First published 18 October 2017. Last updated 19 October 2017.
One important philosophical debate in effective altruism, which doesn’t get as much attention as it should, is the divide between negative and non-negative utilitarianism. Brian Tomasik has outlined 3which can be arrived at by responding in certain ways to a set of 3 inconsistent intuitions (detailed in the same post by Tomasik). In this article, I respond to those versions and intuitions. Then I discuss the prevalence, implications, and importance of negative utilitarianism, and what non-negative utilitarians should do about it.
Tomasik’s 3 Versions of Negative Utilitarianism
- Negative utilitarianism (NU): Suffering is bad; happiness is neutral. The goal is to minimize total suffering.
This is the most reasonable version, but I think the neutrality of happiness as an axiom is irreconcilable with our lived experience. Frankly, it seems just obvious to me that happiness is good.
- Threshold NU: There’s at least one threshold of suffering (and maybe more) such that if someone suffers more intensely than that threshold, that suffering is lexically worse than any amount of suffering less intense than the threshold. For instance, maybe one minute in a brazen bull is worse than stubbing your toe any number of times and has more negative value than any number of happy experiences have positive value.
This one is weird. I don’t see why there should be such a threshold; I think creating one is unparsimonious and calls for a higher level of justification. It seems here like a post-hoc attempt to rationalize the intuition that some experiences are just infinitely painful, or painful beyond words, or whatever. I think instead of trying to rationalize that intuition, we should chalk it up to scope neglect (which is experimentally well demonstrated) and be comfortable with its wrongness.
- Negative-leaning utilitarianism (NLU): Suffering deserves vastly greater weight than happiness. For instance, one minute in a brazen bull might require millions of years of happy life to outweigh morally.
Negative-leaning utilitarians are not negative utilitarians. I may disagree with them on the ideal responses to some cases in applied ethics, but that’s a measurement problem within the same fundamental framework. The fact that many people who call themselves negative utilitarians are actually negative-leaning utilitarians is a problem in its own right, because it muddies the waters.
I think consciousness and hedonic states, however mysterious they may seem, are objective properties of the physical world. I hope that we will over time get much better at measuring them objectively, via improvements in science and philosophy and technology and such. Insofar as we do make such improvements, hedonic measurements by intellectually honest people should converge over time to match the growing science/philosophy. This puts a time limit on the relevance of the question of how negative-leaning one is.
Just as you can be negative-leaning in measurement, you can also be suffering-focused in action, or you can believe that preventing suffering is more “urgent” than creating happiness. You can do any of these things without being a negative utilitarian per se. I agree that preventing suffering should be prioritized over creating happiness in most cases today, for practical reasons. It’s cheaper. But that’s a nontrivial assessment that I’ll explore more fully in a future article.
Tomasik’s 3 Inconsistent Intuitions
Happiness can outweigh small pains: I would accept many of the pains that people normally experience in life in exchange for a sufficient amount of happiness. For example, I would accept mild nausea in exchange for extra days spent with a good friend.
I strongly agree.
Finitude of pains: No pain is infinitely worse than any other pain. One reason we might find this plausible is that we could construct a finite series of pain states — one slightly less intense than the last — starting from a given intense pain and ending with a given mild pain, and if we think each step only increases badness by a finite amount, then the intense pain can only be finitely many times as bad as the minor pain.
I strongly agree.
A day in hell could not be outweighed: I would not accept a day in hell in exchange for any number of days in heaven. Here I’m thinking of hell as, for example, drowning in lava but with my pain mechanisms remaining intact for the whole day. Heaven just wouldn’t be worth it, no matter how long. It seems like there’s no comparison. Nonexistence is fine for me — I wouldn’t be around to miss it — but hell-level suffering is just not something I would accept.
I disagree. I should point out that I definitely sympathize with the intuition! But I think throwing it out, in line with the scope neglect literature, is the least unreasonable way to resolve the inconsistency. In doing so, we become non-negative utilitarians, with whatever amount of negative leaning.
Prevalence in the EA Community
A lot of very smart and effective people are negative utilitarians, especially in relatively fringe areas like(which I do think is a real and important problem). It’s precisely because of how smart and effective they are in important areas that I think it’s so valuable to confront them on this particular idea. However, I would guess that the plurality of effective altruists are non-negative utilitarians, and that there are even more non-utilitarians in EA than true negative utilitarians.
Certain implementations of negative utilitarianism would be really disastrous, so it’s not a trifling technical issue. For example, under certain assumptions, a negative utilitarian might opt to explode the world even if it’s net happy, because the suffering is all that matters.
The question of negative utilitarianism tends to get swept under the rug because these implementations are corner cases, but as technology continues to raise the stakes across the board, we can’t keep ignoring corner cases. Also, to be clear, the importance of whether you’re a negative utilitarian or not isn’t 0 in non-corner cases; it’s just smaller. It always impacts the calculus.
What Should Non-Negative Utilitarians Do About This?
- Talk about it more often. Challenge negative utilitarians on their framework, and remember that the question matters. Do this politely and constructively, of course! Even though I think the “these are corner cases” argument isn’t watertight, it remains true that they are our allies in most day-to-day matters.
- Remember that life is not pain. (Wait, come back – I’m not selling anything!) This is a methodological/linguistic issue I’ve noticed a lot. It’s common to add up the number of farm animals an American omnivore eats in a year, multiply by their average lifespan, and say that the diet has created X minutes of suffering. I’m being a bit pedantic, but this just isn’t true. I would be surprised if any farm animal experienced net suffering for every minute it was alive (including minutes it was asleep!). It sounds trifling, but it matters, because utilitarianism is nothing without accurate measurement.
- I have been guilty of this! “X years of suffering” is catchy, and it’s a hard habit to deprogram.
- Redo hedonic calculations on your own terms. This is a good idea in general, but especially if the calculations you’ve been reading come from negative utilitarians or strongly negative-leaning non-negative utilitarians. A lot hangs on the question of whether there is net suffering in a given ecosystem, for example, and you should, if you feel capable, do your own sanity checks on the calculations involved.
I have written an article about other flavors of utilitarianism.