The Heart Has Its Reasons That Reason Knows Not Of


Psychoanalysts argue that sons are attracted to women who look like their mothers, because they imprint on their mothers and use them as a schema for their ideal woman.

(and probably something similar for daughters and their fathers, though the psychoanalysts don’t usually get around to talking about that)

I’ve counterargued that sons get half their genes from their fathers, who apparently were attracted to women who looked like the sons’ mothers. Since study after study shows genes having unexpectedly big effects on behavior, and early childhood experiences with parents having unexpectedly small effects, maybe this would be a better explanation for the effect (if it even exists).

But the research doesn’t really bear me out.

Goose-ologist Konrad Lorenz raised goslings from birth. When the goslings grew up, he found they tried to mate with humans, especially “Caucasian men with white beards”. He concluded that they imprinted on their adoptive parent (him) and learned to prefer mates who looked like that parent.

More formally, Bischof et al got some male zebra finches and arrange to have them raised by a closely-related species, Bengalese finches. Then they put them in cages with both female zebra finches and female Bengalese finches and observed which females the birds tried to court. The results were pretty striking; they overwhelmingly went for the Bengalese finches who looked like their mothers, not the zebra finches who were genetically more suitable. Spence & Smith replicated this finding with zebra fish raised by differently-colored zebra fish.

So the research shows conclusively that sexual selection is based on learned imprinting, at least in animals whose names start with the string “zebra fi”. What about humans?

I can’t find the study itself, but multiple reviews cite Jedlicka 1984, who looked at children of mixed-race couples (white and native Hawaiian). They found that both men and women were more likely to marry someone of the race of their opposite-sex parent than of their same-sex parent (eg if you’re a woman with a Hawaiian mother and white father, you’re more likely to marry a white person). This is consistent with some kind of social imprinting where your opposite-sex parent serves as a template for future romantic interest. It’s not consistent with a simple genetic theory where you just get both parents’ genes. It might be consistent with a more complicated genetic theory where mate preferences are on a sex-appropriate chromosome or get chromosomally imprinted such that you only care about your father’s preferences for women and your mother’s preferences for men, but this is hard and I haven’t seen any analysis of whether it’s evolutionarily worth it.

Enquist, Aronsson, Ghirlanda, Jansson, and Jannini (2010) starts its Methods section with “We obtained data through newsgroups and”, so you know it’s gonna be interesting. They test another feature of men sexually imprinting on their mothers: suppose you’re a man with a sibling a few years younger than yourself. That means your mother was pregnant or lactating during the supposed critical sexual imprinting window. So if men with younger siblings are more likely to have pregnancy and lactation fetishes, that suggests that sexual imprinting on mothers is really a thing. This is indeed what they found: when a person with a pregnancy or lactation fetish only had one sibling, there was a 66% chance (compared to expected 50%) that their sibling would be younger than they, p < 0.0001 in their sample of 560 such people. This was true if and only if they were between age 1.5 and 5 during their sibling’s birth, hinting at the span of the imprinting window. Of course, this is still a really poor predictor: 33% of such people got the fetish without any younger siblings, and of course most people with younger siblings don’t end up with the fetish at all. But it does look like something is going on.

(I wonder what’s up with adult baby fetishes; if you could find a similar pattern among them that would suggest they’re imprinting on the baby sibling instead of the mother, which would be fascinating. Maybe I should try to survey the appropriate subreddit.)

Hefferman and Fraley did a very similar study. They find that people born to older parents, when compared against people born to younger parents, find older faces more attractive. This was true even after controlling for the age of the participants themselves. The effect size was small but pretty consistent across different groups and measurements. I am not quite as happy about the quality this study as in the ones above, but nothing raises huge red flags.

So there’s a lot of circumstantial evidence suggesting some kind of imprinting process is going on. But what about the original question – do people choose mates who look like their opposite-sex parent? And how genetic versus environmental is it?

Bereczkei, Gyuris, and Weisfeld addresses this question directly. They get a sample of grown-up adopted daughters and their adoptive fathers. They show subjects pictures of the adoptive fathers taken when their daughters were 2-8 years old, and then photos of four similar-aged men, one of whom is the daughter’s current husband. The subjects’ task is to guess which of the photos is the husband based on which of them looks most like the adoptive father. And subjects are in fact able to do this better than chance; they pick the correct husband 37% of the time compared to 21% of the time for (each of the three) incorrect husbands, which in their sample of 242 subjects is a p-less-than 0.001 result.

Somebody else check if I’m wrong here, but I’m a little concerned about the way they calculated p-values in this study. Suppose that their 242 subjects are always accurate in identifying the husband who looks most like the father; no randomness or noise here. But now the chance for randomness and noise comes in how many of the women might have picked husbands who look like their fathers by pure coincidence. It looks like there were only 26 women used in the study, so there’s a lot more opportunity for those twenty-six people to coincidentally choose husbands who happened to look like Dad than for the 242 subjects to coincidentally choose the correct husband even if they all looked alike. Yet as far as I can tell, the p-value calculation was only done on the latter possibility.

(an analogy: suppose that I want to prove I have ESP. I predict that a coin will land heads, then flip the coin. It does indeed land heads. Then I survey one thousand observers to ask them to judge whether the coin landed in accordance with my psychic prediction. All 1,000 of them say yes, a p = 0.000001 result that could not possibly be by coincidence. This does not prove at the p = 0.000001 level that I have ESP!)

Rantala and Marcinkowska do a really excellent review of this field that found a lot of the same papers I did. They acknowledge the Bereczkei/Gyuris/Weisfeld study as very interesting, but note that the first author “replicated” these results in a 2009 study using more scientific facial measurements. That study was found to be extremely flawed (found by Rantala himself, in fact) and was retracted under a cloud of accusations of misconduct and/or inexcusable error. It’s still not clear exactly what happened there, and nobody has formally accused the original father/adopted-daughter study of anything, but it’s kind of awkward to hang your theory on a study by somebody accused of scientific misconduct for faking a very similar study that supported the same hypothesis.

(side note: I wish that people would explain retractions better. Right now I don’t feel like I have a good understanding of exactly how the 2009 study got retracted, so I haven’t learned anything that will help me spot similar problems in the future)

So in summary: there’s a lot of good evidence that animals learn mate preferences by imprinting on their parents. There’s some circumstantial evidence that humans might do this too, across traits as diverse as race, age, and lactation status. There is, however, as yet no smoking gun.


Wait, what about twin studies? Those are usually pretty great! Can we just do a twin study and sort this out once and for all?

Zietsch et al look at “a large community-based sample of twins and their partners and parents (N > 20,000 individuals) to test for genetic and family environmental influences on mate choice”. They find:

…near-zero genetic influences on male and female mate choice over all traits and no significant genetic influences on mate choice for any specific trait (!!!!!). A significant family environmental influence was found for the age and income of females’ mate choices, possibly reflecting parental influence over mating decisions.

This might be the first twin study I’ve ever seen which unambiguously breaks Turkheimer’s First Law Of Behavioral Genetics (every trait is somewhat heritable). Such a striking finding should increase our confidence in all of the above experiments a lot. So okay, I guess this issue is solved, it’s definitely just sexual imprinting on the opposite-sex parent, thank goodness, for once we have a perfectly clear noncontradictory result and we can all just go home and –

We also tested for evidence of sexual imprinting, where individuals acquire mate-choice criteria during development by using their opposite-sex parent as the template of a desirable mate; there was no such effect for any trait.


Okay, fine, let’s look at this a little more closely. They were analyzing a bunch of data from a big Australian survey of twins. This included the twins, the twins’ family, and the twins’ mates “height, BMI, age, education, income, personality, social attitudes, and religiosity”, which were the dimensions along which they tried to predict mate choice. They figured out whether imprinting was involved by checking whether twins’ mates were more like their opposite-sex parent than like other members of their family (their same-sex parent, their other twin). There was no sign of this being true, not even a nonsignificant trend.

If we want to play the dangerous game of trying to explain differences between contradictory studies instead of just dismissing everything as noise, I might argue that this looked at some pretty different variables compared to the last set. Instead of looking at facial similarities, it’s looking at things like social attitudes and religiosity; young children trying to imprint on their mother’s image can maybe be forgiven for not knowing her opinion about Asian immigrants (one of the “social attitudes” questions they asked).

The authors took a different tactic and pointed out that most of us don’t marry the first person we have a crush on, or even the second or third. Sometimes we marry the tenth person we really like, sometimes we settle for people we only like a little, and sometimes we get drunk, have sex in a cheap motel, and have the person’s parent threaten us with a shotgun unless we go to the chapel right now. So maybe the person we end up marrying isn’t a good proxy for our mate preferences per se.

(this is starting to get kind of depressing)

They investigate this hypothesis in a followup study where they directly ask twins about their preferences for an ideal mate. Here they’re able to get some more immediately visual data – preferences like tall/short, long-hair/short-hair, beards/clean-shaven, even big-breasts/small-breasts. They find…well, some things come out heritable, but the confidence intervals are really wide. There are a lot of suspicious things like hair-length preferences being impressively and significantly heritable for women but not heritable at all for men, and on one hand the two findings’ 95% confidence intervals do overlap, but on the other hand that’s because all of the confidence intervals are super-wide anyway. Overall I guess it’s nice that this study doesn’t blatantly break Turkheimer’s First Law, but I’m not going to draw any sweeping conclusions off of it.

I have one more twin study here, Lykken and Tellegen (which coincidentally opens with the same Pascal quote as this post). This experiment has an interesting design – they ask dizygotic and monozygotic twins to rate how attracted they are to their co-twin’s spouse! (I hope these researchers went above and beyond in keeping all their data confidential!) Despite those human interest stories where two identical twins separated at birth rediscover each other and find that they have both married blonde Corgi-owning optometrists named Theo, in this study there was very little resemblance between the spouses of either identical or fraternal twins. On average identical twins correlated with each other at r = 0.57 on some long list of variables, but their spouses correlated with each other only at r = 0.14 (identical) and r = 0.11 (fraternal), and most of this was just similar religious and educational backgrounds that don’t surprise us (we already know people tend to marry others of the same religion and social class, and twins are no exception).

This is interesting because it suggests a minimal role of either genetics or shared environment in mate choice, even in the most extreme circumstance (identical twins raised together). And in fact, the authors remind us that people who have had multiple long-term relationships often choose people who don’t resemble each other in more than the most superficial ways (if the authors had exposure to polyamory, they might note that most people’s simultaneous partners aren’t very similar either).

Maybe this just supports Zietsch et al’s hypothesis that mate preferences, whether imprinted or genetic, don’t really matter because we don’t have a whole shopping aisle of mates lined up to choose from and we’ve got to take whatever we can get. Yet even so there’s room for surprise. Even if only 5% of the opposite sex is interested in you, that still leaves the average person hundreds of different choices over their lifetime. Surely there should be some degrees of freedom for people to pick who they end up with. So what’s going on? If not by genes or shared environment, how do we choose our significant others?

