by Brian Tomasik
First written: 23 Mar. 2013; last update: 8 Aug. 2016

Summary

This piece discusses three intuitions about the badness of suffering that can't all be true. Depending on which is rejected, the result is either pure negative utilitarianism, threshold negative utilitarianism, or negative-leaning utilitarianism. I don't know which view I subscribe to, but fortunately, the choice isn't important, because all three flavors of NU yield roughly the same practical conclusions.

Definitions

  • Negative utilitarianism (NU): Suffering is bad; happiness is neutral. The goal is to minimize total suffering.
  • 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 is also worse than any number of happy experiences.
  • 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.

Assumptions

  • Assume I subscribe to some sort of total utilitarianism.
  • The (dis)value of experience-moments scales linearly with the number of them.
  • (Dis)utilities can be represented by (negative) real numbers or, if needed for lexicographic rankings with threshold NU, by (negative) hyperreal numbers. This assumption implies a total order over outcomes, which implies transitivity.

Three inconsistent intuitions

Following are three claims that may all seem intuitive.

  1. 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.
  2. 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.a
  3. 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.

These intuitions can't all be true. A day in hell is some extremely intense pain. But by intuition #2, it's only finitely many times worse than an extremely minor pain. But enough instances of happiness can outweigh enough instances of an extremely minor pain by intuition #1. Hence, enough instances of happiness can outweigh the day in hell, contrary to intuition #3.

NU theories differ on which intuition they exclude

  1. Happiness can outweigh small pains: Discarding this intuition leads to regular NU because it implies that happiness can't outweigh suffering.
  2. Finitude of pains: Discarding this intuition yields threshold NU, since there are some forms of suffering that are infinitely (lexically) worse than others.
  3. A day in hell could not be outweighed: Discarding this intuition yields NLU, since according to NLU, a day in hell could in principle be outweighed by a huge enough amount of happiness.

Which intuition do I reject?

  1. Happiness can outweigh small pains: It seems like small pains are fine. That said, sometimes our thinking is distorted: When we can't have something we want, we feel bad, and our desire causes suffering. If I didn't exist, I wouldn't really mind staying that way rather than popping into existence, experiencing some happy moments, and then popping back out. It still does seem like I'd rather have good moments with my friends than not exist, but the desire is actually quite weak, and it may be biased by the fact that I already exist, so I'm tempted.
  2. Finitude of pains: Consider the pain of burning. Construct a sequence of experiences starting from 1 day in hell at, say, 1000°C, to 10 days at 999°C, to 100 days at 998°C, and so on down to 10950 days at 50°C (which is still an uncomfortably hot temperature). It seems as though each next step in that chain is overall worse than the previous step. On the other hand, perhaps a brain transitions to fundamentally different dynamics at some point (or several points) along the way, and those qualitatively different dynamics at lower temperatures might be seen as lexically less bad than the dynamics at higher temperatures.
  3. A day in hell could not be outweighed: The obvious reply to this intuition is "scope neglect!". Maybe so. But if you actually asked me what it would take to outweigh a day in hell, I would say there's nothing that could compensate, and this feeling doesn't go away no matter how much I think about the question.

All told, I don't know which intuition to reject. Rejecting intuition #1 can avoid this particular conflict, but the remaining intuitions would still be vulnerable to the "Torture vs. Dust Specks" dilemma, which directly pits intuition #2 against intuition #3. In other words, the torture-vs.-dust-specks thought experiment makes it plausible that the real problem lies with either intuition #2 or intuition #3, not #1 per se. (Of course, a Buddhist or NU might reject intuition #1 anyway even if doing so doesn't solve the fundamental problem.)

Contradictions are expected

My brain is a jumble of different impulses and subroutines, so it's unsurprising that my intuitions don't all cohere. The famous quote from Walt Whitman is

Do I contradict myself?
Very well then I contradict myself,
(I am large, I contain multitudes [of brain subprocesses].)

David Eagleman makes the same point in Incognito: The Secret Lives of the Brain, Ch. 5: "The Brain Is a Team of Rivals".

Practical implications aren't much affected by the choice

The ratio of expected pleasure to expected pain in humanity's future is perhaps on the order of 1:1 to 10:1 and certainly isn't higher than 100:1 or 1000:1. Even if I adopted the NLU position, I would require more than 1,000,000 times as much happiness to outweigh suffering (where, for concreteness, we might say that magnitudes of "happiness" and "suffering" are defined as being those that the median classical utilitarian would use; this specification is necessary because there's no objective answer to how much happiness or suffering a given organism experiences). So regardless of which of these theoretical NU stances I adopt, the practical conclusions should be roughly comparable.

Likewise, even if I were NU rather than threshold NU, I might still care vastly more about extreme suffering relative to mild suffering, and since the amount of extreme suffering in the world is not astronomically smaller than the amount of mild suffering, a focus on preventing extreme suffering might still be most important.

Acknowledgements

MichaelExe very helpfully pointed out a major flaw in the original version of this piece. This thread contains more discussion of the original flaw.

Original version of this piece

"Am I NLU or NU?" on Felicifia.

Footnotes

  1. Larry Temkin responds to this kind of argument by rejecting transitivity, but I see this solution as too absurd to consider as a possibility.  (back)