Tag Archives: State

“The purpose of a system is what it does,” Stafford Beer once said: not what its charter claims, not what its founders intended, not what its defenders insist it’s secretly working toward once the numbers look bad. The line comes from Beer’s work in cybernetics and systems theory, but it travels well, because it was really a warning against taking any institution’s word for itself. Apply it to the state and the sentence stops being quippy and starts being uncomfortable. A state does not do what its Constitution’s preamble promises. It does not do what its bare acts say it must. It does what its politicians ignore until an election forces a promise out of them. It does what a police force using violence on citizens with impunity does. It does the slow torture bureaucrats in government offices administer by attrition. It does the routine ignoring of injustice courts specialise in. Everything else is marketing.

A state’s default setting isn’t a neutral one, but it isn’t a fixed conspiracy either: it simply doesn’t move unless it’s made to. Every gain anyone has pulled out of a state came after someone made not conceding more expensive than conceding.

Reservation itself carries the memory of that fact in its own founding story. Ambedkar’s original demand, at the Round Table Conferences, was a separate electorate for the Depressed Classes: our own constituencies, our own voters, our own representatives answerable to no one else’s veto, a political existence independent of Hindu society rather than embedded inside it. The Communal Award of August 1932 granted exactly that. Gandhi, in prison, announced a fast unto death to force its withdrawal, blackmailing the Depressed Classes and its leaders as well as the British rulers, on the ground that a separate electorate would permanently divide Hindu society. Ambedkar held out against a starving man for days before the political pressure, not the argument, became unbearable on practical grounds, and what emerged from the negotiation at Poona was not the separate electorate but reserved seats inside a joint electorate: more seats than the Award had granted, 148 in the provincial legislatures instead of the original allotment, but folded back into the same political body Ambedkar had tried to exit. Reservation was not the original ask. It is what remained after the original mandate was made impossible by someone else’s leverage and the Hindu moral bankruptcy behind it. The instrument defended for ninety years since it has itself been a settlement, arriving at under duress, from a wider claim.

Rahul Sonpimple, founder of the All India Independent Scheduled Castes Association (AIISCA), made a version of this point at a press meeting on private sector reservation in Delhi on 27th September, framing reservation as something closer to a treaty than an entitlement: a thin, fragile thread and the only one tying Dalits, Adivasis and OBCs to the Indian state, functioning not as generosity but as a negotiated peace. It buys the state and the dominant castes who still hold most of the country’s assets a measure of calm from a majority whose rightful share has otherwise been denied, while giving the numerically dominant Bahujan population some working proxy for the democratic weight their numbers should carry. A contract, not a gift. Except a contract is only as good as what a party can be made to perform under it, and a state that already treats its own preamble as marketing will treat a contract no differently.

A contract implies parties, and a state, unlike a private employer, actually has one with its citizens. It demands loyalty, taxation, submission to its law, alongside the monopoly to dictate the political fates of those within its territory, in exchange for a set of standing obligations back. A private firm asks for labour under a contract either side can in principle at least walk from, and owes nothing back to someone who was never bound to it in any comparable way. That asymmetry is a large part of why a claim on the state’s own hiring is different in kind from a claim on a firm’s, not because the state behaves more responsibly in practice, as we are all well-aware, but because it owes something on the terms of the relationship itself. Although an obligation the state owes but doesn’t perform is worth exactly as much as a preamble it doesn’t act on.

A private firm is not identical to the state, and the case for reaching it anyway doesn’t depend on pretending otherwise. It rests on two separate facts. The economy itself has changed since the framers drew the line at public employment, and a social justice framework confined to the public sector now leaves most of economic life untouched, because most of it no longer sits inside the public sector at all, a gap widened further by the same state’s ongoing privatisation of what public capacity remains, shrinking the one terrain where proportional representation currently carries legal force. But the second fact runs deeper and doesn’t depend on how recent liberalisation is: a private sector built on land, credit, licenses and infrastructure the state provides, privatising its profits while routinely and unsparingly socialising its costs through subsidy, tax concession and public investment, has never sat outside the state’s reach in the first place, this was true of Indian capital before 1991 as much as after it. A firm that depends on the state which itself takes from the public commons for that much can never be allowed to answer to nothing beyond its own balance sheet.

Even so, there’s a real limit to what representation alone produces. Elevated individuals from a movement or caste often become instruments of the institution’s own legitimation rather than being accountable to the base that elevated them, and when they try to push further beyond the constraints and the logics of those systems, they get cut down by the sheer weight of social forces unwilling to tolerate the progress already made. Despite that, political reservation has still converted into something more durable than individual token presence, an independent structure of Bahujan power that didn’t exist before it, even as oppressor-caste violence against that rise has grown alongside it. What government service has that the private sector doesn’t is exactly this: decades of organised structure built up around the reserved seat, structure that keeps contesting the ceiling even where it doesn’t break through it. The private sector has never had an equivalent foothold, and no amount of employer voluntarism has ever built one. That’s the case for statutory reservation even in its thinnest form: not that a quota alone delivers equality, especially substantively, but that it’s the minimum legal anchor around which durable, organisable power can be built at all.

Against that backdrop, private sector reservation is not a new demand groping for its first hearing. It has been formally raised for close to two decades. The All India Confederation of SC/ST Organisations was pressing the case as early as 2012, organising a march on Delhi backed by three dozen SC/ST organisations, arguing plainly that no nation stays strong while its economic and social levels remain this split. But even before that, in 2003, then Prime Minister Atal Bihari Vajpayee addressed a meeting of SC and ST MPs and said that the private sector cannot exist in isolation and has to be responsive to the needs of the society, making it essential to have reservation in the private sector. This was followed by a broad consensus among electoral parties across the ideological spectrum on the matter, even though it has not yet been translated into political will or substantive action. One concrete effort came in 2006, under Prime Minister Manmohan Singh, when he appealed directly to Indian industries to become the government’s partner on the question. However, rather than legislate a mandate, the government let industry regulate itself: CII, FICCI and ASSOCHAM each adopted a Voluntary Code of Conduct built around four pillars, Employability, Education, Entrepreneurship Development and Employment. That October, the PMO had already constituted a Coordination Committee on Affirmative Action in the Private Sector, chaired by the Principal Secretary to the Prime Minister, drawing on the Department of Personnel and Training, the Ministry of Social Justice and Empowerment, the Ministry of Tribal Affairs, and serviced by the Department of Industrial Policy and Promotion. The Committee has met nine times since. By 2010, the government’s own review found adoption of the voluntary codes was slow; by 2013, it was still describing the same dialogue as ongoing; by 2015, industry associations were reporting hundreds of member companies signed on. Twenty years of this arrangement has produced a voluntary code without enforceable duties, and industry associations professing over a lakh of beneficiaries with no verifiable representation data to back the figure. Now we cannot sit once again with a renewed appeal for goodwill. What is needed is to go beyond an empty promise and fight for a statutory guarantee: a binding quota, backed by rule of law, not another code any signatory can stop honouring when they feel like it.

