The Only Part With Right Answers
This is the easiest part of the series to write and the one most likely to change somebody's mind, for the same reason: the Constitution is a document, and documents can be read.
Everywhere else in this series I have had to weigh evidence, mark things contested, and tell you where the honest limits are. Here, for once, a large number of the disputed claims have simply correct and incorrect answers, and you can go and check them in about ten minutes.
The most repeated line in the entire Indian caste argument — that reservation was meant to last only ten years — is one of them.
Chapter 2 does that properly, including the detail that decides it: in 2020, one part of that ten-year clause was quietly allowed to expire, permanently, for the first time. Which tells you the sunset provision was never decorative — Parliament has used it. Just not on the thing everybody argues about.
Two other things in this part were new to me.
Ambedkar himself set the limit that EWS broke. On 30 November 1948 he told the Assembly that reservation had to be confined to a minority of seats, or the rule of equality would be destroyed. That sentence is where the 50% ceiling comes from. In 2019 Parliament went past it, and the people most likely to quote Ambedkar approvingly did not mention that they were crossing his line — and neither did the people who supported the amendment.
And there is a clause almost nobody argues. Article 335 says that the claims of Scheduled Castes and Tribes must be considered "consistently with the maintenance of efficiency of administration." The entire efficiency argument, which the abolition side makes constantly, has a home in the constitutional text — and they almost never cite it. Chapter 6 asks why.
One warning. This part contains more legal machinery than any other, and I have worked hard to keep it readable. Every Article is given in one plain paragraph, in ordinary English, with the actual wording alongside where the wording matters. If a page feels dry, it is doing load-bearing work for Parts 9, 10 and 14.
— Lovepreet Singh
July 2026
1The Room
Who was actually in the Constituent Assembly, and why that matters for how you read what it produced.
The Constituent Assembly of India first met on 9 December 1946. It adopted the Constitution on 26 November 1949, and the document came into force on 26 January 1950.
Just under three years, roughly 300 members after Partition, and a Drafting Committee of seven chaired by B. R. Ambedkar, appointed on 29 August 1947.
Three things about the room
The Assembly was not a cross-section of India. It was dominated by the Congress party, and its membership was drawn disproportionately from the educated, propertied and upper-caste sections of society — the same sections that had staffed colonial administration and the national movement's leadership.
This matters, and it cuts in an unexpected direction. A body composed largely of the beneficiaries of the caste order wrote provisions to dismantle parts of it. Whatever else that was, it was not self-interest.
He was first elected to the Assembly from Bengal. Partition removed that constituency. He was then re-elected from Bombay — with Congress support, at a time when his relations with Congress were hostile and the memory of the Poona Pact was fifteen years old.
He was subsequently made chairman of the Drafting Committee.
Both sides use this and neither uses it well. It is not proof that Congress had converted, and it is not proof that Ambedkar was co-opted. It is evidence that the leadership recognised the document would lack authority without him.
Members were chosen indirectly, by provincial legislatures elected on the restricted franchise of the 1935 Act. The population of India did not vote for this Assembly.
This is a genuine limitation on any argument of the form "the people decided in 1950." They did not. Roughly three hundred people did, on behalf of about 340 million.
The sceptical reading: an unelected body of elites produced a document that has bound a billion people for seventy-six years, and treating its intentions as sacred is a form of ancestor worship — the exact thing Ambedkar warned about in the same room.
The other reading: constitutions are made this way nearly everywhere, and the test is not the procedure that produced them but whether the society afterwards accepts and lives by them. India has, through seventeen general elections.
Both are fair. What neither supports is the move you will see constantly in the next two chapters — treating "the founders intended X" as though it settles a question that the founders explicitly left open for Parliament to decide.
2The Ten-Year Myth
The most repeated claim in the Indian caste argument. It is half true, which is why it survives — and the half that is true has a very interesting recent history.
You have heard it a thousand times. Ambedkar said reservation was only for ten years.
Here is exactly what is true and what is not.
The clock is real. It is in Article 334.
As originally enacted, Article 334 provided that the reservation of seats for Scheduled Castes and Scheduled Tribes in the House of the People and in the Legislative Assemblies of the States, and the nomination of members of the Anglo-Indian community to those bodies, would cease after ten years from the commencement of the Constitution.
Ten years from 26 January 1950 is 26 January 1960.
That is the entire scope of the famous clock. Reserved seats in legislatures. Politics.
What has no time limit, and never had one
Article 15(4) permits the State to make special provision for the advancement of socially and educationally backward classes and for Scheduled Castes and Tribes — the basis of reservation in education.
Article 16(4) permits the State to reserve appointments or posts in favour of any backward class of citizens not adequately represented in the services — the basis of reservation in government jobs.
