ABSTRACT

The disciplines of linguistics, archaeology, anthropology and genetics each have a bearing on the reconstruction of human prehistory, and it is hardly surprising that in their attempts to match findings across disciplinary boundaries the practitioners of one discipline sometimes do violence to the methodology of another. Historical linguists commonly present their findings in the shape of ‘family trees’, the very simplicity of which lends them to misinterpretation. Cavalli-Sforza et al. (1988) attempt to correlate trees of worldwide prehistory drawn on the basis of genetic and linguistic data. Quite apart from the content of their linguistic tree, which represents the position of extreme ‘lumper’ linguists, its use is questionable because it assumes that the kinds of continuity represented by trees in genetics and linguistics are qualitatively similar. The basis of this assumption is that ethnic groups, of which genetic markers and language are taken to be manifestations, have diachronic stability and speciation mechanisms of a kind which can be appropriately depicted as a tree.1