[gwern]

'But that's not how how genotypes and phenotypes work for recessive alleles.'

Recessive alleles aren't immune to natural selection. It's just harder to purge them because they get exposed as phenotypes less often. As they become rarer and rarer (it'll manifest as a phenotype as the square of its frequency), selection weakens. Homosexuality, however, is common enough and the fitness penalty large enough that any recessives are exposed to a lot of selection, so it should be disappearing fast, but it isn't, which raises a lot of questions of what exactly causes homosexuality.

[noobermin]

Exactly at the heart of the problem. IANAB (I am a physicist though) and from what I understand, it is a question currently being researched. I mean, great example is if it is so easy, why is there homosexuality in the animal kingdom where they don't have social push for heterosexual mating?

[int_19h]

Yes, there is extensive research on this subject in ethology.

One hypothesis that seems to be fairly strong is that homosexuality is, essentially, just one manifestation of the broad altruism strategy.

Think about what makes altruism in general persist, despite it been seemingly harmful for its carriers long-term: they might not maximize the likelihood of survival of their own progeny, but in sacrificing that, they increase the likelihood of survival of progeny of people from the same family, tribe or larger community - which have a certain amount of shared genes. If you think of the natural selection game as fundamentally centered on the genes rather than their carriers ("selfish gene" etc), this strategy makes perfect sense - your altruism may result in a lot more people surviving and carrying genes shared with you, and down the line, the end result is that population in the future will be genetically more common with you than it has been otherwise - so your genes "win". If so, the tradeoff is in favor of altruism over selfishness, and altruism becomes a selected trait.

Now, homosexuality can be seen as an extreme example of that. Every homosexual pair that adopts children (and there are a lot!) is, essentially, forgoing spreading their genes directly, and spending the effort that would normally go into that on the genes of some other person. Now, in this day and age, children often get adopted across large distances; but historically, and obviously in nature, adoption would be from geographically and genetically close populations; so it's a form of altruism that can be favored by natural selection. If you unwind back even further, before family was a thing, any tribe members that don't have children of their own have more time to spend watching, feeding and protecting others' children.

Now, you might wonder, why homosexuality rather than asexuality, since the latter would produce the same result? Evolution, due to the mechanisms that drive it, generally takes the path of the least resistance at any given point, even if it results in a very long-winded trek long-term. And it's simpler, from an evolutionary perspective, to "neutralize" the sexual drive by changing its target, than it is to switch it off completely. On top of that, there are some benefits other than children to be derived from sexuality in social species - bonding and mutual assistance, facilitating communication (as e.g. bonobos often use sex) etc.

[A_COMPUTER]

Exclusive homosexuality analogous to human homosexual relationships is rare in nature. There are animals that are extremely promiscuous and have sex with anything else regardless of its sex, and there are animals that form homosexual relationships in the absence of enough sexual partners, and revert to heterosexual behavior when partners become available.

[nikdaheratik]

From a scientific standpoint, there's been alot of theory about an Aunt/Uncle advantage which is similar to a Grandparent advantage as far as it being able to help improve survival through the family of any genes responsible even if it is at the cost of an individual.

Of course, there's also the strong possibility that it isn't purely genetic but a mix of certain genes, which may be linked to other important traits, with a certain environment (possibly even in utero).

[toasterlovin]

The problem with the aunt/uncle theory is the math. A non-breeding sibling would have to enable their siblings to have two additional children (above and beyond the children they would have had anyway) for each child that the non-breeding sibling forgoes having themselves for the strategy to even hit break even, from a genetic point of view.

That's a steep hill to climb.

[kilotaras]

Because the number of siblings is close to number of children it actually comes to each breeding sibling to have extra two children survive into adulthood. This becomes even less steep if we remember that cousins can also be included in the math.

For example the gay person becomes a village healer and can spent time on that precisely because he/she doesn't have own children to look after.

[Snargorf]

The hypothesis I put forth (implicitly) is that homosexual genes are still around because of the massive social pressure to have children and to never admit being gay.

