Contra Cotra Concerning Consequentialism

Tightening the grip of utilitarianism

Okay, okay. I just couldn’t resist alliteration in the title. I’m not really against Cotra, except maybe for the fact that…

I’m a utilitarian and she’s not.

Ajeya Cotra’s recent article about loosening the grip of utilitarianism is a very interesting read, which explores the evolution of her thoughts on morality. She explains how, as a naïve undergrad (my words, not her’s), she “latched onto” utilitarianism, but she gradually shifted away from the framework engaging with its many “repugnant conclusions” and thinking that maybe “the dream of a straightforward bullet-biting utilitarianism was probably incoherent”.¹ She laments having to apply what she calls moral grit to accept those repugnant conclusions, explaining that once the “intellectual grip” of utilitarianism was loosened she had some “breathing room”. She concludes by stating that “I am not quite a utilitarian.”

I have a lot I want to say about this article. But first, why should you even listen to me? Cotra is a big-shot EA working full-time on solving x-risks and other pressing problems. I… well, I’m that naïve undergrad she was ten years ago. There is, however, one thing I have over her:

I am a utilitarian, to the grave.

Wait wait wait, doesn’t this just emphasize my naïveté? You might think that, just like with Cotra, the ‘grip’ of utilitarianism will eventually wane on me. Of course, I can’t be epistemically certain that this won’t ever happen, but making that assumption straightaway misses some crucial context of how exactly I came to associate with utilitarianism - my story is completely different from hers. When Cotra first learned about utilitarianism, I assume she did lots of reading to fully understand the position. When I learned about it, I just thought “Wait, there’s a name for the way in which I see things? Why didn’t I know about this sooner?”.

Seriously. I am a Utilitarian.

Some of my youngest memories include thoughts on this subject. And I really do mean youngest; I pretty much rediscovered the drowning child argument, as a child myself, when I asked my parents why they wouldn’t donate money to save the lives of those less fortunate than us in India. I always fantasized about what it would like to “be” everyone, essentially conducting the veil of ignorance thought experiment. To that end, I would lament the unfairness of the world by wondering why ‘bad’ people deserve to be punished - after all, from their point of view, they’re the good guys!

A particularly vivid memory was when I reasoned to myself in primary school something along the lines of “Hitler was vegetarian, so despite all the bad things he did, he must have been an okay person”.² Ignoring the many flaws in this thinking, the point is that getting myself to care about nonhumans took no more moral work than getting myself to care about humans. While Cotra describes how “the utilitarian logic quickly drove [her] … to change [her] focus from global poverty to animal welfare”, I didn’t need the utilitarian logic to drive me since I already agreed with the conclusion!

In fact, when I first heard about Effective Altruism last year, I was put off by it. Not because it was too utilitarian, but because it wasn’t utilitarian enough. Why was everyone working on x-risks when it would be better if we all went extinct, considering the amount of suffering in the world? Why did 80,000 Hours have so many human-centric cause areas? If these people cared so much about impact, why were they working on areas which just didn’t matter? It was only when I dove deeper into the movement that I found those more obscure ideas that actually resonated with me - I’ll never forget the euphoria I felt upon first landing on Brian Tomasik’s essays.³

It was only around this time when I actually learnt the definition of the word ‘utilitarianism’.

And yes, of course I didn’t have all the answers at the time. I didn’t really consider caring about soil nematodes, bacteria, or the plausibility of suffering in fundamental physics. But the difference is that when I heard about these issues, I happily accepted them, not even feeling a twinge of caution about what they actually entailed. My views on morality haven’t gotten any more confused after considering how weird caring about these issues may seem, and I certainly didn’t feel like I needed to apply any ‘moral grit’ to do so.

So, unlike Cotra, I was born a utilitarian, through and through. I didn’t need to have it argued into me; It’s just the way I’ve always seen things.

So, Are You Actually A Utilitarian?