On 26 July 2024, that demand reached Parliament directly. Chandrashekhar Azad, MP for Nagina, introduced a private member’s bill titled the Reservation for the Scheduled Castes, the Scheduled Tribes and the Other Backward Classes in Private Sector Act, 2024, one of three private member’s bills he introduced in the Lok Sabha that day. It covers private firms, educational institutions and other establishments employing at least twenty people with no government financial stake. Even this bill stops short of a binding hiring quota: its mechanism is incentive-based, requiring the central government to induce private firms toward reservation through special concessions and lower-interest loans via nationalised banks, with an annual implementation report tabled in both Houses. Structurally, a private member’s bill has almost no chance of becoming law: enactment is the exception, not the rule, and has been for decades. That a bill this cautious, relying on incentives rather than quotas, remains a private member’s bill and is still pending says something about how far Parliament remains from treating this as settled business.

But the case for the demand itself rests on documented discrimination. Sukhadeo Thorat and Paul Attewell’s 2007 correspondence study sent matched applications under upper-caste Hindu, Dalit and Muslim names to private-sector job advertisements in the English dailies and found discrimination operating at the very first screening stage, before any interview, before any assessment of merit beyond a name on a page. Zahra Siddique’s 2011 work went further, quantifying the size of the gap and who was driving it: low-caste applicants had to send 20% more resumes than high-caste applicants to get the same callback rate, with the disparity sharpest when the recruiter was male or Hindu, and more pronounced at smaller firms than larger ones. Axmann, Swanson and Cuspinera Contreras, using Indian Human Development Survey data in 2016, found significant wage discrimination against Dalits and Adivasis specifically inside the private sector. Madheswaran and Singhari, drawing on National Sample Survey employment data the same year, found SC workers earning less than equally qualified forward-caste workers in urban labour markets, with the gap wider in private employment than public, and unequal access to occupations itself doing much of the work in producing the disparity. Pattayat’s 2024 study of rural non-farm workers, using NSSO and PLFS data spanning 2004-05 to 2019-20, found the earnings gap persisting across that entire period and examined how much of it discrimination itself accounted for.

Article 15(5) already gives the state constitutional room to act on half of this. Inserted by the 93rd Amendment in 2005 and in force since January 2006, it permits special provision by law for socially and educationally backward classes, Scheduled Castes and Scheduled Tribes in admission to educational institutions, explicitly including private institutions, aided or unaided, with the sole carve-out for Article 30(1) minority institutions. Because education sits on the Concurrent List, both Parliament and state legislatures can act under it, and several states have already legislated on private educational admissions. Tamil Nadu enacted a 2006 law providing 69% reservation for Backward Classes, Most Backward Classes and Denotified Communities, Scheduled Castes and Scheduled Tribes in private educational institutions. Andhra Pradesh had already applied SC, ST and Backward Class reservations to admissions in private professional institutions through its admission rules. Karnataka’s 2006 Professional Educational Institutions Act provides for 50% reservation for SCs, STs and OBCs in unaided non-minority professional institutions. Maharashtra enacted a 2006 law providing reservation for SCs, STs, Denotified Tribes, Nomadic Tribes and OBCs in private professional educational institutions, while Haryana’s 2016 Backward Classes Act provides reservation for Backward Classes in admissions to educational institutions, including private institutions.

What that record shows is that Article 15(5) has been given statutory effect at the state level, but unevenly and without a nationwide framework. A Parliamentary Standing Committee report presented to the Rajya Sabha in August 2025 said that private educational institutions are currently under no national legal obligation to implement reservation because there is no central statute mandating it. It recommended that Parliament legislate 27%, 15% and 7.5% reservation for OBCs, SCs and STs respectively in private higher educational institutions, with government reimbursement of the associated costs on the model of the Right to Education Act. Although a sincere Caste Census may tell us that the necessary numbers should be much higher, the constitutional ceiling on reservation itself also needs to be revisited.

What Article 15(5) doesn’t reach at all is employment, and that gap isn’t small: education and employment were never separable; a degree that doesn’t convert into a job is a credential, not a livelihood. One state that tried a binding private-sector quota by its own legislation ran straight into that trap. Haryana’s 2020 State Employment of Local Candidates Act mandated 75% reservation for domiciled candidates in private jobs under a salary threshold, with a 10% district sub-cap built in, not a caste quota, a domicile one. The Punjab and Haryana High Court struck it down in 2023 as unconstitutional and violative of Part III of the Constitution; the challenge to that judgment remains pending before the Supreme Court. The deeper problem predates the ruling: the 93rd Amendment gave the state constitutional room to reach private education specifically and never extended that room to private employment. Domicile was the classification that failed here, not caste, but replacing domicile with caste would not by itself resolve the Article 19(1)(g) problem, because no amendment has yet done for employment what the 93rd did for education. That is what this demand still needs and hasn’t gotten: not just a statute, but a constitutional amendment opening the same room for private employment that the 93rd opened for private education, without which any statute meets the wall Haryana did.

Winning that room would still only secure a narrower claim than the one Ambedkar himself eventually made. Private sector reservation, at its most comprehensive, means reservation in private employment, reservation in private educational institutions, and an Equal Employment Opportunity Act as organizations like AIISCA and others have been demanding: legislation with a working complaint mechanism, an investigation process, ongoing monitoring, and real accountability, reaching recruitment, wages, promotion, workplace treatment, termination, occupational allocation, contract employment and professional advancement, not just the moment of hire. That last piece is the one twenty years of voluntary codes never touched, because a Voluntary Code of Conduct had no enforceable complaint mechanism and no statutory consequence for ignoring it. A statute would. But even this full slate, a foothold at every stage of the employment relationship, asks only for fair terms of entry into an economy whose ownership stays exactly where it already is. It is a claim to be treated fairly inside the room. It has never been a claim on who owns the room.

Ambedkar’s own economic writing sits alongside all of this as one part of a longer record, that expands the horizon of economic fight itself. In States and Minorities, the memorandum he prepared for the Constituent Assembly on behalf of the Scheduled Caste Federation, he proposed a state control over economy in which key industries would be owned and run directly by the state and basic industries owned by the state and run through its own corporations, with land held under state ownership and cultivated through cooperative farming, aimed at dismantling landlordism and the agrarian exploitation caste had built itself around. That program was never incorporated into the Constitution in the form Ambedkar proposed. Although a version of that was found in the Nehruvian socialist project, which was itself eventually given up in the face of global neoliberal onslaught. What it shows, regardless of anyone’s own view of state ownership as a remedy, is how early and how directly Ambedkar located the actual target in who owns the economy, not merely who holds a fair position inside it.

