Selection Runs Vertically Through Time, Not Just Horizontally Across Maps
2026-08-04 · AI-generated insight
Winegard and Carl argue that racial classification tracks something real: visible differences that "reflect their divergent geographical origins." The implicit model is horizontal — populations spread across space, diverging as they go. The ancient DNA study complicates this by adding a vertical axis. Within a single region, West Eurasia, allele frequencies for complex traits shifted by a full standard deviation on the scale of modern variation in just ten millennia. The people who lived where today's Europeans live were, genetically, substantially different from their own descendants.
This means the divergences Winegard and Carl attribute to geography are partly a snapshot of an ongoing temporal process. Two populations can differ not because they went to different places but because selection kept working on them after they arrived. Geography is not destiny so much as a starting condition for locally varying Darwinian pressure — and that pressure has been surprisingly strong and recent, contradicting the older picture of human evolution as mostly finished.
The more pointed rhyme is a caveat the review would do well to absorb. The ancient DNA authors stress that their trait predictions — body fat, schizophrenia, cognition — "were measured in industrialized societies, and it remains unclear how these relate to phenotypes that were adaptive in the past." The genetic signal is real; its phenotypic meaning across time is not transparent. That warning applies with equal force across space. Skull morphology may classify populations at 80% accuracy, but classifiability is a claim about statistical structure, not about what any given allele did in the environment that selected it.