Where We Left Off
Before we begin
Part Two went looking for where the rules about women came from, and found a chain that runs through property rather than through morality. A field cannot be carried. It has to be inherited. A man wants it to go to his children and — until 1985 — could not verify which children those were. Societies solved that verification problem in one of three ways, and the one that spread was the one that let a man pass wealth to children he raised while staying confident they were his. That solution is called controlling the woman.
We also found the limits of that story. The biology is real and explains almost none of the variation: societies facing identical facts built opposite institutions, including several in India where property passed through daughters and nobody cared who a woman slept with. We found that a farming tool still predicts women’s employment four thousand years later, even among the grandchildren of migrants. We found that dowry is a bid in a marriage market, which is why banning it in 1961 made it grow. And we found that religion did not create these rules — it arrived thousands of years too late — but converted negotiable local customs into non-negotiable ones.
Then Part Two ended on an audit, and the audit left two problems standing. One was old age: India has no real pension, so a son is the pension and a daughter is a cost whose return goes to another family. The other was stranger. Paternity became verifiable in 1985 and not one society relaxed anything. Which means paternity was never the only load on the machine — and in India, it may never have been the main one.
This part is about the other load. It is the part where the series stops describing rules that exist nearly everywhere and starts describing the machinery that exists here.
Two things in it will be unwelcome, and they will be unwelcome to opposite people. The first is that the mechanism which keeps caste alive was identified in 1916, by an Indian, in precise terms, and it runs entirely through the control of women’s marriages — which means the thing most Indians argue about when they argue about caste is not the thing that reproduces it. The second is that a substantial part of what is now defended as timeless Indian modesty can be dated, and some of it has an English birth certificate.
How to Read the Boxes
The notation, in case this is where you started
Six kinds of box run through the series. Here is one live example of each.
Endogamy: the rule that you must marry inside your own group. Its opposite is exogamy, marrying outside.
Most societies have some version. What makes the Indian case unusual is not that endogamy exists but how many separate groups are simultaneously enforcing it, how small some of them are, and how long it has been running.
Why it matters here: this one word is the subject of the entire part. If you take nothing else from these ten chapters, take this: the Indian machine is an endogamy machine, and women are the component it operates on.
A Word Box appears the first time a hard word does — never later, never only in a glossary.
Roughly one Indian marriage in twenty crosses a caste line. That figure has moved very little across decades of surveys.
Put it in a room. Take a hundred married couples at a wedding hall. Ninety-five of them married inside the group they were born into. Five did not.
Now put it against something familiar. In most large countries, the share of marriages crossing whatever the main social boundary is — religion, race, class — has risen substantially over the same period. India’s has barely moved. That is not a leftover. That is a working machine.
Any number too big to picture gets a body attached — a room, a wage, a walk, a hundred couples.
You would expect endogamy to be measurable only through surveys, which people can lie to. It turns out to be measurable in blood.
When a group marries only within itself for many generations, everybody inside comes to share a larger portion of their DNA. Geneticists can detect this, estimate how long it has been happening, and identify when a population stopped mixing with its neighbours.
What it shows: across large samples of Indian populations, widespread mixing ends and strict within-group marriage begins somewhere in the range of two to four thousand years ago, and has continued since. Many groups show founder effects — the genetic signature of descent from a small number of ancestors — stronger than those found in Ashkenazi Jewish and Finnish populations, both of which are studied precisely because those effects are pronounced.
What it cannot show: who enforced it, why, or what it felt like. Genetics gives you the fact of the boundary and its approximate age. It says nothing about the machinery, which is what Chapter Two is for.
This box names the actual evidence and then says what it cannot show.
A framing question for the part. Almost all public argument in India assumes the first answer.
Caste is a system of ranked occupations and inherited disadvantage. That is where the damage is: in who gets which job, which school, which house, which treatment from the police. Public policy, reservation, and the entire modern politics of caste address this, correctly, because this is what ruins lives.
Occupational ranking is the output. The thing that reproduces caste across generations is the rule against marrying out — because a boundary that is never crossed is a boundary that never dissolves. Occupations have changed enormously in a century; the marriage rule has barely moved. Whatever survived that change is the load-bearing part.
What would settle it: whether caste identity weakens where inter-group marriage rises. Where it has risen, it does. That is a small dataset and it points one way.
An Argument box appears where people who have studied something genuinely disagree. Each side gets its strongest case.
Both sides above assume that caste is a thing a person has — a property of an individual, like a height or a surname.
But nobody has a caste alone. It exists only as a relation between families who will and will not exchange daughters. A single person on a desert island has no caste; a hundred families with a rule about marriage have one between them.
The general form: treating a relationship as a property. It matters practically. If caste is a property of persons, then policy addresses persons — quotas, protections, penalties. If caste is a rule about exchange between families, then it survives every policy aimed at individuals, indefinitely, which is roughly what has happened.
This is the signature box. It digs out something both sides take for granted while they fight. There are five in this part.
Every chapter closes with one of these, restating it in the plainest words available, key terms in bold.
Read only these and you should still finish holding the whole argument.
A Note on This Part
Read this before Chapter One
Three warnings, and this time they matter more than usual.
This material is politically live in India in a way the previous parts were not. Caste, colonial history and the origins of Indian sexual norms are all currently contested territory, and every finding in this part is already being used as a weapon by somebody. I am going to report what the evidence shows and decline to hand the conclusion to any camp. Where a finding is popular with one side, I will say what that side gets wrong with it too.
The chronology in Chapters Five and Six cuts in more than one direction, and I will point that out where it happens. The finding that some Indian conservatism has colonial origins is used by liberals to argue that the conservatives are not authentic, and by nationalists to argue that pre-colonial India was superior. It supports neither conclusion, and Chapter Nine turns the whole argument back on itself, including on me.
My own position, again. I am an Indian man. I was born into a caste, as everyone reading this in India was, and I have not been on the receiving end of a single rule in this part. In Chapter Two I am going to set out an argument made by B. R. Ambedkar, who was on the receiving end of most of them. His analysis is the spine of this part. Nothing about my reproducing it makes me a good judge of it, and I have tried to present his reasoning rather than my opinion of his reasoning.
1The Machine Nobody Names
India argues about caste constantly. It argues about jobs, quotas, discrimination and insult. It very rarely argues about the one rule that keeps the whole thing reproducing itself — and that rule is about women.
1.1 — Not a ladder. A set of closed circles.
Most people, Indian and foreign, picture caste as a ladder. Groups ranked from top to bottom, with occupations attached and privileges distributed accordingly.
That picture captures the harm and misses the machine.
Look instead at what a caste actually is in daily practice. Ask what a person cannot do because of theirs, and you find that most of the historical prohibitions have weakened. People of different castes now work in the same offices, eat in the same restaurants, travel in the same carriages, drink the same water. Enormous violations of custom happen in every Indian city, hourly, and nothing collapses.
One prohibition has not weakened. You marry inside.
That is the circle, and the circle is the thing. A caste is not primarily a rank. It is a group of families who will exchange daughters with each other and will not exchange daughters with anybody else. Everything else — the occupations, the ritual rules, the food restrictions — is decoration on that, and much of the decoration has come off in the last century while the frame has not moved at all.
Jati: the actual unit people belong to and marry within. There are thousands of them, often specific to a region and a language, and they are what people mean in ordinary speech when they say caste.
Varna: the fourfold classical scheme found in ancient texts. Four categories for a country containing thousands of marrying groups.
Why it matters here: arguments about caste in India constantly slide between the two. The texts describe varna. Life is organised by jati. A defence of the texts is not a defence of the system anybody actually lives in, and an attack on the texts does not touch it either.
1.2 — The number that has not moved
How closed are these circles, in measurable terms?