Ambedkar’s economic target, and the demand for reservation in the private sector, concern forms of structural injustice that are harder to see than the moments when that injustice erupts into visible violence. That shapes not only who gets to have this conversation and on what terms, but who is willing to listen to it. Oppressor-caste liberals attend to Dalit-Bahujan life most readily through the news cycle of atrocity, because atrocity is visceral and immediately legible, while the guilt of a grieving conscience often springs into action only when Dalit-Bahujan rage forces that attention. A Dalit-Bahujan body demands attention in a way a hiring statistic does not. A wage-discrimination dataset spanning fifteen years, a Coordination Committee that has gone nowhere, a bill pending in Parliament since 2024, none of these produces the same urgency because none offers an easy villain, a headline image or a clean moment of collective horror to perform. Grieving a corpse asks nothing ongoing of anyone. But there is a real villain here, even if it is harder to identify in any single moment: the state, which owes more than what it does and routinely shows where its priorities lie, alongside the structural violence produced through years of oppressor-caste apathy at best and antipathy at worst.

The historical record here is a caution, not a reason to wait. What survived Poona wasn’t Ambedkar’s original demand: it was a narrower settlement that still had to be fought for afterward, seat by seat, for decades, before it was more than ink. The same will be true of a private-sector statute and an Equal Employment Opportunity Act alongside it. Winning them will not finish anything on its own; a seat is captured as often as it’s built on. Beer’s sentence applies once more, now to the law: a statute’s purpose isn’t what its statement of objects and reasons says either. It’s what the Act does. Whether the complaint mechanism gets used, whether a violation carries any real cost, whether the oversight committee holds its next meeting or goes missing in action the way it already has. None of that follows from any bill becoming an act. It follows only if the pressure that forces the act into existence keeps being applied after it exists. Because a state, as established, only holds what it’s made to hold.

Paresh Hate is a Bahujan queer researcher, writer, documenter and activist and the Lead Content Manager at Dalit Voice.

Nationalist Demands in the New Law

The Prevention of Insults to National Honour (Amendment) Bill, 2026 was introduced by Amit Shah in the Rajya Sabha on 24 July 2026. Its Statement of Objects and Reasons recalls that Rajendra Prasad, as President of the Constituent Assembly, said on 24 January 1950 that Vande Mataram would be “honoured equally” with Jana Gana Mana. It notes that no legal provision existed to prevent interference with the singing of the National Song and proposes to address this by amending Section 3 of the 1971 Act, which already made intentionally preventing the singing of the National Anthem, or disturbing an assembly engaged in singing it, punishable with imprisonment of up to three years, a fine, or both, with a mandatory minimum of one year for a second and subsequent conviction. The amendment extends that same offence, and the same penalties, to the National Song. Article 51A(a) names the Constitution, the national flag and the national anthem as objects of citizens’ respect; it has never named the National Song.

The amendment arrived alongside a separate, longer-running push to change what gets sung. Vande Mataram has six stanzas; since 1937, only the first two have been used at national gatherings, a convention set by the Congress Working Committee. The Ministry of Home Affairs had already moved to reverse that convention on its own, issuing a protocol on 28 January 2026 requiring the full six-stanza version at government, ceremonial and cultural events, well before this amendment existed. That protocol was advisory and, as the courts would go on to hold twice over the following months, carried no penalty of its own. Only after those two challenges failed did the government tighten the position twice more: a July 9 MHA letter to all states specifying that only the official six-stanza version could be recited whenever the song was sung in mass singing, and then, on 24 July, the amendment itself. Where both the song and the anthem are performed at one event, Vande Mataram now goes first. Schools are directed to open assemblies with it, although many schools already practiced this of their own volition. It is to be played at state functions, civilian award ceremonies, flag-hoisting events, and the arrivals and departures of the President and Governors.

The Political Fight Over the Song
The push for the full six-stanza version followed a ten-hour special discussion in the Lok Sabha in December 2025 marking the song’s 150th anniversary, in which the Prime Minister accused the Congress of having “sowed the seeds of division” by trimming the song. In 1937, the Congress Working Committee, after consultation with Mohandas Gandhi, Jawaharlal Nehru, Rabindranath Tagore, Subhas Chandra Bose, Rajendra Prasad and Vallabhbhai Patel, decided that only the first two stanzas would be used at national gatherings because of objections to the religious imagery in the later verses. Tagore had written to Nehru that the first two stanzas had “acquired a separate individuality and an inspiring significance of its own,” while the later verses, read in context, “[are] liable to be interpreted in ways that might wound [Muslim] susceptibilities.” Congress rejected the Prime Minister’s charge, pointing to this 1937 decision and Tagore’s letter and accusing the government of manufacturing the controversy for present gain in the context of then upcoming elections, a charge Priyanka Gandhi Vadra made directly. The Congress Working Committee has since reaffirmed the two-stanza position for its own party programmes.

The dispute has now also moved into a question of federal authority after the Karnataka government ordered on 8 September that only the first two stanzas be sung at state government programmes, except when the President, Vice-President, Prime Minister or Governor is present. Within days, advocate Girish Bharadwaj filed a PIL challenging the state’s authority to prescribe a shorter version, arguing that the matter falls within Parliament’s residuary legislative powers and that the state order conflicts with the 2026 amendment to the Prevention of Insults to National Honour Act. The petition also challenged the procedure by which the Karnataka Cabinet took the decision, alleging that it was approved as an “Informal Discussion” without a Cabinet Note or Law Department advice. Governor Thaawarchand Gehlot then wrote to Chief Minister D.K. Shivakumar, invoking Articles 162, 256 and 257(1) and asking the government to reconsider its order on the ground that it conflicted with the Union’s statutory and executive framework. On 16 September, BJP MLA S.N. Channabasappa walked out of a government programme in Shivamogga after only the first two stanzas were sung.

The opposition, however, has also been weak in resisting the Centre’s position. In Kerala, the Congress-led UDF government initially circulated the Centre’s instruction that Vande Mataram be sung in full during the Har Ghar Tiranga campaign from 9 to 17 August. This put a Congress-led government in the position of appearing to implement the same six-stanza version promoted by the BJP-led Union government, despite Congress’s objections to treating the full version, with its references to Hindu deities, as a compulsory expression of national identity. The CPI(M) criticised the UDF’s position, with Leader of the Opposition Pinarayi Vijayan arguing that the song “in its entirety is not compatible with secularism” and that “there is nothing wrong with reciting the opening stanzas”, and accusing the UDF of surrendering to an RSS agenda. The UDF subsequently clarified that it would not follow the full version, and the state government ultimately omitted Vande Mataram from its official Independence Day function.

The song’s meaning has remained contested within the communities whose religious commitments are most directly implicated by it, a history the government’s “everybody knows what is national song” formulation simply passes over. The distinction made in 1937 remains central to the dispute: the Congress Working Committee accepted the first two stanzas for national gatherings while recognising objections to the later verses and their religious imagery. Muslim religious authorities have continued to make this distinction. Darul Uloom Deoband and Maulana Mahmood Madani have objected to the later verses on the grounds that their depiction of the motherland as divine conflicts with the Islamic doctrine of tawhid, while the All India Sunni Ulema Board ruled in 2006 that Muslims could sing the first two stanzas.

