Your System of Ethics Needs a Line of Retreat

Everybody hates an ideology you can't reject in good conscience 🙅

Detail from the Wo kou tu juan (Japanese Pirates Scroll), traditionally attributed to Qiu Ying (1494-1552) but more likely late Ming or early Qing (c. 17th century)

People who aren’t effective altruists are quite busy hating on effective altruism these days. To be honest it’s always been like this, but lately it got worse. Some are openly calling it evil.

There are immediate reasons for this reaction — the SBF thing, and the OpenAI thing, and I’m not going to give any more detail on either because it doesn’t matter for this post. But there are also deeper reasons, like disagreements over whether utilitarianism is correct. And the deepest of all, the spring from which all the hate on effective altruism ultimately traces its origins, yet one that is somehow under-discussed,¹ is… that the vast majority people just aren’t that altruistic, and don’t want to be.

Usually, this basic fact of life isn’t that concerning. Everyone knows that Man is imperfect, and that’s why God / a bunch of prophets / moral philosophers have to come down from the sky / the Holy Land / your local department of philosophy to teach morals to Man. They won’t fully succeed, but at least they’ll make things marginally better. Crucially, any imperfectly moral Man can still feel good about himself even if he doesn’t uphold every moral principle laid down by God / the prophets / the philosophers. Because religions and philosophical theories are themselves imperfect. You can reject them. You can pinpoint their failures and say, “I try to be a good moral Christian / vegetarian, but there are a number of things in the Bible / Peter Singer that we probably don’t want to interpret literally, and also I know a church lady / vegan activist who’s an absolute asshole,² so I think it’s fine if I’m not a maximally ethical Christian / vegetarian 100% of the time.”

In practice, people don’t say that; they want to think of themselves as moral. What they absolutely do not want is to be constantly informed by their ethical system that they’re failing — that they’re not maximally ethical.

Effective altruism is about being maximally ethical. It’s not a system of ethics among others; it’s the system of ethics, the one that claims to be better than the rest. To be clear, it doesn’t claim to be perfect, and it in fact gladly accepts criticism. But it claims to try to be as good as possible and continuously self-improve. It tries to become more perfect. That’s what people hate.

(Religions are the opposite: they kinda claim to be divinely perfect and they don’t seek to self-improve. The Word of God is the Word of God and that’s that. That seems worse, except that everyone outside of their hardcore believers knows it’s fake and is therefore free to ignore it partially or entirely.)

In other words, the problem with effective altruism is that it doesn’t have a line of retreat. It doesn’t really allow you to escape and still feel good about yourself. Upon learning about its tenets, three things can happen:

1. you wholeheartedly agree, become an effective altruist and start donating 10% of your income to GiveWell;

2. you grudgingly agree, don’t do anything, and then are constantly reminded you’re not a maximally ethical person;

3. you disagree, come up with reasons why, and then either:

   1. you get destroyed in internet arguments with effective altruists, realize you’re wrong, and go back to picking one of the 3 options;

   2. or, you actually convince the effective altruists, make it self-improve, and then go back to picking one of the 3 options.

I’m exaggerating for effect: it’s also possible to sort of stay at step 3. You can find some disagreement that effective altruists can neither destroy or accept, and stick with that. For instance, say that the malaria nets are good but the focus on AI risk is bad. Or claim that the whole movement is a cult. Or (my preferred strategy) say that there’s something wrong with effective altruism’s aesthetics. These various positions are what a lot of people eventually land on. But it’s an uneasy balance.

It’s finding a small stronghold on the edge of the battlefield and hanging on for your life, rather than retreating with dignity. That option is simply not possible, when the effective altruist memes start invading, unless you’re willing to admit you’re not a good person. And you’re not willing to do that.

Ironically, I’m borrowing the line of retreat metaphor from rationalist high priest Eliezer Yudkowsky, who in turn borrowed it from Sun Tzu:

When you surround the enemy
Always allow them an escape route.
They must see that there is
An alternative to death.

Sun Tzu discusses it in the context of military strategy: it’s easier to defeat an enemy on the battlefield if they think they can escape than if they’re fighting for their lives. Yudkowsky discusses it in the context of Bayesian reasoning: it’s easier to correctly assess a piece of information if you first visualize a world where the piece of information is true. (Quite frankly I’m not that impressed by Yudkowsky’s metaphor, but whatever.)

