A Proposal at the Edge of Method

Nostratic proposes that Indo-European, Afro-Asiatic, Uralic, Dravidian, Kartvelian and Altaic all descend from a single ancestor spoken perhaps twelve to fifteen thousand years ago. It was developed in the Soviet Union in the 1960s by Vladislav Illich-Svitych and Aharon Dolgopolsky, and has been pursued since by a school in Moscow and, on a somewhat different basis, by Allan Bomhard working in the United States. Its advocates have published reconstructed roots and proposed sound correspondences, which is more than most long-range proposals offer. The majority position rejects it, and the objection is methodological rather than a matter of taste.

The comparative method works by identifying regular sound correspondences in basic vocabulary, and its resolution decays with time, because sounds keep changing and words keep being replaced. Somewhere between eight and ten thousand years the signal falls below the level at which chance resemblance between unrelated languages produces the same number of apparent matches. Critics hold that Nostratic operates past that horizon, and that its cognate sets are indistinguishable from noise. Whether a deeper relationship exists is a separate question from whether it can be demonstrated.