Secularism in the Indian Constitution is usually taught as a single principle with a single adjective attached: the Indian variant, warmer and more accommodating than the Western wall of separation, yet still one idea. The Constituent Assembly Debates of December 1948 tell a different story. Across the sittings of the first week of that month, the members who debated Draft Articles 19 to 22B built three distinct settlements between state and religion and placed them side by side in the constitutional text: a state that establishes no religion and funds none from taxation, a state that respects every religion as a bearer of rights, and a state armed with authority to enter religion and reform it. The three settlements were never harmonised. This article defends a single claim about that record: Indian secularism is not one principle; it is three layers that the founders did not fully harmonise, and the contests that followed have been arguments about which layer governs a dispute, conducted in the vocabulary of secularism as a whole.

The claim matters because the standard treatments flatten the record in opposite directions. Rajeev Bhargava, in The Promise of India’s Secular Democracy in 2010, reads the founding as a coherent and distinctive achievement: a principled distance between state and religion that permits engagement or disengagement as principle requires. The opposing treatment, common in polemic after 2014, reads the founding as a failed copy of Western separation, incoherent because the Indian state refused to leave religion alone. Both treatments share one assumption: that the founders enacted a principle, which later history either honoured or betrayed. The debate record refuses that assumption. When the Assembly rejected K. T. Shah’s attempt to write the word secular into Article 1 on 15 November 1948, and when the same Assembly, three weeks later, guaranteed a right to propagate religion while empowering legislatures to throw open Hindu temples to excluded castes, the members were not applying a principle. They were stacking settlements, each answering a different fear, and leaving the ranking among them undecided. Recovering that stack is the work of the sections that follow, and it begins from the widest frame the founding offers: Tension One, the religious-secular settlement, that this article works out was the first of the three permanent tensions around which the whole founding turned.

A layered settlement explains evidence that a single-principle account must explain away. It explains why Article 27 forbids a tax for any particular religion while Article 25(2)(b) singles out Hindu institutions for reform by name. It explains why the Supreme Court could hold in 1954 that a religious denomination manages its own affairs, and hold in 1958 that a denomination’s temple could be opened to worshippers the denomination excluded. It explains why the Citizenship (Amendment) Act of 2019 and the Sabarimala judgment of 2018 could each claim constitutional warrant while pointing at different clauses. None of these pairs is a contradiction to be deplored or a paradox to be dissolved. Each pair is two layers operating at once, and the legal and political fight, in every generation, has been over priority. The founders built the arena for that fight in December 1948. The fight itself they left to their successors, and the successors have obliged them, case by case and statute by statute, for seventy years.

The Word the Assembly Refused to Write

The most quoted absence in the Indian Constitution is a word. The Preamble adopted on 26 November 1949 described India as a sovereign democratic republic and said nothing about secularism; the word secular entered the Preamble only through the Constitution (Forty-second Amendment) Act of 1976. An absence that large invites two lazy inferences: that the founders rejected secularism, or that they forgot it. The debate record disposes of both. The Assembly considered the word three times in the winter of 1948 and refused it three times, on stated grounds, while enacting the substance that later generations would assemble under the word. The refusal was about the label, and about what a label would freeze. Understanding that distinction is the first step into the layered reading, because a principle can be labelled; a stack of settlements resists a single name.

What did K. T. Shah ask the Assembly to insert into Article 1 on 15 November 1948?

On 15 November 1948, K. T. Shah moved that Article 1 describe India as a secular, federal, socialist union of states. The Assembly negatived the amendment after debate. B. R. Ambedkar, replying, argued that a constitution is a mechanism for regulating state organs, not a place to prescribe the social and economic order future voters must inhabit.

Shah’s amendment deserves a slower reading than it usually receives, because Shah was no marginal figure and his proposal was no slogan. Khushal Talaksi Shah was an economist, a member of the Advisory Committee, and one of the most persistent amendors in the Assembly, a man who returned to the charge twice more that winter: an attempt touching the word secular on 25 November 1948 was negatived after a vote, and a third attempt, aimed at Draft Article 18 on 3 December 1948, was set aside. The pattern across the three attempts is the evidence. The Assembly did not brush past the word in a hurry. The Assembly heard the proposal in three forms, from a respected member, and declined it in three forms, while the clauses on religion were still being drafted in committee and debated on the floor. A body that intended a secular state under a single principle, and merely neglected the vocabulary, does not vote the vocabulary down three times in nineteen days.

Ambedkar’s stated reason repays attention, because it is a reason about constitutional method rather than about religion. Answering Shah on 15 November 1948, the Law Minister argued that the Constitution should not tie the hands of future electorates on the organisation of society: what the social and economic order should be was a decision for the people of each generation, taken through their legislatures, and embedding one generation’s answer in the constitutional text would be undemocratic in the precise sense that it would let the dead govern the living. The argument was aimed mainly at the word socialist, yet the word secular fell with the same stroke, and the stroke is instructive. For the Drafting Committee, the religion clauses were machinery, guarantees and powers whose operation later politics would direct. A label in Article 1 would have pretended that the machinery had a fixed direction. The committee preferred the machinery without the pretence, and the Assembly, voting, agreed.

Two further considerations pressed in the same direction, and both appear in the record around the Shah debates. The first was Partition. The Assembly that sat in the winter of 1948 had watched a subcontinent divided on religious lines fifteen months earlier, and members who spoke on the religion clauses carried that memory openly. A state proclaimed secular in the vocabulary of European church and state history risked being heard, by the minorities who had stayed and by the majority that had demanded Pakistan, as a state estranged from religion itself, in a society where religion structured birth, marriage, death and property. The second consideration was drafting economy. Nehru’s Objectives Resolution, moved on 13 December 1946 and adopted on 22 January 1947, had already guaranteed freedom of thought, expression, belief, faith and worship as justice to be secured for all citizens. The founders expected the enforceable clauses to do the work that a proclaimed adjective could not. Events proved the expectation half right: the clauses did the work, and the adjective arrived in 1976 anyway, carried by an amendment whose own legitimacy the Emergency had compromised, naming a settlement that had been operating, unnamed, for twenty-six years.

Volume VII and the Making of the Religion Clauses

The religion clauses reached the floor of the Assembly with two years of committee work behind them, and the committee record explains the clauses’ layered shape better than any speech. The Constituent Assembly Debates, Volume VII, covering the sittings from November 1948 into January 1949, record the Assembly taking up Draft Article 19, the future Article 25, in the first week of December 1948, and then the companion clauses on denominational rights, taxation and religious instruction. What the volume preserves is not a debate between secularists and believers. Almost nobody in the chamber defended a state religion, and almost nobody proposed a state indifferent to religion. The argument ran along a different axis entirely: how far the state’s respect for religion should extend, and where respect should yield to reform. Every layer of the finished settlement can be traced to one side of that argument prevailing on one clause while losing on the next.

Which committee drafted the religion clauses before the Assembly debated them?

The Advisory Committee on Fundamental Rights, chaired by Sardar Vallabhbhai Patel and appointed in January 1947, drafted the religion clauses through its Sub-Committee on Fundamental Rights and the Minorities Sub-Committee chaired by H. C. Mookerjee. The committees reported in the spring of 1947, and the Drafting Committee carried their clauses into the Draft Constitution of February 1948.

The committee route matters for the layered reading because the clauses were assembled by bodies with different mandates, and the mandates show in the text. The Sub-Committee on Fundamental Rights, where J. B. Kripalani and others weighed the claims of conscience, approached religion as an individual entitlement to be shielded from the state. The Minorities Sub-Committee approached religion as a collective possession: the denominations, schools and personal laws through which communities organised their lives, to be shielded from majoritarian legislation. The Advisory Committee, under Patel’s chairmanship, held both approaches together and added a third concern that neither sub-committee owned: the reform of religious practices that the national movement had already condemned, above all untouchability and the exclusion of castes from temples. When the three concerns reached the Drafting Committee, chaired by Ambedkar and served by members including K. M. Munshi and Alladi Krishnaswami Ayyar, the committee did not reconcile them into a doctrine. The committee translated the individual concern into Article 25(1), the collective concern into Article 26, and the reform concern into Article 25(2), and let the three stand. The layers were born in committee as a division of labour before they were debated on the floor as a philosophy.

When did the Assembly debate Draft Article 19, the future Article 25?

The Assembly debated Draft Article 19 across the sittings of the first week of December 1948, in the session recorded in Volume VII of the Constituent Assembly Debates, with the companion clauses on denominational autonomy, religious taxation and religious instruction following in the same December sittings before the winter session moved to other parts of the draft.

The floor debate of December 1948 is the primary evidence for everything argued in the sections ahead, and its texture deserves emphasis because later commentary often treats the clauses as the Drafting Committee’s solitary handiwork. They were not. The record shows amendments moved, pressed and defeated from every direction: Tajamul Husain of Bihar seeking to strip religion of its public marks, Lokanath Misra of Orissa seeking to confine the guaranteed freedom within narrower bounds, K. M. Munshi defending the widest reading of the rights from inside the Drafting Committee itself, and Ambedkar closing debate after debate with the committee’s reasons. The record also shows the smaller machinery of settlement at work: the kirpan Explanation that protected a Sikh practice by name, the drafting choice that extended the reform clause’s definition of Hindu to cover Sikh, Jain and Buddhist institutions, the careful subordination of every religious right to public order, morality and health. A settlement assembled this way, clause by clause and amendment by amendment, could not have emerged as one principle even if the Assembly had wanted one. The Assembly’s own procedure, majoritarian and textual, produced layers, because each layer was the residue of a different vote.

One further feature of the December record must be fixed before the layers are examined, because the artifact table later in this article depends on it. The members knew they were writing for a society in which religion was not a weekend affiliation but a legal system. Personal law governed marriage, succession, adoption and guardianship for Hindus and Muslims alike, administered by the colonial courts as religious law. Temples held land and commanded revenues. Denominations ran schools that the state aided. Any clause that touched religion touched property, family and education at once, and the members said so, repeatedly, when resisting amendments that treated faith as a private hobby to be protected by a sentence. That is why the settlement has the density it has: the founders were not decorating a secular state with religious exceptions. They were distributing authority over the most heavily governed sphere of Indian life among the believer, the denomination, the legislature and the courts, and the distribution, not any slogan, is the secularism they made.

