The Germ of the Gaps

Posted by Sappho on May 8th, 2014 filed in DNA, Science, Sexuality

A couple of months ago, one of my favorite genetic genealogy blogs linked a couple of blog posts from a blog that focuses on HBD, and a firestorm broke out in the comments. I made a mental note to write about the HBDosphere, but put it off for other things, till Rod Dreher’s post about race, genetics, and Nicholas Wade reminded me. So, here goes.

HBD is short for Human Biodiversity, which is described by its supporters as the investigation of human biological diversity, but which I would describe as the discussion of human biological variation by people who think that Steve Sailer has ideas on the topic worth hearing. Steve Sailer originated the term, and people who disagree with his views, even if they have an interest in the genetic history of hominids, don’t tend to self-label as people who are interested in Human Biodiversity. There’s a political slant to the label.

The other reason I describe the label in terms of people who like Steve Sailer is that it seems more accurate to me than the common liberal description of HBD as being all about thinking black people are stupid. While it does seem to be true that nearly all commenters in the HBDosphere think that black people are genetically predisposed to have a lower IQ than white people (apparently strongly so), and that much blogging effort is expended arguing that this is so, HBD extends beyond the IQ argument. It includes HBDchick, who rarely writes about IQ (though her politics align just fine with Steve Sailer’s), and Heartiste, whose main concern appears to be pickup artistry. It also includes some variation in attitudes on immigration, all the way from, say, Ron Unz to Steve Sailer. But it definitely doesn’t include Stephen Jay Gould, Jared Diamond, or Richard Lewontin.

I’ve encountered the HBDosphere, recently, mainly because of the intersection between HBDish interests and those of genetic genealogy. We all use, though for different purposes, Dienekes Pontikos’ nifty set of admixture calculators. We occasionally appreciate the same books (the link that caused the firestorm was to a series of maps of American regional origins, based in part on David Hackett Fischer’s in Albion’s Seed and Colin Woodard’s American Nations: A History of the Eleven Rival Regional Cultures of North America). General information about population migrations is of obvious interest to both genetic genealogists and HBD bloggers, though we do very different things with that information. And so, in my search for genetic genealogy information, I’ve encountered the HBDosphere, particularly through Razib Khan, source of all kinds of information about how to analyze your 23andme data, and of politics that don’t agree much at all with mine.

It’s an encounter that leaves me unsure how seriously to take this bunch of people. By which I mean, is HBD one of those fringe ideas that you really only encounter on the Internet? Am I spending more of my attention on it than it deserves? After all, I haven’t in the past twenty years had even one person in real life tell me that he or she believed that genetics are a significant driver of ethnic differences in test scores. And often the HBD bloggers themselves seem to describe themselves as an embattled minority, rejected by conservatives as well as liberals. Or is it my real world experience that’s the bubble, and do these views have more influence than I realize? Nicholas Wade’s book, I gather just about to be published so most of us can’t see it yet, has been getting many advance cheers in the HBDosphere for support it’s purportedly going to give their views. Nicholas Wade is a science writer for the New York Times, and the favorable review that Rod Dreher linked is in the Wall Street Journal, written by HBD friendly Charles Murray (who isn’t nearly as discredited in the conservative world as he is in the liberal world).

Mostly, the gulf is about race. But I want to start with a different topic, since it figured heavily in the firestorm at the genetic genealogy blog (which blog I won’t link, because I don’t want to plague the blogger, whom I admire, with controversy not much related to her blog). That topic is the “gay germ.” The blogger to whom she had linked, you see, is Jayman, a strong advocate of the Greg Cochran’s “gay germ” theory. Jayman has called the gay germ “by far the most likely explanation for male homosexuality,” writing that

If the “gay germ” is real – which it almost certainly is.

I think that “almost certainly” is, let’s say, a severe overestimate of the degree of evidence that this theory has been able to muster.

But I need to begin my story about the gay germ with a discussion of the theory that it’s meant to counter, the theory of the “gay gene.”

Though the “gay gene” and the “gay germ” are rival theories, it should be remembered that they aren’t, in principle, completely at odds. Dean Hamer, a leading advocate of the “gay gene” theory, says that he’s actually talking about many genes, not a single one, combined with environmental factors, that could include (his own example) something like getting measles as a child. (I don’t expect that he actually believes disease is a significant factor; he probably used measles just to show that “environment” doesn’t necessarily mean “social environment.” But in principle you could both have one or more “gay genes” and one or more “gay germs.”) Greg Cochran, the leading advocate of the “gay germ,” does acknowledge the possibility that genetic factors could influence response to that pathogen. Still, as theories of the main biological cause of sexual orientation, the theories are starkly opposed.

They also, obviously, have a different emotional feel. Germs are icky, and if you find homosexuality icky, you may find it easier to believe in a gay germ. Genes are natural, and if your gay, lesbian, or bisexual orientation (or that of people dear to you) feels altogether natural, normal, and healthy, a gay gene has more appeal. But I’ll save the feeling side of the story for another post. For now, let’s pretend that I’m viewing this debate with Spock-like detachment, just looking at the evidence. I’m not, of course; I have feelings the same as everyone else. But you don’t have to care about my feelings to evaluate my argument.

I’ll start, though it’s a major oversimplification, by discussing a hypothetical single gay gene. If we have a gay gene, why hasn’t evolution bred it out of us, given the obvious reproductive disadvantage?