Large household surveys in India have asked married people whether their spouse belongs to their own caste. The answer, consistently, is that around one marriage in twenty crosses a caste line. Estimates differ a little by survey and definition, and the figure has moved only slightly across decades.
Set that beside almost anything else about India over the same period. Literacy transformed. Urbanisation transformed. The share of people working in agriculture collapsed. Life expectancy nearly doubled. Television, then mobile phones, then the internet arrived in villages that had no electricity a generation earlier.
Through all of it, ninety-five marriages in a hundred stayed inside the circle.
Anything that survives that much change is not a residue. It is load-bearing, and something is actively holding it there.
1.3 — A boundary you can read in blood
Here is the part that turns an argument about culture into a matter of physical evidence.
When a population marries only within itself for many generations, its members come to share more of their DNA with each other than with outsiders. This is measurable. From the pattern, geneticists can estimate how long a group has been marrying inside, and when it stopped mixing with its neighbours.
The work has been done on large numbers of Indian populations, and it produces two findings that no survey could give you.
First, there is a date. Widespread population mixing across the subcontinent continues for a long period and then, somewhere in the range of two to four thousand years ago, largely stops. After that, groups marry inside themselves. The boundary has a beginning, and the beginning is not lost in the mists of time; it is roughly datable.
Founder effect: what happens genetically when a population descends from a small number of ancestors and does not mix with outsiders afterwards. Rare inherited conditions carried by one of those ancestors become common inside the group, because everybody inside is descended from them.
Why it matters here: it converts a social rule into a medical fact. A marriage boundary maintained for eighty generations is not only a social arrangement — it is a mechanism that concentrates inherited disease, and the concentration falls on children who had no part in the rule.
Second, there is a cost. Many Indian groups show strong founder effects — the genetic signature of a population descended from a small number of ancestors. Some show effects more pronounced than in Ashkenazi Jewish and Finnish populations, both of which are studied intensively precisely because their founder effects produce elevated rates of certain inherited diseases.
The method compares stretches of DNA shared between individuals. Two people from a group that has married inside itself for eighty generations share long, distinctive segments that two random Indians do not.
What it can show: the fact of a closed boundary, an approximate date for when it closed, and how severe the founder effect is. Independent teams using different samples get compatible answers, which is the strongest kind of support available.
What it cannot show: anything about why. Genetics reports that the doors shut. It has nothing to say about who shut them, what arguments were used, or who objected. And the dates carry real uncertainty — a range of two to four thousand years is a range, not a year, and different groups closed at different times.
The finding that gets least attention: the recessive disease burden. Marrying inside a small group for a very long time concentrates rare inherited conditions. This is not a moral argument against endogamy. It is a medical consequence of it, it is measurable, and it falls on children.
1.4 — What India argues about instead
Now hold that against the shape of Indian public debate.
Caste is discussed constantly and passionately. Reservations in education and employment. Discrimination in housing and hiring. Violence. Political mobilisation. Representation. Insult and dignity.
Every one of those is real, every one matters, and every one concerns the output of the system — the harm it produces in a given generation.
Almost none of it concerns the mechanism by which the system arrives in the next generation. There is no serious mass politics of inter-caste marriage in India. No major party campaigns on it. It is not a manifesto item. Individual couples who cross the line are frequently in danger, and the danger is discussed as a law-and-order problem rather than as the central fact about how the whole structure sustains itself.
I want to be careful here, because this can be read as saying that fighting discrimination is a distraction. It is not, and I do not think that. Discrimination is where the suffering is, and addressing suffering does not require first addressing its cause.
The observation is narrower and stranger: India has spent a century arguing about caste while leaving almost untouched the one practice without which caste could not exist for another two generations.
Somebody noticed this in 1916. He was twenty-five, he was sitting in a seminar room in New York, and his explanation of it is the spine of this part.
Caste is usually pictured as a ladder. That captures the harm and misses the machine. Most of the old prohibitions — on shared work, food, water, travel — have weakened enormously. One has not: you marry inside.
A caste is best understood as a set of families who will exchange daughters with each other and with nobody else. Everything else is decoration on that frame, and much of the decoration has come off while the frame has not moved.
Roughly one marriage in twenty crosses a caste line, and that number has barely shifted through a century in which literacy, cities, work and life expectancy all transformed. Whatever survives that much change is load-bearing.
The boundary is visible in DNA. Widespread mixing stops somewhere between two and four thousand years ago, and many groups show founder effects stronger than those in populations studied specifically for their inherited disease burden. That cost falls on children.
India argues about caste constantly and almost never about marriage — which is to say, it argues about the output and leaves the reproduction mechanism alone.
2Ambedkar’s Answer
In May 1916 a young Indian student read a paper to an anthropology seminar at Columbia University. In it he worked out why a closed marriage circle produces sati, enforced widowhood and child marriage — not as three cruelties, but as one solution to one arithmetic problem.
2.1 — The problem he started from
B. R. Ambedkar was twenty-five. He would later be the principal architect of India’s constitution and the most consequential critic of caste the country has produced. At this point he was a graduate student, and the question he set himself was mechanical rather than moral: what actually keeps a caste in existence?
His answer began with a distinction. Many societies have groups. What makes an Indian caste different, he argued, is that the group is closed by marriage — and that this closure is not a side effect of the group but the thing that constitutes it. Strip away everything else and a caste is an enclosed circle of marriage. Remove the enclosure and there is no caste left to have.
Then he asked the question that makes the paper remarkable. If you wanted to keep such a circle closed, what would you actually have to do?
2.2 — The arithmetic
Here is the problem, and it is worth working through slowly because everything follows from it.
A closed marriage circle needs its marriageable men and marriageable women to stay roughly in balance. Everyone must be able to find a partner inside. If they cannot, some of them will go outside, and the circle breaks.
Death disturbs this balance constantly. Whenever a married person dies, their spouse becomes surplus — a person of marriageable status, inside the circle, with no partner.
Ambedkar’s insight was that the two kinds of surplus create different problems and therefore attract different solutions.
The surplus woman. A widow. She is a problem in two ways. If she remarries outside the group, the boundary is breached. If she remarries inside it, she competes for the limited supply of available men, which leaves some other woman unmarried and pushes her towards the boundary. Either way the circle is under strain.
The surplus man. A widower. He creates the same imbalance in reverse. But — and this is the pivot of the whole argument — he is a man, in a society where men hold the power, and a group cannot easily impose a sacrifice on the people who enforce its rules.
2.3 — The three solutions
So what does a society do? Ambedkar answered by looking at the customs that actually existed and asking what problem each one solved.
Sati: the burning of a widow on her dead husband’s funeral pyre. The word is also used for the woman herself.
It was never a universal Indian practice. It was concentrated in particular regions and particular communities, and the numbers involved were small relative to the population — though a small number of very public deaths is exactly what a deterrent requires. It was outlawed in 1829.
Why it matters here: read on its own it looks like an extreme religious observance. Read as part of the argument that follows, it is one of two available answers to a specific problem, and the other answer was applied far more widely.
For the surplus woman, solution one: remove her physically. A widow burned on her husband’s pyre is no longer surplus. The practice was never universal in India and was concentrated in particular regions and particular groups — which, as §2.5 shows, is itself evidence for the argument rather than against it.
For the surplus woman, solution two: remove her socially. If you cannot kill her, you can end her existence as a marriageable person. Enforced widowhood does exactly this and does it with precision.
What enforced widowhood meant in practice, in the communities that enforced it strictly, was a set of rules applied to a woman who might be fourteen.
Her head shaved, sometimes repeatedly. Coloured cloth and ornaments removed, replaced with plain white. Restricted food, often without spices, sometimes one meal a day. Exclusion from weddings, births and festivals, on the grounds that her presence brought misfortune. Frequently confinement to the household of her dead husband’s family, working in it without status.