Layer One: A State With No Church, No Religious Tax and No Catechism

The first layer is the one that most resembles the Western model, and the Assembly built it from three clauses that limit the state rather than the believer. Article 27 provides that no person shall be compelled to pay any taxes whose proceeds are specifically appropriated in payment of expenses for the promotion or maintenance of any particular religion or religious denomination. Article 28 provides that no religious instruction shall be provided in any educational institution wholly maintained out of state funds, that institutions administered by the state under a trust requiring religious instruction are excepted, and that in state-recognised or state-aided institutions no student shall be required to take part in religious instruction or worship without consent. Article 25 itself opens the sequence with a freedom held against the state, and the Preamble’s eventual silence on establishment completes the picture. Read together, the clauses describe a republic with no official church, no tithe, and no state classroom catechism. This is non-establishment, Layer One, and nobody in the December debates contested its core.

The precision of Article 27 repays a close reading, because the clause forbids less than a wall of separation would forbid, and the difference was deliberate. The clause does not bar the state from spending general revenue in ways that benefit religions: a police force guarding a procession, a grant to a denominational college under Article 28’s consent regime, a tax exemption for charitable endowments. What the clause bars is the fiscal signature of establishment, a levy earmarked for a named religion, the arrangement by which a state announces that one faith is its own and bills the taxpayer for the announcement. The members who shaped Article 27 in December 1948 had the colonial precedent in front of them: the East India Company had administered Hindu and Muslim endowments, paid temple servants, and managed religious establishments as a department of government well into the nineteenth century. The clause retires that precedent. The Indian state would regulate religious institutions as it regulates other public trusts; it would not keep a religion as a householder keeps a family priest, on the public payroll, by a dedicated tax.

Article 28 shows the same precision under greater pressure, because education was the sphere where the state’s money and the denomination’s mission most thoroughly mixed. The Assembly in December 1948 faced a school system in which a large share of recognised institutions were run by religious bodies, Christian missions, Arya Samaj trusts, Muslim anjumans and Hindu endowments among them, many receiving state aid. A clause barring religious instruction from every aided school would have detonated that system; a clause permitting instruction everywhere would have made the state a paymaster for catechisms. The settlement split the difference by funding source. Wholly state-funded institutions: no religious instruction at all. State-administered institutions held under a trust or endowment requiring instruction: the trust governs, and instruction continues. Recognised and aided institutions: instruction permitted, attendance voluntary, consent required. Critics then and since have called the threefold scheme untidy. The scheme is better understood as Layer One drawn with a surveyor’s care: the state may not teach religion where the state is the school, and may not compel a child where the state merely helps pay. Establishment was the target. Education was not.

The layer’s afterlife in the courts shows its strength and its limits with equal clarity. In Bijoe Emmanuel v State of Kerala, decided on 11 August 1986, the Supreme Court protected three Jehovah’s Witness children expelled for refusing to sing the national anthem while standing respectfully during it, holding that their conscientious objection fell within Article 25 and that no law required the singing. The judgment is Layer One and Layer Two working in harness: the state school could no more compel a patriotic liturgy against conscience than a religious one. Yet the same decades show the layer’s boundary. The state continued to administer temples through Hindu religious and charitable endowment departments in the southern states, a practice the Supreme Court sustained as regulation of secular administration rather than establishment, and Article 290A of the Constitution itself charges annual payments from the Consolidated Funds of Kerala and Tamil Nadu to the Devaswom Funds of Travancore and Cochin. A strict separationist reads those survivals as hypocrisy. The layered reading absorbs them without strain: non-establishment in India was always a bar on a state church and a religious tax, never a promise that the state would keep its hands off religious institutions. The hands appear in Layer Three, and the founders put them there in the same week.

Layer Two: Equal Respect Written as Enforceable Rights

The second layer reverses the direction of the first. Where Layer One limits the state, Layer Two obliges it: the state must hold itself in a posture of respect toward every religion, and the respect is not a sentiment but a set of enforceable rights held by persons and denominations against the state itself. Article 25(1) guarantees to all persons, not citizens alone, freedom of conscience and the right freely to profess, practise and propagate religion, subject to public order, morality and health and to the other provisions of Part Three. Article 26 guarantees to every religious denomination the right to establish and maintain institutions for religious and charitable purposes, to manage its own affairs in matters of religion, and to own and administer property. Article 29 protects the language, script and culture of any section of citizens, and Article 30 guarantees minorities the right to establish and administer educational institutions of their choice. The posture these clauses compose has no exact Western counterpart. The American First Amendment disables Congress from respecting an establishment of religion and protects free exercise, a grammar of distance. The Indian clauses speak a grammar of regard: religions appear in the text as rights-bearers whose practices the state must accommodate, denominations as lawful polities whose internal affairs the state must respect. Equal respect, Layer Two, is the layer Bhargava’s reading describes best, and the layer that the founders’ critics most often mistake for the whole.

The drafting choices inside Article 25(1) mark the layer’s character at the level of the sentence. The Assembly guaranteed not only practice but profession, the public declaration of faith, and not only profession but propagation, the active commendation of faith to others. Each verb widens the right beyond the interior forum of conscience, and each widening was debated. A constitution of private belief would have stopped at conscience. The Assembly’s text protects a religion that speaks, dresses, assembles, endows and persuades. Explanation I to Article 25 writes the point into the margin: the wearing and carrying of kirpans shall be deemed to be included in the profession of the Sikh religion. A five-inch blade, carried by a baptised Sikh, sits inside the fundamental rights chapter by name. No wall-of-separation constitution carries such a sentence, because no such constitution imagines the state’s task as the Assembly imagined it: not to clear a religion-free public square, but to guarantee each faith room to appear in the square in its own dress.

The courts built the layer’s doctrine early, and the two founding judgments of 1954 still organise the field. In Commissioner, Hindu Religious Endowments, Madras v Sri Lakshmindra Thirtha Swamiar of Sri Shirur Mutt, decided on 16 April 1954, Justice B. K. Mukherjea held for the Supreme Court that Article 26 secures to a denomination the right to manage its own affairs in matters of religion, and that what constitutes an essential part of a religion is to be determined with reference to the doctrines of that religion itself. The Shirur Mutt judgment thus gave denominations a protected interior, a sphere of practice into which the state’s regulatory power could not follow merely because the practice struck legislators as superstitious or wasteful. Six weeks of the same season produced Ratilal Panachand Gandhi v State of Bombay, where the Court sustained the denominational rights of the Jain community in the administration of its temples against a Bombay public trusts statute, while accepting state supervision of the secular administration of trust property. Read together, the 1954 judgments draw Layer Two’s map: belief and ritual at the centre, protected; property and finance at the edge, regulable; and the boundary between centre and edge to be patrolled by courts asking what the religion itself holds essential. The essential-practices doctrine has been criticised for seventy years for making judges into theologians. Its origin is nonetheless a Layer Two decision, a choice to let the denomination’s own self-understanding, rather than the state’s reforming agenda, define the protected core. Layer Three would contest exactly that choice, which is why the argument between the layers became the argument of Indian secularism itself.

Why did the Supreme Court hold in 1977 that propagation stops short of converting another person?

In Rev. Stanislaus v State of Madhya Pradesh (1977), the Court upheld the Madhya Pradesh and Orissa anti-conversion statutes, holding that the right to propagate religion under Article 25 means transmitting or spreading one’s religion by exposition of its tenets, not a right to convert a person, since a converted conscience would violate the convert’s freedom of conscience.

The Stanislaus holding is usually taught as a restriction, and so it is; yet its reasoning reveals Layer Two’s inner logic better than any commentary. Chief Justice A. N. Ray’s judgment treated freedom of conscience as a symmetrical possession: the propagator holds it, and so does the person addressed, and the second holding limits the first. Propagation is protected because conscience is respected; conversion by force, fraud or allurement loses protection because the conscience at the receiving end is disrespected. Honour every faith, and the honouring itself draws the line. The later career of the anti-conversion statutes, enacted in state after state and litigated into the present century on allegations that they license harassment of minorities, shows the cost of letting a Layer Two symmetry be administered by majoritarian police forces, a cost the founders did not price. The settlement they wrote nonetheless contains the resources for its own correction, in the same Article 25 that Stanislaus read narrowly: the freedom belongs to all persons equally, and a statute administered to burden one faith’s propagation while indulging another’s would fail the equality that the layer exists to secure.

The Propagation Fight: Tajamul Husain’s Amendment and What Its Defeat Decided

Every layered settlement has one debate that exposes its joints, and for the religion clauses that debate is the fate of the amendments moved by Tajamul Husain of Bihar in the December 1948 sittings. Husain was a barrister, a Muslim member elected from Bihar, and a secularist of the assimilationist school: he held that religion in India had to become a private matter between the individual and God if the new nation was to cohere, and he moved amendments to Draft Article 19 to make the text say so. His best-known proposal would have replaced the kirpan Explanation with a provision that no person should bear any visible sign or mark or name, and no person should wear any dress, whereby his religion may be recognised. A companion line of amendment sought to cut back the propagation limb of the right, the verb that turned private conviction into public recruitment. Taken together, the Husain amendments proposed a republic of unmarked citizens: faith protected precisely to the extent that faith stayed invisible, and a state whose secularism consisted in the public square’s neutrality of appearance.

The Assembly refused, and the manner of the refusal is the evidence that matters. Maulana Hasrat Mohani answered Husain’s proposal on the floor with a joke that carried a doctrine inside it, suggesting that the mover might begin by changing his own name, since Tajamul Husain announced a religion as plainly as any dress. Beneath the laughter sat the chamber’s settled instinct: in a society where names, dress, diet, calendar and law were saturated with religion, an unmarked public square was not neutrality but a demand that citizens amputate their social selves at the door of citizenship. K. M. Munshi, replying from the Drafting Committee’s side, put the doctrinal answer on the record: the question of professing, practising and propagating one’s faith, he told the House, is a right which the human being had from the very beginning of time and has been recognised as an inalienable right of every human being, not only in this land but the whole world over. The amendments were not accepted. Article 25 went forward with propagation inside it and the kirpan Explanation intact, and the layered settlement acquired, by that defeat, one of its defining features: Indian secularism would protect the public presence of religion, not purchase civic peace with religion’s disappearance.

