The Part Where The Paper Trail Starts
Something changes in this part, and it is worth naming before we begin.
Everything up to now has been inference. Hymns that might be prescriptive or descriptive. Archaeology with no readable script. Genetics that gives dates but no motives. Law books nobody can prove were ever enforced.
From roughly 1860 onward, that stops. The British ran India on paper, and they kept the paper. Census volumes. Gazetteers. Legislative debates. Court records. Commission reports. Petitions from caste associations. Private correspondence between officials arguing about what to do. For the first time in this series, we can watch people make decisions about caste, in writing, with their reasons attached.
Dangerous, because the sudden abundance of records tempts everyone into a false conclusion. When caste appears everywhere in the documents from 1871 and almost nowhere in the documents from 1471, the obvious inference is that something happened in between. Sometimes that is right. Often it just means somebody started writing things down.
The other reason this part matters is that it is where the modern argument is physically manufactured. Reservation is not an idea somebody had in 1950. Its vocabulary, its categories, its legal machinery and its central political bargain were all built between 1871 and 1936, and Independence inherited them almost intact.
Two things in here surprised me enough to change how I think about the whole subject. The first is that the earliest reservation policy in the world was Indian, not British — issued by an Indian king in 1902, decades before the colonial government did anything comparable. That fact is missing from almost every argument on both sides.
The second is Chapter 9. The Poona Pact of 1932 is remembered as a footnote about Gandhi's fast. It is not a footnote. It is the moment the shape of Indian reservation was decided, under duress, by two men who each believed they were preventing a catastrophe. Everything since has been an argument about that week.
— Lovepreet Singh
July 2026
1The Claim
"The British created caste." An argument you have heard many times, put the way its best defenders put it.
Part 4 disposed of the idea that invaders brought caste to India. Foreign observers documented it centuries before any Sultan reached Delhi.
But there is a second, much more sophisticated version of the outsider argument, and it deserves separate treatment because a large part of it is correct and it is supported by serious academic work.
Before the British, "caste" was loose, local and negotiable. A community's rank varied by district. Groups argued about their own placement and sometimes changed it. There was no single list and no authority that could settle a dispute nationally.
The colonial state changed that for administrative reasons that had nothing to do with religion. An empire that governs by paper needs categories. So from 1871 it counted every Indian by caste, decade after decade, and printed the results in official volumes.
Counting did three things at once. It fixed identities that had been fluid. It ranked them in print, with government authority behind the ranking. And it made caste a political unit — once the state allocated anything by caste, every community had reason to organise and fight over its classification.
Then Britain built the machinery: courts applied a Sanskrit law book as if it were a national code; a law declared whole communities criminal by birth; and the state created official schedules of castes entitled to benefits.
So the caste system Indians argue about today — enumerable, national, legally defined, politically mobilised — is substantially a colonial product. Not the hierarchy, but its modern form.
Why this version is harder to dismiss
Because unlike the "invaders brought it" claim, this one rests on documents you can read. The census volumes exist. The Criminal Tribes Act exists. The petitions demanding reclassification exist in their thousands. In moderate form it is the majority position among historians of colonial India: the question is not whether the colonial state transformed caste, but how much.
2The Census
Every ten years from 1871, the British asked 250 million people what caste they were, and wrote the answers down.
The first all-India census was conducted in 1871–72. It was an enormous undertaking, and it included a question about caste.
That sounds administratively dull. It was one of the most consequential things anyone did to Indian society in the nineteenth century, and understanding why requires thinking about what counting does.
Four things the census changed
A census form has one box. But caste identity in pre-colonial India was frequently contextual — a community might be described one way by its neighbours, another way by itself, and another way again in the next district. Sub-groups shaded into each other.
The form does not permit that. It requires a name, and once the name is recorded, the ambiguity is gone from the record — and eventually from the argument.
A weaving community in one district and a weaving community six hundred kilometres away, speaking a different language, with no contact and no history of intermarriage, could be recorded under a single heading — and thereby became, for the first time, a single thing with a single number.
Several communities later organised themselves along exactly the lines the census had drawn.
Successive censuses did not merely list castes. They attempted to arrange them in order of "social precedence." The 1901 census, under Herbert Hope Risley, made this explicit and systematic — and that is Chapter 3.
