Meteuphoric

(Mistakes #1, #2, #3, #4)

I used to be a practicing atheist. I figured I had strong arguments against God’s existence. I talked to some Christians, and found that they were both ill-prepared to defend their views and shockingly uninterested in the fact that they couldn’t. This made them look like the epistemological analogues of movie villains; trivial to scorn.

Alas, this made me less likely to wonder if I was mistaken about the whole topic. If a person responds to criticisms of their beliefs with fluster and fascination with all other subjects, my natural  response is not to back down and think about why I am wrong.

Yet I should have been confused. If a person is apparently doing a host of things because of fact X, and the balance of evidence doesn’t seem to support X, and the person don’t appear to care about that, one should probably question one’s assumption that X is a central part of their worldview. I still think I wasn’t wrong about X, but I was probably wrong about all these people toiling under peculiar and willingly misinformed views on X.

Thinking about it now, it seems unlikely that the existence and exact definition of God is anywhere near as central to religion as it seems to a literal-minded systematization-obsessed teenager with little religious experience. Probably religious people mostly believe in God, but it’s not like they came to that conclusion and then reluctantly accepted the implications of it. It’s part of a big cluster of intersecting things that are appealing for various reasons. I won’t go into this, because I don’t know much about it, and this post isn’t about what religion is about. (If you want a post that is about that, at least a bit, Scott Alexander wrote two good ones recently that seem about right to me.)

This post is about winning arguments. If you repeatedly win an argument too easily, I claim that you should be less sure that you know what is going on at all, rather than smug. My boyfriend points out that being perturbed by the weakness of your opponents’ arguments is perhaps the smuggest way to be unsure of yourself, so maybe I just think you should be less sure of yourself as well as smug.

[Epistemic status: speculation]

If moral progress is so important, probably we should try to improve it.

1. Why have ordinary people been immoral en masse?

From a previous post:

I would like to think I wouldn’t have been friends with slave owners, anti-semites or wife-beaters, but then again most of my friends couldn’t give a damn about the suffering of animals, so I guess I would have been. – Robert Wiblin

I expect the same friends would have been any of those things too, given the right place and period of history. The same ‘faults’ appear to be responsible for most old fashioned or foreign moral failings: not believing that anything bad is happening if you don’t feel bad about it, and not feeling bad about anything unless there is a social norm of feeling bad about it.

That is, I claim that the procedure individuals use for morality has these key components:

1. Conformist moral affect: people have moral feelings, and these mostly reflect what their peers deem right or wrong.
2. Dictatorship of moral affect: moral feelings directly determine what people endorse.

So for instance if everyone around tortures puppies, most people consequently feel ok about puppy torture. And then if you feel ok torturing puppies, you assume that you are in fact ok with it, rather than for instance doing an extra step of conscious deliberation to check this.

(You might wonder what you would be consciously deliberating here: I’m not taking a stance on ethics or meta-ethics, but I think many popular stances do not equate moral correctness with ‘what a person feels like’ so it should often be intelligible to check that one endorses the output of one’s moral feelings.)

This is all a complicated way of saying ‘people do bad because they copy other people who do bad’.

I think it is valuable to say it in the complicated way, because it helps with seeing what might be done differently. It also makes it clearer why things are not so bad—if people only ever copied other people, human morality would be random, which I think is false.

I could say more about why I believe these things, but I probably won’t unless anyone especially disagrees.

2. Should everyone use a different procedure instead?

I claim that while these procedures lead to terrible moral failings by otherwise nice people, they also lead to virtually all nice moral behavior by nice people. So I wouldn’t want to abandon them hastily.

Plus, the obvious alternative seems worse. I’d probably much rather live in a society largely comprised of sheep who follow others’ lead on moral issues than one where every individual reasoned about morality themselves from first principles—in whatever time they decided to allocate to the project—and then took their conclusions seriously.

But I expect that there are mild variations on the status quo that are improvements. If we look at how change usually happens, possibly we can direct it a bit.

3. How do morals change?

On the story here, moral views should be basically stable apart from gradual drift. They would change faster if people sometimes have anomalous moral feelings (i.e. those that don’t reflect the existing consensus around them), or if some people think about what is right independent of their own feelings.