Read both. There is no date, no period, no sunset and no review clause. There never was.
So the popular claim is a real constitutional provision applied to the wrong subject. It is not invented. It is misfiled.
And now the part that decides the argument
If the ten-year clock were a dead letter — a promise nobody ever intended to keep — the abolition side would have a strong point about bad faith.
It is not a dead letter. Parliament has taken it seriously enough to legislate on it seven times.
The two Anglo-Indian nominated seats in the Lok Sabha, and the equivalent nominations in state assemblies, had existed since 1950. The 104th Amendment let them expire.
The abolition side must stop saying reservation had a ten-year limit. It did not. The limit was on legislative seats, it is in a different Article, and repeating the claim after reading Article 334 is not a mistake but a choice.
But they gain something better than the myth. The Anglo-Indian lapse proves that a constitutional reservation can be ended by simple parliamentary decision, without upheaval, when the political cost is low enough. The mechanism exists and was used six years ago. That is a far stronger fact than the one they normally use.
The retention side must stop treating the clock as fictional. The founders did write an expiry into one part of the design, which means they were capable of writing one and chose not to elsewhere — a choice that has to be explained rather than assumed.
And the awkward question for everyone: a provision written to last ten years has now been extended for eighty, in ten-year instalments, each time by a Parliament that found it politically impossible to refuse. Whatever that process is, it is not the periodic review the drafting implies.
3The Articles, One Paragraph Each
Every constitutional provision that matters to this argument, in ordinary English. Keep this chapter; the rest of the series refers back to it.
The equality provisions
| Article | What it does |
|---|---|
| 14 | Equality before law. The State shall not deny to any person equality before the law or equal protection of the laws. The base rule everything else is measured against. |
| 15(1) | No discrimination. The State shall not discriminate against any citizen on grounds only of religion, race, caste, sex or place of birth. |
| 15(4) | The education exception. Nothing in this Article prevents the State from making special provision for the advancement of socially and educationally backward classes, or for SCs and STs. Inserted by the First Amendment, 1951. |
| 15(5) | Private colleges. Extends that power to admissions in private educational institutions, aided or unaided, except minority institutions. 93rd Amendment, 2005. |
| 15(6) | EWS in education. Special provision for economically weaker sections among those not covered by 15(4) or 15(5), up to 10% of seats. 103rd Amendment, 2019. |
| 16(1) | Equality of opportunity in matters of public employment. The rule. |
| 16(4) | The jobs exception. Nothing prevents the State from reserving posts for any backward class of citizens which, in the opinion of the State, is not adequately represented in the services. Note two things: it is enabling, not mandatory; and the test is representation, not poverty. |
| 16(4A) | Promotions. Permits reservation in promotion with consequential seniority for SCs and STs. 77th Amendment, 1995; seniority added by the 85th, 2001. |
| 16(4B) | Backlog. Unfilled reserved vacancies carried forward are treated as a separate class and do not count against the 50% ceiling for that year. 81st Amendment, 2000. |
| 16(6) | EWS in jobs. Up to 10% of appointments for economically weaker sections. 103rd Amendment, 2019. |
| 17 | Untouchability abolished. Its practice in any form is forbidden, and enforcing any disability arising out of it is a punishable offence. One of very few Articles that binds private individuals, not just the State. |
The rest of the machinery
| Article | What it does |
|---|---|
| 46 | The duty. The State shall promote with special care the educational and economic interests of the weaker sections, particularly SCs and STs, and protect them from social injustice and exploitation. A Directive Principle — not enforceable in court, but the stated purpose behind the rest. |
| 330 | Reserved seats in the Lok Sabha for Scheduled Castes and Scheduled Tribes. |
| 332 | Reserved seats in State Legislative Assemblies for SCs and STs. |
| 334 | The clock — on Articles 330 and 332 only. Extended seven times; currently to 25 January 2030. |
| 335 | The efficiency clause. The claims of SCs and STs shall be taken into consideration in appointments "consistently with the maintenance of efficiency of administration." See Chapter 6. |
| 338 / 338A / 338B | The watchdogs. National Commissions for Scheduled Castes, Scheduled Tribes and Backward Classes, each given constitutional status by amendment — 1990, 2003 and 2018 respectively. |
| 340 | The commission power. The President may appoint a Commission to investigate the conditions of socially and educationally backward classes. This is the Article that produced Kalelkar in 1953 and Mandal in 1979. |
| 341 / 342 / 342A | The lists. The President specifies, by public notification, which castes are Scheduled Castes (341) and which tribes are Scheduled Tribes (342); 342A does the equivalent for socially and educationally backward classes. Only Parliament can amend these lists. |
1. Reservation is permitted, not required. Article 16(4) says nothing "shall prevent the State." A government that chose not to reserve posts would not be violating the Constitution. That is a much bigger fact than it appears, and Chapter 10 returns to it.