The dispute, however, has not been confined to Muslim objections. During the 2006 centenary controversy, the Shiromani Gurdwara Parbandhak Committee initially instructed its schools not to sing Vande Mataram, with its president arguing that it promoted a particular religion and did not fulfil the aspirations of minorities, before subsequently backing away from that position. In 2026, the Andhra Pradesh Christian Leaders Forum objected to the newly mandated full version, specifically citing its references to Durga and their implications for Christian beliefs. The question is therefore not simply whether Vande Mataram is the national song, but what religious content the state is asking people to affirm when it mandates the complete text. The Home Ministry’s circular treats that underlying question as though it had never been contested. Where it does allude to the religious objection, it reduces it to a question of Muslim reluctance, allowing a disagreement over religious belief to be recast as a question of patriotism.

The Text Behind the Mandate
The distinction between the first two stanzas and the later verses can be traced to how the song was written. Bankim Chandra Chattopadhyay had written the first two stanzas in the 1870s, describing a land of rivers, fruit and gentle winds. He added four more when he incorporated the song into his 1882 novel Anandamath, and those later stanzas carry the divine imagery: the third stanza calls the motherland “protector of millions and destroyer of enemies,” the fourth names her Durga, Lakshmi and Saraswati outright, while the fifth and sixth complete the poem as a vow of devotion.

Anandamath provides the setting in which those later stanzas acquire their political and religious meaning. Bankim Chandra Chattopadhyay is one of the key figures of proto-Hindutva politics in British India, and the novel is set during the Sannyasi Rebellion. Its protagonist, Mahendra, a zamindar separated from his family amid famine and unrest, falls in with the Santans, a band of ascetic soldiers who have taken up arms to liberate the motherland from Muslim rule, with the British entering the story only later. One of Satyananda’s disciples wants to “pull down the city of the renegades” and “throw it into the river.” Bhavananda, inducting Mahendra into the cause, tells him “unless we drive these tipsy long-beards away, a Hindu can no longer hope to save his religion,” and answers the question of method with “we will beat them.” He then leads Mahendra to the order’s temple, where, standing before the idols of Durga and Kali, he tries to fill him with religious fervor for the fight ahead. Elsewhere in the novel, the Santans send emissaries into villages to gather bands of Hindus, then raid Muslim settlements, which are burned “to ashes.” Characters imagine the day a mosque will be razed to build a temple to Radha-Madhava in its place. The four later stanzas of Vande Mataram were written for, and inside, this story. This is the context of the text now mandated by law.

The Limits of the Law’s Imagination
That January 28 protocol was challenged twice. In February, Muhammed Sayeed Noori approached the Supreme Court, arguing that requiring the complete version raised questions of religious freedom and conscience. On 25 March, the Court declined to entertain the petition, holding that the protocol was advisory and carried no penalty for those who chose not to sing, while leaving open the possibility of intervention if the advisory hardened into actual discrimination or coercion. A parallel challenge in the Karnataka High Court, filed by Bengaluru-based advocate Somashekhar Rajavamshi, objected specifically to the inclusion of stanzas three through six and their references to Durga, Saraswati and Lakshmi; the High Court dismissed it on 9 April on the same ground, that the protocol did not make singing the full version mandatory.

The 2026 amendment changes that position. The January protocol could prescribe the full version’s use; it could not, on its own, criminalise a refusal to sing it. The amendment extends the National Song to the Section 3 offence applicable to the National Anthem since 1971: intentionally preventing its singing, or disturbing an assembly engaged in singing it.

On 22 September 2026, the Supreme Court heard a petition by T. M. Krishna, with senior advocate S. Muralidhar appearing for him. The petition challenges the 2026 amendment insofar as it extends the existing criminal provision to the National Song. It also challenges the Home Ministry’s directions requiring the complete six-stanza version at official functions. The petition argues that the amended provision and the directions violate Articles 14 (equality before the law), 15(1) (prohibition of discrimination on specified grounds), 19(1)(a) (freedom of speech and expression), 21 (protection of life and personal liberty) and 25(1) (freedom of conscience and the right to profess, practise and propagate religion).

The Court refused to decide whether the National Song should consist of two stanzas or six, indicating that this was for the democratically elected state to determine rather than for the Court to prescribe. Justice Bagchi also questioned the suggestion that religious references in all six stanzas are, by themselves, incompatible with the Constitution’s secular character, pointing to the reference to God in the US national anthem. That another country’s anthem contains a religious line does not make it a benchmark for what India’s Constitution should tolerate; it only shows that one other state has also let religious language sit inside a national symbol. Whether a lingering religious motif of that kind is compatible with secularism is exactly the question in front of the Court, and pointing to a second, unexamined instance of the same thing does not answer it. The Centre has two weeks to respond, and the Court has not yet ruled on the constitutional validity of the amendment or the six-stanza directions.

At the hearing, however, the Court drew a distinction between the State’s choice of a national symbol and the use of criminal law in relation to participation in it. Nothing in the amendment says that simply refusing to sing Vande Mataram is itself an offence; the crime, as under the original Section 3, is preventing others from singing or disturbing an assembly engaged in it. The question before the Court is therefore whether the protection recognised in Bijoe Emmanuel, which concerned refusal to sing the National Anthem under that same original section, also protects a person who refuses to sing the National Song on grounds of conscience or religion while leaving others free to do so.

That leaves an odd gap. If the amendment truly does not touch a refusal that neither prevents others from singing nor disturbs an assembly, the state loses nothing by saying so outright, and Krishna’s petition should fall outside the law’s reach without a fight. The Centre resisted even the issuing of notice, and Solicitor General Mehta pressed the Court not to entertain the petition at all. That resistance points to a state interest that Section 3’s text does not itself supply: the discretion to withhold, for as long as possible, any confirmation that non-participation is safe.

Bijoe Emmanuel v. State of Kerala, which the Court said still governs the field and has never been challenged, is the case that established that refusal to sing the National Anthem, where the person neither prevents others from singing nor disturbs the assembly, does not constitute an offence under Section 3. In 1986, the Court considered three schoolchildren in Kerala who belonged to the Jehovah’s Witnesses faith. They believed that singing the National Anthem was incompatible with their religious convictions, although they had no objection to the words or sentiments of Jana Gana Mana. During their school’s morning assembly, they stood silently while the anthem was sung. After the issue was raised, they were expelled from school, despite an inquiry finding that they were law-abiding and had shown no disrespect. The Supreme Court held that their expulsion violated their fundamental rights. It also held that there was no law obliging anyone to sing the National Anthem and that their standing silently while it was sung did not constitute an offence under Section 3 because it neither prevented the singing nor disturbed the assembly.

The Court in Bijoe Emmanuel answered for the facts it had: the children stood, in silence, while others sang, and that was respect enough. Whether sitting, or some other form of non-participation, would count the same way was never argued before it, and the judgment does not say. The Court’s later National Anthem jurisprudence went further than Bijoe Emmanuel by imposing a general requirement to stand, but went no further in deciding whether an able-bodied person who chose not to comply had thereby committed an act of disrespect. In Shyam Narayan Chouksey v. Union of India, the Court initially ordered, in 2016, that the National Anthem be played before every feature film and that everyone present stand. It subsequently recognised exceptions for people unable to stand and for occasions when the anthem appeared within a film, newsreel or documentary. In 2018, it made the playing of the anthem before films optional and handed the wider question of its regulation to the government. What remained unresolved was whether an able-bodied person who deliberately remained seated when the National Anthem was played had thereby shown disrespect.