For instance, a world that doesn’t care about animal welfare would likely remain so until enough people have strong empathy toward animals that causes them to feel bad about animal suffering in spite of popular indifference, or until some people think about whether they endorse animal suffering from some abstract standpoint (such as utilitarianism), and condemn it in spite of having few feelings about it. This sounds about right to me as key ways that moral change happens, but I don’t know a lot about the history of this so I could easily be wrong.

4. How could morals change more and better?

There are probably lots of things to say about this, but I’ll say some random ones that I thought of.

I said that society moves away from existing moral equilibria by people having anomalous feelings, or people deciding to think about what is right independent of feelings. So things are likely to change more both when more people do those things more, and when the people who do those things have an easier time affecting anything. For instance, an initially uncaring society is more likely to come to care about animal welfare if more of its members find themselves empathising with animals in spite of common norms, or if that minority is respected more, or at least has more ways to barrage people who disagree with them with videos that might change their feelings.

This says nothing about the direction of change however. It isn’t obvious whether more or less change is good, or whether there are many directions change happens in, or how many of them are good. And perhaps we can say something more specific about what kinds of feelings or independent moral thought helps?

5. Separating moral feelings and moral positions

My guess is that thinking about ethics instead of acting directly on ethical feelings is usually good. Even if you think ethical feelings are a good basis for decisions, thinking about ethics seems useful because it tends to take a bunch of feelings related to different situations as inputs, and look for consistent positions across a range of questions. My guess is that if there are some ethical views that you would endorse after much thought, this method gets more information about them out of your ethical feelings than acting on each ethical feeling in a one-off fashion does.

I might be failing to think of kinds of ethical thought  that people do. The ones I’m familiar with seem to focus on trying to come up with general principles that unite a bunch of moral feelings (including feelings about how morality should involve general principles and not depend on arbitrary things like spatial coordinates).

6. Having better moral feelings

There are probably lots of ways to go about getting unusual moral feelings. You can pick them up from other cultures, or make unusual conceptual associations, or take drugs, or have some sort of weird morally relevant synesthesia. So I wonder if you can disproportionately try to cause aberrant moral feelings that are useful for moral progress. My guess is yes, but before discussing that, let’s consider common ways moral feelings do change.

First, I wonder if anomalous feelings are often just from changing which group your moral feelings are trying to conform with. If you begin to think of foreigners or women or animals as being in your social sphere, and you imagine that they don’t approve of being treated badly in certain ways, then you come to think treating them badly is immoral just by the usual process of conforming with local moral consensus.

Another kind of moral feeling seems to come from generalizing moral feelings you already have. For instance, if you have a strong sense that pain is bad, and also a sense that it is ok to whip people as punishment, then you watch someone getting whipped and see that it involves pain, you probably end up with some conflicting feelings. And perhaps if you grew up away from people getting whipped, so you have unusually weak feelings about whether it should be allowed, your sense that causing pain is wrong might win out, where it didn’t for other people in your society. So that’s another way you might end up with unusual moral feelings.

I think there are a large class of cases like this where people have moral feelings about the badness of internal states like suffering or indignity, and moral feelings about it being ok to take certain external actions, but where the external actions cause the internal states for someone else. For instance, it might feel wrong for innocent people to live in destitution and danger, and also it might also feel right to be able to control who enters one’s country. And both of these might be prevalent views. Which feelings you end up having about the overall issue of refugee quotas is then not very determined. I think in situations like this people often have unusual feelings relative to people around them because they are in a slightly unusual position—for instance, one where refugees are unusually salient.

7. A specific suggestion for having better moral feelings

I propose that a good way to have novel and useful moral feelings is to try to experience the situations and feelings of the people involved in the relevant situation, in accurate proportions. For instance, if you are making decisions about animal welfare, I expect your feelings to be different to most people’s, and also to more accurately track the ethical views you would want to have, if you have interacted with distressed chickens, and happy chickens, and competing farm-owners, and people who do somewhat better on a meat-based diet, and have spent as much more time with the chickens than the farm owners as is appropriate to their scale.

