A shopkeeper refuses to serve a customer because of the customer’s caste. A landowner’s men beat a labourer who asked for the statutory wage. A candidate who stood first in her reserved category is denied a post because the vacancy was treated as unreserved. Public discussion files all three events under one heading, caste discrimination, and reaches for one body of law to answer them. That instinct is the single largest source of error in this field, because the social justice laws of India are not one body of law. They are three distinct legal layers that happen to share a subject: a layer of constitutional guarantee, which declares rights and empowers the State to act; a layer of criminal prohibition, which makes specified conduct an offence and sends the complainant to the police and the criminal courts; and a layer of affirmative provision, which distributes seats, posts and benefits through administrative machinery governed by service law and admission law. Each layer has its own texts, its own authorities, its own forum and its own remedy, and a claim framed in the wrong layer fails for reasons that have nothing to do with its merits.

The refusal of service belongs to the criminal layer, and within it to the Protection of Civil Rights Act, 1955. The beating belongs to the criminal layer as well, but to a different statute, the Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989, with its own list of offences, its own special courts and its own relief machinery. The denied post belongs to neither: it is a question of affirmative provision, answered by Articles 15 and 16 of the Constitution, by the reservation orders made under them, and by the courts that hear service and admission disputes. A reader who can sort a fact situation into its layer holds the key to the whole field, because almost every dispute in this area turns on which layer is engaged, and the arguments, the evidence and the authorities that matter differ completely from layer to layer. This guide maps the three layers in turn, then maps the machinery that connects them: the lists that decide who belongs to a protected class, the commissions that watch over the safeguards, the special governance arrangements for tribal areas, and the wider perimeter of protective statutes for groups the Constitution treats under other heads.
What counts as a social justice law
The expression social justice appears in the Preamble to the Constitution, which resolves to secure to all citizens justice, social, economic and political. The Constitution nowhere defines the term, and no single statute bears it as a short title. The field is therefore identified by function rather than by label: a law belongs to this field when its purpose is to correct a disadvantage that attaches to membership of a group, whether the disadvantage is exclusion from public space, exposure to violence and humiliation, denial of education and employment, bondage in labour, or dispossession from land. Measured by that test, the field includes provisions of the Constitution itself, a small set of criminal statutes, a larger set of statutes that confer benefits or recognise rights, a body of executive orders that operate the reservation system, and the constitutional schedules that govern tribal areas. It also includes institutions: three constitutional commissions and a family of statutory commissions whose task is to investigate, report and recommend.
Two boundaries keep the map honest. The first is that general welfare law is not automatically social justice law. A universal scheme of food subsidy or rural employment serves the poor of every group and is analysed elsewhere in this series as welfare legislation; it enters this field only where a statute or a constitutional provision targets a group defined by historical disadvantage, such as the Scheduled Castes and the Scheduled Tribes, or by a protected characteristic, such as disability or gender, and confers a remedy or an entitlement that the general law does not. The second boundary is that personal law, which governs marriage, divorce, succession and adoption within religious communities, raises distinct questions of equality and reform that this series treats in its own cluster. The two fields touch, because a rule of personal law can distribute rights unequally, but the instruments differ: social justice law in the sense used here works through constitutional guarantees, criminal prohibitions and affirmative provision directed at group disadvantage, and it is that machinery this guide describes.
The vocabulary of the field is itself a map feature, because the Constitution and the administration use different words for overlapping sets of people. The constitutional text speaks of classes, not castes: Article 15(4) speaks of socially and educationally backward classes of citizens and names the Scheduled Castes and the Scheduled Tribes beside them, and Article 16(4) speaks of any backward class of citizens. From M. R. Balaji onwards the courts have treated caste as capable of constituting a class and as a legitimate starting point for identifying one, while insisting that caste cannot be the sole criterion and that the identification must survive the classification test. The administrative vocabulary adds the Other Backward Classes, the residual category of backward classes other than the Scheduled Castes and the Scheduled Tribes, usually abbreviated to OBC, and the economically weaker sections, abbreviated to EWS, for the branch added in 2019. The expression weaker sections, used in Articles 38 and 46, is the widest of all and has never been given a closed definition; the Supreme Court in N. M. Thomas treated it as embracing the Scheduled Castes and the Scheduled Tribes at its core while extending beyond them. A reader who keeps the vocabularies apart, the text’s classes, the administrator’s categories and the directive’s weaker sections, will avoid the common error of assuming that a term in a judgment, a roster and a scheme guideline necessarily denotes the same set of people.
The constituting materials can be stated shortly, and the rest of this guide is an expansion of the list. The constitutional layer consists of the equality code in Articles 14 to 18, the prohibition of trafficking and forced labour in Article 23, the guarantee against denial of admission in Article 29, the directive principle in Article 46, the balancing provision in Article 335, the legislative reservation provisions in Articles 330 to 334, the local government reservation provisions in Articles 243D and 243T, the specification provisions in Articles 341, 342 and 342A, the commission provisions in Articles 338, 338A and 338B, and the Fifth and Sixth Schedules. The criminal layer consists principally of the Protection of Civil Rights Act, 1955 and the Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989, supported by the Bonded Labour System (Abolition) Act, 1976 and the Prohibition of Employment as Manual Scavengers and their Rehabilitation Act, 2013. The affirmative layer consists of the enabling clauses of Articles 15 and 16 as amended, the statutes and orders that quantify shares, among them the Central Educational Institutions (Reservation in Admission) Act, 2006, and the protective statutes that recognise rights in land and self government, principally the Panchayats (Extension to the Scheduled Areas) Act, 1996 and the Forest Rights Act, 2006. Around that core sits a perimeter of statutes for women, children, persons with disabilities, transgender persons and senior citizens, administered in the main by the same ministries.
How the framework was assembled: the sequence from 1950 to 2024
The framework has roots before the Constitution, and three of them still shape its structure. The expression Scheduled Castes entered the law through the Government of India Act, 1935, which provided for schedules of castes to be specified for electoral purposes, and the Government of India (Scheduled Castes) Order, 1936 listed the castes accordingly; the specification technique of 1950, a Presidential Order listing castes State by State, is the direct descendant of that instrument. Legislative reservation descends from the settlement of 1932, when the Communal Award’s proposal of separate electorates for the depressed classes was replaced, by the Poona Pact between B. R. Ambedkar and M. K. Gandhi, with reserved seats within joint electorates, the design Articles 330 and 332 later adopted: the seat is reserved for a candidate of the class, and the electorate that chooses is the whole constituency. The third root is the administrative practice of the provinces, several of which, Madras and Bombay among them, operated communal orders and backward class preferences in services and admissions decades before 1950, so that the First Amendment of 1951 was not inventing affirmative provision but constitutionalising a practice the courts had just held the new equality code forbade. A framework assembled from an electoral schedule, a political settlement and a provincial practice was never going to read as a code; it reads, as this guide has presented it, as layers laid down at different times for different purposes, and the reader who knows the order in which they were laid can predict where the joins, and therefore the disputes, will be found.
The framework was not designed at a sitting. It accumulated in four broad movements, and the dates matter because each layer was built in a different movement. The first movement is the founding settlement, from 1950 to 1951. The Constitution commenced with the equality code, the abolition of untouchability, the prohibition of forced labour, the directive of Article 46, the specification machinery and the two Presidential Orders of 1950 that filled the Scheduled Castes and Scheduled Tribes lists. Within eighteen months the settlement met its first judicial limit, when the Supreme Court struck down the Madras communal order in Champakam Dorairajan, and Parliament answered with the First Amendment, inserting Article 15(4) and thereby founding the affirmative layer. The sequence established the field’s characteristic rhythm at the outset: text, judgment, amendment.
The second movement, from 1955 to 1989, built the criminal layer and the commissions’ ancestry. Parliament enacted the Untouchability (Offences) Act in 1955 on the foundation of Article 17, received the Elayaperumal Committee’s evidence that the abolition was not reaching the districts, and answered in 1976 with the comprehensive amendment that renamed the statute the Protection of Civil Rights Act and strengthened its provisions. The same year, 1976, brought the Bonded Labour System (Abolition) Act, extending the criminal layer to the Article 23 prohibition. The Special Officer under Article 338 continued through this period as the framework’s single watcher, reporting annually to the President. The movement closed with the Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989, Parliament’s judgment, stated in that Act’s objects, that the civil rights statute had not sufficed against the violence the framework existed to prevent.
The third movement, from 1990 to 2008, rebuilt the affirmative layer and the institutional map together. The Mandal Commission’s report was implemented for central services in 1990, the combined National Commission for the Scheduled Castes and the Scheduled Tribes acquired constitutional status by the Sixty-fifth Amendment in the same year, and the Supreme Court’s nine judge decision in Indra Sawhney in 1992 settled the doctrine of the affirmative layer: the fifty per cent ceiling, the exclusion of the creamy layer, the rejection of reservation in promotions and the recommendation of a permanent body for the backward classes lists, which Parliament enacted as the National Commission for Backward Classes Act, 1993. The promotion amendments of 1995, 2000 and 2001 then worked through the points on which Parliament disagreed with the judgment, and the division of the combined commission into separate commissions for the Scheduled Castes and the Scheduled Tribes by the Eighty-ninth Amendment in 2003 completed the institutional structure for those classes. On the protective side, PESA in 1996 carried the Fifth Schedule’s logic into panchayats, the Ninety-third Amendment in 2005 opened private educational institutions other than minority institutions to special provision, the Central Educational Institutions Act followed, and the Forest Rights Act, enacted in 2006 and in force from the last day of 2007, added recognition of forest rights to the framework’s techniques.
The fourth movement, from 2013 onwards, extended the perimeter and then reopened the framework’s foundations. The manual scavengers statute of 2013 replaced the failed prohibition of 1993 with a prohibition joined to rehabilitation. The National Legal Services Authority judgment of 2014 and the Transgender Persons Act of 2019 brought gender identity within the framework, and the Rights of Persons with Disabilities Act, 2016 consolidated the law for disability on the three layer pattern. Then the backward classes settlement was rebuilt twice in three years: the One Hundred and Second Amendment in 2018 constitutionalised the National Commission for Backward Classes and inserted Article 342A, the Constitution Bench in the Maratha case in 2021 read that amendment as transferring identification of backward classes to the President alone, and the One Hundred and Fifth Amendment in the same year restored the States’ power in express words. The One Hundred and Third Amendment of 2019 had already added the economically weaker sections branch, upheld in Janhit Abhiyan by three judges to two, and in 2024 the seven judge bench in Davinder Singh authorised sub classification within the Scheduled Castes list. Read as a whole, the sequence shows a framework that changes by constitutional amendment more readily than most fields of law, because its operative questions, who belongs, what share, which authority, are questions the constitutional text itself reserves, and every generation of litigation has returned them to the text.
Social justice laws in India begin in the Constitution
Every route in this field starts in the constitutional text, because the Constitution does three things at once. It guarantees equality as an enforceable right. It prohibits particular practices outright. And it empowers the State to depart from formal equality in favour of specified classes, within limits the courts police. Article 14 guarantees equality before the law and the equal protection of the laws to every person. The courts read that guarantee through the doctrine of classification: a law may classify, provided the classification rests on an intelligible differentia that distinguishes the persons grouped together from those left out, and provided the differentia bears a rational relation to the object the law seeks to achieve. That two-part test, developed in the early equality cases and restated across the reservation jurisprudence, is the doctrinal foundation on which every affirmative measure in this field stands or falls, because a reservation, a special provision or a protective statute is, in form, a classification of citizens, and it survives challenge under Article 14 only when both limbs of the test are met.
Article 15 prohibits discrimination by the State against any citizen on grounds only of religion, race, caste, sex, place of birth or any of them, and Article 15(2) carries the prohibition into public access, providing that no citizen shall, on those grounds only, be subject to any disability or restriction in access to shops, public restaurants, hotels and places of public entertainment, or in the use of wells, tanks, bathing ghats, roads and places of public resort maintained wholly or partly out of State funds or dedicated to the use of the general public. Article 16 guarantees equality of opportunity in public employment and prohibits discrimination on grounds only of religion, race, caste, sex, descent, place of birth and residence. Article 17 abolishes untouchability in a single sentence of unusual breadth: untouchability is abolished and its practice in any form is forbidden, and the enforcement of any disability arising out of untouchability is an offence punishable in accordance with law. Article 18 abolishes titles. Article 23 prohibits traffic in human beings and begar and other similar forms of forced labour, and makes the prohibition enforceable against private persons as well as the State, a feature it shares with Article 17 and one that matters when the framework is tested against private conduct.