I think my experience might sound surprising to a lot of young EA’s, like Cotra was at the time. From her essay, it sounds like she thinks that utilitarians don’t ‘really’ resonate with it emotionally. She describes her experience with utilitarianism as having “to suck it up and eat bullets for breakfast every day until they started to taste okay.” This is, frankly, just not how some utilitarians think - the ones who’ve been utilitarian the day we were born, not having it argued into us. We resonate with it because we find it intuitive and emotionally appealing, not because we think it’s ‘rational’ and ‘intellectually grounded’. Tomasik describes morality as “crying out at the horrors of the universe and pleading for them to stop.” And Parfit used to weep at the abstract concept of suffering.

Unfortunately, the fact that many conclusions of utilitarianism are classified as “repugnant” doesn’t do much to help. It’s useful to remember that repugnance is a property of one’s perception of a conclusion, not the conclusion itself. The repugnant conclusions are called that, not because they ‘really are’ repugnant in some sense, but because nonutilitarians think they are. And there just so happens to be a heck of a lot more nonutilitarians than utilitarians to classify those conclusions as repugnant.

To illustrate what I mean, I’ll demonstrate my intuitions behind Cotra’s ‘omnibus ultra-repugnant conclusion’:

There is a hell containing a trillion innocent people suffering horribly. You have the resources to rescue all of these beings and give them rich and amazing lives filled with love and joy. Or you could use those same resources to buy a lottery ticket with the following payout:

• 999,999,999,999 in a trillion chance of tremendously increasing the suffering experienced by everyone in that hell.

• 1 in a trillion chance of creating N short-lived copies of Hitler feeling one second of slight satisfaction at the fact that those people are suffering so much.

Classical utilitarianism dictates, of course, that there is some N for which you are morally obligated to buy the lottery ticket.

When I work through thought experiments such as these, I try to evaluate “being everyone”. This pretty much takes care of all the intuitions I have about the scenario Then I counteract scope insensitivity by normalizing the subjective time in the world such that the time it takes to be everyone in all cases is the same time as a human lifespan.

So the thought experiment reads to me as the following.

You will be born, spend Ɛ time in hell, then disappear. Would you rather convert that Ɛ time in hell into Ɛ time in heaven, or take a lottery ticket with the following payout:

• 999,999,999,999 in a trillion chance of tremendously increasing that suffering in hell, but the suffering will still only last for Ɛ units of time, or

• 1 in a trillion chance of living a full human life.

Is there some sufficiently small Ɛ where you would take the ticket?

Now, I personally wouldn’t take the lottery ticket since I’m not a classical utilitarian,⁴ and in particular I don’t think that extreme suffering can be outweighed by a comparatively less extreme amount of good, no matter how many instances of that good occur. However I can certainly see why others would take the gamble, and it’s definitely not something I’d regard as “repugnant”. Perhaps someone reasons that, no matter how horrible the suffering, they’d only have to endure it for a second, or a picosecond, or a plank time.

But, see? I’m rejecting the conclusion based solely on my intuitions. I’m biased by the fact that I’ve never experienced superhappiness, I’m biased by the fact that I’ve only experienced what it feels like to be a human, and I’m biased by the fact that I don’t really experience the world in increments as small as a plank time. And I’m using these biases to inform my conclusions!

In a sense, you kind of have to use your intuitions as a utilitarian - the only things utilitarians take into account are instances of subjective experience, and you can really only evaluate those by actually being in those experiential states. You have to take into account the truth of feeling, not the truth of fact. So utilitarians aren’t biting on bullets until they taste okay, we’re biting on (vegan) chocolate that already tastes divine.

Do You “Feel” That Utilitarianism Is Correct?

As you can probably tell, I don’t feel any of the moral confusion that Cotra does, since I do in fact resonate with it intuitively. Maybe most people reading this aren’t like that. Maybe you do feel some lingering ‘grip’ of utilitarianism.

And to that, I say, just let it go. Just like Cotra did.

Look, I can’t in good faith give you arguments to let go, since I’m really biased towards utilitarianism - you should read other people’s accounts instead. But I can say this.