We know that having a larger number of older brothers increases the probability that a man will be gay, so perhaps we’re actually talking about new random mutations, which increase in children of older parents? It turns out that the answer is, no, we’re not. Paternal age, which is known to be associated with increased mutational load, is not associated with male homosexuality. (Though I myself, obviously, am not a gay man, but a married and faithful bisexual woman, I’m going to be forced to discuss mostly male homosexuality, because that’s what has been most researched. It’s also where most of the “gay gene” theories are; nobody worries much about how a “bisexual woman” gene could survive, if there is such a gene.)

So, if we’re talking about a gay gene, it has to have some reproductive advantage, to counter the obvious reproductive disadvantage, since, infrequent though homosexuality is, relative to heterosexuality, it obviously isn’t being bred out of us.

Perhaps the gay gene is, as E. O. Wilson suggested, a “good gay uncle” gene, which survives through kin selection. J.B.S. Haldane once said, “I would lay down my life for two brothers or eight cousins.” A gene that provides enough reproductive advantage to your close relatives can survive, even if it doesn’t supply you with any children directly. This effect is particularly evident among eusocial insects like bees, and so it’s no surprise that E. O. Wilson, researcher of eusocial insects, was one of the early proponents of kin selection theory. (Wilson has since shifted to arguing for a kind of multilevel selection that includes both individual selection and group selection, but as far as I know there isn’t a non-kin group selection theory explaining how a “gay gene” could work, so I’ll leave group selection alone till I run across it in my other series, where I make my way through blogging Jonathan Haidt’s book.)

Bees are supersisters, sharing more than the usual 50% of DNA that human siblings would share, so worker bees get more evolutionary bang for the buck when they lay down their lives for nephews and nieces than we do. A human good gay uncle has to be a very good gay uncle indeed for his kin selection advantage to outweigh his reproductive disadvantage (unless, I suppose, social pressure leads him to have enough less enthusiastic sex with his wife that he doesn’t suffer that much of a reproductive hit, in which case the gene only starts to get bred out of the population once GLBT relationships are socially accepted). But even a less impressively good gay uncle could find his avuncular virtue to be a partial offset to the evolutionary drawback of his gay gene, provided that the gene also has some other evolutionary advantage.

One such proposed advantage is the “fertile sister” gene, a sexually antagonistic gene that makes brothers gay and sisters more fertile. Maybe it just makes everyone who carries it, male or female, find sex with men really, really hot. Or maybe it has some other mechanism that we haven’t figured out. A 2004 study of 200 Italian families actually did show that sisters of gay men had more children, so maybe this version of the gay gene does exist. The study would need to be replicated in a larger population, though, for us to be sure.

Dizygotic advantage is another way that genes survive, despite a negative effect on survival or reproduction. If people with one copy of a gene have an advantage, while people with two copies have a disadvantage, the gene will reach equilibrium at some level of presence in a population. The most famous example is sickle cell anemia, where a single copy of the gene provides some resistance to malaria, while two copies cause anemia.

Another possibility is that the gay gene makes men prefer sex with men, but simultaneously in some way makes them more attractive to women. Evolution isn’t about making us happy; it’s about making us grandparents. And if, for much of our evolutionary history, a significant part of the male population has failed to reproduce because, despite really, really wanting sex with women, they couldn’t persuade women to have sex with them, then men who didn’t want sex with women all that much, but who were really good at persuading women to have sex with them, may actually have been rather good at producing grandchildren.

Part of the lesson of this exercise is that it’s possible to make up a Just So story for anything. The test is in whether you can produce evidence for your Just So story. Are gay men really more helpful uncles than straight men? Are their sisters really more fertile (can that Italian study be replicated)? Are they more attractive to women?

Proving that a “gay gene” couldn’t exist, though, is harder. Because you’d have to prove that none of the possible explanations for how a “gay gene” could work will hold true. And it gets worse. Because, in fact, nobody really thinks there’s just a single “gay gene.”

To the extent that homosexuality is genetic, it is, like height, the effect of a number of different gay genes. We’re actually not talking about a gay gene, but a gay epistasis, a combination of genes that, if you get enough of them, will make you gay. This could explain why there have been mixed results in replicating Dean Hamer’s finding of an Xq28 gene that is associated with homosexuality. Perhaps Xq28 either is or isn’t a gay gene, depending on whether you’re surveying a population that does or doesn’t have a high frequency of the modifier genes needed to make Xq28 a gay gene.

Now, if we have a gay epistasis, rather than a single gay gene, then something similar to dizygotic advantage could operate at the epistatic level. Perhaps some of the genes involved in this epistasis have reproductive advantages (it needn’t be the same one for each gay gene) that show up even when you don’t have the full set, and the evolutionary disadvantage, homosexuality, only shows up if you have the genes in combination. You can see how it wouldn’t even take that large of a reproductive advantage per gene for the math to work, if you need fewer genes to get the evolutionary advantage than you do to get the evolutionary hit.

Finally, we have to consider the fact that, however many gay genes there may or may not be, they don’t work independently of the environment. Twin studies show a concordance rate of, well, actually twin studies turn out to show a variable concordance rate (I’ll get to that later), but none of them show a concordance rate of anything like 100%. So we’re actually talking about, not a flat out gay gene, but a gay gene expression.

“Environment” here could be either social environment (such as Daryl Bem’s “Exotic Becomes Erotic” theory of the origin of sexual orientation) or biological environment (such as hormonal exposure in the uterus).