The authors’ guess: true love, a mysterious and magical thing totally beyond scientific understanding.

If we provisionally accept our interpretation of these data, we are left with a curious and disquieting conclusion: Although most human choice behavior lawfully reflects the characteristics of the chooser and of the choice, the most important choice of all, that of a mate, seems to be an exception…we outline a theory that is compatible with these interpretations, namely that human pair bonding is relatively adventitious, based on romantic infatuation which, as Stendhal observed, “is like a fever that comes and goes quite independently of the will.”

Sure. Sounds plausible. Let’s just consider this whole matter closed.

This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

205 Responses to The Heart Has Its Reasons That Reason Knows Not Of

  1. anon says:

    Steelman of the true love theory: humans have desires to form long-term romantic attachments that come and go, and which can presumably be measured perfectly well scientifically, but which depend significantly less on the characteristics of the partner than they would like to think. People satisfice for partner characteristics, but within the universe of people they’re about to go on a date with – where you have a significant amount of demographic likeness already established, as with the religion and social class stuff – these are usually satisficed. So pairing is mostly random within these constraints.

    • Deiseach says:

      human pair bonding is relatively adventitious, based on romantic infatuation which, as Stendhal observed, “is like a fever that comes and goes quite independently of the will.”

      Since Western culture has fetishised romantic love (since the Enlightenment and Romantic poets) as the only acceptable reason to marry (no! arranged marriage, no! marrying for wealth – gold digging/dowry hunting, no! marrying because you think you and spouse would manage a nice, companionable relationship) then it’s not surprising people “marry in haste and repent at leisure” when it comes to romantic infatuation.

      When you have your entire society, including popular media, making a big deal out of “You should only marry for True Love and It’s only True Love if both of you are wandering around in a pink, fuzzy daze of sparkles, and if the fizzy pink feeling ever wears off, then you should break up and find someone else to give you the pink fizzy feeling, otherwise you are Trapped In A Loveless (and likely to be abusive) Marriage”, what surprise that people marry for reasons they can’t articulate sensibly? And that studies only get explanations like “I had a tingle in my tummy every time they walked in the room so I knew they were the one for me”?

    • Mary says:

      “Well,” said Charlotte, “I wish Jane success with all my heart; and if she were married to him to-morrow, I should think she had as good a chance of happiness as if she were to be studying his character for a twelvemonth. Happiness in marriage is entirely a matter of chance. If the dispositions of the parties are ever so well known to each other or ever so similar beforehand, it does not advance their felicity in the least. They always continue to grow sufficiently unlike afterwards to have their share of vexation; and it is better to know as little as possible of the defects of the person with whom you are to pass your life.”

      I was reading a review of Pride And Prejudice by a guy from India who quoted that as among his favorite bits, thinking of his own happy and arranged marriage of many decades. (He also really liked the opening line out of keen memory of the day when he and his mother were cornered at a wedding by a woman suggesting it was time he settled down, and he knew she had a niece/daughter/cousin all lined up for him.)

      • Deiseach says:

        From family anecdotes, I remember my mother telling me about a distant cousin who got married because her mother (a widow) went into town to the local mart, met a guy selling his cattle, and told him “I’ve a nice little heifer at home”.

        Then again, my grandfather got disinherited from the family farm and “never darken my doorstep again” (and he didn’t) by his mother because he insisted on marrying for love my grandmother, instead of the woman great-grandmother had lined up for him.

        So, you know, both sides of the coin 🙂 (And a tradition of very strong-minded women on both sides of the family, so I find the notion of women fretting over “should I lose weight to make myself more attractive to men?” rather amusing, as the attitude I was reared with was more “Men? Just tell ’em what you want them to do and if they don’t, to heck with them”).

  2. The Nybbler says:

    And in fact, the authors remind us that people who have had multiple long-term relationships often choose people who don’t resemble each other in more than the most superficial ways (if the authors had exposure to polyamory, they might note that most people’s simultaneous partners aren’t very similar either).

    A likely confounder in the first case would be people deliberately avoiding people who look like their previous estranged or deceased partner. And in the second case, novelty-seeking among polyamorists. (I’m in the “throw up my hands and dismiss it as noise” camp)

    • 2stupid4SSC says:

      All noise all the way.

      It seems to me that the confounders on selection are just too powerful. Maybe if we could simulate humans, and raise a bunch of simulated humans by simulated orcs, and see if the simulated humans are more attracted to orcs or humans when they are picking a mate. But with the kinds of studies we have access to, it seems hard to draw any strong conclusions.

    • Deiseach says:

      i wouldn’t be surprised if, when ready to settle down, marry, start a family, people select spouses that remind them of their parents – for most people, their model of a successful marriage and family life is going to be their parents (not, of course, if the relationship was abusive, ended in acrimonious divorce, etc. but that’s why I said “most”, not “all”).

      If your model for “two people who manage a successful and lasting relationship” is “mom and dad”, then I think it’s plausible to be attracted to someone who reminds you of mom/dad (depending) because you have subconscious good expectations and good associations there. That doesn’t necessarily mean “looks like mom or dad”; they could have a similar sense of humour or fashion choices* or work or the like. Contrariwise, if one of your parents was abusive, crazy, manipulative, or the like, you’re going to pick a partner that is the exact opposite of them. So if “tall blue-eyed blondes look like Mom and Mom was a bitch to me”, you are less likely to end up with a tall blue-eyed blonde wife.

      *”Anecdotes are not data”, I know, but whenever I see a particular photo online of a certain person who is wearing a cardigan identical to one my father would have, and did, wear – boom! I get hit with the warm, comfortable, friendly, affectionate feelings. I feel relaxed and happy. I have good associations with that. Is it so strange to think a like effect is at work if you meet someone who provokes the same kind of response in you for whatever set of associations, and you end up in a relationship with them?

  3. FeepingCreature says:

    Isn’t this post supposed to open with a Pascal quote?

  4. Tibor says:

    Doesn’t that actually somewhat go against the presence of sexual selection (as an evolutionary mechanism) in humans? That sounds rather odd.

    • rational_rob says:

      I don’t really think it goes against sexual selection – for one, these studies were conducted in a modern setting, where there is a wide selection of mates. Also, sometime evolutionary trends make themselves known over generations – maybe people don’t mate with people that look like their parents, but they have a tendency to marry people that look like their great-great-great grandparents? I don’t have enough data to know.

      I remember reading something on genetic sexual attraction, it came to the conclusion that people tended to mate with people that have similar genetic structures – even when certain traits don’t effect physical appearance! I’m not sure how true it is, but it would provide an alternative explanation for how sexual selection could work: genes that promoted incest among descendants would become more prevalent, until the harmful heritable diseases overcame any benefit that was created.

      But then again, Zebra Finches. If any of this were true, you would expect an animal to try and mate with something of its own species!

      • Tibor says:

        I don’t know much about genetics, but it does not make much sense to me. It is a much better strategy that (keeping fitness of the possible mates equal) you choose someone whose genes are very different from yours. That way you minimize the chances of getting two copies of a harmful recessive gene. I think there is at least some casual evidence for that. I remember a few years back when I visited a beach near Lisbon with all the Portuguese girls, every one of them seemed extremely attractive to me. Similarly I talked to some guys from Spain and Ecuador and they were really amazed by the German and Czech girls. Of course, it could just be the novelty or me being (then) a horny teenager 🙂

        In any case, people who look similar to your parents are also on average more likely to carry similar genes. Of course, in one generation it is not a big deal but it could make a difference in evolutionary time. Hence genetically (and whether it is some genes triggering imprinting of sexual preferences based on your parents or genes that influence it directly is not that important) it should be something selected against.

        It would make some sense that the imprinting serves a useful function (recognizing your parents) but has a side effect which influences your mate choice and the negative effect of that are outweighed by the positive effects of parent recognition.

        But as I said, this is all very amateur genetics speculation.

        By the way, there is a great book about sexual selection called “The Mating Mind”. It is a bit more speculative than Dawkin’s Selfish Gene, but I’d say it is equally good and enlightening (and it does not contain mildly annoying self-congratulatory atheist rants which make up some 20 pages of the Selfish Gene).

    • Deiseach says:

      It does raise interesting questions about the incest taboo: is this a hard-wired response or is it only socially and culturally encoded, because otherwise we would be marrying (or having sex with) our parents/siblings/aunts and uncles?

      Sex and romance – so messed up.

      • gwern says:

        It can be both. The Westermarck effect coexists with the reunited-sibling effect; consider the famous Taiwanese arranged-marriage thing where the adopted girl and son are expected by everyone to get married (so the social expectations are the opposite of anti-incest) but the marriages still fail at very high rates, versus the problem for siblings reunited as adults that they often find each other dangerously sexy ( The mechanism is probably something like neutering sexual response to anyone you were kids with.

    • vV_Vv says:

      Doesn’t that actually somewhat go against the presence of sexual selection (as an evolutionary mechanism) in humans?

      Unless the characteristic they tested for correlated strongly with fitness, no.

  5. hlynkacg says:

    True love? No no-no-no, he clearly said to “to bluff”.

  6. Anonymous says:

    >They found that both men and women were more likely to marry someone of the race of their opposite-sex parent than of their same-sex parent (eg if you’re a woman with a Hawaiian mother and white father, you’re more likely to marry a white person).
    >This is consistent with some kind of social imprinting where your opposite-sex parent serves as a template for future romantic interest, but not with genetic theories; why would genetic tendencies preferentially come from the opposite-sex parent?

    I think you interpreted this backwards, and that this finding is compatible with genetic theories. If the woman’s father is white, then (maybe) the woman’s mother had a genetic tendency to prefer white men; the daughter inherited this genetic tendency from her mother. So, from the same sex patent.

    (Isn’t that what your original theory was?)

    • Scott Alexander says:

      Ah, you’re right. I’ll fix.

      • Sniffnoy says:

        While this would be coming preferentially from the same-sex parent rather than the opposite one, it’s still a little surprising that it should be coming preferentially from either, right?

        Edit: Oh, nevermind, you’ve mentioned this now.

    • S_J says:

      A piece of statistical anecdata, from within my extended family…

      A cousin of mine is from a mixed-race marriage. Most of the women he has pursued romantically are the same skin color as his mother (White), rather than his father (Black).

      Confounder: the area in which the family grew up has a very low percentage of Black residents, much lower than the national average.
      Thus, the availability of women the same skin color as his father might be very low.

  7. Randy M says:

    I’m sorry, I refuse to draw any final conclusions until I see studies of humans raised by wolves. Or zebra finches. But the wolves might eat the finches.

    (Fascinating post by the way)

    • Scott Alexander says:

      I’d be interested in seeing cross-race adoption studies, since that’s about as close to a kind of global shock to similarity of appearance as you’re likely to get. But I imagine all of the cross-race adoptees are already sick of participating in the last two million studies they were recruited for.