Sometimes it is possible to experience the interests of one side much more strongly than the other. For instance, you might one day be able to see that a genetically modified person is well off, but it will be harder to really experience the badness of playing God. So the proposed heuristic for honing moral feelings might seem inherently utilitarian, in that it only accounts for the feelings of conscious entities. I don’t think that’s true though. You can still set out to experience the things that might most viscerally elicit the feeling of badness of playing God. I can’t actually think of anything that would make me feel conflicted about playing God in the relevant way, so maybe I should find out what makes someone else feel bad about it, at least before I play God. My guess is that there are situations that will make me feel more uneasy about playing God, and I’m suggesting that I will have better moral feelings in expectation if I try to actually viscerally experience those.

Softer, easier, less technical subjects: the ones about algorithmically sophisticated self-replicating nano-machinery-based robots with human-level intelligence that were constructed using selection effects, and their elaborate game theoretic interactions. e.g. sociology, economics, psychology, biology.

Harder, more difficult, more technical subjects: the ones about numbers, shapes, simple substances, rocks, making and moving macro-objects, algorithms. e.g. math, physics, chemistry, geology, engineering, computer science.

Why are the easy subjects about super-complicated, hard to understand things and the hard subjects about relatively simple things?

The first theory that comes to mind (perhaps because I’ve heard it before) is that the ‘easy’ subjects are just too hard. Nobody can get anywhere in them, which does two things. It means those subjects don’t accrue any hard-to-learn infrastructure of concepts and theories. And it completely undermines their use as a costly signal of ability to get somewhere in a subject. This leaves these subjects disproportionately popular among people who wouldn’t have been able to send that particular signal in any case, and empty of difficult concepts and theories. Worse, once the capable people leave, the body of useful science grows even more slowly and interest in the subject becomes a worse signal of competence.

Or less cynically, the capable people reasonably go to subjects that are feasible to make progress on, where they can contribute social value.

At any rate, the easy subjects are seen as hard because they have more sophisticated science, and are full of impressive people. They are hard to play at a socially acceptable level, because the frontier is more sophisticated and the competition is stiff.

On this theory, in ancient times rocket science was probably left up to the least capable members of the tribe, while pointy stick science was the place for impressive technical expertise. Which sounds pretty plausible to me.

I’m not sure if this theory really makes sense of the evidence. The kinds of subjects that are too hopeless for a capable person to perceptibly outperform a fool in are the ones like ‘detailed turbulence prediction’. People do actually make progress in soft sciences, and it would be surprising to me if those people were not disproportionately capable. It might be that the characteristic scale of progress is smaller relative to the characteristic scale of noise, so a capable person can less surely show their virtue. But it is less clear that that generally aligns with subjects being harder. For instance, if you need a certain level of (skill + luck) to find breakthroughs, and breakthroughs become harder to find, then more skilled people would at least sometimes be at an advantage.

Another explanation is that everyone feels like they understand subjects relating to humans much more than they feel like they understand physics, because (as humans) human-related things come up a lot for them, so they have relevant intuitions and concepts. These intuitions may or may not constitute high quality theories, and these concepts may or may not be the most useful. However they do make soft subjects look simple and feel understandable.

I have heard this theory before, but I think mostly as an explanation by social scientists for why people are annoying. If it also explains why the hard sciences are easy, that would nicely simplify things.

Are there other good theories?

Steven Pinker describes ‘the euphemism treadmill’:

People invent new “polite” words to refer to emotionally laden or distasteful things, but the euphemism becomes tainted by association and the new one that must be found acquires its own negative connotations.

“Water closet” becomes “toilet” (originally a term for any body care, as in “toilet kit”), which becomes “bathroom,” which becomes “rest room,” which becomes “lavatory.”

“Garbage collection” turns into “sanitation,” which turns into “environmental services.”

I think a similar thing can happen with actions that are intended to carry meanings that are contrary to the context in which they are used.

