Sometimes it can feel to me like the worst enemy of utilitarians are utilitarians. Normies are going to get turned off to utilitarianism when they see things like Nikole Hannah-Jones hurting her own daughter's education, SBF trying to advance EA through financial fraud—regardless of the underlying truth of such claims. There's certainly a time and place for rationalists to engage with any argument, no matter how extreme the hypothetical, but at the same time, it can be wise to tailor your arguments towards your intended audience. I support e.g. Bentham's Bulldog arguing that insects matter more than people, but there's one argument I'd like to see retired in particular, and that's the extreme tail of the Repugnant Conclusion: The idea that large enough amounts of pinpricks or dust specks can outweigh any finite amount of torture.
I've actually come around to believing the RC is true. My objection is about its literal edge case.
A quick recap:
The positive form of the RC states that if you like the universe with N people who are each deliriously happy, you should like even better the universe with 2*N people who are each 99.999% as happy as the people in the first universe. If you repeat this step many times, you’ll eventually end up with an utterly massive universe of mildly happy people that you should prefer to the original utopic universe.
Similarly, the negative form of the RC starts with a universe of N people in terrible torture and says you should disprefer the universe with 2*N people each in 99.999% as much pain. Repeat this step many times, and you end up with a mathematical proof that a huge number of pinpricks can be worse than any finite amount of torture (or instead of pinpricks, imagine individual specks of dust that land in eyes for the most transient and minor of inconveniences).
Counter 1?
Let’s see if this holds up with a toy example. Say the worst possible existence is -10 utility and 10 is pure joy. At zero we cross the threshold to lives that the people living them would consider worth living.
World 1 has <<one person>> in anguish (-5 utility).
Compare that to world 2, which has <<10 people>> each in misery (-3 utility), for a total of -30 utility. We can say world 2 is worse than world 1.
Compare that to <<100 people>> each in severe depression (-1 utility), and we can say that world 3 is further worse.
Compare that to <<10,000 people>> each in mild depression (+1 utility). These people are struggling, yet they prefer life to nonexistence (even after disregarding fear of death, obligations to others, or the hope that their situations will improve). We cannot say this world is worse.
At each step we multiplied the people in pain by a factor of 10 and slightly reduced that level of pain. Subsequent worlds get worse and worse, until we hit a certain point and the direction flips. (The incredible difficulty of identifying exactly where that point lies was the focus of my last post on the RC, about its relation to the Sorites paradox, and this I stand by.)
This toy example shows that there are limits on what the RC logic can prove, but it doesn’t actually object to the RC logic itself; after all, world 3 was agreed upon to be worse than world 1. Worse, though, this example is deceptive. If every step of the process spawns entirely new lives, then each life added will come with all sorts of new joys and misery. If we want to standardize comparisons across worlds, it’d be better to fix the population size: Always start with a baseline of 10,000 happy people, then add a certain amount of torture, pinpricks or depression to subsets of those 10,000.
Counter 2?
What if the average pinprick was not bad? Not just “not that bad”, but actually non-negative over the course of a person’s life? The memory of surviving a pinprick can make me genuinely more appreciate and enjoy future moments, to an extent that might make me better off than if I’d never been pricked. (Personally, this idea feels more plausible when coupled with the observation that pain is infinitely more manageable when I know it’s only temporary.)
Sunny days wouldn’t be special if it wasn’t for rain
Joy wouldn’t feel so good if it wasn’t for pain
— 50 Cent
The above is one of the corniest rap lyrics I’ve ever heard—yet it’s been permanently lodged in my brain since 2003. That’s because though it’s common for movies, books and songs to describe pain as a necessary part of life (especially as something that builds character), it’s far less often that I see lyricists or authors waxing poetic about the hedonic treadmill. (If there's any LW poets or songwriters reading this, I would be grateful if you could come up with a few lines to dislodge this bit of 50 Cent silliness from my brain.)
