Recently I asked for feedback on two versions of a new post, 'Rationality Lessons from Romance'. 'Version 2' was my original draft. 'Version 1' was a more recent draft edited in response to comments from a collaborator. We wanted to test whether my collaborator's comments genuinely improved the post. Version 2 now sits at 1 upvote and version 1 now sits at 16 upvotes, and I take this to be some evidence in favor of the hypothesis that my collaborator's comments improved the post.

My thanks to those who participated in that experiment; we got the information we wanted!

Thanks also to those of you who provided feedback on either version of the post. Most comments were critical, but this is often the case even with massively upvoted posts like Build Small Skills in the Right Order. My hope is that relationships posts on Less Wrong merely need to be better and more sensitive to a wide variety of sensitivities than my first attempt was, not that sex and relationships are inherently mind-killing topics. There is an amazing interplay between rationality and relationships, and I feel it would be a shame to leave them unexplored.

So I'm trying to learn how LessWrongers can discuss relationships productively.

One difficulty in extracting lessons from the feedback on 'Rationality Lessons from Romance' is that different people had different complaints. Some were 'seriously skeeved-out' by the overtly personal nature of the post, even though earlier posts of a similar personal nature have fared well on Less Wrong, and even though others didn't mind the personal-ness of the post. Some didn't like the promotion of polyamory, others didn't mind. Some thought my interactions with women were unethical, others didn't. Some disagreed with a premise of the post, that sexual jealousy and monogamy are suboptimal for some people. Some felt the brief lessons pulled out from the stories were effective, others didn't. Some thought I was, like Socrates, being skewered for looking at things with too much clarity, others thought my post was weird and confusing.

I may try to rewrite 'Rationality Lessons from Romance', incorporating as much of the feedback as I can. In the meantime, I'd like to ask the Less Wrong community to very different questions:

  1. What standard or non-standard rationality lessons can you draw from your own journeys in romance?
  2. How have your intimate relationships been affected by your implementation of rationality skills?


New Comment
48 comments, sorted by Click to highlight new comments since: Today at 5:53 AM

"Rationality" -- or, really, just a thoughtful attitude -- has some major advantages in relationships, I've found. Some important relationship skills that require rationality:

  • Being able to consider hypotheticals without bumping into ugh fields.

"If I ever cheated on you -- not that I would -- I would use a condom." You need a rational person to say this, and a rational person to hear it. An irrational person would never want to consider the possibility of cheating, and so wouldn't want to think of contingency plans; such a person might be offended by the very suggestion that a bad situation might happen. In the same vein, rational couples can discuss things like "if my business fails," "if we break up," "if this crazy feeling of limerence fades," and so on. You can't deal with bad contingencies well if you can't think about them or discuss them at all.

  • Being able to suspend the Typical Mind Fallacy.

People are different. Something could matter a lot to your partner that doesn't matter at all to you. Or vice versa. You can make each other happier if you can put effort into the things that the other person is passionate about even when you can't imagine feeling that way yourself. This is also a great way to diffuse fights before they even start; before you assume your partner has a nasty thought or attitude, find out what he's actually thinking. It can be surprising.

  • Learning from experience.

It's a rationalist skill just to be able to notice what your significant other does and predict that he'll keep doing it. So many relationship problems stem from the failure to notice patterns. Notice a behavioral pattern and you can learn to trust someone, learn his strengths and weaknesses, or learn to do what consistently makes him happier.

  • Understanding that statistics predict but don't control our behavior.

Some stereotypes are roughly true in a statistical sense -- people like high status and attractiveness in their mates, for example. If you want to keep a relationship going well, you should try to do things that statistically make relationships go well. On the other hand, people are individuals, and if it's a good idea to modify yourself to be different from the statistical majority in some way, you can and should. Statistics are evidence, but not destiny.

  • Curiosity.

Rationalists tend to have the attitude that the world is interesting. This means you will never be boring company. It also means you'll be willing to seek out new experiences with your partner.

