Satish Deshpande—whom I hold in high regard as one of our finest academics, a fast-vanishing tribe today—has done me the honour of writing an extensive review of my book Caste Con Census.
This article is not meant to indulge in sterile semantics, least of all to be a rejoinder to Deshpande’s review. It is to discuss an interesting point that he makes at the end to support his contention, albeit in, a speculative manner, that a caste census can unleash attrition which would lead to a weakening of castes::
A caste census is an eminently and imminently possible event in today’s world. And though it will not itself annihilate caste, and may even aggravate problems initially, it will also feed the forces that will begin the war of attrition that can weaken caste.
He does not specify which forces, supposedly unleashed by the caste census, would ignite a war of attrition capable of weakening caste. Elsewhere Deshpande also says: “The very problems and troubles that Teltumbde rightly claims the census creates will also accelerate attrition, initiate a reductio ad absurdum of vested interests, and take us slowly but surely towards ‘abolition’.” He does not spell out how.
I have, on my part, argued that historically caste enumeration has tended to reinforce caste consciousness, generate internecine friction among castes (and sub-castes)—as the Mala–Madiga divide amply illustrates—and erode what little solidarity exists among them. These are hardly forms of attrition that would lead to the abolition of caste.
Nonetheless, this thesis of attrition, precisely because its mechanism remains undefined, is intriguing and has the potential to bolster the confidence of the pro-census constituency. I will therefore engage with it in good faith and with due seriousness.
The claim that enumerating caste will, through the turbulence it generates, gradually erode caste as a social institution rests on an optimistic teleology that is not supported by either historical evidence or the internal logic of the argument. . The more accurate reading — both sociologically and historically — is the opposite: that the administrative enumeration of caste systematically hardens what it purports merely to count, and that whatever turbulence it generates will express itself not as attrition of caste identity but as its intensification, fragmentation, and competitive entrenchment.
What Social Attrition Requires
Before examining whether caste enumeration can produce attrition, it is necessary to be precise about what attrition, as a sociological process, actually means and requires. Social attrition — the gradual erosion of a group identity or institutional form — occurs when the boundaries that define a group become increasingly costly to maintain, decreasingly useful to members, and progressively less legible to the surrounding social environment. Attrition is not conflict. It is not fragmentation. It is not internal contestation. These can all occur within a fully intact identity structure, and even strengthen it. Attrition requires, at a minimum, that the identity in question becomes less salient to the individuals who bears it — that the rewards of membership diminish, that the penalties of boundary-crossing decline, and that the category itself loses its organising power over social life.
The conditions under which social attrition occurs are well understood in the sociological literature. They include the weakening of endogamy, which is the reproductive boundary that most decisively sustains group distinctiveness over generations; the erosion of occupational segregation, which removes the economic foundation of group identity; the decline of residential clustering, which removes the spatial basis of group solidarity; and the weakening of the ritual and symbolic systems that mark group boundaries as morally significant. None of these conditions is produced by enumeration. Several of them are actively impeded by enumeration.
The foundational argument against the attrition thesis begins with the sociology of knowledge and the relationship between administrative categories and social reality. Bernard Cohn’s work on the census in British India demonstrated conclusively that the enumeration process did not passively record a pre-existing social reality but actively produced and stabilised it. Castes that existed as loose, regionally variable, often internally contested clusters of practice and identity were compelled, through the logic of the census questionnaire, to present themselves as bounded, named, and enumerable groups. The census demanded clarity where social life offered ambiguity; it demanded singularity where social life offered multiplicity.
Nicholas Dirks, extending Cohn’s argument in Castes of Mind, showed that this administrative crystallisation had durable social consequences. Groups that had been liminal, mobile, or internally differentiated were frozen into singular administrative categories. The colonial state, in seeking to know and govern caste, made caste knowable in a new and more rigid way — and that new legibility fed back into social practice, providing groups with a template for self-organisation, political claim-making, and boundary enforcement that had not existed in the same form before.
Rogers Brubaker’s theoretical framework in Ethnicity Without Groups provides the general sociological mechanism here. Brubaker argues that the categories through which states count populations tend to become the categories through which populations organise themselves politically, claim resources, and construct solidarity. The census does not merely reflect groups; it reifies them, lending them an objective, quantified existence that strengthens rather than weakens their claim on members’ loyalty. Applied to caste, this means that enumeration provides jatis and caste clusters with a new kind of political existence — a numerical weight, a measurable constituency, a legible interest — that gives their leaders powerful incentives to maintain, police, and if necessary reinvent caste boundaries.
This is not attrition. It is institution-building.
Six Decades of Enumeration, Zero Attrition
The proposition that caste enumeration will lead to attrition is not merely theoretically implausible — it is empirically falsified by the only sustained historical experiment we have. The colonial census operations from 1871 through 1931, conducted over six decades, provide a natural test of the attrition thesis. If enumerating caste produces turbulence, and turbulence produces attrition, six decades of systematic enumeration should have produced some measurable erosion of caste identity and salience. The historical record shows the opposite.