One known environmental factor is the number of older brothers a man has. The more older brothers you have, the more likely you are to be gay. The study that showed this correlation reported that each older brother increases the odds of homosexuality by 33%. Father’s and mother’s age are no longer significant when siblings are controlled for (unlike some other traits that, due to the greater mutational load in aging parents, increase with age). It’s just older brothers. Not age. Not older sisters. The researchers who found this effect, Blanchard and Bogaert, theorize that the increase has something to do with a mother’s immune response being more likely to react to male specific proteins if the mother has carried previous male babies.

So, let’s suppose that we have a gay gene that only gets turned on under particular prenatal circumstances, ones likely to apply in women who have had previous pregnancies involving sons. But this gene might, hypothetically, convey some evolutionary advantage on everyone else. Our gay uncle has nephews and nieces, not only from his fertile sisters, but from his fertile older brothers. With enough siblings, enough nephews and nieces, and enough of an evolutionary advantage (of some kind or another) for the straight members of the family, the math could work. (Of course, it’s also possible, even, I suppose, likely, that the older brother effect isn’t mediated by any such gene at all. My point here is simply that whatever genetic influence is involved, there’s no particular reason to expect its “gay gene” effect to be activated in all environments.)

And here is where the case for the “gay germ” breaks down.

Now, the way the “gay germ” theory works is that Cochran takes each possible explanation for a “gay gene” in turn, and argues for why he thinks it won’t work. Here’s a round up, by Jayman, of Cochran’s various posts debunking various possible mechanisms by which a hypothetical gay gene could work.

And that’s it. That’s all the evidence for this “almost certainly” real gay germ.

I could imagine positive evidence that could point in the direction of a pathogen being a major factor. For instance, sexual orientation could have a prevalence that fit with some disease epidemiology. We could have a seasonal gay germ effect if, for instance, some seasonal virus, like the flu virus, altered sexual orientation if mothers catch the bug in a particular stage of pregnancy. Or if we had a gay germ that, like rubella, came in epidemics every seven years, we could trace its path. Perhaps a gay germ might show correlation with some disease vector. If it’s toxoplasmosis, for instance, you might be more likely to be gay if your mother changed the kitty litter herself. Perhaps children are more likely to be exposed if they live in crowded environments, such as cities. Or perhaps there’s a geographical “gay germ” belt, like the chlamydia related infertility belt in Central Africa. And of course, if we ever found actual antibodies that systematically differed between gay and straight people, that would be a big clue.

But no, we don’t have any positive evidence for a gay germ. It’s a germ of the gaps, a germ whose sole justification is that, well, we don’t know that anything else causes homosexuality.

It also seems a particularly improbable germ of the gaps, because, as Chris at The Lure points out, it would have to act like no pathogen we know.

This points to the biggest weakness of the gay germ theory: it fundamentally misunderstands why pathogens exists. When non-biologists consider pathogens (fun fact: most of the vocal gay-germ theorists are non-biologists), the temptation is to focus on the symptoms rather than on the pathogen itself. Thus viruses are simply bad things that happen to people. But pathogens are fairly indifferent to their host’s status. The only metric selection judges them on are their ability to reproduce and to find new hosts. The cost to the host comes as a service towards one of those two ends. For instance, the open sores and blisters caused by herpes are in part the result of the virus hijacking skin cells to produce so much protein components for future herpes viruses that the cell ruptures (the body’s own response to the infection also plays a role). These sores also aid the herpes virus in infecting new hosts, as the virus requires skin-to-skin contact with the sores to transmit. Indeed, there is a relationship between the parts of the body a pathogen typically infects and the mode of transmission: airborne pathogens typically infect the respiratory system, water-borne pathogens the GI tract, STDs genitalia and their related emissions. Biologists have also postulated a similar positive relationship between a germ’s virulence and its ease of transmission: the more a pathogen requires of its host to find a new host the less it can afford to impair its function. For instance, even though HIV and ebola viruses attack similar cells and transmit through contact with body fluid, Ebolaviruses transmit much much more easily (and produce copious bodily fluids externally) and thus can afford to kill their hosts within a matter of days rather than the years an HIV infection takes.

The hypothetical gay germ violates all of these principles. Rather than maximize its short-term viability, its plan seems to be:

1. Burrow down into the inner portion of the brain, probably the hardest place for an external pathogen to reach
2. Infect a small batch of neurons that control sexual orientation (but little, if anything, else), which, if the germs hits shortly after birth, will have no effect on the host’s behavior for at least a decade
3. ?????
4. Profit!

It’s one thing to be a pathogen that manages to alter host behavior in a way that helps with your own replication. A sexually transmitted disease that makes the crickets it infects horny, or a toxoplasmosis induced loss of fear of cats that allows the Toxoplasma gondii parasite to get from its mouse host to the cat gut that is the only place where it can reproduce, both make sense. But a germ whose one effect is to change its host’s sexual behavior at least a decade in the future?

Gay genes, in contrast, almost certainly exist; the question is, how big of an effect do they have? The gold standard for determining how much a trait is genetic and how much it’s environmental, you see, is twin studies, in which you compare identical and fraternal twins. Twin studies seem to be all over the map in how large a genetic effect they report. At the high end, a 1991 study reported 52% concordance in identical twins and 22% in dizygotic cotwins, but other studies have reported lower concordance.

