Recency Does Not Guarantee Triviality

Gould builds his case against biological race on a temporal argument: because races arose recently, the genetic differences between human groups must be minor, and the visible ones (skin color, hair form) are "literally" skin deep. The unstated premise is that a short span of time cannot do much evolutionary work—recency licenses the inference of triviality.

The ancient DNA study attacks precisely that premise, though from an angle Gould never anticipated. Over "the past ten millennia"—a window even shorter than the one Gould invokes—the authors find "many hundreds of alleles" reshaped by strong directional selection, including shifts of a full standard deviation in polygenic predictors of body fat, schizophrenia, and cognitive performance. In other words, recent time turns out to be ample for pervasive, powerful selection on complex traits, not just on cosmetic surface markers.

This is a real tension worth sitting with. Gould's rhetoric leans on an old population-genetics intuition—hard sweeps are rare, most variation is neutral—and the paper itself concedes classic sweeps were indeed uncommon "over the broad span of human evolution." But it then documents a different mode: subtle, polygenic, sustained frequency shifts that classic sweep-detection methods missed entirely. Gould could plausibly have been right about the mechanisms available to 20th-century detection and still wrong about the conclusion, because the relevant selection was invisible to those tools.

Two cautions keep this from collapsing into a simple refutation. First, the paper is explicit that its trait effects "were measured in industrialized societies" and it "remains unclear how these relate to phenotypes that were adaptive in the past"—the polygenic scores are correlational scaffolding, not a claim about what selection was optimizing. Second, the study covers West Eurasia through time, not differences between contemporary continental groups, which is Gould's actual target. So the finding does not resurrect the ranking of races Gould was rightly demolishing.

What it does dismantle is the specific move from recent to minor. Selection is faster and more thorough than Gould's argument assumes. The honest reading is that biological race can be a poor scientific category and human populations can be substantially, recently shaped by selection—both at once. The comforting syllogism, at least, no longer holds.