The problem is: “pinpricks are good actually” doesn’t work as an objection to the RC. The negative RC says that any number of small negatives can sum to greater than a smaller number of large negatives. If the pinprick is considered a net good, then it wouldn’t count! The argument was never about literal pinpricks, but about minuscule decrements of utility. Pinpricks should be considered in isolation, any further negative or positive consequences removed from the equation.
For instance:
Instead of starting with 10,000 people, imagining spawning in lives, but instead of full human lives, imagine lives that only exist for the duration of their torture or pinprick experiences. What’s worse: 1 person who knows nothing but hours of torture, or a million people who know nothing but a goldfish’s-memory-moment of pinpricked pain?
(Note that this also avoids my Counter 1: These pinprick existences clearly have negative utility lives. Though I don’t yet know how to formalize this, I’d suggest that “worlds with lives that disprefer their own existences” are more often than not worse than worlds with fewer such lives.)
If we want a hypothetical that’s easier to intuit about, we can preserve the RC argument simply by scaling back a bit:
Instead of comparing 1 person in torture to a million receiving pinpricks, compare 1 person in torture to a million getting punched in the face.
Getting punched in the face probably builds character, but I’m not going to argue that they’re positive utility for the average human life.
And when framed this way… yeah. I can finally buy the argument. Enough millions or billions of people getting punched in the face does sound plausibly worse than a single person getting water-boarded or flayed for 72 hours, as terrible as that sounds.
Strength in trauma
I think there’s a couple lessons in here.
First, when it comes to persuading people (especially of philosophical, political or religious schools of thought), intuitions matter. You can be right about the facts, but if you can’t figure out a way to pierce people’s counter-intuitions about a fact, you’re not going to get anywhere with them. I resisted the RC for a long time because I got hung up on the pinpricks thing.
That's it for the more logical side of this post; now for the more wishy-washy side:
Second regards that feeling of mine, with pinpricks. Maybe I’m a bit odd for thinking so positively of pinpricks, but I think it’s part and parcel to humanity’s resilient nature—there’s a whole lot we can take in stride, or that we can learn and grow from, so long as we’re not forced past our breaking points.
I gave the example of pinpricks distracting from migraines, but when writing those words, I’d really been thinking about chemotherapy.
At one point, I’d been given a prognosis of more than two in three odds that my cancer would kill me, probably in a few years. This was, how do you say? Less than ideal. And chemotherapy wasn’t exactly a joy ride, either. There was the nausea, the lethargy, the psychological symptoms of depression which would hit me like clockwork (always the first three days after every infusion), and feelings of enervation which weren’t quite lethargy and weren’t quite depression, but rather a general sort of sucky feeling I can’t put into more specific words.
Yet I look back on my chemo days rather fondly. Some of the best mental health I’ve ever had were during the non-depression days of chemo. Because at the time, I didn’t just feel like a helpless bystander to my own fate. I was trying to tolerate as much chemo as I could handle, urging on the chemo to kill cancer faster than the cancer could kill me. It didn’t matter if the odds were against me—if I was gonna go down, better to to down fucking fighting.
No better way to die than sword in hand — Roddy the Ruin
The hardest part of chemo wasn’t any anxiety over mortal peril. It was those three days of depression at the start of every cycle. I wasn’t in any kind of agony, yet those interminable, joyless hours made me fully realize: I’d strongly prefer nonexistence to living every day like this. The intermittent nausea, though worse than pinpricks, had become a preferable distraction.
I was in a state of “-1 utility” during those particular days, but positive utility in those overall weeks. The cancer took from me a peace of mind I can never get back, gave me a severely bolstered base level of anxiety I’ll probably never untangle, and caused me and my family far more pain than the sum of all its silver linings.
But. The chemotherapy part? Remembering myself as a fighter, not a goner? It made me stronger (and even led to the creation of this blog). It was bad, and a whole lot worse than pinpricks, but I’m grateful for it.
So in the interest of celebrating our victories over smaller pains—survivable, learnable, appreciable pains—I want to request that we retire the whole “pinpricks can be worse than torture” hypothetical. We can make the same RC point through other means.
