India’s democratic experiment has long carried a difficult burden: the state must recognise caste to address inequality, and also seek to transcend caste so citizenship is not permanently bound to birth. It needs caste data to govern justice, but it also needs civic norms that reduce caste as a public marker.
That contradiction has acquired a sharper administrative form. The Union government’s decision to undertake a nationwide caste census, and the Uttar Pradesh government’s directive to remove caste identifiers from police records and public signage after an Allahabad High Court order, appear to move in opposite directions. One expands caste enumeration for policy.
The other reduces caste visibility in public life. Yet both draw legitimacy from the Constitution: recognition in the service of equality, restraint in the service of fraternity. The Home Minister’s statement in Parliament that caste questions will be formally notified in September 2026 ahead of the second phase of the forthcoming Census only underlines that the institutional design is still being worked out.
READ I Caste quotas in private universities back on centre stage
Caste census and constitutional equality in India
This is not a peripheral contradiction. It is built into Indian democracy. Post-independence statecraft has moved repeatedly between two imperatives: acknowledging caste as an empirical category for affirmative action, and trying to weaken caste consciousness as a social force. Social reformers in the late nineteenth century confronted the same problem. Caste could not be dismantled without first naming the inequalities it produced. But freedom also required that caste not become a permanent civic identity.
The current debate on caste census is therefore larger than a contest between political parties. It raises a constitutional question: how can a democracy pursue justice without freezing identity, and pursue fraternity without ignoring inequality? The more useful inquiry is not whether enumeration or erasure should win, but whether institutions can accommodate both. That question becomes especially relevant as India prepares for a caste census.
READ I Caste census is just first step — Govt must act on the data
Caste census history and logic of enumeration
Caste is much older than the modern state, but official counting began with the first full decennial Census of 1872 under British rule. Enumeration was presented as administration. In practice, it also became classification. Over time, the census did not merely record social reality. It helped reorganise it.
The 1881 and 1891 Censuses expanded occupational and sub-caste classifications. Fluid, local identities were often hardened into official hierarchies. Bureaucratic categories acquired legal and political weight. Communities began to treat census labels not only as descriptions, but as claims to status.
Resistance came early. Reformist groups such as the Arya Samaj and Brahmo Samaj urged followers to avoid caste labels and identify instead as “Hindu,” “Arya,” or “nil.” This was not a technical objection. It was a moral and political protest against hereditary divisions. By the 1931 Census, nearly 19 lakh people reportedly entered “no caste” or equivalent responses, suggesting that some Indians already saw enumeration as capable of reproducing the very divisions it counted.
Independence reframed this problem, but did not settle it. The Constituent Assembly recognised caste as a basis for compensatory policy, but did not require universal caste enumeration. After 1951, the Census restricted mandatory caste identification largely to Scheduled Castes and Scheduled Tribes. The state thus acknowledged caste disadvantage while avoiding universal caste classification. The choice reflected an ideological position as much as an administrative one: address historic exclusion without deepening caste consciousness across all citizens.
The 2011 Socio-Economic and Caste Census tested wider enumeration but exposed the limits of weak design. Data quality problems, duplication and classification inconsistencies made the exercise politically combustible and analytically unreliable. What looked like a technical exercise quickly became a dispute over how caste should be recognised in a democracy. Since then, caste enumeration has been less a statistical issue than a constitutional and political one.
READ I Caste and power: Census 2027 prepares for a historic reckoning
UP caste identifiers and the policy paradox
Recent developments bring that old tension back into view. In September 2025, the Uttar Pradesh government issued a directive, following an Allahabad High Court order, to remove caste identifiers from police records, vehicle stickers and public signage, except where such reference is legally required under the Scheduled Castes and Scheduled Tribes (Prevention of Atrocities) Act. The court’s concern was not abstract. It addressed the routine public display of hereditary identity in civic and official spaces.
At the same time, the Union government moved toward a nationwide caste census, the first comprehensive exercise of this kind in decades. The stated rationale was developmental and administrative: better data for welfare targeting, and a clearer picture of the social and economic distribution of backward and intermediate castes. Regional parties and social coalitions have long argued that without updated caste data, claims to representation and resource allocation rest on outdated estimates.
Set side by side, these two moves appear ideologically inconsistent. One seeks to minimise caste in public life. The other seeks to measure it more fully. But both emerge from the same constitutional terrain. The UP directive reflects an aspiration toward civic fraternity. The caste census reflects the state’s obligation to address structural inequality with evidence.
Political positions have been less symmetrical. The Bharatiya Janata Party, in power at the Centre and in UP, has historically shown caution on caste census demands, but electoral and welfare realities have pushed it toward accommodation. Parties that strongly support caste enumeration for representational justice, meanwhile, have often viewed restrictions on public caste display with suspicion.