      • Jeff Kaufman says:

        Cross race adoption is heavily confounded by social segregation. If someone of race A is adopted by a race B couple then they’re much more likely than someone raised by a race A couple to a go to race B typical schools, churches, and events, more likely to go into race B typical careers, and more likely to live in race B typical areas.

    • arbitrary_greay says:

      So what about which parent species mules prefer? Or ligers? Or what about mixed-breed dogs, or cats?

      What if you have an antelope raise a mule? A cow raised by buffalo? that cat who grew up with huskies? (wasn’t it simply scarcity driving the rise of polar bears getting with grizzlies?)

  8. Erik says:

    Aren’t twins supposed to have incredibly close relationships – to the point of inventing their own languages? Would that strong relationship be enough to reduce the effect of the imprinting? We need a study comparing their spouses to different gendered fraternal twins.

    • AxiomsOfDominion says:

      Its incredibly uncommon for twins to invent their own languages and they usually grow out of them at an early age if they do. Also twins tend to have just as much relationship variance as fraternal twins and regular siblings. I’m a twin and I don’t have a super close relationship with my twin, or older sister for that matter.

      • Erik says:

        The imprinting is supposed to happen “between age 1.5 and 5”, which would be the age those languages would exist though, right?

  9. Jordan D. says:

    The Pascal quote is good, but I hope that any deeper meta-analyses open up with the entirety of William Blake’s “The Pebble and the Clod.”

  10. Eponymous says:

    (Mildly depressing) hypothesis: there’s a general factor of attractiveness, and people pair up with people at about their level. Other factors don’t matter much, beyond the big ones that define class (i.e. income, education, race, language, religion, etc). Since these studies are mostly looking at traits orthogonal to this general factor, they don’t find much.

    • 2stupid4SSC says:

      Would a general factor of attractiveness be stronger or weaker in humans vs other animals? My gut says other animals would be more effected if it was real, but they show the strongest signs of imprinting.

      So what explanation is there for why Humans would respond to this general factor of attractiveness more?

    • Simon says:

      And that general factor of attractiveness is some complex/nonlinear combination of individual attributes (which are measured as a linear combination in the studies).

      So if two boy twins go to about the same quality of school, both get respectable jobs, they each have a roughly similar factor. They then go to the dating market and exchange that factor for the best they can find. One finds a cute short brunette with nice tits who works as an elementary school teacher, and the other a Chinese girl who teaches yoga.

      Despite the two spouses having little in common in terms of a linear combination of traits, the interaction of them maps to some general attractiveness/value factor based on the male biological algorithm that overall tends to heavily weight a nonlinear interaction of very base preferences (boobs, butt, fat, symmetry, disposition).

      Also, how many of us date/marry our ‘ideal’ women? I think, particularly for men, our optimal utility function looks like a pyramid with a very flat top. If you’re a guy who prefers tall blonde women, but a really hot short brunette girl who matches your preferences comes along you’ll probably go with it. Because the difference between an ideal and someone else isn’t particularly huge.

      Of course, this means that regression type models really just won’t capture this shit. I know we like to think linear data generating functions approximate everything close enough, but in this case I think there is a strong reason to believe they don’t. Particularly because as humans our base algorithms aren’t linear functions, and we can all simulate what it’s like to be a human making choices perfectly.

    • Scott Alexander says:

      I assume something like this is true (and it’s mentioned in the last paper I cited), but it’s still strange that you can’t get some extra gains out of factors where you put disproportionate value on certain things compared to other people.

      My spending ability is governed by my salary, but I still spend it on very different things than some other people who make the same amount of money I do. Even if there’s some mating “budget” we use to buy desirable features, I would still expect to see people spending it in specific idiosyncratic ways.

      • Stezinech says:

        Maybe it’s hard to pinpoint those idiosyncratic mate choice things.

        In terms of the general factor of attractiveness, I imagine that preference for this has gone nearly to fixation in population, so maybe it’s not so surprising that it’s not heritable. Contra Eponymous, the things they did measure in Zietsch et al. are correlated with the general factor: income, height, desirable personality, etc. It’s a shame they didn’t report the variation in these measures, or really discuss the impact on the heritability estimates.

        • Steve Sailer says:

          I suspect that “romantic chemistry” has something to do with the very complex biochemistry of immune systems, but it’s not something our current science is very good at measuring.

        • Anonymous says:

          I suspect so, yes.

          Anecdotally, whether I am physically (not intellectually-personally-socially) attracted to a man is ~completely explained by three factors:

          – Baseline non-hideousness;
          – Vocal quality;
          – Specific shape of part of [redacted facial feature].

          The first two are fairly typical for women; the last targets a trait, um, coincidentally shared by my father. Once I noticed the pattern, the conclusion was inescapable… but I think it is possible that someone with a similar preference and a different personality would never have identified it. It would probably be very difficult for someone else to deduce it using me as an oracle, and practically impossible for a researcher to detect it from a panel of standard faces or a man I happened to marry.

          I do not have an opinion on whether this is genetic or environmental, and I have no idea how common this type of incredibly specific imprinting might (unconsciously?) be. But it definitely exists and it may make these things much, much harder to measure.

  11. ozymandias says:

    And in fact, the authors remind us that people who have had multiple long-term relationships often choose people who don’t resemble each other in more than the most superficial ways (if the authors had exposure to polyamory, they might note that most people’s simultaneous partners aren’t very similar either).

    Is this true? Many people seem to believe that people have types, and my personal observations seem to back it up. When I’ve had the opportunity to meet all the partners of a poly person who is sexually successful enough not to settle and has enough partners that the noise cancels out, I’ve usually found that their partners are disproportionately trans, or disproportionately Jewish, or disproportionately butch girls, or universally people with mental-health conditions that turn them into assholes. (I dated the latter person, which I think says awful things about my character. :P)

    That said, it might be hard to pick up on this trend in studies, given that there’s hundreds of traits a person could have a type for, that people settle, and that many people have types with more than one characteristic. (If your type is long-haired goths, you might date a long-haired metalhead and then a short-haired goth, and it would appear that you didn’t have a type.)

    • rational_rob says:

      I think the biggest problem here is that we’re all relying on anecdote? I imagine that people who live in more isolated communities will be more careful about who they date, both because it might oust a fetish, and because dating the wrong person could ruin your social standing. Larger communities would encourage more specific selection.

    • Whatever Happened to Anonymous says:

      >I’ve usually found that their partners are disproportionately trans,

      How many partners would you need for a single trans partner not to make your pool disproportionally trans?

      • 2stupid4SSC says:

        That would depend on how trans the population you pull potential partners from is. If you are doing online dating then you could be pulling from an exclusively trans population and 100% trans partners would be proportional.

        Assuming the more traditional friend group dating, then it is pretty anecdotal. I don’t know many trans people but my gay friend’s social group is almost exclusively other gay men, I think I might be the only straight man who goes to some of their social gatherings. So similar friend groups that are, 50% trans, don’t seem that strange to me. In which case 1(2) out of 3(4) trans partners would be proportionate from your pool of potential partners.

        Math Edit: My name is accurate

      • ozymandias says:

        If you’re looking at global base rates, about 300.

        But I’m a Bay Area rationalist, so my current model is something like a person with no trans partners has a weird aversion to dating trans people, a person with less than 1/3 trans partners is perfectly normal person, and a person with more 1/3 trans partners is a chaser and/or an egg.

        • Eric Rall says:

          What does “egg” mean in this context?

          • ozymandias says:

            An egg is a self-closeted trans person. Stereotypically, eggs feel a strange desire to spend lots of time with trans people. Other stereotypes include being a sad nerd, only playing female characters in games, and loudly insisting that people of ANY gender can wear nail polish.

          • suntzuanime says:

            That seems incredibly evil by your ethical standards. Who the fuck are you to question my gender just because I’m a sad nerd etc?

          • Hyzenthlay says:

            Who the fuck are you to question my gender just because I’m a sad nerd etc?

            Zie did say “stereotypes,” implying that this might not actually be the case.

            But yeah, I’d say whether someone is a sad nerd, whether they wear nail polish, and what gender character they like to play are probably not reliable indicators of being trans.

          • jsmith says:

            Especially the part about playing female characters seems weird. When I pirated DOAX3, it’s because I am attracted to titties, not that I want to have them.

          • Roxolan says:

            Bayes, people. Remember the base rates. Just because those are stereotypical characteristics of “eggs” doesn’t mean that a majority of people with those characteristics are “eggs”.

          • The Nybbler says:

            What kind of games, though? DOAX3 is a volleyball game. I’d guess that always choosing female characters in a role-playing game would be a better indicator. Though of course there’s all sorts of confounders… e.g. the nail polish thing isn’t going to distinguish an “egg” from someone who is only interested in cross-dressing.

          • Protagoras says:

            Yeah, I think guys playing female characters in games is complicated, and a volleyball game is hardly the central example. Contra jsmith (and many feminists) I don’t think it’s mostly about ogling their characters (though I’m not saying it’s never about that, just that the role of that is exaggerated; I think in most cases people identify with their role-playing avatars more than that). But I also think it’s rarely about being closet trans. I suspect another big part of it (in role-playing games at least) is guys who want, as part of their escapist fantasy, to feel sexy, and who find it much easier to think of a female avatar as sexy than a male avatar.

      • Lambert says:

        OK. I’m a little rusty so tell me if I’ve got something wrong, but here goes.
        Let number of trans partners X ~ Bin(n,p)
        H0: p=underlying population probability of being trans
        H1: p>underlying population probability of being trans
        (In the USA, probability of being trans is 0.003, according to Wikipedia)

        To answer the question, one must find n, given that X=1, for a certain significance level.
        let’s accept a significance level of 0.05.

        Critical value at
        p(X>=1)=0.05 (should maybe be = not >=?)

        for n=17, p(X=0)=0.950
        for n=18, p(X=0)=0.947

        So you need 18 partners for a single trans partner to not make your pool disproportionately trans, at a 5% significance level.

    • onyomi says:

      I would also anecdotally support the notion that while there is significant variability in the type of people I can find attractive, I and people I know well do seem to have “types.” And sometimes in weird, unexpected ways. Like, how did I, a person who grew up in Louisiana, who has spent no significant time in Pennsylvania, and who has had roughly 3 serious LTRs, end up dating two women from the same region of rural Pennsylvania I’d never been to with no connection to one another (and no, neither of them were Amish, nor otherwise signalling some explicitly obvious Pennsylvania quality)?

      • Scott Alexander says:

        That sounds like coincidence unless there’s something specific about that region that could plausibly be attractive.

        • onyomi says:

          I mean, yeah, probably a result of my failure to notice all the other ways in which these two girlfriends are not the same, but the one weird way in which they are the same (though actually not the only way; also have some similarities in terms of personality and appearance) sticks out. Still, having 2 out 3 people come from a sparsely populated area I hadn’t previously been to does make me wonder if there’s something subtler at work. Maybe I was Amish in a past life.