Another detail from the Wo kou tu juan, showing the Ming troops on their way to fighting the Japanese pirates and drive them off of the coasts of Taiwan

What does it mean for a system of ethics or an ideology to allow a line of retreat?

It means to clearly make it okay for people to consider your ethics or ideology seriously, and then reject it anyway. It means not thinking less of people who consciously decide not to adopt it.

Most ideologies don’t do this very well by default. Islam forbids apostasy; in some countries and in rare cases, it can even be punished by death. But like I said, religions are also relatively easy to not take too seriously. They bundle together a lot of stuff: fairy tales, traditional rituals, clearly obsolete moral doctrines and models of the universe, etc. There are a lot of plausible reasons to be a less-than-maximally-devout Muslim or Christian or Buddhist. Especially considering that there are lots of less-than-maximally-devout Muslims and Christians and Buddhists to learn from. Besides, most people are born into their religion, and we tend to tacitly accept that it’s fine not to be a maximally devout anything if you didn’t make a deliberate choice to follow it as an adult.

Or at least, religions are easy to not take too seriously when there are many of them and it’s clear there isn’t one that beats the others. In past societies, like Medieval Europe, where 100% of who you knew or ever heard about where of the same religion, it was probably much harder to find the line of retreat.

Which brings us to the thing that provides a great escape route from most modern ideologies: pluralism. Diversity of thought. You don’t want to be a conservative or liberal or social democrat or libertarian or communist anymore? In most countries, that’s fine! Some of your friends may be mad at you, but everyone accepts that political ideologies other than one’s own exist and can be reasonably adopted by thoughtful people (even if doing so, obviously, is a mistake).

Even in the realm of normative ethical theories, pluralism is valuable. Not long ago I wrote about consequentialism, deontology, and virtue ethics, and although I’m attracted to the latter, I don’t think it’s necessarily superior to the other two. It seems to me that the three theories will always uneasily coexist — and that’s good! If one of them won out, then everyone would know exactly how to be maximally ethical, and yet wouldn’t be maximally ethical. That would be annoying.

Put differently, we don’t really want to “solve” ethics. Because then we’d have to be ethical rather than argue about ethics!

Effective altruists, to some extent, claim to… well, they don’t exactly claim to have solved ethics, but they’re certainly trying their best. Whenever they find an escape route, they do everything they can to shut it down. In theory, that’s how you win. In practice, it means that their opponents — a group which, by default, includes everybody who’s not trying to be maximally ethical — end up fighting more and more desperately.

Can effective altruism, as a movement, do anything about it? I don’t know; the whole reason it is successful is because of its grand claim to do things better than everyone else. If it didn’t do that, it’d just be “altruism” (or effective-ish altruism!). In that sense, all of this might just be a branding problem. Maybe the solution is simply to somehow make people see effective altruism not as a holier-than-thou ideology, but just one way to do charity out of a plurality of approaches.

Or maybe it can’t escape this branding, and will instead just double down, and too bad if their opponents get so mad they manage to defeat them and deal the finishing blow. It would be unspeakably tragic for the people who benefit from malaria nets or are worried about AI risk, but I bet a whole lot of other people would, secretly, breathe a sigh of relief.

Another detail from the Wo kou tu juan. Full scroll here

1. The one person who recently discussed it (and almost made me decide against writing this essay) is everybody’s favorite supercontroversial blogger, Richard Hanania. He just wrote Effective Altruism Thinks You're Hitler, which I don’t necessarily endorse but does make a number of good points.

The piece it responds to, Scott Alexander’s In Continued Defense of Effective Altruism, is very much worth reading too. Despite all the criticism, it must be repeated that effective altruists do a lot of good, and I’m still broadly sympathetic to the movement.

2. The funniest church-lady-as-an-asshole story I know is when my friends got married earlier this year. In preparation for the ceremony, they had left some wedding decorations in the church basement a few days before. Two incredibly petty church ladies then proceeded to throw the decorations into the thrash, for no reason other than thinking they belonged to a third church lady whom they hate and wanted to annoy.