This is not only expediency. It reflects a deeper tension in Indian democracy between social justice and civic abstraction. As Marc Galanter argued in Competing Equalities, the Indian state has long worked in a dual mode: it recognises caste for compensatory justice, while refusing to endorse caste as a normative social ideal. The present moment extends that logic. Caste must remain visible enough to measure inequality, but not so central that it becomes the permanent grammar of citizenship.
That is the real policy problem. Enumeration can support targeted justice, but it can also harden classification. Erasure can advance civic fraternity, but it can also obscure lived disadvantage. The institutional challenge is to recognise inequality without turning caste into a permanent state-authenticated identity.
Constitutional equality, caste choice and civic dignity
The Constitution’s treatment of caste was never meant to reduce citizens to inherited categories. It recognised caste because caste structures exclusion. But its larger aim was to secure dignity and equal citizenship.
Article 15 captures this dual design. It prohibits discrimination on grounds including caste, while also permitting special provisions for socially and educationally backward classes. Equality, in the Indian constitutional framework, is not only non-discrimination. It also requires corrective intervention.
Article 21 adds another dimension. Judicial interpretation has expanded it to include dignity, autonomy and control over personal information. In Justice K.S. Puttaswamy (Retd.) v. Union of India, the Supreme Court affirmed privacy as a constitutional right and recognised individual control over personal data disclosure. Read together, these constitutional commitments suggest that the state may legitimately seek caste information for justice, but cannot treat identity disclosure as normatively neutral in all contexts.
This constitutional logic aligns with the idea of principled distance associated with Rajeev Bhargava: the state intervenes to correct social disadvantage, but does not preserve or impose identity beyond what justice requires. In practice, India has applied this unevenly, but the principle remains analytically useful.
There are also signs, still limited but notable, that some citizens are asserting identity as a matter of choice rather than inheritance. In 2019, Sneha Parthibaraja in Tamil Nadu obtained official recognition as “No Caste, No Religion.” Such cases are exceptional. They do not prove caste has weakened. But they do signal a civic claim worth taking seriously: that a citizen may seek equal membership without affirming inherited identity in personal self-description.
That is why the caste census debate is not only about data architecture. It is also about citizenship. Can the state acknowledge inequality without fixing identity beyond necessity? Can citizens claim dignity and agency within a justice framework that still requires social classification? Those questions now sit at the centre of census design.
Caste census design and the NOTA option
If the Constitution requires both recognition of caste-based inequality and respect for individual autonomy, then institutional design becomes decisive. One limited but conceptually important possibility in the forthcoming caste census is a “None of the Above” (NOTA) option for caste identification.
Such an option would not end caste census. It would record, alongside caste affiliations, the decision of some citizens to withhold caste identity. That distinction matters. A blank field is missing data. A NOTA response is an explicit choice.
Its value is interpretive, not merely procedural. A NOTA option could register a small but meaningful social tendency: participation in a state exercise without endorsement of inherited identity categories. This would allow the census to capture not only caste distribution, but also a dimension of how citizens relate to caste as identity.
The analytical gains could be significant if interpreted carefully. A measurable NOTA response, mapped across age, education, gender, occupation and region, could help researchers study whether changing social conditions affect the salience of caste identity. The same dataset could then show both persistence and refusal, hierarchy and distance, continuity and transition.
But such a response should not be romanticised. NOTA would not automatically signal progressive politics. For some respondents, especially from privileged groups, refusal may reflect discomfort with recognising caste advantage rather than a commitment to equality. The point is not to treat refusal as morally superior. It is to make the choice legible, and therefore open to analysis rather than assumption.
Seen this way, a NOTA option is less a grand reform than a narrow institutional innovation consistent with constitutional pluralism. It allows the state to gather data needed for redistributive justice while acknowledging that some citizens may seek greater autonomy in identity disclosure. It does not resolve the tension between equality and fraternity. It registers it.
Equality, fraternity and caste census
India’s democratic journey has not followed a straight line. It has had to pursue equality by confronting inherited disadvantage, and fraternity by loosening the hold of inherited identity. Caste sits at the centre of both tasks.
The current debate on caste census reflects that unsettled balance. The demand for updated caste data arises from a continuing structure of inequality and the need for evidence-based welfare and representation. Efforts to reduce caste visibility in public life arise from a parallel aspiration: to build a civic order less organised around birth.
Both impulses are real. Both are constitutional. Neither can simply cancel the other.
That is why the present moment should be read not as a choice between counting and erasing caste, but as a test of institutional maturity. A democratic state must know enough to correct injustice. It must also avoid converting categories of redress into permanent civic identities.
A modern census, like a modern republic, has to do both. It must count inherited disadvantage. It must also leave room for agency, including the possibility that some citizens seek distance from inherited labels. The measure of democratic maturity may lie not in resolving this tension, but in designing institutions capable of holding it.
Amit Singh Khokhar is Assistant Professor, Delhi Skill and Entrepreneurship University.