The distinction matters enormously. A local claim about who is higher is an opinion. The same claim printed in a government volume is a fact of record, and it can be cited.
Once rank appeared in an official publication, communities had a concrete reason to contest it. Caste associations formed — many of them for the first time — hired lawyers, gathered evidence, produced genealogies and petitioned the Census Commissioner to be reclassified higher.
The volume of these petitions was substantial enough to become an administrative problem in its own right.
The petitions are the most revealing part of the story, and both sides misread them.
What one side takes from them: caste identity was so fluid that communities could plausibly argue for a different rank, which proves the categories were artificial and the British froze something that had been in motion.
What the other side takes from them: the communities were not petitioning to abolish the ranking. They were petitioning to be placed higher within it. Nobody wrote in to say the whole scheme was meaningless. That tells you the hierarchy was real, deeply held, and desperately mattered to people — which is not what you would expect if it were a colonial invention.
Both readings are correct at the same time. The census found a real hierarchy with fuzzy edges, and made the edges sharp. That is genuinely different from creating it, and genuinely different from leaving it alone.
The census of 1931, conducted under J. H. Hutton, was the last to enumerate castes comprehensively. After Independence, India continued to count Scheduled Castes and Scheduled Tribes but stopped counting other castes.
Which is why, as covered in Part 1, the Mandal Commission's estimate of the OBC population in 1980 was built on arithmetic applied to 1931 data — and why the caste enumeration in Census 2027 is the most consequential dataset in this entire argument.
3Risley's Ruler
The senior colonial official who believed caste rank could be read off the width of a person's nose — and who ran the census of 1901.
Herbert Hope Risley was an officer of the Indian Civil Service, an ethnographer, and Census Commissioner for the census of 1901. He is the single most consequential individual in the colonial construction of caste, and his theory was nonsense.
What he believed
Risley held that Indian caste was fundamentally racial. The population, in his account, descended from a mixture of two stocks — one supposedly Aryan and one supposedly aboriginal — and a community's position in the caste hierarchy reflected how much of each it carried.
And he believed this could be measured on the body. Working within the anthropometric fashion of nineteenth-century European race science, he took physical measurements of members of different communities, and treated the nasal index — the ratio of the width of the nose to its height — as a key indicator of racial type and therefore of caste rank.
Risley's census did not simply record caste names. It set out to rank communities in order of social standing, as he understood that standing to be recognised locally, and it published the result with the authority of the Government of India behind it.
The exercise produced a flood of protest and petition from communities objecting to their placement — and a large body of caste-association literature arguing for reclassification.
Why this matters more than the silliness suggests
It is tempting to file Risley under Victorian pseudoscience and move on. Three reasons not to.
- He was not a crank on the margins. He was the Census Commissioner and later a senior official of the Government of India. His framework organised the data collection for the most comprehensive survey of Indian society ever attempted to that point.
- His race theory outlived his methods. The nasal index was abandoned. The idea it was built to support — that caste is fossilised race, fair conquerors above dark natives — entered Indian textbooks, colonial administration and, as noted in Part 2, parts of the anti-caste movement, where it remained influential for most of a century.
- The rankings he printed were used. Once a hierarchy exists in a government publication, it can be cited in a court, in a petition, in a land dispute and in a school admission. Print gives an opinion the texture of a fact.
Risley's racial account of caste is now rejected by essentially everyone. The genetic evidence in Part 2 does not show the population split it requires; the linguistic case for reading varna as skin colour is weak.
But look at what remains standing after you remove it. Risley was wrong about why the hierarchy existed. He was not wrong that he found one. He did not have to persuade anybody that some communities were ranked above others — the difficulty he had was that everybody agreed there was a ranking and disagreed violently about their own place in it.
This is the pattern for the whole of Part 5. The British were consistently, embarrassingly wrong about the explanation, and consistently correct that there was something there to explain.
4Born Guilty
In 1871 the Government of India passed a law declaring that certain communities were criminal by birth. Its shadow has not lifted.
Of everything the colonial state did to Indian society, this is the one that is hardest to argue about, because there is nothing to interpret. The law says what it says.
The Act empowered the government to notify entire communities as criminal. Once notified, the effects were sweeping.
