The Citizenship (Amendment) Act 2019 changed who may belong to India by changing the place of religion inside the law of belonging itself. Passed by the Lok Sabha on 9 December 2019 and by the Rajya Sabha on 11 December 2019, assented to by President Ram Nath Kovind on 12 December 2019 and brought into force on 10 January 2020, the Act amended the Citizenship Act 1955 so that Hindus, Sikhs, Buddhists, Jains, Parsis and Christians who entered India from Pakistan, Bangladesh or Afghanistan on or before 31 December 2014 would no longer be treated as illegal migrants and could reach citizenship through a shortened naturalisation period of five years rather than eleven. Every element of that sentence carries an argument. The six named faiths, the three named countries, the fixed entry date and the shortened waiting period together compose a statute that sorts applicants by religion for the first time in the history of the 1955 Act, and the debate that followed, in Parliament in December 2019, in the Supreme Court through more than two hundred petitions, and in the streets of Assam and Delhi through the winter of 2019 and 2020, turned on a single architectural question: whether religion may serve as a criterion of citizenship inside the constitutional order that the Constituent Assembly built between 1947 and 1949 and that the Citizenship Act 1955 then codified.

Two Readings of One Statute

A statute of eleven sections and a few hundred operative words generated two distinct literatures within weeks of its passage, and keeping those literatures separate is the first condition of understanding either. The first literature reads the Act politically. It treats the winter of 2019 and 2020 as a contest over the direction of the Indian state, in which the governing Bharatiya Janata Party pressed a programme its opponents described as majoritarian and its defenders described as corrective, and in which the protests at Shaheen Bagh in Delhi, the marches across Assam and the resolutions of state assemblies formed a single arc of mobilisation and counter-mobilisation. That reading has its own full treatment in this series, in the constitutive-opposition pattern reading, which situates the contest inside the Modi-era renegotiation of the founding settlement. The second literature, the one this article develops, reads the Act architecturally. It asks what the statute does to the structure of citizenship law, which provisions it alters, which constitutional tests those alterations must pass, and what follows for the design of Indian citizenship if religion becomes a lawful sorting device at the point of admission.

The claim defended here is the brief’s namable claim, stated plainly: the CAA debate engages the architectural-substantive question of religion in citizenship, and the architectural-substantive reading is analytically distinct from the political-mobilisation reading. Distinct does not mean detached. The architectural question was argued in political places, by political actors, for political ends, and no account of it can pretend otherwise. The distinction is about what kind of question is being answered. The mobilisation reading answers a question about power: who organised, who resisted, who prevailed electorally. The architectural reading answers a question about design: what does the amended statute say, feature by feature, and can those features stand inside the constitutional architecture that Articles 5 to 11 sketched in 1949, that Parliament legislated in 1955 and that the Supreme Court has policed since the reasonable-classification doctrine took shape in the 1950s and the basic-structure doctrine in 1973? A reader who confuses the two questions will misjudge both, treating a constitutional defect as though it were a campaign setback or a campaign victory as though it settled a constitutional point.

Three bodies of material decide the architectural question, and this article works through each in turn. The first is the statutory text itself: the Citizenship (Amendment) Act 2019 in full, read against the Citizenship Act 1955 it amends, clause by clause. The second is the contestation record: the parliamentary debates of 9 and 11 December 2019, in which Home Minister Amit Shah stated the government’s rationale in its own terms and opposition members stated the constitutional objections in theirs; the Statement of Objects and Reasons appended to the Bill; and the petitions filed in the Supreme Court from December 2019 onward, which converted the parliamentary objections into pleadable constitutional claims under Article 14 and the basic-structure doctrine. The third is the operational record: the Citizenship (Amendment) Rules notified on 11 March 2024, which translated the 2019 text into application forms, district committees and documentary schedules four years after the Act entered into force, and in doing so revealed which parts of the architecture bore weight and which remained paper. A debate conducted across a statute, a court docket and a rulebook is a debate about architecture, whatever else it also is.

The frame for holding that architecture in view comes from the contested-settlement framework, which reads Indian secularism not as a finished achievement of 1950 but as a settlement repeatedly reopened at moments when the state redefines its relation to religion. Citizenship is one of the settlement’s load-bearing walls. The Constituent Assembly debates of August 1949, reconstructed in the post-1947 plural-civilisational settlement, show the founders refusing a religious test for founding citizenship even while Partition’s violence pressed one upon them, and instead writing a jus soli-leaning code for the initial citizenry while leaving future admissions to Parliament under Article 11. The 2019 Act is the most consequential exercise of that Article 11 power in the republic’s history, and its defenders and critics both argued from the founding record, the defenders pointing to Partition’s unfinished business and the critics to the founders’ refusal. The architectural reading takes both invocations seriously by testing them against the text the founders actually left and the doctrines the Court actually built.

The Statute on the Page: What the 2019 Act Changed

The Citizenship Act 1955 is the parent statute, enacted under Article 11 to govern acquisition and termination of citizenship after the Constitution’s commencement provisions in Articles 5 to 10 had fixed the initial citizenry. Its scheme is organised by mode of acquisition: citizenship by birth under Section 3, by descent under Section 4, by registration under Section 5, by naturalisation under Section 6 and by incorporation of territory under Section 7. Two features of that scheme matter for everything that follows. First, the 1955 Act as enacted contained no religious criterion for any mode of acquisition; eligibility turned on birth, parentage, residence, marriage and the qualifications in the Third Schedule, which demanded, among other conditions, residence or government service aggregating eleven years within the fourteen years preceding an application, including the twelve months immediately before it. Second, the Act’s definition of the illegal migrant, inserted into Section 2(1)(b) by the Citizenship (Amendment) Act 2003, barred a foreigner who entered without valid travel documents, or overstayed a visa, from both registration and naturalisation. The 2003 bar made the definition section the gate through which every later liberalisation had to pass, and the 2019 Act passed through exactly that gate.

The 2019 amendment works through three operative moves. Its first move inserts a proviso into Section 2(1)(b), declaring that any person belonging to the Hindu, Sikh, Buddhist, Jain, Parsi or Christian communities from Afghanistan, Bangladesh or Pakistan who entered India on or before 31 December 2014, and who has been exempted by the central government under the Passport (Entry into India) Act 1920 or the Foreigners Act 1946, shall not be treated as an illegal migrant for the purposes of the Act. The proviso does not grant citizenship; it removes a disqualification, converting a class of persons whom the 2003 amendment had placed permanently outside the door into persons entitled to knock. Its second move inserts Section 6B, a special provision for the class the proviso defines. Under Section 6B the central government or a specified authority may grant a certificate of registration or naturalisation to a person covered by the proviso, subject to conditions and restrictions the rules prescribe; proceedings pending against such a person under the Foreigners Act 1946 or the Passport (Entry into India) Act 1920 abate on the grant; and the person is deemed to be a citizen of India from the date of entry into India. The deeming clause is architecturally the boldest line in the Act, because it reaches backward: citizenship, once granted, is backdated to arrival, rewriting the legal character of years the person spent in India as an irregular entrant. Its third move amends the Third Schedule so that, for the class Section 6B covers, the aggregate residence qualification falls from eleven years to five. A fourth, quieter move amends Section 7D to add violation of the Citizenship Act’s provisions to the grounds on which an Overseas Citizen of India cardholder’s registration may be cancelled, a provision whose connection to the main scheme is disciplinary rather than constitutive and which drew less debate than the other three.

What did the Statement of Objects and Reasons record as the Act’s purpose?

The Statement of Objects and Reasons records that the constitutions of Pakistan, Afghanistan and Bangladesh provide for a specific state religion, that many persons belonging to the six named communities faced persecution on grounds of religion there, that many sought shelter in India, and that their applications stalled for want of valid entry documents.

That stated purpose does work inside the constitutional argument, because the reasonable-classification doctrine asks what objective a classification serves and whether the classification bears a rational relation to it. The government’s defenders, beginning with Amit Shah’s speech introducing the Bill in the Lok Sabha on 9 December 2019, held the Statement of Objects and Reasons to be the authoritative statement of that objective: the relief of religious persecution in three neighbouring states whose constitutions establish Islam. Shah told the Lok Sabha that the named communities faced persecution in the three named countries, that the Bill would give dignity and citizenship to people who had lived in India for decades without either, and that the Act took nothing from any Indian citizen of any faith. He added the government’s limiting proposition: Muslims were not named because the three countries’ constitutions make Islam the state religion, so a Muslim applicant from those countries could not claim the persecution ground the Bill addressed, while remaining free to apply under the unamended provisions like any other foreigner. The critics’ answer, stated in the same debate by Congress member Manish Tewari and Trinamool Congress member Mahua Moitra among others, was that a purpose stated in persecution terms cannot sustain a classification drawn in religious terms, because persecution in the named countries also fell on groups the Bill omitted, and because the Indian Constitution’s equality guarantee does not permit the state to sort claimants by faith even for a beneficent end. Both positions were fully formed within the first day’s debate, and the five years of litigation and commentary that followed elaborated them without replacing them.

Two territorial exclusions inside the Act deserve the same close reading as its inclusions, because they show the legislature negotiating with older settlements rather than overriding them. The proviso and Section 6B do not apply to the tribal areas of Assam, Meghalaya, Mizoram and Tripura included in the Sixth Schedule to the Constitution, nor to the areas covered by the Inner Line Permit system under the Bengal Eastern Frontier Regulation 1873, which in December 2019 covered Arunachal Pradesh, Nagaland and Mizoram and was extended to Manipur by a presidential order that same month. Those carve-outs preserved the protective architecture for indigenous communities in the Northeast that the Constitution and colonial-era regulation had layered over the region, and their presence in the Act concedes, within the government’s own text, that citizenship liberalisation in India has never been a single national question but a set of regional settlements of different ages and different terms. The Northeast dimension of the settlement is traced in the Assam-specific architectural-context, and the Assam argument receives its own section below.

