I certainly share enthusiasm for the power of analysis allowed by DNA markers, including of course mtDNA and Y chromosome - especially the latter. One should not neglect, however, the information already available from "classical" markers (proteins and blood groups), which have the advantage that there exists already much information about many populations and many genes, thus allowing a synthetic study of many genes at a geographic level. This is still hard to do with DNA markers, given the lack of information and tendency of most researchers to use different genes. Our analysis of Asian data (see chapter 4 of History and Geography of Human Genes by Menozzi, Piazza and myself) shows four major clusters: Southeast Asians, East Asians (including Mongols and Siberian Uralic speakers), Northeast Asians (showing the greatest similarity to American Natives, of which they are the closest geographic and genetic neighbors), and an heterogeneous group of extra-European Caucasoids, from people of the Indian subcontinent to Iranians, Arabs, and other Middle Eastern groups. Many Central Asians fall in between this group and East Asians. They show a great variation of admixture between these groups, given the number of migrations and major expansions in the region. Some are known historically, like that of Turkic groups, expanding into Anatolia in the 11th century AD and continuing until the 18th century, and earlier that of Mongols, beginning in the 3rd century BC, from an area slightly to the east of the Turkic expansion, but proceeding in all directions and incorporating Turkic groups as well. Before this time, there were expansions of Europeoid peoples (most probably Tocharian speakers), recently attested archaeologically, which may have started around 3800 years ago. Perhaps 1000 or more years earlier, Indo-European speakers from the oases south of the Urals, north of the Black Sea and in western Kazakstan moved in western and in eastern directions; the latter probably moving predominantly to the south, mixing with (presumed) Dravidian speakers of the region including Iran, India and Pakistan.
A few linguistic isolates, according to Ruhlen and Greenberg, survive in refugia - especially in the mountains (Karakorum and Caucasus) and along the Yenisei River in Siberia, and may have a remote relationshipo with other old Eurasian languages. According to Ivanov and other Russian linguists, these may include Sumerian and some European languages such as Basque, and perhaps Etruscan, but also larger surviving families like Sino-Tibetan and American NaDene. Starostin has suggested a very old Eurasian family of which there remain the traces just mentioned, later largely replaced by the expansion of familes belonging mostly to the Nostratic superfamily. It is possible that genetic analysis may help to support some of the hypothetical, remote relationships among these few, highly scattered surviving languages.
Principal component maps show in synthesis: a general gradient from southwest to east is the dominant component, and must reflect the many expansions east to west and west to east, which mixed genotypes that were originallly very different, having been isolated for a time in East Asia, and in West Asia/Europe respectively. The first principal component of such a large continent can hardly resolve the differences in degree of admixture which are found in different populations of Central Asia, like the Uighur, the Kazak and many others. The next component of genetic variation is a north- south, rather regular gradient, which correlates well with anthropometric traits strongly correlated with climate, thought to represent climatic adaptations - size, weight, limb length, etc. Asia has extreme temperatures, among the coldest of the world in Siberia, and very hot ones in the south, and there may well have been biological adaptations to these extreme conditions.
Lower principal components indicate more highly clustered vatiation: a possible expansion from the Japan-Korea region, perhaps toward the northeast - and America? (3rd component). The 4th component shows a strong opposition between an African (?) influence in S.W. Arabia and a Caucasoid one in north India; the 5th is probably a consequence of the demographic expansion of the Turkic peoples, which may be very difficult to distinguish from the adjacent and genetically related expansion of Mongols. New genetic data may help separating the two expansions.
© 1998 by L. Luca Cavalli-Sforza
