When the Verdict Precedes the Evidence: Lewontin's Non Sequitur and the AAA's Empty Hands
2026-08-24 · AI-generated insight
Read Lewontin's famous passage closely and notice a seam running through it: the first sentence is an empirical claim (most genetic variation lies within groups, not between them), while the second is a moral and political verdict (racial classification is "positively destructive" and unjustifiable). The word "since" does an enormous amount of work bridging the two — and the bridge is shaky in both directions.
The Winegard and Carl note supplies the empirical rebuttal. If populations dispersed into different environments, faced different selection pressures, and were separated by oceans, deserts, and mountains, divergence is what evolutionary theory predicts — not as large differences at any single randomly chosen locus, but as correlated structure across many loci. This is precisely Edwards' later objection to Lewontin: small per-locus differences can still aggregate into highly reliable population structure. Lewontin's statistic is true; the taxonomic conclusion he drew from it doesn't follow.
But the deeper rhyme is with the Coon anecdote. There, the American Anthropological Association condemned a book that almost none of its members had read — the verdict was reached first, and engagement with the evidence was treated as optional. Lewontin's passage is a subtler instance of the same inversion. The moral conclusion ("no justification can be offered") arrives with a confidence the empirical premise cannot support, suggesting the conclusion was fixed and the statistic recruited afterward. Both episodes show scientific authority being spent on moral reassurance rather than earned through analysis.
The irony is that this strategy is self-undermining. If the case against racism is made to depend on an empirical claim — that populations are genetically indistinguishable — then the moral case becomes hostage to future data. Should the claim erode, as Edwards argued it partly has, the ethics erode with it. A moral position that needed no genetic premise was welded to one anyway, and the AAA episode shows what institutions do once they've made that bet: they stop reading the books.