The Four Features: An Architectural Map

Read as design rather than as slogan, the 2019 Act has four features, and each feature attracts a distinct constitutional argument. Separating them matters because public debate habitually fuses them, so that a defence of one feature is heard as a defence of all four and an attack on one as an attack on the whole. The first feature is the religion-based eligibility criterion: the proviso to Section 2(1)(b) names six faiths, Hindu, Sikh, Buddhist, Jain, Parsi and Christian, whose adherents from the named countries escape the illegal-migrant bar, and it does not name Islam. The second feature is the country-specific eligibility: the named countries are Afghanistan, Bangladesh and Pakistan, three states whose constitutions establish Islam as the state religion, while other neighbours with documented records of religious persecution, Sri Lanka and Myanmar among them, stand outside the list. The third feature is the time-based cutoff: only persons who entered India on or before 31 December 2014 qualify, so the Act addresses a closed historical class of entrants rather than an open future flow. The fourth feature is the fast-track mechanism: the Third Schedule amendment halves the residence qualification for the covered class from eleven years to five, and Section 6B backdates the resulting citizenship to the date of entry. The table below, labelled here as the Four-Feature Architectural-Substantive Map of the Citizenship (Amendment) Act 2019, sets the four features against the statutory language that creates them, the contestation argument each one draws and the form each took when the 2024 Rules operationalised the Act, so that a reader can cite the map by name and test any claim about the Act against the feature it actually concerns.

Feature	Statutory form in the 2019 Act	Contestation argument it attracts	Form in the 2024 Rules
Religion-based eligibility criterion	Proviso to Section 2(1)(b) names Hindus, Sikhs, Buddhists, Jains, Parsis and Christians; Islam is not named	Article 14 classification challenge and secularism as basic structure	Eligibility certificate attesting the applicant’s religion, issued through a local community institution
Country-specific eligibility	Afghanistan, Bangladesh and Pakistan named; Sri Lanka, Myanmar and other neighbours omitted	Under-inclusion argument: persecution also documented outside the three named states	Documentary proof of nationality of one of the three named countries required
Time-based eligibility cutoff	Entry into India on or before 31 December 2014	Closed-class defence by the government; arbitrariness objection from petitioners	Documentary proof of entry date drawn from a prescribed list of twenty documents
Fast-track mechanism	Third Schedule residence qualification reduced from eleven years to five; Section 6B deems citizenship from date of entry	Whether speed of grant cures or compounds the classification defect	Online application to a district level committee, decision by a state empowered committee

The map clarifies the debate’s geometry. The religion criterion is the feature against which the basic-structure argument is aimed, because that argument concerns the character of the state rather than the fit of a classification. The country list and the cutoff are the features against which the Article 14 under-inclusion argument is aimed, because that argument asks why similarly placed persons, Tamil Hindus from Sri Lanka, Rohingya Muslims from Myanmar, Ahmadiyya Muslims from Pakistan itself, stand outside a classification whose stated ground is persecution. The fast-track mechanism, by contrast, has attracted almost no independent objection: no petitioner argued that five years is too short a residence qualification for a refugee class as such, and the government’s defenders were entitled to note that Parliament had always set residence periods by policy. A debate that treats the Act as one undifferentiated object misses that its features carry unequal constitutional weight, and that a court could in principle sustain the mechanism while striking the criterion, or sustain the criterion while questioning the cutoff. The severability question, what survives if one feature falls, is part of the architecture too, and the petitions addressed it expressly.

The four features also have a genealogy that the architectural reading must record, because the 2019 Act did not arrive from nowhere. A predecessor Bill, the Citizenship (Amendment) Bill 2016, was introduced in the Lok Sabha on 19 July 2016 and referred to a Joint Parliamentary Committee chaired by Rajendra Agrawal, which heard state governments, security agencies and civil society delegations and reported on 7 January 2019. The 2016 Bill passed the Lok Sabha on 8 January 2019 but was not taken up in the Rajya Sabha amid protests in the Northeast, and it lapsed with the dissolution of the Sixteenth Lok Sabha. The 2019 Bill differed from its predecessor in particulars the Committee’s record illuminates: the Sixth Schedule and Inner Line carve-outs were carried into the enacted text after the Committee heard Assamese and tribal delegations argue that an unqualified liberalisation would dissolve the demographic protections their settlements guaranteed. The genealogy matters constitutionally because the classification the Court is asked to review was not drafted in a single night; it passed through a committee process in which the under-inclusion objection and the Assam objection were both placed on record and answered, in the majority’s view adequately and in the dissenters’ view not at all. Notes of dissent appended to the 2019 committee report by opposition members rehearsed, almost word for word, the Article 14 argument the petitions later carried to the Supreme Court.

Behind the 2016 Bill stands a longer administrative history that both sides mined. Successive governments had regularised Partition-era and later entrants through executive exemptions under the Passport (Entry into India) Act 1920 and the Foreigners Act 1946, and notifications issued in 2015 and 2016 by the Ministry of Home Affairs exempted Hindus, Sikhs, Buddhists, Jains, Parsis and Christians from Afghanistan, Bangladesh and Pakistan who had entered on or before 31 December 2014 from penal provisions for irregular entry, the same six faiths, the same three countries and the same cutoff the statute later adopted. The 2019 Act, on the government’s account, converted an executive indulgence into a statutory pathway. On the critics’ account, the same genealogy proves the opposite point: what the executive had done by exemption, tolerating presence, the legislature did by classification, distributing a benefit by faith, and the difference between toleration and classification is precisely where Article 14 begins to bite. The architectural reading does not need to choose between the two accounts of the genealogy to note what the genealogy establishes: the four features were stable across a decade of executive and legislative drafting, which marks them as deliberate design rather than drafting accident.

The Equality Argument: Article 14 and the Classification Test

The first contestation argument, and the one on which the lead petitions principally rest, arises under Article 14 of the Constitution, which provides that the State shall not deny to any person equality before the law or the equal protection of the laws within the territory of India. The petitioners’ opening move is textual and short. Article 14 speaks of any person, not any citizen, so the equality guarantee covers the foreigner standing at the citizenship door, and a statute that distributes access to naturalisation by religion must answer to it. The government’s opening move is equally structural: citizenship is a sovereign function, Articles 5 to 11 commit the subject to Parliament, and the classification doctrine gives the legislature room to choose among classes of entrants by criteria rationally connected to a legitimate purpose. The contest between those two moves is the oldest contest in Indian equality law, and the doctrine that mediates it was built in the republic’s first decade.

How does Article 14 protect a person who is not a citizen?

Article 14 protects any person within the territory of India, words the Constituent Assembly chose over a citizens-only formula, so a foreigner may invoke equality before the law against the Indian state despite holding no right to enter, reside or naturalise, and classifications among foreigners must satisfy the same reasonable-classification doctrine as classifications among citizens.

The doctrine is the reasonable-classification test, stated by the Supreme Court in the State of West Bengal v. Anwar Ali Sarkar in 1952 and systematised by Justice S. R. Das in Ram Krishna Dalmia v. Justice S. R. Tendolkar in 1958. A classification survives Article 14 if it rests on an intelligible differentia that distinguishes the persons grouped together from those left out, and if that differentia bears a rational relation to the objective the statute seeks to achieve. The Dalmia propositions add that the legislature may proceed degree by degree, that under-inclusion alone does not condemn a classification aimed at an evil in stages, and that a presumption of constitutionality attends the legislative choice. Every serious argument about the CAA, for and against, is an argument staged inside that two-limb test, and the architectural reading gains its precision by keeping the limbs separate. The intelligible-differentia limb is barely contested: religion, country of origin and entry date are all perfectly intelligible criteria, and no petitioner argued otherwise. The rational-nexus limb carries the whole fight. The petitioners, led in argument by senior counsel including Kapil Sibal and Abhishek Manu Singhvi in the early hearings, contend that the differentia of religion lacks a rational nexus to the stated objective of relieving persecution, because persecution in the named countries is not distributed by the line the statute draws: Ahmadiyyas, declared non-Muslim by Pakistan’s constitution yet excluded from the proviso’s named faiths as Muslims in the Indian reading of the list, Shia and Hazara communities in Afghanistan and Pakistan, and atheists and dissenters in Bangladesh all faced documented persecution while standing outside the six named faiths. A classification that relieves persecution for some of the persecuted and not others, sorted by the very criterion of faith, is on this argument a classification whose differentia contradicts its objective.

The government’s answer, stated by Amit Shah in the Lok Sabha on 9 December 2019, by Law Minister Ravi Shankar Prasad in the Rajya Sabha on 11 December 2019 and by the Union of India in its affidavits before the Supreme Court, meets the nexus argument on its own ground. The objective, the government says, is not the relief of persecution in the abstract but the relief of religious persecution of minorities in three states whose constitutions establish Islam, states in which the named communities are, by constitutional definition, religious minorities and in which Muslims are not. On that framing the differentia is not religion generally but minority status within a theocratic constitutional order, and the nexus is exact: the class relieved is the class the named states’ own constitutions place outside the state religion. The Dalmia answer to under-inclusion completes the defence: Parliament may address a problem in stages, and the existence of other persecuted groups, in other countries or within the named ones, does not oblige the legislature to relieve all persecution everywhere or none anywhere. Tamil refugees from Sri Lanka and Rohingya from Myanmar raise different foreign-policy and security questions, the government contends, which Parliament may take up separately, as it has handled Tibetan and Sri Lankan Tamil admissions through executive schemes outside the 1955 Act’s naturalisation track.

What route remains open to an applicant outside the six communities under the unamended law?

An applicant outside the six named communities, including a Muslim from Dhaka or Karachi, may still seek citizenship by registration under Section 5 or by naturalisation under Section 6 of the 1955 Act on the same terms as any other foreigner, provided the applicant entered lawfully and is not an illegal migrant under the Section 2(1)(b) definition.

