I just reread Richard Eaton’s book The Rise of Islam and the Bengal Frontier, 1204-1760, and thought I’d make some observations on the theory it purports, and some of the implications. The entire book is available for free: Here.

The Theory

Eaton’s theory of Islamization rests on a much broader theory of how Bengali religion; both Hindu and Muslim was transmitted. It goes more or less as follows: New agricultural technologies, systems of land tenure, and legal/governing institutions were the main drivers behind the spread of both religions. Initially Brahmins, but later Sufis would head into a new region of the delta and establish themselves as local elites responsible for agricultural management (a similar pattern can be seen in the Deccan). Often the Sufi leader and his institution, or a temple institution would get a land grant from the state for this purpose, but this was not always the case. These religiously affiliated colonists brought with them new agricultural technologies from the west, which they would then implement locally. By some arrangement, religious elites became aligned with political elites. This was either because a preexisting regime sent out the Brahmins/Sufis in the first place, or because the regime sought to co-opt them once they had gained a following, and increased local agricultural productivity. Along with religion, the Brahmins and Sufis would bring in with them notions of law, languages, trading customs, etc. which brought eastern regions into the orbit of Indic civilization. In western Bengal this happened in the Epic period. In north Bengal, this happened in the historical period just preceding the Mughals. Due to its lateness, the land tenure system in the north was more primitive, relying on corvée labor rather than taxation in the form of crops or currency which existed in the west of Bengal. In the east and south of the delta however, no organized system of agriculture, government, or religion existed prior to the Mughal period. It was what we would call “aboriginal.” The main driver of this entire pattern was the eastward shift of the Gangetic delta between ~1400 and 1800, which rendered old land less productive, and forced people to confront the task of clearing and farming the forest. This moving river system thus constitutes a frontier in multiple dimensions; ecological, political, technological, and religious. All of these frontiers long predate the presence of Islam in the subcontinent, and can be easily discerned in earlier Sanskrit depictions of the region. Most of Bengal is described as mleccha territory outside the domain of “Aryavarta” in Sanskrit texts like the Baudhayana Dharmasutra (5^th c BCE). Thus, Islamization is just the most recent episode in a phenomenon, which has roots in the Bronze or Iron Age.

Other Theories

This general pattern explains the population drift, and the transmission of new religion and technology eastward over time. But here it might be necessary to back up a step. The broadest problem, which Eaton is trying to solve, is the uncanny distribution of Muslim populations in the subcontinent. It is striking on a demographic map that Muslims are concentrated on the eastern and western flanks of north India, but are thinly spread out in the middle. Eaton lays out the prior 4 theories to explain Islamization before presenting his own. They are as follows: Immigration: This is a theory favored by high class Ashraf Muslims in the past, but it has long been untenable. It states that the majority of Muslims in India are descended from immigrants who arrived to India either on Arab trading ships, as Sufi missionaries, as parts of invading Turko-Persian armies, or as traders. This can explain some cases of Islamization, particularly urban trends. But it cannot explain the conversion of wide swathes of Bengali peasants. It is further rendered suspect by the genetic and anthropometric similarity between low caste Hindus and Muslim peasants, and the genetic difference between both upper caste Hindus and Ashraf Muslims from both aforementioned groups. Patronage: This theory was also favored by Ashraf Muslims, as well as western social scientists who treat religious conversion as a proxy for political change. This theory asserts that the majority of Muslims converted to avoid excessive taxation, or to curry favor with the regime. Aside from question of whether or not Islam was so central to Mughal policies so as to encourage this type of conversion in significant numbers, it also fails to explain mass peasant conversion at the margins of the empire where taxation power was weakest, and the relatively small number of converts in the Gangetic delta. Again, it may well explain some metropolitan trends though. Social Liberation: This was a theory favored by later British ethnographers who sought to portray Hinduism as brutally discriminatory, and maintains popularity amongst Pakistani and Bangladeshi historians. According to this theory the majority of converts were low caste Hindus who sought refuge from the caste system in Islam. However, this theory again may explain some regional cases of conversion, but cannot explain the vast number of converts residing in India’s eastern and western peripheries where the caste system was weakest. If this dynamic were strong, we would expect the largest number of converts to be in the center of India. Furthermore, the degree to which Islam actually liberated people from caste or caste-like structures is highly questionable. Religion of the Sword: This was popular in the earlier British period, in which Crusades influenced historiography was popular. It was also popular amongst Turkic rulers and other Muslims who sought to portray themselves in a self aggrandizing way as victorious and virile conquerors. The theory is fairly straightforward: that the majority of Muslims are the descendants of those who were converted by state coercion. Again, this has regional applicability in places which faced severe religiously motivated violence, but it seems ahistorical to apply it to Bengal, which experienced no such waves of forced conversion of peasantry. If this was the primary mode of conversion we would expect the densest Muslim populations to be in the central Gangetic plain which was the region most subject to invading armies.

Even though Eaton dismisses each theory in turn, they are still worth pointing out for their applicability in specific regions, or in urban centers. Islamization was a complex phonemenon, and even though we are trying to paint a picture which accounts for the trend at the most macro possible level, local histories may have played out very differently.