2. The test is representation, not poverty. "Not adequately represented in the services." Not "poor." This is why the argument in Part 1 — that reservation is not an anti-poverty programme — is a constitutional point rather than a rhetorical one.
3. Untouchability is banned outright; caste is not. Article 17 abolishes a practice. Nothing abolishes caste itself, and nothing forbids a private person from marrying within it. Which returns us to Part 3's conclusion by another route.
4. Almost every operative provision on this list is an amendment. 15(4), 15(5), 15(6), 16(4A), 16(4B), 16(6), 338A, 338B, 342A. The reservation architecture of 2026 is substantially not the document of 1950 — it is seventy-five years of legislative construction on top of it.
430 November 1948
The day the Assembly debated Article 16(4). What Ambedkar actually said — and the limit he set that Parliament crossed in 2019.
On 30 November 1948 the Constituent Assembly debated the draft provision that became Article 16. Ambedkar's speech that day is the single most important document for anyone arguing about how much reservation is too much.
The three positions in the room
He began by describing the disagreement he had to resolve. There were, broadly, three views among members:
- Complete equality of opportunity, with no reservation of any kind — the position that any exception destroys the principle.
- Equality with reservation for communities historically shut out of public employment.
- A strong version of the second, holding that reservation was essential to integrate communities that had never been allowed into administration at all.
His account of the Drafting Committee's task was that it had to produce a formula reconciling these views — that there should be equality of opportunity, and that there should be reservation in favour of communities that had not so far had, in his phrase, a proper look-in to the administration.
And then the sentence that has governed Indian law ever since
His argument was that while equality of opportunity was a great principle, there had to be provision to let communities historically denied entry into public employment actually enter it — but that such reservation must be confined to a minority of posts, so that the principle of equality in clause (1) would not be destroyed.
An exception that consumes more than half of what it is an exception to has stopped being an exception.
When the Supreme Court in Indra Sawhney (1992) held that reservation must normally not exceed 50% of available seats, it was giving legal force to the principle Ambedkar had stated in the Assembly forty-four years earlier.
The ceiling is not a judicial invention. It is the drafting committee chairman's own limit.
In 2019, the 103rd Amendment added 10% for economically weaker sections, taking central reservation to 59.5%.
That is past the majority line. It crosses the limit Ambedkar set in the room, and it does so for a category — economic disadvantage among communities with no other reservation — that Article 16(4)'s representation test does not cover.
Notice who did not object on those grounds. Supporters of EWS did not, for obvious reasons. But nor did most of the people who cite Ambedkar as the authority on reservation, because opposing EWS meant defending a ceiling that also constrains what they want.
The Supreme Court upheld it 3–2 in 2022. Two judges, including the Chief Justice, dissented — substantially on this point.
If you want a single example of how selectively this man is quoted, it is this: the most famous defender of reservation in Indian history also set the strictest limit on it, and the limit was crossed with almost no one invoking him.
One more thing he said
Ambedkar was also clear about what "backward classes" meant in Indian conditions. His position, stated in the Assembly, was that what are called backward classes are in substance a collection of certain castes.
That sentence forecloses a great deal of later argument in both directions. It means the category was understood from the start as caste-based rather than income-based — which supports the retention side's reading of Article 16(4). And it means the founders knew they were building a permanent state apparatus for recording caste, which is exactly what the abolition side objects to.
5Champakam, and the First Amendment
Fifteen months after the Constitution came into force, the Supreme Court struck reservation down. Parliament amended the Constitution rather than accept it.
This is the most under-told episode in the whole legal history, and it settles more than almost anything else in this part.
Madras operated a communal order allocating seats in state medical and engineering colleges among communities — the descendant of the 1921 arrangement described in Part 5.
The Supreme Court struck it down. Its reasoning: Article 15(1) forbade discrimination on grounds of caste, Article 29(2) forbade denial of admission to state-funded institutions on grounds of caste, and no exception then existed for backward classes in education.
Reservation in education, as practised in Madras for thirty years, was unconstitutional.
The response
Parliament — under Nehru, in the Constitution's first full year — responded by amending the document itself, inserting Article 15(4) and expressly permitting special provision for socially and educationally backward classes and for SCs and STs.
The same amendment created the Ninth Schedule, a list of laws placed beyond ordinary judicial review — the device by which Tamil Nadu would later protect its 69% reservation.
Both are still in force.