The adjudication this article offers on the Article 14 contest is architectural rather than partisan, and it can be stated in one sentence: the classification’s survival depends entirely on which statement of the legislative objective the Court accepts, and that dependence is itself the substantive finding. If the objective is read from the Statement of Objects and Reasons at the level of persecution generally, the under-inclusion objection has real force, because the statute’s line leaves named-country persecution unrelieved along a religious boundary. If the objective is read at the level the government urges, persecution of constitutionally defined minorities in states with an established religion, the nexus argument largely dissolves and the Dalmia stage-by-stage answer covers the residue. Indian courts have historically allowed the legislature to state its objective at the level that sustains the statute, and the presumption of constitutionality pushes the same way, which is why several commentators, including constitutional scholar Gautam Bhatia writing in December 2019, treated the Article 14 challenge as necessary but uphill, and why Bhatia and others placed additional weight on a second, less litigated limb: the argument that a classification drawn on religion is not merely under-inclusive but manifestly arbitrary in the sense the Court gave that phrase in Shayara Bano v. Union of India in 2017, where arbitrariness itself became a ground for striking legislation under Article 14. The petitions plead both limbs. The architectural point stands apart from the pleading: a statute whose validity turns on the altitude at which its objective is described has made the description of purpose, a question of drafting and of parliamentary record, into the decisive constitutional act, and the December 2019 debates are therefore part of the statute’s operative architecture and not merely its political scenery.

The Basic Structure Argument: Secularism After Kesavananda

The second contestation argument reaches deeper than Article 14, because it does not ask whether the classification is reasonable but whether the Constitution permits religion to function as a citizenship criterion at all. Its source is the basic-structure doctrine announced by a thirteen-judge bench in Kesavananda Bharati v. State of Kerala on 24 April 1973, in which a majority of the Court held that Parliament’s amending power does not extend to damaging the Constitution’s basic structure. Several of the Kesavananda opinions listed secularism among the basic features, and the nine-judge bench in S. R. Bommai v. Union of India, decided on 11 March 1994, treated secularism as part of the basic structure in the course of upholding and limiting the dismissal of state governments after the Babri Masjid demolition. The petitioners’ basic-structure argument builds from those holdings in two steps. The first step is that a state constitutionally committed to secularism, understood as equal treatment of religions and as the absence of a state religion, cannot make an applicant’s religion the condition of a statutory benefit. The second step is sharper and less commonly stated: citizenship law is not ordinary legislation, because it defines the boundaries of the political community itself, so a religious criterion at that boundary does not merely classify within the community but constitutes the community on a religious principle, which is the very thing the founding settlement refused when the Constituent Assembly rejected separate electorates and religious tests in August 1949.

The government’s answer to the basic-structure argument is jurisdictional before it is substantive. The basic-structure doctrine, the Union of India contends, disciplines constitutional amendments, and its application to ordinary legislation is contested ground: a statute is tested against the constitutional text, including Article 14, and a freestanding basic-structure challenge to legislation untethered to a fundamental right would convert every major statute into a constitutional amendment for review purposes. Where the doctrine has touched legislation, as in the election-law cases, it has operated through enumerated rights. The substantive answer follows: secularism in the Indian settlement has never meant the erasure of religion from law. The Constitution itself contains religion-conscious provisions, from the religious freedom guarantees of Articles 25 to 30 to the state’s power to reform religious practice, and Indian legislatures have long enacted faith-specific statutes, most visibly in the personal-law system whose renegotiations this series traced in the religious-personal-law-state-intervention renegotiation. A law that takes religion into account, the government argues, offends secularism only if it establishes a religion or discriminates invidiously, and a remedial classification favouring persecuted minorities of neighbouring theocracies does neither. Defenders added a historical answer: the founding generation itself distributed citizenship consequences along lines drawn by Partition, in Articles 6 and 7, which addressed migration from and to Pakistan in terms inseparable from the communal character of that migration, so a citizenship law attentive to religious persecution in Pakistan’s successor states works within the founding idiom rather than against it. That Partition dimension of the founding text is part of the integrated three-tensions framework through which this series reads the republic’s recurring renegotiations of its settlement.

The architectural reading adjudicates this contest by separating a strong claim from a weak one on each side. The petitioners’ strong claim is the boundary claim: the point at which a person crosses from outsider to member is the point at which the community declares what kind of community it is, and a declaration made in religious terms at that point has a constitutional significance that a religion-conscious personal law, operating among persons already members, does not. That claim draws genuine support from the founding record, in which the Assembly, even while legislating for a Partition emergency, refused to write a religious test into Articles 5 to 11 and instead wrote migration, domicile and parentage. The petitioners’ weak claim is the freestanding one, that basic structure alone, without Article 14, invalidates the Act, because the Court has not in fact developed a practice of striking ordinary statutes on that basis and the petitions themselves plead Article 14 first. The government’s strong claim is the textual one about Articles 25 to 30 and the personal-law tradition: Indian secularism is a secularism of principled distance and equal respect, not of blindness to religion, and it cannot be read as a general prohibition on religious criteria in law. Its weak claim is the Partition one, because Articles 6 and 7 sorted migrants by territory and date rather than by faith, granting rights to persons who migrated from Pakistan’s territory without asking their religion, which is a precedent for territorial classification under emergency rather than for religious classification in settled times. The net architectural finding is that the basic-structure argument functions less as an independent sword than as a lens for the Article 14 case: it raises the stakes of the nexus inquiry by arguing that where the classification criterion is religion and the subject is membership itself, the Court should scrutinise the fit between criterion and objective with less deference than the Dalmia presumption ordinarily supplies. Whether the Court adopts that heightened scrutiny is the doctrinal question on which the litigation, when it is finally heard on the merits, will turn.

Assam: Where the Architecture Meets a Different Settlement

The third contestation argument is Assam’s, and it differs in kind from the other two because Assam objected to the Act from the opposite direction. Where the Article 14 petitioners argued that the Act included too few, admitting six faiths and not others, the Assamese movement argued that the Act included too many, regularising entrants whom Assam’s own settlement had committed the state to detect and expel. That settlement is the Assam Accord, signed on 15 August 1985 between the Government of India and the leaders of the Assam Movement, the All Assam Students’ Union and the All Assam Gana Sangram Parishad, after a six-year agitation over the names of foreign nationals on Assam’s electoral rolls. The Accord fixed 1 January 1966 as its base date, provided that entrants between 1966 and 24 March 1971 would be detected, struck off the rolls for ten years and then restored, and required that entrants who came to Assam on or after 25 March 1971 be detected and expelled. Parliament wrote the Accord into the Citizenship Act as Section 6A through the Citizenship (Amendment) Act 1985, making Assam the only state in India with its own statutory citizenship cutoffs, a bespoke provision inside a national Act. The provision’s text and the Accord it codifies are the ground on which the Assam argument stands.

How did Assam read the Act against the 1985 Accord?

Assam’s movement organisations read the 2019 Act as a breach of the Assam Accord because the Accord and Section 6A fixed 25 March 1971 as the date after which entrants were to be detected and expelled, while the 2019 Act regularised entrants from Bangladesh who arrived as late as 31 December 2014 under a religious classification the Accord never contemplated.

The collision is exact and can be stated arithmetically. A Hindu entrant from Bangladesh who reached Assam in 1990 was, under Section 6A, a foreigner liable to detection and expulsion, because the entry fell after 25 March 1971. Under the 2019 Act’s proviso, the same entrant is not an illegal migrant at all, is shielded from proceedings under the Foreigners Act 1946, and may be granted citizenship under Section 6B backdated to the 1990 entry. The two provisions cannot both operate on the same person, and the later statute prevails as law, which is why the All Assam Students’ Union, through its president Dipanka Kumar Nath and its adviser Samujjal Bhattacharya, described the Act in December 2019 as an imposition that nullified the Accord’s central bargain and threatened the Assamese language and political weight in the state. The Assamese objection was ethnic and linguistic before it was religious: the movement of 1979 to 1985 had mobilised against Bengali-speaking migration irrespective of the migrants’ faith, and its heirs opposed a statute that welcomed Hindu Bengali entrants while excluding Muslim Bengali entrants, because the settlement Assam had won counted both as foreigners after 1971. The Sixth Schedule and Inner Line carve-outs in the 2019 Act protected the tribal belts and the hill states around Assam, but the Brahmaputra valley, where the Accord’s politics lived, stood fully inside the Act’s operation, a distribution of protection that Assamese critics read as the government safeguarding every settlement except theirs.

The National Register of Citizens supplied the collision’s second term. Assam’s register, updated under Supreme Court supervision in a process running from 2013, published its final list on 31 August 2019, three months before the Act passed: of 33,027,661 applicants, 31,121,004 were included and 1,906,657 were left out. The Bharatiya Janata Party’s national leadership had criticised the register’s outcome for excluding large numbers of Hindus, and the 2019 Act arrived, in Assamese reading, as the corrective that sorted the excluded by faith: the excluded Hindu regained a pathway through Section 6B while the excluded Muslim faced foreigners tribunals with no statutory shelter. The government’s stated position, given by Amit Shah in Parliament and by Assam’s finance minister Himanta Biswa Sarma in Guwahati, was that the Act and the register did different work, the register being an Assam-specific administrative record under the Accord framework and the Act a national humanitarian measure, and that the numbers of Hindu Bengalis affected in Assam were far smaller than movement rhetoric claimed. In the post-2019 constitutional-amendment context of the same parliamentary season, the government had shown it would use large majorities to revise inherited settlements it judged exhausted; Assam’s protesters, who shut the state down in December 2019 and lost five young men to police firing in the disturbances of that month, judged that the Accord had been revised over their objection without the consultation the Accord’s moral authority demanded. The architectural finding for Assam is that the Act created a conflict between two citizenship settlements housed in the same parent statute, the 1985 territorial settlement in Section 6A and the 2019 religious settlement in Section 6B, and that Parliament resolved the conflict by silence, enacting the second without amending the first, leaving the courts to state the relation between them.

Petitions Without a Verdict: The Court’s Procedural Record

The architectural question was placed before the Supreme Court within days of the Act’s passage, and the Court’s handling of it is a study in institutional deferral that the architectural reading must record as faithfully as the arguments themselves. The first petitions arrived in December 2019, filed by political parties, civil liberties organisations, religious bodies, student groups and individual citizens, among them the Indian Union Muslim League, the Congress leader Jairam Ramesh, the Trinamool Congress member Mahua Moitra, the All Assam Students’ Union and the Citizenship Rights in Assam organisation, together with retired officials and academics. On 18 December 2019 a bench headed by Chief Justice S. A. Bobde issued notice to the Union of India on a first batch of fifty-nine petitions and declined to stay the Act, which was then awaiting its commencement notification. The Act entered into force on 10 January 2020. On 22 January 2020 the Court, hearing more than one hundred and forty petitions, gave the Union four weeks to file its reply, declined interim relief again, and indicated that the challenges might require a Constitution Bench. Petitions continued to accumulate, passing two hundred in total, and the Union filed affidavits defending the classification on the minority-in-theocratic-states ground and contesting the maintainability of a basic-structure challenge to ordinary legislation.