    • Subbak says:

      (If your type is long-haired goths, you might date a long-haired metalhead and then a short-haired goth, and it would appear that you didn’t have a type.)

      Well, then you just have to look at their… platonic relationships.
      ( •_•)>⌐■-■

    • gbdub says:

      disproportionately trans, or disproportionately Jewish, or disproportionately butch girls, or universally people with mental-health conditions that turn them into assholes.

      These all strike me as really broad categories – yes, they are all relatively uncommon, but there’s still a huge pool of each type with a ton of variability within them. Are they distinct enough to really label as a “type” in terms of “I am much more physically attracted to people of this characteristic”.

      • ozymandias says:

        (Ashkenazi) Jewish people have characteristic physical traits (dark curly hair, large noses); so do trans people (trans women tend to be taller, have stronger jawlines, have smaller breasts, and have penises). ‘Butch woman’ is primarily an appearance signifier. I admit that mental-health conditions that turn people into assholes do not typically affect one’s physical appearance, although given how good that particular ex-girlfriend was at identifying us I do wonder.

        • 2stupid4SSC says:


          mental-health conditions that turn people into assholes

          could select pretty strongly for people who have children young. People who have children young might as a group be more attractive, or particularly attractive to other teenagers. Being particularly attractive as a teen to other teens could be a distinctive look.

    • Nadja says:

      I’ve always had this (completely made up and unsupported) idea that women have types for evolutionary psychology reasons. If your long term partner looks a certain way, in case you want to cheat, you’d better cheat with a person who looks more or less the same as your partner, or else the looks of your offspring might give you away. (Babies might then not be accepted by the father and less likely to survive.) So having a type might be an adaptation. =)

    • vV_Vv says:

      I’ve usually found that their partners are disproportionately trans, or disproportionately Jewish, or disproportionately butch girls, or universally people with mental-health conditions that turn them into assholes.

      These things may be are correlated with being poly.

  12. James Babcock says:

    I think in this case, it’s very helpful to apply a prior based on knowing what sort of mechanisms evolution is good at building. My engineering-intuition is that, in the language of RNA interactions and proteins, there are mechanisms that can be easily tweaked to cause imprinting, but detecting a race-specific visual feature and linking it to emotions would require a molecular Rube Goldberg machine and would be very unlikely to evolve incrementally.

    On the other hand, race-specific *odors* are something that evolution *does* have access to, because those are already in evolution’s language without having to detour through the visual cortex. So I would expect some heritability there.

  13. Will says:

    You’re absolutely right that the p-value for matching fathers to daughters is fishy (in the extreme case, if there was only one father-daughter pair in the study, which 1 million people all correctly classified, it wouldn’t prove anything except that that one woman married someone who looked a lot like her father).

    I think the right test would be a type of permutation test: compute some test statistic T(X) where X is the full data set (for example, the fraction of subjects choosing the correct husband i.e. 37% in this case). Then create a new fake data set X* where, for each woman, you randomly reassign which of the four men is labeled as the “true husband,” and then recompute the test statistic T(X*) on the fake data. Do this independently many times, then the p-value is the fraction of times that T(X*) is bigger than T(X).

    In my extreme example above, the test statistic would be 100% but the p-value would only be 0.25, because 25% of the time, the man who obviously looks like the husband is randomly assigned to be the true husband.

    (Note the validity of this proposal also depends on the authors’ not doing something silly in picking the decoy men; for example, it would be bad if they always used the same three decoy guys who really don’t look like husband material. Best case is that for each father-daughter-husband group they chose 3 other random husbands from the same data set; then you could randomize by randomly permuting the women).

    • Rin says:

      The permutation test might work, but I think there’s another question here. Just by chance, some proportion of women will marry someone who looks like their father, but if we really want to get at that, we have to figure out approximately what that chance level actually is. which may not be immediately obvious. We could consider an over-simplified case where there are only four clusters of male appearances in the world. In that case, there’d be a 1/4 chance that any individual married woman would have a husband who looks like her father. Using a binomial distribution, we can show that there’d be ~9% probability that more than 9 of the 26 sampled women would have a husband who looks like her father, so p = 0.09, meh. But if we assume that men have slightly more diversity in appearance – let’s say 10 possible appearance clusters – the chance that more than 9 of the women pick their father’s lookalike drops to a .01% (p=.0001). The authors may have assumed that the diversity of male appearance in the world is high enough that there was a negligible chance for two randomly chosen men to look alike, in which case their choice of statistical tests makes sense.

      • Douglas Knight says:

        I don’t think those assumptions justify their test. The experiment can’t detect if the husband is in the same micro-cluster as the father, only if the husband is closer to the father than the decoys. The 4 candidate husbands divide space into 4 clusters and we’re just asking which cluster the father belongs to.

    • Scott Alexander says:

      My cheap-and-stupid way of doing it would just be separate t-tests for the judges and the original women. If they were both highly significant, I’d be pretty happy.

      I’m surprised that Rantala guy who found the much more complicated and well-hidden error in the 2009 paper didn’t catch this one. Then again, I didn’t catch this in the rough draft of this post either even though I spent a really long time looking for flaws in the paper. I’m getting really upset by how many different ways there are to fail and how bad I am at catching most of them most of the time.

      • Will says:

        Yeah good point, your t-test idea would probably be approximately equivalent if there were a lot of women (I’m assuming you mean each woman gets a number, the % of people classifying her husband right, and you test whether the numbers are bigger than 25% on average). CLT kicks in pretty fast, so 26 is probably big enough.

        I’m a stats prof and I’m sure I would have read right past this in the paper, never realizing it was fishy (well, unless I were ideologically pre-committed to disbelieving the result, then I might have caught it). There is no substitute for explicitly thinking all the way through “what is the model” and “what is the null hypothesis” and “what does that logically have to do with my scientific question.” Most people just skip to whatever test first occurs to them and never explicitly think through the logical argument implemented by that test. But that kind of implicit reasoning puts a very heavy burden on your intuition.

        BTW, another possible concern for many of these studies is how much of this is driven by ethnic clustering? White women in Minnesota are more likely to have Scandinavian-looking husbands and fathers because that’s where all the Swedes live, and white women in NYC are more likely to have Mediterranean-looking husbands and fathers etc.

    • tgb says:

      The way I’d do it, if the study gave the full data, would be to take the data for each women separately first. Compute for each of those which was the most popularly chosen possible husband. Then count the number of women for which that one was the actual husband. Under the null-hypothesis, this should be a randomly chosen possibility, and so you’d need 11 or more correct to get p < 0.05, and 13 or more for p <0.01. (This would be a slightly pessimistic count as this assumes that the most-commonly-chosen possible husband is always correctly chosen as the "most similar looking" one.)

  14. suntzuanime says:

    I looked it up, and Romulus seems to have ended up with a human wife, rather than a wolf:

    • Walter says:

      Of course HUMAN sources would say that. You need to check Wolfipedia as well to get both sides of the controversy.

      • Deiseach says:

        Or you could go with the euhemerising version of the myth, whereby the twins’ “wolf” mother was really a human woman who was a prostitute – see Livy, “History of Rome”:

        The tradition is, that when the water, subsiding, had left the floating trough, in which the children had been exposed, on dry ground, a thirsty she-wolf, coming from the neighbouring mountains, directed her course to the cries of the infants, and that she held down her dugs to them with so much gentleness, that the keeper of the king’s flock found her licking the boys with her tongue. It is said his name was Faustulus; and that they were carried by him to his homestead to be nursed by his wife Laurentia. Some are of opinion that she was called Lupa among the shepherds, from her being a common prostitute, and that this gave rise to the surprising story.

        • Furslid says:

          This makes a lot of sense to me.

          The most striking thing about this interpretation is that the language play works in English as well.

          “Romulus and Remus were found and raised by some bitch.” becomes “Romulus and Remus were raised by canines.”

  15. TheWorst says:

    Even if only 5% of the opposite sex is interested in you, that still leaves the average person hundreds of different choices over their lifetime.

    Does it? Most people aren’t polyamorous, and most people don’t meet a very large sample of humanity during their mate-selection years. My math might be wrong (someone correct me) but it’s this: Given that number, in order to have 200 different choices for who to marry, you’d need to have had the kind of interactions that might lead to marriage with four thousand different people. My experience suggests that this is true of nobody.

    I suspect the usual number of marriage options for an average person is much closer to 1 than to 100, which seems like it would explain what we’re seeing.

    • suntzuanime says:

      I’m not sure it’s as bad as that, because people often choose what relationships to pursue based on how attractive their partner is to them. So looking at the number of relationships that got serious enough to consider marriage is the wrong metric, it’s more like looking at the number of people you could have flirted with who would have been receptive and with whom you could have built a serious relationship if you were interested and put in the effort. Which is a lot harder to measure, but that’s how we have to consider “marriage options” if we don’t want to include attraction in the precondition.

      • TheWorst says:

        Yeah, I phrased that incorrectly at first – “relationships that could lead to marriage” was too narrow, but “interactions where marriage is a possibility” is what I was going for. The point is that nobody gets to line up hundreds of compatible partners and pick the one to whom they’re most physically attracted, and the studies cited seem to mostly assume that everyone does.

        Just finding hundreds of compatible partners seems like it would take many, many lifetimes, and getting all of them to wait while you compared them all to each other before picking the hottest one seems like an implausible achievement.

        The choices you make only reflect your real preferences when you have a lot of options and the opportunity to comparison-shop. I don’t think either of those circumstances apply to most human mating.

        I mean… the bathroom you used most recently is probably not a useful basis for determining what your ideal bathroom fixtures would be given an infinite budget, teleportation, and infinite time to evaluate bathrooms.

        • suntzuanime says:

          IIRC the way it works mathematically is you reject the first 1/3 or so of the candidates you meet in your lifetime and then you pick the first candidate you meet that’s better than any of the first 1/3. So no, not quite “line them all up and pick the best”, but still plenty of room for personal attraction to make a difference, which was the only point Scott was making.

          • Tenobrus says:

            1/e in fact. It’s called the secretary problem.

          • Good Burning Plastic says:

            That only applies if you only care about the probability to get the best possible candidate, and are indifferent between getting the second-best and getting the worst.

          • nimim. k.m. says:

            Not to mention that “number of candidates you will meet in your lifetime” is probably very difficult to estimate until it’s too late, and you probably want to maximize the amount of the lifetime spent with each candidate with whom you attempt a serious long-term relationship.

        • ozymandias says:

          But there are lots of situations where one encounters hundreds of people in a relatively short time (crowded malls, large lecture classes, etc.) and physical attractiveness is a very obvious trait. It’s not remotely difficult to follow the rule “I will only hit on you if you are above X percentile in attractiveness”, and many people do. While I’d agree about traits that only appear relatively late in the relationship (e.g. conflict resolution styles), physical attractiveness seems to be a trait where the average person does have lots of options and the ability to comparison-shop.