Against the abolition reading: if the framers had regarded reservation as a marginal, temporary indulgence, the natural response to Champakam was to accept the judgment. Instead the very first amendment to the Indian Constitution was made in order to preserve it. That is not the behaviour of a legislature that considered it optional.
Against the retention reading: the amendment was necessary because the original text did not permit it. The 1950 Constitution, as written and as read by the Supreme Court, did not allow caste-based reservation in education. That is a real point and it is rarely conceded.
The honest summary: reservation in education is not something the founders wrote into the Constitution. It is something Parliament wrote in, immediately, when the courts said the original text did not allow it.
Which is why "what the founders intended" is such a weak instrument here. The founders and the first Parliament were largely the same people, and they intended it enough to amend their own document within fifteen months.
6Article 335 — The Sleeper Clause
The efficiency argument has a home in the constitutional text. Almost nobody who makes that argument has ever cited it.
"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."
Read that carefully, because it does two opposite things in one sentence.
It requires that the claims of SCs and STs be taken into consideration — a positive obligation. And it makes that consideration subject to a condition: consistently with the maintenance of efficiency of administration.
Why this clause should be at the centre of the argument
The single most common argument the abolition side makes is that reservation harms institutional quality. Whatever one thinks of it, it is not a fringe position — it is Nehru's position in the 1961 letter, and it is written into the Constitution.
And the side making it almost never cites Article 335.
For the abolition side, the problem is that "efficiency" is undefined and unmeasured. The Article states a condition without a test. Nobody has ever produced an agreed measure of administrative efficiency against which a reservation policy could be assessed. An argument that cannot be operationalised cannot be won, only asserted — and Part 11 shows how thin the actual evidence is in both directions.
For the retention side, the problem is that the clause exists at all. The framers thought the risk serious enough to write it in. Treating every efficiency argument as bad faith means treating Article 335 as bad faith.
And what Parliament did to it
The proviso permits the State to make provision for relaxation in qualifying marks, or lowering of standards of evaluation, for reservation in matters of promotion for SCs and STs.
In other words, Parliament looked at a clause conditioning reservation on efficiency and expressly legislated an exception to it.
The courts have worked through the same tension. In M. Nagaraj (2006) the Supreme Court required states to demonstrate backwardness, inadequacy of representation and administrative efficiency before providing reservation in promotion. In Jarnail Singh (2018) it removed the requirement to prove the backwardness of SCs and STs, while applying the creamy layer principle to them in the promotion context.
The Constitution contains a brake. Parliament has repeatedly legislated around the brake. The courts have repeatedly tried to give it content and been repeatedly overruled by amendment.
That is not a scandal; it is how a democracy resolves a conflict between a written condition and a political majority. But it does mean something for anyone arguing about the future of reservation, and Part 14 will have to say it plainly: every constitutional constraint on reservation that has been tested has eventually been amended away.
The 50% ceiling by the 103rd Amendment. The efficiency clause by the 82nd. The prohibition on promotion quotas by the 77th. Judicial limits by the 81st and 85th.
Whatever your position, an honest forecast has to account for that pattern.
7Opportunity Or Outcome
The oldest question in this argument, and the Constitution's answer is clearer than either side admits.
The standard framing is that the left wants equality of outcome and the right wants equality of opportunity. The Indian constitutional text does not fit that framing at all, and the reason is structural rather than rhetorical.
Look at the architecture
| Provision | What it says | Its status |
|---|---|---|
| 16(1) | Equality of opportunity in public employment | The rule. Stated first, unconditionally. |
| 16(4) | Nothing shall prevent reservation for backward classes not adequately represented | The exception. Permissive, conditional, and framed as a carve-out from the rule. |
This is not an interpretive claim; it is the grammar of the Article. Clause (1) states the right. Clause (4) begins "Nothing in this article shall prevent…" — the standard drafting form of an exception to what precedes it.
Everything follows from that structure. It is why Ambedkar said reservation must be confined to a minority of posts. It is why the Supreme Court set a ceiling. And it is why the EWS quota crossing 50% was constitutionally significant rather than merely controversial.
Read what Article 16(4) actually conditions the exception on: a class "not adequately represented in the services."
That is not a test of poverty, and it is not a test of outcome equality across society. It is a test of presence in public institutions. The constitutional theory is neither "make everyone equal" nor "ignore history" — it is that a state whose administration is monopolised by a few communities is not delivering equality of opportunity in the first place, whatever its formal rules say.
That is a genuinely third position, and it is the one the text takes. It says: the monopoly is itself the inequality of opportunity, and breaking it is a precondition of the rule rather than a departure from it.