A 2005-2006 survey of all adult twins in Sweden found

Twin resemblance was moderate for the 3,826 studied monozygotic and dizygotic same-sex twin pairs. Biometric modeling revealed that, in men, genetic effects explained .34-.39 of the variance, the shared environment .00, and the individual-specific environment .61-.66 of the variance. Corresponding estimates among women were .18-.19 for genetic factors, .16-.17 for shared environmental, and 64-.66 for unique environmental factors. Although wide confidence intervals suggest cautious interpretation, the results are consistent with moderate, primarily genetic, familial effects, and moderate to large effects of the nonshared environment (social and biological) on same-sex sexual behavior.

So I figure maybe a third of the cause of sexual orientation is genetic, with would make sexual orientation pretty much on a par with many other traits (most personality traits and mental abilities seem to be at least 20% genetic). Some of the environmental cause is probably prenatal. Possible prenatal environmental influences include the immune reaction that Blanchard and Bogaert have proposed to explain the older brother effect, and variation in prenatal hormonal exposure, including epigenetic effects in which epi-marks from the opposite sex parent:

In this model, homosexuality occurs when stronger-than-average SA-epi-marks (influencing
sexual preference) from an opposite-sex parent escape erasure and are then paired with a weaker-thanaverage
de novo sex-specific epi-marks produced in opposite-sex offspring. Our model predicts that
homosexuality is part of a wider phenomenon in which recently evolved androgen-influenced traits
commonly display gonad-trait discordances at substantial frequency, and that the molecular feature
underlying most homosexuality is not DNA polymorphism(s), but epi-marks that evolved to canalize
sexual dimorphic development that sometimes carryover across generations and contribute to gonadtrait
discordances in opposite-sex descendants.

Even identical twins don’t have identical prenatal environments, so prenatal and well as postnatal environment could figure in twin non-concordance.

And there may be additional environmental effects after birth. Some of them may be quite large (given the large supply of identical twins with non-identical orientations), but no one’s figured out yet what they are.

I’m OK with that uncertainty. Given a choice between the gaps and an implausible germ of the gaps, I’ll stay with the gaps, and just figure we don’t fully know yet why some people are straight and some are gay.

17 Responses to “The Germ of the Gaps”

  1. JayMan Says:

    Well, I appreciate your effort to evaluate HBD, and your fairly honest (though frankly poor) attempt to evaluate its claims.

    To that end, it would have been helpful if your facts were correct, especially about the HBD sphere. There is a simple reason the people that embrace HBD do so: the evidence, which is clear. For a primer, see here:

    JayMan’s Race, Inheritance, and IQ F.A.Q. (F.R.B.) | JayMan’s Blog

    Also, it’s probably worth mentioning that I am Black.

    As for your criticism of the “gay germ” theory, I will of course set aside emotion and analyze your claims in a Spock-like manner. But then, that’s what I always do. My first general piece of advice (or a question really) is why do many assume that Cochran hasn’t put a little bit of consideration into this idea? Do you think he hasn’t thought closely about your objections? Do you honestly believe that he hasn’t put more than 5 minutes of thought into it? Would it not have been prudent to check to see if he has in fact addressed these points? In fact he has. You can find an index of the his writing on the matter right here:

    Greg Cochran’s “Gay Germ” Hypothesis – An Exercise in the Power of Germs | JayMan’s Blog

    As an aside, on the bisexual woman issue, it’s also worth mentioning that some of us have in fact considered it. You can read some of my musings on the matter here:

    The Evolution of Female Bisexuality | JayMan’s Blog

    I will quickly address your criticisms of the gay germ theory here:

    1. Cochran had a field day with E.O. Wilson’s “gay uncle” nonsense. Simple mathematics shows why this couldn’t be the case. In order for an allele to survive that essentially zeros fitness in the individual in which its main effect is expressed (as any putative homosexual alleles would do), it would have lead the individual to ensure that at least four more nieces/nephews survive than otherwise would. This is a set of behaviors that would need to be stronger and more effective than mother love. It clearly does not exist.

    2. The “fertile” sister idea has been ruled out by genomic analysis (GWAS). It would have been easy to find there; it wasn’t found.

    3. GWAS, by the way, also rules out heterozygote advantage. Such a gene would stick out like a sore thumb. Nothing found. Further, this is evolutionarily preposterous: heterozygotes would need an enormous selective advantage to lead an allele so deleterious in homozygotes to stick around. What kind of advantage would that be? Certainly malaria isn’t a problem in much of the world, is there some similar selective pressure?

    4. The idea of some sort of compensatory fertility advantage in male homosexuals (say, making them more attractive) is invalidated for a simple reason: how to explain being homosexual? The struggle for mates is a fierce one. Having zero interest in the opposite sex predisposes one not to try to attain a mate. Such individuals would be quickly out-competed by those who were actively trying to land a mate. We’re talking on average here. Even if some gay individuals had exceptional reproductive success, gay men as a whole would be less fecund than straight men as a whole, leading whatever such genes to vanish over evolutionary time.

    5. The idea that homosexuality is polygenic (like most traits) doesn’t help things. Selection acts on each gene, and each would have to have some compensatory fitness advantage to stick around. If you’re saying that it’s a form of “balancing” selection, where the fitness hit of homosexuality is balanced by a fitness boost in individuals who have some of the putative risk alleles, but not all of them, there are two big problems:

    First, that fitness boost in the relatives of homosexuals would need to be huge. It would also need to be common, considering that homosexuals are 3-5% of all males. What could that boost be? Why isn’t it obviously associated with homosexuality in families? Also, why hasn’t natural selection decoupled this effect over evolutionary time, given the huge fitness hit of homosexuality? This is similar to the problem of heterozygote advantage. Without some huge selective force, evolution would have “ironed-out” the problem and would have de-coupled the positive fitness impact of these alleles from their negative fitness impact.

