are what we care about.
(Either that or the rule continues on the new equal utility line, in which case you would always prefer 1 person with slightly higher happiness than any arbitrary number of people that have less happiness than the single person.)
LikeLiked by 1 person
Will comment more substantially later, but shouldn’t the graph be y=c/x, where c is some constant? Assuming you mean something like “twice the population with half the quality of life” that is.
LikeLiked by 3 people
I’ve never seen “a life barely worth living” described in any satisfactory way, but that aside, it seems to me that focusing on total utility waters down some of what’s good about utilitarianism – it should be about making people as well-off ad possible in net, and creating new people with relatively low utility (especially by making existing people worse off) misses the point. For example, at the extreme, if a really happy population of 100 has the choice between making themselves even happier or adding additional people with equal or lower utility, it seems obvious to me that they should do the former. Total utilitarianism divorces well-being from people’s well-being (for lack of any better way to put it).
LikeLike
I will never understand people who try to give counterarguments against utilitarianism, because for that to work, there would have to be utilitarianists who are actually making sense. (sorry for trolling, but I’m somehow really not in on the joke.)
LikeLike
This is on a different point than what you’re talking about (or maybe it isn’t?), but I think Eliezer Yudkowsky’s response is also worth mentioning.
LikeLike
So perhaps I’m not fully understanding the flow of your argument, but I don’t think it really addresses the paradox as I understand it. Your argument seems to show, among other things, that the repugnant conclusion will never in practice forse utilitarian to advocate making numerous barely happy people. However, I don’t see it doing anything to address the concern that utilitarianism nonetheless entails that for every population of a given size and mean happiness, there is a population which is much larger and has very low mean happiness which utilitarianism judges as preferable. The paradox is that utilitarianism makes this judgment, which many regard as intuitively false.
LikeLiked by 1 person
I like this post a lot.
Related thought experiments: instead of imagining creating or removing (isn’t *that* a nice euphemism?) people I like to imagine I’m cruising around the galaxy in my starship, whereupon I stumble upon two asteroids about to hit two planets, and I only have the time/resources to divert one asteroid.
Scenario A: small population of people who are consistently blissfully happy vs. large population of people who are consistently *just* happy enough to not want to kill themselves.
I feel like I should save the small population.
Scenario B: small population of people who are consistently blissfully happy vs. large population of people who are sometimes very happy, and sometimes are very sad, who think “It’s been a very rough life with it’s ups and downs, but I feel *just* good enough, on average, to want to keep living”
I feel like I should save the large population.
My conclusion: I don’t know how to compute the “proper” utility function, but it’s clearly not just raw “happiness”.
LikeLike
Two possibilities come to mind:
1. People often place value accomplishing things in their life, a value that is separate from the experience and emotions generated by that achievement. Most people would prefer actually accomplishing a goal over being tricked into thinking they accomplished it, even if they know they would never find out about being tricked.
In scenario B you therefore perceive the ups as having greater value because people are probably accomplishing and achieving things during those ups. In A, by contrast, it sounds to you like they are just passively existing.
2. You are perceiving Scenario A as being boring for the large population, and forgetting to add in the boredom when you do the calculations for whether or not their lives are “happy” enough to want to live. If you added in the boredom, it would actually make them bored enough to want to die.
LikeLiked by 1 person
I was debating some egoists on an SSC thread and a couple of them pointed out that there is no experience of total utility because everyone only experiences zir own utility. This didn’t convince me of egoism, but it did make me question the idea that adding more people is a solution. It still seems that increasing the utility of the people who already exist is clearly good, but the number of people in existence seems like a separate moral question.
LikeLike
I think it’s pretty obvious that the RC is impractical. In most realistic scenarios helping existing people be happier is usually more practical than adding more marginally happy ones.
But I still find it to be a troubling and annoying problem because years of trolley problems have conditioned me to believe that ethical systems should work in crazy, unlikely scenarios. It’s unlikely I’ll ever have to make choices related to the RC in real life. But it’s also unlikely I’ll ever have divert a runaway trolley. It’s still something I take seriously, even though I’m unlikely to ever encounter it.
I think one of the stronger arguments against the RC is that people seem to oppose it on the micro-level as well as the macro level. When you ask people questions like “Is it okay to kill someone if you conceive a baby to replace them?” or “Which is better, two kids who live to 35, or one who lives to 70?” they tend to answer “no,” and “one to 70.” This is consistent with people placing greater value on concentrated happiness. And since these questions involve only two people, instead of millions, it’s much less likely that their answers are caused by scope-insensitivity.
