Can more knowledge about our society, about the individuals and groups who constitute it, be a bad thing?

I’ve been wondering about this lately, in the context of two government initiatives to gather more knowledge about us Indians, as caste groups and as individuals. Both of these information-gathering exercises—the proposal for a “caste census", which has generated a stormy argument, and the merely desultory discussion over the planned Unique Identification number (UID) for every Indian—has implications for our sense of what it is to be a citizen, and for the terms of the social contract that holds us together as a nation.

Common denominator: (top) A census official visits a family in a village near Guwahati, Assam. Anupam Nath / AP; and people from the Dalit community at a World Dignity Day event in Delhi in 2005. Gurinder Osan / AP

These two debates raise a series of common questions: What do we need to know about our society to make it a better one? What are the dangers and costs to certain types of knowledge? And are we prepared to shoulder these costs? The problem is that we’ve got worked up about only one initiative, the caste census, when the other initiative is the one that speaks more urgently to our future.

The counting of caste groups was first undertaken in a systematic and exhaustive manner by the British, and gave statistical reality to the operative motto of the empire: that India was so fractured by caste that only the grip of imperial rule could keep it together.

Counting castes was a trial for the British census officers. Their questioning elicited many thousands of self-descriptions, including sub-castes, sects, lineages and jatis, which the census men pruned down and ranked as “castes". To some of these castes, the British awarded social and economic privileges, so that politics in the colonial era revolved around caste groups petitioning the British for preferential categorization.

Congregations: (top) Thousands of low-caste Hindus participating in mass conversion to Buddhism in Mumbai in 2007. Rajesh Nirgude / AP; and Rajeev Goswami after he set himself alight during the 1990 anti-Mandal Commission agitation. AFP

At independence, the Indian state decreed caste abolished. Although the 1948 Census Act makes no mention of what categories should or should not be enumerated, the 1951 Census broke with the colonial census tradition and did not count individual castes. The exact reasoning behind the decision remains a mystery of history—in part because the relevant documents weren’t transferred to the National Archives, in yet another instance of our recent history disappearing into ministerial dustbins. But census-makers no doubt wished to reinforce the Constitution’s abolition by fiat of caste, turning a Nelson’s eye to the existence of caste in the hope that it would gradually fade out in favour of a common citizenship.

However, the counting of two social groups subject to particular social and economic deprivation was continued. The scheduled castes, those castes marked by the stigma of “untouchability", and the scheduled tribes, outside the Hindu caste order altogether, were enumerated and made the recipients of state policies of positive discrimination.

Debate about reintroducing caste counts was reopened from the early 1980s, with the invention of new hold-all categories such as the Other Backward Classes (OBC), designed to identify other castes subject to systematic inequality, who therefore had a claim to benefit from positive discrimination. The BP Mandal Commission repeatedly asked governments to compile detailed figures on the OBC population, in order to validate percentages set aside for quotas. But, although legislation was enacted to expand the use of quotas to include OBCs, no new figures at the national level were produced (today, the best estimates put India’s OBCs at around perhaps 45% of the population). So we have a policy, but no clear sense of the people who are the policy’s target.

In the current debate over caste in the census, all parties agree that they wish to see the abolition of caste; and all share a concern with remedying the systemic inequalities of our society: with providing at the very least equality of opportunity, “a level playing field" for all (all agree too that Dalits and tribals should continue to benefit from affirmative action policies). The differences turn on what they judge to be the best means to get there.

There are three broad positions. Some thinkers are entirely opposed to counting caste, and argue that we must move to more universalist policies to address inequalities. According to this view, giving caste groups the imprimatur of the census serves only to harden the identities which are themselves opportunity traps. It does little to bring disprivileged groups into the social and economic mainstream, and reinforces political mobilization along caste lines; as such, it fosters resentment and undermines any sense of common citizenship.

Others argue that the census must be used to produce a detailed caste enumeration of the OBCs. Such data, they argue, will reveal that the OBCs don’t form a homogenous bloc subject to equal deprivation. Some within this broad category—for example, those who own land—are doing quite well, while others are clearly not. More data, it’s plausibly argued, will help to identify the genuine from the spurious claimants to positive discrimination. This is a view that places positive discrimination at the core of India’s efforts to address inequality—but asks that it be more precisely targeted.

Finally, still others are calling for a full caste census, arguing that a complete count of all castes is the only non-discriminatory form of caste enumeration. This view seeks to politicize all caste categories and to disabuse those (upper castes) who believe themselves somehow to be “casteless". Only by counting all caste groups can we come to acknowledge the pervasive reality of caste. Merely to count the lower castes is to perpetuate a discriminatory order. In this view, “annihilating" caste—to use Ambedkar’s verb—requires nothing less than full-on confrontation between the upper and lower castes. It is in the end only political struggle, not law, that can rid us of caste.