For instance, eating ice cream is a happy activity. Snuggling up in bed is a happy activity. So if you are unhappy, you might try cuddling up in bed and eating ice cream. Eventually eating ice cream in bed becomes a depressing activity because it is what you do when you are unhappy. So now you have to move to some other activity that is still happy. Activities don’t become unhappy instantly, so each one can still cheer you up for a bit.

This naturally happens in the direction of good things coming to be associated with badness. The opposite associations happen too—intrinsically bad things come to be associated with goodness—but I think there is no particular treadmill in that direction. At the time of writing this, I am undressed in a half made bed on a mattress on the floor in a darkened room with little furniture at noon, and the floor is decorated with used glasses, and discarded clothes. Such situations are perhaps intrinsically depressing, but right now it just seems like what having a goal looks like, so it seems nice. In this direction, there is a negative feedback instead of a positive one. If sitting in a messy room is less depressing than usual, I’m less inclined to change to a new activity.

I asked:

Suppose you are in the business of making charity recommendations to others. You have found two good charities which you might recommend: 1) Help Ugly Children, and 2) Help Cute Children. It turns out ugly children are twice as easy to help, so 1) is the more effective place to send your money.

You are about to recommend HUC when it occurs to you that if you ask other people to help ugly children, some large fraction will probably ignore your advice, conclude that this effectiveness road leads to madness, and continue to support 3) Entertain Affluent Adults, which you believe is much less effective than HUC or HCC. On the other hand, if you recommend Help Cute Children, you think everyone will take it up with passion, and much more good will be done directly as a result.

What do you recommend?

Here are some of my own thoughts.

First, it depends on what you are claiming to do.

If you claim to be recommending ‘something good’, or ‘something better than EAA’ or anything that is actually consistent with recommending HCC, then probably you should recommend HCC. (This ignores some potential for benefit from increasing the salience of effective giving to others by recommending especially effective things).

If you claim to be recommending the most effective charity you can find, then recommending HCC is dishonest. I claim one shouldn’t be dishonest, but people do have different views on this. Setting aside any complicated moral, game theoretic and decision theoretic issues, dishonestly about recommendations seems likely to undermine trust in the recommender in the medium run, and so ultimately lead to the recommender having less impact.

You could honestly recommend HCC if you explicitly said that you are recommending the thing that is most effective to recommend (rather than most effective to do). However this puts you at odds with your listeners. If you have listeners who want to be effective, and have a choice between listening to you and listening to someone who is actually telling them how to be effective, they should listen to that other person.

Perhaps there should just be two different recommendations for different groups of people? An ‘effective’-labeled recommendation of HUC for effectiveness-minded people who will do it, and a something-else-labeled recommendations of HCC for other people. (Some readers last time suggested something like this).

I think this makes things better (modulo costs and complication), but doesn’t resolve the conflict. Now you have two categories of people, and for each category there is a most effective thing to suggest to them, and a most effective thing for them to do.

The main conflict would disappear if the most effective thing for you to recommend on the values you are using was also the most effective thing for your listeners to do, on their values and in their epistemological situation.

I think a reasonable approximation of this might be to choose the set of values and epistemological situation you want to cater to based on which choice will do the most good, and then honestly cater to those values and epistemological situation, and say you are. If your listeners won’t donate to HUC because they value feeling good about their donations, and they don’t feel good about helping ugly children, and you still want to cater to that audience, then explicitly add a term for feeling good about donations, say you are doing that, and give them a recommendation that truly matches their values.

This will probably often run into problems. For instance, the general problem that sometimes (often?) people’s values are too terrible to be spoken aloud, and they certainly don’t want to follow a recommendation that endorses it. e.g. perhaps they are sexist, and will in fact devalue recommendations that help girls. Yet they are not going to follow recommendations that are explicitly for effective sexist giving. This seems like a different kind of general (though closely related) problem that I won’t go into now.

In sum, I think it is dishonest to advertise HCC as the most effective charity, and one shouldn’t do it. Even if you don’t have a principled stance against dishonesty, it seems unsustainable as an advice strategy. However you might be able to honestly advertise HCC as the best charity on a modified effectiveness measure that better matches what your audience wants, and something like that seems promising to me.

Related: consequentialist-recommendation consequentialism