The idea that linguistic and genetic diversity within the human population are correlated has received strong interest in genetics and archaeology, but has been generally ignored or even derided from within linguistics. There are several reasons for this. The linguistic classifications used in the original research did not represent an accepted characterisation of world linguistic diversity, but rather a geographical ordering of heterogeneous loose groups of languages, many of which were defined by mere contiguity, rather than convincing linguistic reconstruction. Thus the apparent congruence of linguistic and genetic trees represented nothing more than the fact that both sides of the diagram were ordered by geographical proximity.
Current scholarship suggests that approximately 250 deep language families can be identified before the phylogeny of world languages becomes effectively star-shaped. These families are unevenly distributed across the globe, with fewer than 20 in Africa, but as many as 150 in the Americas. This is a paradoxical result, in view of the respective ages of those two populations, and represents the exact opposite of the pattern observed in several genetic systems. I argue that the pattern can be explained by considering the different rates of change and modes of transmission of genes on the one hand and languages on the other. Languages diverge very fast, which means continents like the Americas can harbour great diversity within a few thousand years of settlement. However, languages are greatly subject to extinction and homogenisation, because of social conformity effects within their largely horizontal inheritance system. Thus diversity in language has been more completely erased than genetic diversity by the agricultural and political expansions of the Holocene. These homogenising processes affected Africa and Eurasia disproportionately. Thus, the oldest continent is today the least diverse, and youngest the most diverse, in linguistic terms.
Copyright. All Rights Reserved.