You can disagree with that theory. What you cannot do is describe the Indian Constitution as choosing equality of outcome, because it does not, or as choosing formal equality alone, because it explicitly does not.
Shahu Maharaj in 1902, from Part 5, gave exactly this justification: not compensation for history, but the observation that a handful of communities held nearly every post in his administration.
The oldest reservation policy in the world and the Indian constitutional text are making the same argument, seventy years apart. Whether it works is Part 11.
8Eighteen Amendments
How the reservation architecture was actually built — one legislative decision at a time, over seventy years.
Count the amendments passed specifically to overrule a court decision that had limited reservation: the First (after Champakam), the 77th (after Indra Sawhney on promotions), the 81st and 85th (after subsequent limits), the 82nd (qualifying the efficiency clause), the 105th (after the Maratha judgment).
The courts have repeatedly attempted to constrain the scope of reservation. Parliament has repeatedly amended the Constitution to remove the constraint — under governments of every political description, across seventy years, usually with overwhelming majorities and often unanimously.
For the abolition side this is the strongest structural argument available, and they rarely make it: the constitutional limits on reservation are not limits, because the body that can remove them always does.
For the retention side it is a reminder that the policy's foundation is political rather than constitutional. It survives because it commands parliamentary majorities, not because the text protects it.
Which means both sides are wrong to argue about what the Constitution requires. The Constitution says what Parliament decides it says, and has for seventy-five years.
9Seven Judgments
The Supreme Court decisions that built the system you actually live under — each in plain English, with what it settled and what it did not.
Chapter 8 was Parliament's half of the story. This is the courts' half. Seven decisions, in order, and by the end of them you will understand nearly every technical term used in this argument.
What it held: the Madras communal order allocating college seats by community violated Articles 15(1) and 29(2). There was no exception in the 1950 text permitting it.
What followed: the First Amendment, within months, inserting Article 15(4).
Why it still matters: it is the proof that caste-based reservation in education was not in the original Constitution. Both sides should know this and only one usually does.
What it held: Mysore had reserved 68% of seats. The Court struck it down, observing that a special provision under Article 15(4) should be less than 50% — the first judicial statement of the ceiling. It also held that caste alone could not be the test of backwardness, and warned against the more advanced members of a backward group capturing the benefit.
What it did not settle: almost everything. The 50% figure was framed as guidance rather than a hard rule, and the relationship between caste and backwardness stayed contested for another thirty years.
What it held, in four parts:
(a) 27% OBC reservation is constitutional. Caste can be the starting point for identifying a backward class, because in India caste and social backwardness substantially overlap.
(b) Total reservation must normally not exceed 50% — Ambedkar's "minority of seats," now binding law.
(c) The creamy layer — the advanced sections within a backward class — must be excluded. This is where the OBC income test comes from.
(d) Article 16(4) does not extend to promotions.
What followed: Parliament overruled (d) by the 77th Amendment in 1995.
What it held: the amendments enabling reservation in promotion were valid, but a state wishing to use them must first collect quantifiable data showing the backwardness of the class, the inadequacy of its representation, and that the measure would not damage administrative efficiency under Article 335.
Why it mattered: for the first time a court asked for evidence before permitting a quota. Many states could not produce it.
What it held: the requirement to prove the backwardness of Scheduled Castes and Tribes was removed — their backwardness is presumed by their listing under Articles 341 and 342. But the creamy layer principle was applied to SCs and STs in the promotion context, having previously been confined to OBCs.
Why it matters: it is the first authoritative acceptance that advancement within SC and ST communities is real enough to require exclusion — the finding that returns, much larger, in 2024.
What the majority held: reservation on purely economic criteria is permissible; excluding SCs, STs and OBCs from the EWS category does not violate equality, because they already have their own provision; and the 50% ceiling is not inviolable, applying to Articles 15(4) and 16(4) rather than binding a new provision.
What the dissent held: that excluding the poor among SCs, STs and OBCs from a poverty-based quota is discrimination on its face, and that breaching the ceiling opens a door that cannot be closed. Two judges, including the Chief Justice.
One vote the other way and the entire structure of Indian reservation would look different today.
What it held: states may sub-classify Scheduled Castes, splitting the SC quota to give the most deprived sub-castes a protected share within it — overruling an earlier decision that had treated the Scheduled Castes as a single homogeneous class.
What it means: the highest court in the country has now formally accepted that the benefits of reservation have been captured unevenly within reserved categories.
And note the lineage: Balaji warned about it in 1963. Indra Sawhney built the creamy layer around it in 1992. Jarnail Singh extended it in 2018. Davinder Singh completed the arc in 2024. Elite capture is not a talking point; it is sixty years of judicial finding.