The Husain debate also disciplines a tempting misreading of the founding, the misreading that treats the Assembly’s secularism as the ideology of a Westernised elite imposed on a reluctant society. The chamber that defeated Husain was the same chamber that would, six months later, abolish separate electorates and reserve the state’s reforming hand over Hindu institutions. Its members were not choosing between secularism and religion. They were choosing among secularisms, and the one they rejected in December 1948 was the secularism of erasure: the French-inflected model in which the republic recognises no faith because the citizen displays none. The one they built instead was a secularism of recognition, in which the republic sees every faith, guarantees each its rites and its voice, and reserves, in Layer Three, the right to discipline what any faith does to its own adherents. Husain’s amendments survive in the record as the road not taken, and the road’s clear signposting is a gift to the historian: because the alternative was moved, debated and voted down, the Assembly’s choice of recognition over erasure cannot be mistaken for drift. The founders knew the other model existed. They had it in amendment form, with a mover to defend it. They buried it, and built the layers.

Layer Three: The State as Reformer of Religion

The third layer is the one that startles readers formed on Western models, because it runs the state’s power into the interior of religion itself. Article 25(2) provides that the guarantee of religious freedom shall not prevent the state from making any law regulating or restricting any economic, financial, political or other secular activity which may be associated with religious practice, and, in clause (b), any law providing for social welfare and reform or the throwing open of Hindu religious institutions of a public character to all classes and sections of Hindus. Explanation II adds that the reference to Hindus shall be construed as including persons professing the Sikh, Jain or Buddhist religion. Pause over the architecture of that clause. A constitution that had proclaimed a wall between state and religion could not contain it. A constitution of principled distance might contain its first half, the regulation of secular activity associated with practice. Clause (b) goes further: it names a religion, classifies its institutions as reformable in the interest of social welfare, and constitutionalises the throwing open of its temples as a legitimate legislative object. Layer Three is not an exception smuggled into the settlement. It sits in the same article as the freedom, in the same sentence structure, debated in the same December week. The founders did not hide the reforming state in a schedule. They enthroned it beside the right.

Ambedkar had announced the layer’s philosophy a month before the December sittings, in the speech with which he presented the Draft Constitution to the Assembly on 4 November 1948, a speech whose treatment by Ambedkar and the Drafting Committee that carried the religion clauses only partly captures its reach. Defending the draft’s religion clauses against the charge of interference, Ambedkar observed that the religious conception in India covers every aspect of life, from birth to death, and that if the state were barred from legislating wherever religion spoke, Parliament would be left with almost no field of legislation at all. The argument inverted the liberal presumption. In a society where religion was a thin private creed, a state barred from religion loses little governing range. In India, where religious law and custom governed marriage, inheritance, caste status, temple access and charitable wealth, a state barred from religion would be barred from social legislation itself. Ambedkar, who had burned the Manusmriti at Mahad on 25 December 1927 and who would resign from Nehru’s cabinet on 27 September 1951 over the stalling of the Hindu Code Bill, understood the reform clause as the constitutional warrant for the social revolution the national movement had promised and the priesthoods had resisted. Layer Three is Ambedkar’s layer, and its presence in the text is the strongest single proof that the settlement was layered by design rather than by accident.

The clause’s first great judicial test showed the layers in open combat. The Madras Temple Entry Authorisation Act of 1947 had thrown open Hindu temples in the Madras Presidency to castes excluded by custom. When the hereditary trustees of the Gowda Saraswath Brahmin temple at Coondapur resisted a state notification admitting excluded worshippers, the resulting litigation reached the Supreme Court as Sri Venkataramana Devaru v State of Mysore, decided in 1958. The denomination argued Layer Two: Article 26(b) protected its right to manage its own affairs in matters of religion, and exclusion was its affair. The state argued Layer Three: Article 25(2)(b) authorised exactly this throwing open. The Court held for the state on the entry question while preserving the denomination’s right to exclude non-Hindus and to conduct its rituals undisturbed, splitting the case along the seam between the layers. Temple doors: reform governs. The denominational core of worship: respect governs. The Devaru judgment is the layered settlement’s classic exposition, and it arrived within a decade of the founding, from a bench reading the same December text. Later adjudication repeated the pattern. In Sardar Syedna Taher Saifuddin Saheb v State of Bombay, decided in 1962, the Court struck down the Bombay Prevention of Excommunication Act of 1949 as a violation of the Dai-ul-Mutlaq’s denominational rights under Article 26(b), refusing to let a reform statute reach the power of excommunication that the Dawoodi Bohra faith treated as its head’s religious prerogative. Reform lost there, because the practice in question was classified as religion’s interior rather than society’s welfare. Classify the practice, and the layer follows: that rule, unstated in the text and unavoidable in the cases, is the operating system the founders installed.

The reform layer’s legislative career runs through the statute book of the 1950s and straight into the founding’s largest unfinished project. The Hindu Marriage Act of 1955, the Hindu Succession Act of 1956, the Hindu Minority and Guardianship Act of 1956 and the Hindu Adoptions and Maintenance Act of 1956, the four statutes into which the Hindu Code Bill was finally divided, are Article 25(2)(b) translated into a code: monogamy imposed on a religion that had permitted polygamy, divorce introduced into a sacrament that had forbidden it, daughters written into inheritance lines from which custom had excluded them. The politics of that translation, the opposition of President Rajendra Prasad, the campaign of Swami Karpatri, Ambedkar’s resignation when the code stalled, are traced in the Hindu Code Bill that operationalised Article 25(2)(b) in the post-founding period. For the argument here, the statutes matter as proof of concept. Within six years of the Constitution’s commencement, the reform layer had rewritten the family law of the majority community in defiance of orthodox religious authority, and the courts, in State of Bombay v Narasu Appa Mali in 1952, had already held that personal laws were not laws in force for the purposes of Article 13 and so stood largely beyond fundamental-rights challenge. The settlement’s asymmetry was thereby doubled: the state could reform Hindu personal law by statute, yet the reformed law itself was shielded from rights review. Muslim personal law, meanwhile, stood outside the reform project altogether, an abstention whose politics would detonate in the Shah Bano litigation a generation later. The layers were never harmonised, and neither were the communities’ exposures to them.

The reform layer’s reach has a judicially patrolled boundary, and its clearest modern marker is a temple the state tried to take over and had to hand back. The Nataraja temple at Chidambaram in Tamil Nadu had been administered for centuries by the Podu Dikshitars, a denominational community of priest-trustees, until the Tamil Nadu Hindu Religious and Charitable Endowments department appointed an executive officer to the temple in 2009 under the state’s endowments statute. In Subramanian Swamy v State of Tamil Nadu, decided on 6 January 2014, the Supreme Court ordered the takeover undone, holding that a denomination’s right to administer its institutions under Article 26 is not a concession the state may withdraw on a general plea of mismanagement, and that state intervention in denominational administration must answer a proven breakdown and remain proportionate and temporary. The judgment did not weaken Layer Three; the endowments framework itself, with its audits, budgets and supervision of temple property across the state, stood untouched, as it has stood since the Madras legislation of the 1950s. What the judgment policed was the seam. Administration of property is regulable without limit of time; administration of the denomination’s religious life is not annexable at all. A state that audits a temple’s accounts acts within the settlement. A state that appoints itself the temple’s trustee in permanence has crossed from reform into establishment in reverse, a state religion of administration, and the Court in 2014 said so in the founders’ own vocabulary of clause and denomination. The layers held, because a bench was willing to read them as layers and to assign the disputed power to the right one.

The Three-Layer Matrix: The Findable Artifact

The argument of this article can be carried away in a single reference object, the Three-Layer Matrix of the Indian Secular Settlement, set out in the table below. Each row names one layer, the constitutional text that carries it, the founding-era debate that produced it, and the post-Independence cases in which the layer was invoked or contested. A reader who wants to test any later controversy against the founding settlement can begin by asking which row the controversy belongs to, and which row its opponents answer from.

Layer	Constitutional text	Founding-era debate reference	Post-Independence invocation or contest
Layer One: non-establishment	Article 27 (no tax for any particular religion); Article 28 (no religious instruction in wholly state-funded institutions)	Volume VII debates on the taxation and instruction clauses, December 1948; rejection of a state religion across the Advisory Committee reports of 1947	Bijoe Emmanuel v State of Kerala (1986), conscience against a compelled anthem; Article 290A Devaswom payments sustained as an exception by design
Layer Two: equal respect	Article 25(1) (profess, practise, propagate); Article 26 (denominational autonomy); Articles 29 and 30 (culture and minority institutions)	Draft Article 19 debates, first week of December 1948; defeat of the Tajamul Husain amendments; K. M. Munshi on propagation as an inalienable right	Shirur Mutt (1954) and Ratilal Panachand Gandhi (1954), the protected denominational interior; Sardar Syedna Taher Saifuddin (1962), excommunication protected; Rev. Stanislaus (1977), propagation bounded by the hearer’s conscience
Layer Three: reform authority	Article 25(2)(a) (regulation of secular activity associated with religion); Article 25(2)(b) (social welfare, reform, throwing open Hindu public institutions), with Explanation II	Ambedkar’s speech of 4 November 1948 on religion covering every aspect of life; the December 1948 adoption of clause (2) beside the freedom in clause (1)	Sri Venkataramana Devaru (1958), temple entry sustained; the four Hindu Code Acts of 1955 and 1956; Sabarimala judgment (2018), reform read through equality; Shayara Bano (2017), instant triple talaq set aside
The unresolved seam	Article 25 read as a whole: clause (1) rights limited by clause (2) powers, with no ranking stated	K. T. Shah’s three rejected attempts to name the settlement, November and December 1948; no harmonising clause moved or adopted	Kesavananda Bharati (1973), secular character listed among basic features; S. R. Bommai (1994), secularism held a basic feature and a ground for reviewing state action

The matrix also records what the settlement is not. No row corresponds to a separation wall, because no clause builds one. No row corresponds to a state religion, because every attempt to move the text in that direction failed in committee or on the floor. The fourth row is the honest one: the seam, the place where clause (1) and clause (2) meet without a stated ranking, is itself part of the artifact, the founders’ most consequential bequest precisely because it was left open. Commentators who treat the seam as a drafting failure must explain why the Assembly, which voted on hundreds of amendments that winter, never adopted a harmonising clause though the tension was visible in its own chamber. The layered reading has the simpler explanation. The seam was the settlement. Ranking the layers in 1948 would have required the founders to decide, in advance and forever, whether a denomination’s temple door matters more than the excluded worshipper’s equality, whether propagation matters more than the hearer’s peace, whether the state’s reforming mandate outruns the believer’s conscience. The Assembly decided those questions case by case in debate and declined to decide them wholesale in text. Posterity inherited the questions in the only form in which a plural society can honestly hold them: open, structured, and assigned to courts and legislatures to answer one dispute at a time.