Article 51A(a) requires citizens to respect the Constitution, the national flag and the national anthem. That language concerns respect for specified national symbols, while Section 3 concerns conduct that interferes with their public performance. The Constitution speaks of an attitude of respect; Section 3 defines specific acts that the law prohibits. Bijoe Emmanuel is the only case here that actually examined a specific act of non-participation and ruled on whether it counted as respectful: the children stood silently while others sang. Chouksey, by contrast, imposed a general rule about what everyone present in a cinema hall had to do when the anthem was played, and later withdrew the requirement that the anthem be played there at all. The distinction between respectful and disrespectful non-participation has therefore never actually been adjudicated on its own terms. That leaves the discretion to decide sitting with whoever is enforcing the law in the moment, not with any settled rule a citizen could rely on in advance. And even supposing that gap could be closed, and the state produced a clean definition of respectful non-participation tomorrow, a further question is why a person should have to demonstrate respect for a national symbol through participation of any kind, rather than simply being left alone by the law altogether.

Article 51A(a) itself answers none of this. The Fundamental Duties, inserted into the Constitution by the 42nd Amendment in 1976, are not enforceable in court. No citizen has ever been prosecuted for an abstract failure to respect the Constitution, the flag or the anthem. Section 3 punishes specific conduct: preventing a performance, disturbing an assembly, even though the boundaries of that conduct can still be interpreted and policed according to the discretion of whoever is enforcing the law. Article 51A(a) supplies the word “respect” and nothing else. It does not say what respect consists of, and it does not say respect must be shown rather than merely left unviolated. The state has built a criminal statute that reaches only interference, and rested it on a duty clause that was written to describe an attitude, not to compel an act.

The Right Not to Sing
The present controversy asks a larger question about the limits of state power over national rituals. A person may have a religious objection to the text, reject the use of a national symbol altogether, or simply choose not to participate. If that refusal does not prevent anyone else from singing, what is the state entitled to demand from the person who refuses? Even as the space for conscientious objection shrinks in the shadow of ultranationalism, we must continue to ask: why should the state require that refusal to take a prescribed form of respect, and why can it not leave room for people to decline participation altogether?

The UN Committee on the Elimination of Racial Discrimination (CERD) has raised serious concerns over the treatment of some of India’s most marginalised communities, including Dalits, Adivasi and tribal peoples, Rohingya, Bengali-speaking Muslims, migrants and asylum-seekers. Its findings follow the first review of India by the Committee since 2007 and examine the country’s combined 20th and 21st periodic reports under the International Convention on the Elimination of All Forms of Racial Discrimination.

The findings were issued after CERD’s 118th session, which also examined Finland, Honduras and Kuwait. This was part of the Committee’s regular review of states that have signed the Convention. CERD’s concluding observations are not court judgments and do not have the force of law. They assess how governments are implementing their obligations under the Convention and identify areas where the Committee finds serious problems.

In India’s case, CERD raised concerns spanning caste, ethnicity, religion, citizenship and policing. One of its gravest concerns was the alleged conduct of law-enforcement agencies towards Indigenous and Tribal Peoples, Scheduled Tribes, Scheduled Castes, particularly Dalits, ethnic and ethno-religious groups and non-citizens. CERD recorded reports of excessive force, racially motivated violence, extrajudicial killings, arbitrary and prolonged detention, torture, ill-treatment and sexual violence. It called for prompt, thorough and impartial investigations, with accountability where violations are established.

The reference to Dalits and Scheduled Castes also brings an older disagreement into the review. The Indian government maintains that caste should not be treated as equivalent to race or “descent” for the purposes of CERD. Its submissions nevertheless contain information on Scheduled Castes and Scheduled Tribes and measures intended to address their disadvantages. CERD’s treatment of caste within its examination of descent and discrimination places caste exclusion within the scope of the Committee’s scrutiny, even as the Indian state rejects that equivalence.

CERD also drew particular attention to Rohingya, Bengali-speaking Muslims, migrants and asylum-seekers. It noted an increase in law-enforcement operations involving these groups, particularly after a 2017 Home Affairs order and the April 2025 Pahalgam attack. Reports of police stops and identity checks based on racial profiling allegedly resulted in arbitrary arrests and detention, torture and other mistreatment. The Committee also raised concerns about deportations and forcible returns involving people who may require international protection.

During the review, CERD Country Rapporteur Stamatia Stavrinaki referred to independent monitoring that recorded more than 1,300 anti-Muslim hate-speech incidents in 2025, including recurring rhetoric attributed to senior officials. India’s delegation responded that legislation enacted in 2023 criminalises acts amounting to hate speech and hate crimes, with additional protections applying when offences target Scheduled Castes and Scheduled Tribes. It also pointed to freedom of religion as a constitutional guarantee.

But CERD’s observations focused on what these protections mean in practice. Alongside the government’s account of its legal safeguards, the Committee recorded concerns about discrimination, hate speech, violence and the conduct of law-enforcement agencies. It called on India to address hate crimes against Rohingya, Bengali-speaking Muslims, migrants and asylum-seekers, refrain from collective expulsions and ensure access to international protection. This includes respecting non-refoulement, the principle that people should not be returned to a place where they face a serious risk of persecution or other grave harm.

The findings also address citizenship and documentation. CERD noted the absence of adequate disaggregated data on Indigenous and Tribal Peoples, ethnic and ethno-religious groups, migrants, refugees, asylum-seekers and stateless people. It highlighted the categorisation of Bengali-speaking Muslims as “non-original inhabitants” and reports that they face more stringent scrutiny during National Register of Citizens verification. The Committee also noted that much of India’s available data on Scheduled Castes, Scheduled Tribes and refugees still derives from the 2011 Census, while acknowledging the delayed 2026 Census and its proposed collection of caste data through self-identification.

The treatment of Bengali-speaking Muslims in the NRC process raises a broader question about how citizenship is administered. CERD’s concern is not simply that documentation is required, but that particular ethnic and religious groups are reportedly subjected to different levels of scrutiny. In a political environment where Muslims are repeatedly cast as outsiders or demographic threats, administrative categories such as “non-original inhabitants” can acquire consequences beyond paperwork. The NRC thus becomes part of a wider system in which belonging is subjected to heightened state verification.

The concerns around Rohingya also sit within a longer history of UN scrutiny. A 2025 communication by UN Special Procedures carried allegations that deportation operations against Muslims, including Rohingya, intensified after the Pahalgam attack, involving raids, arrests, ill-treatment and forcible returns. Those allegations came through a separate UN process and were not findings by CERD or a judicial determination.

India is not a party to the 1951 Refugee Convention or its 1967 Protocol and has no comprehensive national refugee law. UNHCR therefore carries out protection-related functions for refugees and asylum-seekers in the country. This absence has become more significant alongside changes to India’s citizenship and immigration regime.