Members could be registered — recorded by name in police lists. Their movement was restricted; they could be required to report to the police, sometimes at fixed intervals, sometimes daily. They could be confined to settlements. Their children could be removed. Later amendments extended the reach and the surveillance.
No individual offence was required. Membership of the community was the qualifying condition.
Read the logic of that. The state took the caste principle — that birth determines what you are — and gave it the force of criminal law and the machinery of a modern police system.
Independent India repealed the Criminal Tribes Act, and the affected communities were formally "denotified" — which is why they are known today as Denotified Tribes, alongside Nomadic and Semi-Nomadic Tribes.
They number in the region of 150 communities and tens of millions of people. Successive commissions have documented that they continue to face police suspicion, difficulty obtaining documents, and severe disadvantage in housing and employment.
Their name in Indian administrative language is still defined by a law repealed three-quarters of a century ago.
It proves that the colonial state did not merely record caste. In this instance it manufactured a category, gave it legal force, and inflicted enormous, documented, lasting harm. Anyone who says the British were passive observers of Indian society has to explain this Act.
It does not prove that Britain created caste. The Act worked because the underlying idea — that birth determines character — was already the organising principle of the society it was imposed on. A law declaring communities hereditarily criminal would have been unintelligible in a place without hereditary categories.
The colonial state's most damaging interventions were nearly always of this kind: taking an existing Indian principle, and giving it the reach of a modern state.
Because Denotified Tribes are among the most deprived communities in India and among the least visible in this argument. They are frequently not on the Scheduled Caste list, frequently not on the Scheduled Tribe list, and often classified as OBC where they compete with far more powerful communities.
When Part 13 asks who the current categories miss, this is one of the largest answers.
5The Colonial Construction Thesis
The academic argument that modern caste is substantially a British product — at its strongest.
The most influential statement of this position is Nicholas Dirks's Castes of Mind: Colonialism and the Making of Modern India (2001), building on earlier work by Bernard Cohn on the census and colonial knowledge.
1. Before colonialism, the king mattered more than the priest. In pre-colonial India, political power — kings, chiefs, local rulers — was central to social standing. Rank was awarded, contested and revised through political relationships. It was not a fixed religious ladder.
2. Colonial rule removed the political layer. The British dismantled or hollowed out indigenous kingship. What was left, once the political dimension was gone, was the ritual dimension — and colonial observers, seeing only that, concluded that caste was essentially a religious phenomenon. Their description then became the reality, because they administered on the basis of it.
3. Knowledge was administration. Censuses, gazetteers, ethnographic surveys and tribal handbooks were not neutral descriptions. They were instruments of rule, produced to make a vast society legible to a small bureaucracy, and they simplified enormously in order to do so.
4. The courts made one book into the law. As covered in Part 3, colonial courts adopted Sanskrit texts, above all the Manusmriti, as the basis of "Hindu law" — giving a single treatise an authority across a subcontinent that no pre-colonial Indian state had given it.
5. Modern caste politics is a colonial creation. Caste associations, caste-based demands, reserved seats and quota lists all emerge in this period — and emerge because the state had made caste the unit through which resources were distributed.
Conclusion: caste as a national, enumerated, legally defined, politically mobilised system is a modern object with a colonial origin — however old the underlying hierarchy may be.
Points 3, 4 and 5 are simply true
Worth saying plainly before Chapter 6, because this thesis is often waved away as fashionable revisionism. The census did simplify and fix. The courts did elevate Manu. Caste politics did take its modern form here, for exactly the reason given. None of that is disputed by anyone who has read the documents. The argument is about points 1 and 2 — and how far the conclusion follows.
6The Rebuttal
Equally strong. And a fairly precise verdict on which parts of each case survive.
1. The eyewitnesses. Part 4 laid these out. Megasthenes around 300 BCE, Fa-Hien around 400 CE, Xuanzang around 630 CE, Al-Biruni around 1030 CE — all describing hereditary rank, restricted marriage, inherited occupation and people required to live outside settlements. None of them was British, and none had a colonial administrative interest.
2. The genome. Endogamy — the actual lock — closed roughly two thousand years ago. Whatever the British did in 1871, they did not persuade Indians to start marrying within their own communities. That was already eighteen centuries old.