LikeLiked by 1 person
I base most of my population ethics around a rejection of the repugnant conclusion (see eg http://lesswrong.com/lw/mkk/integral_vs_differential_ethics_continued/ ).
But I think you’re missing the point of the RC here. The point of hypotheticals like this is to test theories to destruction, to see the edge cases where they may not work. The total utilitarian, when faced with a simple binary choice “lots of people with lives barely worth living” versus “small population with flourishing lives”, concludes the first. And that, in my view, is ethically wrong. And don’t get me started on the Very Repugnant Conclusion.
(I notice you’re “testing total utilitarianism to destruction” in the *other direction* by mentioning the ultra happy single person.)
But in both cases, you’re saying that practical considerations prevent these “bad” outcomes from happening. And that’s fine, if our argument is “actually implementing total utilitarianism wouldn’t be bad, in practice”. Which is good for practical AI-programming purposes and similar, but terrible for determining a system of ethics. What are values are people using to say that RC, the VRC, or the single happiest person aren’t ideal? Obviously not total utilitarianism.
I’d much rather people said “I’m a total utilitarian, with a few patches to prevent the extreme ridiculous outcomes”, than “I’m a total utilitarian, but, don’t worry, the extreme ridiculous outcomes won’t happen for practical reasons, so I won’t have to re-examine my stated ethical system, phew!”
LikeLike
I think there are a few problems with this.
First, you can include a rule to pick the happier population if total utility is equal. But that doesn’t solve the problem. You can add one person to the large population, and suddenly total utility isn’t equal.
Second, I don’t think you’re describing a life barely worth living. If you think that a New Yorker is living a ife barely worth living, then you should think that anyone living a worse life than the average New Yorker could improve the world by committing suicide. That seems very insulting to a hell of a lot of people.
Furthermore, New York is not maximally resource-efficient for population. If it has great art and music, that uses resources which could be devoted to making more people. If it has good food, that uses resources which could be devoted to making more people. If people are taking vacations, that implies that there is somewhere to go on vacation, which could be repurposed to hold more people.
Third, I agree with Stuart Armstrong that this isn’t about what will happen, but what total utilitarianism implies.
LikeLiked by 1 person
I don’t think New Yorkers have lives barely worth living, and I have no idea how you could get that from my essay, because that is literally the opposite of my point. My point is that New Yorkers use a third of the resources of other Americans, and have lives that are just as happy, if not happier; therefore, it seems plausible that we can have lives that don’t use up very many resources but are still meaningful and happy by anyone’s definition.
I am not a population maximizer. I am a per capita happiness * population maximizer. What matters is not whether the resources used to create music could also be used to have children; what matters is whether the resources used to create music would be better used having children. Even in the ridiculous situation where it costs a child’s worth of resources to create an album– which isn’t true–it would take you 140 years to listen to all the music that’s on iTunes right now. There is no plausible situation where Total Utilitarian Utopia won’t have music.
If my morality gives me reasonable outcomes for the situations that happen, I am not sure why I ought to care that it gives me bad outcomes for ludicrous situations. “If everyone was made happier by torturing someone to death, utilitarianism says you should publicly torture people to death!”– well, yeah, but they *aren’t* made happier by torturing people to death, and that’s kind of important.
LikeLike
Something important I forgot to mention: Robin Hanson’s em scenario is one way the repugnant conclusion could literally come true. Trillions of copies of minds, constantly forked, created, and destroyed, stripped down to nothing but the minimum that allows them to work productively.
LikeLike
I am not certain that creating unhappy ems is less resource-intensive than creating happy ems. It seems plausible to me that creating a single brain is sufficiently resource-intensive that the resources to make that brain *happy* are essentially negligible.
LikeLike
And if that isn’t the case, what would you decide about the em world?
LikeLike
I don’t know. I haven’t seen it. How should I know if it is good or not?
LikeLike
But we probably have to decide now, while we’re in the unique position of being able to affect the probability of the outcome somewhat. We can’t be sure if it’s good or not, but we have to estimate it.
LikeLike
What is there to decide, and why do you think we here are in any position to influence it?
LikeLike
Being at the FHI, in close contact with Robin Hanson, involved with people who have some policy influence, and doing FAI research… I seem to be in a relatively good position to have some influence. And the whole rationalist community is in this orbit, and has at least some probability of influence.