The Citizenship (Amendment) Act, 2019 was defended by BJP ministers as addressing the unfinished business or “wrongs” of Partition. Yet the CAA was not a refugee law. It created a religion-based route to accelerated citizenship for specified non-Muslim communities from Pakistan, Bangladesh and Afghanistan. It did not establish a general system of refugee protection based on persecution or the risk of serious harm.

And in 2025, Parliament enacted the Immigration and Foreigners Act, replacing four earlier laws governing the entry and treatment of foreign nationals. The new law consolidated the state’s immigration powers without creating a corresponding refugee protection framework. CERD’s call for protection against collective expulsions and respect for non-refoulement therefore comes alongside a strengthened legal machinery for controlling and removing foreign nationals, while India still lacks a dedicated process for determining who requires refugee protection. For Rohingya and other asylum-seekers, that leaves a basic question: what domestic procedure requires the danger they may face after removal to be considered before the state deports them?

India’s delegation, led by Solicitor General Tushar Mehta, defended the country’s legal framework and its claims of pluralism and equality. The government reiterated its position that caste should not be equated with race or “descent” under the Convention and pointed to legislation addressing hate speech and protections for Scheduled Castes and Scheduled Tribes.

CERD’s findings place those claims alongside reports of how state institutions operate on the ground. The Committee’s concerns cover law-enforcement violence against Dalits and Adivasi and tribal communities, anti-Muslim hate speech, racial profiling of Bengali-speaking Muslims and Rohingya, and the treatment of people whose citizenship or immigration status is under scrutiny.

For Dalit and Adivasi communities, and for Muslims and other minorities whose belonging is increasingly subjected to state scrutiny, the findings also expose how Hindu nationalism operates through state power. Caste discrimination, communal targeting, citizenship enforcement and policing are often treated as separate questions, even as they converge in the lives of communities whose relationship with the state is shaped by unequal power. CERD’s observations bring these questions into the same examination of discrimination and direct attention towards the institutions through which unequal treatment is carried out.

India’s Foreign Contribution (Regulation) Act was introduced in 1976, during the Emergency, to regulate foreign influence in Indian political life. The 2010 Act replaced that framework with a broader system for regulating foreign contributions to organisations and associations. Over time, FCRA has also given the state increasing power over whether organisations can access foreign funding and continue their work. The latest changes take that process further.

The Foreign Contribution (Regulation) Amendment Bill, 2026 was introduced in the Lok Sabha on 25 March and referred to a Joint Parliamentary Committee on 12 August. It is not law yet. The 2026 Rules, however, are already in force. Together, they significantly tighten the conditions under which organisations receiving foreign contributions can operate.

The Bill proposes a Designated Authority to supervise foreign contributions and assets after an organisation’s FCRA registration is cancelled, surrendered, expires or is not renewed. The Authority can take provisional custody of the foreign contribution and relevant assets. If the registration is not restored or renewed within the prescribed period, the assets can vest permanently in the Authority and subsequently be transferred or disposed of, with the proceeds going to the Consolidated Fund of India. The proposed framework also covers assets created partly from foreign contributions.

There is an important detail here. The principle of asset vesting is not entirely new. Section 15 of the existing FCRA already provides for the vesting of assets created out of foreign contribution when an organisation’s registration is cancelled or surrendered. What the 2026 Bill does is build a much more elaborate administrative machinery around that principle, giving the state greater control over what happens to those assets. And this matters because the state already has considerable power to make an organisation incapable of functioning long before anyone arrives to take possession of its assets.

The state does not need to seize an organisation’s property to make it unviable. The cancellation or non-renewal of an FCRA registration can cut off its access to foreign funding, while investigations and regulatory proceedings can tie up its finances and consume the administrative capacity needed to keep the organisation running. By the time an organisation is able to challenge such action, it may already have lost the ability to function as it once did.

The 2020 amendments established much of the architecture that made this possible. The ceiling on administrative expenditure was reduced from 50% to 20% of foreign contributions. That expenditure includes basic organisational costs such as personnel, salaries, travel and management. At the same time, organisations receiving foreign contributions were prohibited from transferring those funds to other organisations, effectively ending the sub-granting arrangements through which larger organisations had supported smaller grassroots groups.

The consequences for smaller organisations were particularly severe. A national organisation could previously receive foreign funding and pass some of it to a local group working in a remote district. After 2020, that channel was closed. The local organisation had to deal with the compliance burden itself, even when it lacked the staff, money or administrative capacity to do so.

The 2020 amendments also required foreign contributions to be received through an FCRA account at the State Bank of India’s New Delhi Main Branch on Sansad Marg. For organisations working in rural areas or distant states, even a basic banking requirement became another layer of centralised administrative control.

This has happened alongside a dramatic contraction in the number of active FCRA organisations. Thousands of registrations have been cancelled or allowed to lapse over the years. Organisations including Amnesty International India, Oxfam India and the Centre for Policy Research have faced cancellation, suspension or non-renewal.

The 2026 Rules add further conditions. Organisations must now identify the purposes for which they intend to use foreign contributions and the states or Union Territories in which they intend to operate. Their FCRA registration is consequently tied more closely to specified purposes and geographical areas. The Rules also expand disclosure requirements concerning activities, office bearers, publications and social media accounts. An organisation must demonstrate that it has undertaken “reasonable activity” by utilising at least ₹10 lakh in foreign contribution over the preceding two financial years. For organisations receiving foreign contributions in instalments under prior permission, the release of a subsequent instalment is tied to utilisation of at least 75% of the previous instalment and a field inquiry.

These requirements are presented as compliance measures. But compliance itself is a mechanism of control where the state has the power to decide whether an organisation has done enough, worked in the right place, spent enough money or remained within the purposes recorded on its certificate.

This is particularly useful to a government that wants to curb progressive work while abandoning its own responsibilities to the people. Organisations working with communities affected by caste violence, displacement, environmental destruction, poverty and state violence can be weakened through the same machinery that is supposed to regulate their funding. The state retreats from its public obligations while strengthening its ability to constrain those who organise in the spaces it leaves behind.

The implications extend beyond organisations conventionally described as NGOs into spaces that are squarely political. Civil society includes legal aid groups, research institutions, health organisations, women’s collectives, workers’ organisations, environmental groups, educational initiatives and organisations working with Dalit, Adivasi and minority communities. These institutions document violence, gather evidence, challenge government policy and provide services where state institutions are inaccessible or inadequate.

It also includes religious and minority institutions. The government has stressed that the 2026 Rules continue to permit specified faith-based welfare and religious activities, including religious education, maintenance of places of worship and charitable work. At the same time, the Rules place foreign-funded activities within increasingly specific categories and exclude proselytisation from the permissible religious purposes.

This is pertinent in a country where schools, hospitals, shelters and charitable institutions are run by religious and minority organisations, especially in a context where successive governments have failed to carry out their public and institutional duties effectively. Why should the state possess such extensive administrative power over institutions that provide social services while it fails to do the bare minimum?

As Azam Khan writes in Maktoob, the changes represent “a systematic attempt to control and subjugate the social fabric of India.” An independent organisation that researches caste violence, documents environmental destruction, provides legal aid to workers or challenges a government policy creates an institutional space from which the state can be questioned. That space matters precisely because it exists outside direct government control.