  • Not being a moron.

I never hear "judgment" among the top-ten qualities people desire in their partners, but I think it's more important than people realize. You really want to be with someone whose judgment you can trust and whose opinions you can respect. That doesn't mean "sober and boring," it means that you could trust your partner to be responsible for a choice that would affect your life profoundly (and that often happens in long-term relationships). Rationalists, if they're doing it right at all, have better judgment.



Short version:

  • There are patterns in human sexual behavior and preferences that are perceivable anecdotally, or discoverable through research. These patterns can be used to make useful predictions.

  • In order to create a certain outcome, it helps to fulfill the criteria necessary to produce that outcome. There is a strong incentive to self-modify towards traits that are attractive to the people you want to date, unless such self-modification is sufficiently costly in terms of time, money, ethics, and sense of self.

  • Be as pragmatic as you can let yourself get away with, when needed. Don't be impragmatic unless you have a good reason!

  • There is an incentive to self-modify one's sense of self in order to satisfy the requirements for the type of sexual or romantic outcomes that you desire, and in order to allow pragmatic self-modification in other areas.

  • Learning how to date taught me more about my sense of self than my sense of self taught me about learning to date.

  • Notice instances where an attempt at self-modification fails, or feels overly strained. If you can't figure out a way to make that form of self-modification work, stop it, and try something else.

  • If you never strain your sense of self, you probably aren't being curious enough (or your current self is already sufficiently successful).

  • Try to cultivate a self that will be successful for your goals (up to limits of time, energy, ethics, and empirically-discovered traits) without feeling that you are trying hard. Instead of constantly trying to micromanage your behaviors and traits, create yourself an identity that manifests those behaviors and traits for you. (e.g. if you want to date people who like athletic partners, then, within the limits mentioned in previous points, try to cultivate an identity and sense of self that makes you want to engage in exercise and feel that exercise is an authentic expression of your values.)

  • The effectiveness of explicit verbal communication is limited, especially when not dating people who are nerdy with ridiculous IQ.

  • Examine your view of ethics for false-negatives (i.e. things that you don't currently consider unethical, but should). A useful sort of question might be "if I was to try behavior X with 100 people, how many times would they have a negative response, and am I comfortable with that?"

  • Examine your view of ethics for false-positives (i.e. things that you currently consider unethical, but shouldn't). A useful sort of question is "does this ethical principle literally ban a sexual or romantic behavior that the majority of the population currently engages in?" If you are holding yourself to a vastly more restrictive ethical standard than the vast majority of the population, then either everyone else are pigs, or you are being overly idealistic.

  • Consider the outcome of other ethical theories than your current one, and the views of people whose ethics differ from yours. (Both for enabling or prohibiting behaviors.)

  • The goal is not to maximize the average desire for you in the population, it's to exceed a threshold of people (with characteristics you want) who desire you strongly enough to actually date you.

  • Self-modify to temporally disable analytical thinking (such as throughout this post) to enjoy being in the moment.

  • Successful self-modification often involves forgetting what you were like before, and all these other bullet points.

If you are holding yourself to a vastly more restrictive ethical standard than the vast majority of the population, then either everyone else are pigs, or you are being overly idealistic.

Can't see why everyone else couldn't be pigs. "Slaves are people" was a true positive.

Certainly everyone could be pigs. But are they? Moral philosophy is heavily based on human intuitions. The fact that other people feel differently from oneself is (at least weak) evidence against one's own moral theories. There are many reasons to doubt moral prohibitions that are being slung around in public discourse about sexuality.

Declaration of morals is a social act. Isn't it a jolly coincidence how our promotion of certain morals just happens to make us look and feel better than other people? Isn't it convenient that morality gives us excuses to derogate our competitors? Perhaps even an excuse to go to war on them and take their land? While I believe that moral discourse is valuable and I engage in myself all the time, I do consider moral proposals to be suspect, because there is such a strong incentive for people to advocate morals for self-serving reasons without actually thinking them through in the context of moral theories that get fairly applied to everyone.

