On 24 March 2026, the Supreme Court dismissed the appeal of Chinthada Anand, a Madiga pastor from Guntur (Madiga being a Scheduled Caste), who had lodged a case of being assaulted and abused by his caste name. Before the court weighed the gravity of the allegations, it set itself a different task, framed in its own words as deciding “the conditions under which a person, who has undergone religious conversion, may avail the statutory benefits granted to the members of the Scheduled Castes [SC] and Scheduled Tribes [ST]”. One question, asked of two categories—the SC and the ST—that the law does not treat as one.

Anand’s own facts were not in dispute. He was “originally born into the Madiga community of Scheduled Caste”, the court found. A tahsildar had already issued him a certificate to that effect. He had for years professed Christianity and served as a pastor. The court decided that his professing Christianity extinguishes the legal weight of birth and caste certification: “the caste status [...] stood eclipsed in the eyes of law”.

Being a pastor is not evidence against Anand having suffered caste violence; it is a plausible account of why he suffered it.

Built into that sentence is a reality the court never examined: that professing Christianity and being Madiga are not mutually exclusive. The religious bar itself is not the court’s invention. Clause 3 of the 1950 Presidential Order states it directly: “no person who professes a religion different from the Hindu, the Sikh or the Buddhist religion shall be deemed to be a member of a Scheduled Caste”; and a line of precedent (Guntur Medical College (1976), M. Chandra (2010), and C. Selvarani (2024) has settled this position. What the court originates on its own authority is narrower: the judgement’s declaration in Postulate (d), that being Christian and being Madiga are “mutually exclusive” as a matter of law, and the three-part reconversion test at Postulate (e), which treats a tahsildar’s certificate as worthless absent full ritual renunciation and community re-acceptance. Having built that machinery, the court spends its attention on the question those postulates manufacture: whether Anand’s profession of Christianity extinguished a caste status the tahsildar’s certificate had never put in doubt.

 The Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989 (SC/ST Atrocities Act), did not require the court to ask and answer this question. The violence itself, meanwhile, was handled at the level of technicalities: whether the injury was simple or grievous, whether the allegations of assault and restraint were corroborated. Being simply hurt does not make an offence less serious, and whether the corroboration was sufficient is a question for a trial to answer. That trial never took place.

Here is what the Supreme Court had the occasion to understand, and did not: that being a pastor is not evidence against Anand having suffered caste violence; it is a plausible account of why he suffered it. Visible religious difference—moving away from what a dominant-caste Hindu community expects—is exactly the kind of departure that invites the atrocities the SC/ST Atrocities Act exists to prevent. The court’s causal arrow runs backwards. What it treats as disqualifying him from protection is, on a more careful reading, evidence of exactly what that protection is for.

Trace how the court poses the question of disqualification, and the drift becomes visible. “The question as to whether a person is a member of Scheduled Caste or Scheduled Tribe,” it states, “remains an absolute question of fact”. Yet, by the time the judgment reaches its postulates, “member” has changed to “belong”. One of these postulates puts it directly: “the claimant must demonstrably belong to a caste or tribe which is specifically notified and recognised [...] and such status must be established by clear, cogent, and unimpeachable evidence”.

The Word That Disqualifies

The word that disqualified Anand was “belong.” It is a word that appears nowhere in the sources that create the protection—not in the SC/ST Atrocities Act, not in the Constitution, not in the Presidential Order of 1950. Section 2(1)(c) of the act defines “Scheduled Castes and Scheduled Tribes” by reference to clauses (24) and (25) of Article 366 of the Constitution. The judgment never considers this. Not once in this chain of definitions does the law say “belong”. 'Belonging' is not a word the law inherited. The courts introduced it.

Rohith Vemula’s case shows [...] how unstable the 'discovery' actually is [...] there was no fact of the kind “belonging” claims to name—only a series of authorities, each empowered to declare an answer, arriving at different ones.

Of the law’s words, “deemed” is the more precise one to hold onto. Article 341 of the Constitution empowers the president to specify which castes “shall [...] be deemed to be Scheduled Castes”. The court’s own postulates in this case still use the words from the 1950 Order applicable to the category: “no person who professes a religion other than Hindu, Sikh or Buddhist shall be deemed to be a member of a Scheduled Caste”. Belonging has not replaced deemed. It has been layered on top of it, as a further condition.

Tracing the word ‘belong’ back from the Chinthada Anand judgment leads not to a statute but to a retired judge’s book. Justice S.B. Wad's Caste and the Law in India (1984) is cited as background in Punit Rai v. Dinesh Chaudhary (2003), a Supreme Court decision on an election dispute over a reserved Bihar assembly seat. The winning candidate had claimed to be Pasi, a Scheduled Caste, based on his mother’s caste. His claim was disputed because of his father’s caste, Kurmi, which is an Other Backward Class (OBC). The court held that caste descends through the father under customary law, unless displaced by statute.