Look at that list as a design problem rather than as cruelty, and notice what every item does. Shaving, white cloth, no ornaments, no festivals — each one removes a signal of sexual availability. This is not a set of punishments that happens to be harsh. It is a uniform, and the uniform announces to every family in the circle that this woman is not in the marriage market.
She might live another fifty years inside it.
For the surplus man: give him a girl. Celibacy could be prescribed and largely could not be enforced. So the practical solution was to supply a partner — which means the group needs unmarried girls available at short notice, at any time, which means girls must be married very young, which means marrying them before they are old enough to have any view on the matter.
That is child marriage, arrived at not through any theory about children but as the supply side of an inventory problem.
2.4 — What the argument actually claims
Set out plainly, the claim is this. Sati, enforced widowhood and child marriage are not three separate expressions of contempt for women. They are three components of one system, and the system exists to keep a marriage circle closed. Each solves a specific failure mode in the arithmetic.
And that is a much more disturbing account than the one it replaces.
Nearly everybody discussing these customs — those who condemn them and those who explain them away — assumes that the cruelty was the point.
The critic reads sati and enforced widowhood as expressions of hatred or contempt for women. The apologist replies that there was no hatred, that these were spiritual practices, misunderstood by outsiders. Both are arguing about attitude.
Ambedkar’s account makes attitude irrelevant. In his reading the suffering is not the objective. It is a by-product of solving an inventory problem in a closed group. Nobody has to hate anybody. The arithmetic produces the outcome regardless of what anyone feels.
Notice why this is worse rather than better. Contempt can be argued with — you can change what people feel about a group. Arithmetic cannot be argued with. As long as the circle must stay closed, the surplus problem recurs every time somebody dies, and something has to be done with the surplus. Improve everybody’s attitude and the problem is exactly where it was.
The general form: mistaking a by-product for a purpose. It is the most common error in reading any old institution, and it produces solutions aimed at the wrong thing.
2.5 — Does it hold up?
Ambedkar was working with the anthropology available in 1916, and parts of his framework have not aged well. It is worth separating what has survived from what has not.
His paper makes two distinct claims: one about how caste began, one about what maintains it. They have fared very differently.
Three predictions follow from it and all three check out. The customs should be strongest where endogamy is strictest — and they were, being enforced far more severely in groups with the most rigid marriage boundaries and the most property, while many other communities permitted widow remarriage throughout. The marriage rule should outlive the occupational system — and it has, comprehensively. And the genetic evidence should show boundaries closing and staying closed — which it does. A theory that predicts three independent things correctly is doing real work.
Ambedkar proposed that Indian society was originally one people, that a priestly group closed itself off first, and that caste spread outward by imitation. Modern genetics and history describe something far messier — a long period of mixing between distinct ancestral populations, boundaries hardening at different times in different regions, and no single point of origin. His account of how it started is a reasonable 1916 hypothesis and is not what the evidence now shows.
The functional argument can be correct and still be incomplete. Not everything about widowhood or child marriage reduces to marriage-market arithmetic; ritual purity, property law and simple inherited habit all contributed. Explaining a custom’s persistence by its function is a move that can be applied to almost anything, and it needs the specific predictions above to be more than a story.
What would settle it further: systematic comparison of widowhood and marriage-age customs against endogamy strictness across a large number of groups, using historical records. Some of this exists in fragments; nobody has done it at scale.
Why people care so much: because of what follows from it. If the mechanism claim is right, then the remedy for caste is not primarily economic or legal — it is marriage. Ambedkar said exactly that twenty years later, arguing that inter-caste marriage was the only real solvent. That conclusion is inconvenient to nearly everybody, which may be why the paper is honoured far more often than it is used.
2.6 — Why this is barely quoted
Ambedkar is among the most cited Indians in Indian public life. Statues, quotations, parties, holidays. And this particular argument — arguably his sharpest single piece of analysis — is almost absent from the conversation.
Consider who it inconveniences.
It inconveniences anyone defending tradition, because it reframes revered customs as inventory management for a closed group.
It inconveniences caste-based political movements, because it locates the mechanism in the family rather than in the state, and families do not respond to legislation.
It inconveniences reformers who prefer economic explanations, because it says the load-bearing element is who marries whom, not who earns what.
And it inconveniences almost every Indian family, including mine and probably yours, because the remedy it implies is not a policy anyone else has to enact. It is a decision about a daughter, taken at a kitchen table.
In 1916 Ambedkar asked a mechanical question: what keeps a caste in existence? His answer was that a caste is a closed marriage circle, and everything else follows.
A closed circle needs marriageable men and women in balance. Death breaks that balance constantly. Every death produces a surplus person who must be dealt with or the boundary leaks.
The surplus woman was removed physically (sati) or socially (enforced widowhood — the shaved head, white cloth, no ornaments, no festivals: a uniform announcing she is not in the market). The surplus man could not be made to sacrifice, so he was supplied with a girl — which requires girls married young. That is child marriage as an inventory problem.
Three customs, one system, one arithmetic problem. And that is worse than the cruelty explanation, because contempt can be argued with and arithmetic cannot.
His mechanism claim holds up and predicts three separate things correctly. His account of how caste started does not, and should be treated as superseded. His critics attack the second to dismiss the first.
He is among the most quoted Indians alive or dead, and this argument is almost never used — because the remedy it implies is not a policy anyone else has to pass. It is a decision taken at a kitchen table about a daughter.
3How the Rules Travel Downwards
The obvious picture is that the powerful impose restriction on the powerless. In India the movement runs the other way at least as often: groups adopt restrictions on their own women, voluntarily, in order to climb.
3.1 — The picture that does not fit
If a rule is imposed from above, you expect to find it strongest at the top and weakest at the bottom, with the bottom resisting.
That is not what the Indian record shows. For much of the period we can document, communities lower in the caste order permitted things that communities higher up forbade. Widow remarriage was accepted in a great many groups and prohibited in a smaller number near the top. Women in poorer and lower-status communities worked in fields, in markets and in trades, moved in public, and were not secluded — because their labour was needed and the household could not afford to do without it.
So restriction was, in a real sense, a luxury good. Keeping a woman out of the fields costs a household her earnings. Only a household that can afford to lose them can buy the status that comes with it.
Which sets up the question this chapter answers. If restriction was concentrated at the top, and if the practices at the bottom were looser, then why did those practices not simply survive as the country changed? Why did they spread downwards instead of the freer arrangements spreading up?
3.2 — Climbing by imitation
An Indian sociologist working in the 1950s gave this process a name after observing it in a district of southern India.
Sanskritisation: the process by which a lower-status group raises its position by adopting the customs, rituals and lifestyle of a higher-status one.
Named by the sociologist M. N. Srinivas in 1952. The word points to the practices associated with Sanskrit textual tradition, but the process is about status rather than about texts.
Why it matters here: it is the mechanism by which restrictions on women spread rather than merely persisting. A group that wants to be treated as higher does not petition for it. It changes its behaviour to look like the group it wants to be counted among.
What does a group actually adopt? The list is short and consistent.
Giving up meat. Giving up alcohol. Adopting the ritual markers of higher groups. And — the item that concerns us — changing what its women do.
Specifically: prohibiting widow remarriage where it had been permitted. Withdrawing women from paid work outside the home. Adopting seclusion norms. Switching from receiving a payment for a daughter to paying one, which is the shift from bridewealth to dowry described in Part Two.
Notice that most items on the list are cheap. Giving up meat costs a household very little. Giving up a woman’s earnings costs it a great deal — which is precisely what makes it a credible signal. Anybody can stop eating meat. Only a household with something to spare can stop needing a woman’s wages, and everybody watching knows it.