Alongside the rights stand the directives. Article 46, in Part IV, requires the State to promote with special care the educational and economic interests of the weaker sections of the people, and in particular of the Scheduled Castes and the Scheduled Tribes, and to protect them from social injustice and all forms of exploitation. Part IV is not enforceable by any court, and the history of the field is in part the history of Parliament and the courts negotiating the relation between the directives and the rights: the First Amendment of 1951, which inserted clause (4) into Article 15 after the Supreme Court struck down a communal order of seats in the State of Madras v. Champakam Dorairajan, was the first such negotiation, and it set the pattern of constitutional amendment answering judicial reading that recurs throughout this guide, in the promotion amendments of the 1990s, in the educational institutions amendment of 2005 and in the backward classes amendments of 2018 and 2021.
Article 46 does not stand alone in Part IV. Article 38 directs the State to promote the welfare of the people by securing and protecting a social order in which justice, social, economic and political, shall inform all the institutions of the national life, and, in clause (2), to strive to minimise the inequalities in income and to endeavour to eliminate inequalities in status, facilities and opportunities, not only amongst individuals but also amongst groups of people residing in different areas or engaged in different vocations. The group language of Article 38(2) is the directive counterpart of the classification doctrine in Part III: where Article 14 permits the law to treat groups differently for a rational object, Article 38 instructs the State to aim its policy at the inequalities that separate groups. The Supreme Court’s mature position, stated in Minerva Mills v. Union of India in 1980, is that Parts III and IV are to be read in harmony, neither having primacy over the other, and the social justice statutes are the practical deposit of that harmony: a prohibition statute such as the Protection of Civil Rights Act enforces a Part III right in the service of a Part IV objective, and a recognition statute such as the Forest Rights Act converts a Part IV commitment into Part III enforceable rights by the act of enactment itself.
Two further constitutional provisions complete the first layer’s perimeter. Article 15(3) provides that nothing in Article 15 shall prevent the State from making any special provision for women and children, a clause that predates the backward classes clauses and that carries the large body of protective legislation for those groups, from maternity benefit to the prohibition of child labour in hazardous work. Article 29 protects the cultural and educational rights of minorities: clause (1) guarantees any section of citizens having a distinct language, script or culture the right to conserve it, and clause (2) provides that no citizen shall be denied admission into any educational institution maintained by the State or receiving aid out of State funds on grounds only of religion, race, caste, language or any of them. Article 29(2) is the provision the communal order in Champakam was also held to violate, and it is the textual reason Article 15(4) had to be drafted to override it expressly. Beside it stands Article 30, which guarantees religious and linguistic minorities the right to establish and administer educational institutions of their choice and protection against discrimination in the grant of aid. The minority provisions cut across the affirmative layer in both directions: they are themselves a form of group protection, addressed to cultural survival rather than to backwardness, and they mark the boundary at which reservation stops, because Article 15(5) and the admission statutes made under it exclude minority educational institutions from their reach. A map that omits Articles 29 and 30 cannot explain why two colleges on the same street, both private and unaided, may stand under different admission obligations.
The horizontal reach of Articles 17 and 23 deserves emphasis, because it distinguishes them from most of Part III. The fundamental rights are, in the main, guarantees against the State, and Article 15(1) and Article 16 bind State action. Articles 17 and 23 are framed against the practice and the conduct, whoever engages in them, and the Supreme Court confirmed the consequence in People’s Union for Democratic Rights v. Union of India in 1982, the case arising from the conditions of workers on the Asian Games construction projects in Delhi: the prohibitions in Articles 17, 23 and 24 are enforceable against private persons as well as the State, and a writ may issue for their enforcement accordingly. The point matters to the architecture of the whole framework. The criminal layer’s statutes bind private conduct because the constitutional prohibitions they enforce already did; a shopkeeper, a landlord or an employer who enforces a disability arising out of untouchability violates Article 17 itself and not merely a statute, and the bonded labourer held by a private moneylender holds a constitutional right against the moneylender that the 1976 Act then operationalises. The affirmative layer, by contrast, binds the State as distributor and reaches private institutions only where an enabling clause, Article 15(5) or Article 15(6), has expressly carried it there, and then only by law. The framework’s treatment of private power is therefore asymmetrical by design: prohibition runs against everyone, distribution runs through the State, and most disputes about private conduct in this field begin by asking which of those two propositions governs the facts.
How many layers does the social justice framework have?
Three. The constitutional layer declares rights, prohibits specified practices and empowers the State to make special provision. The criminal layer makes defined conduct an offence and routes the complaint through the police and the criminal courts. The affirmative layer distributes seats, posts and benefits through reservation and recognition statutes administered by governments, universities and employers.
The three layers differ in everything that decides a real dispute. The constitutional layer is enforced principally by writ petition to the High Courts and the Supreme Court under Articles 226 and 32, and its characteristic remedies are the declaration that a law or an order is invalid and the direction that an authority act according to law. The criminal layer is enforced by prosecution, its burden of proof is proof beyond reasonable doubt, and its characteristic outcomes are conviction, sentence and the statutory relief that follows a registered case. The affirmative layer is enforced through the administrative process itself: a candidate applies, a roster is maintained, a list is published, and the dispute, when it comes, is about whether the roster, the list or the eligibility rule was correctly drawn and correctly applied. Confusing the layers produces characteristic errors. A complainant who treats a refusal of service as a service matter writes to a department and waits; a candidate who treats a roster error as an atrocity files a criminal complaint that no court can convert into an appointment. The layers also differ in who may invoke them: the criminal statutes in this field protect members of the Scheduled Castes and the Scheduled Tribes against offenders who are not members of those classes, while the affirmative provisions can be invoked only by a person who belongs to the class for which the provision is made and who satisfies its conditions.
Is reservation an exception to the right to equality?
No, in the settled reading. In State of Kerala v. N. M. Thomas, (1976) 2 SCC 310, the majority held that clause (4) of Article 16 is not an exception to the equality of opportunity guaranteed by clause (1) but an emphatic statement of a classification that clause (1) itself permits.
The Constitution Bench in Indra Sawhney carried the same approach into the general doctrine: the reservation provisions are facets of the equality code, instruments for achieving substantive equality, and not derogations from it that must be confined to the narrowest possible scope.
The distinction is not verbal. If reservation were an exception, it would be construed strictly, the burden would rest on the State throughout, and every extension of it would begin under suspicion. Treated as a facet of equality, reservation is construed as one of the means by which the equality code pursues its own object, and the questions a court asks are the classification questions: whether the class identified is genuinely backward or inadequately represented, whether the share fixed bears a rational relation to that condition, and whether the measure stays within the limits the Court has drawn, including the ceiling on total reservation and the exclusion of the creamy layer within backward classes. Article 178 of this series examines the Indra Sawhney holdings in detail; what matters for the map is the location of the doctrine. The argument about reservation is an argument inside the equality code, conducted with the code’s own tools, and the frequent description of reservation as an exception to equality misstates the doctrine that the courts apply.
Why is untouchability not defined anywhere in the constitutional text?
Because the Constituent Assembly chose prohibition over definition. Article 17 abolishes untouchability and forbids its practice in any form, without defining the term, and neither the Protection of Civil Rights Act, 1955 nor any other statute supplies a definition. The courts have therefore construed the word in its historical sense.
In that sense the word means the practice of imposing social disabilities on persons by reason of their birth in specified castes, rather than the literal imposition of physical contact.
The choice has consequences that run through the criminal layer. A prosecution under the 1955 Act must fit the facts within one of the disabilities the Act enumerates, such as preventing a person from entering a place of public worship, from offering prayers, from using a public conveyance or from accessing a shop, because the Act punishes the enforcement of specified disabilities rather than the abstraction of untouchability itself. Conduct that humiliates without enforcing a listed disability may therefore fall outside the 1955 Act entirely, and it was part of the stated case for the stronger statute of 1989 that a different instrument, organised around offences against members of the Scheduled Castes and the Scheduled Tribes, was required. The definitional gap also explains why the boundary between the two criminal statutes is a working question for complainants and police officers rather than a theoretical one: the same episode can disclose an enumerated disability under the earlier Act, a listed offence under the later Act, both, or neither, and the classification chosen at the stage of the first information report shapes the investigation, the court and the relief that follows.
What does Article 335 add to the reservation provisions?
A balancing instruction. Article 335 provides that the claims of the members of the Scheduled Castes and the Scheduled Tribes shall be taken into consideration, consistently with the maintenance of efficiency of administration, in the making of appointments to services and posts in connection with the affairs of the Union or of a State.
A proviso added by the Constitution (Eighty-second Amendment) Act, 2000 permits the State to relax qualifying marks and standards of evaluation in favour of such members in promotion.
Article 335 is the point at which the affirmative layer acknowledges a competing constitutional value in its own text. The Supreme Court has read the efficiency clause neither as a veto on reservation nor as decoration: in the promotion cases, of which M. Nagaraj v. Union of India in 2006 is the leading authority, the Court treated the clause as one of the conditions that discipline reservation in promotion, alongside quantifiable data on backwardness and inadequacy of representation, and in Jarnail Singh v. Lachhmi Narain Gupta in 2018 the Court removed the requirement that the State collect quantifiable data to prove the backwardness of the Scheduled Castes and the Scheduled Tribes for promotion purposes, while retaining the application of the creamy layer principle. Article 178 examines that line of authority in full. For the map, the point is structural: the Constitution does not present equality, reservation and administrative efficiency as provisions at war with one another. It places them in adjacent articles of the same Part and instructs the State to hold them together, and the litigation in this field is largely litigation about what that instruction requires in a given service, in a given year, on a given record.
The constitutional layer also carries a political branch of affirmative provision that is easy to overlook because it operates through elections rather than appointments. Articles 330 and 332 reserve seats for the Scheduled Castes and the Scheduled Tribes in the Lok Sabha and in the Legislative Assemblies of the States, in proportion, as nearly as may be, to their population. Article 334 fixed a sunset for that reservation: the provisions were to cease after ten years from the commencement of the Constitution, and Parliament has extended the period by successive amendments, the Eighth in 1959, the Twenty-third in 1969, the Forty-fifth in 1980, the Sixty-second in 1989, the Seventy-ninth in 1999, the Ninety-fifth in 2009 and the One Hundred and Fourth in 2019, which carried the reservation to 25 January 2030. The One Hundred and Fourth Amendment also ended the nomination of members of the Anglo-Indian community to the Lok Sabha and the State Assemblies, which had shared the sunset clause, so that the periodic review the sunset requires produced two different answers for two different provisions in a single amending Act. Articles 243D and 243T carry reservation into the third tier: seats for the Scheduled Castes and the Scheduled Tribes, and not less than one third of the total number of seats for women, in panchayats and municipalities, with offices of chairpersons reserved on the same pattern. The local provisions have no sunset clause, a textual difference that reflects the different design of Part IX and Part IXA, and challenges to them are governed by the same classification doctrine applied to a different institutional setting.
The local tier produced its own leading case. In K. Krishna Murthy v. Union of India, (2010) 7 SCC 202, a Constitution Bench upheld the reservation scheme of Articles 243D and 243T, including reservation for backward classes in local bodies where State law provides it, while attaching conditions drawn from the service and admission doctrine: the State must rest backward class reservation in local bodies on quantifiable data regarding the inadequacy of representation in those bodies, and the aggregate of vertical reservation must observe the fifty per cent ceiling. The judgment illustrates how the doctrine travels down the layers: the classification test, the ceiling and the demand for data, developed for posts and seats in Articles 15 and 16, were applied to a third institutional setting with its own constitutional text, and the differences the Court recognised, the political character of local office and the absence of an efficiency clause comparable to Article 335, were differences of application rather than of principle. For the aspirant, the local provisions are also the point at which the framework’s three branches of reservation can be compared in a single answer: legislative reservation under Articles 330 and 332 is proportional to population and time limited by Article 334, employment and admission reservation under Articles 15 and 16 is share based and doctrine limited by the ceiling, and local reservation under Articles 243D and 243T combines proportionality for the Scheduled Castes and the Scheduled Tribes with a mandatory one third share for women across all categories, the only place in the constitutional text where women’s reservation in elected bodies is enacted rather than proposed.
The federal division: who makes and who administers this law
The framework is national in its texts and State administered in its operation, and the division follows the Seventh Schedule. Criminal law and criminal procedure stand in the Concurrent List, which is why Parliament could enact the Protection of Civil Rights Act and the Atrocities Act for the whole country, the former resting additionally on Article 35. Police and public order stand in the State List, which is why the investigation of every offence under those Acts is the work of State police forces answerable to State governments, and why the quality of the criminal layer varies with the administration of the State in which the offence occurs. Education was moved from the State List to the Concurrent List by the Forty-second Amendment in 1976, and admission reservation in State institutions is State policy within the constitutional doctrine, while the central institutions are governed by Parliament’s statute of 2006. Land stands in the State List, which is why the Forest Rights Act, a Union statute recognising rights over forest land, is implemented through State machinery and why its interface with State land and forest administration generates a large share of its disputes. Social security, social insurance and the welfare of labour are Concurrent subjects, and the subjects of the perimeter statutes are distributed on the same pattern.