Direction one — the courts keep trying to bound it. Balaji's ceiling, Indra Sawhney's 50% and creamy layer, Nagaraj's evidence requirement, Jarnail Singh's extension of creamy layer, Davinder Singh's sub-classification. Every one narrows or conditions the benefit.
Direction two — Parliament keeps widening it. The 1st, 77th, 81st, 82nd, 85th, 93rd, 103rd and 105th Amendments, each removing or qualifying a judicial constraint.
Seventy-five years of a court applying brakes and a legislature releasing them. That is the actual constitutional history, and neither side describes it accurately: the abolition side treats the courts as helpless, and the retention side treats each judicial limit as an attack rather than as the system working.
10Two Readings Of The Same Transcripts
Both cases, at full strength, from the identical record.
They wrote a clock. Article 334 sets an expiry. The framers were capable of writing sunset provisions and did write one. A body that intended something permanent does not draft a termination date for part of it.
Ambedkar set a ceiling. On 30 November 1948 he said reservation must be confined to a minority of posts or the equality principle would be destroyed. He was not describing a permanent redistribution; he was describing a bounded exception.
They wrote a brake. Article 335 conditions the whole thing on the maintenance of administrative efficiency. That is the language of people worried about cost.
They made it optional. Article 16(4) is enabling, not mandatory. A mandatory permanent entitlement would have been drafted as one.
And the man who ran the government disliked it. Nehru's 1961 letter, from Part 7, is what the political leadership actually thought once it was operating.
They put no clock on the things that matter. Articles 15(4) and 16(4) have no expiry. The framers demonstrably knew how to write a sunset clause, wrote one for legislative seats, and deliberately did not write one for jobs or education. Omission by people who used the device elsewhere is a choice.
They amended the Constitution within fifteen months to save it. When the Supreme Court struck down reservation in education in 1951, the first amendment to the Indian Constitution was passed to reinstate it. Not a statute — the Constitution.
They built for the future, not the interim. Article 340 empowers commissions to investigate backward classes with no time limit. Article 46 states an ongoing duty. These are the provisions of people expecting a long project.
The condition was never met. Article 16(4) is triggered by inadequate representation. It ends when representation becomes adequate, not when a calendar page turns. Nobody has argued the condition has been satisfied.
And Ambedkar told them how long it would take. On 25 November 1949 he warned that India was entering a life of contradictions — political equality alongside social and economic inequality — and asked how long that could last. That is not a man expecting the problem solved by 1960.
Reading A is right that the framers intended limits, and its strongest evidence is Ambedkar's own ceiling — which has since been crossed.
Reading B is right about the drafting. People who wrote a sunset clause in one place and not in another were making a distinction, and the distinction is exactly the one the popular myth erases.
What I think the record actually shows is that they intended reservation to be bounded but open-ended: limited in quantity, unlimited in time, and ending when a condition was met rather than when a period expired.
Which is an uncomfortable finding for everyone. It denies the abolition side its deadline. And it denies the retention side the 50%-plus arrangement that now exists, because the boundedness was the part the framers were explicit about — and it is the part that has been eroded.
The design was: a minority of seats, for as long as necessary. India currently has a majority of seats, for as long as necessary. Both halves of the original settlement have been abandoned, and the two sides each defend the half they like.
11Nine Questions Nobody Asks
The blind spots in this stage of the argument.
1. Article 16(4) is enabling. Could a government simply stop?
The text says nothing "shall prevent the State" from reserving posts. It does not say the State must. On a plain reading, a government that ceased reserving central government posts would not be violating Article 16(4).
Nobody has tested it, and no government has come close to trying. Which tells you the constraint operating here is political, not legal — a point Part 14 has to build on.
2. Why does nobody cite Article 335?
The efficiency argument is the abolition side's most-used claim and it is written into the Constitution. They cite Nehru's private letter instead.
The likely reason: citing 335 means engaging with what "efficiency of administration" would actually be measured by, and no one has an answer. The letter requires no evidence; the Article does.
3. Why did the founders never define "backward classes"?
Article 16(4) turns on a category the Constitution does not define, leaving it to the opinion of the State. Article 340 provides for commissions to investigate it.
Every subsequent fight — Kalelkar, Mandal, creamy layer, EWS, the Maratha and Jat claims — flows from that single undefined term. It may be the most consequential drafting decision in the document.
4. Did they anticipate OBC reservation?
Structurally, yes. Article 340 exists solely to provide for investigating socially and educationally backward classes, and 15(4) and 16(4) both refer to backward classes as a category distinct from SCs and STs.
The claim that Mandal was an unconstitutional expansion beyond the founders' scheme does not survive reading Article 340.
