This is not a scientific work.

Behavioral genetics may be science or it may not be, and some of the statements in this book may or may not be true---I honestly do not know enough about the field to say---but there is no sense of scientific inquiry in this book. There is no sense of doubting, which is the essence of science. In fact, the behavioral psychologists have a way of writing which offends my delicate physicist's sensibilities. I'll tell you what I mean: Turkheimer, Harden's supervisor, in a commonly cited paper in the field and this book, posits Three Laws of Behavioral Genetics. Like Newton, you see. The gall of these people. Laws? "The nature-nurture debate is over. The bottom line is that everything is heritable... All human behavioral traits are heritable... The point is that now that the empirical facts are in and no longer a matter of serious debate, it's time to turn attention to what [it] means." If you are an academic, ask yourself whether this kind of pronouncement would be acceptable in your field, or whether you'd get laughed out of the room.

There's a lot of good salesmanship here: Another technique, which is not coincidentally also common to evangelical proselytizing, is convincing the reader that they already know what Harden is telling them. Mothers of multiple children, she claims, have a good sense of the heritability of various traits. People in general, one of the graphs shows, do a good job of estimate the amount of heritability for various traits, from psychological to sexual to pathological. But how? People have sequencers in their eyeballs? If it takes this enormous machinery, millions of genomes, and hundreds of authors to tease out small effects, what the hell is this folksy stuff supposed to mean? How a scientist can write this kind of thing, I don't know.

In the first section, Harden goes over the various methodological techniques that her field depend on, the main attraction being the Genome Wide Association Study, which the book takes pains to tell you is pronounced Gee-Was, saving its readers the embaressment of calling it something dumb like Guh-wuss or Gee-Double-U-A-Ess in front of their friends.

GWAS is the method that produces all these nifty "polygenic indices," and on whose authority the majority of the book's conclusions rests. After decades of chasing false leads about individual genes contributing to intelligence, as Harden herself points out, the field was rescued by "big data,"one of the 21st-century's biggest scams. Throw about 1,000,000 genomes and associated traits into a big black box that finds correlations, and you get polygenic scores. Wouldn't you know it, it correlates which just about any darn thing you want it to correlate with. Very good for publishing and academic careers! ``Everything is heritable,'' as they say, which translates to ``everything is publishable.'' And indeed, close to everything has been: Polygenic index papers have become a small industry.

Here is a telling example: Height is supposed to be the canonical polygenic trait, and Harden indeed mentions height and the GWAS studies involved. Yet there is no mention of the series of papers that were published in 2019 that found that previous polygenic scores associated with height were overestimated, and the polygenic signal was gone. Previously undetected levels of stratification were partly at fault---The method inflates even very small errors. Another big source of error was... a computer bug, which was fortunate enough to produce the results the researchers were looking for.

I'm not suggesting height isn't polygenic; it almost certainly is. I am suggesting that this method is not as robust as Harden would have you think. Yes, Harden addresses the method's most well-known (and honestly, convenient) shortcoming: Its results so far only apply to those with recent European descent; the method requires homogenous populations, so results cannot be easily ``ported'' to other populations. But otherwise, there is no sense that this method is anything other than perfect. There is no sense that there are problems with the methodology, or that something less than total certainty in its power is warranted. I've read several of the papers cited in this book, some of which count Harden herself as co-author, and they tend to be much more sanguine about the power of their method.

You see, I'm not an expert on GWAS; it's not my field. I also suspect that many who claim to be experts in it, and who use its results, don't really understand it either. But I am enough of a scientist to be left deeply skeptical of its claims. That skepticism is my right, as it should be everyone's. Hell, not "right"---duty. And it should be this book's responsibility to address and share that skepticism, even to nourish it. This book fails to do that.

In fact, it tends to do something rather ugly instead. It tries to psychoanalyse why you may not agree with it. I found this a particularly bizarre aspect of the book. Harden tacitly dismisses the idea that there could be healthy scientific skepticism of this method that is independent of its moral conclusions. No: You don't agree with it either because you are clinging to the idea of meritocracy or because you want to believe it, but you don't want to be associate with Bad Men like Murray and Harris. I don't find this kind of rhetorical device very impressive in a "scientific" text.

So we come to the moral aspect of the book, its second and weakest part. Harden produces three positions, in analogy with the current thinking of how we should deal with race: Eugenic, Genome-Blind, and Anti-Eugenic. If you remember your three little bears, you know how this goes: Too much genetics, not enough, and juuuuust right.

But the Anti-Eugenic recommendations are pretty weak. There's one example of where polygenic scores might help, essentially showing that abstinence education is pointless, and that's pretty much it. There's a truly bizarre section which criticizes GINA, a law that prohibits discrimination based on genetic information as being the equivalent of "race-blind" policies. Is tracking down every last bit of variance in social experiments important for academics in those fields? Sure, why not. Is it as important for society? I don't see it, and I think the book makes a weak case for it.

Rating: 1 turd,