 Implications for Bengali Islam’s Character

Eaton’s theory seems to view Islamization as essentially a dependent variable of the expansion of medieval state forms and agricultural techniques into the eastern deltaic frontier. In this sense, it is a patronage theory of sorts, but one much more distantly related to state power than the standard iteration. It appears as though the Mughals were merely interested in settling the new land, and in obtaining revenue from it. Furthermore, it appears as though the Mughals did delegate Zamindar status to Brahmins to settle much of the new land, but that the Brahmins delegated the task in a further level of infudation to Sufi pirs, whom had the title of Tuluqdar. Thus the patronage is not of peasants to the state, but of peasants to Sufi Taluqdars to urban absentee landlord Brahmins to the local Mughal governor, to the central Mughal state. As a result of this bottom up process of Islamization, the character of Islam in Bengal was unique in many respects. It was not as though the population converted from aboriginal religion to something recognizable as mostly-Islamic within a single generation. The word “conversion” seems misplaced in this context. Islam was incorporated into Bengali society through a process which Eaton identifies as inclusion, identification, and displacement. First Islamic concepts were introduced and believed in along side Hindu, Buddhist, or Bengali aboriginal beliefs. For instance Allah might be worshipped as one of a number of deities and spirits, and Muhammed seen as one of a number of venerated sages. This is analogous to how in the process of Sanskritization, the major Hindu deities enter the tribal pantheon of deities (and visa versa). Then identification takes hold. Here, local religious concepts are seen to be expressions of fundamentally Islamic ideas. For instance, Brahman might be seen as the local description of Allah. This is analogous to the Hindu process of identifying Shiva and Shakti with tribal religions who worship a divine mother and father. Finally, displacement occurs. Here the Islamic concept is retained while the other concepts are excluded. In Bengal this was an extremely slow process, and in my view was still drastically incomplete even in the 1800s. In parts of Bangladesh it is still incomplete. Though, with modern communication technology, Salafism, and the conscious self-identification of Bangladeshis as part of the broader “Muslim World” displacement is accelerating. The Hindu parallel for this would be the homogenization and “reform” of Hindu religion which has been ongoing since the 19^th century, though even the “end-state” of Hinduism after the displacement phase seems more internally diverse than that of Islam.

There are certain problems with this three phase model. For one, it seems to imply that there is an “end-state” of a religion after displacement has occurred. This seems debatable. It also implies that during the identification phase, the prior religious terminology and conceptual structure is seen as subservient to the new religion’s conceptual structure. Again, this is debatable. It seems possible that the two conceptual structures might become merged, or might produce something entirely new when combined. But these are minor problems. As a result of this slow (and still incomplete) process of transmission, Bengali Islam has adopted a remarkable number of features of preexisting Bengali religion. Preexisting tiger deities of the south have become associated with Sufi pirs, and the faces of tigers still appear on some Sufi shrines as a result. Tantric practices have entered the Bengali Sufi tradition. Ayurveda is popular. Themes of divine motherhood have been strongly carried over into Islamic characters like Amina and Fatima. In parts of rural Bangladesh, Islam seems to have barely entered the displacement phase at all, with Islamic and non-Islamic cosmologies, figures and deities being clearly present in the same religious system. Islam in Bengal is thus not an alien imposition, but a part of the organic religious development of the region, which might explain why the tradition is so deeply rooted in Bangladeshi consciousness.

 A Problem

            Obviously, I think this theory is fairly persuasive. Still, I find it incomplete in one crucial respect. Eaton goes to great lengths to explain the precise mechanics of Islamization on the ground, in terms of Sufis organizing agricultural production. What is not satisfactorily explained is why it was Islam (or more precisely, Sufi pirs), which took up this task. Eaton shows satisfactorily why it couldn’t have happened to the north in Kuch Bihar or Assam or in West Bengal where Hinduism and its associated feudal institutions was stronger, but this does not suffice to explain why Hinduism stopped spreading and Islam started. A superficial answer might be that since an Islamic regime controlled India and Bengal at the time, local (initially non-Islamic) religious leaders who administrated land were co-opted into the State’s religious paradigm. However this explanation can only explain the Islamization of truly aboriginal religious leaders. It doesn’t explain why Sufis (most of whom who were evidently acting independently of state direction) and not Brahmins were the ones to venture into the forest and administrate the practical, grassroots agricultural spread in Bengal during the Mughal period. Eaton himself points out that even under Buddhist rule, it was Brahmins who had settled amidst Bengali indigenous peoples around the 5th century AD to spread their religion and agricultural technology. Thus being part of a different religious group from the official state religion is not sufficient explanation for why Brahmins failed to colonize the delta to the degree which Sufis did. Furthermore, the presence of substantial, though not majority populations Hindus in the east shows that some Hinduization was occurring simultaneously with Islamization. What was the difference between these trends? Why was one trend so much greater than the other?

