alan little’s weblog

more indian genetics

20th April 2004 permanent link

I’ve picked up pointers from gene expression for two newer papers on the genetic origins of Indian populations, following on from the one by Bamshad et al that I mentioned a few months ago.

The first one is The Genetic Heritage of the Earliest Settlers Persists Both in Indian Tribal and Caste Populations by T. Kivisild et al in the American Journal of Human Genetics 2003. The authors argue that the DNA of Indian populations – both caste and tribal – appears to go back to the first wave of human settlement from Africa in the Pleistocene era and that the genetic impact of subsequent immigration, if any, has been very limited.

The second paper, Independent Origins of Indian Caste and Tribal Paternal Lineages by Richard Cordaux et al (Current Biology, February 2004) supports Bamshad et al and points out that, although there are no genetic markers that are unique to either caste or tribal populations – suggesting that there has been a lot of mixing between the two groups – nevertheless the frequencies of a number of key markers are very different. When this is taken into account the male DNA of the caste population resembles central Asians far more than it resembles the assumed-to-be-indigenous tribal populations. This appears to be true both in north and south India.

We conclude that paternal lineages of Indian caste groups are primarily descended from Indo-European speakers who migrated from central Asia 3,500 years ago

I’m inclined to agree with this conclusion. But it is far too sweeping to make just based on the evidence presented in this paper, and depends on some major unexamined and questionable assumptions.

• Cordaux et al assume, for no stated reason whatsoever, that there is a connection between genetic lineage and language. “[P]aternal lineages of Indian caste groups are primarily descended from Indo-European speakers …”. “Haplogroup O-M95 … might be associated with the spread of Austro-Asiatic languages, which are exclusively spoken by tribal groups in India”. This, remember, in a paper that discusses how Indo-European-speaking north Indians and Dravidian-speaking south Indians are genetically almost indistinguishable; how both are closely related to mostly Turkic-speaking central Asians and not so closely related to Indo-European speaking Europeans. Hmm.
• In addition to the completely unexplained assumption about genetics and languages, they say nothing about where they get their “3,500 years ago”. There is a passing reference to the “relative homogeneity” of caste DNA being an argument for “recent” arrival, but no attempt to quantify this. I strongly suspect they are just echoing what they call “archaeological and linguistic evidence” (actually, as far as I have been able to tell, pretty much a purely linguistic hypothesis for which there’s no clear archaeological evidence one way or the other).
• Another question that occurs to me as a non-specialist is: does a genetic resemblance between Indian and central Asian populations necessarily tell us anything about the direction of migration? What are the reasons for thinking the shared genetic markers originated in central Asia not in India? Would the evidence also be consistent with a population that originated in northern India and moved outward in both directions, northwards into central Asia and southwards into southern India?

I’m not suggesting that anybody should be expected, in a five page paper, to examine every possible nuance and counter-argument for every assumption they make. But some mention of these rather large and obvious questions might not have been a bad idea. While I’m niggling, a couple of smaller niggles:

• They cite Cavalli-Sforza as an authority on “archaeological and linguistic evidence”. Cavalli-Sforza may be an important pioneer of genetic research, but I’m not aware that anybody would accuse him of being an authority on archaeology or linguistics.
• (and this is a really minor nitpick, but I can’t help it. I care about language and the blighting effect political correctness has on it). There is a very interesting discussion of several reasons why the maternal DNA of caste and tribal populations is much more alike than the paternal; one being that women are generally far more likely than men to marry upwards in social hierarchies. Good point, interesting. But then they say this is an alternative to “the proposal that the Indo-European migration was male mediated”. I presume this is some sort of bizarre PC-speak for “the proposal that the immigrants were mostly men”. So why not just say “the proposal that the immigrants were mostly men”?