    Second, this idea runs into another huge problem: the low heritability of homosexuality. Twin-registry studies, which give us our best representative samples to estimate the heritability of homosexuality find that it is low, less than 0.22. If any genetic impact was behind it, we would see much higher heritability.

    6. Congenital/prenatal effects just don’t work. Homosexuality is simply far too common (3-5%). Similar fitness-reducing congenital defects, like speechlessness or malformed genitalia are nowhere near this common.

    7. Due to sampling and other issues, it’s still not clear if the fraternal birth order effect is real. One study (with N ~ 1,000,000) failed to find one.

    7. Lund’s criticism of the mode of action of the pathogen is laughable (good thing “misdreavus” isn’t here – who himself is gay, but has a deep understanding of infectious pathology). It is easily debunked when one thinks of polio. Paralysis clearly doesn’t help the polio virus spread, but it is a side effect of the virus’s mode of action. Cochran talks about the likely damage to brain cells that may be an autoimmune effect in defending against the pathogen. In the arms race between pathogen and host, all manner of defenses and assaults have evolved, in which pathogens always have the upper hand due to their much faster rate of evolution.

    Even more broadly, in a paper Cochran co-authored with Paul Ewald, it is discussed about the many yet-to-be discovered pathogens and the ailments they cause. The discovery of pathogens suffers from a low-hanging fruit/”looking for your keys under a street light in the dark” problem. The ones we have found are the easiest ones to find. They say little about the ones that are harder to find, observe, culture, etc. That we have not found such a pathogen is not a criticism.

    But no, we don’t have any positive evidence for a gay germ. It’s a germ of the gaps, a germ whose sole justification is that, well, we don’t know that anything else causes homosexuality.

    You don’t see the explanatory strength of this? It’s hardly like you’re here posing a workable alternative yourself. As I stated at my post: “when you have eliminated the impossible, whatever remains, however improbable, must be the truth.” That’s where we’re at with this. Eliminating potential alternatives is itself a fine process for getting to the truth.

    You say you would analyze this with “Spock-like detachment.” I would advise that when you’re trying to get to the truth, you should always approach the situation that way, without exception. It is only way to proceed.

  2. Peter Says:

    After all, I haven’t in the past twenty years had even one person in real life tell me that he or she believed that genetics are a significant driver of ethnic differences in test scores.

    Which is hardly surprising. It’s much too taboo a topic for people to discuss in real life except maybe in very trusted circles. Only with the relative anonymity of the Internet do people feel safe enough to speak.

  3. Sappho Says:

    Oops! Was away from the blog for a few days and left Jayman’s comments in moderation, so as to give him the impression I was moderating him out entirely. That would hardly be fair. I don’t have time to respond to this discussion this morning, but just have one quick comment.

    “You say you would analyze this with “Spock-like detachment.” I would advise that when you’re trying to get to the truth, you should always approach the situation that way, without exception. It is only way to proceed.”

    I think that the way to proceed to get at the truth is to frankly acknowledge your bias, which we all have, and not pretend you actually *have* that “Spock-like detachment” (much less that you have it and others don’t), but to make some attempt to separate out gut feelings and intuitions from more methodical reasoning. So that’s the approach I try to take.

    @Peter: It wasn’t always too taboo to discuss in real life, though. When I was young, it wasn’t out of the bounds of normal discourse for the older generation. And I find it significant that it’s out of bounds much of the time in conservative and Republican circles as well as in liberal and Democrat circles (Republicans say many things that Democrats consider racist, without being put off by the opinions of Democrats). At the same time, the discussion in even mainstream newspapers and magazines following the publication of a book like The Bell Curve or like Nicholas Wade’s new book suggests that more people believe these things than are willing to say them.

  4. Lynn Gazis-Sax (Sappho) Says:

    OK, now I’ve actually found time to read the long comment I approved, and want to respond to “when you have eliminated the impossible, whatever remains, however improbable, must be the truth”. I see two problems with that line of reasoning:

    1) There are just too many theories (with many of the studies too preliminary and conflicting) to eliminate all but one as impossible. That’s why I end with an unknown.

    2) If I did go the “eliminate the impossible” route, the “gay germ” is one of the options that I’d eliminate as impossible, leaving one of the others as the merely improbable theory that I’d be obliged to accept.

    Here’s how I see it:

    Gay gene: Not just possible, but probable. There are plenty of genes that survive despite sometimes imposing a genetic disadvantage. And concordance among identical twins is significantly higher than among fraternal twins. So, there are multiple gay genes. But 50% genetic is the upper bound possible, if you take the most “gay gene” favorable twin concordance studies, and something like 30% is more likely. So there are other factors.

    Gay germ: Not with the decided lack of epidemiological evidence pointing to any kind of pathogen. And I still find the mechanism implausible.

    Immune factor: This one’s possible. It has a plausible mechanism (maternal immune reaction, sensitized by prior pregnancies, affecting hormonal sensitivity), some “older brother effect” support, and fits with the observed gender atypicality of children who grow up to be gay. It’s hardly conclusive, but if I have to reject the impossible in favor of the improbable, this one goes in the “at least possible” bucket.

    Freudian stuff: Pretty much a dead horse, can be dropped in the impossible bin.

    Choice: Also can be dropped in the impossible bin.