Where the Layers Collide: The Internal Tension, Faced Directly

A layered reading earns its keep only if it faces the settlement’s hardest feature without blinking: the layers contradict each other, in text, on questions that matter. Place Article 25(1) beside Article 25(2)(b) and the contradiction is plain. The first clause guarantees every person the right to practise religion, and Article 26 guarantees every denomination the management of its religious affairs. The second clause authorises the state to legislate for social welfare and reform and to throw open Hindu religious institutions to all classes and sections of Hindus. If equal respect were the settlement’s whole content, clause (2)(b) would be indefensible: a state respecting all religions symmetrically does not single out one religion’s institutions for statutory opening, and does not define that religion, through Explanation II, to sweep in three neighbouring faiths for reform purposes while leaving Islam and Christianity outside the clause’s reach. The asymmetry is real, it was noticed in 1948, and the brief for this article names it as the tension the article must adjudicate. Adjudication begins by refusing two evasions. The first evasion calls clause (2)(b) an aberration, a lapse from the founders’ secularism. The text forbids that reading: clause (2) was debated, defended by Ambedkar and Munshi, and adopted in the same article as the freedom. The second evasion calls the asymmetry a proof of bad faith, evidence that the settlement was majoritarian from birth. The record forbids that reading too: the reform clause’s target was the majority community’s own orthodox establishment, its first legislative fruits fell on Hindu practice alone, and the members who carried the clause, Ambedkar above all, were the orthodox establishment’s most determined enemies.

The honest adjudication is harder and more interesting. The founders distributed their distrust unevenly because the social facts before them were uneven. Hindu society in 1948 presented the Assembly with a case that no other community presented in the same form: a caste order, sanctified by religious text and custom, that excluded millions from temples, wells, schools and occupations, and a national movement that had spent three decades promising to dismantle that order, from the Vaikom Satyagraha of 1924 to the Travancore Temple Entry Proclamation of 12 November 1936. The reform layer was aimed at that case. Its asymmetry records a judgment, contestable but intelligible, that the gravest religion-linked unfreedom in India was internal to the majority community and could be ended only by state power, while the gravest danger to the minority communities was state power itself, from which Layer Two’s rights were built to shield them. One community received a reforming state because its oppressions were exercised through religion’s institutions. The others received a shielding state because their vulnerability ran the other way. A reader may dispute the judgment; the Constituent Assembly’s Muslim members did, on the floor, when they asked why reform’s warrant should stop at one community’s boundary, and the uniform civil code was the founders’ deferred answer, placed among the Directive Principles in Article 44 where Nehru’s government could approach it when politics allowed, a placement whose own tensions are traced in the directive-principles settlement that housed the uniform civil code aspiration. What the reader cannot honestly do is describe the asymmetry as an oversight. The Assembly saw it, debated it, and enacted it, because the alternative, a symmetry of pure respect, would have frozen caste inside the temple door and called the freezing tolerance.

The strongest counter to the layered reading, stated at its strongest, must also be faced: in the ordinary run of Indian life, the layers do not collide at all. Temples are administered, denominations govern their affairs, processions proceed, schools teach, taxes are levied for general purposes, and no one invokes Article 25. A settlement whose tensions surface only in litigation and agitation might be described, with equal fidelity, as a single workable arrangement with friction at the edges. The counter is correct about the ordinary run and wrong about what it proves. Constitutional settlements are defined by their contested cases, because the uncontested cases never reach the forums where the settlement’s meaning is fixed. The disputes that reach the Supreme Court, the legislatures and the streets are precisely the disputes where the layers diverge: a denomination excluding women or castes, a state banning a slaughter or a conversion, a personal law colliding with an equality claim. In each such dispute, the parties do not argue about secularism in the abstract. They argue, in the founders’ own grammar, about which layer governs: the denomination’s Article 26 autonomy against the reformer’s Article 25(2)(b), the propagator’s Article 25(1) against the hearer’s conscience, the taxpayer’s Article 27 against the endowment department’s budget. That the argument always takes the layered form, seventy years after the text was written, is the layered reading’s best evidence. A settlement is what its arguments are made of.

Bhargava’s Principled Distance: The Strongest Rival Reading

No adjudication in this field is honest unless it states the rival reading at full strength, and the rival reading here is Rajeev Bhargava’s. Across the essays collected and extended in The Promise of India’s Secular Democracy in 2010, Bhargava argues that Indian secularism is a coherent, sui generis doctrine unjustly measured against Western models it never tried to copy. Its organising idea is principled distance. The state may engage with religion or disengage from it, may intervene in one religion while abstaining from another, may treat religions differently, provided every engagement, abstention and difference is justified by constitutional values: freedom, equality, fraternity, the ending of religion-linked hierarchy and exclusion. On Bhargava’s account, the features that look like incoherence to a separationist, the temple-entry laws, the personal-law pluralism, the state aid to denominational schools, are the doctrine working as designed, because principled distance authorises a differentiated posture and demands only that the differentiation answer to principle. The reading has genuine explanatory power. It captures why the Indian state’s interventions cluster around hierarchy and exclusion rather than around religion as such. It rescues the founding from the condescension of the failed-copy school. It has shaped two decades of scholarship and commentary, and any layered reading that cannot say what Bhargava gets right has not understood the settlement it claims to describe.

What Bhargava gets right is Layer Two’s moral centre, generalised. The founders did hold that the state’s posture toward religion must answer to the values of liberty and equality rather than to a geometry of distance; the December debates are full of exactly that reasoning, from Munshi’s defence of propagation as a human right to Ambedkar’s insistence that religion’s social empire could not be immunised from law. Where the principled-distance reading overreaches is in its claim of coherence. A doctrine is coherent when its parts answer to one ordering idea under pressure, and the founding record shows the parts answering to three different ideas whose ordering the founders declined to state. Principled distance, applied to the temple-entry question, justifies state intervention to end exclusion. Applied to the excommunication question in the Taher Saifuddin litigation, the same values justified the opposite result, the protection of a religious head’s power to expel a reformist follower from the faith, a power whose exercise is hierarchy in its purest form. Bhargava’s framework can narrate both outcomes after the fact, because its values are capacious enough to bless either. A framework that blesses either outcome in advance decides neither, and the founders’ text, unlike the framework, at least marks the battlefield honestly: clause against clause, layer against layer, with the ranking left to the institutions that must live with the result. The layered reading therefore absorbs Bhargava’s insight and denies his conclusion. Principled distance is a fine description of how the settlement’s custodians should behave. It is a poor description of what the settlement is. What the settlement is, textually and historically, is three commitments in permanent, productive, unresolved tension.

A word must also be said about the second rival, the failed-Western-copy reading, because it draws strength from the same coherence assumption in reverse. On that reading, Indian secularism promised separation, failed to deliver it, and the failure explains the settlement’s later troubles, from the Shah Bano reversal to the communal legislation of later decades. The reading fails at the record. The Assembly never promised separation. The Husain amendments offered a separationist-adjacent model and were defeated; the Shah amendments offered the secular label and were defeated; the text the Assembly enacted protects propagation, denominational autonomy and minority institutions in terms no separationist constitution uses. One cannot betray a promise one refused three times to make. The troubles of the later decades are real, and the sections ahead take them seriously, yet they are troubles of priority among the layers, fought by actors who invoke the founding text because the text genuinely supplies their weapons. That is a harder story than betrayal, and a more useful one, because it tells the reader where to watch: not at the distance between the founding’s promise and the present’s performance, but at the seam between the layers, where every generation’s real argument occurs.

Mehta, Jayal and the Case for the Layered Reading

The reading defended here descends from two scholars whose work the article’s brief names, and the descent should be made explicit, because the layered reading is an adjudication among authorities as much as among clauses. Pratap Bhanu Mehta, in the essay What is Constitutional Morality?, published in Seminar in 2010, and in a sequence of Indian Express essays after 2014, reads the Indian Constitution as a document whose moral ambitions are carried by institutions perpetually negotiating the gap between the text’s promises and the society’s arrangements. Mehta’s constitutional morality, a phrase he recovers from Ambedkar’s own usage, is not a doctrine that resolves tensions; it is a discipline for inhabiting them, a demand that public actors argue from the Constitution’s values rather than from power or sentiment when the text underdetermines the answer. The secularism settlement is the discipline’s home ground. Where the layers conflict, Mehta’s work suggests, the Constitution expects the conflict to be argued in the open, on reasons that the losing side can recognise as constitutional, and the expectation is itself a founding choice, the choice that a layered text institutionalises. Niraja Gopal Jayal, in Citizenship and Its Discontents, published in 2013, supplies the complementary insight from the direction of membership. Jayal shows that Indian citizenship was layered from the founding: a uniform legal status laid over communities whose religious and customary memberships the state continued to recognise, govern and, in the personal-law system, enforce. A layered citizenship and a layered secularism are the same settlement viewed from two sides. The citizen of Jayal’s account holds rights against the state as an individual, claims protections as a member of a denomination or minority, and stands exposed to the state’s reforming power as a subject of personal law, all at once, without the three standings ever being merged. The founders who voted the religion clauses in December 1948 were building Jayal’s citizen as surely as they were building Mehta’s arena.