3.3 — The census made it worse
There is a specific historical accelerant, and it is one of the stranger consequences of colonial administration.
From the 1870s the British began conducting decennial censuses that recorded caste — and, in several editions, attempted to rank castes in order of social precedence. A written, official, national ranking now existed where before there had been thousands of local understandings that varied from district to district.
The response was immediate and rational. Caste associations formed. They petitioned census officials for higher classification. And to support their claims they submitted evidence of their respectability, which meant evidence about their customs — that they were vegetarian, that they did not drink, and that their widows did not remarry.
Some associations passed formal resolutions prohibiting practices their members had followed for generations, specifically in order to make the claim. A woman’s remarriage stopped being a family matter and became a datum in a national status competition.
The evidence here is administrative paperwork, which is the most reliable kind because nobody generating it was trying to make a point that would matter a century later.
What it is: census reports and correspondence, and the petitions caste associations submitted to census officials, along with the resolutions those associations passed and printed. These documents state, in the associations’ own words, which customs they were adopting and why.
What it shows: that the tightening was deliberate, argued for in writing, and explicitly connected to a claim for higher classification. This is not an inference about motives from a pattern of behaviour. The motive is written down.
What it cannot show: how far the resolutions were obeyed. An association printing a rule is not a village following one, and the gap between the two is unmeasured. It also tells us nothing about what the women in those communities said, because they were not the ones petitioning.
This is worth holding beside Chapter Five, where a great deal of Indian sexual conservatism turns out to have colonial fingerprints on it. Here the mechanism is different and more interesting. The British did not prohibit widow remarriage; they abolished sati and legalised widow remarriage. What they did was build a ranking system, and Indian communities then tightened their own rules to score better on it.
Both camps assume that the people constrained by these rules are reluctant parties to them.
The reformer assumes women submit under pressure and would choose otherwise. The traditionalist assumes women accept the rules as right and proper. Both are pictures of a subject being acted upon by a rule.
The evidence in this chapter shows something neither expects. Whole communities adopted restrictions on their own women that they had not previously had, voluntarily, at real cost, because doing so raised the standing of everybody in the group — including the women. A woman in a community that has just stopped permitting widow remarriage has gained something real: her family is now treated as higher, her children marry better, her sons’ prospects improve.
She has paid for it with the freedom of a future version of herself.
The general form: assuming that the person who bears a cost did not participate in choosing it. This is the hardest thing in the part to sit with, and it is why arguments about false consciousness are so common — they exist to dissolve exactly this observation. But dissolving it does not answer it. It says a woman who supports these rules is mistaken about her own interests, which is the identical move the other side makes about women who reject them.
3.4 — The modern version
The process did not end with the census, and this is where it stops being history.
In most of the world, as a country gets richer, more women work for pay. India is a striking exception at the household level: as an Indian family’s income rises, the women in it become less likely to be in paid work outside the home.
That is a strange pattern until you read it as sanskritisation. Withdrawing a woman from work is a purchase. It buys respectability. When a family cannot afford it, she works; when the family can, she stops, and the family’s status rises.
Part Twelve counts what this costs in national terms and the number is large. What matters here is the mechanism, because it explains something that otherwise looks like a paradox: the restriction is not a hangover from poverty. It is something families buy with their first surplus.
Which means economic growth, by itself, does not dissolve it. Growth funds it.
Restriction was concentrated at the top, not the bottom. Widow remarriage was permitted in many lower-status groups and forbidden in higher ones. Women in poorer communities worked, moved and were not secluded, because their labour was needed.
That makes restriction a luxury good. Keeping a woman out of the fields costs her earnings, so only a household with a surplus can buy the status it brings.
Sanskritisation is the process by which a group climbs by imitating the customs of a higher one. What gets adopted is a short list: no meat, no alcohol, ritual markers — and changing what the group’s women may do. Most items are cheap. Giving up a woman’s wages is expensive, which is exactly why it works as a signal.
Colonial censuses from the 1870s ranked castes officially, and communities petitioned for higher standing by submitting evidence of their customs. Some passed formal resolutions banning widow remarriage in order to score better. The British did not impose that. They built a scoreboard.
And it is still running. In India, as a family’s income rises, its women are less likely to work. Restriction is not a hangover from poverty — it is what families buy with their first surplus, which means growth does not dissolve it. Growth funds it.
4What India Was Actually Like
A country whose temples carry explicit sculpture, whose classical literature includes a manual of sexual technique, and whose Tamil poets wrote about lovers meeting in secret, is now widely regarded — by its own people — as among the most sexually conservative on earth. Both facts are real. Neither means what it is usually made to mean.
4.1 — The evidence people reach for
Three things get cited whenever this argument happens, and all three are genuine.
The temples. At Khajuraho in central India, a group of temples built by the Chandela dynasty around a thousand years ago carries sculpture that includes explicit sexual imagery, on the outside, in public view, on religious buildings. Konark in Odisha, a couple of centuries later, does the same. These are not hidden. They are architecture.
The manual. The Kamasutra, composed in Sanskrit somewhere in the early centuries of the common era, is a systematic treatise on desire, seduction, sexual technique and the conduct of relationships, including material on women’s pleasure and a substantial section on courtesans.
The poems. Tamil Sangam literature, from roughly the same era and earlier, includes an entire recognised genre of love poetry with an established set of conventions — among them clandestine love before marriage and elopement, treated not as scandal but as one of the standard situations a poem may describe.
So the claim that Indian civilisation was always sexually restrictive is false, and can be shown false with objects you can go and photograph.
Here is the thing that makes this confusing, put on a single line.
The Manusmriti — the text most often cited for restrictions on women — was compiled roughly two thousand years ago. The Khajuraho temples were built roughly one thousand years ago.
Set that gap against something you can feel. A thousand years is the distance between the Norman conquest of England and today. It is longer than the gap between Khajuraho and the British arrival.
So this is not a civilisation that was one thing and then became another. It is a civilisation that contained a severe legal text and, a millennium later, publicly carved erotic temples, and then a millennium after that arrived at the present. Anybody telling you what “Indian tradition” says about women is choosing which thousand years to stand in.
4.2 — And here is what it does not prove
Now the correction, because the reformist use of this evidence overreaches badly and the overreach is worth naming precisely.
Temple sculpture is not a photograph of anybody’s marriage. The erotic panels at Khajuraho are a minority of the total, they sit within a religious and symbolic programme whose meaning is genuinely disputed, and they were commissioned by a royal dynasty. Elite art tells you what an elite paid for. It does not tell you what happened to a widow in a village fifty kilometres away.
The Kamasutra is an elite urban text. It is addressed to a man of leisure in a city, with money, servants and time. It describes a world of a very small number of people. And it was composed in the same civilisation, in roughly the same era, that produced legal texts of considerable severity towards women. Both existed at once, which is the actual finding.
And “sexually permissive” is not the same as “good for women.” This is the error that matters most. A society can be relaxed about sexual imagery, comfortable with courtesans, unbothered by pre-marital love in poetry, and still marry its daughters at ten, forbid its widows to remarry, and prohibit women from owning property. Those are separate variables and they move independently.
This argument is had constantly in India and almost always badly.
Sculpture, literature and manuals from across a millennium show a civilisation at ease with sexuality, discussing it openly in its highest art and its most learned texts. That ease is gone. Something removed it, and Chapters Five to Eight identify what. The present prudery is therefore not authentically Indian and cannot claim the authority of antiquity.
Every item is elite, urban and produced by men. The same period produced restrictive legal codes. Erotic art in a temple may be ritual, symbolic or apotropaic, and even on the most literal reading it tells you what a king commissioned, not what a farmer’s daughter was permitted. Reading a nation’s sexual culture off its monuments is like reading modern India’s off its film industry.