3. The petitions defeat the thesis they are cited for. Thousands of communities wrote to the Census Commissioner. Not one significant campaign demanded that the ranking be abolished as an alien imposition. They all wanted to be placed higher in it. People do not fight that hard for position within a hierarchy they regard as a foreign invention.
4. The texts predate the empire by two thousand years. The Dharmashastras were not written in Whitehall.
5. Indians were legislating on caste before the British were. Chapter 7. The first reservation policy anywhere came from an Indian king in 1902, in response to a caste monopoly that was already entrenched. He was not reacting to a British invention; he was reacting to his own administration.
6. And the thesis has an uncomfortable political use. "Caste is a colonial construct" can function as a way of saying that Indian society bears no responsibility for it — which is precisely the alibi Part 4 identified. Not every scholar advancing the thesis intends this. Many people repeating it online do.
The verdict
On whether Britain created the hierarchy: the rebuttal wins decisively. The hierarchy is documented by neutral outsiders across thirteen centuries before the East India Company existed, and the endogamy that sustains it is two thousand years old.
On whether Britain created the modern form of caste: the construction thesis largely wins. Enumeration, national categories, fixed lists, a legal definition, official rankings, caste as a unit of political claim-making — all of these are colonial in origin, and all of them are what an Indian in 2026 actually deals with. Your caste certificate is a colonial artefact. The argument about the census is a colonial argument.
The formulation I think the evidence supports: Britain found a social hierarchy and converted it into an administrative system. The injustice was Indian. The bureaucracy is British. And because reservation is a bureaucratic remedy, it inherited the British machinery to address the Indian injustice.
Which is why the two are so hard to separate now, and why "the categories are colonial" is both entirely true and completely useless as an argument for abolishing them.
7The First Reservations Were Indian
Not 1950. Not Mandal. Not the British. An Indian king, in 1902, twenty years before anything comparable existed anywhere in the world.
This chapter contains the fact most often missing from the entire argument, and it is missing from both sides.
Shahu Maharaj, the ruler of the princely state of Kolhapur, issued a notification in the state gazette reserving half of all vacancies in state service for candidates from backward communities.
The order provided that of all posts falling vacant from the date of the proclamation, 50% should be filled from the backward classes — and that in any office where backward-class employees numbered fewer than half, all subsequent appointments should go to them until the balance was reached.
It is widely regarded as the earliest affirmative action programme of its kind anywhere in the world.
Why he did it
Not as a theory. As a response to what he could see in his own offices.
The administration of Kolhapur was overwhelmingly staffed at senior levels by a single community. Shahu Maharaj's conclusion was that a small number of dominant castes had monopolised education and government employment, and that almost everyone else in his state was simply locked out.
His remedy was not persuasion, not scholarships, and not a committee. It was a quota, with a mechanism for filling it.
And he was not alone
It disposes of "reservation is a foreign idea." It was conceived by an Indian ruler in an Indian state, in response to Indian conditions, before Britain had done anything comparable and before most of the world had thought of it.
It disposes of "reservation was invented for vote-bank politics." Kolhapur in 1902 was a monarchy. Shahu Maharaj had no election to win and no constituency to court. Whatever his motives, they were not electoral, because there were no votes.
And it complicates the pro-reservation account too. The original policy was not framed as compensation for historical injustice, nor as a remedy for untouchability. It was framed as breaking a present-day monopoly in employment — a specific, measurable, current problem in a specific set of offices.
That is a narrower and more concrete justification than the one usually offered today, and arguably a stronger one, because it is testable. Part 11 tests it.
8The Road To Poona
Between 1909 and 1932, one question came to dominate Indian politics: when power is transferred, who counts as a group?
As the British began, slowly and reluctantly, to concede elected representation to Indians, a question appeared that had no precedent and no obvious answer. Elections require constituencies. Constituencies require deciding who votes with whom.
And in a society organised by community, that decision was not a technicality. It was everything.
What the Award actually gave
This is the detail that makes the next chapter comprehensible, and it is usually skipped.
Under the Award, Depressed Class voters would vote in their own separate constituencies to elect their own representatives — and would also vote in the general constituencies.
The consequence was that their representatives would be chosen exclusively by them, and could not be selected, funded or removed by the far larger general electorate.
