The Ubiquity Problem for Group Differences in Behavior

The Ubiquity Problem for Group Differences in Behavior

The main point of what I have written about race and intelligence in Vox and elsewhere concerns a misleading intuition about heritability and group differences—individual differences are uncontroversially heritable, why shouldn’t group differences be the same? This is exactly the argument that was adopted by Sam Harris and Charles Murray. Richard Haier recently called it the,

Default Hypothesis—whatever the factors are that influence individual differences in IQ, the same factors would influence average group differences. Since there is overwhelming evidence that genes influence the former, it would not be unreasonable to hypothesize that genes at least partially influence group differences. 

My previous post discussed one reason the default hypothesis is wrong—it actually has no meaning, because there is no clear empirical definition of what it would mean for a group difference to be heritable. As usual, Haier doesn’t even attempt to formulate one. Here I want to discuss another reason that the Haier’s Default Hypothesis (HDH) is incorrect—the intuition applies not only to IQ in blacks and whites, but to every difference in average behavior between every group of people that we could possibly imagine.

Everything is heritable. I have spent the last 20 years trying to talk people out of the intuition that there is something special about intelligence (or schizophrenia, or whatever other behavioral construct you are trying to explain) because rMZ > rDZ. Marital status, to trot out my standard example, is heritable. It’s important to bear in mind that marital status is heritable in exactly the same way that intelligence is. One of the ways that mainstream BG has responded to my arguments following from the fact that all twin studies come out basically the same way has been to move the discussion away from twin studies to new molecular approaches like GWAS and GCTA, which have the advantage of having been conducted for IQ but not for divorce. “Well sure rMZ > rDZ for marital status,” the (unspoken) argument goes, “but for IQ we have GWAS hits! We have SNP heritabilities! That’s really heritable, the real actual biological thing!”

But the fact of the matter is that the only reason we don’t have GWAS hits for divorce is that no one bothers to look for them. Why else would it be heritable? Individual differences are heritable because there is a correlation (important point to be followed up later: not necessarily a causal relationship) between DNA variation and phenotypic variation. So there are GWAS hits, GCTA heritabilities, whatever you want, for everything from IQ to schizophrenia to the five factor model to how much TV you watch to your political persuasion to your marital status. Everything, no exceptions. This fact has a lot of implications for understanding the biological status of GWAS hits, but that isn’t exactly where I want to go right now. For now, we are just observing that the first term of Haier’s Default Hypothesis is true for everything. It doesn’t buy us any insight at all.

Now for the group differences part. The preferred hereditarian view of race nowadays is that any and all geographic or ethnic groups of people count as race, because they are in part the result of real ancestral (on a recent time scale) or evolutionary (on a longer one) processes. I don’t think “race” is a very good word for where you end up from this definition, but I don’t object to the overall idea. Yes, Lithuanians differ genetically from Koreans, and if a person thinks they are of Lithuanian rather than Korean ancestry their DNA will almost certainly bear them out. But one consequence of this is that the number of possible groups we can be comparing is almost infinite. We are so used to worrying about black and white Americans that it seems like an especially natural category to us, but what about Koreans and Lithuanians? Irish-Americans and Jewish Americans? If we are going to undertake the task of trying to understand whether differences between groups are genetic or environmental, we have an enormous number of groups to consider.

And all of these groups differ in behaviors in thousands of ways. Think of all the ways that the average behavior of Koreans differs from the average behavior of Lithuanians. Differences is FFM profiles (which are individually heritable), differences in divorce rates (which are individually heritable). Genetic or environmental? So the HDH buys you literally nothing: there are thousands of heritable individual differences that differ among ancestral groups. Does the HDH imply that they are all group-heritable (which again, is something that needs to be defined), that all of the behavioral differences between Koreans and Lithuanians are the results of some kind of genetic “tuning” (Sam Harris) that dictates that they have to be that way? Or, if that isn’t the case, is there some meaningful scientific methodology for working through the thousands of group differences in individually heritable behaviors, carefully sorting them into more or less genetic piles? Name me one single complex (ie, not sickle-cell) behavioral (ie, not height) trait for which something like this has ever been accomplished.



Eric Turkheimer

Eric Turkheimer is the Project Leader for the Genetics and Human Agency Project. Eric is a clinical psychologist and behavioral geneticist. For thirty years he has been involved in empirical and theoretical investigations of the implications of genetics for the genesis of complex human behavior. Current projects include understanding the interaction between socioeconomic status and the heritability of intelligence, and philosophical analysis of the ethical status of work that purports to demonstrate biologically based differences in behavior among racial groups.

No Comments