The history of FCRA shows how this power has accumulated. A law introduced to regulate foreign political influence has become a mechanism through which the state can determine which organisations receive foreign resources and how long they are allowed to operate with them. The 2020 amendments tightened the financial conditions. The 2026 Rules add further administrative requirements. The 2026 Bill extends state control to what can happen to an organisation’s assets after its FCRA registration ends.

The result is a civil society whose institutional survival is increasingly placed in the hands of the executive. Independent institutions create space for people to organise outside the state and produce knowledge that can challenge official accounts. When those institutions are weakened, the communities that depend on them lose some of their capacity to make their experiences visible and press their demands. And when that space shrinks, so does the space available for dissent.

DalitThinkerToday is a Chartered Accountant by profession from Sangli, Maharashtra, who occasionally dabbles in thinking about politics when people in her oppressor-caste-majority office frustrate her. This is her pseudonym.

The demolition of homes belonging to Adivasi tea-tribe families in Hatikhuli Bagicha near Kaziranga marks the latest escalation in a dispute over land acquired for a proposed five-star luxury hotel and cultural centre backed by the Assam Tourism Development Corporation (ATDC). According to campaigners, the demolitions have displaced families who had long lived and cultivated the area, bringing into effect a project that has faced sustained opposition from local residents, land rights groups and environmental activists.

The dispute centres on land at Ingleng Pathar and Hatikhuli, adjacent to Kaziranga National Park. Local organisations maintain that many families have cultivated the land for generations, continued to pay land revenue, and possess raiyati records or periodic pattas despite the land being classified as government land. In June, the Assam Chah Mazdoor Sangha (ACMS) demanded a high-level inquiry into the land allotment, alleging that the land had been reclassified and transferred to a hospitality company without consulting the communities dependent on it. The union also stated that residents had resisted eviction attempts for four to five years while seeking formal recognition of their land rights.

The controversy deepened after the Assam government signed an agreement with Hyatt Hotels in September 2023 to develop a luxury hotel near Kaziranga’s central range. The proposed project affects land cultivated by more than forty farming families, most of them from tea-tribe communities. Protesters have argued that the project would displace long-settled residents while further expanding hospitality infrastructure in an ecologically sensitive landscape used by wildlife moving between Kaziranga and the Karbi Anglong hills during the annual floods. Environmental groups have similarly raised concerns about the impact of construction on wildlife corridors and habitat connectivity.

Public opposition intensified through demonstrations in June and July, following which indigenous rights activist Pranab Doley and several others were arrested on charges including unlawful assembly, criminal conspiracy and rioting. Police alleged that protesters entered the project site and obstructed construction, while Doley and land rights organisations described the arrests as an attempt to criminalise resistance to displacement. Soon after, a group of UN experts expressed concern over the detention of five indigenous rights activists, warning that the arrests could have a chilling effect on civic space and undermine the ability of indigenous communities to defend their rights. They called for the release of any activists detained solely for the peaceful exercise of their rights and urged the suspension of land acquisition and development until affected communities had been meaningfully consulted and their free, prior and informed consent obtained.

The developments at Kaziranga reflect a pattern that extends well beyond Assam. From the Ken-Betwa river-linking project in central India to the Great Nicobar Island Development Project, large-scale projects justified in the name of development have repeatedly advanced despite concerns over ecological integrity and the destruction of longstanding human habitats. Whether the issue is the displacement of Adivasi cultivators, the future of the Nicobarese community and the survival of the Shompen, or the loss of forest commons, the concerns of affected communities are repeatedly set aside in favour of state-backed commercial and strategic priorities.

Uttar Pradesh’s minimum wages remained unrevised for more than a decade despite statutory requirements for periodic revision, while consumer prices rose sharply and labour protections were progressively weakened through the implementation of the Labour Codes and state-level deregulation. Against this backdrop of stagnant wages and insecure work, workers across Noida’s industrial belt began peaceful demonstrations in April 2026 after Haryana substantially increased its minimum wages, leaving workers performing similar jobs across the state border with significantly lower pay. The mobilisation spread rapidly across dozens of factories through informal worker networks, reflecting long-standing grievances over wages, excessive working hours, insecure contracts and poor working conditions.

The demonstrations were met with an extensive police crackdown. Hundreds of workers, trade unionists, student activists and labour organisers were detained or arrested, multiple FIRs were filed, and authorities characterised the protests as the work of conspiratorial and “external” elements. Civil liberties organisations, lawyers and fact-finding teams have since documented serious concerns regarding arrests, detention procedures and the criminal investigation, while two activists were booked under the National Security Act.

Following the protests, the Uttar Pradesh government announced an interim increase in minimum wages for Gautam Buddha Nagar and Ghaziabad along with other labour measures. While this addressed part of the workers’ immediate demand, the revised wages remained below those in neighbouring Haryana and did not resolve broader concerns over working conditions and labour rights. At the same time, the government continued to pursue criminal cases arising from the protests. Civil liberties groups, lawyers and the defence campaign have argued that many of the FIRs and charges are fabricated, citing inconsistencies in the investigation, contradictory arrest records and a lack of credible evidence. Multiple accused remain imprisoned, while legal challenges to the investigation continue before the courts.

We stand in solidarity with the workers, trade unionists, students, artists and activists who continue to face criminalisation for asserting labour rights. The demand for fair wages, dignified working conditions and the freedom to organise cannot be treated as a law-and-order problem. We call for the release of those unjustly detained, the withdrawal of fabricated and politically motivated cases, and the protection of workers’ democratic and trade union rights. For readers seeking a fuller account of the events, chronology, legal proceedings and ongoing campaign, please follow:

The Campaign for Release of Workers and Activists of Noida – The Carwan 

When Parliament passed the Scheduled Castes and Scheduled Tribes (Prevention of Atrocities) Act in 1989, it was acknowledging a basic truth: caste violence was not an aberration but a structure. Ordinary criminal law had proven incapable of addressing it. Atrocities were not simply crimes between individuals. They were acts meant to discipline entire communities. The Act attempted to name that reality and create stronger legal consequences for it.

More than three decades later, the Act remains one of the most powerful legal instruments available to Dalit and Adivasi communities. Yet it also exists within a criminal justice system that often neutralizes its force. This contradiction sits at the heart of the law’s history. The Atrocities Act represents both a recognition of caste violence and a reminder of how limited legal remedies are in a society structured by caste power.

Upward mobility and backlash

The sociologist Dag-Erik Berg offers a useful way of thinking about the relationship between law and caste violence. He describes the Dalit experience in modern India through what he calls a recurring pattern: upward mobility coexisting with enduring atrocity.

The expansion of education, reservation policies, and political representation has enabled sections of Dalits to challenge older hierarchies. Yet each moment of assertion has also produced violent backlash. Many of the most brutal caste massacres in recent decades have targeted communities that were asserting new social or economic claims.

