The evolution of homosexuality

by Alan Cohen

With this post, I’m going to get back to my roots as an evolutionary biologist, as well as wade into a controversial subject: how could evolution produce homosexuality? (At first glance evolution and homosexuality appear to be contradictory, since lacking a sex drive for the opposite sex should decrease reproductive success.) I am certainly not the first to write or speculate on this topic, and I cannot profess to have read in depth what most other writers are saying. But when the question came up casually the other day, I realized that simple application of evolutionary principles can lead to a fairly nuanced understanding of how this could happen. It’s also a good way to demonstrate the complexity of evolution with a subject likely to pique people’s interest.

First, the caveats and assumptions. This perspective takes no moral stand on homosexuality. Those who know me know that I strongly favor gay rights, but in developing the arguments below the opposite might sometimes seem to be the case. So when I say that something increases “risk of homosexuality” or is “not favored,” I don’t mean it in a normative sense, only in an objective, who-survives sense.

Additionally, I am assuming a biological (though not necessarily genetic) basis for homosexuality. This should not be controversial to anyone with even a rudimentary understanding of biology – many studies have shown differences in hormone levels and brain structures based on homosexuality. I also assume that there is an important (though not exclusive) role for exposure to pre-natal hormones. Studies have shown that male fetuses exposed to low levels of male hormones like testosterone during pregnancy are more likely to become gay. For simplicity’s sake, I consider only the case of male homosexuality, though the same principles can be easily extended to female homosexuality.

I do not believe there is any one explanation for how homosexuality evolves – rather, I believe there are several potential processes, and that some combination of them is responsible. What the relative balance is, I don’t know, but here they are:

1) Lack of precise control over development

Not all individuals turn out the way they are “supposed” to. As much as evolution attempts to standardize the developmental process, things often go wrong. This is partly because environments vary, partly because many processes in development are very sensitive, and partly because genes and gene combinations vary. This is particularly likely to be true for subtle traits like sexual preference, as opposed to things like development of the heart and lungs.

Sexual preference (including “deviant” preferences such as pedophilia, as well as matters of taste, such as tall versus short partners) is determined by many things, likely including genes, pre-natal hormonal environment, early imprinting, and later-life experiences. Other species provide a useful parallel: Many birds and fish also have genetically determined preferences for certain color and feather/scale patterns. Female songbirds are attracted to the songs of males of their species, but only if they are exposed to these songs during a critical developmental period.

This suggests that in fact there are many things that need to go right for sexual preference to become expressed properly. Natural selection may be working hard to ensure this, but the variety of factors involved may be complex enough that things go wrong occasionally. This could result in homosexuality even if it is completely bad for the individual. Under this explanation, homosexuality might arise for similar reason to autism. (Here is where I want to emphasize that this is not a normative judgement – by “things go wrong,” I mean from the perspective of evolution trying to optimize reproductive success, not from the perspective of human morality.)

2) Correlation with other useful personality traits

I don’t want to get into stereotyping, but most of us would agree that there appear to be general personality differences between straight and gay men. This doesn’t mean that all gay men are one way and all straight men another, but it does mean that the probability of having certain personality traits is different between gay and straight men. If the personality traits typical of gay men are good in general, natural selection will favor people with these personality traits. (Yes, natural selection can work on personality, as has been shown in numerous animal studies, including by some of my colleagues here in Sherbrooke.)

Of course, if there were a perfect association between personality and homosexuality and gays had no kids, this wouldn’t work. Everybody with a certain trait would be gay and would have zero reproductive success. But if gays have kids anyway sometimes, or if the personality trait changes the probability of being gay, but not 100%, it would still be possible for natural selection to result in homosexuality as a side effect of good personality traits.

3) Variation in optimal personality traits

The previous argument supposes that the relative value of personality traits is aways the same. This is unlikely to be the case in group-living species such as humans (and the contrary has actually been demonstrated in chipmunks). For example, if everyone in your tribe is highly aggressive and dominant, you may all end up fighting with each other and killing each other, but if no one is dominant, you may end up without a clear leader and unable to defend yourself against other tribes. Thus, a mix of personality types in a population is likely beneficial in many circumstances.

