Saturday, July 16, 2005

Language, genes, etc.   posted by Razib @ 7/16/2005 09:47:00 AM

Luigi Cavalli-Sforza's magnum opus The History and Geography of Genes opened the flood gates in terms of a series of popular books which attempted to glean the movements of peoples by the examination of data from their genes (of late, usually matrilineally transmitted mitochondrial DNA and patrilineally transmitted Y chromosomal sequences). Of course, there is the ever present problem that different genes might have different histories, that selection confounds our inferences, and many of the models make gross simplifying assumptions (ie; no admixture, etc.). In Beyond languagese I highlighted the issues that crop up with the nomenclature used when referring to various language groups, as the names themselves often confuse those who attempt to interpret them transparently.1 The problem with language also crops up in historical genetics, as people want to revert back to the early 20th century tendency to fix on to a particular language group a distinctive genetic history. Cavalli-Sforza is I think responsible for this tendency, because many people were drawn to the section of his book where he addresses this question, and illustrates the concordance between gene and language trees with graphs.

So, getting to the point, there has been some recent work that indicates that the peoples of Madagascar are predominantly Asian, specifically derived from Borneo, on their male genetic line (Y) and African on their female genetic line (mtDNA). The authors of the paper make a big deal about this, but 1) I thought this was assumed from previous studies, if not as definitively, 2) you can find similar asymmetries across the New World. One issue that the authors find surprising is the contrast with the genetic data as far as the language as Malagasy seems almost wholly Asian. Well, again I think the New World example should indicate to us that this can happen, and has happened many a time in the past (non-Indo-European substratum are strewn across Europe via place names and other localized terms). And today I stumble upon this:

...Wood and colleagues tested 40 populations across the continent for associations between genetic, linguistic and geographical distances. Examining Y chromosomes and mtDNA revealed that gene flow patterns have been different for men and women historically: language appears to be passed from father to child, and women may have mixed between populations more than men....

I doubt we will ever find a One True General Answer to these sorts of question, though we may begin to establish an overall preponderance and a knowledge of local patterns. I for one am really curious as to how the languae of the Magyars managed to establish itself as dominant upon the Danube plain (there seems little genetic differentiation from neighboring populations, though that might be due to high levels of genetic exchange after Christianization circa 1000, and so an obscuring the original admixture dynamics).

Related: "Racial Diversity."

1 - One time I mentioned to a friend that 40% of the words in Greek (like the ones with ending in nth) seem to be "Mediterranean." By this I meant a catchall term I have seen used to simply clump all the pre-Indo-European dialects that dominated Greece before the arrival of Greek speakers. But my friend interpreted the term directly and responded, "Well, that makes sense, since Greeks are 'Mediterranean' and Greek is a 'Mediterranean' language."