To a large degree, morality isn't about ethics, it's about status, signaling, tribalism, and policing others. There are many historical examples of over-reaching sexual morality such as religion, stigma against homosexuality, and suspicions about the sexuality of black men. Nowadays, there are other people who attempt moral prohibitions that are controversial, such as militant vegetarians, animal rights activists, environmentalists, and opponents of free speech on college campuses to name a few. Everyone political group wants to tell people what to do.

Since I'm a bit of a Hobbesian, I recognize that people will quickly violate any intersubjective forms of morality if they are left to their own nasty and brutish devices. Consequently, I would recognize that lots of moral prohibitions should be necessary. So, if I hear a moral prohibition articulated, shouldn't I find it credible? Not necessarily, because moral prohibitions have been appropriated by humans engaging in Hobbesian competition and weaponized. Well-founded moral prohibitions compete for our attention among a deluge of self-serving, incoherent, and hypocritical prohibitions.

Like White cards in Magic: The Gathering, moralists of every stripe think they stand on the side of goodness, law, and order, but that's hardly always the case.

In dating, there is also potential for bias proclamations about morality of how people of another gender should treat you. That's because people want to encourage behaviors that they personally find attractive, funnel the people they find attractive towards them, and avoid people they don't find attractive. If allowed to, they will stomp on the preferences of other people of their own gender, and on the preferences of the gender they are trying to date. Cross-gender claims about dating morality ("e.g. women shouldn't do this", "men shouldn't do that") are highly biased by the personal goals of the speaker, which doesn't always make them wrong, but does always make them suspect.

Now, what about moral prohibitions towards oneself? Those can still be self-serving, if you are using them to try to make yourself feel better or increase your status relative to others. See the story of the Fox and the Grapes. "I can't fulfill the desires of people I want to date? Well, I would've had to violate my lofty principles to do so anyway... what's more, the people who are doing better than me are pigs!"

Alternatively, even if your acceptance of restrictive moral principles is not self-serving, it still might show that one of your competitors has duped you and memetically neutered you for their own self-serving goals. Or perhaps you were influenced by someone else being the fox calling the grapes sour who wants you to be their partner in misery.

Moral prohibitions are important, and some of them are coherent and meaningful. Yet the above analysis gives us some priors about the sources of restrictive sexual and romantic morals which should cause us to think twice about moral proscriptions against personal development, including internalized ones.

Certainly everyone could be pigs. But are they?

Well, based on your accurate description, it would be pretty surprising if most of them didn't get morality grossly wrong.

This would have been the best post I've seen on this site in months... if it wasn't wasted as a mere comment!

Agreed, would like to see this again in the form of a top-level post. This cuts at the heart of one of the most important sets of lies we are told by society and expected to repeat.



Weaponized morality is a clearly a fact of history both on a personal and on a societal level yet it seems the best people elsewhere can do is say boo religion and boo ideology, without discussion how the ethical frameworks that support them are themselves memetically selected and spread or even generated precisely for that purpose. Even here on LW this is an unpleasant unspeakable truth despite the fact that basically all of us have the building blocks for this interpretation right in front of our noses (namley selfserving rationalizations, bits of signaling theory and accepting the concept of memes and memeplexes)!

Please make a top-level post of this HughRistik, even if you just copy paste it!



Smart people value exactitude. So, naturally, they like to follow rules... easier to be exact that way, if you have directions to follow.

I don't think that's necessarily the case. In many situations, I think that smart people are those who know when and where to break the rules. Writing is a handy example; the best writers know when not to apply what they were taught.

I for one do not like to be confused, but if I understand what I'm doing well enough, thinking in terms of rules is unnecessary. I think many smart people are the same way.



"Keeper" relationships must be antifragile.

