« Dissolution of a meme | Gene Expression Front Page | "Buying organic 'gives you boost'" »
September 04, 2004

An observation on history & science

I read a lot of history books. I read a lot of science books. There are other sorts of books I read, but these are the two primary types. One thing that I have noted is how persuasive some historians can be, but that it is hard to evaluate phrases like, "it is generally agreed," or "this seems to be the most likely reason." The lack of numbers makes fact-checking hard. You basically have to take the erudition and honesty of the historian at face value (as a lay person). So, reading The Reformation, I take MacCulloch's word that the importance of the various points and facts he is presenting before me are really relevant, but it would be nice if he backed up the importance of a particular theological argument by noting that textual analysis showed that this dispute showed up "45% of the time in letters recovered from this period." Perhaps there is stuff out there like that (I've heard of it, and sometimes it even shows up in the popular books), but I would like more of it. Am I going to "fact check"? Not really, but, I have a more precise understanding of a given percentage than I do "significant."

I am come to this thought for this reason, I have read several books about caste in the Indian subcontinent. The authors often note:

  • Caste is fuzzier, more malleable than most South Asians realize.
  • The authors I have read tend to reject the idea that the British invented caste, and that is, the position that it has a short history as a social feature of the Indian subcontinent.

Well, duh! The recent genetic studies that have been coming out of laboratories seem to point out that:

  • Many markers are found up and down the caste hierarchy.
  • Many markers are shifted toward one end of the caste spectrum.

These two facts tend to argue for a moderate view, in contradiction of the radical "social constructionists" and Hindu fundamentalists. But there is another point that I have found striking: different "high caste" groups tend to cluster together. That suggests that caste has not only been a long standing feature of Indian society, but to some extent, caste mobility has not been high enough to break apart genetic affinities at the top of the spectrum! The persistence of M17 patrilineages over thousands of years for example on the Gangetic plain at the upper ends of social strata really puts into perspective writers who wax on about the importance of social mobility of brigand tribes (adivasis) during the interval between Mughal decline and the British Raj (to be specific). These are plain numbers at work here. It would be nice if I saw this type of material being injected into the new scholarship. Analysis of the archives and oral history to refute a particular model is fine & good, but historians, in this particular context, have some strongly suggestive genetic studies they can appeal to.

This line of thought was precipitated by the fact that an acquaintance of mine, a South Asian graduate student who knows an encyclopediac amount of facts on the region, was totally surprised by possible genetic affinities of high castes and differences between them as a group as opposed to low castes (though he was aware of textual evidence about the migration of Brahmins to South India and so a priori did not assume that caste came about in the 18th century). Unfortunately, when this work does get out, it is through popular press pieces that oversimplify to reenforce popular myths ("Aryan" vs. "Dravidian"). Right now there aren't enough studies that lay persons can appeal to "established opinion." Once you have 500 studies on English vs. Welsh genetics, or North Indian vs. South Indian genetics, non-scientists will have a large field of data to mine and use to buttress or refute models. I hope they make good use of it, not least of which because the scientists themselves and the popular press will mangle the history (if you read the "Discussion" section of many historical genetics pieces you do see some howlers that indicate no historian who specialized in a given region was given a draft of the paper).

Related: Spencer Wells has a new special this fall, Quest for the Phoenicians. I'm skeptical of this level of specificity, but I'll wait until the special comes out.

Posted by razib at 05:12 PM