I think there is force to the case for OBC enumeration. Given that we already have extensive affirmative action policies, it seems essential to have the basic empirical data to help us judge those policies’ effectiveness. How many people qualify for this affirmative action? Who exactly is benefiting from such policies and who is getting left behind? Without this information, it is impossible to assess and improve our policies.

Some proponents of OBC enumeration hope to show that perhaps half of those today classified in the OBC category are doing well enough not to justify being recipients of positive discrimination. They see hard data will be a basis on which to exclude those who are better off, and to direct resources more precisely at the truly needy. However, I wonder if this is a realistic reading of the nature of caste politics.

Is it credible that simply collecting the empirical data will be sufficient to induce the better off OBC groups to renounce their reserved benefits? It’s far more likely that they will mobilize in order to preserve their quotas; and they will certainly find political entrepreneurs willing to defend their interests in return for votes. We will have a more fragmented and more vicious mobilization of politics along caste lines.

Second is the fact that the reserved sector to which the OBCs are struggling to gain access is fast shrinking. Quotas apply to the formal economy, and within that to the public sector; and to places in higher education. The public sector employs around 20 million people; there are around 10 million students pursuing higher education. Positive discrimination in the form of quotas has diminishing returns—and as such, it should be allowed to fade out naturally. But the counting of OBC castes will generate pressure to extend reservations into the private sector, when we ought to be thinking of quite different policies to deal with inequalities.

Third, there is the matter of the social and political costs of enshrining caste counts in the census. The census is, precisely, a “ritual of citizenship": the one moment when the state and every citizen encounter one another. Should we make this encounter one where the majority of our citizenry have to account for themselves in caste terms? Is that the message we want our state to convey: that it’s interested in our caste?

I think not. Rather, I think we need to collect empirical data on OBC castes by means other than the census: by academic studies, special commissions and reports. We may well lose something in accuracy and authority. But I’d argue it’s a necessary discount. As one of our few tangible expressions of citizenship, the census needs to maintain—and to be seen to maintain—an indifference to caste identities. In our politics, it need hardly be said, symbols and symbolism matter.

The Census of India symbolizes a certain way of thinking about what India is, what it is to be a citizen of this society and state. That society and state recognize a range of diversities among its members—and so the census rightly enumerates gender, language, religion, place, occupation, education. These are self-descriptions that make us who we are, and are part of our identity as citizens. It has also enumerated those who have been subject to the most scandalous feature of our social order: the practice of treating some of our fellows as sub-human, through the crime of untouchability—in order to give them a special push towards becoming full citizens. But it never recognized caste, because it saw that as disintegrative to the idea of citizenship.

To advocates of a caste census of OBCs, opposition to such a measure is seen as a peculiarity of anxious liberals, perplexed by the workings of real politics. After all, isn’t caste ordinary—just one more form of identity available to Indians, and one among several indicators of social disadvantage (which would include gender, region, religion, class)? It follows that we shouldn’t essentialize caste—that’s to fall into the trap of advocates of caste politics, those who see it as the only reality of Indian society. Rather, we need to treat caste as a sort of administrative category: and enumerating the OBC castes, in this view, is essentially an administrative matter rather than one that goes to the foundations of identity and citizenship.

This is disingenuous. Caste in India is not just another form of identity, like any other—it does have a pervasive quality, and it does possess the potential to grip our politics in ways paralleled only by religion. In fact, it is caste—much more than religion—that has proved to be the identity around which our democratic politics has organized itself. A caste census would further entrench this; it would deepen the nexus between caste, electoral politics, and the pursuit of legislative favours.

Yet I think the strongest case against a caste census is the fact that persisting with policies of positive discrimination and reserved quotas is no longer the best way to construct a more just society. Instead of continuing to tinker with reservation policies, we’d do better to write a new social contract for ourselves, based on a more universalist approach to justice. Instead of arguing for privileges for some, we should be redesigning the state so that it works towards providing adequate public goods—above all, education—for all.

In working to build a new social contract, founded on a universalist approach, the ability to individuate our citizens is fundamental. And for this, the Unique Identification number is an important tool. That’s not to say that there are not dangers inherent in it. All forms of knowledge, especially those collected by a state and linked to state power, contain the potential for pernicious misuse. But it is that debate, about the utility as well as the perils of the UID project in the task of building a new idea of citizenship, that we should be having today. It’s more future-directed than the argument that distracts us now, about whether or not to revert to a practice that kept the British busy—and us divided—well over a century ago.

Sunil Khilnani is the author of The Idea of India and is currently working on a new book, India in Search of Wealth and Power. Write to him at