Genetics, Linguistics, and Prehistory: Thinking Big and Thinking Straight

Article excerpt

The 'new synthesis'

After the Second World War most prehistorians pursued their own discipline without speculating about the languages and genetic make-up of prehistoric populations. The methodological - and perhaps ideological - objections to synthesizing archaeology, philology, and population genetics seemed too great. Gradually, however, the age of the grand syntheses returned. In the 1970s Ammerman and Cavalli-Sforza correlated the spread of Neolithic farming communities northwestwards from Anatolia with the distribution pattern for 27% of 39 selected modern European genes, the so-called 'first principal component' (Ammerman & Cavalli-Sforza 1984:105-6 - revised in Piazza et al. 1995 to cover 26% of 95 genes). In 1987 Renfrew, referring to the non-genetic aspects of their work as well as drawing on earlier work of his own (1973), correlated the spread of Neolithic agriculture with the hypothetical diffusion of the Proto-Indo-European (PIE) language from Anatolia (Renfrew 1987). Two important differences were that Renfrew did not bring in genetic evidence at all in 1987, still treating it with reserve (Renfrew 1989: 145-9), and that Ammerman & Cavalli-Sforza had preferred to equate the 'Indo-Europeans', not with the Neolithic farmers and the genetic 'first principal component', but with the spread of the Battle Axe culture of the 3rd millennium BC which they tentatively correlated with the distribution of their 'third principal component' (corresponding to 11% or 12% of 39 modern European genes - revised in Piazza et al. 1995 to cover 8.8% of 95 genes). This 'third principal component' peaks north of the Black Sea, in the aree of Marija Gimbutas' putative Indo-European homeland (Ammerman & Cavalli-Sforza 1984: 107-8; Cavalli-Sforza et al. 1994: 299-300). Despite these major differences, Renfrew's line of reasoning could be seen as moving along the same lines as Cavalli-Sforza's ideas about the feasibility of a 'Reconstruction of human evolution: bringing together genetic, archaeological, and linguistic data' (Cavalli-Sforza et al. 1988), and it is therefore not surprising that Cavalli-Sforza and his colleagues hailed what they saw as Renfrew's 1987 conversion to the Ammerman/Cavalli-Sforza theory about the spread of Neolithic agrarian populations through a 'wave of advance' as an important support for their own genetic theories (1994: 108, 257,264 - 65,297; but cf. Renfrew 1973). Renfrew himself, in his 1991 Huxley Memorial Lecture on 'Archaeology, Genetics and Linguistic Diversity', came round to endorsing 'the view that the spread of farming to Europe from Anatolia had significant genetic effects', adding 'These may plausibly be correlated with significant linguistic effects, namely the initial spread of a proto-Indo-European language or languages through much of Europe' (Renfrew 1992: 463,472; also 1997: 88).

Renfrew (1997: 82) notes that 'for some time now there have been indications that a new synthesis is emerging between the disciplines of historical linguistics, prehistoric archaeology and molecular genetics'. This synthesis embraces not only accepted language families such as Indo-European, but also super-families which are dismissed by nearly all historical linguists (see Trask 1996; Dixon 1997; Carstairs-McCarthy 1997), such as 'Nostratic' (putative ancestor of PIE and several other families) and 'Amerind'. Thus Cavalli-Sforza's 1988 'Reconstruction of human evolution', mentioned above, links nearly all the languages of the world, as does the 'Worldwide analysis of genetic and linguistic relationships of human populations' of Chen, Sokal, and Ruhlen (1995), while Renfrew's most recent hypothesis about the dispersal of farming populations from the Near East (Renfrew 1997: 87) (given genetic support by Barbujani et al. 1994a)

would place the respective homelands of the Proto-Indo-European, Proto-Afroasiatic, Proto-Elamo-Dravidian and Proto-Altaic languages within a relatively circumscribed Near Eastern heartland. …