The scholarly dispute, then, adjudicates as follows. Bhargava describes the settlement’s aspiration with unmatched elegance and mistakes the aspiration for the structure. The failed-copy school describes the settlement’s failures with unmatched confidence and mistakes a promise the founders never made for the measure of their work. The layered reading, following Mehta’s discipline and Jayal’s sociology, describes the structure itself: three commitments, each real, each textually grounded, each with its own December 1948 pedigree, arranged without a master rule for their collisions. The reading also explains the settlement’s strange durability better than either rival. A coherent doctrine can be discredited by its failures; a betrayed promise can be mourned into irrelevance. A layered settlement survives its failures because failure in one layer leaves the others standing, and because every political generation finds in the stack a layer that answers its own grievance: the excluded worshipper finds Layer Three, the threatened denomination finds Layer Two, the taxed citizen finds Layer One. Politicians keep returning to the founding text on religion not out of piety but out of inventory. The text still has something for everyone, which is what a settlement, as opposed to a doctrine, is for.

Evidence of the settlement has had to be preserved to be argued over, and the preservation is a scholarly and journalistic achievement that deserves a sentence in its own right. The Constituent Assembly Debates survive in official volumes and in the digital archives that research libraries maintain; the correspondence, drafts and marginalia that would let a future historian test this article’s reading survive only where someone has kept them against pressure and neglect. The same discipline appears wherever contested records are kept under threat: the working methods described in accounts of how journalists use VaultBook to protect sources whose testimony documents communal conflict rest on the identical premise that an argument about the past is only as strong as the archive that survives to be checked. A layered settlement, which assigns its collisions to future argument, is a bet on archives. December 1948 made the bet. Every library that keeps the debates, and every vault that keeps a witness’s papers, is part of the settlement’s machinery.

The Institutions Cases: Article 30 and the Boundaries of Respect

Layer Two’s largest battlefield turned out to be neither temples nor processions but classrooms, because Article 30’s guarantee that minorities may establish and administer educational institutions of their choice placed a denominational interior inside the state’s most regulated public system. The first great test came early. In the Kerala Education Bill reference, opinion delivered in 1958, the Supreme Court, speaking through Chief Justice S. R. Das, considered a Kerala statute that tightened state control over aided schools, many of them run by the Christian churches and the Nair and Ezhava organisations of the state. The Court read Article 30(1) generously toward the minority institutions, holding that the right to administer is real and not illusory, while accepting that the state may regulate aided institutions in the interest of educational standards, provided the regulation does not destroy the minority character the article protects. The formula, protection for character, regulation for standards, is Layer Two doctrine at its most durable, and it governed the field for decades: Ahmedabad St. Xavier’s College v State of Gujarat, decided in 1974, struck down provisions of a Gujarat university statute that would have converted the governing bodies of denominational colleges into state-shaped boards, holding that the right to administer includes the choice of the administering body itself.

The second great test pushed the formula toward its limit. In St. Stephen’s College v University of Delhi, decided in 1992, the Court sustained the college’s preference for Christian applicants up to fifty percent of admissions, treating a measured community preference as an incident of the Article 30 right rather than a breach of the equality code. A decade later, the eleven-judge bench in T. M. A. Pai Foundation v State of Karnataka, decided on 31 October 2002, reorganised the whole field, holding that the right to establish and administer institutions belongs to all citizens under Article 19(1)(g) in its occupational dimension and to minorities with additional protection under Article 30, that unaided institutions enjoy wide autonomy in admissions and fees, and that the state’s regulatory power varies with the aid it gives. Critics across the spectrum found something to mourn in T. M. A. Pai: secularists mourned the autonomy that let denominational and commercial managements escape public norms, and minority educators mourned the doctrinal route that dissolved their special protection into a general right. The layered reading notices something else, the continuity. From 1958 to 2002, through benches of every composition, the Court kept asking the December 1948 question in the December 1948 grammar: where does the denomination’s protected interior end and the state’s legitimate regulation begin? The answers moved. The question, and the clause that poses it, never have. An educational system in which a Jesuit college, a madrasa, an Arya Samaj school and a state university all operate under one constitution is the layered settlement’s daily plebiscite, held every admission season, and the settlement keeps passing it, because each institution in the system can point to the layer that shelters it and each regulator to the layer that empowers it. That is not the profile of a failed secularism. It is the profile of a settlement doing the work settlements exist to do, absorbing a plural society’s friction into argument about clauses, in forums built for argument, under rules that every side in the argument accepts as its own inheritance.

After the Founding: Seventy Years of Arguments About Which Layer Governs

The post-Independence record reads, under the layered lens, as a sequence of priority disputes, and the sequence begins almost at once. In Mohd. Hanif Quareshi v State of Bihar, decided in 1958, the Supreme Court considered the cow-protection statutes of Bihar, Uttar Pradesh and Madhya Pradesh, enacted under the shelter of the Directive Principle in Article 48, against the Article 25 claims of Muslim butchers whose livelihood and festival practice the bans burdened. The Court upheld prohibitions on the slaughter of cows and breeding cattle as reasonable restrictions serving the agricultural economy, while striking down total bans covering aged and useless cattle. Neither layer won outright. The reforming and directive state took the milch and draught animal; the practising citizen kept the rest; and the judgment’s reasoning moved openly between the state’s economic programme and the claimant’s religious freedom, weighing each against the other because the text supplied no scale. Cow protection has returned to legislatures in every decade since, most sharply in the Gujarat Animal Preservation (Amendment) Act of 2017, which raised the penalty for cow slaughter to imprisonment for life. The 2017 statute is a Layer Three instrument, state power over a practice defended by its opponents as religion and condemned by its authors as cruelty and economic sabotage, and its litigation has been argued, predictably, in the founders’ grammar: whose clause, whose layer, whose reading of December 1948.

The personal-law front supplied the settlement’s most instructive reversal. In Mohd. Ahmed Khan v Shah Bano Begum, decided on 23 April 1985, a five-judge bench of the Supreme Court held that a divorced Muslim woman was entitled to maintenance under section 125 of the Code of Criminal Procedure, a secular provision, notwithstanding the Muslim personal law’s narrower rule, and Chief Justice Y. V. Chandrachud’s judgment regretted the state’s failure to enact the uniform civil code that Article 44 had promised. Layer Three appeared to advance: a general welfare statute overriding a denominational rule, with the Court invoking the founding’s deferred aspiration. The political system answered within a year. The Muslim Women (Protection of Rights on Divorce) Act of 1986, enacted by Rajiv Gandhi’s government with its parliamentary majority, displaced the judgment for divorced Muslim women, and the displacement was defended in Parliament as Layer Two politics: respect for a minority community’s own law, against a judiciary accused of majoritarian reform in secular dress. Both sides quoted the founding. Both sides had a layer to quote. The sequel completed the lesson in the settlement’s strange resilience: in Danial Latifi v Union of India, decided in 2001, the Supreme Court upheld the 1986 Act while construing it to require a reasonable and fair provision for the divorced woman’s future, recovering much of Shah Bano’s substance inside the statute that had displaced it. No doctrine was vindicated. The layers were rebalanced, by a court this time, and the rebalancing is the sort of event the layered reading exists to notice: the settlement did not break under its most famous stress. It flexed, as settlements do, and the flexing is the history.

The judiciary’s most consequential intervention was doctrinal rather than distributive. In Kesavananda Bharati v State of Kerala, decided on 24 April 1973, Chief Justice S. M. Sikri’s enumeration of the Constitution’s basic features included the secular character of the Constitution, a phrase that quietly performed an extraordinary operation: it took a word the Assembly had refused to write into Article 1 or the original Preamble and installed it, at the level of unamendable structure, on the authority of the clauses the Assembly had written. Twenty-one years later, in S. R. Bommai v Union of India, decided on 11 March 1994, a nine-judge bench held that secularism is a basic feature of the Constitution and that a state government pursuing unsecular policies could be dismissed under Article 356, converting the settlement from a distribution of rights into a discipline on governments themselves. Bommai is the layered reading’s hardest case and its vindication at once. The word the founders withheld became the ground for dismissing elected governments, which no single layer authorises in terms; yet the bench reached the holding by reading the layers together, Articles 25 to 30 as a system, and the holding has disciplined every subsequent attempt to make a state government the instrument of one faith. The seam held a doctrine that no clause states. That is either a scandal or the settlement working at full stretch, and the founders’ own refusal to rank the layers is what makes the question permanently arguable, which is to say, permanently Indian.

The Post-2014 Contests: Layer Disputes in Modern Dress

The contests that have defined the years since 2014 are best understood, and most honestly argued, as layer disputes, and the layered reading’s value can be tested against the four that the record supplies. The first is cow protection, treated above: a directive-principle programme enforced through criminal law, answered by Article 25 and Article 19 claims, its Gujarati extreme enacted in 2017. The second is the triple talaq sequence. On 22 August 2017, in Shayara Bano v Union of India, a five-judge bench set aside the practice of instant triple talaq by a majority of three to two, Justices Kurian Joseph, R. F. Nariman and U. U. Lalit prevailing by different routes over the dissent of Chief Justice J. S. Khehar and Justice S. Abdul Nazeer, who would have protected the practice as personal law while asking Parliament to legislate. The Muslim Women (Protection of Rights on Marriage) Act of 2019 then criminalised the pronouncement itself, with imprisonment of up to three years. Reform layer throughout: the state entering a religiously sanctioned practice in the name of the women subject to it, exactly as Article 25(2)’s philosophy licenses and exactly as the Hindu Code precedent pioneered, with the standing objection, pressed at every stage, that a reform welcomed by many Muslim women was being carried by a political movement whose solicitude for Muslim women was contested by the women themselves. The layers frame the argument; they do not settle who argues in good faith, and any honest account must record that the good-faith question, not the constitutional grammar, is where that dispute’s heat lives.

What did the Muslim Women (Protection of Rights on Marriage) Act of 2019 change?

The Act of 2019 declared a pronouncement of instant triple talaq void and illegal, made the pronouncement a cognisable offence punishable with imprisonment of up to three years and a fine, and provided the affected wife with a subsistence allowance and custody rights, converting the Supreme Court’s 2017 invalidation of the practice into a criminal prohibition enforced by the state.