There the merits have rested. The petitions were listed, adjourned and re-listed across 2020, 2021 and the years following, through changes of Chief Justice and the disruption of the pandemic, without a final hearing on the constitutional questions and without any interim order suspending the Act. The street contest that formed the mobilisation reading’s subject had its own compressed chronology, which belongs in this article only as the petitions’ context: protests against the Act spread from Assam in early December 2019 to university campuses and city centres across India, the sit-in at Shaheen Bagh in Delhi began on 15 December 2019 and became the movement’s best-known site, violence in northeast Delhi in February 2020 killed more than fifty people in rioting that both sides attributed to the other’s mobilisation, and the Shaheen Bagh site was cleared on 24 March 2020 as the national COVID-19 lockdown began. This article does not retell that winter; the protest narrative, its organisation and its political consequences are the subject of the companion reading in Pillar 5, and the architectural record is complete for present purposes once the sequence from statute to petition to procedural pause is established.

The deferral itself carries an architectural meaning that the complication section of the brief rightly insists on facing. A constitutional question can be fully argued in the country’s public sphere and fully pleaded in its highest court while remaining institutionally undecided, and during the undecided interval the statute is not in suspense: it is law, capable of being operationalised, as the 2024 Rules demonstrated. The petitioners’ strongest institutional argument against deferral was that citizenship granted under Section 6B is practically irreversible, since a grant backdated to entry creates a citizen whose status a later judgment striking the Act would disturb only at enormous human and legal cost, so that each year of deferral converts a contested classification into settled demography. The Union’s strongest institutional answer was that deferral is the ordinary condition of constitutional litigation, that no grant under the Act occurred at all until the Rules were notified in March 2024, so no fait accompli accumulated during the years of pendency, and that the Court’s refusal of a stay, twice recorded, reflects the presumption of constitutionality rather than indifference. Both arguments are honestly available on the record, and the architectural reading reports the standoff rather than resolving it: the Indian system permits a membership statute to be enacted in a week in December, contested in the streets for a season, pleaded in court for years and operationalised by rule-making before any judge has pronounced on its validity. That sequence is a fact about the architecture of review, and it conditions what the eventual judgment, whenever delivered, can practically decide.

The Longer View: Citizenship Scholarship and the Founding Refusal

The architectural question did not begin in 2019, and the scholarship the brief makes mandatory reads the Act against a seventy-year argument about what Indian citizenship is. Niraja Gopal Jayal’s Citizenship and Its Discontents: An Indian History, published by Harvard University Press in 2013, remains the field’s organising work. Jayal traced Indian citizenship as a contested inheritance shaped by three inheritances that never fully reconciled: the colonial subjecthood the republic inherited, in which membership was a status granted from above; the nationalist movement’s promise of equal citizenship as the moral core of independence; and the Partition settlement, which sorted millions by territory under emergency conditions and left faith entangled with belonging in administrative practice even where the constitutional text refused the entanglement. Her central finding, developed across chapters on the founding debates, on naturalisation practice and on the Assam question, is that Indian citizenship has always carried a tension between a jus soli imagination, citizenship by birth on the soil, and a jus sanguinis pull, citizenship by descent and community, and that the statutory history from 1955 onward records the jus soli element being progressively narrowed, in the amendments of 1986, 2003 and 2015, each of which tightened birth-based citizenship for children of persons outside the law’s protection. Read through Jayal’s frame, the 2019 Act is not an aberration from the statutory trajectory but its boldest step: the amendments before it narrowed by descent and documentation, while the 2019 amendment sorted by faith expressly, converting a drift in administrative practice into a stated legislative criterion. Her frame also explains the intensity of the scholarly reaction, since a criterion long present as subtext had been promoted to text, where constitutional doctrine could reach it.

Pratap Bhanu Mehta’s post-2019 essays in the Economic and Political Weekly carried the argument from the scholarly register into the constitutional one. Mehta contended that the Act, read with the prospect of a national register of citizens, altered the basis on which the Indian state presumes belonging: the citizen under suspicion of foreignness would have to prove membership, and the statute’s religious line told one community that its proof would never be assisted by the shelter the law extended to others. The architectural reading extracts from Mehta’s essays the element that survives translation into doctrine: his claim is at bottom a claim about presumptions, about which side of the line the burden of proof and the benefit of doubt occupy, and that claim connects to the statute’s actual machinery, the abatement of proceedings and the backdating of citizenship for the covered class against the foreigners tribunal process for everyone else. Sandipto Dasgupta’s post-2019 essays on the constitutional-architectural question supplied the third scholarly register, a legal-historical one. Dasgupta situated the Act within the long career of the category of the illegal migrant and the executive’s exemption power, arguing that the republic’s citizenship architecture was built as much through notifications, exemptions and tribunal practice as through the 1955 Act’s celebrated text, and that the 2019 amendment’s novelty lay in Parliament claiming for the religious classification a democratic warrant that executive practice had never asserted. The three scholars disagree in emphasis and in politics, and the architectural reading does not conscript them into one camp. Read together, they establish the scholarly disagreement this article was asked to adjudicate in its precise form: whether the Act is best understood as a political act with constitutional side effects, the mobilisation reading’s premise, or as a constitutional act with political side effects, which is the premise the evidence of text, doctrine and rule-making assembled here supports.

The founding record anchors that adjudication. When the Constituent Assembly took up the citizenship articles in August 1949, members pressed amendments that would have written religious and racial qualifications into the founding code, and the Drafting Committee’s leadership, with Jawaharlal Nehru and Vallabhbhai Patel speaking for the settlement, refused them, in a chamber still shadowed by Partition’s trains and camps. The articles the Assembly adopted sorted the founding citizenry by domicile, birth, parentage and migration dates, never by faith. Parliament in 1955 legislated the same silence into the permanent code, and for sixty-four years, through amendments that narrowed birthright and hardened the illegal-migrant bar, the silence held: faith decided nothing in the text of Indian citizenship law. The defenders of the 2019 Act answer that the silence was a luxury of a settlement whose Partition wounds the Act finally treats, and that the founding generation’s own exemption practice toward Partition refugees shows the silence was never as complete in administration as in text. The critics answer that a silence maintained through six decades of amendment is not an oversight but the settlement itself. The architectural reading sides with the critics on the historical fact, the silence was real, textual and repeatedly renewed, while crediting the defenders’ observation that administrative practice under the Foreigners Act 1946 and the exemption notifications of 2015 and 2016 had already built, below the text, the six-faith and three-country scheme the Act adopted. The Act’s architectural significance is exactly that promotion, from notification to statute, from practice to principle, and it is why the debate about it cannot be closed by observing that little changed on the ground in 2020: the ground of principle had already changed in December 2019.

The 2024 Rules: Architecture Becomes Procedure

For four years after its commencement on 10 January 2020, the Citizenship (Amendment) Act existed as law without machinery, because Section 6B grants depend on rules prescribing the authority, conditions and manner of application, and no rules came. The interval itself became an argument: the government’s critics read the delay as evidence that the Act had been passed for its signalling value in the winter of 2019, while the government attributed the pause to the pandemic and to the care required in designing a process for applicants whose documentary lives were irregular by definition. On 11 March 2024 the Ministry of Home Affairs notified the Citizenship (Amendment) Rules 2024, and the Act acquired a working form five weeks before the first phase of the 2024 general election. The Rules are the brief’s under-cited primary source for good reason: commentary treated them as an election-season announcement, while their annexures and schedules quietly settled how each of the four architectural features would operate on actual applicants.

The machinery the Rules built is two-tiered. An applicant applies online through the citizenship portal, filing the application with a set of prescribed documents and an affidavit, and the application passes to a district level committee, chaired by a senior postal officer as designated officer, which verifies the documents and administers the oath of allegiance. Decisions rest with an empowered committee for each state or union territory, headed by the Director of Census Operations, which scrutinises the verified application and grants or refuses citizenship, whereupon a certificate issues and the successful applicant stands deemed a citizen from the date of entry recorded in the file. Lawyers and community organisations preparing applicants for the process assembled dossiers, scanned affidavits and hearing presentations, frequently opening hearing decks through an online presentation viewer to walk applicants and counsel through the committee stages before filing. The documentary schedules are where the four features acquire their operational teeth. The religion criterion is proved through an eligibility certificate issued by a locally reputed community institution, attesting that the applicant belongs to one of the six named faiths. The country criterion is proved through documents issued by the authorities of Afghanistan, Bangladesh or Pakistan, from a prescribed list that includes passports, birth certificates, school records and land documents. The cutoff is proved through any of twenty listed documents showing entry into India on or before 31 December 2014, a list drawn wide enough to include rent agreements, ration cards and school certificates from Indian institutions, in recognition that the covered class entered irregularly and documented its presence informally. The correspondence between the Ministry of Home Affairs and state governments in the weeks after notification, over the constitution of the empowered committees and the staffing of the district committees, forms part of the operational record preserved in the Ministry’s archives, and it shows several opposition-ruled state governments recording objections while the committee structure proceeded on central nominees, a federal friction the Rules’ text anticipates by placing decision authority in committees the centre constitutes.

Two features of the operational record bear directly on the architectural argument. The first is the scale the machinery implied. The application window opened to a class defined by entry before the end of 2014, a class whose size no official figure has ever fixed, and the portal’s early operation in March and April 2024 processed applications in the hundreds in its first reported grants, with the first certificates issued in May 2024 to applicants in Delhi. Whatever the eventual totals, the operational Act addresses a finite historical population rather than a continuing flow, which confirms the cutoff feature’s defence that Parliament legislated for a closed class, while leaving untouched the critics’ objection that a closed class drawn by religion is still drawn by religion. The second feature is the Rules’ treatment of the pending litigation. The Rules contain no clause suspending their operation pending the Supreme Court’s decision, and when petitioners moved the Court in March 2024 to stay the Rules, a bench headed by Chief Justice D. Y. Chandrachud on 19 March 2024 declined an interim stay while issuing notice and recording the Union’s undertaking to file its response, so the operational and judicial tracks proceeded in parallel, grants issuing under a statute whose validity remained unadjudicated. The architectural question thus acquired its final form: not whether the classification could be defended in the abstract, but whether a fully working membership machinery, sorting by faith at district counters across India, is the kind of state action the founding settlement’s citizenship architecture was built to permit or to prevent.