In a concurring opinion in Punit Rai, Justice S.B. Sinha made two distinct moves. He named Article 341’s “deemed to be” mechanism itself a “legal fiction”. The Constitution, on his account, does not discover caste membership but constructs it by notification. Separately, quoting Wad’s book, he reached for the language that “a person belongs to a caste in which he is born”. Sinha did not himself claim that caste membership is a fact courts discover, the way a birth date or a boundary line is discovered. His first move said the reverse: that the status is made, not found. Yet “belonging”, borrowed from Wad without Wad’s own qualifications (Wad treats reconversion, community non-disapproval, and lived acceptance as capable of altering caste status), has hardened, in every judgment since, into something closer to fact than fiction. What was invoked as a construct is now applied as a discovery.

Also read: On Annihilation of Caste, Are We Asking the Right Question?

Rohith Vemula’s case shows what that hardening costs, and how unstable the 'discovery' actually is. After his death by suicide in 2016, four separate official inquiries reached four different conclusions about whether he was Dalit. The Guntur deputy collector and the National Commission for Scheduled Castes both affirmed his Mala (SC) status. A second, state-ordered inquiry—convened without stated reason—found him OBC, weighing two testimonies over five to the contrary, amid allegations of political pressure. A 2024 police closure report went further still, asserting that a scrutiny committee had found his family’s SC certificates “fraudulent”. What these reversals show is not that fact-finding failed, but that there was no fact of the kind “belonging” claims to name—only a series of authorities, each empowered to declare an answer, arriving at different ones.

The Supreme Court lays down that rule to settle a question about the ST category that was absent from the case before it [...] in the same postulate where it explains why the two categories should not be read as one.

A few months after Punit Rai (2003), in State of Kerala v. Chandramohanan (2004), that language of belonging crossed over to the ST category. An eight-year-old girl’s father was a member of the Mala Aryan tribe that appears in the ST category. The Kerala High Court held that the girl was not entitled to protection under the SCST Act because her parents professed Christianity. The criminal proceedings under the act were quashed.

The Supreme Court, reviewing that quashing, hardened the reasoning into settled practice. In its account of what tribal membership requires, it turned to W.H.R. Rivers via the 1961 edition of the Encyclopaedia Britannica and to further ethnographic sources. The result was a judicial test for what an “authentic” tribal member looks like.

Rivers’ 1961 definition supplied the ethnographic checklist: a common dialect, a single government, cooperation in shared purposes such as warfare, a common name, a contiguous territory, a uniform culture, a tradition of common descent. Chandramohanan added a second layer: continued adherence to the tribe’s customary law of succession, inheritance, and marriage, and, decisively, acceptance by the community itself. A claimant has to clear both registers—anthropological description and communal ratification—neither of which the Presidential Orders or Article 366(25) require.

By 2026, the judgment in Chinthada Anand, originating in the SC category, is doing the rule-making for the ST category. The court does not simply borrow across the two categories without noticing—it names the difference between them and reaches across it anyway. Postulate (g) does two things at once. It observes that, unlike the Scheduled Castes Order, “the Constitution (Scheduled Tribes) Order, 1950 does not prescribe religion-based exclusion,” and that tribal status must instead turn on “the essential attributes of tribal identity”—a correctly stated distinction. And in the same postulate, the Supreme Court lays down that rule to settle a question about the ST category that was absent from the case before it: Chinthada Anand claimed to be Madiga, a Scheduled Caste—not a tribal claimant. The court volunteers a rule nobody asked it to settle, in the same postulate where it explains why the two categories should not be read as one.

Reification and Theft

There is a name for what the Chandramohanan judgment did in reaching for outside ethnography to define who counts as authentically tribal, and it comes, unexpectedly, from a jurisprudential source elsewhere. Cheryl Harris, writing on whiteness as a form of property in American law, borrows from Georg Lukács the concept of reification—a social relation that “takes on the character of a thing,” acquiring a “phantom objectivity” that conceals its own origin in a relationship between people. Rivers and Britannica do not describe the Mala Aryan community; they constitute, for the law’s purposes, what a Mala Aryan is.

Judicial belonging entrenches the boundaries fraternity is meant to dissolve. It does so, moreover, by a route Ambedkar warned against.

Harris also identifies a pattern Indian courts have reproduced with striking fidelity. Writing of whiteness, she shows that American law “recognized and codified racial group identity as an instrumentality of exclusion”—using group identity to bar non-whites from asserting rights-claims grounded in that same kind of identity—“while refusing to recognize group identity when asserted by racially oppressed groups as a basis for affirming or claiming rights.” The Indian courts’ version runs in the opposite direction but keeps the same structure: religious identity, once it is Anand’s, is fully legible to the court and instrumentally used—not to affirm a claim, but to extinguish one. Anand’s profession of Christianity is fully legible to the court—“indubitably” so, in the judgment’s own words—the moment it is used to strip him of protection. What the court’s own postulates demand instead, the same “clear, cogent, and unimpeachable evidence” standard set out in Postulate (a), applies only to the status that would have protected him, not to the finding that ended the protection.