What the evidence robustly establishes is variation — across region, century, caste and class, within one civilisation, at a scale that no single account of “Indian tradition” can absorb. That is not a weak conclusion. It is fatal to the claim that these rules are timeless, and it is exactly what Part Two predicts if the drivers are material rather than doctrinal.
What would settle it: evidence about ordinary women rather than elites — property records, court cases, inscriptions naming women as donors. This work is being done and it consistently shows more regional variety than the standard picture allows.
Why people care so much: because both sides want antiquity on their team. That is the genetic fallacy from Part One, and Chapter Nine turns it on this entire part, including on me.
4.3 — The devadasi question
One case deserves separate treatment, because it is the hardest and because both standard versions of it are wrong.
Devadasi: literally a servant of the deity. A woman or girl formally dedicated to a temple, which placed her outside the ordinary system of marriage, widowhood and inheritance described in Chapter Two.
Why it matters here: being outside that system meant different things at different times. In some periods it meant property, training and a public role that married women did not have. In others it meant no protection at all. The word covers both, which is why arguments using it go in circles.
Devadasis were women and girls dedicated to temples. Over a long history the institution meant different things in different periods and regions. In some, dedicated women held property in their own names, were literate, trained in music and dance, participated in public ritual life, and were not subject to the marriage-and-widowhood system described in Chapter Two. For women of that era those were significant advantages, and they are real.
By the nineteenth century, in most places where it survived, the institution had become something else: a route by which girls from poor and low-caste families were dedicated and then sexually available to men of higher standing, with no exit and no protection.
Both descriptions are accurate for different periods and places, and choosing one is choosing an argument rather than reporting a history.
The campaign to end dedication ran from the 1890s and produced legislation in the 1930s and 1940s. Missionary opinion was involved, which is the fact usually cited to dismiss the campaign as colonial interference. The fact usually left out is who led it: among the most prominent figures was an Indian woman doctor and legislator who came from a devadasi lineage herself, and who pressed the case in an Indian legislative body against Indian opposition.
So this is not a story about foreigners ending an Indian practice. It is a story about a genuinely mixed institution, degraded over time, ended by a campaign in which Indian women from the affected communities were central — and one where the language used against it, drawn from Victorian ideas about respectability, also damaged an entire tradition of women’s performing arts that was subsequently reassembled by women of higher caste under a more respectable name.
All of that is true at once. Anybody presenting it as simpler is selling something.
4.4 — The actual finding
Strip away the argument and what this chapter establishes is narrower than either camp wants and stronger than either expects.
Indian practice regarding women has varied enormously — by region, by century, by caste, by class. There were periods and places of considerable openness. There were periods and places of extreme severity. Frequently they coexisted, in the same century, a few hundred kilometres apart, among people who considered themselves to share a religion.
That pattern is not consistent with rules handed down by doctrine, because doctrine did not vary that much. It is entirely consistent with Part Two’s account, in which the drivers are property, farming, marriage markets and boundary maintenance — all of which vary exactly this way.
And it means one specific sentence is false, wherever you hear it: this is how it has always been. It has not. That is all this chapter needs to establish, and it establishes it comfortably.
Khajuraho carries public erotic sculpture on temples a thousand years old. The Kamasutra is a systematic treatise on desire including women’s pleasure. Tamil Sangam poetry treats secret pre-marital love as a standard subject. All three are real and photographable.
And none of them proves what it is used to prove. Temple art tells you what a king commissioned. The Kamasutra addresses a wealthy urban man. The same civilisation produced severe legal texts at the same time.
Above all: sexually permissive is not the same as good for women. A society can be relaxed about erotic art and still marry its daughters at ten and forbid its widows to remarry. Those variables move independently.
The devadasi case resists both standard stories. The institution gave some women property, literacy and a public role — and by the nineteenth century had become a route by which poor girls were dedicated and used. The campaign that ended it was led in part by an Indian woman from a devadasi lineage, against Indian opposition.
The finding is variation — across region, century, caste and class, at a scale no single account of “Indian tradition” can hold. That is fatal to the sentence “this is how it has always been,” and it does not require the past to have been better.
5The Victorians Arrive
A great deal of what is now defended in India as ancient modesty can be dated. Some of it was drafted in the 1830s by an Englishman who had been in the country for a few years, and some of it England itself abandoned decades ago.
5.1 — A criminal code written by a committee
India’s criminal law was not grown. It was written.
A law commission headed by Thomas Babington Macaulay drafted a penal code for India in the 1830s. It was enacted in 1860 and came into force at the start of 1862. It replaced a patchwork of existing legal systems with a single code applying across the territory — one of the most ambitious pieces of legal drafting ever attempted, and by most accounts an impressive one technically.
It was also, unavoidably, a Victorian document. Its assumptions about sex, decency and public morality were the assumptions of the men writing it, and those men were not Indian.
Two of its provisions matter here.
The provision on so-called unnatural offences criminalised certain sexual acts. Its lineage runs back through English law to a statute of the sixteenth century. It was not a codification of any Indian rule. It was an English prohibition, translated into a new code, and applied to a fifth of humanity. It remained in force, applying to consenting adults, until the Supreme Court of India read it down in 2018 — one hundred and fifty-six years.
The provisions on obscenity are the more revealing case, because of what happened to them afterwards.
5.2 — The test England abandoned and India kept
In 1868 an English court laid down a test for what counted as obscene. The formulation asked whether the material tended to deprave and corrupt those whose minds were open to such influence, and it permitted a work to be judged on isolated passages rather than as a whole. It was, by any modern standard, an extraordinarily broad test — under it, almost any serious treatment of sex could be prohibited.
England moved away from it by statute in 1959.
India adopted it in 1965, in a Supreme Court judgment upholding a conviction over the sale of an English novel, and it remained the operative Indian test for roughly half a century afterwards, before the courts began shifting towards a standard based on contemporary community norms.
This is not reconstruction. It is a paper trail of dated statutes and reported judgments, all published, all citable, all in English.
What it shows: a specific Victorian English legal standard, adopted into Indian law after the English had discarded it, and applied in Indian courts for decades. The judgment adopting it cites the English case by name. There is no ambiguity about the source.
What it cannot show: that Indian attitudes were caused by the law. Law and attitude move together and it is genuinely difficult to say which pulls. Indian judges adopting an English test in 1965 were Indians, choosing it, eighteen years after independence. That is a fact about what had already been absorbed, not simply about what was imposed.
There is a further case, and it is the sharpest one in this part, but it belongs in Chapter Six because it concerns a girl who died.
5.3 — Regulating women directly
The colonial state also legislated on women’s bodies in a more direct way.
From the 1860s, laws modelled on British legislation provided for the registration and compulsory medical examination of women identified as prostitutes in the vicinity of military stations. Women could be detained, inspected, and confined for treatment. The purpose was the health of soldiers.
Note the structure, because it recurs throughout this series. A problem arising from the behaviour of men was addressed by an apparatus applied to women. Nobody proposed registering and inspecting the soldiers.
Alongside the statutes ran a campaign of a different kind. From the 1890s an alliance of missionary organisations and Indian social reformers pressed against dance performances by professional women and against the temple dedication described in Chapter Four. The vocabulary of that campaign — respectability, decency, the good woman and the fallen one — is Victorian in origin and became fully naturalised in Indian public speech. It is still the vocabulary people use.
5.4 — And the other side of the ledger
Now the correction, because a chapter that only listed colonial impositions would be doing exactly what Chapter Eight criticises.