The division explains two features of the framework that puzzle readers who approach it as a single code. The first is that the lists and the quotas differ by State without any inconsistency: the Scheduled Castes and Scheduled Tribes lists are specified in relation to each State, each State maintains its own backward classes list for its own purposes, a power the amendment of 2021 wrote into Article 342A(3) expressly, and the quantum of reservation in State services and State institutions is fixed by each State within the ceiling doctrine, so that the total share reserved in one State’s services may lawfully differ from the share in another’s. The second is that implementation is everywhere State business even where the text is Parliament’s: the Protection of Civil Rights Act is implemented by State governments and Union territory administrations, the vigilance committees of the Bonded Labour Act are district bodies, the scrutiny committees for certificates are State committees, and the commissions at the Centre watch, report and recommend over a machinery they do not command. A grievance in this field therefore almost always has a State address and a Union text, and identifying both is part of the routing exercise the closing sections of this guide formalise.
The best known illustration of State variation is Tamil Nadu. That State reserves sixty nine per cent of seats and posts in the aggregate, a share assembled over decades of State policy, and when the fifty per cent ceiling was settled in Indra Sawhney the share stood exposed to challenge. Tamil Nadu answered by enacting the Tamil Nadu Backward Classes, Scheduled Castes and Scheduled Tribes (Reservation of Seats in Educational Institutions and of Appointments or Posts in the Services under the State) Act, 1993, and Parliament, by the Constitution (Seventy-sixth Amendment) Act, 1994, placed that Act in the Ninth Schedule, the schedule whose listed enactments were historically shielded from challenge on fundamental rights grounds. The shield is no longer absolute: in I. R. Coelho v. State of Tamil Nadu in 2007, a bench of nine judges held that laws placed in the Ninth Schedule after 24 April 1973, the date of the Kesavananda Bharati judgment, are open to challenge where they damage the basic structure, including the equality principle. The Tamil Nadu share therefore stands as a working exception with a contested perimeter, applied in the State’s admissions and services while its constitutional shelter is qualified rather than complete. For the map, the episode completes the federal picture: the ceiling is national doctrine, the shares are State choices, and the points at which a State choice exceeds the doctrine are resolved by the amendment and basic structure machinery of the constitutional layer, not by the service law of the affirmative layer.
The criminal layer: prohibition backed by punishment
Article 17 is self executing as a declaration but not as a penal code. Its last clause contemplates legislation: the enforcement of any disability arising out of untouchability shall be an offence punishable in accordance with law. Parliament acted under that clause and under Article 35, which reserves to Parliament the power to make laws prescribing punishment for acts declared to be offences under Part III. The resulting statute, the Untouchability (Offences) Act, 1955, received the assent of the President in May 1955 and entered into force on 1 June 1955 as Act 22 of 1955. It was comprehensively amended and renamed the Protection of Civil Rights Act, 1955 by Act 106 of 1976, with effect from 19 November 1976, following the report of the Committee on Untouchability, Economic and Educational Development of the Scheduled Castes, chaired by L. Elayaperumal, which Parliament had appointed in April 1965 and which reported in January 1969. The renaming itself carries the statute’s theory: the Act protects civil rights, meaning the rights accruing to a person by reason of the abolition of untouchability, and it does so by punishing their denial. The statute’s provisions, its presumption and its record are examined in this series at the Protection of Civil Rights Act, 1955, which owns the statute in detail; this guide takes from it the features that fix its place on the map.
Those features are three. First, the Act protects every person, because the disabilities it lists are disabilities imposed on the ground of untouchability, and its enforcement machinery does not ask whether the victim belongs to a Scheduled Caste. Second, its penalties are modest and were set in the monetary terms of an earlier era: for the offence of enforcing a religious disability under Section 3, the punishment is imprisonment for a term not less than one month and not more than six months, and a fine not less than one hundred rupees and not more than five hundred rupees, a scale introduced by the amending Act of 1976 and not revised in the decades since. Third, the Act provides for collective fines where the inhabitants of an area are concerned in or abetting an offence, a provision aimed at the social enforcement of exclusion, under which a village that collectively boycotts a family for defying a caste prohibition can be fined as a community. Each feature marks a design decision made in the 1950s and consolidated in 1976: a universal civil right, a deterrent pitched at the level of the petty session, and a collective sanction for collective conduct.
By the 1980s the official position, recorded in the Statement of Objects and Reasons of the later statute, was that the existing law had proved inadequate to check and prevent the atrocities that continued to be committed on members of the Scheduled Castes and the Scheduled Tribes despite the 1955 Act. Parliament’s answer was a second and far heavier criminal statute, the Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989, Act 33 of 1989, which received assent on 11 September 1989 and entered into force on 30 January 1990, with rules following in 1995 that prescribed, among other machinery, the norms of relief and rehabilitation for victims. The Atrocities Act is the subject of the next article in this batch, the Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989, and of the article on its amendment of 2018, so this guide states only the contrasts that locate it on the map. Where the 1955 Act protects any person against disabilities imposed on the ground of untouchability, the 1989 Act protects members of the Scheduled Castes and the Scheduled Tribes against offences committed by persons who are not members of those classes, and membership of the specified lists is therefore an element of every charge. Where the 1955 Act enumerates disabilities, the 1989 Act enumerates offences, a catalogue that reaches insult and intimidation, the imposition of social and economic boycott, the wrongful occupation of land, the pollution of water sources and a series of aggravated forms, expanded substantially by the amendment of 2015. And where the 1955 Act is tried in the ordinary magistracy with small penalties, the 1989 Act provides for special courts, enhanced sentences and a statutory scheme of relief that attaches to the registration and trial of the case. The two statutes overlap in territory and diverge in almost every operative particular, which is why the choice between them, or the decision to invoke both, is the first professional judgment in any case on this layer.
The design of the 1989 Act also illustrates how this field builds procedure into substance. An ordinary prosecution asks the criminal process to do two things, try the offence and punish it. A prosecution under the Atrocities Act is designed to do a third: the Rules of 1995 attach a schedule of relief to the case itself, payable to the victim or the victim’s dependants at defined stages, so that registration of the case starts a welfare process alongside the penal one, administered by the District Magistrate. The special court is part of the same design. Cases under the Act were originally tried by courts of session designated as special courts, and the amendment of 2015 provided for exclusive special courts in districts, on the stated ground that delay in trial was itself a mechanism by which the protection failed, witnesses turning hostile or being bought as proceedings lengthened. The framework’s later statutes repeat the pattern in their own fields: the manual scavengers statute couples prohibition to survey and rehabilitation, and the Forest Rights Act builds its recognition through committees rather than courts. The general proposition the map supports is that in this field the legislature seldom trusts a bare prohibition. Each prohibition arrives with an institution attached, a committee, a fund, a court or a schedule of relief, and the institution is not machinery added to the law but part of the law’s theory of why the previous instrument failed.
The criminal layer extends beyond caste violence to the practices that Article 23 prohibits. The Bonded Labour System (Abolition) Act, 1976, Act 19 of 1976, was deemed to have entered into force on 25 October 1975, having first been promulgated as an ordinance, and it gives effect to the prohibition of forced labour by abolishing the bonded labour system outright: every bonded labourer stood freed and discharged from any obligation to render bonded labour, and every bonded debt stood extinguished, by force of the statute rather than by individual application. Its enforcement machinery is administrative as much as criminal, resting on vigilance committees at district and sub divisional level and on the district magistrate, and its history illustrates a recurring feature of this layer, that a prohibition statute in this field must build its own identification machinery, because the victims of the practice are seldom in a position to invoke the law unaided. The Prohibition of Employment as Manual Scavengers and their Rehabilitation Act, 2013, Act 25 of 2013, whose preamble expressly invokes Article 46 and the goal of dignity that the earlier statute of 1993 had failed to secure, prohibits the employment of persons as manual scavengers and the construction and maintenance of insanitary latrines, and couples the prohibition with survey, rehabilitation and alternative occupation provisions. Article 176 of this series owns that statute. On the map it sits at the junction of the criminal and the affirmative layers: it punishes a practice, and in the same text it undertakes the economic reconstruction of the persons the practice employed, acknowledging that a prohibition without rehabilitation leaves the underlying compulsion intact.
Article 23’s first prohibition, traffic in human beings, carries its own statute, which belongs to the women’s rights cluster of this series and appears here only to complete the layer. The Suppression of Immoral Traffic in Women and Girls Act, 1956, enacted in pursuance of the international convention signed at New York in 1950, was renamed the Immoral Traffic (Prevention) Act by the amending Act 44 of 1986, and it criminalises the commercial sexual exploitation of trafficked persons through brothel keeping, living on the earnings of prostitution and procuring, while providing for protective homes and corrective institutions. Its presence in the layer demonstrates the breadth of the Article 23 foundation: the same clause that grounds the abolition of bonded labour grounds the trafficking statute, and the two statutes together mark the criminal layer’s second front, directed not at caste disability but at the conversion of persons into instruments of labour and commerce. The framework’s treatment of exploitation is therefore wider than its treatment of untouchability, and the layers cross at this point, because the victims of trafficking and bondage are drawn disproportionately from the very classes the lists specify, a fact the rescue and rehabilitation provisions of both statutes acknowledge in their design.
The affirmative layer: shares, seats and posts
Affirmative provision is the layer at which the framework stops prohibiting and starts distributing. Its constitutional machinery is a series of enabling clauses, each inserted or widened at a definite point in the argument between Parliament and the courts, and the sequence of the clauses is itself a history of the field. Article 15(4), inserted by the First Amendment in 1951, provides that nothing in Article 15 or in Article 29(2) shall prevent the State from making any special provision for the advancement of any socially and educationally backward classes of citizens or for the Scheduled Castes and the Scheduled Tribes. The clause is enabling, not self executing: it removes a constitutional objection and leaves the content of the special provision to legislation and executive order. Article 16(4) performs the parallel operation for employment, providing that nothing in Article 16 shall prevent the State from making any provision for the reservation of appointments or posts in favour of any backward class of citizens which, in the opinion of the State, is not adequately represented in the services under the State. The difference between the two clauses matters in practice: Article 15 speaks of special provision, a category wider than reservation and capable of covering scholarships, hostels, fee concessions and preparatory schemes, while Article 16(4) speaks of reservation of appointments or posts, and it is the interplay of that width in the one article and that specificity in the other that structures admission law and service law respectively.
Article 16 contains a further provision that belongs on the map because it fixes the boundary between the national equality of opportunity and State claims based on residence. Clause (2) forbids discrimination in public employment on the ground of residence, among others, and clause (3) then provides that nothing in the article shall prevent Parliament from making a law prescribing, for a class or classes of employment under a State or a local authority, a requirement of residence within that State or Union territory before employment. The power is Parliament’s alone: a State legislature cannot impose a residence requirement for public employment by its own authority. Parliament exercised the power in the Public Employment (Requirement as to Residence) Act, 1957, and the interplay of clauses (2) and (3) explains why domicile based preferences in public employment, a recurring demand in State politics, are constitutionally fragile unless enacted by Parliament, while preferences in admissions to State educational institutions travel by the different route of Article 15’s special provisions and the classification doctrine. The clause is a reminder that the affirmative layer is not a general licence to prefer: each departure from the equality guarantee must find its own enabling text, and the enabling texts are specific about who may act, for whose benefit, and within what limits.
Between the enabling clause and the individual appointment stands the roster, the administrative instrument through which reservation is actually delivered, and no map of the layer is complete without it. An appointing authority does not reserve a percentage in the abstract: it maintains a running roster of points, each point in the cycle assigned to a category in the proportions the orders prescribe, and vacancies are filled against the points as they arise, so that over a full cycle the shares are realised exactly and in any single year approximately. Where a reserved point cannot be filled because no eligible candidate of the category is available, the vacancy is carried forward to a succeeding year rather than filled at once from the general pool, and the carried forward vacancies, called backlog vacancies in the administrative vocabulary, are the subject of periodic special recruitment drives. Article 16(4B), described below, constitutionalised the accounting treatment of that carry forward by directing that such vacancies form a separate class for the purpose of the ceiling. The roster explains why so much service litigation in this field is arithmetic before it is doctrine: disputes about whether a point was correctly assigned, whether a vacancy was a backlog vacancy or a fresh one, and whether a candidate was adjusted against the reserved point or on merit against a general point, determine outcomes that the constitutional argument only frames. It also explains the administrative premium on the certificate machinery described in the lists section, because the roster operates on categories, and the category of every candidate must be fixed, by certificate, before the roster can turn.