Agreed. I don't really care how good someone is to be around when things are good, if they're bad to be with when things are bad. When things are bad is when they're the most needed.

If someone wanted lifetime monogamy, and they got married at a somewhat young age to their first serious romantic partner, and they and their partner were very happy with the relationship for several years, up to and including the present day, what would you expect about this person regarding their relationship skills? Would you guess that they just lucked out, or that they are good at partner selection, or that they are good at relationship maintenance, or all/some/none of those?

If the person attributes their relationship success to very good partner selection skills, would you find that believable?

I am such a person*. I feel very lucky, but we've put a lot of thought and effort into our relationship. So a little from column A, and little from column C.

On partner selection, I think Dan Savage nailed it, on finding "the one": "There ain't no one. There's a .67 or a .64 that you round up to one" (Although I think those are conservative numbers - shoot for a .8). More here.

My parents were also such people, and my wife's parents have been married for a long time. I suspect, as children, we internalize relationship heuristics from our parents, but I doubt there's anything unlearnable. Although, if these conjectures are true, and both partners are children of failed relationships, it might make it hard to navigate challanges.

Also, I think GabrielDuquette is on to something with "anti-fragile" in his post below.

* - Qualifiers: depends what you mean by "young" and "serious". Also, we lived together/common-law for 5 years before we were "married."


What a clever quote? Do you have it in verbatim? I'd like to recite it elsewhere.

[This comment is no longer endorsed by its author]Reply

Oh, definitely, partner selection is not about finding "the one". I do have to say, though, that Dan Savage's outlook is much more "settley" than mine. I also think the .67 is way low, and I can't relate to what he says about "relationships built on lies" in that video.

Since you bring it up, my parents have a terrible marriage (though they are still together) and my husband's parents are divorced. His brother also divorced shortly after marrying. My brother had a shotgun wedding and then came close to divorce, but his marriage seems to be ok for now.

Ya, "lies and deceit" seem a bit hyperbolic.

FWIW, our siblings' success/failure ratio is 3/4 - I have one sibling who is having a little trouble. He was in an otherwise good relationship, but they had mismatched long-term goals, and couldn't find a compromise. There's a lot of variables that have to come together, and I think that's where luck comes in...



Yeah, that's pretty much what I hear from most people. I'm so lucky! It can't be that I put careful thought into it at that young age. Even when people don't chalk it all up to chance, their first thought is that I have learned to apply good relationship maintenance (they're less likely to think that the relationship maintenance is my husband's doing).

Not that relationship maintenance isn't important, but I really believe that I have a happy marriage because I had a clear idea of what someone well-suited to be my husband would be like. I'm not sure I'd call it luminosity, but it's along those lines.



As I was starting to hint at with the "luminosity" reference, I think a long habit of introspection helped me to figure out that, yes, marriage is for me. And what particular kinds of marriage are for me. And what sort of spouse would be suited to having that kind of marriage with me. And how I needed to change in order to be a good spouse in return. And which criteria are more crucial than others.

As a teen I was glad to be able to learn from the mistakes of my peers, instead of going through that heartache myself. It seems like not everyone can do this, though? At least, I can think of a few people who always seem to have to touch the hot stove before they learn that it burns.

I was that teenager, too. I didn't want to date until I had observed enough to do it well. It worked out well - it taught me to love without needing to possess. That's been a huge life skill for me.

  1. a. Communication is the key. Neither person, even if they are trying to be completely honest, is ever completely honest. Partially, this is because a person is not always honest with themselves. b. Sometimes being nice is worth a white lie. (Though I am still iffy about this one.)
    c. It's very important to analyze your own emotions. Why are you angry/happy/sad/apathetic?

  2. a. Communication is the key, so "Don't propose solutions," but, rather, find all the variables first.
    b. It's important to understand what both people want from the relationship. Are those goals in alignment? If you have similar goals, then it's likely you can arrange your wants for maximum mutual benefit.