As for what to decide… if Robin is right, do we try and stop this world from happening? Do we try to influence it, through regulation or laws or charitable movements? And we should be thinking of ways of making the bad outcome (if you think it’s a bad outcome, as I do) somewhat less bad. Eg the “slow-paced retirement” idea that Robin began pitching after people complained to him.
LikeLike
Obviously, there are better possible worlds that a singleton could enforce. But the problem is that this would require an ethical singleton with the right kinds of values.
At least an unregulated em scenario would create strong selection for the kinds of mind that prefer existence and reproduction over those who prefer it slightly less. All else equal, those are the ones we should want to exist.
A bad singleton could do a lot worse.
LikeLike
Pingback: Conclusively Repugnant | The Baliocene Doctrine
You need to realize: Many lives are not worth living. Severe suffering is rather common, most pleasures are mediocre at best. Conflict dominates human interaction. Political morality is a con. Even for food, energy and sex, we abuse and torture frequently.
Susebron mentioned suicide as an implication. But suicide is actually hard, unpleasant and socially vilified. Just the other day, I got censored on the “Effective Altruism” forum because I criticized the plan to ban suicide methods in developing countries without compensating the loss of option value. This happens to me frequently. Suicide is only use hypocritically, to pretend suffering is consensual.
The truth to absorb: Life is often terrible, average utility is probably negative, always has been and probably will be for a long time. That solves the RC in practice.
LikeLiked by 2 people
I have a life that is, by many standards, quite unpleasant; ten to fifteen percent of people with my mental disorder die by our own hands, placing us behind only bipolar disorder and schizophrenia, and schizophrenia is hardly a fair example because many schizophrenics kill themselves while in a psychotic episode and therefore cannot be taken to have reasonably assessed the pros and cons of continuing to live. However, I am genuinely happy and quite like being alive. Therefore, I assume that other people are also happy and like being alive– at least in the developed world.
Furthermore, I notice that “everyone’s lives are not worth living” is a position almost solely held by depressed people, and that as soon as they get treatment they change their viewpoint. While it is rude to attribute people’s beliefs to their mental illnesses, I do find this suggestive.
LikeLike
I don’t know of anyone who claims “everyone’s lives are not worth living”. Even Benatar makes a more subtle distinction.
You can have a world where most individual’s lives are worth living and yet average utility is negative and adding another life decreases expected total utility. None of these are mutually exclusive.
I do think most farmed animals have lives not worth living, as do most human torture victims, children dieing before age 3, wild animals and people in absolute poverty. I don’t think the existing happiness studies are good science. I don’t think humanity has a good plan to improve average utility above zero, and I don’t see a consensus that it’s their goal at all.
What I do see is an accidental strategy to destroy wild animals through land use and so perhaps reduce total disutility, but this is merely an accident and could flip at any time, e.g. through technological expansion of the biosphere. I can accept that adding some children, very selectively, can have positive marginal impact though their predicted externalities, but almost certainly not random children and certainly not all children. People don’t have good models of this and of course they never actually make the decision based on such reasoning.
I do agree there is no problem with the RC as a *theoretical* argument, but it tells us nothing useful about practice and causes more confusion than it is worth outside of academia.
It’s good that you like living; I don’t and I don’t think it matters at all for the subject matter.
LikeLike
I like how you demonstrate nicely the absurdness of the kind of arguments used here.
LikeLike
Unless I’m missing something, it seems like you’re addressing Parfit’s conclusions without really addressing the logical steps Parfit takes to get to those conclusions, or treating Parfit’s thought experiment as though it were based immediately on Benthamite utility-summation.
What about each of the individual steps in the RC argument? (High-average-utility initial condition, initial condition with addition of still worthwhile but somewhat less happy lives without affecting lives in IC, equalization and slight increase of average utility between IC lives additional lives, repeat.)
LikeLiked by 1 person
I feel like if we are buying or not buying into a belief system based on the what beliefs it licenses we are deep deep in the weeds already. Super backwards.
The “repugnant” adjective gives the game away, yeah? Whatever is causing such a disparaging adjective to be added to one idea or other, that’s the real ethic. Lots of dancing around to find pretty words to approve of what doesn’t give the yucky feeling is wasted energy. The fix was in from the moment the queasy feeling hit the scene.
LikeLike
Pingback: Ideas for Orthogonal Politics | neurorationalist
Pingback: Resolving Utilitarianism – The Reconstructionist