The Unresolved Court and the Settled Text

The strongest counter to the architectural-substantive reading, stated in the brief and faced here directly, is institutional: a question that the Supreme Court has not answered in more than four years of pendency may be a question the constitutional system does not in fact treat as substantive, and an analytical frame built on an unadjudicated question risks mistaking the commentators’ agenda for the Constitution’s. The counter deserves its full weight. The Court declined a stay in December 2019, declined again in January 2020 and declined to halt the Rules in March 2024. No bench has held the Act unconstitutional in any part, no bench has held it valid on the merits, and the doctrines the petitions invoke, the nexus limb of the classification test, manifest arbitrariness, secularism as basic structure applied to legislation, remain unapplied to the statute that was designed to test them. If substance is measured by decided law, the CAA debate’s architectural content is a set of arguments awaiting a decision that may, when it comes, resolve the matter on a narrow ground that leaves the large questions open, as Indian courts have often done with membership statutes before.

The answer the architectural reading gives is that the question’s substance does not wait on its resolution, for three reasons grounded in the record assembled above. First, the classification operated in the world before any court pronounced on it, through the exemption notifications of 2015 and 2016, through the Act’s commencement in 2020 and through grants under the Rules from 2024, so the architecture the debate concerns is not a proposal but a functioning arrangement whose features can be mapped, as the four-feature map maps them, without predicting any judgment. Second, the debate disciplined actors in ways that are themselves constitutional facts: a Joint Parliamentary Committee heard the under-inclusion objection and answered it on record, two houses of Parliament divided on the classification by recorded votes of 311 to 80 and 125 to 105, more than two hundred petitions converted the parliamentary dissent into pleadable doctrine, and a state movement in Assam forced the federal collision between Section 6A and Section 6B into the open. A question that organises the behaviour of committees, legislatures, movements and litigants over a decade is substantive in the only sense that matters to political analysis, whether or not a court has yet dignified it with an answer. Third, the question outlives the litigation’s outcome in either direction. A judgment upholding the Act would settle the classification’s validity while leaving the architectural precedent, that religion may serve as a citizenship criterion where the nexus to a theocratic-neighbour rationale holds, available for every future membership statute. A judgment striking the Act would settle the precedent the other way. Either way the debate of 2019 to 2024 is the moment the Indian republic first argued, at full constitutional depth, whether faith may draw the boundary of belonging, and that is a fact about the republic’s architecture that no disposition of the petitions can retroactively alter.

Set beside the two rival framings, the architectural reading’s claim to priority is modest and exact. The constitutive-opposition reading explains why the winter of 2019 and 2020 took the political form it did, a government pressing a founding-settlement revision and an opposition discovering in the resistance a unifying grammar, and nothing in this article displaces that explanation on its own ground. The political-mobilisation reading explains the protests’ organisation, symbols and electoral afterlife. The architectural reading explains what the contest was about at the level where its result cannot be reversed by the next election: a four-feature statutory design that sorts applicants for membership by religion, country and date, defended as relief for the persecuted of three theocratic states, attacked as a breach of equality and of the secular boundary of the political community, complicated in Assam by an older territorial settlement the same statute book houses, and operationalised in 2024 through committees and schedules that translated the design into district-level practice. Elections will revise governments, movements will rise and disperse, and the petitions will one day be decided. The question the Act posed to the architecture, whether religion can be a citizenship criterion in the Indian constitutional order, was posed in December 2019 with a completeness that the founding generation avoided and that no subsequent politics can unpose. That is the debate this article set out to explain, and it is why the architectural-substantive reading, distinct from and complementary to the readings built on mobilisation and opposition, is the reading the statute itself demands.

The Parent Act Transformed: Citizenship Law from 1955 to 2019

The architectural reading gains depth from the amendment history of the parent Act, because the 2019 amendment was the fifth major revision of a code whose direction of travel had been consistent for three decades. The Citizenship Act 1955 as enacted was among the most inclusive citizenship statutes of its era: citizenship by birth under Section 3 attached to birth on Indian soil with narrow exceptions, reflecting the founding generation’s jus soli confidence that the republic’s membership should begin from the territory rather than from blood or belief. The founding provisions the Constitution itself supplied, Articles 5 to 11, had sorted the initial citizenry through the emergency of Partition, a settlement whose territorial and migratory logic is traced in the founding settlement of the three tensions and in the account of Partition’s constitutional consequences. Parliament’s 1955 code preserved that territorial logic for settled times.

The amendments then narrowed the code step by step, each step answering a political pressure that the architectural record should name. The Citizenship (Amendment) Act 1986, passed in the shadow of the Assam agitation and the 1985 Accord, ended pure jus soli: a person born in India after its commencement was a citizen by birth only if either parent was an Indian citizen at the time of the birth. The Citizenship (Amendment) Act 2003, enacted under the government of Atal Bihari Vajpayee, hardened the structure further. It restricted birth citizenship again, so that a person born in India on or after 3 December 2004 is a citizen only if both parents are citizens or one parent is a citizen and the other is not an illegal migrant; it inserted the illegal-migrant definition into Section 2(1)(b) and attached the bar on registration and naturalisation that the 2019 proviso would later lift for the covered class; it created the Overseas Citizen of India category; and it inserted Section 14A, empowering the central government to compulsorily register citizens and maintain a National Register of Indian Citizens, the statutory seed from which the later NRC debate grew. The Citizenship (Amendment) Act 2015 merged the Person of Indian Origin card into the OCI scheme. Across the 1986, 2003 and 2015 revisions, the criteria Parliament adjusted were descent, documentation and residence. At no point in that thirty-year narrowing did Parliament write religion into the eligibility text, which is the precise sense in which scholars describe 2019 as a change of kind rather than of degree: the direction of travel, toward a more guarded membership, was continuous, while the instrument, an express religious classification, was new to the statute book.

That history also disciplines the two partisan mythologies that surround the Act. Against the claim that 2019 ruptured an otherwise generous settlement, the record shows a settlement already narrowed by the governments of Rajiv Gandhi in 1986 and Atal Bihari Vajpayee in 2003, with the illegal-migrant bar, the very obstacle the 2019 Act relieved for six communities, itself a product of the earlier narrowing. Against the claim that 2019 merely continued business as usual, the record shows that every prior amendment sorted entrants by facts about their birth, parentage, papers and years of residence, facts an applicant of any faith could in principle supply, while the 2019 proviso sorted by a fact about the applicant’s faith itself, which an excluded applicant could not alter by any length of residence or completeness of papers. The architectural question lives in that difference, and it is visible only when the amendment history is laid out in sequence, which is why the petitions, the committee dissents and the scholarly essays all rehearse the sequence and why this article has done the same.

Two Days in December: The Parliamentary Record

Parliament debated the Bill across two sittings that together constitute the fullest official statement of both cases, and the architectural reading treats the debates as primary evidence because the classification doctrine makes legislative purpose a legal fact. The Citizenship (Amendment) Bill 2019 was introduced in the Lok Sabha on 9 December 2019 by Home Minister Amit Shah. Shah’s opening speech framed the Bill as the completion of a Partition obligation: the constitutions of Pakistan, Afghanistan and Bangladesh establish Islam as the state religion, minorities in those states faced persecution that Indian governments had acknowledged since 1950, the Nehru-Liaquat Pact of 8 April 1950 had failed to protect them, and the Bill would grant citizenship to refugees who had lived in India for years in a condition of suspended dignity. He stated the government’s limiting assurances in the same speech: the Act concerned foreigners alone, it took away the citizenship of no Indian of any faith, and apprehensions connecting the Act to Indian Muslims rested on a confusion the government disclaimed. Congress member Manish Tewari answered with the constitutional case in its parliamentary form, arguing that the Bill offended Article 14 by classifying on religion, that it violated the basic structure, and that a secular republic could not distribute citizenship by faith without altering its own character. Trinamool Congress member Mahua Moitra pressed the nexus objection, asking why a persecution rationale excluded the persecuted of Sri Lanka and Myanmar and the Ahmadiyyas of Pakistan. Asaduddin Owaisi of the All India Majlis-e-Ittehadul Muslimeen opposed the Bill as a measure directed against Indian Muslims and tore a copy of it in the House in protest. The Lok Sabha passed the Bill late that night by 311 votes to 80, with the Janata Dal (United), the Shiromani Akali Dal and the Shiv Sena supporting the government in that House.

The Rajya Sabha took the Bill on 11 December 2019, and the debate there refined the legal arguments because the upper house’s arithmetic forced the government to hold allies whose support was publicly conditional. Law Minister Ravi Shankar Prasad defended the classification on the theocratic-states ground and on the stage-by-stage doctrine, arguing that Parliament addressed the persecution it could document in states with an established religion and remained free to address other cases separately. Congress members Anand Sharma and Kapil Sibal argued the equality and basic-structure cases, Sibal contending that a two-nation logic, rejected at the founding, was being read back into citizenship law. Derek O’Brien of the Trinamool Congress attacked the Bill’s drafting and its Assam consequences. The Shiv Sena, which had supported the Bill in the Lok Sabha while demanding clarifications, was absent at the Rajya Sabha division, and the Bill passed by 125 votes to 105, a margin narrow enough to record that the classification commanded the House without commanding its consensus. Presidential assent followed on 12 December 2019.