Another tip for "communication is the key": try to never punish them for communicating. This is really difficult, and it isn't always an option. Sometimes communicating changes the game in itself.

For instance, imagine a couple, Casey and Pat[1]. Casey has the flu, and Pat's friend suggests they go out for some commonplace but appealing social event. Pat would feel guilty going out, as the convention in their relationship is to be present and nurture the sick person while they feel sorry for themself. What to do here? Even asking could be interpreted as insensitive. Worse, it puts the ball explicitly in Casey's court: it'd make Casey the villain, denying Pat the opportunity to do something fun.

It's possible that asking the question is legitimately going to ruin the comfort of being nurtured for Casey. In this case, perhaps Pat really shouldn't ask at all. On the other hand, if Pat does ask, Casey should try as hard as possible not to resent it. They then both have to try to be as honest as possible about the stakes to them, to try to come up with the highest net utility move.

In my opinion, when the utility is approximately a tie, it's wrong for Pat to ask the question. It's a toss up, so it's not worth making Casey feel guilty for saying no. When the utility's heavily in Casey's favour, asking the question isn't so bad, because Casey won't feel guilty. And when the utility's in Pat's favour, asking the question moves them to a better outcome.

However, Casey can wreck the game by punishing Pat for even asking, no matter what the outcome is --- by being very offended, and wracking up credit in the "grievance bank". If the relationship has this kind of dynamic, then Pat may end up not saying anything, while privately believing it'd be in their collective best interest for Pat to go out and have a good time --- and quietly simmering about Casey's perceived selfishness.

I think these little negotiations of give-and-take are constantly happening in relationships, so it's important to have this long-term strategy of promoting communication as much as possible. Of course, when the topic at hand is emotionally salient, you have this pressure to play an entirely short-term game. So it's difficult.

[1] Care has been taken to avoid assigning a gender to either of them. I hope you don't find this too distracting.

Thanks, you've read my mind. Sometimes best communication is saying nothing, which took me a long time to understand (and I still struggle with it).

The major thing that improved my relationships was reading about feminism. In retrospect, this was a case where I had been generalizing from one example (my own experience), and ignoring the contrary data of others' verbal reports (confirmation bias; one argument against an army).

What were the things you realized you'd been wrong about?

How women perceive the ways men sometimes express attraction. See the recent comments by Rebecca Watson for an example.

Also I was confused by the nice guy syndrome until I read about how women perceive "nice" guys. There's actually a bit in Feynman about this, which makes explicit the implicit beliefs of many "nice" guys -- that women owe sex to men who are "nice" to them.

There's actually a bit in Feynman about this, which makes explicit the implicit beliefs of many "nice" guys -- that women owe sex to men who are "nice" to them.

It should be noted that it is not "Feynman" who makes this 'explicit'. It is the feminist who is declaring that as his or her interpretation of Feynman.

This being the rationality forum, it is worth pointing out that Feynman acted as a consummate epistemic rationalist. He was presented with a testable model of female behavior and he conducted an experiment, which happened to validate the model, rather than falsify it. Moreover, he repeated his experiment and got the same results again.

He may have falsified the model on other occasions, but did not report it in his book, possibly due the positive bias. If so, it is not a big deal, given that he committed it in an autobiography, not in a scientific or technical setting (something he warned against in his report on the Challenger disaster).

The feminist getting mad at Feynman for successfully testing a model he didn't even come up with just looks pathetic.

The trouble is that Feynman learned the wrong lesson - a failure of epistemic rationality. The lesson learned should have been "do what the successful guys do," basically - that night the MC told him what to do, and he did it, and he got laid.

There was more than that to his hypothesis, though. This extra part, which may be just a teeny bit problematic to feminists, was that it was important to think of the women as "worthless bitches" while you were doing this. This extra part was not tested, yet still gets reported in the story as if it was a good idea.

The lesson learned should have been "do what the successful guys do," basically

I haven't followed this thread, but this part sounds familiar.