The Atrocities Act emerged from this historical dynamic. It was not simply the product of humanitarian concern but the outcome of decades of Dalit struggles that forced the state to acknowledge caste violence as a structural problem. Earlier laws, such as the Protection of Civil Rights Act, focused mainly on untouchability practices like temple entry. They proved inadequate for addressing organized violence, social boycotts, and collective punishment. The 1989 law attempted to respond to these realities. It listed specific offences tied to caste domination and created special procedures for investigation and prosecution. In doing so, it brought caste explicitly into the framework of criminal law.

But as Berg points out, the presence of such legislation does not automatically transform the social structures that generate violence. Reservation policies and atrocity laws operate within domains visible to the state, such as education, employment, and criminal punishment. Meanwhile, the deeper mechanisms through which caste reproduces itself remain largely untouched. Marriage practices continue to enforce caste boundaries. Land ownership remains deeply unequal across caste lines. Economic and social power in villages often rests with dominant castes. The law intervenes after violence occurs, but the social order that makes violence possible remains largely intact. This gap between structural inequality and legal remedy explains why atrocities continue even in a constitutional democracy that formally prohibits caste discrimination.

The everyday obstacles of justice

If the Atrocities Act represents recognition of caste violence, its implementation reveals how difficult it is to translate that recognition into justice.

Anthropologist Sandhya Fuchs traces what happens when survivors attempt to turn experiences of violence into legal cases. The journey often begins at the police station, where complaints must first be recorded. Here, the law encounters the everyday functioning of the criminal justice system. Police officers frequently reinterpret or dilute complaints in ways that prevent cases from being registered under the Atrocities Act. Acts of violence are reframed as ordinary disputes, family conflicts, or property disagreements.

Survivors may be pressured to withdraw complaints or accept informal settlements. Sometimes cases never reach the courts at all. Even when cases move forward, survivors must navigate a complex field of actors, including activists, lawyers, community leaders, and political intermediaries. Each of these groups may have their own ideas about how the law should be used.

Fuchs describes how atrocity cases can become sites where competing visions of justice, honor, and political strategy collide. Survivors themselves often make pragmatic choices within these constraints. Out-of-court compromises are common, even though many activists criticize them as evidence that the law is failing. For some families, however, the threat of legal action becomes a bargaining tool to secure compensation, protection, or a measure of dignity in hostile local environments.

These negotiations reveal something important about the law’s role. The Atrocities Act is not simply a legal instrument. It has also become a symbolic resource that marginalized communities use to challenge dominant caste power, sometimes inside the courts and sometimes outside them. Yet the same processes that allow creative uses of the law also expose its fragility. Police bias, political pressure, and social intimidation constantly shape whether justice is possible.

Evidence, memory, and the burden of proof

The obstacles do not end once a case reaches court. Legal procedures themselves often place heavy burdens on survivors and witnesses.

Legal scholar Achintya Anita Gurumurthy highlights this problem through the example of the Test Identification Parade, a procedure used during criminal investigations to confirm whether eyewitnesses can identify suspects. In theory, such procedures are meant to strengthen evidence. In practice, they can be used to discredit survivors. Courts sometimes treat minor inconsistencies in witness testimonies as signs of unreliability. Delays in conducting identification parades or technical procedural lapses can weaken cases, even when violence is widely known within local communities. In atrocity cases, this evidentiary framework often collides with the realities of trauma, fear, and social pressure faced by Dalit witnesses.

The result is a pattern in which the credibility of survivors becomes the central issue, while the broader structure of caste power fades into the background. The legal process turns into an examination of memory rather than an inquiry into domination.

Gurumurthy argues that evidentiary practices cannot be treated as neutral procedures. They operate within social hierarchies. When police investigators, prosecutors, and judges overwhelmingly come from dominant castes, the interpretation of testimony inevitably reflects those power relations.

Recent legal developments have not fundamentally altered this situation. The Bharatiya Sakshya Adhiniyam, which replaced the colonial-era Evidence Act in 2023, largely preserves existing evidentiary structures, including the continued reliance on identification procedures like the Test Identification Parade. The formal language of reform has therefore not significantly transformed how evidence functions in caste atrocity cases.

For survivors, this means that the struggle for justice often involves confronting not only perpetrators but also the institutional habits of the legal system itself. Despite sensitive case law in Malkhansingh v. State of Madhya Pradesh and Kulwinder v. State (NCT of Delhi), which reflects the work of Dalit lawyers and activists, much more remains to be done in criminal justice reform if the goals of the Atrocities Act are to be realised.

Contestation and dilution

The Atrocities Act has always faced political resistance. Dominant caste organizations have repeatedly portrayed it as a law prone to misuse. This narrative gained national attention during the controversy surrounding a 2018 Supreme Court judgment that introduced restrictions on arrests under the Act, citing concerns about false cases. Dalit protests across the country forced Parliament to amend the law and restore its original provisions. The episode demonstrated both the vulnerability of the Act and the power of collective mobilization.

Such struggles are part of a longer history. Every expansion of legal protection for marginalized groups has been accompanied by attempts to dilute it, delay its implementation, or reinterpret it in ways that limit its reach. The Atrocities Act is no exception.

At the same time, the law has become deeply embedded in Dalit political consciousness. Activists, lawyers, and community organizations have built networks around it. Survivors often approach the law not simply as a technical instrument but as a statement that their suffering deserves recognition.

This dual character explains why the Act can appear both powerful and fragile, as Fuchs calls it. It is powerful because it names caste violence explicitly and creates legal consequences for it. It is fragile because its effectiveness depends on institutions that remain shaped by the very hierarchies the law seeks to challenge.

What law can and cannot do

For many survivors of atrocities, the most painful realization is that law alone cannot transform social relations. As Fuchs shows through her study of rural Rajasthan, families affected by brutal violence sometimes describe the law as something distant. Court cases may continue for years without changing the attitudes of the communities that carried out the violence.

This does not mean the law is meaningless. Legal action can bring compensation, punishment, or recognition. It can create new networks of solidarity and political mobilization. It can signal that caste violence is not a private dispute but a public injustice.

But law operates within a society where caste remains deeply embedded in everyday life. Courts cannot reorganize land ownership. Police procedures cannot dismantle endogamous marriage networks. Criminal punishment cannot eliminate the desire for status that sustains graded inequality. The persistence of caste therefore exposes the limits of relying solely on legal remedies.

Beyond the courtroom

The history of the Atrocities Act makes one thing clear: law has rarely initiated social transformation in India. More often, it has been forced into existence by movements that refused silence. Dalit and Adivasi struggles compelled the state to recognise caste violence, defend the law when it was diluted, and push institutions to act when they would rather look away.

Seen from this perspective, the Atrocities Act is not the source of justice but one instrument within a larger political struggle. It provides language, leverage, and sometimes protection. But the force that gives the law meaning lies outside the courtroom.

The fight against caste atrocities has never been won by statutes alone. It has been sustained by collective resistance, by communities that organise, testify, and refuse the social order that treats their lives as expendable. When the law works at all, it is usually because those struggles have compelled it to.

Paresh Hate is a Bahujan queer researcher, writer, documenter and activist and the Lead Content Manager at Dalit Voice.