That is the prize Ambedkar had spent a decade arguing for, and the reason is straightforward. A reserved seat inside a joint electorate is filled by a candidate whom the majority is willing to accept. A separate electorate produces a representative the majority has no say over at all.
He believed, on the evidence of everything in Parts 3 and 4 of this series, that a community which had been excluded for two thousand years could not rely on the goodwill of the people who had excluded it.
Gandhi's position is usually reported as simple opposition. It was more specific than that, and it needs stating properly.
He accepted that untouchability was a monstrous wrong and had campaigned against it. What he refused was the constitutional separation of the Depressed Classes from the rest of Hindu society.
His reasoning: separate electorates would make the division permanent and legal. Caste Hindus would lose any electoral reason to care about Dalit voters, since those voters would no longer be in their constituencies. The reform of Hindu society from within — which he regarded as the only durable solution — would become impossible, because the two groups would never again need each other's votes.
He also, and this is the part rarely acknowledged, feared it would do to Hindu society what separate electorates had already begun doing between Hindus and Muslims. Given what happened in 1947, that fear cannot be dismissed as merely tactical.
Both men were arguing about the same risk from opposite ends. Ambedkar feared that without separation, Dalits would be permanently dependent on the goodwill of their oppressors. Gandhi feared that with it, they would be permanently outside the society that had to change. Neither was being unreasonable. That is what made the collision unresolvable.
9The Poona Pact
A fast unto death in a prison cell, six days of negotiation, and the document from which every argument in 2026 descends.
On 20 September 1932, in Yerwada Central Jail in Poona, Mohandas Gandhi began a fast unto death. His condition for ending it was the removal of separate electorates for the Depressed Classes from the Communal Award.
He was 62 and in prison. Ambedkar was 41, and had spent his life arguing for exactly what Gandhi was now fasting against.
The position Ambedkar was placed in
It is worth being precise about this, because it is the most consequential week in the history of Indian reservation and it is usually narrated as a story about Gandhi.
Ambedkar had won. The Award was announced. Separate electorates for the Depressed Classes were British government policy.
And then the most revered man in India announced he would die unless that policy was reversed — with the whole country watching, and with the entirely predictable consequence that if Gandhi died, the blame would fall on Ambedkar and on the community he represented.
What was signed
Signed at Yerwada Central Jail. Ambedkar signed on behalf of the Depressed Classes; Madan Mohan Malaviya signed on behalf of caste Hindus.
Separate electorates were given up. They do not appear in the document.
In exchange, reserved seats were roughly doubled. The Award's 71 seats in the provincial legislatures became 148 — but as reserved seats inside joint electorates, filled by all voters in the constituency rather than by Depressed Class voters alone. A share of seats in the central legislature was also provided for.
Gandhi ended his fast on 26 September 1932.
The trade, in one line
Ambedkar exchanged political independence for numerical quantity. More seats, less autonomy in filling them.
Everything about modern Indian reservation is downstream of that trade.
The mechanism is the Pact's mechanism. Reserved seats within a common pool, rather than a separate pool — which is exactly the structure described in Part 1. It was settled here, in 1932, and it has never been revisited.
The critique of it is also from here. Ambedkar's own later assessment was bleak: that representatives elected by a general electorate would be acceptable to the majority rather than accountable to the minority, producing what critics still call nominal representation. Whether that prediction came true is a live empirical question, and Part 11 examines it.
And the moral shape of the argument was set here too. The Pact was agreed under a threat that Ambedkar did not create and could not refuse. That is why "Dalits agreed to this arrangement" is such a contested sentence, and why the Pact is commemorated by some as a moment of national unity and by others as coercion.
Both descriptions are accurate. It was a genuine agreement, and it was signed under duress. Those are not contradictory.
Gandhi is quoted by people who want to argue that Dalits should be integrated rather than separately protected. Ambedkar is quoted by people who want to argue for stronger safeguards. Both are also quoted, selectively, by the abolition side.
Part 7 gives each of them a full hearing in their own words, at length, including the parts their admirers leave out. Neither man was simple, and neither is available to be borrowed as cheaply as he usually is.
10What Britain Actually Left Behind
Not an attitude. A filing system — and it is still the one India uses.