The recorded divisions matter architecturally for a reason beyond their drama. When the Supreme Court eventually hears the classification challenge, the parliamentary record will be the principal evidence of legislative purpose against which the nexus limb is tested, and the December debates supply both sides with their materials: the government can point to a purpose stated consistently from the Statement of Objects and Reasons through both debates, the relief of religious minorities in three named theocratic states, while the petitioners can point to the same record to show that the under-inclusion objection was placed before Parliament, by named members in both houses, and answered by assertion rather than by evidence distinguishing the omitted persecuted from the included. A legislature that has heard the precise constitutional objection and proceeded anyway has made a deliberate architectural choice, and the debates fix that choice in a form no later affidavit can soften. The committee stage tells the same story in miniature: the Joint Parliamentary Committee on the 2016 Bill had received the Assam objection and the equality objection in testimony, and its majority report of 7 January 2019 carried the carve-outs as the answer to the first while leaving the second to the plenary, where the December divisions answered it by majority. Purpose, objection and answer are all on the record, which is what makes the CAA debate an architectural debate conducted with unusual completeness, whatever judgment the record finally receives.

The Persecution Question: Evidence, Demography and a Foreign Precedent

Because the classification’s defence rests on persecution in the three named countries, the state of the persecution evidence is part of the architecture, and the architectural reading reports it at the strength each side actually marshalled. The government’s evidence in the committee stage and the debates was demographic and testimonial. The demographic claim is that the Hindu share of population in the territories of the named states collapsed across the republic’s lifetime: Hindus formed roughly a fifth of East Pakistan’s population at the 1951 census and under a tenth of Bangladesh’s population at the 2011 census, while Pakistan’s Hindu minority remained a small and shrinking fraction of its population across the same decades. Successive Indian governments cited that trajectory, and the Joint Parliamentary Committee heard testimony from refugee delegations describing forced conversion, temple desecration, land seizure under enemy-property and vested-property regimes in Bangladesh, and the exodus of Afghan Sikhs and Hindus from a community once counted in the tens of thousands to a remnant of a few hundred families by 2019. The testimonial record before the committee included delegations from the very settlements the Act would serve, Partition refugees and their descendants in Rajasthan, Gujarat, Delhi and Assam, whose irregular status had persisted across generations.

The critics’ evidentiary answer operated at a different level, contesting not the existence of persecution but its distribution. Human rights documentation placed before the public debate recorded that persecution in the named countries also fell heavily on communities the Act omits: Ahmadiyyas in Pakistan, declared non-Muslim by the Second Amendment to Pakistan’s constitution in 1974 and prosecuted under Ordinance XX of 1984, yet excluded from the Indian Act’s shelter because the Act reads them by the faith they profess; Hazara and Shia communities in Afghanistan and Pakistan, targeted by sectarian violence across the same decades; and Hindu Tamils in Sri Lanka and Rohingya Muslims in Myanmar, whose persecution occurred in states whose constitutions do not establish Islam and which the Act therefore leaves outside its map altogether. On the critics’ reading of the census evidence itself, Bangladesh’s declining Hindu share reflects migration driven by a compound of discrimination, land pressure and economic opportunity rather than persecution alone, a compound the Act’s binary classification cannot express. The architectural significance of the evidentiary contest is that the two sides were not disagreeing about whether the classification had a factual basis; they were disagreeing about the unit at which persecution should be measured, the faith of the victim taken one person at a time, or the constitutional character of the persecuting state taken as a whole. The classification doctrine, which asks for a rational relation between differentia and objective, does not itself choose between those units, and the choice between them is the substantive political judgment the Act embodies, dressed in the doctrine’s vocabulary.

One foreign precedent, invoked in the Indian debate by the Act’s defenders, deserves record because it shows that category-based relief is not an Indian eccentricity. The United States’ Lautenberg Amendment of 1989, named for Senator Frank Lautenberg, directed American refugee admissions to presume a well-founded fear of persecution for named categories from the Soviet Union and later Iran, including Jews, Evangelical Christians and members of particular churches, lowering the evidentiary burden for applicants inside the named categories while leaving others to prove their cases individually. The amendment has been renewed across decades and applied through administrations of both American parties. Defenders of the Indian Act cited the Lautenberg model to argue that democracies routinely relieve persecution by category rather than by individual proof, and that India’s classification performs the same operation for its neighbourhood. The critics’ distinction, also placed on record in the essays of December 2019, is that refugee admission and citizenship differ in kind: the Lautenberg categories govern admission to a protection process in a country of asylum, while the Indian Act distributes membership itself, the constitutive status from which all other rights in the republic flow, and a criterion tolerated at the gate of protection may carry a different constitutional weight at the gate of belonging. The architectural reading records the precedent and the distinction together, because the comparison genuinely cuts both ways and any adjudication that suppresses one direction of the cut is advocacy rather than analysis.

The Federal Contest: Assemblies, Resolutions and the Union List

Citizenship in the Indian Constitution belongs to the Union: Entry 17 of the Union List assigns citizenship, naturalisation and aliens to Parliament, and Part II commits the subject to parliamentary law. The 2019 Act therefore required no state’s consent, and no state possessed a legal veto over its operation. The federal contest that followed the Act’s passage was consequently fought in the register of political refusal rather than legal competence, and its record forms part of the architectural story because it tested how a membership statute operates in a federation whose constituent units dispute its premise. The Kerala Legislative Assembly moved first, adopting on 31 December 2019 a resolution demanding the Act’s repeal, carried with support from both the ruling Left Democratic Front and the opposition United Democratic Front against the single Bharatiya Janata Party member. The Punjab Assembly adopted a similar resolution on 17 January 2020, followed by the Rajasthan Assembly on 25 January 2020 and the West Bengal Assembly on 27 January 2020, with Chhattisgarh and other opposition-ruled states adopting comparable positions in assembly resolutions or cabinet decisions through the same season. Chief ministers including Pinarayi Vijayan of Kerala, Amarinder Singh of Punjab and Mamata Banerjee of West Bengal stated that their governments would not cooperate in implementing the Act or any associated register exercise.

The Union’s answer rested on the constitutional text the architectural reading has already noted: citizenship is a Union subject, state assemblies may resolve as they wish, and the Act’s machinery, when it arrived in 2024, was deliberately designed to run through committees constituted under central authority, with the district level committees chaired by officers of the postal department and the empowered committees headed by census officials, so that the grant of citizenship under Section 6B does not depend on the cooperation of a state government that has resolved against the Act. Several state governments also carried the contest to the Supreme Court under Article 131, Kerala filing an original suit against the Union in January 2020 that asked the Court to declare the Act unconstitutional as a violation of equality and secularism, a procedural route distinct from the Article 32 petitions and one that placed the federal objection inside the litigation record alongside the individual and organisational challenges. In the Northeast the federal contest took the different form already described: not refusal but exemption, with the Meghalaya cabinet resolving in December 2019 to demand the Inner Line Permit for the whole state, Nagaland already standing inside the permit regime, and Manipur brought within it by the presidential order of December 2019, so that the map of the Act’s application in the region was redrawn state by state in the same weeks as its passage. The federal record thus exhibits the Act operating through three distinct state postures, compliance in states ruled by the governing party and its allies, symbolic refusal in opposition-ruled states without a legal instrument to give the refusal effect, and negotiated exemption in the Northeast where older protective settlements supplied such an instrument. A citizenship statute that produces three federal postures is exercising the federation as well as the citizenry, and the architectural reading counts that exercise among the Act’s substantive features rather than among its political side effects.

Counting the Covered Class: Numbers the Debate Never Fixed

One quantity the entire debate conspicuously failed to fix is the size of the class the Act covers, and the absence is architecturally informative. The government offered no official estimate of the persons eligible under the proviso in the December 2019 debates, in the committee record or in its affidavits, and Home Minister Amit Shah’s parliamentary case proceeded without a number, while his election-season speeches in West Bengal and Assam in 2019 had spoken of beneficiaries in far larger terms that the parliamentary record never adopted. Estimates in public commentary ranged from the tens of thousands of documented long-stay refugees in recognised settlements to figures in the millions derived from census residuals in Assam and West Bengal, a range so wide that it functioned less as information than as a measure of the two sides’ intuitions: a small class makes the Act a modest humanitarian instrument whose constitutional price seems disproportionate to its defenders’ critics, while a large class makes it a demographic intervention whose modesty its defenders cannot claim. The Intelligence Bureau and the Registrar General published no count, the 2024 Rules were notified without an accompanying estimate, and the first reported grants, certificates handed to fourteen applicants in Delhi by Union Home Secretary Ajay Kumar Bhalla on 15 May 2024, gave the operational record a deliberately modest opening figure. The architectural reading draws the honest conclusion from the missing number: Parliament enacted a membership classification of the first constitutional importance without placing on record the size of the population it classified, and both the defence of the Act as narrow relief and the attack on it as demographic engineering rest partly on a quantity neither side established. That gap belongs in any complete account of the debate, because it marks the point at which the architectural argument and the political argument shared a common evidential failure, each proceeding as though the other’s missing number were known.

The missing number had a prehistory in the committee room, where the question was asked in terms and answered in categories. When members of the Joint Parliamentary Committee pressed officials for the expected scale of the beneficiary class, the record shows the government answering by describing the settlements and camps where long-stay refugees lived rather than by tabulating them, and the committee’s report of 7 January 2019 carried the description forward without converting it into a figure. The pattern repeated in the plenary: Manish Tewari asked in the Lok Sabha how many persons the Bill would benefit, and the government’s reply returned to the persecution narrative rather than to a count. A legislature that classifies by faith for the first time in the republic’s history, and that declines across committee, debate and affidavit to state the size of the class it is creating, has left the proportionality inquiry that any future court must conduct without the denominator that inquiry requires. The petitioners noticed the gap and pleaded it, arguing that a classification whose burden and benefit cannot be quantified cannot be shown to bear the rational relation the doctrine demands, while the government treated the demand for a number as a category error, since a remedial class defined by persecution and date is complete as a legal description whether it holds ten thousand persons or ten lakh. The architectural reading closes the point where it closes the others: the missing denominator is not a debating detail but a structural feature of the Act’s passage, one more instance of the December 2019 pattern in which the decisive constitutional materials, purpose, objection, evidence and scale, were all placed half on the record and half off it, and the debate’s participants argued across the seam.

