The diverse origins of the Bengali gentry

Satyendra Nath Bose, a Kayastha

I’ve been rather busy with other things and the South Asian Genotype Project has fallen a bit by the wayside. But, I plan on allocating a day on a weekend soon to getting through the backlog. But, before that, I thought I would submit something that might clarify or illuminate an aspect of South Asian history and culture that the data have shown.

The public data have a large number of Bangladeshis. Sampled in Dhaka. Unlike most South Asians they don’t exhibit much structure. But unfortunately, we don’t have too many samples from West Bengal. That being said, I do get some through this project. As you can see on the plot (round dark point) the Brahmins of Bengal are genetically very distinct from the Bangladesh samples (I have more samples, but this position is always the same for all of them). In fact, they are closer to Brahmin samples from Uttar Pradesh. Using modeled based clustering though I have become strongly convinced that the Bengali Brahmins do have a minor contribution from the non-Brahmin Bengali population because they have East Asian ancestry.

The triangles on the plot are from Kayastha individuals from West Bengal. I have seen a few Kayasthas from this region, and two trends about them:

  1. closer to Bangladeshis than Bengali Brahmins
  2. a lower fraction of East Asian ancestry than Bangladeshis

The lower fraction of East Asian ancestry is explained by an admixture cline which is a function of geography. My own East Asian admixture is high, and that is easily explained by the fact that my family is from the very east of eastern Bengal. As far as number one, one hypothesis some have presented is in Bengal the influence of Buddhism relatively late in history (down to ~1000 AD) resulted in a relatively panmictic population with low amounts of structure (a problem with this model is that Buddhism was strong in Punjab relatively late, but that region has lots of structure).

In any case, socially this is relevant because the Brahmins and Kayasthas of Bengal have traditionally been the two groups which are the pillars of the gentry class of the region (“bhadralok”). But there have been subtle differences between them. The Kayasthas have traditionally been more willing to work with Muslim rulers as a literate service class, as their ritual status is lower than Brahmins, and so presumably they had less to lose from “pollution” through contact with these foreign groups.

Looking at the genetic differences between the two major castes of the gentry in Bengal is interesting because it is a test of a model of the origin of caste in South Asia. Overall I think Nicholas Dirks’ Castes of Mind has not had a good effect on how we understand caste and jati, but it isn’t totally uninformative for what we see in Bengal. Like South Indian Brahmins, Bengali Brahmins can be modeled as a mixture of Brahmins from the upper Gangetic plain, with a minority local substrate contribution (~25% or less). They are genetically a distinct population whose endogamy is ancient and thorough. In contrast, the Kayasthas are “generic Bengalis,” who likely did increase their social status by becoming a literate bureaucratic class (“clerks”).

7 thoughts on “The diverse origins of the Bengali gentry

  1. This means absolutely nothing until Razib explains the clouded mysteries of the infamous mtDNA M49 found in Khasi Austro Asiatic tribal folk in Bangladesh’s eastern periphery. Once elucidated, the star shaped patriarchy of the R1a1 all father must give way to the matrilineal mastodon of the all evading M49.

  2. “a problem with this model is that Buddhism was strong in Punjab relatively late, but that region has lots of structure”

    “a problem with this model is that Buddhism was strong in Punjab relatively late, but that region has lots of structure”

    I don’t think this is an anomaly. Buddhism departed from Punjab roughly around the same time as rest of India. The kingdoms that Turkic invaders conquered to establish a foothold in Punjab were all Hindu (Shahi kingdom, Chauhana kingdom etc).

    The Buddhist relics that are discovered in Punjab are located in far northwest, and in the region beyond Indus (NWFP), dating back to Gandhara culture. The death knell of this northwestern Buddhist culture was already sounded by Hun tribes many centuries before Islamic invaders came in.

    Buddhism did maintain a tenuous hold in Sindh though. We know this because when the Arabs came to Sindh they noted that even the ruling dynasty was Hindu (Dahir), general population consisted of a large numbers of Buddhists.

    It will be interesting to know if Sindhis display a homogeneity like Bangladeshis, or they have a rich substructure like Indian castes.

  3. What is the significance of the circles labelled “Kayastha_Bengal”? They don’t seem that close to the triangles.

    color scheme is shitty. must be other colors that are similar. i have only two explicitly kayastha samples in there.

    (one of the 1000 genome bangaldesh samples looks to be a half-brahmin half normal bengali, but i pruned that individual)

  4. “The Kayasthas have traditionally been more willing to work with Muslim rulers as a literate service class, as their ritual status is lower than Brahmins, and so presumably they had less to lose from “pollution” through contact with these foreign groups”

    The Kaysthas all over India worked for muslim as well as the British. I would also say that this idea that Brahmins didnt work for the muslims(due to pollution) is wrong. What has happened during the British time was constant removal of this collaboration with the foreign groups. This is not just with Brahmins. Rajput family who intermarried with Mughals constantly paid money to historians like James Todd etc to “rewrite” their histories. So everyone worked with/for everyone notwithstanding the pollution part.

    On Kayasthas i have Kayasta freinds who claim that they are the only group who moved form being “Shudras” to Uppercaste.

  5. Is there a Iran_N vs Steppe cline between various Bengali groups? Brahmins are high steppe which means regular Bengalis have high Iran_N. I have heared regular Bengalis have similar ANI as Reddy/Patel like groups? In term of AASI they seem in between South Indians and North Indian.

  6. For those of us who are not that advanced in Statistics could you suggest links to a site that explains Principal Component Analysis. I’d like to understand what these charts mean.

Comments are closed.