5. EWS crossed Ambedkar's line. Why did nobody say so?
Because the ceiling constrains everyone. The retention side could not attack EWS on ceiling grounds without endorsing a limit on their own claims. The abolition side supported EWS. So the one argument with real constitutional force went unmade by both.
6. Why is there no reservation in the judiciary?
Higher judicial appointments in India are made through the collegium system, and reservation does not apply. The composition of the higher judiciary is correspondingly unrepresentative.
Neither side campaigns on it. The abolition side does not, because it is an example of an unreserved institution that is visibly monopolised. The retention side does not, because it would mean confronting the judiciary directly.
7. What does "adequately represented" mean?
The Article makes the trigger inadequate representation and then leaves the assessment to "the opinion of the State."
No numerical standard exists. Which means the condition on which the entire exception rests has no definition and no test — and therefore no mechanism by which it could ever be declared satisfied. Part 14 has to deal with this, because it is the real reason the policy has no exit.
8. If the Constitution has been amended over a hundred times, what does "founders' intent" mean?
Nine of the amendments in Chapter 8 rewrote the reservation provisions. Several overruled the Supreme Court. The document of 2026 is not the document of 1950.
Appeals to founders' intent are therefore appeals to an early draft of a document that its own authors began revising within fifteen months.
9. Is anyone actually bound by any of this?
The pattern in Chapter 8 is unbroken: every constitutional limit on reservation that has been tested has eventually been amended away.
This is the question that should worry both sides. It means the abolition side cannot win in court, because Parliament will amend. And it means the retention side has no protection either, if a future Parliament with a large enough majority decides otherwise. The whole structure rests on political arithmetic, and both camps talk as though it rests on principle.
12What The Evidence Can And Cannot Settle
The honest scorecard for Part 8.
| Question | Status | Best current answer |
|---|---|---|
| Was reservation limited to ten years? | No | Article 334 clocked legislative seats only |
| Is the ten-year clock real? | Yes | Extended 7 times; now to 25 Jan 2030 |
| Has a constitutional reservation ever been allowed to lapse? | Yes | Anglo-Indian nomination, 25 January 2020 |
| Do Articles 15(4) and 16(4) have expiry dates? | No | They never did |
| Is reservation constitutionally mandatory? | No | 16(4) is enabling — "nothing shall prevent" |
| Is the test poverty or representation? | Representation | "Not adequately represented in the services" |
| Where does the 50% ceiling come from? | Ambedkar | 30 Nov 1948 — "confined to a minority" |
| Was reservation in education in the 1950 text? | No | Added by the First Amendment, 1951 |
| Have courts been able to limit reservation? | Not durably | Every limit tested has been amended away |
| Is there a defined test for "adequate representation"? | No | Left to the opinion of the State |
The three things nobody can tell you
- What the framers would say now. They wrote for a country of 340 million with 12% literacy and almost no private sector. Extrapolating their intent to 2026 is guesswork dressed as authority, whoever does it.
- What "efficiency of administration" means. Article 335 states a condition with no metric. Seventy-five years on, nobody has produced one.
- When "adequate representation" would be reached. No standard exists, so the condition cannot be declared satisfied — by design or by omission, nobody knows which.
The founders designed a bounded exception with no deadline, conditioned on a standard they never defined — and then their successors removed the bound and left the condition undefined.
Which means the argument in 2026 is not about what the Constitution says. It is about who has the votes to change it, and it always has been.
13The Five-Line Nichod
Everything in this book, compressed. If you remember nothing else, remember these.
PART 8 — THE CONSTITUTION ROOM
- The ten-year clock is real and it is in the wrong Article. Article 334 clocked reserved seats in the Lok Sabha and state assemblies — politics. Articles 15(4) and 16(4), which govern education and jobs, never had an expiry date. The clock has been extended seven times, now to 25 January 2030.
- And one part of it really did expire. The 104th Amendment let Anglo-Indian nomination lapse on 25 January 2020 — the first constitutional reservation ever allowed to end. Proof the sunset mechanism works when the political cost is low enough.
- Ambedkar set the 50% ceiling, and EWS crossed it. On 30 November 1948 he told the Assembly reservation must be "confined to a minority of seats" or equality would be destroyed. That is where Indra Sawhney's cap comes from. The 103rd Amendment took central reservation to 59.5% in 2019, and almost nobody invoked him.
- Reservation in education was not in the 1950 Constitution. The Supreme Court struck it down in Champakam Dorairajan (1951). Parliament responded with the First Amendment, inserting Article 15(4) — and creating the Ninth Schedule that later protected Tamil Nadu's 69%.
