Jay Joseph on My Book

Am I not skeptical enough?

One of the ongoing traditions in the nature-nurture debate is absolute rejection of the idea that genes have anything at all to do with human behavior. Back in the day there was Lewontin, Rose, and Kamin’s Not in Our Genes, for example. This point of view, I have always thought, was often more a reaction against hereditarian overreach than an accurate representation of what is really going on with genes and human behavior. (I wrote a little bit about Not In Our Genes here, for example) It would be hard to read my book, I would think, and come away with the impression that I am generally supportive of the behavioral human heritability project and its use for understanding the causes of human differences, but that is the way Jay reads it.

Jay Joseph has been carrying on the absolute rejection tradition in recent years. I will confess that I have always found the absolutist position frustrating, because I am so close to agreeing with them. My frustration led me to write a too-unkind review of Jay’s most recent book. On the bottom line, I think you can have what the absolutists want— rejection of Jensenism, The Bell Curve, over-biologized psychiatry— without turning away from a set of basic facts that need to be reinterpreted, not rejected.

I am analyzing data from the Louisville Twin Study these days. The IQ scores of middle aged MZ twins are correlated in the neighborhood of .8. The DZ twin correlation is just about half of that. Why would that happen? Jay’s hobbyhorse is the equal environments assumption, one of those topics about which we agree in many ways, except that I am not able to be sufficiently absolute. Are you really ready to think that the only reason MZ twins are so similar is that their parents dressed them the same? (I agree, and have said many times, that the EEA is a very good reason not to obsess over whether the heritability of something is .4 or .6. Note also that the EEA doesn’t apply to animal work, because equal environments can be imposed experimentally.) In the book, I put a ton of effort into explaining how two things can be true at the same time: (1) genes “influence” behavioral outcomes in a general way that is especially manifest in identical twins; but (2) there are no genes with anything resembling specific causal pathways linking them to normal-range human behavioral differences.

One question I often ask absolutists is this: suppose you are adopting a child at birth, and are informed that both of the child’s biological parents are severe alcoholics. Do you have any reason to be concerned? I don’t mean the ethical part— certainly it is a wonderful thing to adopt an at risk child. But if your answer is yes, there is reason to be concerned, is any part of that concern genetic? I know there are also non-genetic reasons to be concerned— epigenetics, perinatal environment, a million others. But is there any net genetic effect reflected in the parents genotype that might lead you to expect a heightened risk of alcoholism in the child?

This brings me back to the paradox I sketched above. The phenotypes of the child’s first-degree relatives (or identical twin if we happen to have one) are a way of agglomerating tens of thousands of tiny, non-systematic genetic “influences” (the non-specific term of art) to produce a meaningful prediction of what the adopted child’s phenotype might eventually be. Yet, at the same time, without the relatives, given only the DNA sequence of the potentially adopted child, we have almost no information at all. A polygenic score for alcohol risk in the child would almost certainly not account for as much as 1% of the alcohol-use variance, nowhere near enough for adoptive parents to use for decision making.

This paradox is at the root of the missing heritability problem. It is the great unsolved problem of human behavior genetics. Close relatives give us probabilistic information about the likelihood of phenotypic outcomes, and there is plenty or reason to think that some of that information arises in the genotype. Yet the only way we know how to make use of that information is in the form of the organically grown phenotypes. (See my thought experiment of making predictions using frozen cloned embryos on page S37 here) At the same time, printed out as a DNA sequence genomic information is close to useless for predicting human behavior. My book is an extended attempt to resolve this paradox. Jay may not agree with my solution, but I insist it is necessary to take the problem seriously instead of dismissing it out of hand.