          • TheWorst says:

            Of the people your average person is attracted to, what percentage are willing to date them long enough to evaluate compatibility? I’d be surprised if it was high. There are a lot of different filters, and getting any given person through even one filter takes time.

            In the time between “too young to get married” and “old enough to stop looking if I hadn’t found anyone,” I met maybe 4 people I might have married. This is considered high by most people I’ve known well enough to have a conversation on the topic with.

            The odds that you’ll meet someone who is your physical ideal and is compatible with you and that it’ll happen at the point in time where you’re both mate-shopping (and before you marry someone else) seem extremely low. The studies seem to assume that the odds of that happening are closer to 100% than to 0. I don’t think that assumption’s likely to be true.

            My competing hypothesis is that there are many different metrics people use for evaluating partners, and that physical appearance is only one of them, and that people marry the first one they meet who scores acceptably high in all categories. And that therefore, the partner they marry is not necessarily a suitable proxy for the highest possible score in any particular category.

            Satisficing, basically, not optimization-with-infinite-resources.

          • ozymandias says:

            I feel like there’s a distinction between “is attractive to” and “is the physical ideal of.” Most people do not marry people who are their physical ideal. Most people do marry someone they find attractive; I would not even be surprised by the finding that most people marry someone whom they find more attractive than average. Given this, one would expect a correlation between the traits of a person’s spouse and the traits of their physical ideal, which is all you need for the studies to hold.

          • 2stupid4SSC says:


            That is an important distinction because the point of the research is just to find anything significant. So of the limited choices that a person has, are they able to find something that at least approximates their ideal enough such that we can measure what that ideal is, and maybe figure out where it comes from.

            But TheWorst is, I think, more directly attacking the idea that people have as many options as Scott implies in the post, and on that point I agree with TheWorst totally.

            I would even go so far as to say that the limits on the marriage selection are so significant, the data we gather is always going to end up looking like a bunch of noise. People just don’t have anything close to the amount of freedom of mate selection necessary to see a strong signal.

          • TheWorst says:

            I would even go so far as to say that the limits on the marriage selection are so significant, the data we gather is always going to end up looking like a bunch of noise. People just don’t have anything close to the amount of freedom of mate selection necessary to see a strong signal.

            Said it better than I did. If someone’s making minimum wage, you can’t get very good information about their favorite type of car by looking at how they actually get around, because that’s basically pure noise and almost no signal. And I think that’s the analogy closest to how it works for average people (your 4’s and 5’s and 6’s).

        • Howard Treesong says:

          >The point is that nobody gets to line up hundreds of compatible partners and pick the one to whom they’re most physically attracted, and the studies cited seem to mostly assume that everyone does.

          In the modern world of online dating, I think that happens much more often than you might think. Even in my case (I’m old and no better than average attractiveness), I ran across many many many potential mates in college, grad school, work, and various other social settings — I’d have a hell of time trying to do an actual count, but it wouldn’t surprise me one bit if the number were in the low four digits. I clearly and obviously put FAR more effort in to pursuing the ones I found physically attractive.

          In a highly statistically-relevant anecdote, I’ve never liked women that looked or acted anything like my mother.

          • TheWorst says:

            What percentage of those did you get to know well enough to determine if you were actually compatible? As far as I know, that takes years.

            Of those, how many of them were attractive enough to you that we can use any one of them as a stand-in for a picture of “Howard Treesong’s Ideal Mate?”*

            I think it’s a mistake to look at someone’s actual partner and assume that’s their Optimal Ideal Partner, is the point.

            For comparison, in my entire life I have met one person I can point to and say “That, there, is an avatar of everything I find physically attractive.”

            We didn’t get married. We met about two months before I got engaged to my wife; I later found out that she (Ms. Avatar, not my wife) was wildly incompatible with me. The studies seem to be based on the assumption that both of these last two circumstances are impossible.

          • nimim. k.m. says:

            Apparently there is a problem that we are arguing on the basis of anecdata, and individual experience of meeting “potential mates” seems to vary a lot.

            During my 5 years at university, the number of members of opposite sex I had a realistic chance to attempt at flirting with was around 10, and because I had not much success, I certainly had not an ability to choose.

    • nydwracu says:

      No you wouldn’t. The people you interact with aren’t a random sample of the population.

  16. Brad (The Other One) says:

    Have there been any studies with LGBT individuals? Or people raised by a single parent of the same gender?

    • Scott Alexander says:

      I think I remember that being raised by a single parent doesn’t change sexuality. I haven’t seen any of this kind of study with LGBT people, and I’m glad. I worry that it would be done in a building-upon-previous-work way, eg “Here’s how things work for heterosexuals, now let’s see if it works the same for gays”, whereas I feel like we haven’t even gotten a consistent answer for heterosexuals, and I would be really surprised if someone was so good at this that they could solve the homosexual case (with much less data) before the other researchers could even get the straights straight.

  17. MonkeyBoy says:

    My armchair hypothesis as an uninformed monkey on the internet is that genetically-determined attraction factors might not play well with evolution.

    If we consider monkey->human evolution, and assume that the monkeys had monkey-specific attraction genes, it seems like those genes would either have had to keep up to their changing appearance, or die out.

    The latter seems strictly more likely in the short term. In the long term, new/modified versions of genes could specialize on the new appearance, but this would take time and could only really happen after a trait had reached fixation.

    On the other hand, genes using social order or more universal traits (like body symmetry) for choosing mates seem like they would work pretty well almost all the time, which means that they could survive almost any change.

    So even if we do have genes coded to specific appearance, it seems to me like they would be much more recent, less universal, and secondary to genes using social order and more universal appearance traits for choosing mates.

    This might also explain why we tend to find other mammals “cute”, and why my human-raised dog is scared of every species of animal except humans (including other dogs!).

    As for why the imprinting effect seems to have such large error margins, beyond what others have already said, I’d propose that we don’t just imprint on our parents, but are also affected by everyone else we see and meet while growing up.

    Also, in case it wasn’t clear, I’m not very informed on this subject and this is all speculation. Please take this entire post with a grain of salt.

  18. maznak says:

    Well from my lifelong observation I tend to believe that people often, suspiciously often, marry people who sort of look like them. I know that it is just anecdotical and all kind of bias can play a role too – including confirmation bias. But still, I tend to believe that people more often than can be attributed to chance marry similar looking people. Sometimes it borders on being funny. I have always thought that it may be either slight kin preference, or some narcissistic effect. But imprinting would explain it too (and better): I may look like my mother and for imprinting reasons pick a spouse that looks like my mother and therefore looks sort of like me.

    • Scott Alexander says:

      The last study in (I) looked for that and found that women married people who looked like themselves, but the effect was less striking than the degree to which the person looked like their adoptive father.

  19. Anonymous says:

    This is indeed what they found: when a person with a pregnancy or lactation fetish only had one sibling, there was a 66% chance (compared to expected 50%) that their sibling would be younger than they, p < 0.0001 in their sample of 560 such people. This was true if and only if they were between age 1.5 and 5 during their sibling's birth

    If they were between age 1.5 and 5 during their sibling’s birth, shouldn’t there be a 100% chance that their sibling is younger than them? Should this say something more like “Iff their age difference was between 1.5 and 5 years?”

    • Scott Alexander says:

      I think first they checked the direction of the age difference, then they took the older siblings and established the window.

  20. Rm says:

    So we have studies on humans and non-humans; why are we lumping them together?

    • suntzuanime says:

      Why not? We lump together studies on students at UC Berkeley and students at UCLA.

    • 2stupid4SSC says:

      The impression I got was that they were not really lumped together? The animal studies seem to provide much stronger evidence that, if taken at face value would make a much stronger case than he ends up making. Rather I think they are presented more as an interesting and related data point, for your consideration.

  21. Svejk says:

    These findings are consistent with mate choice having multiple optima along the axis “similarity to self/parent” , both within and between individuals. In that case, availability/chance factors/cost of the choice (which is partially influenced by availability) would influence selection, and could make it seem fairly random. Perhaps some men would be happy to marry a girl just like the girl that married dear old dad, but their local optimum is a more ‘exotic’ woman with higher MHC complementarity.

  22. Alex says:

    No consideration of things like race? If I’m Jewish, my parents will probably make it clear that they prefer that I marry a nice Jewish boy, and I will also have lots more opportunities to socialize with Jews than a non-Jewish person would. I’d be much more likely than average to marry a Jewish boy, and a Jewish boy is more likely to look like my father than a random boy does. Or this whole social priming to like Jewish boys might even mean that I end up liking people who *look like* Jewish boys, even if they aren’t Jewish. None of this would mean that my preferences came from imprinting on my father’s appearance.

    For the fetish stuff — one possible consideration is that exposure to the behavior makes you more likely to develop the fetish just from mere exposure. Someone who has never or rarely seen a woman breastfeed, for example, might never develop a breastfeeding fetish simply because they’ve had no exposure to it, in the same way that a person who lives in a culture where everybody is barefoot all the time probably won’t have a fetish around high heels. It would be interesting to see if, for example, it’s actually the family relationship with the breastfeeding woman that makes a difference, or the amount of breastfeeding that they observe.

    • arbitrary_greay says:

      Yeah, my joking go-to was to wonder about 2nd generation immigrant asian girls dating non-asian guys. If there’s a factor of liking people who look like their fathers, it seems to be overridden by the attraction norms of their peer group.

    • Mary says:

      Heck, people will actively throw you together.

      My sister first attended a Jewish wedding as the maid of honor. This was because one day a friend of hers was grousing about not being able to find a nice Jewish boy, and my sister looked around, thought, Hey, Matt’s Jewish, and said, “Hey, Matt come over here and meet Sarah.”

  23. Callum G says:

    EDIT: never mind, question was asked further up.

    I wonder if studying homosexual couples would reveal anything. Seeing as the genetic theory does seem to be cross-gender attraction, a gay man inheriting his fathers taste for women wouldn’t say a lot about his choice in husband. Would gay men imprint on their fathers? Or would they take the schema of the mother and apply it to a man?

    • Anonymous says:

      More anecdote: I and my brother (both white men) are estranged due to LGBT-&-religion problems and live thousands of miles apart, but both in highly homogeneous white populations. But it turns out all the people we’ve dated (women for him, men for me) have been latinas/os and east Asians with very nearly the same skin tone. So I’m inclined to expect common genetic or environmental influence, and not us having imprinted on opposite parents. Similarly, my husband’s sister’s husband is a white man who looks very similar to me, which may be more interesting since my husband and his sister were both born-and-raised-to-adulthood in Seoul.

  24. Ryan W. says:

    Hm. If the notion is that some patterns are obscured by scarcity, I wonder if those trends are stronger among peple with higher socioeconomic status or those who are rated as more attractive, since they’d be more desirable and have more options.