The colonial period ends with a specific inheritance, and it is more concrete than most of this argument allows.
The Act provided for reserved seats and used the term that India still uses today. The following year, the Government of India (Scheduled Castes) Order 1936 set out the actual schedule — the list of which communities qualified.
That list is the direct ancestor of the Constitution (Scheduled Castes) Order, 1950, which is the instrument discussed at length in Part 4 — the one that still determines who is legally a Dalit in 2026.
The five things independent India inherited intact
- The word. "Scheduled Caste" is a colonial administrative term, coined for a 1935 statute. Nobody chose it as a description of a people; it means "the castes on the schedule."
- The list. A specific enumeration of communities, compiled by colonial officials, largely carried forward in 1950.
- The mechanism. Reserved seats inside joint electorates — the Poona Pact structure.
- The counting apparatus. The census, and the assumption that caste is something the state records and acts upon.
- The habit of legislating by category. Group entitlements attached to group membership, adjudicated by the state.
The Reservation Hatao Andolan's central proposal is that the state should stop recording and recognising caste — that the column should be deleted, and with it the category.
Everything in this book supports one half of that argument and destroys the other.
The half it supports: the categories really are colonial. The list, the word, the schedules, the certificates — all of it is the residue of an imperial filing system built to make Indians legible to their rulers. It is entirely fair to find that grotesque.
The half it destroys: the thing being filed was not created by the filing. Al-Biruni described it in 1030. The endogamy lock closed two thousand years ago. Shahu Maharaj was fighting a caste monopoly in 1902 that no British official had built.
Deleting the record does not delete the thing recorded. It deletes the ability to measure it — which is a real proposal with real consequences, and Part 14 assesses it properly. But it should be argued as what it is: a proposal to stop counting, not a proposal to end caste.
11Nine Questions Nobody Asks
The blind spots in this stage of the argument.
1. Would caste have faded without the census?
The strongest form of the construction thesis implies it might have. Local, negotiable, unenumerated identities might have dissolved under urbanisation and industry, as European estate distinctions largely did.
The reason to doubt it: endogamy was two thousand years old and required no state at all. Nothing in that mechanism depended on being counted. But this is a counterfactual, and anyone who is certain about it is guessing.
2. Why did every community petition to move up rather than to abolish the ladder?
Thousands of petitions, and essentially none arguing that ranking communities was illegitimate.
It is the single most revealing fact about colonial caste, and it is uncomfortable for the construction thesis. People do not litigate for decades over their position in a hierarchy they believe a foreigner made up last Tuesday.
3. Did counting create the interest, or reveal it?
Caste associations formed in response to the census. But an association is an organisational form, not a sentiment. The question is whether the census gave people a new interest or gave an existing interest a new address to write to.
The petitions suggest the latter. What was new was not the desire for rank. It was the discovery that rank could now be appealed.
4. What happened to communities that fell between the categories?
The census demanded exclusive answers. Communities that had been genuinely intermediate — claiming one status in one district and another elsewhere, or shading into a neighbouring group — had to be placed somewhere.
Some of those placements determined, eighty years later, whether a community would be eligible for reservation. Decisions taken by a district officer with a deadline are still producing consequences.
5. Why is the Criminal Tribes Act so rarely mentioned?
It is the clearest documented case of the colonial state actively manufacturing caste harm rather than merely recording it. It should be Exhibit A for the construction thesis.
It is barely cited, in academic argument or in public debate — and the communities it damaged remain among the least represented in Indian politics. Those two facts are related.
6. Why does nobody credit Shahu Maharaj?
The world's first reservation policy was Indian, from 1902, and almost nobody in this argument mentions it.
The reason is that the fact is unhelpful to everyone. It denies the abolition side "reservation is a foreign import" and "it was invented for votes." It denies the retention side its founding narrative, which starts with Ambedkar and the Constitution. A fact that serves no faction gets dropped by all of them.
7. Was Gandhi's fast a moral act or coercion?
The answer people give tracks their prior view of Gandhi almost perfectly, which is a sign that nobody is reasoning about it.
The honest description contains both. He was risking his own life, not threatening anyone else's — that is a real moral distinction. And the practical effect was that Ambedkar had to concede or be blamed for a death, which is a real form of pressure. Insisting on only one half is where the dishonesty starts.