    Exotic Becomes Erotic: We have a winner! It allows for the significant genetic factor shown by the fact that there’s *some* significant twin concordance, while allowing enough environmental factor to explain the non-concordance in identical twins. It can involve early imprinting, such that you don’t remember your sexual orientation as chosen. I can, if I need to, think of benefits to a general “exotic becomes erotic” tendency in humans. Usually, it would pull people toward the opposite sex: an evolutionary win. It could explain why women tend to be attracted to men whose immune system differs from their own: another evolutionary win. And it could tie in to the Westermarck effect, in which children who grow up together (as in kibbutzim or the institution of minor marriage in Taiwan) seem to show a reverse sexual imprinting, another evolutionary win in that that effect discourages mating with close kin. Finally, it can explain what “gay genes” might be doing in the first place. If they are, not so much genes directly for homosexuality, but genes for gender non-conformity, then they could be “alternate mating strategy” genes. Many species, humans included, have alternate mating strategies. A range of genes, leading to varying levels of masculinity/femininity, and only occasionally, when you get a particular genetic load in combination with a particular environmental imprinting, making you gay, at least strikes me as possible, so I’ll pick it as the theory left, however improbable, after I’ve rejected the impossible ones.

    I’m not particularly wedded to “Exotic Becomes Erotic,” though; it could well be thoroughly debunked tomorrow. I prefer just acknowledging our limited knowledge, and figuring that several theories might prove possible, though some can be rejected.

  5. JayMan Says:

    OK, now I’ve actually found time to read the long comment I approved, and want to respond to “when you have eliminated the impossible, whatever remains, however improbable, must be the truth”. I see two problems with that line of reasoning:

    1) There are just too many theories (with many of the studies too preliminary and conflicting) to eliminate all but one as impossible. That’s why I end with an unknown.

    2) If I did go the “eliminate the impossible” route, the “gay germ” is one of the options that I’d eliminate as impossible, leaving one of the others as the merely improbable theory that I’d be obliged to accept.

    Well, if you are going to engage in that type of reasoning, you need to do it properly. Unfortunately, you are failing on that front.

    Here’s how I see it:

    Gay gene: Not just possible, but probable. There are plenty of genes that survive despite sometimes imposing a genetic disadvantage.

    Name one. We’re talking a net disadvantage (and remember that genetic evidence has ruled out heterozygote advantage and sexual antagonistic selection).

    And concordance among identical twins is significantly higher than among fraternal twins.

    No, it isn’t really:

    Neither Genes nor Choice: Same-Sex Attraction Is Mostly a Unique Reaction to Environmental Factors

    Comb through any twin registry study (a requirement to minimize selection bias) and take a look at the concordance rate of MZ vs. DZ twins.

    Gay germ: Not with the decided lack of epidemiological evidence pointing to any kind of pathogen. And I still find the mechanism implausible.

    I pointed it to you previously, but here it is again. You might want to read Greg Cochran’s & Paul Ewald’s 2000 paper on pathogens. A good fraction of pathogens that affect humans, and their actions, are actually unknown – possible a majority.


    Infectious Causation of Disease: An Evolutionary Perspective

    Immune factor: This one’s possible. It has a plausible mechanism (maternal immune reaction, sensitized by prior pregnancies, affecting hormonal sensitivity)

    Nope. This is essentially impossible given the nature of evolution: natural selection has a HUGE incentive to ensure that low-fitness phenotypes do not occur. Hence, very low fitness phenotypes are all rare. Homosexuality is not. It just doesn’t work.

    Exotic Becomes Erotic: We have a winner! It allows for the significant genetic factor shown by the fact that there’s *some* significant twin concordance, while allowing enough environmental factor to explain the non-concordance in identical twins. It can involve early imprinting, such that you don’t remember your sexual orientation as chosen.

    Again, as per above, this absolutely does not work. Natural selection has a huge incentive to make sure that this sort of thing happens as little as possible.

    Finally, it can explain what “gay genes” might be doing in the first place. If they are, not so much genes directly for homosexuality, but genes for gender non-conformity, then they could be “alternate mating strategy” genes.

    You’re good at coming up with stuff, but not good at evaluating them. Think this trough. Why would any alternative mating strategy – apparently defined here as even lack of attraction to the opposite sex – be at all selectively advantageous? Especially in men?

    OK, I have explained the situation more than sufficiently, and given you references to evidence on the matter. I have to say, you see unfortunately ignorant on basic evolutionary theory. One good resource on that is:

    Keller, Matthew C.; Miller, Geoffery (2006). Resolving the paradox of common, harmful, heritable mental disorders: Which evolutionary genetic models work best? Behavioral and Brain Sciences, 28, 285-452.

    My time is limited, so I’m going to leave this off here. If you have any questions, feel free to ask, but the best way to get a hold of me is through e-mail or via my blog.

  6. Sappho Says:

    “Alternative mating strategy” is *not* defined here as lack of attraction to the opposite sex. I’m defining “alternative mating strategy” as “less stereotypically masculine behavior.” My assumption is that *most* such people would still be heterosexual (after all, identical twin concordance appears to be around 30%), just less macho heterosexual men, and that in some circumstances it would work to be a more macho heterosexual man and in other circumstances to be a less macho heterosexual man. Makes more sense to me than a germ which follows no epidemiological pattern whatsoever and whose *only* effect, apparently, is to make men gay.