The third contest is Sabarimala. On 28 September 2018, in Indian Young Lawyers Association v State of Kerala, a five-judge bench held by four to one that the exclusion of women aged ten to fifty from the Sabarimala temple violated Articles 14, 15 and 25, and that the Ayyappa devotees did not constitute a separate religious denomination entitled to sustain the exclusion under Article 26. Justice Indu Malhotra, dissenting alone, argued that in a pluralist democracy the courts should not rationalise religion or dictate whose beliefs are essential, and that the denomination’s own understanding of its deity’s celibate form deserved Layer Two’s protection. Majority and dissent divide along the settlement’s seam with diagrammatic neatness: the majority read the case through reform and equality, asking what the practice did to the excluded; the dissent read it through respect and autonomy, asking what the state may say about a faith’s self-understanding. Both readings are available in the December 1948 text. That availability is not a defect the founders overlooked. It is the settlement’s structure, visible in a single judgment’s split, and the review petitions that followed, referred to a larger bench in November 2019, show the argument continuing inside the judiciary itself, which is where the founders, who gave the courts the essential-practices doctrine and the reform clause in the same breath, evidently expected it to live.

The fourth contest is citizenship. The Citizenship (Amendment) Act of 2019 amended the Citizenship Act of 1955 to offer an expedited path to Indian citizenship for Hindu, Sikh, Buddhist, Jain, Parsi and Christian migrants from Afghanistan, Bangladesh and Pakistan who had entered India on or before 31 December 2014, and omitted Muslim migrants from the benefit. The Act’s defenders argued Layer Two in a majoritarian key: India as the natural refuge of the subcontinent’s persecuted non-Muslim minorities, a reading with founding-era echoes in the promises made to minorities who stayed and to refugees who crossed in 1947. The Act’s challengers argued Layer One and the equality code: a citizenship law that classifies by religion abandons the non-establishment neutrality of the founding, and read the omission against the dimensional disagreements among Patel, Nehru, and Ambedkar that the secularism debates revealed, in which the founders had refused, against considerable pressure, to define the republic by faith. The litigation over the Act, and the street argument that accompanied it, were conducted as an argument about the founding because they were one. Each side’s founding is genuinely in the record. The layered reading does not pretend otherwise, and its honest verdict on the citizenship contest is the verdict it returns on the whole settlement: the text distributes weapons to both camps, and the outcome will be decided, as the founders arranged, by institutions arguing at the seam, not by the discovery that one camp’s layer was the true secularism all along.

A pattern completes the survey. In each contest, the political work is done by naming: each side claims the whole word secularism for its layer, and accuses the other side of betraying the word itself. The accusation is the tell. Where a settlement is layered, the struggle for the master noun is the struggle for priority among the layers, conducted by other means. The founders, who refused to write the noun into Article 1 in 1948 and watched Parliament write it into the Preamble in 1976 without touching a single religion clause, built a system in which the noun would always be available for capture and never sufficient for victory. The clauses decide. The clauses are layered. The argument, therefore, is permanent, and its permanence is not the failure of Indian secularism but the form of its life.

The Minorities Question Inside the Settlement

The religion clauses cannot be understood apart from the minorities settlement that the Assembly built around them, because for the members of 1948 the two projects were one project. A state emerging from Partition, with some thirty-five million Muslims remaining in India as the largest minority, had to decide how religious communities would appear inside its political architecture: as corporate bodies with guaranteed shares of power, or as citizens whose religious freedom the state protected while their political claims ran through common institutions. The colonial order had chosen the corporate route, through separate electorates in 1909 and their extension thereafter, and the national movement had blamed that route for the subcontinent’s division. The Constituent Assembly dismantled the corporate route in stages, and each stage illuminates a layer. The Advisory Committee under Patel, reporting in 1947, accepted reserved seats for minorities in legislatures as a transitional safeguard while rejecting separate electorates. Then, in May 1949, the Assembly went further: on the Advisory Committee’s recommendation, moved by Patel on 25 and 26 May 1949, the members abolished reserved seats for religious minorities altogether, retaining reservation only for the Scheduled Castes and Scheduled Tribes, whose disadvantage the Assembly classified as social and economic rather than religious. Begum Qudsia Aizaz Rasul, the only Muslim woman in the Assembly, supported the abolition, arguing that reservation kept minorities in a perpetual tutelage and that the community’s future lay in common citizenship. Muslim members who resisted, among them Z. H. Lari and Tajamul Husain in an earlier phase of the debate, warned that a minority stripped of guaranteed representation would depend wholly on majority goodwill.

The layered reading gives that May 1949 decision its proper place, as the political twin of the December 1948 clauses. What the minorities surrendered in guaranteed power, the settlement returned in guaranteed standing: Article 25’s freedom held by all persons, Article 26’s denominational autonomy, Article 29’s protection of language, script and culture, Article 30’s right to establish and administer educational institutions, the last of these a corporate right of exactly the kind the Assembly had refused in the electoral sphere. The pattern is too consistent to be accidental. Communities would not be constituencies; they would be rights-bearers. Their schools, endowments, rituals and internal affairs received constitutional armour thicker than the armour of any political party, while their claim to a fixed share of legislatures was retired as a colonial device that had ended in the country’s division. A reader who asks whether the founding was generous or grudging toward minorities is asking a question the settlement deliberately splits. In the register of power, the founding was austere: no separate electorates, no reservation, no weightage. In the register of faith and culture, the founding was lavish: rights that a simple majority could not touch, administered by courts instructed to ask what the denomination itself held essential. Both registers belong to the same design, and the design is the layered settlement wearing its political face. The later career of Article 30, litigated from the Kerala Education Bill reference of 1958 through St. Xavier’s College in 1974 to the minority-institution cases of the present century, shows the lavish register holding firm even as the austere register was contested in every generation’s reservation debates.

Personal Laws: The Chamber the Settlement Left Quiet

One chamber of the settlement was left deliberately quiet, and its quietness has generated as much argument as any clause: the personal laws. At the founding, marriage, succession, adoption and guardianship for Hindus, Muslims, Christians and Parsis were governed by distinct religiously sourced codes, administered by the state’s courts as law. The Constituent Assembly neither unified those codes nor expressly entrenched them. Instead, the founders fought the question once, on 23 November 1948, over Draft Article 35, the future Article 44, which directs the state to endeavour to secure a uniform civil code throughout the territory of India, and placed the aspiration among the non-justiciable Directive Principles. The debate of 23 November 1948 is a miniature of the whole settlement. K. M. Munshi defended the article as an instrument of national consolidation and of women’s emancipation from discriminatory custom. Members including Mahboob Ali Baig Sahib and Hussain Imam moved to protect personal law from the article’s reach, arguing that the codes were religion itself to their adherents. Ambedkar, closing, refused the premise of immutability: he pointed to the matrilineal law of the North Malabar Muslims, the Marumakkathayam system, as proof that what passed as unchanging Muslim law in one district was unknown in another, and he assured the House that any future code could begin as an optional instrument, binding only those who declared themselves under it. The article passed as a directive. The codes stayed. The promise and the postponement were enacted in the same gesture.

The courts then shielded the postponement. In State of Bombay v Narasu Appa Mali, decided in 1952, Chief Justice M. C. Chagla and Justice P. B. Gajendragadkar of the Bombay High Court held that the personal laws were not laws in force within Article 13 and therefore could not be struck down for inconsistency with the fundamental rights, while upholding the Bombay Prevention of Hindu Bigamous Marriages Act of 1946 as valid reform legislation. The two holdings together define the personal-law chamber’s architecture: the state may reform a community’s family law by statute, exercising Layer Three, yet the unreformed residue of that law stands beyond the reach of the equality and freedom clauses that govern every other statute. Ambedkar’s own trajectory measures the chamber’s cost. The Law Minister who had assured the Assembly in 1948 that codification could proceed gently spent 1948 to 1951 driving the Hindu Code Bill through select committee and floor, watched the Bill stall under orthodox and presidential resistance, and resigned on 27 September 1951 with the code unenacted, telling Parliament that to leave the reform where it stood was to make a farce of the Constitution’s promises. The four Acts of 1955 and 1956 enacted most of his code for Hindus after his resignation. No equivalent statute has ever been enacted for Muslim personal law by a legislature of the republic; the changes that came arrived through courts, in Shah Bano in 1985 and Shayara Bano in 2017, and through the contested criminal statute of 2019, each arrival reenacting the founding’s distribution of burdens. A settlement that reforms the majority’s law by legislation and the minority’s law by litigation has not malfunctioned. It has reproduced, decade after decade, the asymmetry the founders enacted when they aimed Layer Three at the community whose orthodox establishment sat inside their own chamber, and wrapped the other communities’ codes in Layer Two’s silence.

The Word Arrives by Amendment: 1976 and the Preamble

The word the Assembly refused in 1948 arrived in 1976, and the manner of its arrival is the layered settlement’s strangest chapter. During the Emergency, the Congress government of Indira Gandhi appointed the Swaran Singh Committee in 1976 to recommend constitutional changes, and the Constitution (Forty-second Amendment) Act of that year, the most sweeping amending statute in the republic’s history, rewrote the Preamble’s description of India as a sovereign socialist secular democratic republic, adding secular beside socialist, and inserting the word integrity into the fraternity clause. No religion clause was altered. Articles 25 to 30 stood exactly as December 1948 had left them. The amendment thus performed, at the level of the Preamble, the opposite of the founding’s operation: the founders had enacted the substance and refused the label; the Emergency’s Parliament enacted the label and touched none of the substance. Defenders of the 1976 change reply that the label merely named what the clauses already built, and that the Supreme Court in Kesavananda Bharati in 1973 had already counted the secular character among the basic features, so the amendment was declaratory rather than inventive. Critics reply that a word rejected after debate in 1948, inserted by a Parliament whose term had been extended and whose opposition sat in prison, carries a legitimacy deficit that no declaratory theory cures. The layered reading stays with the institutional facts, which support a narrower conclusion: whatever its legitimacy, the insertion changed the settlement’s politics without changing its law. After 1976, every contest among the layers could be fought for possession of a constitutional word, and the word’s presence in the Preamble gave courts from Bommai onward a textual peg for doctrines the clauses alone had carried since 1950. The Janata Parliament’s Forty-fourth Amendment of 1978 repealed or softened much of the Forty-second, and left the word secular standing in the Preamble untouched, a ratification across the political divide that counts for more, as evidence of the word’s acceptance, than the circumstances of its insertion count against it. The founders’ stack acquired a name, twenty-six years late, from the hands of an Emergency. The stack did not notice. The arguments continued at the seam, exactly where the founders had left them.