The employment clause was then widened by amendment in stages. The Constitution (Seventy-seventh Amendment) Act, 1995 inserted Article 16(4A), permitting reservation in promotion for the Scheduled Castes and the Scheduled Tribes where the State is of the opinion that they are not adequately represented. The Constitution (Eighty-fifth Amendment) Act, 2001 recast that clause to provide for reservation in promotion with consequential seniority, with retrospective effect from the date the original clause took effect. The Constitution (Eighty-first Amendment) Act, 2000 inserted Article 16(4B), providing that reserved vacancies carried forward from a preceding year are to be treated as a separate class of vacancies and not counted with the vacancies of the year in which they are filled when the ceiling on reservation is applied. The Constitution (Eighty-second Amendment) Act, 2000 added the proviso to Article 335 already described. Each of these amendments answered a judgment: the 1995 amendment answered Indra Sawhney’s holding that reservation does not extend to promotions, and the amendments of 2000 and 2001 answered the holdings on carry forward vacancies and seniority that had developed in the case law of the 1990s. The pattern is the one first set in 1951, and it is the mechanism by which the affirmative layer has been built: a judicial statement of the doctrine, a legislative judgment that the doctrine should be otherwise, and a constitutional amendment drafted to the width of the disagreement.
The education side followed a parallel course a decade later. The Constitution (Ninety-third Amendment) Act, 2005 inserted Article 15(5), in force from 20 January 2006, enabling the State to make special provision by law for the advancement of socially and educationally backward classes, the Scheduled Castes and the Scheduled Tribes in so far as such provisions relate to admission to educational institutions, including private educational institutions, whether aided or unaided by the State, other than the minority educational institutions protected by Article 30(1). Under that power Parliament enacted the Central Educational Institutions (Reservation in Admission) Act, 2006, Act 5 of 2007, which received the assent of the President on 3 January 2007 and reserves, in Section 3, fifteen per cent of seats for members of the Scheduled Castes, seven and a half per cent for members of the Scheduled Tribes and twenty seven per cent for the Other Backward Classes in the central educational institutions to which it applies. The Act does not apply to minority educational institutions within Article 30(1), to institutions in the tribal areas of the Sixth Schedule, or to the institutions of excellence, research and national and strategic importance specified in its Schedule. A Constitution Bench upheld the Act and the amendment behind it in Ashoka Kumar Thakur v. Union of India in 2008, sustaining the twenty seven per cent share for the Other Backward Classes in central institutions while requiring the exclusion of the creamy layer. The shares themselves, fifteen, seven and a half and twenty seven, predate the statute: they express the proportions that executive orders had applied in central services following the implementation of the Mandal Commission’s recommendations in the years after 1990, and the statute carried the same proportions from employment into admissions.
The doctrine that disciplines the whole layer is the doctrine of limits, and its canonical statement is the Indra Sawhney reservation ceiling: reservation under Articles 15(4) and 16(4) should not exceed fifty per cent of the seats or posts in a year, save in exceptional circumstances, the advanced sections within a backward class, described in the judgment as the creamy layer, must be excluded, and the identification of backward classes is a justiciable exercise that must rest on evidence rather than assertion. The fifty per cent figure did not begin with Indra Sawhney; it runs back through M. R. Balaji v. State of Mysore, where the Court warned that a special provision should be less than fifty per cent and struck down a Mysore order reserving sixty eight per cent of seats, and the Constitution Bench that decided the Maratha litigation in 2021 declined, unanimously on this point, to refer the ceiling to a larger bench. The newest branch of the layer departs from the group logic of the rest. The Constitution (One Hundred and Third Amendment) Act, 2019 inserted Articles 15(6) and 16(6), enabling reservation of up to ten per cent for economically weaker sections other than the classes already covered, a criterion of individual economic condition rather than group backwardness; the amendment was upheld by a majority of three to two in Janhit Abhiyan v. Union of India, and this series examines it in the posted article on the 103rd Amendment. Its presence on the map matters because it shows the layer’s enabling technique being applied to a different definition of disadvantage, and because challenges to its design, which turn on the exclusion of the classes covered by the older clauses and on the breach of the fifty per cent ceiling, are argued with the same classification tools the older doctrine supplies.
The mechanics of the economically weaker sections branch repay a closer statement, because they show the enabling technique at its most complete. Article 15(6) empowers the State to make special provision for the advancement of any economically weaker section of citizens other than the classes mentioned in clauses (4) and (5), including reservation of seats in educational institutions, private as well as State institutions, other than minority educational institutions, subject to a maximum of ten per cent of the total seats in each category, and expressly in addition to the existing reservations. Article 16(6) performs the same operation for appointments, again capped at ten per cent and again in addition to the existing reservations. In Janhit Abhiyan, the majority upheld the amendment against the basic structure challenge, holding that reservation founded on an economic criterion is not in itself alien to the equality code and that the fifty per cent ceiling, developed for reservation under clauses (4) and (5) of the two articles, is not an inflexible rule of the basic structure that the new clauses transgress; the minority held the exclusion of the classes covered by the older clauses from the new benefit, and the breach of the ceiling, to be destructive of the equality principle. The split itself is instructive for the map. Every judge in the case argued from within the classification doctrine, and the disagreement concerned the doctrine’s limits: whether a class defined by the absence of group backwardness can be a valid class, and whether the ceiling is a rule of application or a rule of constitutional identity. Those are the questions Article 178 and the posted article on the 103rd Amendment examine; the map records that the affirmative layer, seventy years after Article 15(4), was still being extended by the same technique of an enabling clause drafted against the background of the case law, and still generating the same sequence of enactment, challenge and doctrinal restatement.
The lists: who belongs and who decides
Every provision described so far depends on a prior question of identity: which communities count as Scheduled Castes, which as Scheduled Tribes, and which as socially and educationally backward classes. The Constitution answers that question by specification, not by definition. Article 341 empowers the President, after consultation with the Governor of the State concerned, to specify by public notification the castes, races or tribes, or parts of or groups within them, which shall be deemed to be Scheduled Castes in relation to that State or Union territory. Article 342 does the same for Scheduled Tribes. The instruments issued under those powers, the Constitution (Scheduled Castes) Order, 1950 and the Constitution (Scheduled Tribes) Order, 1950, as amended from time to time, contain the lists, and the lists are State specific: a caste specified in relation to one State is not thereby specified in relation to another, and a person who migrates carries the status only if the caste is specified in relation to the State of residence as well. Article 341(2) and Article 342(2) reserve the subsequent power of inclusion and exclusion to Parliament, which may by law include in or exclude from the lists any caste, race, tribe or part or group; neither the President by a second notification, nor a State government, nor a court can add an entry, and the courts have repeatedly declined invitations to treat an unlisted community as equivalent to a listed one, on the ground that the specification is conclusive of identity however sympathetic the claim.
Two features of the Scheduled Castes Order repay attention on the map. The first is paragraph 3, which provides that no person who professes a religion different from the Hindu religion shall be deemed to be a member of a Scheduled Caste. The Order was amended by Parliament to extend the specification to persons professing the Sikh religion in 1956 and to persons professing the Buddhist religion in 1990, so that the condition in its present form admits Hindus, Sikhs and Buddhists and excludes Christians and Muslims, whatever the caste of origin and whatever the social treatment the convert’s community receives. The Scheduled Tribes Order carries no corresponding religion condition. The asymmetry is one of the most contested features of the listing system, and the litigation about it is litigation about the text of the Order and the competence of Parliament, not about the facts of any community’s condition, which no court is empowered to weigh against the specification. The second feature is that the Order specifies castes in relation to areas as well as States, so that a caste may be scheduled in one district of a State and not in another, a refinement that matters at the point of certification, where the issuing authority must match the applicant’s caste, religion and area of residence to the entry in the Order for that State.
The backward classes stand on a different footing, because there is no single constitutional Order for them. Article 340 empowers the President to appoint a commission to investigate the conditions of socially and educationally backward classes and to recommend the steps that should be taken to improve their condition and the grants that should be made for the purpose; the commission’s report is laid before each House of Parliament with a memorandum explaining the action taken on it. The power was exercised twice at the national level. The first Backward Classes Commission, chaired by Kaka Kalelkar, was appointed on 29 January 1953 and reported on 30 March 1955, identifying, in the report’s own enumeration, 2,399 backward castes; the report was not accepted, and its chairman’s covering letter recorded his own reservations about caste as the test of backwardness. The second, chaired by B. P. Mandal, was appointed on 1 January 1979 and reported in December 1980; it estimated the population of the Other Backward Classes at around fifty two per cent of the total, a figure drawn from the census of 1931, the last census to enumerate caste comprehensively, and recommended reservation of twenty seven per cent in central services and public sector undertakings. The announcement by the Government of V. P. Singh on 7 August 1990 that the Mandal recommendations would be implemented produced the litigation decided in Indra Sawhney, and with it the modern doctrine of the affirmative layer. Article 177 of this series examines the Mandal recommendations and their implementation; the point for the map is that backward class identification, unlike Scheduled Caste and Scheduled Tribe specification, grew up through commissions, executive lists and litigation, with the Union and each State maintaining lists of their own.
The dual structure has a practical consequence that candidates meet every year and that the amendments of 2018 and 2021 were fought over. A community may stand in a State’s backward classes list and not in the Central List, or in the Central List with a different entry and area, so that the same person is entitled to reservation in the State’s services and institutions and not entitled in the Union’s, or the reverse. Eligibility is therefore never a property of the person alone: it is a relation among the person, the list and the purpose, and the certificate that proves it must state the list to which it refers. The Maratha litigation turned on exactly this relation, because the Maharashtra statute of 2018 had proceeded on the State’s power to identify a class for its own purposes, and the constitutional question was whether that power survived the centralising text of Article 342A. The restoration in 2021 preserved the dual structure deliberately, writing into clause (3) that a State list prepared for the State’s own purposes may differ from the Central List, so that divergence between the lists is not an anomaly of the system but its enacted design.
That dual structure was placed under strain by the Constitution (One Hundred and Second Amendment) Act, 2018, which made the National Commission for Backward Classes a constitutional body under the new Article 338B and inserted Article 342A, under which the President may, after consultation with the Governor, specify the socially and educationally backward classes in relation to a State or Union territory, with subsequent inclusion and exclusion reserved to Parliament. The amendment was widely read, at the time, as standardising backward class specification on the model of Articles 341 and 342. In the Maratha case, Jaishri Laxmanrao Patil v. Chief Minister of Maharashtra, decided on 5 May 2021, a Constitution Bench held by a majority of three to two that the amendment had that effect in full: after the insertion of Article 342A, the power to identify socially and educationally backward classes lay with the President alone, and the States had lost the power to maintain lists of their own for the purposes of the Constitution. Parliament answered within the year. The Constitution (One Hundred and Fifth Amendment) Act, 2021 amended Article 342A by inserting clause (3), under which every State or Union territory may, by law, prepare and maintain for its own purposes a list of socially and educationally backward classes, entries in which may differ from the Central List prepared for the purposes of the Central Government, and made corresponding changes to Article 338B and to the definition in Article 366(26C). Article 178 owns the Maratha judgment and the doctrine around it. Stated as a sequence, which is all the map requires, the episode is the clearest illustration in the field of the amendment and interpretation cycle: a constitutional amendment in 2018, a judicial reading in 2021 that gave it a wider effect than many of its proponents had stated, and a restoring amendment in the same year that wrote the States’ power into the text expressly.
The listing question has one further branch, on which the law was restated in 2024. In State of Punjab v. Davinder Singh, decided on 1 August 2024, a bench of seven judges held by a majority of six to one, with Trivedi J. dissenting, that a State may sub classify the castes within its Scheduled Castes list for the purpose of reservation, distributing the reserved share among groups within the list on the basis of empirical evidence of differing backwardness and representation, and overruling the contrary holding in E. V. Chinnaiah v. State of Andhra Pradesh. The majority reasoned that Article 341 fixes the identity of the listed castes but does not declare the list a homogeneous class for all purposes, so that a classification within the list must satisfy the Article 14 test like any other. The judgment belongs to Article 178’s territory and is stated here only to complete the map of the lists: specification decides which communities are in the list, Parliament alone can change the list, and, after 2024, a State may differentiate within the list for the distribution of benefits, on evidence, without altering the list itself.