If the prospect of caring about soil nematodes scares you, rather than excites you, you shouldn’t worry about cause areas you just don’t care about. Seriously. I sympathize with you. I likewise just can’t get myself to care about global health, not when a single square meter of soil holds nematodes whose aggregate complexity may far exceed a human being's. I can’t get myself to care about x-risks, not when we don’t even know whether each of the billion trillion trillion trillion trillion trillion trillion atoms in the observable universe could have their own experience, no matter how simple they might be.

That being said, it’s still fine if ‘maximizing utility’ guides your thinking. Utilitarianism is only radical because it stipulates that qualia is the only source of value, but everyone agrees that qualia is a source of value. The “E” in “EA” doesn’t stand for utilitarianism.

All the prominent EA cause areas are mainstream not because they are contingent on utilitarianism, which only a small minority of EAs actually resonate with, but because they are robust to a wide variety of moral theories. As a rough guess, I would say that anything less extreme than wild invertebrate welfare (including farmed animal welfare, wild vertebrate welfare, AI welfare, and longtermism) has meaningful grounding in legitimate nonutilitarian theories - This aligns with the listed cause areas on 80,000 Hours. You should still deliberate which causes to prioritize among all of those. It’s only cause areas more extreme that require at least some utilitarian logic to care about.

And just be aware that there are EA’s who truly think our effects on soil nematodes dwarfs everything else, or who think that x-risks aren’t a bad thing. Don’t be scared of us. We’re just as emotionally attached to children dying of malaria or chickens tortured in cages as you are. The only thing is, we think there is the chance of there being exponentially worse problems in the world than even that, and we can’t sleep easy until the possibility of such problems existing is essentially zero.

We’re here to fight the injustice in the world, just as you are.

You Do “Feel” That Utilitarianism Is Correct?

But maybe, you’re in the same boat as me. Maybe you can feel in your bones that utilitarianism is the way to go, and everything else is just layering unnecessary abstraction upon unnecessary abstraction. I’m glad you exist in a tiny corner of an already tiny community. Please, no matter what, keep your mindset alive - no one else can treat the subject of suffering in nematodes, bacteria, or fundamental physics with the respect it deserves.

If you’re scared that so many people, even within EA, even former utilitarians, are against our position, just know that our position is coherent. The true nature of sentience is complicated, and we haven’t even come close to formulating a mathematical structure that captures it exactly. But that doesn’t mean there isn’t one out there, somewhere. And for the time being, we have pretty good heuristics to work off.

You’d expect, as Cotra says, that “the wise elders who had thought hardest about the intricacies of moral philosophy had never really been utilitarians in the first place.” Those people who think the hardest about moral philosophy do it because they care about those weird abstract concepts, like ‘morality’. They don’t care about the true source of value in the world. You are one of the few people who do. Never change.

1. As someone who writes a blog literally titled “The Dream of Hedonia”, this line really stood out to me. It might be one of the reasons I felt like I ‘had’ to respond to this article, in some attempt to intellectually ‘defend’ myself. Unfortunately a bit of that emotion has leaked into my writing.

2. True story by the way! I think my rationale at this time was something like factory farming kills more than Hitler did, and Hitler didn’t ‘support’ factory farming. Of course, I was too young to realise that (a) the story of Hitler’s vegetarianism is likely fabricated, and of course (b) that vegetarianism doesn’t magically counterfactually prevent all of factory farming.

3. Just to be clear, yes, I was horrified upon hearing about plausible cause areas I hadn’t even considered. AI sentience, invertebrate sentience, and the plausibility of suffering in physics, to name a few. The world was already terrible with factory farming and wild animal suffering, and there was a possibility that it could get worse? No, the reason I was euphoric was just because I finally found someone who thought like me. I didn’t feel intellectually alone for once!

4. Here I’m assuming Cotra means some sort of simple structure S for a utility function when she refers to “classical” utilitarianism. I’m assuming this because her intuitions later rely on this assumption. For example, “a massive population of rich joyful lives can be outweighed by a big enough pile of the barest pleasure” already assumes that S is at least a totally ordered semigroup under addition (it has to be totally ordered since it assumes that those two things are even comparable in the first place). I’m not a classical utilitarian under this definition since I don’t believe anything about S except that it must have a partial order.