The same colonial legal apparatus abolished widow-burning in 1829. It legalised widow remarriage in 1856. It raised the age of consent, twice. It restricted child marriage. Every one of these ran against entrenched Indian custom and every one was resisted by substantial Indian opinion.
And Indians were central to all of them. The campaign against sati was driven by Indian reformers. The widow remarriage legislation followed years of argument by an Indian scholar who marshalled scriptural evidence for it. These were not measures done to India by foreigners; they were Indian arguments that used a foreign state as an instrument, which is a different thing and a more complicated one.
So the honest ledger has entries on both sides. The colonial state imported a body of sexual prudery in law and vocabulary. It also forced through liberalisations that Indian society had not managed on its own. Anybody using this history to argue that colonialism was simply the cause of Indian conservatism has to explain the widow remarriage act, and anybody arguing the reverse has to explain the obscenity test.
The entire argument in this chapter — and the argument it is answering — assumes that a tradition has one origin.
The nationalist says: this is ours, it is ancient, it comes from our civilisation. The critic says: no, this is Victorian, it came from England in 1860. Both are arguing about a single point of origin, as though the practice had a birth certificate with one country on it.
Nothing in this part has one origin. What an Indian family enforces about its daughters today is a composite: an endogamy rule that is thousands of years old, marriage-market arithmetic from the medieval period, a status-competition mechanism sharpened by a colonial census, a vocabulary of respectability imported from Victorian England, a legal architecture drafted in the 1830s, and a nationalist meaning attached in the 1920s that Chapter Seven describes.
Those layers came from different centuries and different continents, and they do not agree with each other. What holds them together is not a common source. It is that they all point at the same person.
The general form: demanding a single origin for a composite object. The demand is not innocent. Both camps want one, because one origin can be attacked or defended, and a composite cannot.
5.5 — How much of it is colonial?
This claim is made constantly in India, by people who disagree about everything else.
The penal code is dated and its drafters are named. The obscenity test has an English case citation on it. The vocabulary of respectability, the campaigns against performing women, the medical surveillance of prostitutes — all arrive with the colonial state and all can be dated. Meanwhile the erotic sculpture, the manuals and the love poetry described in Chapter Four are indigenous and pre-colonial. Something changed, and the timing is not subtle.
Every load-bearing element predates the British by a very long way. Caste endogamy closes two to four thousand years ago. Seclusion practices, child marriage, enforced widowhood and dowry are all attested long before any European arrives. The British contributed a legal vocabulary and a set of statutes on public obscenity — which is a real contribution and a shallow one. The machine in Chapters One to Three is entirely indigenous, and blaming colonialism for it is a way of not looking at it.
The most consequential colonial contribution may not be any statute. It is that colonial criticism of Indian treatment of women made defending that treatment into a patriotic act. That is not a law and cannot be repealed, and Chapters Seven and Eight are about it.
What would settle it: comparing Indian regions with different depths of colonial administration on measures of sexual conservatism today, holding other things constant. The princely states versus directly ruled provinces offer a natural experiment nobody has fully exploited.
Why people care so much: because the same finding is used by opposing camps for opposite purposes. Liberals use it to argue that conservatives are not authentically Indian. Nationalists use it to argue that pre-colonial India was superior and that the present should be restored to it. The finding supports neither conclusion — which does not stop either from being drawn.
One thing is worth holding from that verdict before we go on. A finding that both camps can use is usually a finding neither camp has understood, and Chapter Nine returns to this one with a much sharper instrument.
India’s criminal law was written, not grown — drafted in the 1830s by a commission headed by an Englishman, enacted 1860. Its assumptions about sex and decency were Victorian assumptions.
The obscenity standard is the clearest case. An English court laid down a very broad test in 1868. England abandoned it by statute in 1959. India adopted it in 1965 and applied it for roughly half a century — and the judgment adopting it cites the English case by name.
Colonial law also registered and medically inspected women near military stations, for the health of soldiers. Nobody proposed inspecting the soldiers. That structure — a problem created by men addressed through an apparatus applied to women — recurs throughout this series.
The ledger has entries on both sides. The same apparatus abolished sati, legalised widow remarriage, and raised the age of consent, all against entrenched Indian opinion — and Indians drove every one of those campaigns.
Nothing here has one origin. What a family enforces today is a composite: a rule thousands of years old, medieval marriage arithmetic, a colonial census, a Victorian vocabulary, an 1830s legal code, and a nationalist meaning added in the 1920s. Those layers do not agree with each other. What holds them together is that they all point at the same person.
6Phulmonee
Calcutta, 1890. A girl of about ten died of injuries inflicted by her husband on their wedding night. He could not be convicted of rape. The law that followed produced the fiercest political mobilisation of its decade — and the fury was not about the girl.
6.1 — What happened
Phulmonee Dasi was married, as girls in her community were, before she reached puberty. Her husband was a grown man in his thirties.
On the night the marriage was consummated she suffered internal injuries and bled to death.
The case came to court in Calcutta. The medical evidence was clear about the cause of death. The facts were not in dispute. And the court could not convict him of rape.
The reason was a provision in the penal code described in Chapter Five. Sexual intercourse by a man with his own wife was not rape, provided she was above a stated age. That age was ten. Phulmonee was above it.
He was convicted of causing death by a rash or negligent act and sentenced to twelve months.
6.2 — Where that provision came from
The marital exemption: a rule in criminal law stating that intercourse by a man with his own wife does not count as rape, whatever she wanted.
The reasoning behind it is that marriage itself constitutes consent, given once and not withdrawable. Note what this implies: not that she agreed on a particular night, but that the question of her agreement stopped being askable on the day she married.
Why it matters here: it is the single clearest case in this part of a rule that is neither ancient nor Indian, is defended as both, and is still in force.
The exemption in question — that a wife cannot be raped by her husband — was not an Indian rule. It came from a doctrine formulated by an English judge in the seventeenth century, holding that a woman gives her consent irrevocably at the moment of marriage and cannot withdraw it.
That doctrine travelled into the Indian Penal Code in 1860 along with the rest of the English legal furniture.
England abolished it in 1991.
India has not. The exemption survived the replacement of the colonial penal code in 2023 — the new criminal law reproduces it, raising the age below which it does not apply to eighteen but otherwise leaving it intact. Petitions challenging it have been before the Supreme Court for years and the matter remains unresolved as this is written. The government’s stated position has been that criminalising sex within marriage is unsuited to Indian conditions.
Lay the dates out in a line, because the shape of it is the point.
A judge in England writes the doctrine in the seventeenth century. It reaches India in 1860. England throws it out in 1991. India still has it now.
Put a lifetime against that gap. A woman married in England the year the doctrine fell has an adult daughter today. In India, over the same span, three generations of married women have lived and are living under a rule that the country which wrote it has not applied for thirty-five years.
And when India replaced the colonial code in 2023 — explicitly, in the language used to introduce it, to remove a colonial legacy — this particular colonial legacy was copied across.
I want to be careful about what that fact does and does not show. It does not show that the exemption is unpopular in India and imposed from outside; it plainly has substantial domestic support, and that support is the reason it survives. What it shows is something narrower and stranger: a rule that arrived from England, was discarded by England, and is now defended in India as suited to Indian conditions.
6.3 — The Act, and the storm
Phulmonee’s death drove a change in the law. In 1891 the age below which the exemption did not apply was raised from ten to twelve.
Twelve. That was the reform.
It produced the largest political mobilisation India had seen on any social question. Public meetings across Bengal and western India. Newspapers on both sides. The most prominent nationalist voice of the era campaigned hard against the measure.
And here is the thing that has to be understood correctly, because getting it wrong makes the whole episode unintelligible. The opposition was not, in the main, a defence of intercourse with ten-year-olds. Many opponents said plainly that the practice was indefensible.