Now that this pressure is gone, we could see a huge drop in these traits over the next few centuries.

[gwern]

I don't know how that is supposed to work. What sort of social pressure could possibly produce near-zero fitness impact when it makes you not want to have sex with women, in an environment where the men who do want to have sex with women are like 95% of the population and will happily put enormous efforts into doing that and taking your place? In addition, as I said, given the apparent fitness impact and frequent childlessness of gay men, we should be able to see homosexuality rates plummeting over the past century, but as far as I know, there's no evidence. (Plus the wild animal thing, yeah.) There's something stranger going on. The 'gay germ' hypothesis, as unpopular as it may be, at least adequately reconciles all of the observations.

[toasterlovin]

Maybe you're misunderstanding the parent (or I'm misunderstanding you!). I believe the parent is saying that the genes which predispose somebody to being gay remain in the population, despite their seemingly zero reproductive fitness, because gay people face enormous pressure from their relatives to not be gay and reproduce. And they face this pressure because their family members have a genetic interest in the gay person's offspring and so have evolved to coerce gay family members into having children.

[gwern]

I understand what they're saying. My point is that it's highly unlikely: coercion just doesn't work that well when it comes to something extremely expensive and requiring a lifetime of work to maybe succeed, when the task is not just not pleasant but outright aversive. It's difficult to see how such coercion could be so extraordinarily successful as to reduce the average fitness penalty to something so tiny that homosexuality could still exist at current frequencies like 5%; it'd be like you'd have to have such pervasive and super-effective coercion that not a single gay man out of 100 fails to reach his quota, and this would have to obtain in all societies forever, effectively, or else eventually the coercion would slacken and the homosexuality genes would almost immediately vanish, permanently. This is pencil-balancing-on-its-point-for-millennia territory.

[toasterlovin]

Right, I agree that if the genes which predispose somebody to being gay had no benefits whatsoever, then they could not even exist; they would have a purely negative effect on fitness. So they must have some positive effect on genetic fitness to even exist at all. And they must be relatively significant and intractable from the side effect that some of their carriers turn out gay, given how common gay people are.

But that is separate from whether or not pressure from relatives to reproduce can be used as a mitigating strategy for the cases where these genes do result in someone being gay. And this strategy, then, would lower the amount of fitness those genes would need to provide to be viable.

[gwern]

"So they must have some positive effect on genetic fitness to even exist at all."

That's not true. There's a number of ways variants can exist without being fit. They can be regularly created by mutation, or they can increase due to genetic drift, especially in a bottleneck scenario. And there's even more ways that a phenotype can persist while being highly unfit - if it's a side-effect of a co-evolving pathogen, being one of them, as then the human natural selection is constantly fighting it but the pathogen easily evolves even faster.

[JoeAltmaier]

Many animal populations have non-reproductive members. It takes a village to raise a child. In such an extremely social animal as Homo Sapiens, many of our adaptations are social. E.g. it only takes a few men to guarantee another generation, yet 50% men are born in each. Another: we don't die the instant we give birth; we're there (and non-reproductive) for decades afterward, competing for resources. Its the social benefits that allow for this.

[makomk]

There's something called Fisher's principle that explains why the sex ratio of most species is 50:50. In short, if less than 50% of humans are men, then men must have more children on average than women and it becomes reproductively advantageous to have more make kids in order to take advantage of it.

[JoeAltmaier]

Still, in many animal populations the ratio 1:1 doesn't hold. Something else is going on then.

[internaut]

I think it's not quite 1:1 in humans, the males have a slight lead.

http://www.scientificamerican.com/article/is-a-pregnant-woma...

[int_19h]

Your hypothesis assumes that, historically, all human societies had a massive social pressure against homosexuality. But it contradicts what we know about our history - there were numerous societies that treated it ambivalently (e.g. China, Japan, Native Americans) - and it didn't seem to produce any noticeable effect.