Tuesday, April 03, 2018

A comment from @bennedose

From here (copied in full):

This paper has some of the following issues which I find interesting. I believe that geneticists are mistaken when any good work that they do is mixed with the unscientific hypotheses about language spread that linguists have established for the last 150 years. Genes do not code for language. Genetic fact should not be mixed up with tenuous linguistic hypotheses.

1.The linguistics dates of IE spread are taken from David Anthony (ref no 46) whose assumptions linking the archaeology of Andronvo grave pits and Sintashta culture with grossly erroneous translations of the Rig Veda are untenable. In particular Anthony’s linking of Rig Veda 10:18 with Kurgans is based on an incorrect translation, apart from other assumptions.

2. This paper reinforces what was earlier suggested by Reich et al and Priya Moorjani (refs 44 and 45) that ASI-ANI mix occurred in the last 2000-4000 years ago. Both this paper and the earlier Reich paper suggest that there was lack of mixing of ASI and ANI before that. In fact this paper says that ASI and ANI were "largely unformed) 4000 years ago (page 14).

Now here is what is odd. If ANI and ASI ancestry represent IE speakers and Dravidian language speakers (respectively) and if they started mixing as recently as in the last 4000 years when the migrations are said to have occurred, but the admixture had NOT occurred before that date - it tears down all assumptions of a "Caste system" having been created by the migration of ANI/IE speakers displacing ASI/Dravidian language speakers. If they were mixing, there were not remaining separate. It is as simple as that.

If one assumes (probably wrongly) that IE languages came to India about 1500-2000 BCE then it appears that the real mixing of ASI and ANI started AFTER those people came contrary to all assertions by linguists that endogamy commenced in India with the arrival of IE speakers in India.

3. Of course all this ignores and fails to explain how the oldest IE language Sanskrit has references to hydronyms and the riverine geography of IVC area from 3000 BCE and the start of aridity in 2000 BCE - adding to the growing body of evidence that linguistic hypotheses have depended upon dubious assumptions.

It is notable that German and Greek have perhaps 25-40 % Indo-European language content, the rest of the words having a non Indo-European or pre- IE origin. However Sanskrit, the oldest known IE language has 97% Indo-European word content. This anomaly cannot be explained by saying that no language was spoken in India before the arrival of IE languages in 1500 BCE.

4. This paper makes statistical predictions about the entire Indian subcontinent with zero samples from IVC or anything east or south of that (as clearly admitted in the paper). The closest "south Asia" samples are from Swat in Pakistan. Swat is closer to "Turan" (Iran, Uzbekistan, Turkmenistan) than to major IE speaking areas of India like Bengal, south Gujarat, Maharashtra and the Eastern Gangetic plain. So linguistic conclusions connected to genetics need to be viewed with caution.