What Mandal demonstrates, most devastatingly for the attrition thesis, is that enumeration-backed reservation creates a permanent institutional incentive for the reproduction of caste identity.
What the colonial census produced was the proliferation of caste sabhas and caste associations across the subcontinent — formally organised bodies whose explicit purpose was to define, defend, and advance the interests of particular caste communities as enumerated administrative units. Groups that had been internally differentiated began to aggregate and homogenise around census categories. Groups whose ritual rank was ambiguous or contested began to organise political campaigns to improve their census classification — a process that Susan Bayly and Rosalind O’Hanlon have documented in rich detail. The turbulence of enumeration expressed itself not as the dissolution of caste but as the modernisation of caste — its transformation from a locally variable, customarily reproduced institution into a politically organised, formally constituted interest group capable of operating in the arenas of electoral politics, administrative petition, and legal claim.
There is no point in this six-decade history of census enumeration at which attrition is visible. There is abundant evidence of the opposite: the hardening of boundaries, the growth of caste political organisation, and the increasing salience of caste identity in public life. The experiment has been run. The attrition thesis failed. Another concrete case is of the Scheduled Castes, whose enumeration, instead of producing attrition, has only led to internecine sub-caste clashes.
The Post-Colonial Evidence
The post-colonial history of caste and enumeration in India confirms the colonial pattern. The Mandal Commission’s enumeration of OBC communities and its subsequent implementation in 1990 is the most instructive case.
Political scientists including Yogendra Yadav have documented how the Mandal moment inaugurated what he termed the “second democratic upsurge”. But this upsurge was organised precisely around rendering caste identity more coherent, more politically legible, and more administratively consequential by enumeration and reservation. The numerically significant OBC communities of the Hindi belt — Yadavs, Kurmis, Koeris, and others — did not experience the Mandal process as the beginning of attrition of caste identity. They experienced it as the beginning of their political arrival. Caste identity, far from eroding, became the primary currency of democratic politics across north and central India.
What Mandal demonstrates, most devastatingly for the attrition thesis, is that enumeration-backed reservation creates a permanent institutional incentive for the reproduction of caste identity. A group that has secured reservations on the basis of its caste classification has an ongoing material interest in maintaining that classification — in ensuring that its members continue to identify as members of the group, that boundaries against outsiders are maintained, and that the group’s numerical weight is not diluted by inter-caste marriage, or the weakening of caste endogamy. Enumeration and reservation together create precisely the opposite of the conditions required for attrition.
Fragmentation Is Not Attrition
It might be objected that caste enumeration, even if it does not produce attrition of caste identity as a whole, produces fragmentation: the breaking up of large, aggregated caste categories into competing sub-caste and jati identities. And that this fragmentation is a form of the reductio ad absurdum that the attrition thesis anticipates. This objection must be addressed directly, because it conflates two entirely different social processes.
Fragmentation and attrition are not only distinct in the context of caste politics, they are near opposites. When a large, aggregated OBC category fragments into competing sub-caste claims — when Yadavs, Kurmis, and Nishads begin to contest each other’s share of the reservation pie, or when Jatavs and Valmikis within the Scheduled Caste category develop competing political identities — what is happening is not the erosion of caste identity but its intensification at a finer grain. Each sub-caste or jati that enters the political arena as a distinct claimant is asserting, not dissolving, its caste identity. The fragmentation of large census categories into smaller competing units does not move toward annihilation; it moves toward proliferation.
Indeed, the logic of competitive enumeration — in which groups seek to maximise their counted numbers to secure greater political and administrative weight — creates powerful incentives for boundary maintenance at every level of the caste hierarchy. A jati that allows its boundaries to blur, that tolerates inter-marriage with adjacent groups, or that fails to maintain the solidarity of its members, risks losing numerical weight and therefore political influence in a system that distributes resources on the basis of enumerated group size. Fragmentation under these conditions does not lead to the collapse of caste; it leads to the emergence of a more finely grained, more intensely competitive caste politics in which more groups, organised at a lower level of aggregation, are more deeply invested in the reproduction of caste identity than before.
Internecine conflict between castes and sub-castes — the actual observable consequence of competitive enumeration — is not a symptom of caste’s decline. It is a symptom of caste’s vitality. Groups do not fight bitterly over the boundaries of a dying institution; they fight bitterly over the boundaries of one that still distributes significant rewards.
Logical Structure of the Error
The attrition thesis, when examined closely, turns out to rest on a logical error that is worth making explicit. The argument runs like this: enumeration creates turbulence; turbulence creates contradictions; contradictions accelerate the internal collapse of vested interests; collapse leads to annihilation. Each step in this chain is asserted rather than demonstrated, and at least two of the links fail under scrutiny.
Enumeration does not expose the irrationality of caste interest; it gives caste interest a rational, quantified, administratively recognised foundation on which to stand.
