Main

Diamond reports Bellwood and Renfrew's argument that major demographic upheavals (‘steamrollers’) at different times in the past 10,000 years, resulting from domestication of certain species of plants and animals, have erased “the products of previous tens of thousands of years of language evolution” in some parts of the world (for example, Europe) more strongly than in others (for example, New Guinea and ‘native’ California). But there is no generally accepted theory of language diversification2,3; indeed, Renfrew and Bellwood's views are strongly contested4,6. Measuring the degree of isolation between human groups and estimating the length of time since their presumed separation (isolation) does not adequately predict either number of languages per unit area or their taxonomic (historical) diversity7.

These are a few of the reasons why we are reluctant to agree that modern New Guinea or aboriginal California can be taken as models for other places and times in history. Nobody knows what the linguistic diversity of Europe at the end of the Pleistocene was like and one can only guess at the linguistic situation in Europe even as late as the early Bronze Age, in the second millennium BC. Were one to accept Diamond's metaphor of the language steamroller as a heuristic device and his argument that speakers of (an unknown number of) early Indo-European languages moved “as a steamroller” over Europe, there is no way to tell whether these hypothesized migrants (or just the Indo-European language) had to flatten a sharply dissected linguistic landscape or rolled over a basically level field.

The solution to the mystery of historical linguistics will come only with better historical data and better sociolinguistic models8.