Additionally, the best personality traits may change over time. David Brooks recently wrote that in the modern educational system, agressive, hyperactive boys don’t do as well; this was not the case 100 years ago. What type of personality is best for a chipmunk (exploratory versus conservative) depends on what time of year you are born. What this means is that there could be strong selection for homosexuality-associated personality traits under some conditions.

4) Gay men have some reproductive success

While it is certainly true that being gay is likely to decrease a man’s reproductive success in many contexts, gay men also have some reproductive success. In traditional societies, they were often forced to marry. Many gay men are at least partly bisexual and may have some heterosexual relationships during their lives. And they can help their families, aiding the reproductive success of brothers and sisters, nieces and nephews. All of these factors will decrease the force of natural selection against homosexuality, and allow the other factors above to operate even when they are not so strong.

In my opinion, this explanation is probably the least important so far. There is homosexuality observed in many animal species, including some where males provide little aid to their relatives and where there is no social pressure to reproduce.

5) Correlated traits across sexes (a side effect of sexual conflict)

Some gene copies (alleles) may produce traits that are good in males and bad in females, or vice versa. When this happens, there is evolutionary conflict between the sexes. Sometimes, but not always, this gets resolved by controlling the expression of the genes so that both sexes have it their way. But if, for example, there is a gene for personality traits that are good in women but bad in men, some men may end up with this personality trait just because they get genes from their mother as well as their father. And if these genes are associated with sexual preference…

6) Genetic complexity

We tend to like simplistic explanations of genes: a gene for blue eyes, a gene for being religious, a gene for being aggressive, etc. Most of the time, however, genes have many, complicated effects. There are many versions (alleles) of each gene, and various combinations of different alleles for different genes can have unpredictable effects. For example, some types of hybrids can end up bigger or smaller than either of the parent species.

There are many ways this could result in evolution of homosexuality. Here is one hypothetical example. Gene A has alleles A1 and A2, and having A1 makes you more sociable than A2. Gene B has alleles B1 and B2, and when you have two copies of B1 or two copies of B2, you have a less resistance to atherosclerosis than if you have one copy of each. Gene C has alleles C1 and C2, and C1 gives you lower testosterone levels than C2, which can be good for health and diminish the risk of dangerous behaviour. So the best genotype might be A1A1, B1B2, C1C1, all else equal. But maybe this particular combination dramatically increases the chance of becoming homosexual. Because this combination is somewhat rare and the relevant alleles are good for other things, these alleles will not be selected against and homosexuality will exist as a side-effect.

7) A combination of these factors

The lesson to take from all of this is that the details of evolution are highly complex and depend on many things. In the long run, evolution usually succeeds in having its way: most males are heterosexual, after all. But in the short run, things are messy, and there are lots of ways that evolution can produce unintended consequences and sub-optimal results.

Given that we see some homosexual individuals in many species, my best guess is that explanations 1 and 6 are the most important (though all of them likely play some role). But there are many scenarios in which homosexuality could depend on particular combinations of these explanations, in ways that are mind-bogglingly complex. I’ll give one completely hypothetical example just to scare you off:

Perhaps too much testosterone during pregnancy is bad for women’s health long term, and some women have an allele that decreases this. Normally, lower levels of testosterone might increase the probability of being gay, but when the son also has this allele, there is actually a lower risk of becoming gay because it is linked to other protective alleles. However, the protective alleles can have negative consequences on development if the testosterone level during development is too high, so many men whose mothers don’t have the low-testosterone allele don’t have the protective allele either. When a male fetus without the protective allele is exposed to low testosterone during pregnancy, the probability of becoming gay may still be low, except when the mother has been consuming lots of estrogen-like chemical in her diet (phytoestrogens in soy, for example), which will somewhat augment the chances. But the probability might become very high only when the fetus also has certain alleles predisposing him to certain personality traits that are under strong pressure from natural selection.

This is all hypothetical, but there is good reason to believe that the reality is often this complex. For many things like this, we are unlikely to ever fully understand how the chance of becoming gay changes with every different combination of genes and environmental conditions. An understanding of the general principles that can lead to this complexity may be sufficient, even if our brains long for a simple answer (it’s the Devil!).