The Settlement the Founders Actually Left

Return, at the end, to the chamber in the first week of December 1948, where the settlement can be watched in the act of being made. A member from Bihar proposes that citizens shed every visible mark of their faiths, and the chamber, laughing at a Maulana’s joke about the mover’s own name, declines to unmark India. A member from Orissa presses to narrow the guaranteed freedom, and the chamber keeps the wide verbs: profess, practise, propagate. The Drafting Committee’s spokesman rises to say that the right to propagate is as old as humanity and inalienable, and in the same sittings the chamber writes into the same article the state’s power to reform religion and open its temples. No one in that chamber stood up to announce a doctrine of Indian secularism, and the absence of the announcement is the fact that the scholarship keeps circling. Doctrines were available: separation from America and France, establishment from the departing British practice, theocracy from the Pakistan being built next door. The Assembly enacted none of them. The Assembly enacted clauses, and the clauses, read in the order the Assembly adopted them, describe a state forbidden to keep a church or levy a religious tax, bound to honour every faith’s conscience, propagation and denominational self-rule, and authorised to enter religion itself where religion’s practices wound the citizens the Constitution elsewhere protects. Three layers. No harmonising rule. A working machine for arguing about religion and power in a society where both are everywhere.

The claim this article set out to defend can therefore be restated with its evidence attached. Indian secularism is not one principle; it is three layers that the founders did not fully harmonise. The layers are textual: Articles 27 and 28 for non-establishment, Articles 25(1), 26, 29 and 30 for equal respect, Article 25(2) for the reforming state. The non-harmonisation is documentary: the Shah amendments negatived, the Husain amendments defeated, no ranking clause moved or adopted across the December sittings of Volume VII. The consequence is historical: from Devaru in 1958 to Sabarimala in 2018, the contests that define the field have been priority contests among the layers, argued in the founders’ grammar by parties who find their weapons in the founders’ text. Readings that promise coherence, whether Bhargava’s principled distance or the separationist lament, purchase their elegance by discounting two-thirds of the evidence. The layered reading keeps the evidence and pays for it with an untidy conclusion, which the record amply justifies: the founders were not confused about religion and the state. They were precise about three different things at once, and they trusted their successors to fight over the ranking. This pillar of the founding, the Constitution and founding settlement whose other faces include the drafting committee’s labours and Tension One, the religious-secular settlement, in its widest form, rests on that trust. The argument continues in the pillar devoted to religion and politics, where each layer meets the movements, parties and judgments of the decades that inherited the seam, and where the question the founders left open is asked again in every generation: not whether India is secular, but which secularism, whose layer, governs this dispute, in this court, on this day.

Frequently Asked Questions

Q: What does the Indian Constitution say about secularism?

The Constitution distributes secularism across clauses rather than stating it as one doctrine. The Preamble, as amended in 1976, calls India a secular republic. Article 25 guarantees freedom of conscience and the right to profess, practise and propagate religion, while Article 25(2) lets the state regulate secular activity associated with religion and legislate for social welfare, reform and the throwing open of Hindu religious institutions. Article 26 protects denominational autonomy, Article 27 bars taxes earmarked for any particular religion, Article 28 restricts religious instruction in state-funded institutions, and Articles 29 and 30 protect minority culture and educational institutions. The Constituent Assembly debates of December 1948 show the founders building three layers, non-establishment, equal respect and reform authority, without ranking them, so the text answers most secularism questions by asking which clause, and therefore which layer, governs the dispute at hand.

Q: Is India officially a secular country?

Yes, in the sense the Constitution gives that word. The Preamble has described India as a secular republic since the Constitution (Forty-second Amendment) Act of 1976 inserted the word, and the Supreme Court held in S. R. Bommai v Union of India in 1994 that secularism is a basic feature of the Constitution, beyond the reach of parliamentary amendment. The Indian meaning of the word differs from the French or American separationist meanings. The Indian state has no official religion and may not levy a tax for any particular faith, yet the state regulates religious endowments, administers some temples, aids denominational schools under consent rules, and legislates to reform religious practices under Article 25(2). Officially secular in India therefore means bound to equal respect and non-establishment, not bound to keep a wall between the state and religion, because the founders in December 1948 rejected the separationist amendments and built a layered settlement instead.

Q: What does Article 25 say about religion?

Article 25 guarantees to all persons freedom of conscience and the right freely to profess, practise and propagate religion. The guarantee is expressly subject to public order, morality and health and to the other fundamental rights in Part Three. Clause (2) then preserves the state’s power to regulate or restrict any economic, financial, political or other secular activity associated with religious practice, and to make laws for social welfare and reform, including the throwing open of Hindu religious institutions of a public character to all classes and sections of Hindus. Explanation I counts the wearing and carrying of kirpans as part of the profession of the Sikh religion, and Explanation II provides that references to Hindus include Sikhs, Jains and Buddhists for the reform clause. The article therefore carries two of the settlement’s three layers in a single provision: an expansive freedom in clause (1), and a reforming state preserved in clause (2).

Q: What is principled-distance secularism?

Principled-distance secularism is Rajeev Bhargava’s account of the Indian model, developed across his essays and in The Promise of India’s Secular Democracy in 2010. Bhargava argues that the Indian state need not keep a fixed distance from religion, as separationist models demand, and need not treat all religions identically. Instead, the state may engage with a religion or abstain from it, intervene in one faith’s practices while leaving another’s untouched, provided the engagement, abstention or difference is justified by constitutional principle, above all freedom, equality and the ending of religion-linked hierarchy and oppression. The account explains why Indian practice combines personal-law pluralism with temple-entry legislation and state aid to minority schools without formal contradiction. Critics of the account, including the layered reading defended in this article, accept its description of the state’s permitted posture while denying that the posture amounts to a single coherent doctrine, because the founding text arranges the principles in layers whose ranking the founders left undecided.

Q: Did Ambedkar support a separate-religious-personal-laws system?

Ambedkar accepted the personal laws’ survival at the founding as a transitional fact, not as a permanent ideal. In the debate on the uniform civil code on 23 November 1948, he defended Draft Article 35, the future Article 44, against amendments that would have immunised personal law, pointing out that Muslim law itself varied across India, since the matrilineal Marumakkathayam system governed Muslims of North Malabar. He assured the Assembly that a future code might operate optionally at first, applying only to those who declared themselves bound by it. As Law Minister he then spent 1948 to 1951 driving the Hindu Code Bill toward enactment, resigning on 27 September 1951 when the Bill stalled. His position was therefore neither separatist nor immediately unificationist. Personal laws could persist while the republic matured, the state retained full authority to reform them, and the uniform civil code stood in the Directive Principles as the destination the settlement pointed toward without fixing a date.

Q: What is Article 25(2)(b)?

Article 25(2)(b) is the clause preserving the state’s power to make laws providing for social welfare and reform, or the throwing open of Hindu religious institutions of a public character to all classes and sections of Hindus, notwithstanding the religious freedom guaranteed by Article 25(1). It is the constitutional seat of the settlement’s third layer, the reform authority. The clause was debated in the Constituent Assembly in December 1948 as part of Draft Article 19, with K. M. Munshi and T. T. Krishnamachari among the members testing its scope, and Ambedkar defending the state’s need for a legislative field that religion could not fence off. Its Explanation II extends the reference to Hindus to Sikhs, Jains and Buddhists for this purpose. The clause underwrote the temple-entry legislation of the period and the reform philosophy behind the Hindu Code Acts of 1955 and 1956, and courts have read it against denominational rights under Article 26 in cases from Sri Venkataramana Devaru in 1958 to Sabarimala in 2018.

Q: How is Indian secularism different from Western secularism?

The difference lies in what each model asks the state to do. The American model, built on the First Amendment, combines non-establishment with free exercise and polices a boundary between religious and state action, while the French model of laïcité expects the public sphere itself to be cleared of religious display. Indian secularism, as the Constituent Assembly debates of December 1948 constructed it, asks for three things at once. The state establishes no religion and levies no religious tax, under Articles 27 and 28. The state guarantees religions a public presence, protecting profession, practice and propagation and denominational self-government under Articles 25 and 26. And the state keeps authority to reform religious practices and open religious institutions under Article 25(2), a power no separationist constitution grants in terms. The Assembly considered and rejected the alternatives: Tajamul Husain’s amendments would have stripped religion of public marks, and K. T. Shah’s would have written a single secular label into Article 1.

Q: What does Article 26 guarantee to religious denominations?

Article 26 guarantees to every religious denomination, or any section of one, four rights, subject to public order, morality and health. A denomination may establish and maintain institutions for religious and charitable purposes. It may manage its own affairs in matters of religion. It may own and acquire movable and immovable property. And it may administer that property in accordance with law. The Supreme Court’s judgment in the Shirur Mutt case in 1954 read the article as protecting a denomination’s interior life, holding that what counts as an essential religious matter is to be judged with reference to the denomination’s own doctrines. The same line of cases allowed the state to regulate the secular administration of denominational property, as in Ratilal Panachand Gandhi in 1954. Article 26 is the strongest textual home of the settlement’s second layer, equal respect, because it treats the denomination, not only the individual believer, as a bearer of constitutional rights against the state.

Q: What does Article 27 say about taxes and religion?

Article 27 provides that no person shall be compelled to pay any taxes whose proceeds are specifically appropriated in payment of expenses for the promotion or maintenance of any particular religion or religious denomination. The clause is narrow by design. It forbids the fiscal signature of establishment, a levy raised for a named faith, the arrangement under which the East India Company had once administered religious endowments as a department of government. It does not forbid general public spending that incidentally benefits religious bodies, such as policing a procession, funding a denominational college under the consent rules of Article 28, or exempting charitable trusts from taxation. Courts have accordingly sustained state administration of temples and even the annual Devaswom payments charged by Article 290A to Kerala and Tamil Nadu, treating those as regulation or inherited obligation rather than as a tax for a religion. Article 27 is Layer One in its purest form: the state may touch religion in many ways, but it may not bill the citizen for one faith’s upkeep.