First, the move from turbulence to the weakening of vested interests assumes that conflict and contradiction erode the institutions around which they are organised. But this is not what the sociology of institutions shows. Conflict frequently strengthens the institutions it appears to threaten, by increasing their salience, reinforcing members’ investment in group identity, and generating the organisational infrastructure — leaders, associations, legal strategies, political parties — that makes the institution more durable, not less. Caste conflict has historically produced not the erosion of caste but the political organisation of caste.
Second, and more fundamentally, the argument assumes that the vested interests nurtured by caste are weakened by visibility and numerical legibility. The opposite is demonstrably true. A caste group that does not know its own size and distribution is politically weaker than one that does. The census provides precisely that knowledge, and, with it, the rational incentive to invest more heavily in the maintenance of caste identity. Enumeration does not expose the irrationality of caste interest; it gives caste interest a rational, quantified, administratively recognised foundation on which to stand.
The expectation that counting caste will lead to its abolition is, at the level of logic, equivalent to expecting that counting religious communities will lead to the abolition of religion, or that counting ethnic groups will lead to the dissolution of ethnicity. The historical evidence from every context in which states have enumerated ascriptive group identities points in the same direction: enumeration strengthens, organises, and perpetuates the identities it counts.
The Only Honest Premise for Caste Enumeration
This analysis points toward a conclusion that advocates of the caste census are often reluctant to state plainly, but which the internal logic of their position requires: the argument for caste enumeration can only be coherently premised on the recognition that caste is a permanent or at least durable feature of Indian social life, and that, given its permanence, the groups it defines deserve their proportionate share of state resources, political representation, and administrative power.
This is a legitimate argument. It broadly reflects the social justice tradition in Indian politics—from Dr Ambedkar’s insistence on the material reality of caste oppression to the Mandal Commission’s administrative operationalisation of that insight. It takes caste seriously as a system of inherited disadvantage and demands that the state do the same in the distribution of remedies. There is no intellectual dishonesty in this position.
However, it should not be forgotten that while Ambedkar’s pragmatic concerns led him to demand reservations, his ultimate ideological objective remained the annihilation of caste.
What is intellectually dishonest — or at best intellectually confused — is the attempt to graft onto this argument the additional claim that enumeration is also a road toward caste’s eventual abolition. The two positions are not merely in tension; they are contradictory. A policy instrument premised on the recognition and administrative entrenchment of caste identity cannot simultaneously be a mechanism for that identity’s dissolution. The road built for one destination does not arrive at the other.
Annihilation of caste — if it is a serious goal rather than a rhetorical aspiration — requires instruments that work in the opposite direction from enumeration: the dismantling of endogamy, the universalization of economic opportunity irrespective of caste, the delegitimisation of caste as a category in social and administrative life, and the construction of a public culture in which caste identity becomes progressively less meaningful to the individuals who bear it. None of these instruments is the census. The census is the instrument of the social justice argument, and it should stand or fall on the merits of that argument alone, and not on the borrowed credibility of an annihilationist vision it cannot, by its own logic, advance.
At a pragmatic level, support for caste enumeration may be conceded. But it should not be burdened with the untenable claim that it will generate attrition and lead to the abolition of caste. The short-term appeal of “social justice” must not eclipse the long-term goal of annihilating caste. The former clearly serves the political class; it is incumbent on public intellectuals to ensure that it does not override the long-term interests of the people.
Postscript
To return to Deshpande’s review of my book, I was pleasantly surprised to find that his opening remarks focused not on the book but on me—suggesting that I have “a streak of contrariness, the need to be distinctive,” and that I approach issues from an oblique angle. Deshpande, who has followed my work since the 1990s, may also recall that I have long explained the driving force behind my writing: the stark absence of adequate intellectual engagement with the pressing issues of peoples’ lives. Mere interpretive work in the jargon-laden language of social science does not constitute genuine intellectual engagement; it must reflect a dialectic between theory and practice to uncover hidden dimensions, and at the very least demonstrate analytical rigour.
This realisation developed gradually—first as someone subjected to the system from the wrong side, and later as a civil rights activist committed to challenging it. It is therefore natural that my views would appear contrarian to those whose perspectives are shaped by conventional wisdom. Indeed, that has been the very raison d’être of my interventions.
I have repeatedly stated, in both speech and writing, that had others addressed these concerns with the seriousness they demand, I would have been content to pursue my own professional work—grounded, unlike most, in systems science and frontier technologies, which I practised for a living and taught for a time in academic institutions. The very fact that I write or speak is to foreground questions that remain unasked or unanswered.
Therefore, while Deshpande’s observation that I possess a contrarian streak is not inaccurate, the insinuation that accompanies it does not hold. I do not write to differ for its own sake; I write because the issues compel it. I would never reproduce what others have already said—least of all to validate the state’s position. That, to my mind, is not the role of a public intellectual, and certainly not mine.
Anand Teltumbde is a scholar, writer, and civil rights activist.
