Isn't that just an example of poor sampling/ test hypothesis? Heritability of eyes should be 100% (as predicted) if you include people with genetic blindness.

There obviously are genes for eyes, and they have a causal relationship with humans having eyes. They may have many variations that produce working eyes, but that doesnt mean the relation isnt causal.

Maybe i miss how this relates to your greater point that genetic associations are just coincidental, not causal.

Seems easy enough to test for the case of genetic intelligence. If you find a subset of genes within genetically diverse populations that cross culture, geography, and societies, then you would need a very strong rationale for coincidental association.

It's a practically useless concept outside of extremely narrow fields of biology, where you can control causes or otherwise eliminate genetic variation.

In the vast majority of cases where people use 'heritability', it is useless. Knowing how much a trait varies, or fails to vary, with genes tells you nothing about whether its inherited.

Under stable genetic equilibrium near universally inherited traits will have a heritability of zero; and under cultural, social, geographical etc. selection of mates, 0% inherited traits will be arbitrarily close to 100% heritability.

It's a statistics designed for cases where the experimenter is actually controlling the genetic variance (ie., breeding plants) across populations; it's otherwise meaningless.

Consider a security guard who forces all past shoplifters to wear a tag. Under such an induced causal relationship, wearing a red tag becomes predictive of future criminal behaviour. Otherwise, it's useless.

Likewise, when an experimenter is inducing genetic varience, then it becomes explanatory and predictive; otherwise, it's useless.

> Knowing how much a trait varies, or fails to vary, with genes tells you nothing about whether its inherited.

I am not sure what this is trying to convey. Because I think there's clear cases where stuff is inheritable, and clear cases where stuff isn't, and that's a very meaningful distinction to make from a sociological perspective.

For eg. consider skin color and accent. The former is fully inheritable, the latter is completely uninhabitable.

A Scottish kid raised by Africa foster parents in Nigeria will still look like his biological parents and have their skin complexion, while he will sound entirely like how people sound in Nigeria.

One would be an idiot for trying to force him to have darker skin, or to have a Scottish accent, because we understand that one is genetically controlled, and the other is socially / culturally controlled.

Now in these cases it was as absolute as you can get with these things. But can we not imagine that there's an aspect of genetic influence on many other aspects that can also be somewhat affected culturally?

I don't care what we call it, but can we at least agree that this distinction can exist and might be sociologically salient?

isn't saying that someone's character is inherited from their parents such a claim?

For eg. if you have dark skinned parents, you are likely to be dark skinned yourself.

So 'skin color' is a trait that is being claimed to have heritability.

Of course a scottish accent isnt inheritable. The point is that the heritability statistic is near 100%.

There is no statistics for inheritability. We have no idea, in general, what is inheritable and what isnt. It requires the kinds of experiments which are practically impossible or unethical.

Heritability doesnt mean inheritability. A trait is "heritable" only if it varies with the genes of some population, so a scottish accent is nearly 100% heritable.

Heritability was a statistic invented for when the experimenter was in control of what genes were in what population (for breeding plants). It is meaningless when genes arent being controlled.

Heritability is the idea that if you control which plant has which gene, and the plants are different, then it's probably because of the gene. But its applied to cases where no one is in control of the gene, so it's meaningless.

When I said scottish accents are nearly 100% heritable, that's correct, as a matter of fact. If you compute the heritability stat for scottish accents it'd be nearly 100%.