Interesting link - and fair point. So yeah , "do what seems to cause success" is a better idea.

What about the "worse than a whore" bit? To me, that is the central issue.

What about the "worse than a whore" bit? To me, that is the central issue.

The real problem there is the insulting of prostitutes. Providing sexual services in exchange for money is not something worthy of derision.

The 'worse' seems to refer to manipulation. The exploitation of knowledge of another's instincts and the creative use of deception to make them do something that you know that, on reflection, they would prefer not to have done. Some consider that not-ok.

Oh, that's out of context:

“See, I flunked. The master gave me a lesson on what to do, and I flunked. I bought her $1.10 worth of sandwiches, and hadn’t asked her anything, and now I know I’m gonna get nothing! I have to recover, if only for the pride of my teacher.”

I stop suddenly and I say to her, “You… are worse than a WHORE!”

He screwed up his experiment and was trying to salvage it. Feynman certainly never seriously thought that a woman who doesn't put out after dinner is worse than a whore. If you doubt that, read the rest of the book and "What do you care what other people think?". He absolutely liked to have fun when not in a committed relationship, but there was never a hint of disrespect. And that is probably one reason he never used this way to get laid outside of this experiment.

I have read both books. And I've also read other things about Feynman's attitude towards women (search for soup).

Noticing womens' reactions to attitudes like Feynman's (that is, not 1940s women, but modern women), and treating them seriously, has helped me in my relationship.


I have read the whole book, and I have always found this exchange breathtakingly disrespectful.

I agree with you, but this experiment could not have been ethically performed on lab mice or monkeys.

And yeah, he should have abandoned it right there, rather than pushing it to this limit, as a proof that it was the model that worked, not his personal charms or maybe the bar ambiance, and conduct a new trial. Certainly would not satisfy the standard required for publication in relevant psychological journals. Then again, he was not fond of repeating it.

I agree with you, but this experiment could not have been ethically performed on lab mice or monkeys.

But mere women, well, surely they could suffer for the sake of an experiment.

Suffer is perhaps not the correct word here. I think shminux is thinking of experimenters wining and dining (or not) and then having sex with monkeys.

However there is a totally viable experimental design with monkeys, if any monkeys exhibit gift-giving as a part of mating - just make sure that the males have some resource, train some to give the gift and some not to, put them in a position where they're trying to win the attentions of a female, and watch what happens.

Ahh... so that was the difference in the two. There didn't seem to be any systematic difference. I'm glad you finally told us. :-) For what it's worth, I agree with the revisions in almost every case (I read them side by side eventually, and almost always liked Version 1 better).



[This comment is no longer endorsed by its author]Reply

My rationality lesson: sometimes even rational people need to be irrational and just vent. Learning to tell venting from rational discourse was essential for me. The difference is not always obvious to either side, and an explicit clarification is usually in order. (I suppose that it should be agreed upon first that this question is OK.) It is counterproductive to argue with your partner at this stage.

Once s/he lets out the steam and is in a condition to think more or less clearly, (and maybe a few hugs and affirmations of mutual love and respect later) a reasonable discussion might begin. If a dialog turns into a fight, then more often then not at least one side (and usually both) did not get a chance to vent.

As usual, YMMV.

It strikes me that, if you want to use personal examples, they seem best received when they are teaching something new, rather than merely illustrating something previous. Equally, if you're going to use personal examples, it is important to try and make them non-controversial so as not be distracting.

I'll also note that Eliezer has had his slip-ups as well. It does not seem exclusive to either you, or the topic of relationships :)


What standard or non-standard rationality lessons can you draw from your own journeys in romance?

Not that this is a communicable lesson, but the best practice I've ever gotten in wanting to believe what's true.

[This comment is no longer endorsed by its author]Reply

How have your intimate relationships been affected by your implementation of rationality skills?

In my review of Schelling I described a situation where knowing game theory helped me solve a relationship problem (sorta).