Tuesday, October 20, 2009

Language(s), then and now   posted by Razib @ 10/20/2009 02:31:00 PM

One of the trends which L. L. Cavalli-Sforza pointed to years ago is that linguistic and genetic cladograms exhibit a great deal of similarity. More recently geneticists such as Marcus Feldman have suggested that the reason behind this is that people tend to marry those who they can communicate with. Once the genetic data becomes more granular in Europe it will be interesting to see if this works on the national level. Though it is true that languages such as French or German are relatively new standards within the boundaries of their nation state, it is also true that people from village to village were intelligible with each other, until presumably they hit upon a major linguistic boundary such as between Romance and Germanic languages (though presumably here bilingualism would have been common).

But one of the problems which crops up now and then trying to think of the past, especially with all the new results coming out of historical genetics, is that we project the linguistic homogeneity of the present back. For example, in ancient Mesopotamia there were two languages which are linguistic isolates to the best of our knowledge, Sumerian and Elamite. If we didn't have extant written records, or, more accurately if writing hadn't arrived in Mesopotamia so early, we'd have no idea that there were non-Semitic languages in that region today. Not only that, but the ancient records from that period in Mesopotamia hint at other language groups which were not literate and so left no record. In the Roman Empire the vast majority of our extant records are in Greek and Latin. We know the existence of other languages, such as Celtic dialects, Punic, and such, but there are also indications of the flourishing of local lingua franca such as "Iberian" in southern Spain which have left only a marginal literary impact. The emergence of large polities led to linguistic homogenization; I have read that the dominance of Quechua, the language of the Incas of Cuzco and its environs, actually dates to the Spanish colonial period. In the chaos after the fall of the Inca Empire the language of the Incas gave the local peoples are a sort of solidity.

I'm thinking about this because I am corresponding with some people about the David Reich paper on Indian genetics, and people are asking questions such as "did the ancestral South Indians speak Dravidian languages?" I doubt it, but I also doubt that the language families of ancient India were as cleanly sorted out as they are today. Even the languages of the Andaman Islanders may not be one family! Over the past 10,000 years, and especially the last 2,000 years, much of the World Island has seen cultural positive feedback loops which have resulted in linguistic homogeneity. The rise of nation-states, literate elites, class stratification, etc., are I think part of the whole package which has given rise to this dynamic. Some of these are general phenomenon which presumably apply to H. sapiens 30,000 years ago, but I suspect the powerful magnitude is a new feature.