The lists also rest on criteria that are nowhere stated in the Constitution, and the criteria in use have a history. For the Scheduled Tribes, the indications applied in the specification exercises descend from the census practice of the colonial period and were restated by the Advisory Committee on the revision of the lists, chaired by B. N. Lokur, in 1965: indications of primitive traits, distinctive culture, geographical isolation, shyness of contact with the community at large and backwardness. The criteria have been criticised as dated in vocabulary and as difficult to apply to communities that have changed across two generations, and proposals to replace them with contemporary indicators have been examined by official task forces without displacing them from the process. For the Scheduled Castes, the operative test is the historical one embedded in the 1950 Order itself: castes which suffered the disability of untouchability, a test that explains both the religion condition in paragraph 3 and the recurring claims by communities that practise other religions that the same disabilities followed their ancestors across conversion. For the backward classes, the criteria are the commissions’ own, Kalelkar’s and Mandal’s indicator sets, refined by the courts into the requirement that the identification rest on social and educational backwardness supported by evidence. The three criteria sets differ because the three lists do different work: the tribal list identifies communities by their distinctness and isolation, the caste list by their historical exclusion from the ritual order, and the backward classes lists by measurable social and educational condition, and a claim that satisfies one test may fail the others, which is another way of stating why the lists cannot be merged into a single schedule of disadvantage.
Because only Parliament can change the lists, a claim for inclusion follows a settled administrative route before it becomes a Bill. In the procedure the Union Government follows for such claims, a proposal originates with the State government, which examines the claim of the community for inclusion in the list for that State; the proposal is then examined by the Registrar General of India, whose office comments on the ethnographic basis of the claim, and by the concerned national commission, the National Commission for Scheduled Castes or the National Commission for Scheduled Tribes, whose concurrence is sought; the inter ministerial process concludes with the approval of the Cabinet, after which a Bill to amend the Order is introduced in Parliament. Claims that fail at any stage are returned with reasons, and the volume of pending claims at any time is large, because inclusion in a list carries admission, employment and electoral consequences and every excluded community adjacent to a listed one has an incentive to press its case. The route explains a feature of the system that litigants find frustrating: a court asked to declare that a community is, in substance, the same as a listed community must refuse, not because the ethnographic claim is weak, but because the Constitution has assigned the question to a different process, and the refusal is itself a holding about the separation of powers within the framework.
Between the list and the individual stands the certificate. Membership of a listed class is proved, for every practical purpose in this field, by a caste or tribe certificate issued by the revenue or welfare authorities of the State, and the certificate is the document on which admission, appointment, election to a reserved seat and the protection of the criminal statutes all ultimately depend. The Supreme Court addressed the integrity of that machinery in Kumari Madhuri Patil v. Additional Commissioner, Tribal Development in 1994, laying down a procedure for the issue and verification of certificates, including scrutiny committees to examine claims, after cases in which admissions and appointments had been obtained on certificates for castes to which the claimants did not belong. The judgment’s directions made verification a standing feature of the system: a certificate may be scrutinised, a claim may be rejected after inquiry, and benefits obtained on a certificate found false may be withdrawn, with the appointment or admission founded on it treated as void. The machinery distributes a heavy practical burden. The applicant must prove membership of an entry that is defined by State, district and religion, often from records of parents and grandparents maintained by the same administration that issues the certificate, and errors at this stage present themselves later as service disputes, election petitions or prosecutions for false claims, each governed by a different layer’s rules. When commentators describe the framework as difficult to navigate, the certificate is a large part of what they mean: the lists are national in authority and local in proof, and the distance between the Presidential Order and the tehsil office is where a substantial share of the field’s litigation is generated.
The commissions that watch the safeguards
A framework of guarantees and provisions requires a watcher, and the Constitution supplies three. Their history begins with a single office. Article 338 as enacted provided for a Special Officer for the Scheduled Castes and the Scheduled Tribes, appointed by the President, whose duty was to investigate all matters relating to the safeguards provided for those classes and to report to the President on the working of those safeguards. The officer, known as the Commissioner for Scheduled Castes and Scheduled Tribes, produced annual reports that became a principal public record of the framework’s operation. The Constitution (Sixty-fifth Amendment) Act, 1990 replaced the single officer with a multi member National Commission for the Scheduled Castes and the Scheduled Tribes, and the Constitution (Eighty-ninth Amendment) Act, 2003 divided that combined commission in two: the National Commission for Scheduled Castes under Article 338 and the National Commission for Scheduled Tribes under the new Article 338A, the separate tribal commission taking shape in February 2004. The Constitution (One Hundred and Second Amendment) Act, 2018 completed the trio by inserting Article 338B for the National Commission for Backward Classes, replacing the statutory commission that had existed since 1993 under the National Commission for Backward Classes Act, enacted after Indra Sawhney recommended a permanent expert body to examine requests for inclusion in and exclusion from the backward classes lists.
Each constitutional commission is composed of a Chairperson, a Vice Chairperson and three other members, appointed by the President, and each is charged with a matching set of functions: to investigate and monitor all matters relating to the safeguards provided for its class under the Constitution and under law, to inquire into specific complaints of the deprivation of rights and safeguards, to participate and advise in the planning process of socio economic development of the class, to present reports to the President annually and at other times, and to make recommendations for the effective implementation of the safeguards. While investigating, a commission has the powers of a civil court trying a suit, including summoning and enforcing the attendance of persons, requiring the discovery and production of documents and receiving evidence on affidavit. Reports go to the President, who causes them to be laid before Parliament with a memorandum explaining the action taken or proposed on the recommendations and the reasons for any non acceptance; reports touching a State are forwarded to the Governor and laid before the State Legislature on the same pattern. The Union and the States are required to consult the commissions on all major policy matters affecting the classes within their charge, and the consultation requirement inserted for the backward classes commission by Article 338B was qualified by the amendment of 2021, which removed the need for a State to consult the National Commission for Backward Classes when preparing its own list under Article 342A(3).
The reports deserve a word of their own, because they are the framework’s memory. A commission’s annual report reviews the working of every safeguard within its charge: the operation of the criminal statutes in the districts, the filling of reserved vacancies, the distribution of scholarships and hostels, the condition of the lists. Because the reports are laid before Parliament or the State Legislature with the government’s memorandum of action taken, they create a public, dated and answerable record against which the claims made for the framework can be tested, and researchers use them as the standard source for the State by State picture. The commissions may also take up matters on their own motion, a power they exercise where an incident is reported widely and no complaint has been filed, and their inquiries in such matters proceed with the civil court powers described above. Several States maintain commissions of their own for the Scheduled Castes, the Scheduled Tribes or the backward classes under State law, with functions modelled on the national template; a complaint may lie to either level, and the choice between them is a routing question of the same kind as the choice between the criminal statutes, turning on whose report will carry weight with the authority that must act.
The powers define the institution, and they also define its limit: a commission investigates, inquires, reports and recommends, and it cannot quash an order, award compensation as a court can, or prosecute an offender. Its reports bind no government; what they do is place findings on the parliamentary record in a form the government must answer, which is the characteristic enforcement technique of a constitutional watchdog. Confusion among the three commissions is common in public discussion and is worth correcting on the map, because the three bodies have different classes within their charge, different constitutional bases and different consultative roles in the listing process, and a complaint about the denial of a safeguard to a member of a Scheduled Tribe lies to the commission under Article 338A, not to the commission under Article 338.
Around the constitutional trio stands a family of statutory commissions built on the same template for other protected groups: the National Commission for Women, constituted in January 1992 under the National Commission for Women Act, 1990, Act 20 of 1990; the National Commission for Minorities, under the National Commission for Minorities Act, 1992; the National Human Rights Commission, under the Protection of Human Rights Act, 1993; and the National Commission for Protection of Child Rights, constituted on 5 March 2007 under the Commissions for Protection of Child Rights Act, 2005, Act 4 of 2006, whose field this series treats in its child protection cluster. The statutory bodies differ from the constitutional three in their source and their security: they exist because Parliament enacted them, they can be restructured by ordinary legislation, and their reports travel to the Central Government rather than to the President under a constitutional duty. The template, investigation, complaint inquiry, review of safeguards, recommendations and an annual report, is common to both families, and the practical difference between them appears at the point of institutional conflict, where a constitutional commission’s mandate cannot be narrowed by the government of the day without amending the Constitution.
Tribal areas: two schedules and the statutes that follow them
For the tribal areas of the north east and of the mainland, the Constitution does not rely on the general machinery of guarantees and lists. It supplies two distinct systems of governance, in the Fifth Schedule and the Sixth Schedule, and a family of statutes that follow from them. Both schedules are reached through Article 244, which provides that the administration of Scheduled Areas and tribal areas in States other than the Sixth Schedule States shall be governed by the Fifth Schedule, and that the administration of the tribal areas in Assam, Meghalaya, Tripura and Mizoram shall be governed by the Sixth. A Scheduled Area under the Fifth Schedule is an area the President declares to be such by order, applying criteria of tribal preponderance, compactness, economic backwardness and underdevelopment that the scheduling practice has settled; the areas so declared lie across the mainland States with substantial tribal populations. The Sixth Schedule’s districts, by contrast, are named in the schedule itself and may be altered by Parliament by law. The distinction between a declared area and a named district is the first of many between the two systems, and it is characteristic: the Fifth Schedule is a discretionary protective overlay on ordinary State administration, adjusted by executive order, while the Sixth Schedule is a settled constitutional settlement with named institutions, whose alteration requires legislation. The Fifth Schedule applies to the administration and control of Scheduled Areas and Scheduled Tribes in States other than the four Sixth Schedule States of the north east. Its design is supervision from the Union: the executive power of the Union extends to giving directions to a State as to the administration of its Scheduled Areas, the Governor reports to the President annually on their administration, a Tribes Advisory Council advises on matters of welfare and advancement referred to it, and the Governor may direct that a law of Parliament or of the State Legislature shall not apply, or shall apply with modifications, to a Scheduled Area, and may make regulations for the peace and good government of such an area, including regulations prohibiting or restricting the transfer of land by or among members of the Scheduled Tribes. The Sixth Schedule, by contrast, constitutes Autonomous Districts and Autonomous Regions in Assam, Meghalaya, Tripura and Mizoram, each with an Autonomous District Council or Regional Council that is itself a law making body: the councils may make laws on the allotment, occupation and use of land, the management of forests other than reserved forests, the use of water for agriculture, the regulation of jhum and other shifting cultivation, the establishment of village and town committees, the appointment and succession of chiefs and headmen, the inheritance of property, marriage and divorce, and social customs, and they administer justice through village and council courts for cases between parties who belong to Scheduled Tribes. Where the Fifth Schedule governs by advice, direction and gubernatorial exception within the ordinary State structure, the Sixth Schedule governs by devolving legislative power to elected tribal institutions. The legislative competence for the subjects these arrangements touch is distributed between the Union and the States by the Seventh Schedule lists, and the two schedules qualify that distribution for the areas they cover.
Two features of the schedules deserve emphasis because they determine how the protection operates in practice. The Tribes Advisory Council under the Fifth Schedule is a body of not more than twenty members, of whom, as nearly as may be, three fourths are representatives of the Scheduled Tribes in the Legislative Assembly of the State, and its duty is to advise on such matters pertaining to the welfare and advancement of the Scheduled Tribes as may be referred to it by the Governor: it is an advisory organ inside a supervisory design, with no power to bind the State government whose measures it reviews. The Governor’s regulation making power under the same schedule is, by contrast, a power of substance, extending to the prohibition or restriction of transfers of land and the regulation of money lending to members of the Scheduled Tribes, and it may be used to repeal or amend any Act of Parliament or of the State Legislature in its application to the Scheduled Area. Under the Sixth Schedule, the corresponding power of adaptation runs in the other direction: the Governor may direct that an Act of the State Legislature on a subject within the councils’ competence shall not apply to an Autonomous District, or shall apply with exceptions and modifications, while Acts on subjects outside the councils’ list apply as the Governor directs. The contrast completes the answer to how the two areas differ. In a Fifth Schedule area, the ordinary law applies unless the Governor displaces it to protect the tribe; in a Sixth Schedule district, the council’s law governs the enumerated subjects unless the constitutional machinery applies the ordinary law, so that the default rule of the two systems points in opposite directions, and a person advising on land, forest or inheritance in a tribal area must first establish which schedule, if either, governs the place.