Q: What does Article 28 say about religious instruction in publicly funded schools?

Article 28 sorts educational institutions into three classes and gives each class a different rule. In institutions wholly maintained out of state funds, no religious instruction may be provided at all. In institutions administered by the state but established under an endowment or trust requiring religious instruction, the trust’s terms govern and instruction may continue. In institutions recognised by the state or receiving state aid, religious instruction and worship are permitted, yet no student may be required to take part without consent, or, for a minor, the guardian’s consent. The scheme was the Assembly’s answer, in the December 1948 sittings, to a school system in which missions, samajes, anjumans and endowments ran much of the recognised provision with public money. A total ban would have destroyed that system; total permission would have made the state a paymaster for catechisms. Article 28 instead makes compulsion and public money the tests, keeping the state’s own classrooms neutral while leaving aided denominational schools room to teach their faiths voluntarily.

Q: Why did the Preamble adopted in 1949 not contain the word secular?

The omission was deliberate, and the record shows the word being refused rather than forgotten. On 15 November 1948, Professor K. T. Shah moved to describe India in Article 1 as a secular, federal, socialist union of states, and the Assembly negatived the amendment after B. R. Ambedkar replied that a constitution is a mechanism for regulating the organs of the state, not a device for prescribing the social order future electorates must accept. Further attempts followed on 25 November and 3 December 1948 and met the same fate. The founders expected the enforceable clauses, Articles 25 to 30, to do the work a label could not, and many members also distrusted the European church and state associations the word carried in a society where religion structured family, property and education. The word entered the Preamble only in 1976, through the Forty-second Amendment enacted during the Emergency, naming a settlement that had operated without it for twenty-six years.

Q: What did Tajamul Husain propose in the December 1948 religion debates, and why did the Assembly reject it?

Tajamul Husain, a Muslim member from Bihar, moved amendments to Draft Article 19 that would have made religion a private matter in the text itself. His best-known proposal would have replaced the kirpan Explanation with a rule that no person should bear any visible sign or mark or name, and no person should wear any dress, whereby his religion may be recognised, and his amendments also sought to cut back the propagation limb of the guaranteed freedom. Husain argued that Indians in Europe or America could not be sorted by faith from their clothes or names, and that national integration required the same unmarked citizenship. The Assembly declined. Maulana Hasrat Mohani answered that the mover’s own name announced his religion, and K. M. Munshi replied for the wider freedom that professing, practising and propagating a faith is an inalienable human right. The defeat settled a defining feature of Indian secularism: the republic would protect religion’s public presence rather than purchase civic peace with its disappearance.

Q: What work does the word propagate do inside Article 25?

The word propagate converts Article 25 from a charter of private belief into a charter of public religion. Profession declares a faith; practice performs it; propagation commends it to others, through preaching, teaching, publication and persuasion. The Constituent Assembly kept the word against attempts in December 1948 to narrow the clause, and K. M. Munshi defended it on the floor as part of an inalienable right of every human being. Its presence explains why conversion, missionary work and religious publishing stand inside the fundamental rights chapter, and why the Indian settlement cannot be described as privatising faith. The word carries its own limit, drawn by the Supreme Court in Rev. Stanislaus v State of Madhya Pradesh in 1977: propagation protects the exposition of one’s tenets, not a right to convert another person, because the hearer’s freedom of conscience stands equal to the speaker’s. Layer Two thus contains its own boundary, respect for the propagator bounded by respect for the person addressed.

Q: What did the Supreme Court decide about secularism in S. R. Bommai v Union of India in 1994?

In S. R. Bommai v Union of India, decided on 11 March 1994, a nine-judge bench of the Supreme Court held that secularism is a basic feature of the Constitution, that a state government which pursues policies contrary to secularism acts against the Constitution, and that a proclamation under Article 356 dismissing such a government is open to judicial review, with the court empowered to restore the dismissed government where the proclamation fails. The case arose from the dismissal of state governments after the demolition at Ayodhya on 6 December 1992, including the Bharatiya Janata Party governments of Madhya Pradesh, Rajasthan and Himachal Pradesh, whose dismissals the Court upheld, while striking down the dismissal of the Karnataka government that gave the case its name. Bommai converted the founding settlement into a discipline on governments: the religion clauses, read with the Preamble’s secular description, limit not only how the state treats believers but which political projects a state government may pursue and remain in office.

Q: What is the essential religious practices doctrine, and where did the Supreme Court first state it?

The essential religious practices doctrine holds that Articles 25 and 26 protect the practices a religion itself regards as essential and integral to it, while leaving the secular activity associated with religion open to state regulation. The Supreme Court first stated the test in Commissioner, Hindu Religious Endowments, Madras v Sri Lakshmindra Thirtha Swamiar of Sri Shirur Mutt, decided on 16 April 1954, where Justice B. K. Mukherjea held that what constitutes an essential part of a religion is primarily to be ascertained with reference to the doctrines of that religion itself. The doctrine was meant as a Layer Two safeguard, letting the denomination’s self-understanding define the protected core against reforming legislatures. Its career has been double-edged. The same test that shielded denominational autonomy in 1954 allowed later courts to hold practices non-essential and therefore reformable, in temple entry, in the Sabarimala case of 2018, and in challenges to personal law, making judges arbiters of doctrine in ways the 1954 bench had hoped to avoid.

Q: Which temple-entry reforms did Article 25(2)(b) make possible during the 1950s?

Article 25(2)(b) constitutionalised the throwing open of Hindu religious institutions of a public character to all classes and sections of Hindus, and the 1950s converted that warrant into practice across the temple-entry front that the freedom movement had opened with the Vaikom Satyagraha of 1924 and the Travancore Temple Entry Proclamation of 1936. The Madras Temple Entry Authorisation Act of 1947, already in force when the Constitution commenced, survived challenge because the new text expressly authorised its object, and the Supreme Court in Sri Venkataramana Devaru v State of Mysore in 1958 upheld state-ordered entry against a denomination’s Article 26 claim, while protecting the denomination’s rituals and its right to exclude non-Hindus. State temple-entry and endowment statutes of the decade proceeded on the same footing, opening public temples administered under the Hindu religious endowments framework to worshippers whom custom had excluded for caste. The reforms were the reform layer’s first legislative harvest: a religious disability abolished by statute, under a clause written for that purpose in December 1948.

Q: What secular activity associated with religion may the state regulate under Article 25(2)(a)?

Article 25(2)(a) preserves the state’s power to regulate or restrict any economic, financial, political or other secular activity which may be associated with religious practice. The clause rests on a distinction the courts elaborated from 1954 onward: a practice may be religious in motive and setting while remaining secular in its operative character, and that character is regulable. Temple finances and endowment administration are the standard case, supervised by state Hindu religious and charitable endowment departments whose validity the Supreme Court sustained in the Shirur Mutt line of cases, even while denominational worship stayed protected. Processions, loudspeaker use, slaughterhouse licensing, the employment conditions of temple staff, the audit of religious trusts and the management of religious property all fall on the regulable side of the line when they present as public order, health or administration questions. The clause is the settlement’s workhorse, deciding more disputes than the celebrated reform clause, because most friction between a religious institution and the state concerns money, land and noise rather than doctrine.

Q: What did the Sabarimala judgment of September 2018 decide, and which layer of the settlement did it invoke?

On 28 September 2018, in Indian Young Lawyers Association v State of Kerala, a five-judge bench of the Supreme Court held by four to one that the exclusion of women aged ten to fifty from the Sabarimala temple in Kerala violated the rights to equality and religious freedom in Articles 14, 15 and 25, and that the devotees of Lord Ayyappa did not form a separate religious denomination entitled to maintain the exclusion under Article 26. The majority read the case through the settlement’s third layer, reform, asking what the practice did to the excluded women and treating the entry bar as a disability the Constitution empowered courts and legislatures to remove. Justice Indu Malhotra’s lone dissent answered from the second layer, equal respect, arguing that a pluralist democracy should not let courts rationalise a faith’s self-understanding. Review petitions were later referred to a larger bench in November 2019, leaving the priority argument between the layers formally open, as the founders’ unranked text predicts.

Q: What did Chief Justice Sikri list among the basic features of the Constitution in Kesavananda Bharati in 1973?

In Kesavananda Bharati v State of Kerala, decided on 24 April 1973, Chief Justice S. M. Sikri, setting out the features of the Constitution that Parliament’s amending power could not destroy, listed among them the supremacy of the Constitution, the republican and democratic form of government, the secular character of the Constitution, the separation of powers, and the federal character of the Constitution. The appearance of the secular character in that list performed a quiet constitutional operation. The Constituent Assembly had refused in November and December 1948 to write the word secular into Article 1 or the Preamble, choosing to enact clauses instead of a label. Sikri’s enumeration read those clauses, Articles 25 to 30 with the equality code, as amounting to a structural commitment deep enough to limit even constitutional amendment. The word the founders withheld as a slogan returned a generation later as unamendable structure, an outcome the layered reading treats as the settlement’s own logic completing itself in court.

Q: How did the abolition of reserved seats for religious minorities in May 1949 complete the founding settlement on religion?

In May 1949, acting on the Advisory Committee’s recommendation moved by Sardar Vallabhbhai Patel on 25 and 26 May 1949, the Constituent Assembly abolished reserved seats in legislatures for religious minorities, keeping reservation only for the Scheduled Castes and Scheduled Tribes, whose disadvantage it classified as social rather than religious. Begum Qudsia Aizaz Rasul supported the change, arguing that reservation kept minorities in tutelage. The decision completed the December 1948 clauses by fixing the political register of the settlement. Communities would not appear in the Constitution as constituencies with guaranteed shares of power, the colonial device of separate electorates and weightage being retired after Partition’s verdict on it. They would appear instead as rights-bearers, protected in conscience, denominational autonomy, culture and educational institutions by Articles 25, 26, 29 and 30, armour a majority could not vote away. Power was made common; faith was made inviolable. The two moves together, austerity in representation and generosity in religious standing, are the founding settlement in its political form.