As a final note, please don’t go around pinpricking people because I said pinpricks are positive utility. Veracity of that claim aside, other negative externalities will surely arise.
I can't tell which parts you're questioning, or whether you're just saying "this extreme conclusion isn't compelling, so stop following axioms to their conclusions"? Personally, my values are jagged and not perfectly legible, and do NOT aggregate linearly (or simply). I value variety and have declining marginal value for most things. All of these prevent me from worrying too much about extremes or rule-based large-scale ethics.
HOWEVER, if you accept the premises (that entities have intrinsic value based on their experiences, and they are additive in some way), you REALLY need to accept the conclusion or find a flaw in the argument.
Sometimes it can feel to me like the worst enemy of utilitarians are utilitarians. Normies are going to get turned off to utilitarianism when they see things like Nikole Hannah-Jones hurting her own daughter's education, SBF trying to advance EA through financial fraud—regardless of the underlying truth of such claims. There's certainly a time and place for rationalists to engage with any argument, no matter how extreme the hypothetical, but at the same time, it can be wise to tailor your arguments towards your intended audience. I support e.g. Bentham's Bulldog arguing that insects matter more than people, but there's one argument I'd like to see retired in particular, and that's the extreme tail of the Repugnant Conclusion: The idea that large enough amounts of pinpricks or dust specks can outweigh any finite amount of torture.
I've actually come around to believing the RC is true. My objection is about its literal edge case.
A quick recap:
The positive form of the RC states that if you like the universe with N people who are each deliriously happy, you should like even better the universe with 2*N people who are each 99.999% as happy as the people in the first universe. If you repeat this step many times, you’ll eventually end up with an utterly massive universe of mildly happy people that you should prefer to the original utopic universe.
Similarly, the negative form of the RC starts with a universe of N people in terrible torture and says you should disprefer the universe with 2*N people each in 99.999% as much pain. Repeat this step many times, and you end up with a mathematical proof that a huge number of pinpricks can be worse than any finite amount of torture (or instead of pinpricks, imagine individual specks of dust that land in eyes for the most transient and minor of inconveniences).
Counter 1?
Let’s see if this holds up with a toy example. Say the worst possible existence is -10 utility and 10 is pure joy. At zero we cross the threshold to lives that the people living them would consider worth living.
At each step we multiplied the people in pain by a factor of 10 and slightly reduced that level of pain. Subsequent worlds get worse and worse, until we hit a certain point and the direction flips. (The incredible difficulty of identifying exactly where that point lies was the focus of my last post on the RC, about its relation to the Sorites paradox, and this I stand by.)
This toy example shows that there are limits on what the RC logic can prove, but it doesn’t actually object to the RC logic itself; after all, world 3 was agreed upon to be worse than world 1. Worse, though, this example is deceptive. If every step of the process spawns entirely new lives, then each life added will come with all sorts of new joys and misery. If we want to standardize comparisons across worlds, it’d be better to fix the population size: Always start with a baseline of 10,000 happy people, then add a certain amount of torture, pinpricks or depression to subsets of those 10,000.
Counter 2?
What if the average pinprick was not bad? Not just “not that bad”, but actually non-negative over the course of a person’s life? The memory of surviving a pinprick can make me genuinely more appreciate and enjoy future moments, to an extent that might make me better off than if I’d never been pricked. (Personally, this idea feels more plausible when coupled with the observation that pain is infinitely more manageable when I know it’s only temporary.)
Sunny days wouldn’t be special if it wasn’t for rain
Joy wouldn’t feel so good if it wasn’t for pain
— 50 Cent
The above is one of the corniest rap lyrics I’ve ever heard—yet it’s been permanently lodged in my brain since 2003. That’s because though it’s common for movies, books and songs to describe pain as a necessary part of life (especially as something that builds character), it’s far less often that I see lyricists or authors waxing poetic about the hedonic treadmill. (If there's any LW poets or songwriters reading this, I would be grateful if you could come up with a few lines to dislodge this bit of 50 Cent silliness from my brain.)