Parliament carried the Fifth Schedule’s logic into local government by the Panchayats (Extension to the Scheduled Areas) Act, 1996, Act 40 of 1996, commonly called PESA, which received assent on 24 December 1996. PESA extends Part IX of the Constitution, on panchayats, to the Scheduled Areas referred to in Article 244(1), with the exceptions and modifications set out in Section 4, and its centre of gravity is the Gram Sabha: State legislation on panchayats in Scheduled Areas must be in consonance with the customary law, social and religious practices and traditional management practices of community resources, every Gram Sabha is competent to safeguard and preserve the traditions and customs of the people, their cultural identity, community resources and the customary mode of dispute resolution, and the Gram Sabha or the panchayat at the appropriate level must be consulted before land is acquired in a Scheduled Area and before persons affected by a project are resettled, with its recommendation made mandatory in advance of the grant of a prospecting licence or mining lease for minor minerals. Article 175 of this series owns PESA; on the map, the statute is the bridge between the Fifth Schedule’s protective supervision and the affirmative layer’s recognition of rights.
The recognition strand reaches its fullest form in the Scheduled Tribes and Other Traditional Forest Dwellers (Recognition of Forest Rights) Act, 2006, Act 2 of 2007, which entered into force on 31 December 2007 and is administered by the Ministry of Tribal Affairs as the nodal ministry. The Act recognises and vests forest rights in forest dwelling Scheduled Tribes and other traditional forest dwellers, reversing in statutory form the position under the forest laws that occupation of forest land without a recorded title was encroachment: it recognises rights of occupation in land under individual or common occupation for habitation or self cultivation, subject to a ceiling and to the condition of occupation on 13 December 2005, community rights of use and access, rights over minor forest produce, and the right to protect, regenerate, conserve and manage community forest resources, with the Gram Sabha as the authority that initiates the process of determining rights and the other traditional forest dweller category defined by residence in the forest for three generations, a generation being twenty five years, before that December date. The statute is examined in full at the Forest Rights Act, 2006, the third article of this batch. Its place on the map is distinctive: it is affirmative provision in the form of recognition, a statute that does not reserve a share of a benefit but declares that a class of persons holds rights the record had failed to state, and its disputes are disputes of evidence and procedure before committees and the Gram Sabha rather than disputes of classification before the constitutional courts.
The wider perimeter: disability, gender identity, age and the household
The core of the field is built around caste and tribe, because the Constitution’s own specification machinery is built around them, but the same three-layer technique has been extended to other forms of disadvantage, and a complete map marks the extensions. For persons with disabilities, the Rights of Persons with Disabilities Act, 2016, Act 49 of 2016, combines all three layers in one statute: it prohibits discrimination, it creates enforcement machinery through the Chief Commissioner and State Commissioners for persons with disabilities, and in Section 34 it makes affirmative provision by reserving not less than four per cent of the total number of vacancies in the cadre strength of government establishments for persons with benchmark disabilities, distributed at one per cent for each of the specified groups of disability, with Section 33 providing for the identification of posts suitable for each category. Disability reservation also differs from caste reservation in its geometry, a point the wider series develops: it operates horizontally, cutting across the vertical categories of the general, Scheduled Caste, Scheduled Tribe and backward class compartments, so that a candidate with a benchmark disability is adjusted within the category to which the candidate otherwise belongs.
For transgender persons, the Supreme Court’s judgment in National Legal Services Authority v. Union of India, delivered on 15 April 2014, recognised transgender persons as a third gender, held that gender identity rests on self identification, and directed that transgender persons be treated as a socially and educationally backward class for the purposes of admission and appointment. Parliament legislated in the Transgender Persons (Protection of Rights) Act, 2019, Act 40 of 2019, which entered into force on 10 January 2020: the Act prohibits discrimination against a transgender person across education, employment, healthcare, access to goods and services, movement, residence and the holding of office, provides for a certificate of identity issued through the District Magistrate, and constitutes a National Council for Transgender Persons to advise the Central Government. The statute has been debated on the point where it departs from the judgment, the requirement of a revised certificate following surgery for recognition as male or female, and the debate is an instance of the framework’s recurring pattern, a judicial statement of principle followed by a legislative text that implements the principle in a narrower administrative form.
The disability statute’s definition work is worth a sentence, because it shows the perimeter using the core’s techniques. A person with benchmark disability is defined by reference to a threshold of not less than forty per cent of a specified disability, certified by the certifying authority, and the reservation in Section 34 attaches to that certified status just as caste reservation attaches to the caste certificate, with the same practical consequence that the certificate, and the medical board that issues it, become a forum of the framework in their own right. The Act also extends the prohibition side beyond the State, providing that no establishment shall discriminate against any person with disability on the ground of disability unless the act or omission is a proportionate means of achieving a legitimate aim, a formulation that imports the classification structure of Article 14 into a statutory duty owed by private employers.
For senior citizens, the Maintenance and Welfare of Parents and Senior Citizens Act, 2007, which received assent on 29 December 2007, works mainly on the affirmative and protective side of the line: it obliges children and, in defined circumstances, relatives to maintain parents and senior citizens who cannot maintain themselves, creates Maintenance Tribunals for summary orders, and requires State Governments to provide old age homes and medical support. The women’s and children’s clusters, which this series treats in their own articles and which include the Protection of Women from Domestic Violence Act, 2005, Act 43 of 2005, share the field’s protective logic while resting on distinct constitutional foundations in Articles 15(3), 21 and 39, and they are mapped there rather than here. What unifies the perimeter with the core is administrative: the statutes of the perimeter are administered by the same small group of ministries, and their entitlements are claimed through the same district machinery as the core statutes, so that a district officer’s desk may hold, in one week, a caste certificate appeal, a forest rights claim, a disability certificate and a maintenance application, each governed by a different layer of the same framework.
The administrative map deserves a paragraph of its own, because knowing which ministry owns a statute is part of knowing how the framework works. The Ministry of Social Justice and Empowerment is the nodal ministry for the Scheduled Castes, the Other Backward Classes, senior citizens, persons with disabilities through its department for their empowerment, and the protection and rehabilitation statutes described in this guide. It began as the Ministry of Welfare, formed in 1985 by drawing together subjects that had been handled by the Ministry of Home Affairs and others, and it took its present name in May 1998. In October 1999 the Tribal Development Division was moved out of it to form the Ministry of Tribal Affairs, which is the nodal ministry for the Scheduled Tribes, for the Forest Rights Act and for the administration of the Scheduled Areas, and in January 2007 the minorities division left to form the Ministry of Minority Affairs. The pattern of the reorganisation is itself instructive: as a class within the framework acquired a distinct statutory programme, it acquired a distinct ministry, and the commissions described earlier watch the ministries from outside the executive structure altogether.
Overlaps, gaps and the two arguments the framework attracts
No map of this field is complete without the points at which the layers fail to meet, because the gaps are where the disputes collect. The first gap is evidentiary: the framework distributes by group membership, and its distributional questions, which castes are backward, which groups within a list are least represented, what share a class should receive, all demand data, yet the decennial census has not enumerated caste comprehensively since 1931, and commissions from Mandal onwards have worked from that ageing base supplemented by surveys of varying quality. The Supreme Court’s insistence, from Indra Sawhney through Nagaraj to the Maratha case, on quantifiable data as the price of a sustainable classification is the judicial response to the gap, and the political argument about a caste census, which Article 177 takes up in connection with the Rohini Commission and the enumeration debate, is the legislative response. The gap is not for want of collection alone: the Socio Economic and Caste Census of 2011 gathered caste returns across the country, and those returns were not published, the government stating that the data contained errors on a scale that made release unwise, so that the framework’s most consequential distributional decisions continue to be argued from a 1931 base, from commission estimates of varying vintage and from State surveys conducted for particular litigations. A reader evaluating any statistic in this field should therefore ask the questions the courts ask: who enumerated, when, by what criteria, and for what proceeding was the figure produced. Figures in this guide are confined to the shares and scales the texts themselves prescribe, and to estimates expressly attributed to the commissions that produced them, for exactly that reason. The second gap is the one the counter reading in every serious account of this field identifies: the distance between the criminal layer’s text and its outcomes. Commentators across positions describe a body of protective law that is extensive on paper and weakly enforced in the police station and the trial court, citing delayed registration, low conviction and the attrition of witnesses; the figures most often quoted are those of the National Crime Records Bureau’s Crime in India series, and they are contested in detail, but the pattern they describe is not seriously disputed. The third gap is internal to the affirmative layer: benefits distributed to a class are not necessarily distributed within it evenly, and the argument that advanced sections and better placed communities within the lists capture a disproportionate share is the argument that produced the creamy layer doctrine within backward classes and, in the Davinder Singh litigation, the judicial acceptance of sub classification within the Scheduled Castes list.
The sub classification development belongs to the third gap. Within the Davinder Singh majority, four of the seven judges wrote separately to express the view that the principle of the creamy layer, which excludes the advanced sections of a backward class from its benefits, should in time be applied within the Scheduled Castes and the Scheduled Tribes as well, so that the benefits of reservation reach the most disadvantaged within the lists; the view was expressed as a direction for future policy rather than as a condition of the sub classification the Court upheld, and the creamy layer’s application to those classes in admission and appointment generally remains, after the judgment, an open question of the kind Article 178 examines. The episode shows the framework’s internal correction mechanism at work: an argument that begins as a critique of distribution within a class becomes, through litigation, a doctrine of classification within the class, and the evidentiary burden the Court attached to it, empirical demonstration of inter se backwardness and inadequacy of representation, returns the argument to the data gap with which this section began.
Is the weakness of the civil rights statute a failure of enforcement or of drafting?
Both, on the record, and the drafting half of the diagnosis is the one usually missed. The Protection of Civil Rights Act, 1955 carries penalties whose monetary scale was set by the amending Act of 1976, with a fine for a first offence under Section 3 that runs between one hundred and five hundred rupees.
A deterrent fixed in the prices of that period deters little in any later period, and no amendment has revised the scale since.
The enforcement half of the diagnosis is equally well supported: the Elayaperumal Committee, reporting in 1969, had documented the gap between the abolition declared in Article 17 and the practice in the districts a decade before the renaming amendment, and every succeeding review has described the same mechanics of failure, complaints not registered, offences compounded informally, and prosecutions withdrawn. What the drafting point adds is an explanation of why the statute fell into disuse even where the machinery was willing: a claim under the 1955 Act promises a small penalty at the end of a criminal trial, while the same facts pleaded under the Atrocities Act of 1989 carry enhanced sentences and statutory relief for the victim, so complainants and their advisers rationally choose the later statute, and the earlier one atrophies. The lesson generalises across the criminal layer. A prohibition statute in this field lives or dies by the relationship between its penalties, its procedure and the alternatives available to the complainant, and a framework that adds new statutes without revising the penalties of the old ones will find its older instruments abandoned in practice while remaining in force on paper.
The second argument the framework attracts is the argument about the affirmative layer’s duration, and it is conducted between two positions that must be stated in their own terms because this guide adjudicates neither. The first position holds that affirmative provision was designed as a corrective of limited duration, that the Constitution itself expressed the instinct in Article 334’s original ten year sunset for legislative reservation, that decades of operation have produced substantial representation in legislatures, services and admissions, and that a permanent distributive system organised by caste entrenches the categories it was meant to dissolve; proponents of this position point to the repeated extension of the Article 334 sunset, seven times across seventy years, as evidence that the corrective has become a fixture, and to the concentration of benefits within lists as evidence that the fixture no longer reaches its intended beneficiaries evenly. The second position holds that the conditions the corrective addresses, measured by representation in the higher services, in the judiciary, in university faculties and in ownership of land and enterprise, persist at levels far from parity, that the sunset in Article 334 governs legislative seats alone and says nothing about the duration of Articles 15 and 16 provision, and that withdrawal of the corrective before the underlying distribution of educational and economic capital changes would freeze an unequal order rather than complete an equal one. The constitutional text permits the second position’s premise: Articles 15(4) and 16(4) carry no sunset, and the courts, while policing the ceiling, the creamy layer and the demand for data, have never held that the passage of time itself invalidates a special provision. What the text does not settle, and what the evidence genuinely leaves disputed, are the empirical questions on which the argument turns: how representation has moved within each listed class, how benefits are distributed inside the lists, and what the counterfactual distribution would have been. Those questions belong to the census debate and to Articles 177 and 178, which take up the Mandal settlement and the ceiling doctrine in full; the map’s duty is to show where the argument lives, which is in the affirmative layer, and what would settle it, which is data the framework has not generated since 1931.