The Exemption Decade: How the Executive Built the Class Before Parliament Did

Between the parent Act’s text and the 2019 amendment stands a decade of executive action that the architectural reading must reconstruct in detail, because both the defence and the attack draw their strongest inferences from it. The machinery of toleration was built from two colonial-era and early-republic statutes: the Passport (Entry into India) Act 1920, which penalises entry without a passport, and the Foreigners Act 1946, which empowers the central government to regulate, restrict and deport foreigners and under which foreigners tribunals in Assam determine status. Long-stay refugees from Pakistan and Bangladesh lived under those statutes in a condition of renewable precarity, their presence tolerated through periodic exemptions and their applications for long-term visas handled case by case by the Ministry of Home Affairs. In September 2015 the Ministry issued notifications under both statutes exempting Hindus, Sikhs, Buddhists, Jains, Parsis and Christians from Bangladesh and Pakistan who had entered India on or before 31 December 2014 from the penal consequences of irregular entry, and a further notification in July 2016 extended the exemption framework to Afghanistan and simplified the regularisation of the same class. The six faiths, the three countries and the 2014 cutoff that Parliament enacted in 2019 are all present in the 2015 and 2016 notifications, four years before the statute, in instruments that no legislature debated and no court reviewed.

The government read that genealogy as continuity and therefore as reassurance: the Act changed the legal quality of a toleration the executive had already extended, converting precarious exemption into a citizenship pathway for a class the state had long recognised in practice. On that reading the classification’s constitutional novelty is small, because the state had sorted this population by faith for administrative purposes since 2015 without judicial objection. The critics read the same genealogy in the opposite direction and with equal documentary support. An exemption under the Foreigners Act is an act of executive grace that confers no right and distributes no status; it tolerates a presence while reserving every sovereign claim. A statutory citizenship pathway is an act of the legislature that confers the republic’s constitutive status and engages the equality guarantee at the point where status is distributed. The step from grace to right, the critics argued in the December 2019 essays and in the petitions, is precisely the step at which Article 14 acquires jurisdiction over the criterion, so the exemption decade does not launder the classification but explains its origin, in an administrative practice that grew up below the waterline of constitutional scrutiny and was then promoted, untested, into the statute book. The architectural reading finds the critics’ account of the legal difference persuasive and the government’s account of the political continuity accurate, and notes that the two are compatible: a practice can be continuous as administration and discontinuous as law, and the debate’s heat arose because each side insisted on describing only the continuity or only the discontinuity. The finding matters for the doctrine as well, because a court testing the classification’s nexus will encounter the exemption record as the state’s own prior statement of the class’s boundaries, boundaries drawn by the same criteria the statute adopted, which the government will cite as deliberation and the petitioners as path dependence. Either way, the exemption decade is not background to the Act; it is the Act’s first draft, written by the executive in the only register the executive commands, and then transcribed by Parliament into the register that courts can review.

Severability and the Shape of the Doctrinal Question

Because the Act is built from separable features, the doctrinal question before the Court is not a single question, and the architectural reading closes its analysis by separating the outcomes the doctrine makes available, as an exercise in mapping the architecture rather than forecasting a result. The first possibility is full validation: the Court accepts the government’s statement of objective, applies the Dalmia presumptions, holds the basic-structure challenge to legislation inadmissible or unmade out, and the four features stand together, establishing that a religious classification in citizenship law survives where the state ties it to the constitutional character of the countries of origin. The second possibility is invalidation of the criterion with severance of the machinery: the Court holds the religion-based proviso to Section 2(1)(b) unconstitutional while leaving the rest of the 1955 Act untouched, a disposition the severability doctrine, stated in R. M. D. Chamarbaugwalla v. Union of India in 1957, makes formally straightforward since the proviso is a self-contained insertion and the parent Act functioned without it for six decades. The third possibility is a reading that preserves the text while altering its operation, of the kind Indian courts have used to save statutes by construction, for example by treating the six-faith list as illustrative of a persecution class rather than exhaustive of it, a construction the government’s own counsel would have to accept or resist on the record of the December debates, where the list was defended as deliberate and exhaustive. The fourth possibility, the one the deferral years have made familiar, is continued pendency across further changes in the political cycle, during which the Rules operate and the class the Act created acquires the settled expectations that any eventual invalidation would have to unwind.

Each outcome distributes consequences differently across the architecture, and the differences are the substance the mobilisation reading cannot capture. Full validation constitutionalises a technique: future Parliaments would hold a judicially approved template for faith-conscious membership legislation tied to foreign theocracies, and the founding silence on religion in citizenship law would stand replaced by a doctrine of permitted exception. Invalidation restores the silence as a judicially enforced rule and confines the exemption technique to the executive register, where it confers toleration without status. A saving construction splits the difference in the manner most characteristic of the Indian courts, preserving legislative supremacy in form while relocating the classification’s real boundary from the statute to the doctrine that interprets it. Continued pendency, the state of affairs across the whole period this article covers, distributes the consequences to time itself: the architecture operates as enacted while its validity remains an open file, and the political branches learn that a membership statute can be fully operationalised inside the interval between enactment and adjudication. The architectural-substantive reading of the CAA debate is, in the end, the claim that these four dispositions and their consequences are what the debate was always about beneath its slogans, and that a reader who can state them can state the debate, while a reader who can state only the winter’s marches and counterslogans has described the debate’s weather and missed its climate.

A Close Reading of the Proviso: The Words That Carry the Argument

The whole debate finally rests on the wording of one proviso, and the architectural reading owes the reader a slow pass through its language, because several of the largest arguments turn on small words. The proviso to Section 2(1)(b) speaks of any person belonging to Hindu, Sikh, Buddhist, Jain, Parsi or Christian community in Afghanistan, Bangladesh or Pakistan. The legislature’s chosen noun is community, not religion, and its chosen verb is belonging, not professing or practising. Defenders of the Act drew on that wording to argue that the classification is communal and historical rather than theological: the named communities are the historically settled non-Muslim populations of the three states, identified as social facts rather than as creeds, and an applicant qualifies by belonging to such a population, a matter of birth and social identity, rather than by demonstrating belief or observance. Critics answered that the distinction does no work at the point of application, because the 2024 Rules operationalise belonging through an eligibility certificate attesting the applicant’s religion, issued by a community institution, so that the statute’s communal noun is cashed out, in the machinery, as a religious fact certified by a religious body. The gap between the statute’s noun and the rule’s certificate is a small, precise instance of the debate’s largest pattern, in which each side describes the same provision in the register that favours its case, the text as sociology or the machinery as theology, and the architectural reading records both descriptions because a court will have to decide which register governs the classification’s character.

The proviso’s second load-bearing phrase is its condition of exemption: the covered person must be one who has been exempted by the central government under the Passport (Entry into India) Act 1920 or the Foreigners Act 1946 or the provisions of those statutes as applied to the person. The covered class is therefore not self-executing on the face of the citizenship statute; it is defined by reference to an executive act under two other statutes, the exemption notifications of 2015 and 2016, so that the legislature in 2019 classified by incorporating an executive list drawn in 2015. Petitioners used that structure to argue that Parliament had abdicated the classification’s real work to the executive’s prior paperwork, while the government used the same structure to argue deliberateness, a class tested in administration for four years before its promotion into the Act. The third load-bearing element is Section 6B’s deeming clause, under which a person granted a certificate under the section is deemed to be a citizen of India from the date of entry into India. Deeming clauses are the statute book’s instrument for rewriting legal history, and this one rewrites the grantee’s entire Indian past: years lived as an irregular entrant become, in law, years lived as a citizen, with consequences for the status of children born in that interval, for the validity of documents obtained and for any proceedings abated along the way. No other provision of the 1955 Act backdates citizenship to a moment before the state’s own recognition of the person, naturalisation and registration both operating from the date of the certificate, and the clause’s presence in Section 6B is the clearest textual evidence that the legislature understood the covered class as wronged by the interval itself, as persons whose belonging the law was correcting retrospectively rather than conferring prospectively. Critics replied that a retrospective correction distributed by faith compounds rather than cures the classification’s defect, since the Muslim entrant of the same year, on the same road, under the same conditions, receives no correction at all. In that exchange over a deeming clause, conducted in the vocabulary of drafting, the entire constitutional argument of December 2019 reappears in miniature, which is the architectural reading’s final demonstration: in this debate the drafting was the politics, and the politics, at every level from the street to the court, kept returning to the drafting.

Frequently Asked Questions

Q: What is the CAA?

The CAA is the Citizenship (Amendment) Act 2019, which amended the Citizenship Act 1955 to provide that Hindus, Sikhs, Buddhists, Jains, Parsis and Christians from Afghanistan, Bangladesh and Pakistan who entered India on or before 31 December 2014 are not illegal migrants, and to let them seek citizenship under a new Section 6B with a residence qualification of five years rather than eleven. Parliament passed the Act in December 2019, it entered into force on 10 January 2020, and the Citizenship (Amendment) Rules notified on 11 March 2024 supplied the application machinery through which grants are made.

Q: Why does the CAA exclude Muslims?

The Act excludes Muslims because its stated ground is relief for religious minorities in three states whose constitutions establish Islam, and the government argued in Parliament that Muslims, as adherents of the state religion in Afghanistan, Bangladesh and Pakistan, cannot claim that ground. Home Minister Amit Shah stated that position in the Lok Sabha on 9 December 2019 and added that Muslim applicants remain free to apply under the ordinary provisions of the 1955 Act. Critics answered that persecution in the three countries also fell on Muslim communities such as Ahmadiyyas and Hazaras, so the exclusion contradicts the Act’s stated purpose.

Q: What is the basic-structure-doctrine challenge?

The basic-structure challenge holds that the Act violates the Constitution’s basic structure because secularism is one of that structure’s features, as several opinions in Kesavananda Bharati held in 1973 and as the Supreme Court affirmed in S. R. Bommai in 1994, and a statute that makes religion the criterion for access to citizenship offends secularism at the boundary of the political community itself. The government’s response is that the basic-structure doctrine disciplines constitutional amendments rather than ordinary legislation, that a statute must be tested against enumerated provisions such as Article 14, and that Indian secularism has never barred religion-conscious law.

Q: When was the CAA implemented?