Harris identifies a second, related pattern: protected status runs only one way. Whiteness, she shows, functions as a reputational interest courts protected in its own right. To falsely call a white person Black was once an actionable defamation, but no equivalent protection ever ran the other way, because the law only guards the status that confers privilege. In Harris’s account, that privileged status is whiteness, named and directly protected. In Chinthada Anand, the equivalent status is never named. There is no certificate for “the Reddy community (OC category)”, the way there is for Madiga or Mala Aryan. What the law’s harder tests for SC and ST protection actually confer is not a named, protected status, but social impunity on whoever stands outside the two categories. This juridically authorised impunity is a caste privilege structurally similar to the protected, unaccountable status Harris traces in whiteness as property.

Also read: How Common Sense Sustains Caste

The court’s marking is administrative; it is a ruling on who qualifies for statutory protection, but its effect is social: the same postulate that removes Anand’s protection simultaneously removes his assailants’ exposure to it. One ruling, two effects, aimed at opposite parties.

Ankit Kawade’s reading of Baburao Bagul’s 1963 story ‘Jevha Mi Jaat Chorli Hoti’ supplies the sharper way to put this. Bagul’s protagonist, an educated Mahar man, does not disclose his caste to the Brahmin and Kshatriya colleagues who assume him to be a Brahmin. When they discover he is an Untouchable, they attack, assault and abuse him. In his own words, he is “robbed” by the same colleagues who had assumed him to be a Brahmin. Caste privilege, Kawade writes, is never something the privileged pay for: “it does not even come at a ‘price’[...] It can only be stolen”.

Belonging presumes an identity freely held—available only from a position of privilege that is never purchased and is never audited. Step outside it, and ‘belonging’ stops describing an identity and starts imposing one: the individual becomes not a person but property — something that belongs, in the sense of ownership, to the rigid boundaries of a category. What is owned is bound, in turn, to perform that category’s caste-prescribed function of subordination.

Fraternity, Inverted

Harris’s framework shows what ‘belonging’ does structurally, how it converts a status into an instrument of exclusion. That does not say what belonging is supposed to do instead. Ambedkar supplies what’s missing.

Ambedkar's fraternity is meant to enlarge who counts as kin—to build the common brotherhood a shared national life requires. Judicial belonging does the opposite: it rewards those who stay legible to a fixed idea of caste and tribe, and penalises exactly the departures.

Annihilation of Caste (1936) treats caste as the thing that makes a shared social life impossible. “You cannot build up a nation”, Ambedkar wrote, “you cannot build up a morality. Anything that you will build on the foundations of caste will crack, and will never be a whole”. His remedy is not incremental. Abolishing sub-castes may only consolidate fewer, larger castes; occasional inter-dining, he observed, “has not succeeded in killing the spirit of caste”. The real remedy, he argued, is intermarriage — “fusion of blood can alone create the feeling of being kith and kin”—and beneath even that, the harder work of destroying the religious sanction that makes caste feel sacred rather than chosen: “the real remedy is to destroy the belief in the sanctity of the Shastras”.

Thirteen years later, addressing the Constituent Assembly on the eve of the Constitution’s adoption, Ambedkar returned to the same doubt, extended now from Hindu society to India as a nation. He reached for an American analogy: a 19th-century church convention that could not agree to pray for “our nation”. Because the word “nation” seemed to claim more unity than there was, they replaced the phrase “our nation” with “these United States”. If Americans could not feel they were one nation, he told the Assembly, it was even harder for Indians: “I am of opinion that in believing that we are a nation we are cherishing a great delusion. How can people divided into several thousands of castes be a nation?”

But he did not leave the thought there. Political democracy, he warned, would mean little without social democracy beneath it—a way of life recognising liberty, equality and fraternity not as separate values but as “a union of trinity”, each one hollow without the others. And fraternity, he said, was the one still to be built: “fraternity can be a fact only when there is a nation. Without fraternity, equality and liberty will be no deeper than coats of paint.”

Belonging, as the courts have built it, works against fraternity rather than for it. Ambedkar's fraternity is meant to enlarge who counts as kin—to build the common brotherhood a shared national life requires. Judicial belonging does the opposite: it rewards those who stay legible to a fixed idea of caste and tribe, and penalises exactly the departures, religious or otherwise, that a person’s life might take. Instead of building fraternity, it entrenches the boundaries fraternity is meant to dissolve. It does so, moreover, by a route Ambedkar warned against on the same day he named fraternity as the republic’s unfinished work: a life, as he put it, of contradictions—political equality granted in form, social equality still to be secured in fact.