Nevertheless, niggles aside, it’s becoming clear that the genetic evidence points to Indian caste populations – in the male line at any rate – having originated as migrants from central Asia, whereas tribal populations appear to be indigenous, quite possibly dating back to the first wave of “Out of Africa” migrations in the Palaeolithic. This would be consistent with the general view held by linguists that the Vedic speakers were immigrants to India. Note that I say “consistent with”, not “proves”: firstly because there is no provable direct correlation between genetics and language and secondly because I haven’t seen anything that seriously argues a date for the arrival of the immigrants based on the genetic evidence. For linguistic and archaeological reasons (chariots, mainly) I don’t believe the Vedic speakers can have arrived much before about 3500 years ago; Cordaux et al also give this date, but there’s nothing in their paper to say whether they base it on the genetic evidence or are just parroting the linguists’ standard geusstimate.

When I first wrote about this whole subject of trying to reconstruct ancient history in the face of sparse evidence and political agendas, I said “I find this both fascinating and rather depressing … wondering whether there’s any chance of anything other than rival tribes shouting ideology & speculation at each other.” Razib, in this excellent essay at gene expression, thinks genetics may be that chance and talks eloquently about how genetic studies have the potential to give us some sort of firm scientific basis for studying prehistory and the origins of present-day “peoples”, whereas until now we have had really quite limited scraps of linguistic and archaeological evidence used to back up all kinds of racist and nationalist fantasies.

In the past ten years genetics has begun to highlight very sharp divergences from national myths and even linguistic and historical analyses (the latter often influenced by and used by the mythicists)
…
First, there are the National Essentialists. These are people who are nationalists at the least, and racialists at the most. Historical questions are crucial to the identities of these people. They come in all sorts. From Native Americans that reject the Siberian origin of their peoples on the grounds of spiritual chauvinism to upper caste Indians who react with fury when told that they are genetically closer to their black-skinned Dalit neighbors than other “Aryans.” Genetics tends to throw the “Great Chain of Being” of these people out of kilter since they invest so much in the current state of knowledge, reverse engineering the facts to show their own heritage to be the most prestigious. If they hold that “group A” brought civilization to the world, and it turns out they are a member of “group B,” it takes great effort to reinterpret the facts so that group B are the jewels of God’s creation. Rather, instead of throwing out years of self-serving “scholarship,” this group will reject the genetic evidence as long as possible.
…
Second, there are the National Idealists. The statistical nature of genetics gives them the room to reject all assertions of the movement of genes and people. Though a given set of results is provisional and subject to revision, they tend to deny the results a priori (or evince a divine skepticism) because they prefer to think that culture, ideas, are the prime movers, not peoples.
…
Both groups are concerned with norms and re-creating idealized utopias. Their values might be different, one group worships the primacy of Blood and another idolizes the power of the Idea, but their enemy is the same, facts, reality, tightly constructed deductive models buttressed by empirical evidence, subject to provision, de-sacralized and reduced to the bare necessities that science demands , but no more. They stand united as romantics against the unfeeling march of science and scholarship. Without the aid of natural science the human sciences have traditionally been hijacked and used as tools in the furtherance of ideological crusades. But now that natural science has joined the fray (the scientists have the “back” of empirically oriented scholars and the reverse), the ideologues are terrified, their utopian visions always threatened by the encroachment of reality. The scholarship is strengthened and more difficult to dismiss when buttressed by genetics or skeletal morphology (augmented by computation).
…
there are people out there that live in terror of facts.

Read the whole thing, it’s excellent. However, as Razib also mentioned in an email, the time of the great genetic light-shedding has not yet come because we simply don’t know enough yet:

I suspect that we'll know what's going on in about 10-15 years after we have 100 studies on every given question. right now we have about 10 years of fine precision laboratory work (the PCR era), so there are “ground-breaking” and “revolutionary” results coming out because the sample is so small that the variance is all over the place (sampling error)

These three papers on India, coming out two-to-one in favour of major immigration from central Asia based on largely the same sample of a few hundred individuals, certainly backs up Razib’s view. Not only sampling error, but apparently competent and serious researchers coming to almost directly opposite conclusions by applying different statistical techniques to the same samples. Still, it does look as though the genetic evidence is starting to point in something like the same direction as the linguistic evidence.

Discussion ...

related entries: Yoga

all text and images © 2003–2008