Their argument was about sovereignty. A foreign government, of another religion, answerable to nobody in India, was legislating on the domestic and religious practice of a subject population. Accept that, they argued, and you have conceded that Indians cannot govern their own households — which is the whole colonial claim in miniature. Reform was necessary and must come from Indians.
That is a serious argument. It is not a stupid one and it was not made only by cynics. It also had a consequence its makers did not intend.
6.4 — What changed permanently
Before 1891 it was possible in India to be a reformer on the treatment of women and a patriot at the same time, without contradiction. Several prominent figures were both.
After 1891 that combination became difficult, and the difficulty has never gone away.
The reason is structural. Once opposing a reform becomes an act of national self-assertion, supporting the reform acquires the opposite meaning — whatever the supporter intends. An Indian arguing for a higher age of consent was now, whether he liked it or not, arguing on the same side as the colonial government. His opponents did not need to answer his argument. They only needed to point at who else was making it.
That move is the ancestor of the word this series began with. When somebody today says a woman has become westernised, they are performing the manoeuvre first assembled at scale in 1891: not answering the claim, but locating the claimant on the foreign side of a national boundary.
Notice also what this cost the reformers. The 1891 Act was a real improvement on a horrifying baseline. It was also, in the political ledger, a defeat — it confirmed that changes in the treatment of Indian women arrived through foreign power, which is precisely what the opposition said and precisely what made the next reform harder.
The next substantial change came in 1929, when the minimum age for marriage was raised. By then the legislature had a large Indian membership, and the measure was carried by Indians. That is the sequence, and it took thirty-eight years.
Calcutta, 1890. Phulmonee Dasi, about ten, died of internal injuries on the night her marriage was consummated. Her husband could not be convicted of rape, because the code exempted a husband where the wife was above ten. He got twelve months for a rash act.
The exemption was not Indian. It came from a seventeenth-century English doctrine holding that a wife consents irrevocably at marriage. It entered Indian law in 1860. England abolished it in 1991. India has not — and when the colonial code was replaced in 2023 expressly to shed colonial legacies, this one was copied across.
The 1891 Act raised the exemption age from ten to twelve, and produced the largest political mobilisation India had yet seen on a social question.
The opposition was mostly not a defence of the practice. It was an argument about sovereignty: a foreign government legislating on Indian households concedes that Indians cannot govern their own. That is a serious argument.
Its unintended consequence was permanent. After 1891, arguing for reform in the treatment of Indian women meant standing on the same side as the colonial state — so opponents no longer had to answer the argument, only point at who else was making it. That manoeuvre is the direct ancestor of the word “westernised.”
7The Inner Room
Facing a coloniser who was materially superior, Indian nationalism struck a bargain: concede the outside world, and declare the home the one place where the nation remained undefeated. That bargain is the single most important thing ever done to Indian women, and almost nobody who lives under it knows it was made.
7.1 — The problem
Consider the position of an educated Indian in the second half of the nineteenth century.
The evidence in front of him was humiliating and hard to argue with. The British had better guns, better ships, better medicine, a more effective bureaucracy, railways, telegraphs and a functioning legal system. Indians were being taught in English about the superiority of the civilisation that had conquered them, and much of the specific evidence for that superiority was visible from the window.
To be a nationalist he needed a claim. Not just that foreign rule was unjust — that is easy — but that there was something worth being sovereign for. Some domain in which Indian civilisation was not merely equal but better.
Where do you locate it, when the material evidence is all going the other way?
7.2 — The bargain
The answer, worked out across the nineteenth century and articulated most clearly by later historians looking back at it, was to divide the world in two.
The outer and the inner: in Bengali and other Indian languages the distinction is often put as bahir and ghar — the world outside and the home.
The nationalist settlement assigned different rules to each. In the outer domain — economy, technology, statecraft, science — the West was superior and Indians should learn from it without shame. In the inner domain — home, family, religion, and above all women — Indian civilisation was declared spiritually superior and must be protected from contamination.
Why it matters here: this is the moment a woman’s conduct stopped being a family matter and became the frontier of a civilisational contest. Everything modern India argues about regarding women runs through this settlement.
Look at how efficient it is as a political move. It concedes exactly the ground that cannot be held, keeps exactly the ground that cannot be inspected, and converts a defeat into a division of labour. The coloniser has the guns; we have the soul. The soul lives at home. Women are at home.
It also produced a specific new figure: the modern Indian woman as nationalism required her. Educated — because an ignorant wife was an embarrassment before the coloniser’s gaze — but modest. Capable in the home, active in public life to a controlled degree, and defined by a double contrast. She was not the Western woman, who was portrayed as immodest and selfish. And she was not the coarse woman of the lower classes, who was portrayed as vulgar.
She was, in other words, the nation itself in a sari. Which sounds like a compliment and is a job description.
7.3 — What it cost
The immediate consequence is visible in the historical record and it is startling once you look for it.
Count the decades of Indian public argument about women.
From the 1810s to the 1890s, the treatment of women is one of the central public controversies in India. Widow-burning. Widow remarriage. Female education. Age of consent. Newspapers, pamphlets, mass meetings, legislation, furious disagreement. Roughly eighty years of it.
Then, through the main decades of the mass nationalist movement, it goes quiet as a subject of first-order political dispute. Women enter the movement in very large numbers — that is not in question, and it mattered. But the question of what is owed to women inside the household largely leaves the agenda.
Something that occupied eighty years of argument stops being argued about, during the period of the most intense political mobilisation in Indian history. That is not a coincidence and it is not neglect. It was a settlement. The question had been answered in advance: the home was already the one part of India that was not defeated, so there was nothing there to reform.
That is a strong claim about an absence, and absences are the easiest thing in history to imagine. So it is worth being explicit about what is actually being counted.
A claim that a subject stopped being discussed sounds unfalsifiable. It is not, because the discussion left records.
What the evidence is: newspapers and periodicals, pamphlet literature, the proceedings of political organisations, and legislative records. What was printed, argued and voted on is countable, decade by decade.
What it shows: a subject that generated sustained controversy through the nineteenth century, and then far less as a first-order political dispute during the mass movement, even as women’s participation in that movement grew enormously.
What it cannot show: that nobody was discussing it. Household conversation leaves no archive, and women’s organisations continued working throughout. The claim is about what the national political argument was about, which is narrower than what people cared about — and this is interpretation by historians reading a record, not a measurement.
And the longer consequence is the one that concerns this series.
If the home is the last sovereign territory of a colonised nation, then a woman changing her behaviour inside it is not making a personal choice. She is opening the final gate. Her clothing, her marriage, her movement, her work — every one becomes a national security matter, in the most literal sense the phrase allows.
This is why arguments about Indian women’s freedom are so wildly out of proportion to their apparent subject. A dispute about a garment is not really about a garment. Both parties know, without either saying it, that the garment is standing in for something enormous.
Everyone in the modern argument assumes the home is a private space.
The liberal says: what happens in a family is a private matter, so a woman’s choices are hers and the community should stay out. The conservative says: what happens in a family is a private matter, so the state and the courts should stay out.
Both are invoking privacy, and both are working inside a settlement that made the Indian home the most public space in the country — the designated site of a contest over civilisation, watched by everybody, standing in for the nation.
Which means a woman appealing to her privacy is appealing to a status that was deliberately removed a century before she was born. It is not that her family fails to respect her private life. It is that in the settlement they inherited, her private life is the national frontier, and defending a frontier is what families understand themselves to be doing.
The general form: claiming a right that the arrangement in force has already spent. The argument cannot be won by appealing to privacy, because the privacy was traded away for something else — and the thing it was traded for was the self-respect of a colonised people, which is not a small thing and is why the trade has never been reversed.