8. What would India look like if the Communal Award had stood?
Dalit representatives elected only by Dalit voters, for ninety years, entirely independent of the parties. It is arguably the biggest counterfactual in modern Indian political history and it is almost never explored seriously.
It might have produced a genuinely autonomous Dalit political bloc. It might equally have produced the permanent constitutional separation Gandhi feared. Nobody knows, and both possibilities should trouble whoever finds one of them comforting.
9. If the categories are colonial, what would replacing them look like?
"These categories are a colonial artefact" is true and is usually where the argument stops.
But any replacement needs its own categories. Economic criteria require income thresholds, verification and an administrative machine — as EWS already demonstrates. There is no way to distribute anything to anyone without defining who. The choice is never between categories and no categories. It is between one set and another.
12What The Evidence Can And Cannot Settle
The honest scorecard for Part 5.
| Question | Status | Best current answer |
|---|---|---|
| Did the British create the caste hierarchy? | No | Documented long before them; endogamy 2,000 years old |
| Did they create its modern administrative form? | Largely yes | Census, lists, legal categories, caste politics |
| Did the census fix fluid identities? | Yes | Exclusive answers, national categories, printed rank |
| Was Risley's racial theory of caste correct? | No | Abandoned; genetics does not support it |
| Did colonial law manufacture caste harm directly? | Yes | Criminal Tribes Act, 1871 |
| Where did reservation begin? | Kolhapur | 26 July 1902, Shahu Chhatrapati, 50% |
| What settled the mechanism India uses? | Poona Pact | 24 September 1932 — 148 seats, joint electorates |
| Where does the term "Scheduled Caste" come from? | 1935 Act | List drawn up in the 1936 Order |
| Would caste have faded without colonial rule? | Unknowable | Counterfactual; endogamy needed no state |
The three things nobody can tell you
- How much the census changed behaviour rather than records. We can see identities becoming fixed on paper. Whether the same families behaved differently at home is a separate question, and the household evidence does not exist.
- What Ambedkar would have accepted without the fast. He never had the chance to find out, and neither do we.
- Whether separate electorates would have worked. The largest unrun experiment in modern Indian politics.
Britain did not build the house. It surveyed it, numbered every room, printed the plan, and handed the keys to a government that has been living in it ever since.
Which means "the categories are colonial" is true, and is an argument for redesigning them — not for pretending the building is not there.
13The Five-Line Nichod
Everything in this book, compressed. If you remember nothing else, remember these.
PART 5 — THE BRITISH MACHINE
- The census did not create caste — it created the caste system you actually deal with. From 1871, exclusive answers, national categories and printed rankings turned local, negotiable identities into fixed legal ones. Your caste certificate is a colonial artefact. The thing it certifies is not.
- Risley was wrong about everything except that there was a hierarchy. The Census Commissioner for 1901 believed caste rank could be read from the width of a nose. The theory is dead. The pattern for the whole colonial period is the same: embarrassingly wrong about the explanation, correct that there was something to explain.
- The Criminal Tribes Act of 1871 declared whole communities criminal by birth. Repealed in 1952; the affected communities are still called Denotified Tribes, still number tens of millions, and are still among the least visible people in this entire argument.
- The world's first reservation policy was Indian. 26 July 1902 — Shahu Chhatrapati of Kolhapur reserved 50% of state posts for backward classes, in a monarchy with no elections, to break a monopoly he could see in his own offices. Neither side in 2026 mentions him, because the fact helps neither.
- Everything modern descends from six days in September 1932. Gandhi fasted from the 20th; the Poona Pact was signed on the 24th; he broke his fast on the 26th. Ambedkar gave up separate electorates and got 148 reserved seats inside joint electorates instead. That trade — more seats, less autonomy — is still the mechanism, and still the argument.
Coming next
The part this whole series has been building towards, and the question that started it. No inference, no hymns, no genetics. Court records, land records, temple records, missionary accounts, travellers' accounts and colonial files — primary documents, quoted directly. Mahad 1927, Vaikom 1924–25, Kalaram 1930, dated and sourced. What the breast-tax story does and does not support. Modern measurement: what surveys find about untouchability still being practised in Indian homes, and what the atrocity data shows along with its honest caveats. Field experiments with identical CVs and different surnames.
