Diamond Wins the Question Winegard Refuses to Ask
2026-08-05 · AI-generated insight
Place these two side by side and you find them arguing about different things while appearing to argue about the same thing — which is exactly what makes the pairing illuminating.
Winegard and Carl land a sharp methodological blow against the Lewontin tradition: looking at one genetic locus at a time is like sorting faces by nose size alone. Variation within groups swamps variation between them at any single trait, but correlations across many traits produce clusters you can classify with startling accuracy. Statistically, this is correct. Human populations do carry a real, multivariate structure. So the point stands: "race" is not pure fiction at the level of genetic correlation.
But notice what this argument establishes and what it quietly leaves untouched. It shows that populations are distinguishable. It says nothing about whether those distinguishable differences explain the outcomes we actually care about — who conquered whom, who got writing, who got steel. That is Diamond's entire question, and his answer is that history's asymmetries trace to environments, not to the genomes of the peoples in them.
Here is the synthesis the notes don't state: you can grant Winegard and Carl their multivariate clustering in full and it does no work for the conclusions race science reaches for. Being able to guess someone's ancestry from correlated markers is a claim about detectability. Claiming ancestry causes the fate of civilizations is a claim about causal potency. The face analogy is revealing precisely because it's honest about this — knowing a face is female tells you nothing about what that person will achieve.
So the correlation argument, valid as far as it goes, functions as a kind of rhetorical bait-and-switch when deployed against Diamond. It proves the trivial thing (groups differ, and the differences are correlated) and lets the reader import the momentous thing (the differences matter for outcomes). Diamond's geography-as-destiny thesis is the missing rebuttal filed one note over: even if every human population were biologically identical, the maize and the horses and the mountain ranges would still have produced the same lopsided history.