Genetic analysis indicates that the vast majority of people in India today are descended from two populations that Reich dubbed Ancient North Indians (ANI) and Ancient South Indians (ASI). Different populations have different mixes of ancestry from ANI and ASI, but most are composed almost entirely of just those two groups – with some exceptions in eastern India, where you see some people that expanded out of Southeast Asia and speak Austroasiatic languages related to Vietnamese, and people with East Asian ancestry speaking Sino-Tibetan languages, living in the mountains. ASI and ANI were pretty different – about as genetically different as Europeans and East Asians today. Note: ANI and ASI, as acronyms, were created in order to obscure certain points that many Indians don’t want to think about. By the way, in the sense we’re talking about, biogeographical, ‘India’ includes Pakistan and Bangladesh.
Let me give a more familiar example: most people in Latin America are, to a first approximation, descended from a mix of Amerindians and Spaniards. The mix varies: you can find populations that are 100% Amerindian, populations that are 60-40 like Mexico, 20-80 like Colombia, and so on. If you plotted ancestry on a principal components plot, most people in Latin America would fall somewhere along a line. India is like this.
On the whole, people from North India have more ANI ancestry, while people in southern Indian have more ASI ancestry. The proportions generally range from 80% ANI to 80% ASI. There are actually a few populations that are close to unmixed ANI (Kalash) or unmixed ASI (tribal populations in South India such as Palliyar, Ulladan, Malayan, and Adiyan). Groups that speak Indo-European languages typically have more ANI, while those speaking Dravidian have more ASI. Populations (including castes) with higher social status generally have more ANI ancestry. The Y-chromosome and mtDNA patterns show that ANI contributed a disproportionate fraction of male ancestry, while ASI accounts for the great majority of female ancestry – again, much like Latin America.
Outside India, populations with much ASI ancestry are practically nonexistent. Sure, you could find a touch of it in Indonesia ( particularly Bali ) and mainland South East Asia (Cambodia) from traders and missionaries, but that’s about it. Not counting 20th century movements, of course.
Who the hell are these mysterious ANI and ASI populations? Basically, they are mixtures of three older, more fundamental populations:
A. hunter-gatherers that had lived in India for a long time ( barring undocumented prehistoric replacements as have occurred in Europe). The closest living population ( not that close, but distant cousins) are the Negrito inhabitants of the Andaman Islands. Ancient Ancestral South Indian ( AASI )
B. Farmers from Iran – specifically, goat-herders from the Zagros mountains, sampled from around the 8th millenium B.C. (Iranian agriculturalist-related)
C. Steppe ancestry related to the Yamnaya
ANI, the ancient north Indians, are about half steppe and half Iranian agriculturalist-related. ASI, the ancient southern Indians, are about 25% Iranian agriculturalist-related and 75% old Indian hunter-gatherers (AASI)
Timeline: AASI are there first. ASI forms as Iranian farmers move east & is probably the basis of the spread of the Dravidian languages [ Elamo-Dravidian?] . ANI comes last, and carries with it the Indo-European languages.
Question: who made up the Indus civilization? Probably a mix of Iranian agriculturalists and AASI. Some Indus skeletons have been sequenced, but the results have not been released, probably because the researchers involved find them upsetting.
Another odd point: it looks as if there wasn’t much R1a in the region when you’d think there already would have been. There is a possible explanation, which also would explain why today it is so very widespread in this subcontinent: it confers a fitness advantage, like that Y-chromosome in horses. I call this the Topsy hypothesis.
Back to caste, jatis and all that.
Reich found caste interesting, partly because he grew up in a caste, Ashkenazi Jews. He feels empathy for all the Romeos and Juliets held apart by caste over the millennia – was there perhaps a little red-haired girl in his past?
The situation in Europe is different from India, because there practically everyone in India is a member of a caste, while in Europe most people aren’t.
There is an overall classification called varna, which divides all of society into four ranks : priestly (Brahmins) , warriors (Kshatriyas) , middle classes ( Vaishya) , and laborers (Shudra), plus the Untouchables ( Ness ). Varna probably influences the selective pressures that populations in a particular caste experience. For example it wouldn’t surprise me if Untouchables have developed specialized adaptations for dealing with certain pathogens, and of course surviving hot lead.
But those castes are made of thousands – maybe tens of thousands – of jatis, small endogamous groups. Generally one must marry within the jati- it is usually impermissible even with another jati of the same caste. Jatis have sometimes changed caste, but few individuals change jatis.
How did this very odd system originate? Varna sounds quite a lot like old patterns in PIE society, if you believe Dumezil. But Jatis are another story. Some revisionist historians, led by the anthropologist Nicholas Dirks, have argued that the Brits strengthened the caste system, which they say was not very important in parts of India. Of course
Megasthenes was writing about the Indian caste system two thousand years ago, but who reads?
Reich’s genetic analysis showed that the many jatis were founded by small groups and have experience near-zero gene flow for two to three thousand years. Other jatis with larger founding groups would not show the same tell-tale genetic signature, but most of them could also have been almost perfectly endogamous. I expect that is true for the great majority of jatis: the long-term degree of of endogamy was probably not strongly related to the formation process.
Partitioning the population into tens of thousands of gene-tight cells is unusual and had to have some interesting ( probably bad) genetic consequences. For one thing, it would greatly slow the transmission of new favorable alleles. No Fisher waves in India ! Pre-existing variation, for example the lactose-tolerance allele brought in by the Indo-Europeans, could increase in frequency if favored, but new favorable mutations would be locked-in to the jati they arose in. These results show Nicholas Dirks and company were utterly wrong, unless they think that perfidious Albion reached back in time to make India what it is today. Being a fountain of nonsense is presumably why Dirks makes the big bucks ( currently on leave, getting $434,000 per annum ).
A lot of people in India are fans of an Out-of-India theory, usually positing India as the origin of the Indo-European languages, but that can’t possibly be true, since as I mentioned you simply don’t see much of those Andaman-like lineages outside of India. Nor does it fit any of the linguistic evidence. or anything else. So they’re crazy, but who isn’t?
The Aryan-Invasion-Theory sure looks to be basically correct. As for the archaeologists saying that there’s not enough evidence of devastation, Reich points out that they can’t really detect the fall of the western Roman Empire, which hardly means it didn’t happen. War and migration are well-known important factors in written history – why not in prehistory? Because many contemporary archaeologist and historians think that wishing can make it so. They should be paid accordingly.