The standard explanation for Neanderthal DNA in modern human genomes has been demographic: small populations meeting at territory boundaries, hybrid incompatibilities filtering the results, genetic drift doing the rest. The interbreeding happened because the populations overlapped. The asymmetries in the signal — more Neanderthal DNA on autosomes than on X chromosomes — were attributed to differential migration, population structure, and the well-documented fragility of hybrid X chromosomes.
Platt, Harris, and Tishkoff (Science 2026) analyzed the X chromosomes of Neanderthals themselves, not just modern humans, and found a 62 percent excess of modern human ancestry. The modeling shows that pure demographic processes cannot account for this asymmetry. The best-fitting explanation is mate preference: interbreeding was predominantly between Neanderthal males and modern human females, and hybrid females with more human ancestry were preferred as mates.
The hidden variable was preference itself.
Demography predicts symmetric patterns with noise. Selection predicts asymmetric patterns with structure. The X chromosome excess in Neanderthals has structure — it is consistently biased in one direction, consistently sex-biased, and the magnitude is too large for drift. When you model what demographic processes can produce and compare it to what you observe, there is a residual. The residual has the shape of choice.
This does not tell us anything about the psychology of the choice. It tells us that the genomic signal is incompatible with the null hypothesis that interbreeding was random with respect to sex and ancestry. Something was non-random. The simplest model that fits the data is preference. Whether that preference was cultural, aesthetic, or something else is invisible to the genome. The genome records only that the mating pattern was not what chance would produce.