    I also think you overestimate the reproductive hit of homosexuality, through most of our evolutionary history. It’s true that, for those men in whom the gene expression winds up as a “gay gene,” there’s a hit, since they’re not so strongly motivated to have the one kind of sex where they can reproduce. So, on the whole it’s an evolutionary disadvantage in that case. But you’re assuming a *huge* evolutionary disadvantage, that can only be made up by enormous advantages elsewhere. I’m assuming that, humans living in communities, motivated would be grandparents, for much of human history, would still be marrying those men off, and they would then have relatively passionless, not all that happy marriages, in which their wives wonder why their husbands don’t find them attractive while the husbands cheat on the wives with other men, but which still manage to produce some children. That is, after all, what happened in our own culture, for the most part, in the pre-Stonewall era. I personally know multiple men who are gay, not bisexual but distinctly preferring men, who had marriages and children in this fashion.

    You can call this ignorance of evolutionary theory, if you like; I call it a realistic assessment of how the closet actually worked.

  7. Sappho Says:

    So that, on balance, for the “gay gene” (or, actually, genes) to eke out a modest survival to a sufficient degree that a small minority of the male population is predominantly gay, you don’t need a *huge* advantage to having it, just some moderate compensating advantage that will be spread out over the set of men who carry the genes and aren’t gay (a larger set than those who carry the genes and are gay). And that set of people may be even larger than the 30% concordance rate among identical twins would imply, since, in addition to the people who aren’t gay even with the same “gay gene” genetic load as those who are, there would also be people with a smaller “gay gene” load.

  8. Sappho Says:

    The link between “less stereotypically masculine behavior” and “gay gene” being, of course, Daryl Bem’s “Exotic Becomes Erotic” theory, based on the observation that gay people have a tendency to gender non-conformity as children. Not that I’m wedded to that theory. Just that it strikes me as less impossible than a “gay germ.”

  9. Sappho Says:

    Incidentally (and my last word on this today, apologies to anyone reading this thread for not pulling my thoughts into a single comment), I note that Matthew C Keller, recommended to me by Jayman as an expert on heritable mental disorders (and really useful to me here, since I’m actually more interested in the genetic factor in bipolar disorder than I am in the genetic factor in homosexuality), writes, on his biosketch page (, that, “However, and importantly, while much evidence appears consistent with a mutational role in the genetic variation of traits, we have recently reported several findings that at face value appear inconsistent with predictions of a mutational model. For example, we have reported that several predictions of a polygenic mutation hypothesis are not born out for homosexuality (ZIETSCH et al. 2008), suggesting that perhaps balancing selection accounts for heritable variation in this trait.” So, evidently Keller himself doesn’t think that only profound ignorance of the nature of evolution could lead a person to suggest that there’s some balancing advantage carried by the genes that predispose to homosexuality (this in contrast to schizophrenia and bipolar disorder, which he describes in the same biosketch as mutations that haven’t yet been bred out, correlated with things like paternal age and ionizing radiation).

    Here is the reference that Keller himself supplies:

    ZIETSCH, B. P., K. I. MORLEY, S. N. SHEKAR, K. J. H. VERWEIJ, M. C. KELLER, S. MACGREGOR, . . . N. G. MARTIN, 2008 Genetic factors predisposing to homosexuality may increase mating success in heterosexuals. Evolution and Human Behavior 29: 424-433.

    I’ll note that “genetic factors predisposing to homosexuality may increase mating success in heterosexuals” is *exactly* what I meant by the words “alternative mating strategy.” I think I win this round.

  10. misdreavus Says:

    Jayman has already refuted some of your most egregious errors, so I have I do not belabor this point.

    1) Male sexual orientation follows a “J shaped” distribution, meaning that the overwhelming majority of adult males cluster around “1″ and “6″ on the Kinsey scale, with relatively few men who identify as bisexual. (And I guarantee you a large percentage of those putative “bisexual men” are lying.) Whatever it is, it is NOT a quantitative variable, which means that the statistical distribution of hundreds or thousands of alleles, each of which contribute to a minute portion of adult male variation for sexual orientation, ought to be regarded as a scientific non-starter. Dean Hamer’s findings have consistently failed to replicate. It just doesn’t work out.

    Just imagine if 95% of men were 6’1″ to 6’4″ in stature, while virtually all of the rest were 4’2″ or below – and GWAS repeatedly failed to identify putative SNPs that explain the persistence of this unfortunate, vertically challenged minority, which happens to resurface generation after generation (I.e. midgets who somehow reproduce, more often than not, have kids who are taller than 6’1″). Wouldn’t you find that just *slightly* odd? Which brings me to the following:

    2) The overwhelming consensus from twin studies with *randomized* cohorts is that male homosexuality has an extremely low heritability, usually on the order of 30% or less – far lower than height or IQ (roughly 70-90%) or the Big 5 (around 50%). So whatever is responsible for the overwhelming cases of male homosexuality, it’s got to be some sort of environmental trigger – one that happens to cause homosexuality despite persistent selection against the trait.

    3) Male homosexuals suffer a loss in reproductive fitness that is roughly equivalent to that of achondrplasic dwarfs. In other words, from an evolutionary point of view, it’s better to be a 3’10″ midget than it is to be Ian McKellen. Whether or not the fitness value of homosexuality has been this low across all cultures and epochs is a moot point – not wanting to screw your wife is a big problem, even in cases of forced arranged marriage. Even a fitness loss as small as 5% would be enough to drive selection against this trait, and in the case of environmental triggers, enable selection for any and all behaviors that mitigate the likelihood of encountering said “trigger” in the first place. (For instance, if consuming lemon juice in utero were a potential trigger for homosexuality, selection would favor women who dislike the taste of lemons – and every single one of these behavioral proclivities is heritable.)