Until 1990, anyone who followed Ambedkar into Buddhism was, by the letter of the law, similar to Chinthada Anand: exit into a religion the Presidential Order did not yet recognise, on this practice’s own logic, excluded them from the SC category.

Nothing in the statutes this practice claims to interpret asks for this. The preamble to the 1989 SC/ST Atrocities Act is unambiguous about its purpose: “An Act to prevent the commission of offences of atrocities against the members of the Scheduled Castes and the Scheduled Tribes”. Its predecessor, the Protection of Civil Rights Act, 1955, exists to give effect to Article 17’s abolition of untouchability, and defines “civil rights” by direct reference to that article. Neither statute mentions ‘belonging’. Both define who is protected by pointing back to the Constitution itself, at Article 366, not to any test a court has since invented.

A framework built to prevent atrocities and enforce equality—Articles 14, 15, 16 and 17 together are its constitutional sources—cannot coherently be read to narrow, case by case, who is covered by it. The narrower the reading of membership, the more atrocity and discrimination it leaves unreached. That is not a neutral interpretive choice. It is retrenchment, dressed as legal reasoning.

Also read: Renaming Caste, Retaining Privilege

Ambedkar himself treated religious exit as a live route out of caste, invoking Buddha’s and Nanak’s rejection of Hindu scriptural authority as the model for what annihilating caste would require. Twenty years later, on 14 October 1956, his own embrace of Buddhism made the same point in his life. Yet the postulates governing the SC category bar only conversion to a religion “other than Hindu, Sikh or Buddhist.”That formulation looks like a stable, principled line. It is not. The original 1950 Order recognised only the Hindu religion, with a narrow proviso for four caste groups in Punjab. Sikh religion was added to the general exemption in 1956—the same year Ambedkar embraced Buddhism. The Buddhist religion itself was not added until 1990, by an amendment a further 34 years later.

For those intervening years, until 1990, anyone who followed Ambedkar into Buddhism was, by the letter of the law, similar to Chinthada Anand: exit into a religion the Presidential Order did not yet recognise, on this practice’s own logic, excluded them from the SC category. The religious boundary the court now treats as self-evident is a patchwork of amendments made six and 40 years respectively after the 1950 order came into force. It is not a fixed line at all, but one that has already moved twice, and could move again.

A Judicial Retrenchment of Protection

A word absent from the Constitution, absent from Parliament’s statute, and absent from the two Presidential Orders that define who is protected, has become the instrument by which protection is withdrawn. This is not a doctrine the courts have reasoned their way to. It is an invention—a word one judge called a fiction, picked up and repeated by courts that never asked why, until repetition gave it both the appearance of settled law and the power to retrench.

The case was resolved on a threshold finding—that Anand had “ceased to be a member of the Scheduled Caste on his conversion to Christianity” and that he “cannot be a person aggrieved” under the SC/ST Atrocities Act.

Chinthada Anand had already done everything the actual law asked of him: he held a certificate issued by the tahsildar, confirming exactly what he claimed to be. The court asked a further question anyway, and answered it with a word that does not belong to the statute. But the finding returns to the very word it started with. That word is “membership”, not belonging. Anand’s membership of the SC category is withdrawn all the same. What the courts owe the law they are applying is a reading as expansive as its own preamble, read together with the Constitution that authorises it; not narrowed into hyper-technicality, and not reinvented into tests like “belonging”.

The judgment in Chinthada Anand is where that choice was made, and lost. The court never reached the assault, the abuse, the caste it turned on; the case was resolved on a threshold finding—that Anand had “ceased to be a member of the Scheduled Caste on his conversion to Christianity” and that he “cannot be a person aggrieved” under the SC/ST Atrocities Act—and went no further. Read on the merits, this was an opportunity to address an unresolved question: the Constitutional status of caste-subordinated persons “who profess a religion different from the Hindu, the Sikh or the Buddhist religion” and, on that profession alone, are read out of the preventive framework of the SC/ST Atrocities Act.

That opportunity was an ambitious one. A narrower one was declined in the same judgment: Postulate (g), in the same breath that names the Scheduled Tribes Order as textually distinct from the Scheduled Castes Order, extends a rule built on the SC category to govern the ST category anyway—a question nobody before the court had asked it to settle. Superimposing what applies to one category onto a category it does not fit needs no ambitious theory of statutory reading to count as damage; it is damaging on its own terms.

The court had both opportunities. It took, instead, every opportunity to advance a juridical framework of retrenchment.

Sumit Baudh is a professor of law at O.P. Jindal Global University, and the author of Law at the Intersection of Caste, Class and Sex (forthcoming, Routledge).