And the "it is all exaggerated" case, steelmanned properly — where it has a real point, and where it collapses.
A closing thought: the most useful thing in this part is the smallest. A king in 1902 looked at a list of who worked for him, noticed that one community held almost everything, and did something about it. No theory of history, no compensation for the past — just a monopoly, observed and broken. That is a narrower argument for reservation than the one usually made, and it is the one that can actually be tested. — L.S.
Sources & further reading — Part 5
Every factual claim in Part 5, and where it comes from.
The census and colonial ethnography
First all-India census, 1871–72; decennial censuses thereafter. · H. H. Risley, Census Commissioner for the 1901 census; the attempt to classify castes by social precedence; anthropometric method including the nasal index; The Tribes and Castes of Bengal (1891) and The People of India (1908). · The formation of caste associations and the volume of petitions for reclassification are documented in the census reports themselves and in the subsequent historiography. · The 1931 census under J. H. Hutton was the last comprehensive caste enumeration; caste enumeration returns in Census 2027.
The Criminal Tribes Act
Criminal Tribes Act, 1871, and subsequent amendments — notification of communities, registration, restriction of movement, reporting requirements and settlements. Repealed in 1952; affected communities formally denotified. Denotified, Nomadic and Semi-Nomadic Tribes number in the region of 150 communities; successive government commissions have documented continuing disadvantage and police suspicion.
The colonial construction debate
Nicholas B. Dirks, Castes of Mind: Colonialism and the Making of Modern India (2001) — the fullest statement of the construction thesis, including the argument about the displacement of kingship by ritual as the organising principle. · Bernard S. Cohn on the census and colonial forms of knowledge. · Susan Bayly, Caste, Society and Politics in India from the Eighteenth Century to the Modern Age (1999), for a more qualified account. The rebuttal in Chapter 6 draws on the pre-colonial eyewitness evidence set out in Part 4 and the genetic dating set out in Part 2.
Early reservation
Shahu Chhatrapati of Kolhapur, notification in the state gazette dated 26 July 1902, reserving 50% of vacancies in state service for backward classes, with the provision that where backward-class employees numbered fewer than half, subsequent appointments should go to them. Widely described as the earliest affirmative action programme of its kind. · Mysore: the committee under Sir Leslie Miller, 1918. · Madras Communal Government Order, 1921.
The road to Poona, and the Pact
Morley–Minto reforms, 1909 — separate electorates for Muslims. · Montagu–Chelmsford reforms and the Government of India Act 1919. · Simon Commission, 1927–28; Ambedkar's evidence. · Round Table Conferences, 1930–32. · Communal Award, 16 August 1932, announced by Prime Minister Ramsay MacDonald — separate electorates for Muslims, Sikhs, Europeans, Anglo-Indians, Indian Christians and the Depressed Classes, with 71 reserved seats in the provincial legislatures for the Depressed Classes, together with the right to vote in general constituencies. · Gandhi's fast begun 20 September 1932 at Yerwada Central Jail. · Poona Pact, 24 September 1932 — signed by Ambedkar on behalf of the Depressed Classes and by Madan Mohan Malaviya on behalf of caste Hindus; separate electorates abandoned; reserved seats in the provincial legislatures increased from 71 to 148 within joint electorates, with provision for the central legislature. Gandhi ended the fast on 26 September 1932.
The inheritance
Government of India Act 1935 — introduction of the term "Scheduled Castes" and provision for reserved seats. · Government of India (Scheduled Castes) Order, 1936 — the schedule of communities. · Constitution (Scheduled Castes) Order, 1950 — discussed in full in Part 4.
A note on what is missing — and on corrections
Part 5 makes no claim about how severely caste rules were applied to real people, which is Part 6, nor about the Constituent Assembly debates, which are Part 8. Gandhi's and Ambedkar's positions are summarised here only as far as the Poona Pact requires; both men are given full and lengthy treatment in their own words in Part 7. Anything marked CONTESTED CLAIM stays contested until evidence is produced.
Where accounts of the Poona Pact differ slightly on seat numbers for the central legislature, this book gives only the provincial figures, which are not in dispute. If you find an error, I want to know. Corrections will be carried in later parts.