    4) knowing that male homosexuality imposes such severe fitness costs to its bearer, the only kind of environmental trigger that can cause it, generation after generation, must be something that can evolve *on its own* to combat the defenses of the host organism. We know that homosexuality isn’t very heritable, and that sexually antagonistic selection / heterozygote advantage are not credible genetic explanations. What else is supposed to make sense? “Germ of the gaps” indeed.

    ADDENDUM: As for whether or not homosexuality follows any significant epidemiological patterns, of course it does. In the United States, at least, boys born in urban areas are substantially more likely to identify as gay during adulthood than boys who spent their formative years in rural areas. Some of the gap might be attributed to greater intolerance of homosexuality among religious conservatives, but there is a noticeable pattern here.

    We also know that homosexuality is strangely absent among hunter gatherer groups such as Inuits and Aka pygmies. This has been carefully documented in the anthropological literature, and you can research it for yourself. (

  11. misdreavus Says:

    And in case I haven’t made my point clear – appealing to environmental factors that operate in *tandem* with gay genes is a scientific non starter. As I mentioned, even in that scenario, selection would favor behaviors that discourage any encounters with said environmental trigger. Even if it were something as capricious as the weather.

    Were we were to identify any putative gay SNPs that account for all the genetic variance, that still leaves unexplained the mystery of WHAT IS CAUSING EVERYTHING ELSE. There’s got to be some sort of trigger
    *in the absence of which* male homosexuality becomes improbable.

    Random methylation patterns in the genome that occur during fetal development (I.e. “epigenetics”) is scientifically untenable. The development of sight, vision, hearing, etc. occur more or less robustly, and are largely insensitive to major environmental insults – with the noticeable exception of pathogens (e.g. rubella causes deafness, river blindness destroys vision). Absolutely no trait that is as valuable to fitness as the human sex drive is going to f*** up generation after generation, among a certain percentage of the population for no good reason.

  12. Sappho Says:

    Few men fall in the middle of the Kinsey scale, but the proportion of Kinsey 5s is far from negligible. Really strong social pressure on men to marry and really strong limitations on women divorcing their philandering husbands (straight or gay) do “work” in the sense that they get a lot of gay men to be husbands and fathers. So, relative to, say, Tay Sachs or cystic fibrosis, yes, the hit on evolutionary fitness is moderate. (Relative to, say, hay fever, I’d figure the hit to be large.) Genes do exist that have a worse effect on fitness than this one.

    Mainly, the take away that I get from going through this exercise is that, of all the mechanisms for explaining a “gay gene” that people have produced, the “good gay uncle” looks weakest (I don’t see any evidence that gay uncles are *that* good). And that the most robust clue to the cause or causes of sexual orientation seems to be the correlation between gay or lesbian sexual orientation and being a somewhat more feminine man or somewhat more masculine women, especially in childhood (but still, most gender atypical children seem to wind up straight).

    Since I see that Jayman tweeted this link on his feed, and since that may produce more comments than I have time to reply to, I won’t promise a response to comments, but I’ll approve them as they come, with minimal moderation, and keep an eye out for any interesting references to peer reviewed literature (regardless of what I think of the arguments they’re attached to). The thread will automatically close two weeks after I published the post, as all my threads do.

    Off to do more conference blogging.

  13. misdreavus Says:

    “So, relative to, say, Tay Sachs or cystic fibrosis, yes, the hit on evolutionary fitness is moderate. (Relative to, say, hay fever, I’d figure the hit to be large.) Genes do exist that have a worse effect on fitness than this one.”

    Right. Which is why they are either extremely rare (genetic disorders are constantly being purged through purifying selection, despite being continually generated though random mutation), or they carry a heterozygote advantage that is commensurate with their frequency throughout the population.

    The prevalence of the sickle cell allele had dropped by an order of magnitude among African Americans since the early days of slavery – precisely because it simply doesn’t boost fitness the same way here that it does in west Africa. Bad genes will diminish over time. There is really no known exception to this rule.

  14. Sappho Says:

    Oh, I don’t disagree that genes that provide a net reproductive disadvantage decline. I’ve never argued otherwise. That would be why homosexuality is also rare, and, to the degree that it’s genetic it may well also decline, particularly as social pressure to get those Kinsey 5s into heterosexual marriages declines. Unless use of assisted reproduction replaces that social pressure as a mitigating factor.

  15. misdreavus Says:

    But male homosexuality isn’t that rare, in most industrialized countries – figures vary from 1-5%, depending on opinion polls. That’s exceedingly high for a trait that diminishes your fitness by 80%.

  16. Sappho Says:

    Reference for that 80% figure? Homosexuality as measured how, by whom, in what time and place? It strikes me that what may be going on is a mixing of apples and oranges, and that the entire set of people who passed the genes on likely didn’t have their fitness reduced by nearly that high a percentage. (That’s setting aside the fact that only maybe 15% of the people who carried the genes, even at a load equal to those who are gay, would themselves have been gay, figuring a combination of 30% twin concordance with a set of gay genes that differ for each sex, and so you’d have some people getting the counterbalancing pleiotropic effects without getting the hit.) I’d expect a much lower reduction in descendants than that, for more closeted times and places (though still some noticeable reduction). On the other hand, 80% fewer children for gay men among, say, a particular set of people surveyed in the San Francisco Bay Area during my lifetime, wouldn’t surprise me.

  17. misdreavus Says:

    You’d expect it to be substantially lower, even among men who live in a place like Saudi Arabia.

    being disinclined to screw your own wife is a big problem.