    In terms of genetics, I’d look at OR7D4 receptor alleles (olfactory steroid receptors) and tendencies like novelty seeking or a tolerance for physical aggression.

  25. AoxyMouseOnArgo says:

    Meh. Want to know what’s interesting from a Psychoanalytical perspective?

    Go on the top porn sites and erotica sites, and sort by “Most Viewed”

    What stands out to you?

    Now *that* is what surprises me that there isn’t more papers on. Too bold, maybe.

    • Captcha says:

      I think that most people are completely fine with vanilla porn, but if all the best pornstars/production teams are working on something fetish then most people will settle for it anyway. In this way doing a vanilla porn scene only reaches vanilla people, but, if the fetish isn’t too abrasive, then a fetish scene may reach vanilla people + fetish people.

      It is a tad disquieting that family relations seems to be the most demanded fetish that also catches vanilla users.. I guess for people with step families it would be a legitimate source of sexual confusion. Although, in saying that, sometimes I feel like the word ‘step’ is just shoehorned into a scene to meet legal requirements. It would be quite interesting to talk to someone knowledgeable about this.

      • ozymandias says:

        Oh, was that what we were supposed to get out of it? I thought that they were darkly hinting about the popularity of teen porn.

        • Samantha says:

          I thought they were darkly hinting about the popularity of white women in porn.

          • suntzuanime says:

            Maybe we should all try to hint a little lighter so we can see what we’re hinting at.

          • Sandy says:

            Really? I figured they were rather pointedly hinting at the popularity of “stepmom”, “stepsister” and “stepdad” porn.

          • AoxyMouseOnArgo says:

            Naked people, actually.

            Why are they so common in porn?

          • Captcha says:

            You’re right. There is no clothed porn on the most viewed page of pornhub. Case closed.

          • Autolykos says:

            Yup. That is incredibly weird, considering that most people I meet in my day-to-day life are not naked. It seems the imprinting theory has quite a coincidence to explain here…

          • Scott Alexander says:

            Your mother was naked when she gave birth to you, so maybe the critical imprinting window is between one to three minutes after birth.

            (I actually don’t find nakedness attractive, but I know I’m in the minority here)

          • nydwracu says:

            Maybe we should all try to hint a little lighter so we can see what we’re hinting at.

            Incest porn. It’s about incest porn.

            For some reason, a lot of narrative-driven porn is incest porn. Porn video sites are pushing narrative-driven porn to attract women.

  26. MugaSofer says:

    I really expected this to be about heart disease.

  27. Eric says:

    Speaking of mate choice, I have been collecting data about what personality survey items show similarity between pairs of best friends (and thus presumably indicate a causal effect to relationship formation):

    Its mostly what you expect, two big trends are religious/political views and straightedge lifestyle.

  28. Coach McGuirk says yes, evidence being FFS (Fat Father Syndrome).

    • Skef says:

      I can’t watch anything else with voice work by Jon Benjamin, including Archer, because all I hear is Coach McGuirk.

  29. Douglas Knight says:

    The animal models tell us nothing about humans.

    The animals are imprinting on a different species, while the humans are putatively imprinting on intraspecific variation. You might think there are mechanisms to imprint on gross variation, then there should be triggered by fine variation. But, no. The imprinting mechanisms that are fooled by other species are kludges to imprint on the right species. Their environment of adaptation is only of the same species. Interspecific adoption is usage out of specification. They may well be designed to be robust to intraspecific variation.

    • Scott Alexander says:

      Not sure I agree. Zebra finches and Bengal finches potentially look pretty similar; same with the differently colored zebrafish. Evolution doesn’t understand the concept of “species”, so it’s got to be something like “mate with someone who looks like this”. Given that it can differentiate a zebrafinch from a Bengal finch, it’s got to be at least sort of fine-grained. My guess is that a Swede and a Nigerian are at least as different-looking as a zebra finch and Bengal finch, although of course most people’s mate decisions aren’t that dramatic. But things like height, BMI, hair color, etc also seem like the sort of thing that could get caught in the mix.

      I’m not sure how you say that the animal models tell us *nothing* about humans. I agree you can tell a story about why they might not transfer, but surely knowing that animals imprint on the opposite-sex parent in general makes imprinting on the opposite sex parent for details more likely, especially in the presence of other evidence for this.

      • Douglas Knight says:

        Let me add a refinement. If the purpose of imprinting is to make sure the animals mate with the right species then it (1) needs to distinguish from species nearby and (2) needs to be robust to variation in its own species, but only in the local population. So if bengal finches share a range with zebra finches, they should imprint on details that distinguish them. Humans might imprint on details that happen to vary between races, since there is no selection to be robust to them.

        Anyhow, if you want to know whether animals imprint on intraspecific variation, it seems a lot easier to do zebra finch adoption studies than human experiments.

        knowing that animals imprint on the opposite-sex parent

        Why are you emphasizing the opposite-sex part? Do you actually care about that detail? In any event, I don’t think that detail is tested in the animal models. Didn’t both the males and females imprint on Lorenz?

        • Steve Sailer says:

          Back in the 1980s-1990s, when Edward O. Wilson was promoting the importance of saving the Amazon rainforests to preserve biodiversity, I thought a lot about the the importance of species purity, which is, by the way, a big legal issue regarding the Endangered Species Act.

          The impression I picked up from Wilson is that preserving species boundaries is very important in small-brained species such as Amazonian insects who have their behavioral patterns hard-coded. Wilson pointed out that beetles species, for which the Creator has an inordinate fondness, have extremely specific ecological niches, such as the vertical versus the horizontal surfaces of the same tree.

          Larger-brained species, which have to survive both summer and winter in temperate zones, tend to have more adaptability, so temperate species are fewer and more widespread. They have to learn to be able to deal with different seasons, so they can also learn to deal with different locations.

          Humans don’t have to worry much about making sure to mate within the species since there isn’t much else like us and we are the greatest geographically imperialist species on Earth.

      • Steve Sailer says:

        As others have mentioned, the advantages of avoiding incest might work against imprinting romantically on a parent since that would tend to lead to attraction toward a sibling.

  30. Maryana says:

    Hi Scott,

    First-time caller, long-time listener.

    and on one hand the two findings’ 95% confidence intervals do overlap

    Not a big deal, but like people in love, confidence intervals interact in odd ways. A difference may still be statistically significant at the X% confidence level even if the X% confidence intervals overlap. See, e.g., this.

    • Autolykos says:

      Yup, that is way too conservative. It’s usually enough if the average of A is outside the 95% confidence interval of B and vice versa.
      Or, as a rough estimate, if the standard errors do not overlap (assuming they are about equal).

  31. reytes says:

    Mate preference seems (anecdotally/facially) like something complex which could easily be the result of a combination of several different factors (for instance it seems fairly reasonable that sexual and romantic desire are not necessarily identical things)! It seems plausible that different elements of attraction are affected in different ways by different factors!!! For instance – hypothetically – the qualities that someone finds sexually attractive could be a result of imprinting while the qualities that are emotionally satisfying might be a result of genetic coding!!!!!!!

    Again this is just off the top of my head, and probably incoherent due to external factors (viz whiskey)! But when you have something that our existing analysis is unable to explain, it seems at least worth considering that we are glossing over underlying distinctions and complexities.

  32. Steve Sailer says:

    “The authors’ guess: true love, a mysterious and magical thing totally beyond scientific understanding.”

    Yes, that’s been my vague hunch from the social science literature. The usual nature or nurture factors aren’t very useful at explaining romantic chemistry.

    For example, I talked to Nancy Segal, the twin expert at Cal State Fullerton, about this. She says that identical twins tend to approve of their twin’s choice of a spouse, but generally don’t find their twin’s spouse romantically attractive. I’m sure there must be lurid exceptions to this pattern, but in general our culture doesn’t have a stereotype of identical twins pining away for the other’s beloved.

    • Steve Sailer says:

      Identical twins aren’t even that concordant on general sexual orientation. When J. Michael Bailey redid his identical twin sexual orientation study on the Australian twin registry (which gets around selection effects), he found that a homosexual identical twin would have a heterosexual twin about 78% of the time.

      For example, of the playwright Shaffer twins, Peter (“Amadeus”) was gay, while Anthony “Sleuth” was straight.

      By the way, somebody should write a play about the Shaffer twins.

      • Steve Sailer says:

        It would be easier to make a story about the Shaffer twins into a movie, like Charlie Kaufman’s “Adaptation,” in which Nicholas Cage expertly plays identical twins with inexplicably opposite personalities. But the story of the Shaffers deserves to be a play more than a movie.

        There aren’t, as far as I know, any famous movie star pairs of identical twins. I found one list of 10 movie stars who have a twin:

        9 have a fraternal twin. The only identical twin is Jon Heder of “Napoleon Dynamite.” (It’s been theorized that identical twins aren’t as ambitious as the average person. How to test that I don’t know.)

        However, there are a lot of former child star identical twins. The entertainment industry hires a lot of twins as babies or children to get around California laws about hours worked for minors. The Olsen Twins are a famous example. And I’ve met a pair of identical twins who had their own TV show when they were tweens.

        So there are more than a few identical twins in Southern California and New York with professional acting experience, who might be able to take on the roles of the Shaffer Twins in a stage production.

        • Steve Sailer says:

          Identical twins are somewhat more common in rock bands than in movies: e.g., The Proclaimers, Good Charlotte, The National, and Bros.

          “500 Miles” makes me happy:

        • IamtheTarpitz says:

          Linda Hamilton also is/has an identical twin, which was helpful in filming Terminator 2.

          Sadly, a worthwhile play about the Shaffers would require not merely twins who were professional actors but twins who were good actors and a reasonable fit for the parts, a much rarer breed. It might even require multiple such pairs, depending on the writer’s approach.

      • Squirrel of Doom says:

        Wow. That seems to invalidate the much popular idea that sexual orientation is in your DNA.

        78% is a lot less than the~95% you’d expect if it was fully independent of DNA, but it still can’t be the major factor.

        Or am I missing something?

        • Ricardo Cruz says:

          I think the popular idea is that sexual orientation is biological. Biological does not imply genetic. Maybe some people think it is coded in the DNA, but I would think most would say something like it is related to some random process in the environment of the uterus. But nobody knows. What people do know is that gays feel their orientation from an early age, and that they cannot change it, so it is some biological factor.

        • Steve Sailer says:

          Nah, it’s a pretty striking result.

          It would be good for somebody to try to replicate it. You might be able to do a quick and dirty look at celebrity identical twins where at least one is gay, but there aren’t that many celebrity identical twins.

  33. Lucky-O Mishima says:

    I wouldn’t a priori expect genetics to make sense of this. Imprinting is a bit simplistic, too. What about people who are sexually attracted to cars, balloons, cartoon ponies, or the idea of being transformed into a plastic doll by a witch who is also their mom?