- Every constitutional limit that has been tested has been amended away. Champakam by the 1st, the promotions bar by the 77th, the efficiency clause by the 82nd, judicial limits by the 81st and 85th, the ceiling by the 103rd, the Maratha judgment by the 105th. The argument is not about what the Constitution says. It is about who has the votes.
Coming next
The Kaka Kalelkar Commission of 1953 and why its own chairman ended up doubting it. The Mandal Commission of 1979–80: its method explained simply, its eleven indicators, its reliance on 1931 census arithmetic, and a fair account of the criticism. Then 1990 — V. P. Singh's announcement, the nationwide protests, the self-immolations, and the permanent redrawing of north Indian politics. And Indra Sawhney (1992), the nine-judge bench that gave us the 50% ceiling, the creamy layer, and the exclusion of promotions that Parliament overturned three years later.
A closing thought: the framers' actual design was a minority of seats, for as long as necessary. India now has a majority of seats, for as long as necessary. Each side defends the half of that settlement it prefers, and neither will say the other half aloud. — L.S.
Sources & further reading — Part 8
Every factual claim in Part 8, and where it comes from.
The Assembly
First sitting 9 December 1946; Constitution adopted 26 November 1949; commencement 26 January 1950. Drafting Committee of seven chaired by B. R. Ambedkar, appointed 29 August 1947. Members were elected indirectly by provincial legislatures chosen on the restricted franchise of the Government of India Act 1935. Ambedkar was first elected from Bengal, lost that seat at Partition, and was re-elected from Bombay.
The Articles
Constitution of India, Articles 14, 15, 16, 17, 29(2), 46, 330, 332, 334, 335, 338, 338A, 338B, 340, 341, 342 and 342A. Article 335 is quoted in full; Article 16(4)'s enabling formulation is quoted in Chapter 3 and discussed in Chapters 7 and 11.
The Constituent Assembly debates
Constituent Assembly Debates, 30 November 1948 — Ambedkar on the draft of what became Article 16: the three positions in the Assembly, the Drafting Committee's task of reconciling them, and his statement that reservation must be confined to a minority of seats so that equality of opportunity under clause (1) is not destroyed. This is the basis of the 50% ceiling later adopted in Indra Sawhney. · Debates of 25 November 1949 — the "life of contradictions" speech, quoted in Part 7.
Cases
State of Madras v. Champakam Dorairajan (1951) — Madras communal order struck down under Articles 15(1) and 29(2). · M. R. Balaji v. State of Mysore (1963) — 68% reservation struck down; the first judicial statement that special provision should be under 50%; caste alone not the test of backwardness; early warning about advanced sections capturing the benefit. · Indra Sawhney v. Union of India (1992) — 27% OBC reservation upheld; 50% ceiling; creamy layer; promotions excluded. · M. Nagaraj v. Union of India (2006) — quantifiable data on backwardness, inadequacy of representation and administrative efficiency required before reservation in promotion. · Jarnail Singh v. Lachhmi Narain Gupta (2018) — backwardness no longer to be established for SCs and STs; creamy layer applied to them in promotion. · Janhit Abhiyan v. Union of India (2022) — EWS upheld 3–2, with two dissents including the Chief Justice. · State of Punjab v. Davinder Singh (2024) — sub-classification of Scheduled Castes permitted, 6–1. · Jaishri Laxmanrao Patil v. Chief Minister (2021), the Maratha reservation case, which prompted the 105th Amendment.
The amendments
1st (1951) — Article 15(4); Ninth Schedule. 8th (1959), 23rd (1969), 45th (1980), 62nd (1989), 79th (1999) and 95th (2009) — successive extensions of Article 334. 65th (1990) — constitutional status for the National Commission under Article 338. 77th (1995) — Article 16(4A). 81st (2000) — Article 16(4B). 82nd (2000) — proviso to Article 335 permitting relaxation in qualifying marks and standards of evaluation. 85th (2001) — consequential seniority, retrospective to 1995. 89th (2003) — Article 338A. 93rd (2005) — Article 15(5). 102nd (2018) — Articles 338B and 342A. 103rd (2019) — Articles 15(6) and 16(6), EWS. 104th (2019) — Article 334 extended to 25 January 2030; Anglo-Indian nomination not extended, ceasing on that date; signed 21 January 2020, effective 25 January 2020. 105th (2021) — restoration of state power to identify socially and educationally backward classes.
A note on what is missing — and on corrections
Part 8 describes the constitutional design, its amendment history and the leading judgments. It makes no claim about whether reservation has worked (Part 11), about Mandal and its aftermath (Part 9), or about the recent changes (Part 10). Judgments are summarised in outline and are not a substitute for the texts. Amendment numbering follows the official series. If you find an error, I want to know.