Twenty years ago, I wrote a paper (Turkheimer, Reference Turkheimer2000) that began with the sentence, “The nature–nurture debate is over.” That assertion has sometimes been misinterpreted to mean that one side had been victorious, but that is not what I meant. It's true that I predicated the idea on the results of the twin studies of the era, which showed over and over again that identical twins are more similar than fraternal twins for pretty much everything, from height to how much TV they watch. My conclusion, however, was not that the causally vague genetic “influence” implied by r _MZ > r _DZ actually demonstrated that television watching is in some important sense “genetic.” Instead, my point was that nature has certain methodological advantages over nurture. Genes are relatively discrete and developmentally stable, whereas the environment is amorphous and ever-changing. Genes are not more influential; rather the effects of genes are easier to detect and quantify than the effects of the environment.

This entire field of study was undertaken by Francis Galton, who gave the nature–nurture debate its unfortunate name. The discussion, then as now, was conducted in terms of a number called a heritability coefficient. The quantitative notion of heritability is based on ideas developed by R. A. Fisher, who derived ratios of genotypic over phenotypic variances that expressed, in some eternally fraught sense, the extent to which phenotypic differences could be “attributed” to genotypic differences. Galton and Fisher interpreted their primitive statistics, as more modern scientists interpret their structural equation models, as methods for keeping score in the nature–nurture tournament. The modern twin study era was conducted in this spirit. Summaries of behavior genetics by the twin researchers of the time (e.g., Bouchard, Reference Bouchard2004), fairly bristle with certainty that the estimation of big heritability coefficients proves that genetic differences are the dominant force shaping human differences.

The target article, with its interest in non-genetic forms of generational transmission, allows itself some skepticism about the old heritability concept, but remains committed to an old-fashioned interpretation of it. The authors refer to heritability as both a number (once, unfortunately, as a “score”), and as a biological process in which differences in phenotype are passed from one generation to the next with differing degrees of determination. There is nothing wrong with documenting the vicissitudes of heritability coefficients, but noting developmental and socioeconomic trends in heritability coefficients is one thing, and understanding the causal interplay of biological and cultural processes in the determination of phenotypic differences is another. Although I am skeptical whether the developmental processes underlying human heritability coefficients can ever be worked out causally or quantitatively, I do think that the rich appreciation of genetically and culturally informed human complexity has led somewhere very important.

In the years since Bouchard (Reference Bouchard2004), behavior genetics has been revolutionized by the onset of the human genome project. The roller-coaster of scientific events is by now familiar: linkage analysis didn't find any genes of large effect; candidate gene studies didn't find any genes of medium effect; genome-wide association study (GWAS) didn't find any genes of tiny effect; GWAS was able to estimate human heritabilities without twins, but they were much smaller than twin heritabilities; GWAS was used to sum tiny DNA effects into polygenic scores (PGSs), which were modestly correlated with behavioral phenotypes; those PGSs are able to make genetic discriminations within families in ways that twins are not; in particular they are able to separate genetic effects on parents transmitted environmentally to offspring from genetic effects originating directly in the offspring genome; and all of these analyses are confounded in complex ways by genetic, ethnic, and phenotypic clustering, that is, culture.

But in the newest genomic studies of human behavior, something remarkable has happened. Chastened, perhaps, by the absence of any actionable genes, by the diminishing heritabilities, by the elusiveness of meaningful biology, by the less than practical performance of PGSs, yet spurred on by the deep fractal complexity of modern genomics, the most recent behavior genetic papers include no nature–nurture content whatsoever. Consider, for instance, the recent GWAS of sexual orientation (Ganna et al., Reference Ganna, Verweij, Nivard, Maier, Wedow, Busch and Zietsch2019). A few single-nucleotide polymorphism (SNP) hits were identified, some small heritability was quantified, and on a charitable reading some interesting biological and behavioral pathways were suggested. All of which is fine, and may or may not lead to insight in the long run. But in the meantime, the paper contains absolutely zero Bouchardian litigation of nature–nurture issues as regards sexual orientation.

When I first wrote in these pages (Turkheimer & Gottesman, Reference Turkheimer and Gottesman1991), we were responding to a target article by Plomin and Bergeman (Reference Plomin and Bergeman1991) that, ironically, had a great deal in common with the current one. Where the current article notes that the transmission of differences across generations often occurs along cultural rather than strictly genetic pathways, Plomin and Bergeman (Reference Plomin and Bergeman1991) argued that ostensibly environmental modes of transmission often encompass genetic variance. From the point of view of a social scientist trying to sort it all out, it is the same conclusion viewed through the nature–nurture looking glass. In response, Gottesman and I concluded,

We now know where it led: To the end of nature–nurture as a serious question to be debated by genetically informed social scientists. Good riddance.