One set of facts, four proceedings: how the layers interact
The layers are easiest to see in combination, and a composite example, assembled from the recurring patterns of the reported cases and stated without reference to any individual matter, shows why the routing question is not academic. A family belonging to a Scheduled Caste cultivates a plot at the edge of a village and draws water from the common well. The village council announces that the family will not be permitted to draw from the well, because the family’s son has taken a job in the town against the council’s wishes. When the family draws water regardless, its members are assaulted, the crop on the plot is damaged, and the son is told that the caste certificate on which his appointment rests will be challenged. One set of facts has generated four distinct legal proceedings, each in a different layer, and the family’s advisers must run all four correctly because no one of them can deliver the remedies of the others.
The exclusion from the well is the enforcement of a disability on the ground of untouchability, and it belongs to the Protection of Civil Rights Act: the complaint is a criminal complaint, the forum is the magistrate’s court, and the outcome, if the charge is proved, is the punishment of the persons enforcing the disability and, where the village acts collectively, the Act’s collective fine. The assault and the damage to the crop belong to the Atrocities Act, provided the offenders are not members of a Scheduled Caste or a Scheduled Tribe and the acts fall within the Act’s catalogue of offences: the case proceeds before the special court, the victim’s family becomes entitled to the relief the 1995 Rules prescribe as the case advances, and the prosecution carries the enhanced sentences of the 1989 Act rather than the small penalties of the 1955 Act. The challenge to the son’s certificate belongs to neither statute: it is a question of the lists, governed by the Scheduled Castes Order for the State, by the scrutiny committee procedure that Kumari Madhuri Patil required, and by service law so far as the appointment is concerned, and it will be decided on records and entries rather than on the events at the well. The plot itself may raise a fourth proceeding: if the land is claimed as ancestral occupation, or if the village asserts a competing right, the dispute moves into the revenue and, where forest land is involved, the Forest Rights Act machinery of Gram Sabha claim and committee determination. A single lawyer’s letter that treats the affair as one wrong will misfire in at least three of the four forums. A complaint under the 1955 Act cannot recover the crop; the atrocities prosecution cannot validate the certificate; the scrutiny committee cannot punish the assault. The layers interact, in the sense that the findings in one proceeding become evidence in another and the family’s position is the sum of all four, but they do not merge, and the discipline of keeping them separate is the practical content of the thesis this article has advanced.
A fifth thread completes the picture. If the police delay registering the case, or the relief under the Rules is not paid, the family may complain to the National Commission for Scheduled Castes, which can summon the district record, inquire into the failure of the safeguard and report upon it, and whose intervention frequently moves the district machinery in a way the family’s own representations did not. The commission’s proceeding is not a fifth remedy competing with the four: it delivers no conviction, no title and no appointment. It is the watchdog layer operating exactly as designed, converting an individual failure into a finding on the parliamentary record and a direction of attention from the government that must answer the report. Composite cases of this kind, in which a single village dispute engages the civil rights statute, the atrocities statute, the certificate machinery, the land machinery and the commission, are not exotic. They are the ordinary texture of practice in this field, and they are the reason the map, rather than any single statute taken alone, is the unit of understanding the field requires.
Routing a question: three layers, one question
The practical use of the map is routing. A person with a grievance in this field, and anyone advising such a person, begins not with the identity of the wrongdoer or the wording of a complaint but with the situation, and the situation selects the layer, the layer selects the text, the text selects the authority and the forum, and the forum determines the remedy that can actually be obtained. The question of which authority to approach, which is the question people in this field ask most often and answer worst, dissolves once the layer is fixed. A criminal complaint goes to the police and from there to the magistrate or the special court, and no representation to a ministry can substitute for the first information report that starts the case. A complaint about the denial of a safeguard goes to the commission for the class concerned, whose inquiry can summon the record and whose report the government must answer, but the commission cannot order the appointment or admission that only the appointing authority and, in challenge, the High Court can deliver. A claim to a benefit goes to the authority that administers the benefit, the admitting institution, the appointing authority or the Gram Sabha and committees under the forest statute, and it is decided on the certificate, the roster and the list. A challenge to the list itself goes, in the end, to Parliament, by the administrative route described in the lists section, and no authority below Parliament can grant it. Each authority is powerful inside its layer and powerless outside it, and the map’s routing table is simply that proposition set out by situation. Professionals who keep source material, interview notes and case papers in order rely on disciplined records, and the habits described in accounts of how journalists use VaultBook to protect sources illustrate the same discipline from the documentation side: a claim in this field succeeds or fails on the record assembled for it, the certificate, the roster entry, the first information report, and keeping that record accurate and protected is part of the work. The table below states the routing for the five situations that recur most often. It is a statement of the general legal position, not advice on any individual case, and a person pursuing a claim needs advice on the facts of that claim, because limitation, local rules and the evidence available change outcomes that the table cannot foresee.
| Situation | Layer | Statute or constitutional provision | Authority | Forum | Remedy |
|---|---|---|---|---|---|
| Denial of access to a shop, well, temple, road or public service on the ground of untouchability | Criminal prohibition | Protection of Civil Rights Act, 1955, Articles 17 and 35 | Police of the district, State Government | Magistrate’s court | Conviction and punishment of the offender, collective fine where a community enforces the denial |
| An act of violence, insult or humiliation against a member of a Scheduled Caste or a Scheduled Tribe by a person outside those classes | Criminal prohibition | Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989 | Police, District Magistrate, special public prosecutor | Special Court under the Act | Conviction with enhanced sentence, and statutory relief and rehabilitation for the victim |
| A claim to a reserved seat in admission or a reserved post in public employment | Affirmative provision | Articles 15(4), 15(5), 16(4) and the Central Educational Institutions (Reservation in Admission) Act, 2006 | Admitting institution, appointing authority, State or Union Government | Administrative process, then the High Court under Article 226 | Correct application of the roster and the list, admission or appointment where the claim is made out |
| A claim to land or forest occupied or used by a forest dwelling community | Affirmative provision by recognition | Forest Rights Act, 2006, Fifth and Sixth Schedules | Gram Sabha, Sub Divisional and District Level Committees | Committees under the Act, then the High Court | Recognition and vesting of the forest right and issue of title |
| A question whether a community is on a list, or whether a benefit reaches a group within a list | Constitutional specification | Articles 341, 342 and 342A, the Presidential Orders, Article 340 | President and Parliament for the lists, State Government for sub classification | Certification authorities, then the constitutional courts | Inclusion or exclusion by law of Parliament; within a list, sub classification on evidence by the State |
The examinable core: how this field is tested and taught
For the aspirant reader, whose payload the brief for this article names expressly, the framework’s examinable weight sits in the constitutional layer and the commissions, and the three layer structure is the most reliable organising device for an answer in this field. The questions that recur across examinations follow the map’s joints. The specification questions ask who decides membership of the lists and by what instrument, and they are answered from Articles 341, 342 and 342A with the Orders of 1950 and the rule that Parliament alone amends them. The doctrine questions ask how reservation is reconciled with equality, and they are answered from the classification test, the facet reading in N. M. Thomas and Indra Sawhney, the ceiling and the creamy layer. The institutional questions ask how the three commissions differ, and they are answered from Articles 338, 338A and 338B with their appointment, functions and reporting pattern, against the statutory commissions that imitate them. The amendment questions ask what changed in 2018 and 2021, and they are answered as a sequence: constitutional status and presidential specification, the Maratha reading, the restoration of the States’ lists. An answer that begins by locating its question in a layer, then states the governing text, then the doctrine and the institution, will cover the examinable ground in the order the examiner’s own marking scheme follows, because the scheme, like this guide, is built on the framework’s actual structure rather than on the chronology of the syllabus. The errors to avoid are the three this guide has corrected throughout: describing reservation as an exception to equality, treating the three commissions as interchangeable, and stating the law of backward class identification as it stood between the Maratha judgment and the amendment of 2021 without the restoration that followed.
Verdict: the layer is the map
The framework this guide has mapped is often described as a maze, and the description is unfair to the design. The design is legible once its organising idea is seen: the Constitution declares the guarantee and names the classes, the criminal statutes police the boundary of conduct that the guarantee forbids, and the affirmative provisions distribute opportunity to the classes the guarantee names, with commissions watching all three operations and reporting to the President and to Parliament. The maze appears only when the layers are collapsed into one another, when a criminal complaint is asked to deliver an appointment, when a roster dispute is pleaded as an atrocity, or when a question about a list is argued as if it were a question about a share. Each of those collapses is avoidable, and avoiding them costs nothing but the discipline of asking, at the start of any question in this field, which layer the question belongs to. The series thesis for this cluster holds that a field composed of constitutional provisions, criminal statutes and administrative orders cannot be read as one body of law, and this article’s answer to its One Test is the same proposition stated as a skill: identify the layer, and the text, the authority, the forum and the remedy follow. One further test of the map is worth applying before leaving it. Take the questions with which public argument in this field is conducted and ask of each which layer it belongs to. Whether a caste should be added to a list is a specification question, and it will be decided in the administrative route to a parliamentary amendment, not in a courtroom and not by a commission. Whether a share is too large is a doctrine question, and it will be decided by the classification test, the ceiling and the data, in the constitutional courts. Whether an offence was committed is a criminal question, and it will be decided on evidence in a trial in which the complainant’s belief about the framework’s fairness is irrelevant. Whether a commission acted is an institutional question, answered by its reports and the government’s memoranda upon them. Argument in this field goes wrong in a predictable way: a proposition established in one layer is carried into another where it proves nothing, so that the undoubted fact of an atrocity is offered as proof that a share should rise, or the undoubted antiquity of a list entry is offered as proof that no offence occurred. The layers are connected by subject and separated by logic, and holding both truths at once is the whole of the discipline this guide has tried to teach.
The articles that follow in this cluster take up the individual instruments in the depth they require: the Atrocities Act and its amendment, the Forest Rights Act, PESA, the manual scavenging statute, the Mandal settlement and the Indra Sawhney doctrine, and the Protection of Civil Rights Act in its turn. The map is drawn so that each of those treatments can be read as the exploration of one layer, one text and one machinery, in a field whose whole structure is three layers deep.
Frequently Asked Questions
Q: Which laws form India’s social justice framework?
The framework has three layers. The constitutional layer is the equality code in Articles 14 to 18, Articles 23, 29, 46 and 335, the specification provisions in Articles 341, 342 and 342A, the commission provisions in Articles 338, 338A and 338B, the reservation provisions in Articles 330 to 334 and 243D and 243T, and the Fifth and Sixth Schedules. The criminal layer is principally the Protection of Civil Rights Act, 1955 and the Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989, supported by the Bonded Labour System (Abolition) Act, 1976 and the manual scavengers statute of 2013. The affirmative layer is the enabling clauses of Articles 15 and 16 with the statutes and orders made under them, and the recognition statutes for tribal self government and forest rights.
Q: What does Article 17 of the Constitution abolish?
Article 17 abolishes untouchability. Its text is short: untouchability is abolished and its practice in any form is forbidden, and the enforcement of any disability arising out of untouchability shall be an offence punishable in accordance with law. The Article does not define untouchability, and no statute defines it either; the courts read the term in its historical sense, as the imposition of social disabilities on persons by reason of their birth in specified castes. The Article is enforceable against private persons as well as the State, and its penal clause is the foundation on which Parliament enacted the Untouchability (Offences) Act, 1955, renamed the Protection of Civil Rights Act, 1955 by the amendment of 1976.
Q: Which constitutional commissions protect disadvantaged groups?
Three commissions have constitutional status. The National Commission for Scheduled Castes sits under Article 338, the National Commission for Scheduled Tribes under Article 338A, inserted by the Eighty-ninth Amendment in 2003 when the earlier combined commission was divided, and the National Commission for Backward Classes under Article 338B, inserted by the One Hundred and Second Amendment in 2018. Each consists of a Chairperson, a Vice Chairperson and three other members appointed by the President. Each investigates and monitors the safeguards for its class, inquires into specific complaints, advises on planning and policy, and reports to the President, whose reports are laid before Parliament with the government’s memorandum of action taken. Beside them stand statutory commissions for women, minorities, human rights and child rights, which rest on Acts of Parliament rather than on the Constitution.
Q: What is the difference between Scheduled Castes, Scheduled Tribes and backward classes?
The three expressions denote classes identified by different tests and listed by different instruments. The Scheduled Castes are the castes specified in the Constitution (Scheduled Castes) Order, 1950 under Article 341, historically the castes that suffered the disability of untouchability, and subject to the religion condition in paragraph 3 of that Order. The Scheduled Tribes are the tribes specified in the Order of 1950 under Article 342, identified by indications of distinctive culture, geographical isolation and backwardness, with no religion condition. The backward classes, administered as the Other Backward Classes beside the two scheduled classes, are socially and educationally backward classes identified through commissions under Article 340 and through Union and State lists, without a single constitutional Order; the expression socially and educationally backward classes is the constitutional term for them.