  34. Revealed Preferences says:

    Here they’re able to get some more immediately visual data – preferences like tall/short, long-hair/short-hair, beards/clean-shaven, even big-breasts/small-breasts

    Okay, this is an aside, but why is the last of those considerations phrased as if it were trivial relative to the others? For me, it’s probably the single most important variable in determining a woman’s visual attractiveness, certainly dominating height or hair length* (at least over normal ranges of each) … and I don’t expect I’m that unusual?

    *I’ve never actually been presented with a direct choice between a bearded, large-breasted woman and a smooth-faced, small-breasted one and all else being equal I think I’d probably be more attracted to the former, but I honestly couldn’t say with full confidence)

    • Mary says:

      No, the “even” indicates that you can bluntly ask about a question that would have to be more delicately phrased, even if feasible, in most situations.

  35. Anonymous says:

    A couple of ideas for making sense of this:

    1) Relative ease of evolving a genetic “what my species looks like” thing vs evolving a learning mechanism to get in imprint off the almost inevitable[1] environment. Learning mechanisms seem to be ubiquitous and it seems easier to adapt one to the task than to figure out how to encode an image in kinases and transcription factors. In short “good-enough and easy to get to” beats “theoretically optimal but relatively inaccessible” every time.

    2) Evolvability. If your species does mate selection via imprinting rather than a genetically-encoded ideal-mate image, then it doesn’t have to keep the ideal-mate code up to date. This means that a) the species can evolve in appearance and b) all of the visual hardware and image-processing stuff in the brain can evolve too, without having a brake on the whole thing.

    [1] Unless some pesky biologist is on your case, in which case all bets are off, or unless your parents die, in which case you’re probably dead anyway.

  36. Agronomous says:

    Am I the only one who was sure this post was going to end with a zebra pun?

  37. IamtheTarpitz says:

    Anecdata: I sometimes have a beard, and sometimes don’t. I have noticed a distinct correlation between women who think I look better with a beard and women whose fathers have beards.

  38. HeelBearCub says:

    One of the things that is bugging me in these conversations is the assumption that the model is a “shopping” model, where the shopper has all the control in selection and the chosen mate has none.

    What if one reason one ends up with mates that look like your opposite sex parent is that they have raised you to be attractive to people like them?

    Alright, I admit it seems unlikely that brunette women can raise sons who are more attractive to other brunette women, but I do think there is something missing when we don’t account for the fact that this is a two way selection process.

    • onyomi says:

      I think it can make a difference. My mother, for example, bought most of my clothes until I graduated high school and still gives me clothes for Christmas. She has strong opinions about my haircut and facial hair, is not shy about telling me I’m looking a little fat, etc. etc. In other words, consciously or unconsciously, she may be making me more attractive to women who share her taste. This might even make a certain amount of sense evolutionarily: a way of helping people get along with the in-laws.

      • HeelBearCub says:

        Right, but do those things correlate with her external features?

        I guess, if we think about an extreme case like Hassidic Jew, sure. Maybe that means that we should expect that we can pick up on external features that are socially mediated, and therefore the the parent can influence the child to be more attractive to people who like those socially mediated external features.

        • nimim. k.m. says:

          In very broad sense of external features, it might, but then it might make more sense to lump that particular explanation with other environmental factors (such as culture or class or religion).

          For example, if your aesthetic sense of fashion is influenced by your mother, is that you mother in particular, or just the wider social group you grew up in.

    • nimim. k.m. says:

      I’m too tired to pay any serious amount of mental effort to think this through, but should one take into account that the even the simple effect *would* be two-way when it comes to e.g. ethnic groups.

      Consider that Swedes are statistically different looking than Italians. (There are more natural blondes in Sweden, even though not so as many as the stereotype goes, but let’s assume for the sake of the argument…) Suppose a stereotypical person has preference towards people who look like their opposite-sex-parent. Now let us assume a heterosexual female Swede who has equal amount of opportunities to date with Swedish and Italian males: even if this particular specimen of the Swedish did not have any preferences at all, assuming the effect exists amongst the other candidates would mean that she’s more likely to have success with the Swedish candidates (who prefer partners who look like her) than Italians (who would prefer females who look their Italian mothers).

      Of course the example scenario difficult to play out in real life, as Swedes live in Sweden and Italians in Italy, and even with EU and freedom of movement, there is a question of availability. And it goes very much against the “hot foreigners” trope. [1] However, you can interchange the nationalities for any ethnic group.

      It also does not work for all kinks and features one might be attracted to.

      [1] However, the Swedish culture is sufficiently different from the Italian so that it probably has some effect on the probability of a successful long-term relationship.

      • Autolykos says:

        A thing to note about the “hot foreigners” trope is that foreigners don’t tend to age very well (subjectively, compared to people from one’s own tribe). Which is probably working as intended. Spreading your genes to different tribes can be a good idea, but actually living with them may be a lot less beneficial.

        • Creutzer says:

          A thing to note about the “hot foreigners” trope is that foreigners don’t tend to age very well

          I assume you mean that in a somewhat metaphorical sense. Would you elaborate? What do we know about the relative stability of inter- and intracultural relationships?

          • Wander says:

            I think they mean that different ethnic groups (or even just groups in different environments) tend to change in different ways as they age, like the severity of shrinking in old age in people from East Asian countries, or the leathery skin of elderly white people in relatively equatorial regions.

  39. Desertopa says:

    Human sexuality seems like something that has to be at least partly environmentally determined. Twin studies indicate that even identical genes won’t determine with anything near certainty that two people will both have the same target gender of sexual attraction.

    There isn’t enough information on the genome to code for all our complex behaviors.Our genes provide a template, but that template has to develop in response to other inputs. A lot of it might not be inputs we normally think of as social, but the extra information has to come from somewhere.

    I’d recommend checking out the book “Perv: the Sexual Deviant in All of Us” for some more research along these lines. Here’s one interesting thing I picked up from it; in a study where sheep were raised alone among goats and vice versa, male goats preferred to mate with female sheep when they grew up, and vice versa for male goats, but female goats would mate with either male goats or male sheep, and so would female sheep. In humans as well, both men and women have paraphilias which involve being aroused by something that’s not normally considered sexually titillating, but people who are only able to be aroused by this unusual thing, and not also by normal sex, are almost all men.

  40. gwern says:

    I can’t find the study itself, but multiple reviews cite Jedlicka 1984, who looked at children of mixed-race couples (white and native Hawaiian).

    Isn’t it right there in Google Scholar? Searching “Jedlicka 1984” takes me to as the second hit, so both a fulltext and a domain where you know you can get it trivially through Libgen or requests.

  41. nimim. k.m. says:

    >Even if only 5% of the opposite sex is interested in you, that still leaves the average person hundreds of different choices over their lifetime

    Erm. Urm. I can’t even.

    5% sounds like terribly optimistic. By an order of magnitude. Unless you are a very good-looking celebrity known in large parts of the globe.

    • Roxolan says:

      More charity is useful.

      I assume he meant something like “5% of the opposite sex, if they were single/open and met you under conditions where flirting was welcome, would not be opposed to the idea”. That seems conservative enough to me, for an average person.

      • Autolykos says:

        Even then, you’d probably need to tack on a few additional conditions, like “same social circle”, “similar age” and “compatible interests” to get anywhere close to a 5% figure. I wouldn’t even consider dating 99% of the opposite-sex people I encounter on the street if they were single/open, and that sample is already somewhat preselected by class and culture (and I didn’t have a chance to sort them out based on anything they’ve said).
        And experience has not given me reason to assume it’s any different from the other side.

        • HeelBearCub says:

          I wouldn’t even consider dating 99% of the opposite-sex people I encounter on the street if they were single/open

          That would appear to be typical-mind fallacy?

          Unless what you are trying to say is “too old or too young”, but I don’t think that’s actually terribly relevant.

          The relevant age cohort to do the calculation is probably two years up or down from you given prime mating age, which is probably 16 to 22? Spitballing as an idea I mean.

          Within that, I think there is a fair amount of evidence that most people have attraction to a wide range of the cohort.

          • Autolykos says:

            Similar age is of course relevant and would get you some of the way, but even if you take that as a given, that would be nowhere near enough to get to 5%. OTOH, perhaps I am just weird.

  42. dtsund says:

    > Even if only 5% of the opposite sex is interested in you, that still leaves the average person hundreds of different choices over their lifetime.

    Sure, that’s true today. It wasn’t true during the EEA, and that’s what shaped our behavior.

    • Mary says:


      Having a type can be evolutionarily very unsound behavior. Unless your type is a real striking evolutionary advantage, and sometimes even then, it’s going to hinder passing on the genes.

  43. Jill says:

    “do people choose mates who look like their opposite-sex parent?”

    It is interesting to me, the questions that researchers ask. Of all the questions that would affect people’s lives the most, what their mate looks likeseems pretty low on the scale.

    What I’d like to see is lots of research on do people choose mates who have similar neuroses to one or both parents e.g. addicts. Just in clients I’ve worked with, and in acquaintances and friends I’ve had, this does seem to be so a high percentage of the time.

    I think the research result that people are little influenced by their parents’ behavior toward them, is only obtained by asking the wrong questions.

    If you look for environmental influences only of certain types, or in certain areas, then presto! You’ve got it. Your “proof” that everything must be genetic and so there must be no environmental influences. But the full specrtrum of the underlying reality may be different.

    Researchers need to have wider experience with other human beings and their issues than they do. Maybe researchers in psychology or psychiatry should be required to first spend a certain number of years as a psychotherapist, so that they are familiar with what problems humans have, and can better think of what research questions might be useful to ask about them, with the goal being to discover information that would decrease human misery.

  44. Squirrel of Doom says:

    Could there be evolutionary pressure to have kids “imprint” on whoever raised them, as a way to incentivize adoption?

    I mean, from a selfish gene perspective, raising someone else’s kid is (hyperbolically) genetic suicide. But if it results in the kid procreating with people genetically similar to you, it might come out as a win-win anyway.

  45. k3w says:

    I could see narcissism as a driving force.
    To paraphrase Alone of thelastpsychiatrist, maybe partner choice is guided by the question “What kind of woman would someone like me be married to.” In a ‘western’ society where work is a defining part of persons life (How often is ‘what do you do for a living’ among the first questions at social gatherings?) that would mean that e.g. that lawyers wives, doctors wife (maybe in different disciplines) all are alike (in terms of looks, character, tone of voice etc.

    This would also allow to explain the religious/class preferences, at least to a degree. Separation could be read as a person a) not fulfilling the spousal role they were cast in (sole provider, HPOA, smart, embodiment of niceness) or b) people wanting a change of image/identity and see ‘a lawyers spouse’ as a hindrance to reframing as a crime novelist.

    *Excuse me, if i’m hard to understand, I’m german and on a mobile device.

  46. Outis says:

    Is it common for preferences to change during life? For example, a childhood friend of mine says he used to find black women thoroughly unattractive, until he met one he liked in his 20s. Now he is much more attracted to black women, or women with dark skin in general. What happened there?