I have some speculations as to the answers. Perhaps Eaton underplays the role of the Mughal regime in intentionally favoring Sufis and pushing them into the delta, as opposed to Brahmins. He does briefly seem to weakly imply that this was a matter of policy in relation to Sylhet, where Brahmins received land grants, but not grants of the type useful for settling new uncultivated land. Even if Islam was separate from the state’s civic religion, a subtle bias towards Islam might have played a role in the granting of land titles. This objection breaks down though, when one considers that many Sufis were operating under the direction of Brahmins. Or, perhaps his efforts to avoid “misreading” self edifying Islamic histories extends too far in the other direction, and the effects of Turkic conversion by the sword are understated. However this objection also fails when chronology is taken into account, as most conversion took place under Mughal, not Turkic or Afghan rule). As an alternative theory, perhaps there was something unique about the labor intensive, decentralized and local nature of medieval East Bengali rice production, or of the medieval evolution towards decentralized infudation, which favored Islam over Hinduism as a driving ideology. This theory seems suspect though, as it relies on ossified categories of Islam as outwards looking and egalitarian, and Hinduism as inward looking and elitist. The idea that any religion “inherently” pushes any concrete ideological position has gone out of fashion, but if it is true then the answer might lie in the nature of Islam itself. As a third theory, perhaps Hinduism had already expanded to its “natural limit” on the subcontinent, and had solidified its perception of easterners as mleccha. This just pushes the question back one step further though: What sets the constraints on the “natural limit” of the expansion of a religious tradition? Most likely, there was some incentive for heading into the jungle to organize agrarian cultivation, which applied to Sufis and not to Brahmins, though I cannot discern what it was. One speculation could be that Brahmins by the Mughal period had already entirely latched onto the royal institutions they helped establish in the north and especially the west of Bengal, and so felt no economic need to dive into the forest any longer, unlike the more recent and less wealthy and influential Sufis. However, I don’t know how well this matches the empirical facts, as there always seems to be some surplus population of Brahmins in any given region whom are poor and unconnected with royal power.

 Modern Implications

            I’m a bit surprised that I haven’t seen much reference to Eaton’s work in Pakistani or Bangladeshi nationalist historiography. Perhaps I just have’t looked hard enough. It seems as though Eaton is arguing for a sort of primordiality of Bangladesh (and perhaps Pakistan as well is the general argument is carried over). In his narrative Bengal was always substantively different from the Indian core. He emphasizes the non-Aryan character of indigenous Bengalis and identifies proto-Munda speakers as the substratum population. This is not particularly controversial. He observes that despite the heavily sanskritized western region, eastern Bengalis adhered primarily to indigenous religion and Buddhism followed by Islam, with Hinduism playing a minor role. Swadeshi and Indian/Hindu Nationalist ideas of Bengali (and Indian) unity are seriously undercut by this viewpoint. Hindutva tends to frame Bangladesh as “lost” Hindu territory. Of course, if Hinduism is defined as any non-Islamic Indian religion, this is true. But if Hinduism is defined more specifically by reference to the Vedas, or worship of specific deities like Ram, Vishnu, Shiva, etc. then it is much less true (though it seems likely to me that Mother Goddess worship was always indigenous to Bengal, even in the south and east). This narrative seems to be an anachronistic projection of the 20th and 21st century killings and persecutions of Bangladeshi Hindus into the medieval past. Yet, I’ve not seen Hindutva constructions of history confronted with Eaton’s arguments. On a less antagonistic note, I am optimistic that if framed properly, Eaton’s theory could do work towards breaking the tense ossification which has set in between the two religious communities since the colonial period. After all, if Islam and Hinduism both spread according to the same pattern, and if Bangladeshi Islam has an indigenous substratum, the distance between the two traditions seems to lessen.

The theory puts Bangladeshi nationalism in an odd position as well. Bengali language ideology (including some Bangladeshi variants) posits all of Bengal as a primordial sociological unit, but Eaton seems to suggest that the divide between east and west Bengal is ancient. This in turn could strengthen some forms of particularist East Bengali nationalism which emphasizes the distance between Bangladesh and India, and assimilates ethnically non-Bengali Bangladeshis into the narrative of East Bengal’s past. If the logic of Eaton’s theory applies elsewhere in the subcontinent, then other South Asian Muslim nationalist ideologies could also be bolstered by it. The narrative would be “We’ve always been on or beyond the periphery of Hindu civilization. The difference between us and them is primordial.” The logic of the theory can also be carried further to justify some forms of Dalit nationalist ideology, which also emphasizes persistent historical differences between urban, agrarian Hindu society, and the societies which existed in the hills or on the margins of society. Yet surprisingly, I’ve not seen Eaton’s work mobilized to political ends. Eaton’s theory is the most plausible one for explaining the distribution of Islamic populations, though it needs further elaboration to be a truly complete theory. It has the strength of being in step with the most ancient and persistent trends in Indian history, especially at a time when Indian historians have eschewed such “big picture” explanations. The theory’s continuity with long run patterns makes it compelling, but this only leaves the discontinuity (i.e. the halting of Hinduism’s expansion and the dawn of Islam) even more puzzling.