Q: Which social justice laws are criminal and which are welfare?
The criminal laws are those that make conduct an offence. The Protection of Civil Rights Act, 1955 punishes the enforcement of disabilities arising out of untouchability. The Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act, 1989 punishes a catalogue of offences against members of those classes by persons outside them. The Bonded Labour System (Abolition) Act, 1976 and the manual scavengers statute of 2013 also carry offences, though both join rehabilitation to prohibition. The welfare and affirmative side of the framework works by entitlement rather than punishment: the reservation provisions of Articles 15 and 16 and the statutes made under them, the Central Educational Institutions (Reservation in Admission) Act, 2006, and the recognition statutes, PESA in 1996 and the Forest Rights Act, 2006, which vest rights and shares instead of prosecuting offenders. Several statutes, the 2013 Act among them, stand on both sides of the line.
Q: What did the amendments on backward classes commissions change?
Before 2018 the National Commission for Backward Classes was a statutory body, created by the National Commission for Backward Classes Act, 1993 after Indra Sawhney recommended a permanent expert body for the lists. The Constitution (One Hundred and Second Amendment) Act, 2018 repealed that footing and made the commission a constitutional body under the new Article 338B, with the composition, functions and civil court powers of the older commissions, and it inserted Article 342A, under which the President specifies the socially and educationally backward classes for each State and Union territory. A Constitution Bench in the Maratha case in 2021 read Article 342A as removing the States’ power to identify backward classes. The Constitution (One Hundred and Fifth Amendment) Act, 2021 answered by inserting clause (3) in Article 342A, expressly preserving every State’s power to prepare and maintain its own list for its own purposes, and by qualifying the commission’s consultation role for State lists accordingly.
Q: Who decides which communities are listed as Scheduled Castes or Tribes?
The President decides in the first instance, and Parliament decides every change. Under Article 341 the President, after consultation with the Governor of the State concerned, specifies by public notification the castes deemed to be Scheduled Castes in relation to that State, and Article 342 does the same for Scheduled Tribes; the Orders of 1950 contain the lists. Under clause (2) of each Article, Parliament alone may by law include a community in the list or exclude one from it. A State government cannot add an entry, the President cannot revise the notification, and a court cannot treat an unlisted community as equivalent to a listed one. Claims for inclusion travel administratively from the State government through the Registrar General of India and the concerned national commission to the Cabinet before a Bill is introduced.
Q: How do Fifth and Sixth Schedule areas differ?
The Fifth Schedule governs Scheduled Areas in the mainland States through supervision inside the ordinary State structure: the Union may direct the State on administration, the Governor reports annually to the President, a Tribes Advisory Council advises the Governor, and the Governor may disapply or modify laws for a Scheduled Area and make regulations, including restrictions on land transfer among members of the Scheduled Tribes. The Sixth Schedule governs the tribal areas of Assam, Meghalaya, Tripura and Mizoram through Autonomous District and Regional Councils that hold legislative power of their own over land, forests other than reserved forests, water for agriculture, shifting cultivation, village administration, inheritance, marriage, divorce and social customs, and that administer justice through their own courts for cases between members of Scheduled Tribes. The Fifth system protects by exception from the ordinary law; the Sixth devolves law making to tribal institutions.
Q: What is the difference between Article 15 and Article 16?
Article 15 is the general guarantee against discrimination in civil life, and Article 16 is the specific guarantee for public employment. Article 15(1) forbids the State to discriminate against any citizen on grounds only of religion, race, caste, sex or place of birth, and Article 15(2) protects access to shops, restaurants, hotels, wells, tanks, roads and places of public resort. Article 16(1) guarantees equality of opportunity for all citizens in matters of employment or appointment to any office under the State, and Article 16(2) adds descent and residence to the forbidden grounds. Their enabling clauses differ as well: Article 15(4) permits special provision, a wide category, while Article 16(4) permits reservation of appointments or posts for a backward class not adequately represented in the services, so admission law is built on the first and service law on the second.
Q: What does Article 340 empower the President to do?
Article 340 empowers the President to appoint a commission, consisting of such persons as the President thinks fit, to investigate the conditions of socially and educationally backward classes within the territory of India and the difficulties under which they labour, and to recommend the steps that should be taken by the Union or any State to remove those difficulties and to improve their condition, together with the grants that should be made and the conditions of those grants. The commission’s report, with a memorandum explaining the action taken on it, is laid before each House of Parliament. The power has been exercised twice at the national level: the Kalelkar Commission, appointed in 1953 and reporting in 1955, and the Mandal Commission, appointed in 1979 and reporting in 1980, whose recommendations produced the modern reservation settlement for the Other Backward Classes.
Q: What does Article 29(2) guarantee?
Article 29(2) guarantees that no citizen shall be denied admission into any educational institution maintained by the State or receiving aid out of State funds on grounds only of religion, race, caste, language or any of them. The guarantee belongs to every citizen individually, whether or not the citizen belongs to a minority, although the Article sits in the group of provisions protecting minorities, clause (1) of the same Article securing to any section of citizens with a distinct language, script or culture the right to conserve it. The provision mattered decisively in 1951, when the Supreme Court in the Champakam Dorairajan case held the Madras communal order of seats to violate Articles 15(1) and 29(2), and the First Amendment answered by drafting Article 15(4) to operate notwithstanding both provisions, which is why the text of clause (4) names Article 29(2) expressly.
Q: What is the difference between a special provision and a reservation?
A reservation is one species of special provision. Reservation sets apart a defined share of seats or posts for a class and fills that share only from the class, operating through rosters in services and through seat matrices in admissions. A special provision is the wider constitutional category used in Articles 15(3), 15(4) and 15(5): it covers any measure for the advancement of the class, including scholarships, hostels, fee concessions, preparatory and coaching schemes, age and attempt relaxations and the establishment of institutions, most of which reserve nothing. The distinction carries legal consequences. A measure that is a special provision but not a reservation is not counted against the fifty per cent ceiling that the courts apply to reservation, and challenges to it are tested by the classification doctrine without the ceiling analysis. Drafting a benefit as one or the other is therefore a substantive choice, not a verbal one.
Q: How does reservation in legislatures differ from reservation in jobs and education?
Legislative reservation reserves constituencies, not shares of a service. Under Articles 330 and 332, seats are reserved for the Scheduled Castes and the Scheduled Tribes in the Lok Sabha and the State Legislative Assemblies in proportion, as nearly as may be, to their population; the reserved seat is filled by election from the general electorate of the constituency, so the representative is chosen by all voters and sits for the whole constituency. The provision is time limited by Article 334 and has been extended by amendment at ten year intervals, and by the Constitution (One Hundred and Fourth Amendment) Act, 2019 to 25 January 2030. Reservation in jobs and education, by contrast, distributes posts and seats among categories of candidates under Articles 15 and 16, has no constitutional sunset, and is disciplined by doctrine, the fifty per cent ceiling, the creamy layer and the requirement of data, rather than by an expiry date. No legislative reservation exists for the backward classes.
Q: How are seats reserved in panchayats and municipalities?
Articles 243D and 243T reserve seats in every panchayat and municipality for the Scheduled Castes and the Scheduled Tribes in proportion, as nearly as may be, to their population in the area, and reserve not less than one third of the total number of seats for women, including within the seats reserved for the two scheduled classes, so that women’s reservation operates across categories. The same pattern extends to the offices of chairpersons, which are reserved for the scheduled classes and for women in the manner State law provides. Reserved seats are allotted by rotation to different constituencies at successive elections. State law may also reserve seats and offices for backward classes, and the Supreme Court in K. Krishna Murthy in 2010 upheld the scheme while requiring quantifiable data for backward class reservation in local bodies and observance of the fifty per cent ceiling in the aggregate. Unlike legislative reservation under Article 334, the local provisions carry no sunset clause.
Q: Does a person remain a member of a Scheduled Caste after converting to Christianity or Islam?
Under the Constitution (Scheduled Castes) Order, 1950, no. Paragraph 3 of the Order provides that no person who professes a religion different from the Hindu religion shall be deemed to be a member of a Scheduled Caste. Parliament amended the Order to extend it to persons professing the Sikh religion in 1956 and the Buddhist religion in 1990, so the specification covers Hindus, Sikhs and Buddhists, and a person who converts to Christianity or Islam is not deemed a member of a Scheduled Caste for the purposes of the Order, whatever the caste of birth and whatever the social treatment the person’s community continues to receive. The Scheduled Tribes Order contains no religion condition, so conversion does not affect Scheduled Tribe status. Challenges to the paragraph 3 line are challenges to the text of the Order and to Parliament’s competence, and the courts have treated the condition as conclusive within the Order’s scheme.
Q: Does Scheduled Caste status follow a person who moves to another State?
Not by itself. The Scheduled Castes and Scheduled Tribes lists are specified in relation to each State and Union territory, and an entry operates only where it is specified: a caste scheduled in relation to one State is not scheduled in relation to another unless the Order contains an entry for it there as well, and some entries are further limited to specified districts or areas within a State. A person who migrates therefore holds the status, for the purposes of the State of destination, only if the caste is entered in that State’s list, and the certificate that proves status must be issued with reference to the list of the State whose benefit is claimed. The Supreme Court has sustained this State specificity against the argument that status, once acquired in the State of origin, travels with the person for all Union and State purposes, reasoning that the specification in Articles 341 and 342 is by its terms in relation to a State.
Q: Why are minority educational institutions left out of the reservation provisions?
Because Article 30(1) guarantees religious and linguistic minorities the right to establish and administer educational institutions of their choice, and the enabling clauses were drafted to stop at that guarantee. Article 15(5), inserted by the Ninety-third Amendment in 2005, empowers special provision for admission to educational institutions including private institutions, aided or unaided, but expressly other than the minority educational institutions referred to in Article 30(1). The Central Educational Institutions (Reservation in Admission) Act, 2006 repeats the exclusion, as does Article 15(6) for the economically weaker sections. The design treats minority administration of an institution as itself a protected constitutional interest, of a different kind from the backwardness the reservation provisions address, and the courts have read compulsory reservation in a minority institution as a direct burden on the Article 30 choice. The exclusion is therefore textual and structural, not an oversight, and it applies to aided and unaided minority institutions alike.
Q: What can the constitutional commissions do when a safeguard is violated?
A commission can investigate, inquire, summon and report, and it cannot punish or award relief itself. Each of the commissions under Articles 338, 338A and 338B may inquire into a specific complaint that a person has been deprived of rights and safeguards, and while investigating it holds the powers of a civil court trying a suit: summoning and enforcing the attendance of any person, requiring the discovery and production of documents and receiving evidence on affidavit. It may take up a matter on its own motion, call for records from the authority concerned, and place its findings and recommendations before the President, who causes the report to be laid before Parliament with a memorandum explaining the action taken or proposed and the reasons for any recommendation not accepted. The enforcement is institutional rather than coercive: the government must answer the report on the record, and in practice a commission’s inquiry is frequently what moves a district administration to register a case, pay relief or correct a roster.
Q: Which ministries administer social justice legislation?
Two ministries carry the core of the field. The Ministry of Social Justice and Empowerment is the nodal ministry for the Scheduled Castes, the Other Backward Classes, senior citizens and persons with disabilities, and for the protection and rehabilitation statutes, including the Protection of Civil Rights Act, the Atrocities Act and the manual scavengers statute; it began as the Ministry of Welfare in 1985 and took its present name in May 1998. The Ministry of Tribal Affairs, formed in October 1999 when the Tribal Development Division was moved out of the social justice ministry, is the nodal ministry for the Scheduled Tribes, for the administration of the Scheduled Areas and for the Forest Rights Act. The Ministry of Minority Affairs, formed in January 2007, administers the minorities field. Implementation of the criminal statutes rests with State governments and their police and district administrations, under the nodal ministries’ monitoring.
Q: Do social justice provisions bind private persons, or only the State?
Both, depending on the provision. Most of Part III binds the State, and Articles 15(1) and 16 operate against State action. Three provisions reach private conduct directly. Article 15(2) forbids subjecting any citizen to disabilities in access to shops, public restaurants, hotels, places of public entertainment and public resorts and water sources, whether maintained by the State or open to the general public. Article 17 abolishes untouchability and forbids its practice in any form, by anyone. Article 23 prohibits traffic in human beings and forced labour against all persons. The Supreme Court held in People’s Union for Democratic Rights in 1982 that Articles 17, 23 and 24 are enforceable against private persons. The criminal statutes of the framework are built on that reach, which is why a private shopkeeper, landlord or employer can commit the offences they define. Affirmative provision, by contrast, binds the State as distributor and reaches private institutions only through an enabling clause and a law.