The problem is: “pinpricks are good actually” doesn’t work as an objection to the RC. The negative RC says that any number of small negatives can sum to greater than a smaller number of large negatives. If the pinprick is considered a net good, then it wouldn’t count! The argument was never about literal pinpricks, but about minuscule decrements of utility. Pinpricks should be considered in isolation, any further negative or positive consequences removed from the equation.
For instance:
(Note that this also avoids my Counter 1: These pinprick existences clearly have negative utility lives. Though I don’t yet know how to formalize this, I’d suggest that “worlds with lives that disprefer their own existences” are more often than not worse than worlds with fewer such lives.)
If we want a hypothetical that’s easier to intuit about, we can preserve the RC argument simply by scaling back a bit:
Getting punched in the face probably builds character, but I’m not going to argue that they’re positive utility for the average human life.
And when framed this way… yeah. I can finally buy the argument. Enough millions or billions of people getting punched in the face does sound plausibly worse than a single person getting water-boarded or flayed for 72 hours, as terrible as that sounds.
Strength in trauma
I think there’s a couple lessons in here.
First, when it comes to persuading people (especially of philosophical, political or religious schools of thought), intuitions matter. You can be right about the facts, but if you can’t figure out a way to pierce people’s counter-intuitions about a fact, you’re not going to get anywhere with them. I resisted the RC for a long time because I got hung up on the pinpricks thing.
That's it for the more logical side of this post; now for the more wishy-washy side:
Second regards that feeling of mine, with pinpricks. Maybe I’m a bit odd for thinking so positively of pinpricks, but I think it’s part and parcel to humanity’s resilient nature—there’s a whole lot we can take in stride, or that we can learn and grow from, so long as we’re not forced past our breaking points.
I gave the example of pinpricks distracting from migraines, but when writing those words, I’d really been thinking about chemotherapy.
At one point, I’d been given a prognosis of more than two in three odds that my cancer would kill me, probably in a few years. This was, how do you say? Less than ideal. And chemotherapy wasn’t exactly a joy ride, either. There was the nausea, the lethargy, the psychological symptoms of depression which would hit me like clockwork (always the first three days after every infusion), and feelings of enervation which weren’t quite lethargy and weren’t quite depression, but rather a general sort of sucky feeling I can’t put into more specific words.
Yet I look back on my chemo days rather fondly. Some of the best mental health I’ve ever had were during the non-depression days of chemo. Because at the time, I didn’t just feel like a helpless bystander to my own fate. I was trying to tolerate as much chemo as I could handle, urging on the chemo to kill cancer faster than the cancer could kill me. It didn’t matter if the odds were against me—if I was gonna go down, better to to down fucking fighting.
No better way to die than sword in hand — Roddy the Ruin
The hardest part of chemo wasn’t any anxiety over mortal peril. It was those three days of depression at the start of every cycle. I wasn’t in any kind of agony, yet those interminable, joyless hours made me fully realize: I’d strongly prefer nonexistence to living every day like this. The intermittent nausea, though worse than pinpricks, had become a preferable distraction.
I was in a state of “-1 utility” during those particular days, but positive utility in those overall weeks. The cancer took from me a peace of mind I can never get back, gave me a severely bolstered base level of anxiety I’ll probably never untangle, and caused me and my family far more pain than the sum of all its silver linings.
But. The chemotherapy part? Remembering myself as a fighter, not a goner? It made me stronger (and even led to the creation of this blog). It was bad, and a whole lot worse than pinpricks, but I’m grateful for it.
So in the interest of celebrating our victories over smaller pains—survivable, learnable, appreciable pains—I want to request that we retire the whole “pinpricks can be worse than torture” hypothetical. We can make the same RC point through other means.
As a final note, please don’t go around pinpricking people because I said pinpricks are positive utility. Veracity of that claim aside, other negative externalities will surely arise.