  47. vV_Vv says:

    And in fact, the authors remind us that people who have had multiple long-term relationships often choose people who don’t resemble each other in more than the most superficial ways (if the authors had exposure to polyamory, they might note that most people’s simultaneous partners aren’t very similar either).

    Or all these studies are just failing to identify the relevant characteristics for mate selection.

  48. Tykebomb says:

    Whats the reading list for learning how to read studies?

  49. TomA says:

    Perhaps the unpredictability of mate selection is advantageous to evolutionary success. In other words, the lack of causality is the mechanism that maximizes robustness in the species.

    • Roxolan says:

      That hypothesis sounds suspiciously like group selection. Which generates nice stories but doesn’t actually happen because genes are too selfish. The animal kingdom has plenty of sexual-selection downward spirals.

      • TomA says:

        Not necessarily. At the individual level, randomness in mate selection might be positively correlated with genetic robustness; otherwise we might spiral toward sameness of appearance, particularly among isolated populations.

      • AoxyMouseByArgo says:

        What? If my memory serves correct, group selection has been shown in bacterial models to be true in certain circumstances due to examples of it.

        Group selection gets too mangled up with morality, me thinks.

        Its clearly true in some bacteria studies. In mammals its obviously more complex, and with human emotions,children(closely linked to genetics) and learning its really just hard to separate anything that approaches clean group selection from the rest of everything(like being socially rewarded for ‘good’ behavior) I do think a lot of the debate ends up as semantics and battles of how words are defined, endlessly.

        In the end, its how the math works out. If there end up only minor average reproduction rewards for selfishness, with large benefits to “sharing”, altruism results, and things get all wacky with the human brain and its memes.

        This of course, reminds me to renew my subscription to scientific american and another popular magazine.

  50. Your criticism of the Bereczkei et. al. p value looks correct, but there should be a simple test. The probability that all 27 daughters happened by chance to choose husbands who resembled their father is very low. If what is going on is that an above random fraction of them did, you should be able to identify which and how many of the daughters those were by which husbands the subjects more than randomly identified. You can then calculate the p value for that number of daughters having, by chance, married husbands who looked like their fathers.

  51. The great thing about the human brain is that it does allow for huge variation beyond environmental exposure and genetics. The exploration of human sexual consciousness has been very limited by mechanical approaches and limited imagination. As recently as 5 years ago I attended a conference as asked the lecturer (on intimacy, sex, and marriage) who was doing work in the area currently. I was referred to a psychoanalyst who I had corresponded with many years earlier and who was then deceased. Degrees of freedom at the anthropological and neurobiological is most definitely the issue at hand and the answer is as complex as the fact that if there are 7 billion people on Earth, there are 7 billion unique conscious states. People can certainly be ranked in terms of attractiveness and there are some scientific findings on that subject. But unique conscious states are always generated by the processing power of the brain interacting with the environment.

    The possibilities in this situation are large and they are much larger in modern times. It is safe it say that at no point in human history have prospective partners known more about one another than they do today. At no point in human history has the decision making been more elaborate. In my lifetime, men and women were often instructed on who they should date and marry. Today the advice is less explicit but it still goes on to some extent. There are an entire set of decisions based on that advice (implicit or explicit) that is not considered in this research. I think that a research program closer to human consciousness that examines sexual fantasy and daydreaming may be a more productive approach to learn more about the process than just focusing on demographic characteristics – but that is undoubtedly my bias. That is also difficult information to get at and collect over time.

    I think that mate preferences matter only to the beholder. There are always bystanders shocked by some of the choices ready to utter: “Well I guess there IS somebody for everybody.” Human choice behavior for mates is complex and it should be. It flows from a very rich conscious state.

  52. Jesse says:

    Longtime reader, rare commenter here.

    Others have briefly mentioned looking at this issue in terms of homosexual attraction, but I see a bigger possibility here – homosexual couples with heterosexual children. I’d like to see a study that gathers data from, say, couples of two females who have heterosexual male children. If one of the pair is both the genetic mother and the primary caregiver during ages 1-5, this might provide useful data. Are the heterosexual male children more likely to marry someone who looks like their genetic mother/early primary caregiver, which would be evidence for imprinting, or to marry someone who looks like their genetic mother’s wife/partner, which would be evidence for genetics playing a larger role. The same would work for daughters of homosexual male couples, or children of any couple who used a sperm or egg donor to conceive, I suppose.

    Political events and social changes throughout many countries in the last 20 years should be providing us with a fairly large potential sample size either now or in the next few years.

  53. Heinar says:

    Did the papers on zebra fish & zebra finches take into account the possibility of more general preferences in appearance?
    I’d argue that something like a’ preference for individuals who look and/or behave more similar to the individuals the animal had most interaction with while growing up’ would need to be counterchecked to distinguish the imprinting of mate-choosing preferences and the imprinting of general contact-choosing preferences.

    I’ll try to find the time to read them after finishing work, but maybe someone can enlighten me ? 🙂

  54. Derannimer says:

    Apologies for the below, but I have OCD, and incest is my most recurring doubt/fear, so:

    Ew ew ew, does this mean Freud was right and I’m subconsciously attracted to my father? In that case I must never have sex with anyone, because I might symbolically be having sex with my father, which is revolting.

    • Derannimer says:

      Also, this is one reason I have trouble with the concept of trigger warnings. *I* sure the hell never get one. >.>

    • The Nybbler says:

      I think if there’s any effect here, it’s going on at a different level than what Freud would call the “subconscious”. And it wouldn’t be about being attracted to your father; it would be attracted to people who look vaguely like your father did when you were very young — that is, it’s about keeping it in the tribe, not the family. And in any case it’s a small effect, and if you’re that worried about it you could deliberately avoid romantic relationships with people who look like your father. Or talk about it with a professional (preferably one who also does not look like your father)

      • There is more to Freudian theory than what your father looks like. The physical and interpersonal dimensions of your father, how he treated you, how you reacted to that treatment, and how you resolved any central conflicts in that treatment. The central conflict is resolved by identification with the same sex parent. There is certainly plenty of research to point out that adult neurosis can be treated without psychoanalysis and that child development does not proceed according to Freud.

    • TAP says:

      You should treat this possibility as an irrational symptom of OCD. Something to be rejected as a valid concern. Not as additional fuel for the OCD fire. Do you have a process for dealing with crazy OCD thoughts? You should use it here. (And if you don’t have one you should get one.)

  55. baconbacon says:

    How do the animal studies differentiate between physical and behavioral preferences? Social animals pick up on lots of small cues and isn’t it plausible that two slightly different bird species would have slightly different behaviors? My father in law and I are not physically similar but we are both libertarian/anarchists who like to argue and also randomly drift off to spend time alone. That has to be a small pool of potential mates.

    It also seems like you could determine people’s (males) physical preferences through porn. Porn has incredibly specific categories now, and there isn’t the trade off problem that you get with spouses.

  56. Matthieu says:

    You’re right about the pvalue. Standard p values are incorrect because guesses are correlated for each woman. To compute correct standard errors, you should first average all the guesses for each woman, and then consider each average as one observation. Alternatively/Formally, you should use standard errors clustered at the woman level.

  57. ryanwc4 says:

    Another example that is amazing and amusing. The population of whooping cranes dwindled to about 20 half a century or so ago. Most were from a population in Texas with dangerously limited genetic diversity. One crane had been raised from a chick in a zoo, and she was from a Louisiana flock, so getting her genes back into the gene pool would be extraordinarily important to the future of the species. But she had imprinted on humans, and no matter how they tried to get her to mate with male cranes she wouldn’t. So finally, her closest human companion, George Archibald, began mimicking crane courtship dances, waving his arms, jumping while arching his neck back. He was able to get the bird to ovulate, and the egg was artificially inseminated, and this is one big reason that whooping cranes have been able to rebound.

  58. joscha says:

    This may have a rather straightforward explanation. While our individual need dimensions are genetically encoded, the objects of our needs are cortical representations that have to be learned. For instance, while we are genetically predisposed to be interested in faces, the actual face representations are built by looking at lots of faces.

    There should be some evolutionary benefit in getting mate seeking individuals to look for face-havers that are similar to their parents, but coding suitable face recognition code into the genome (or epi-genome) is going to be tricky. (Smells might be easier.) Thus, imprinting is likely as good of a proxy as it gets. We have to build face representations first, before we can associate them with our (usually innate) preferences.

  59. slovakmum says:

    There was a famous chimpanzee Lucy Temerlin, who was raised in human household. She masturbated at pictures of human males and never found her own species attractive, never reproduced, too. There is a wikipedia article about her.

  60. N.K Anton says:

    This is weird because I tend to have ‘seasonal types’ imprinted by one very specific girl.

    For the longest time, I was really into pale redheaded women because I knew one really attractive one and always disliked blondes until recently. After working with one of the most gorgeous women I ever been in the same room with, I only seem to notice and be attracted to blondes. My first crush was a Filipino classmate in like grade 3 and I was only attracted to East Asian girls until the end of high school.

    The kicker of all of this is my mother is not only blonde or redheaded or East Asian but a stay-at-home traditional Desi woman. The only real influence from her end is probably personality and an appreciation for “striking” or “exotic” eyes.

  61. Tyrant Overlord Killidia says:

    I tend to date girls with dark curly hair. Even if their hair is straight, I always later find out that they straighten it on purpose because left to its own devices would get curly. My first I guess proto-girlfriend was in 1st grade who had dark, curly hair.

    This means that I tend to date lots of Jewish girls. This year alone it’s been all Jewish girls, with the exception of one Chinese girl.

    Oh and no… I do not have a curly haired Jewish mother.

  62. Captain Ford says:

    Genes don’t directly code for appearance, so it seems really odd to me that you would expect genetics to tell us who we should seek as a mate. From an evolutionary standpoint, it’s much easier to evolve a general imprinting mechanism that works fine no matter how genes mutate in the future than it is to keep two sets of code in sync.

    Let’s say one day a donkey suffers a mutation that changes it’s coat color from gray to bright pink. But the gene that tells it what to look for in a mate still tells it to look for gray donkeys. This is fine. That’s actually what it should be looking for.

    But say instead that the code that tells it what it should look for in a mate suffers a mutation instead, and now its genes say to only look for pink donkeys. Well, there aren’t actually any pink donkeys, so this is a problem.

    There definitely are genetic factors that tell us what to seek in a mate, but these are really vague things like “emits this pheremone”, or “has a shape vaguely like this”, which are simple and redundant. If a random mutation destroys one of these, it doesn’t matter much, since there’s still a dozen-ish more that work. But something as specific as “shape of the nose” or “distance between the eyes” can change in just a few generations. I just don’t think it could evolve quickly enough to be able to target your specific ethnic group, since these systems only emerge at random, and the only way they can be made accurate is when failed genes are selected out (which would require a really big body count), they’re not nearly as useful as a system that just looks at what your parents actually look like and uses that.