The Act was passed by the Lok Sabha on 9 December 2019 and the Rajya Sabha on 11 December 2019, received presidential assent on 12 December 2019 and was brought into force on 10 January 2020. Implementation in the operational sense waited more than four years, because grants under Section 6B required rules prescribing the application process, and the Ministry of Home Affairs notified the Citizenship (Amendment) Rules 2024 only on 11 March 2024. Applications then opened through the online portal, district level committees began verification, empowered committees began deciding cases, and the first certificates under the Act were issued in May 2024.

Q: What is the relationship between CAA and NRC?

The CAA and the National Register of Citizens are separate instruments that public debate joined because of sequence and statements. The Assam NRC, an administrative update of the citizen register for one state under the Assam Accord framework, published its final list on 31 August 2019 and left out 1,906,657 applicants. The CAA followed in December 2019 and offered the excluded who belonged to the six named faiths and three named countries a pathway that excluded Muslims lacked. Critics argued the two together would render excluded Muslims stateless, while the government maintained that no nationwide NRC had been decided and that the Act took no citizen’s status away.

Q: What was Shaheen Bagh?

Shaheen Bagh was a sit-in protest in a south Delhi neighbourhood that began on 15 December 2019, days after the CAA passed, and continued until the site was cleared on 24 March 2020 as the COVID-19 lockdown began. Muslim women from the locality led the occupation of a stretch of road, reading the Constitution’s Preamble and displaying the national flag and portraits of the founders, and the site became the best-known symbol of the winter’s resistance to the Act. For the architectural debate, Shaheen Bagh mattered because its participants framed their claim in constitutional rather than communal terms.

Q: What does Article 14 say about CAA?

Article 14 provides that the State shall not deny to any person equality before the law or the equal protection of the laws within the territory of India, and its application to the CAA turns on the reasonable-classification doctrine: a classification must rest on an intelligible differentia bearing a rational relation to the statute’s objective. Petitioners argue that religion fails that test because persecution, the Act’s stated objective, also struck groups the Act omits. The government answers that the differentia is minority status in three theocratic states and that Parliament may legislate in stages. The Supreme Court has not decided the question.

Q: Which definition in the Citizenship Act 1955 did the 2019 amendment change first?

The amendment’s first operative change was to the definition of illegal migrant in Section 2(1)(b) of the Citizenship Act 1955. That definition, inserted by the 2003 amendment, covers a foreigner who entered India without a valid passport or travel documents or who stayed beyond the permitted period, and it barred such a person from citizenship by registration or naturalisation. The 2019 Act added a proviso declaring that members of the six named communities from the three named countries who entered on or before 31 December 2014, and whom the central government had exempted from the Passport (Entry into India) Act 1920 and the Foreigners Act 1946, are not illegal migrants under the Act.

Q: What documents must an applicant file under the Citizenship (Amendment) Rules 2024?

The 2024 Rules require an online application supported by an affidavit and three families of documents. An eligibility certificate from a locally reputed community institution attests that the applicant belongs to one of the six named faiths. A document issued by the authorities of Afghanistan, Bangladesh or Pakistan, drawn from a prescribed list that includes passports, birth certificates, school records and land documents, proves the applicant’s country of origin. A third document, chosen from a list of twenty that includes rent agreements, ration cards and Indian school certificates, proves entry into India on or before 31 December 2014. The district level committee verifies the set before the empowered committee decides.

Q: Which regions are exempt from the CAA under the Inner Line Permit system and the Sixth Schedule?

The Act does not apply to the tribal areas of Assam, Meghalaya, Mizoram and Tripura that are included in the Sixth Schedule to the Constitution, which gives autonomous district councils authority over land and local affairs in those areas. It also does not apply to the areas covered by the Inner Line Permit system under the Bengal Eastern Frontier Regulation 1873, which at the Act’s passage covered Arunachal Pradesh, Nagaland and Mizoram and was extended to Manipur in December 2019. The exemptions preserve older protective settlements for indigenous communities in the Northeast, so the Act’s citizenship pathway operates in the region’s non-tribal, non-permit areas, including the Brahmaputra valley.

Q: How did the two Houses vote on the Citizenship (Amendment) Bill in December 2019?

The Lok Sabha passed the Bill on 9 December 2019 by 311 votes to 80, after a debate in which Home Minister Amit Shah moved the Bill and opposition members including Manish Tewari and Mahua Moitra argued that it violated Article 14 and the secular character of the Constitution. The Rajya Sabha passed it on 11 December 2019 by 125 votes to 105, with Law Minister Ravi Shankar Prasad defending the classification as relief for persecuted minorities of three theocratic states. President Ram Nath Kovind gave assent on 12 December 2019, the Act was published in the Gazette the same day, and the government later notified 10 January 2020 as its commencement date.

Q: What argument did Gautam Bhatia make about the CAA and equality?

Gautam Bhatia, a constitutional scholar who analysed the Bill on his Indian Constitutional Law and Philosophy forum in December 2019, argued that the classification fails Article 14 because it is under-inclusive against its own stated purpose: if the objective is relief from religious persecution, a statute that excludes persecuted Ahmadiyyas, Shias, Hazaras and atheists in the named countries has chosen a differentia, the applicant’s religion, that does not track the harm it claims to remedy. Bhatia also developed the manifest-arbitrariness limb drawn from the Shayara Bano judgment of 2017, arguing that a religious line in citizenship law is arbitrary in the sense that doctrine condemns. His essays shaped the petitions’ pleading.

Q: How does Niraja Gopal Jayal describe the competing ideas of citizenship in India?

In Citizenship and Its Discontents: An Indian History, published in 2013, Niraja Gopal Jayal describes Indian citizenship as a field of tension between a jus soli imagination, membership by birth on the soil, and a jus sanguinis pull, membership by descent and community. She shows that the statutory history after 1955 progressively narrowed the jus soli element, through the amendments of 1986, 2003 and 2015, and that administrative practice kept faith entangled with belonging even where the constitutional text refused a religious test. Her account frames the 2019 Act as the moment a long administrative drift became an express legislative criterion, which is why scholars across positions cite her book in the CAA debate.

Q: Why did the 2019 Act choose 31 December 2014 as its entry cutoff?

The cutoff repeats the date used in the Ministry of Home Affairs exemption notifications of 2015 and 2016, which regularised the position of the six named communities from the three named countries who had entered India on or before 31 December 2014 under the Passport (Entry into India) Act 1920 and the Foreigners Act 1946. By adopting the same date, the Act converted an existing executive class into a statutory class and kept the beneficiary group closed and historically defined, which the government used to answer the objection that the Act opened an indefinite future flow. Petitioners replied that a closed class drawn by religion remains a religious classification whatever its date.

Q: What do the district level committees do with a Section 6B application under the 2024 Rules?

Under the Citizenship (Amendment) Rules 2024, the district level committee is the verifying tier of the process. Chaired by a senior postal officer acting as designated officer, the committee receives the online application after filing, scrutinises the affidavit, the eligibility certificate and the documents proving country of origin and date of entry, and administers the oath of allegiance to the applicant. The committee then forwards the verified file to the state empowered committee, headed by the Director of Census Operations, which holds the power to grant or refuse citizenship. A successful applicant receives a certificate and is treated as a citizen from the date of entry into India recorded in the application.

Q: Does the CAA change how a person born in India acquires citizenship?

The CAA does not amend Section 3 of the Citizenship Act 1955, so citizenship by birth continues to operate on the terms the earlier amendments set: a person born in India on or after 3 December 2004 is a citizen by birth only if both parents are citizens, or one parent is a citizen and the other is not an illegal migrant. The 2019 Act touches birth only indirectly, through the definition in Section 2(1)(b): because a covered entrant is declared not to be an illegal migrant, the entrant’s children are not disqualified by the parent’s status in the way the 2003 framework would otherwise have required. The grant under Section 6B is also backdated to entry.

Q: How many applicants were left out of the Assam NRC final list published on 31 August 2019?

The final list of the Assam National Register of Citizens, published on 31 August 2019 after an update process supervised by the Supreme Court, included 31,121,004 of the 33,027,661 persons who had applied, and left out 1,906,657 applicants, roughly 1.9 million people. Exclusion from the list did not itself declare any person a foreigner, since appeals lay to foreigners tribunals, but the figure set the human scale against which the CAA was read three months later: the Act offered the excluded who belonged to the six named faiths a statutory pathway, while the excluded who were Muslim faced the tribunal process without one, which is the asymmetry at the centre of the Assam argument.

Q: What procedural step did the Supreme Court take on the CAA petitions on 22 January 2020?

On 22 January 2020 a bench headed by Chief Justice S. A. Bobde heard the accumulated challenges to the Act, by then numbering more than one hundred and forty petitions, and directed the Union of India to file its reply within four weeks. The bench declined to stay the Act, which had entered into force on 10 January 2020, and observed that the questions raised might require consideration by a Constitution Bench. The hearing set the pattern that followed: notice, affidavits from the Union defending the classification, repeated listings and adjournments, and no interim order suspending the statute while its merits awaited a final hearing.

Q: What power over Overseas Citizen of India registration did the 2019 Act add?

Alongside its main provisions, the 2019 Act amended Section 7D of the Citizenship Act 1955, which lists the grounds on which the central government may cancel the registration of an Overseas Citizen of India cardholder. The amendment added a new ground: violation by the cardholder of any provision of the Citizenship Act or of any other law the central government notifies in the official gazette. Cancellation on that ground requires that the cardholder be given a reasonable opportunity of being heard. Critics in the December 2019 debates treated the clause as a disciplinary provision widening executive discretion over a large diaspora status, while the government presented it as an enforcement measure aligned with the Act’s scheme.

Q: What separates citizenship by registration from citizenship by naturalisation in the 1955 Act?

Registration under Section 5 and naturalisation under Section 6 are the 1955 Act’s two discretionary routes for adults, and they differ in their qualifying links. Registration serves persons with a defined connection to India, including persons of Indian origin ordinarily resident for seven years, spouses of citizens, and minor children of citizens. Naturalisation serves any foreigner who meets the Third Schedule qualifications, centred on residence or government service aggregating eleven years in the fourteen years before application, knowledge of a language in the Eighth Schedule and intention to reside in India. Section 6B, inserted in 2019, lets the covered class use either route with the residence period reduced to five years.