7.4 — Why it is still running
Colonial rule ended in 1947. The settlement did not, and there are reasons.
Settlements outlive their occasions. Part Two, Chapter Nine established this in general terms: a rule’s original purpose and its current supports are different sets of hands. The inner-outer division was built for a specific humiliation, and it has since become simply how a great many Indians understand the difference between what is theirs and what is borrowed.
It is also continuously renewed, because the material asymmetry that produced it has not gone away. The economically dominant civilisation is still elsewhere. Its films, its platforms, its consumer goods and its opinions arrive daily and unbidden. Every one of those arrivals reactivates the original logic: we take their technology and we keep our home.
And the settlement has a defender in every household, because it hands ordinary people something valuable. A man with little power in the outer world — no capital, no influence, a job that does not respect him — retains one sovereign domain in which he is the custodian of a civilisation. That is not nothing. Taking it away is not a small ask, and pretending otherwise is why so many arguments about this go nowhere.
Nineteenth-century Indian nationalism faced a problem: the coloniser was materially superior, and a movement needs a claim to something worth being sovereign for.
The answer was to divide the world in two. The outer domain — technology, economy, statecraft — conceded to the West and learned from without shame. The inner domain — home, family, religion, women — declared spiritually superior and protected from contamination.
It is an efficient move: it concedes the ground that cannot be held and keeps the ground that cannot be inspected. And it produced a new figure — educated but modest, defined against both the Western woman and the coarse one. The nation itself in a sari, which sounds like a compliment and is a job description.
The immediate cost is visible. Roughly eighty years of fierce public argument about women’s treatment goes quiet during the most intense political mobilisation in Indian history. Not neglect — a settlement. The home was the one undefeated part of India, so there was nothing there to reform.
Which is why a dispute about a garment is never about a garment. If the home is a nation’s last sovereign territory, a woman changing her behaviour inside it is opening the final gate — and both sides know it without either having to say so.
8Mother India
In 1927 an American journalist published a book attacking Indian society through its treatment of women. It was propaganda with a political purpose, and most of its specific facts were true. What it did to Indian public life is still doing it.
8.1 — The book
Katherine Mayo was an American writer. In 1927 she published a book about India that became a bestseller in the United States and Britain and one of the most notorious publications in modern Indian history.
Its method was to describe, in detail, the physical condition of Indian women and children. Child marriage and early consummation. Death and injury in childbirth among girls whose bodies were not ready. Maternal and infant mortality. The condition of widows. The treatment of those outside caste society.
Its conclusion was that these conditions arose from Hindu religion and social practice, that they were characteristic of Indian civilisation rather than incidental to it, and that a people organised this way was not fit to govern itself.
That last step is what the book was for.
8.2 — Who wrote it, and why
Mayo had form. She had earlier published a book arguing that Filipinos were unfit for self-government. Her Indian research trip was facilitated by officials of the colonial administration, and her book appeared at a moment when Indian claims to self-rule were under active political consideration in Britain and the United States.
This is not a hidden history requiring detective work. Historians who have gone through the correspondence and the publication record have documented the political context in detail, and the pattern is clear: the book was an intervention in a live argument about whether Indians should govern India, and the condition of Indian women was the instrument.
Gandhi reviewed it and described it, memorably, as the report of a drain inspector — a person who goes through a country looking only at its sewers and then publishes an account of the country. The phrase stuck because it was accurate about the method.
8.3 — And the facts were largely true
Here is where the honest account gets difficult, and where most Indian discussion of this episode stops.
Child marriage was widespread. Girls were being married and their marriages consummated at ages that caused serious injury and death. Maternal mortality was appalling. Widows in many communities lived under the regime described in Chapter Two. Phulmonee had died thirty-seven years earlier and the law had responded by raising the age to twelve.
Indian reformers had been saying all of this, in public, in Indian languages and in English, for a century. The specific horrors Mayo catalogued were not her discovery and were not disputed by the people best informed about them.
So the book was propaganda, written for a political purpose, by an author with a record of writing exactly this kind of thing — and the facts in it were substantially accurate.
Both. At once. That combination is what made it so effective and so damaging.
Ninety years later this is still fought over in India, and the two standard answers are both incomplete.
The book’s purpose was to justify continued colonial rule, its author had a documented record of writing against self-government for colonised peoples, and her access was arranged by the administration. It presented the worst of Indian practice as the whole of it, attributed to religion what was produced by poverty and by law, and omitted everything that did not serve the argument. A work built to mislead is a lie regardless of whether its individual sentences check out.
Every practice she described was real, documented, and had been denounced by Indian reformers for decades. Girls were dying. The response of Indian public opinion was not to dispute the facts but to attack the author’s motives — which is the manoeuvre this entire series began with, applied on a national scale. If the facts were true, the author’s purpose is irrelevant to their truth.
Accurate facts assembled to support a false conclusion. The facts were real; the claim that they arose from Hinduism as such, characterised the whole society, and proved Indians unfit to govern, was not. And because the facts were real, they could not be refuted — so the defence had to be conducted against the author instead, which is exactly what happened and exactly what a work like this is designed to produce.
What would settle it: nothing further. The facts are documented and the political purpose is documented. What remains is a disagreement about what to call a work that is accurate in its parts and false in its whole.
Why people care so much: because the episode is the founding case of a pattern that has run in India ever since — external criticism of the treatment of Indian women being met with an examination of who is criticising.
8.4 — What it produced
Two effects, running in opposite directions, and both are real.
The first was reform. The controversy mobilised Indian women’s organisations, which used it — and used the international embarrassment — to press for change. Two years after the book appeared, Indian legislators raised the minimum age of marriage. Some of the impulse behind that was a determination to demonstrate that Indians could reform themselves without being told to.
So the book, written to prove India unfit to govern itself, contributed to India legislating a reform in order to disprove it. That is genuinely how it happened and it is worth holding on to, because it is the most hopeful thing in this part.
The second effect was a reflex, and it has outlasted the reform.
After 1927, criticism of Indian treatment of women carried a permanent second meaning. However the criticism was framed, and whoever made it, it now had a shape that everybody recognised — the shape of an outsider using Indian women to argue that Indians were backward.
That recognition is not paranoid. It happened, at scale, in a bestseller, for a documented political purpose. The reflex was earned.
But a reflex does not check credentials. Once it exists, it fires at any criticism with a similar shape, including criticism from Indians, including criticism that is entirely correct, including — and this is the part that matters most — criticism from Indian women about their own lives.
Combine that with Chapter Seven and the machine is complete. The home is the nation’s last sovereign territory. Criticism of what happens inside it has the shape of a colonial attack. Therefore a woman raising a complaint about her own household is, structurally, in the position of Katherine Mayo — and she will be answered the way Katherine Mayo was answered, by an examination of whose side she is really on.
In 1927 an American journalist published a book cataloguing the physical condition of Indian women and children — child marriage, injury and death in childbirth, the treatment of widows — and concluded that Indians were unfit to govern themselves.
It was propaganda with a documented political purpose. The author had already written against self-government for another colonised people, and her access was arranged by the administration. Gandhi called it a drain inspector’s report, which was accurate about the method.
And the facts were largely true. Indian reformers had been saying them for a century. Both things at once — accurate parts, false whole — is what made it so effective.
It produced reform. Indian women’s organisations used the embarrassment, and two years later Indian legislators raised the marriage age, partly to prove they did not need to be told.
And it installed a reflex that outlasted the reform: criticism of how India treats its women now carries the shape of a colonial attack. That reflex was earned. But a reflex does not check credentials — it fires at correct criticism, at Indian criticism, and at a woman complaining about her own household, who will be answered not on the facts but by an examination of whose side she is on.
9The Word “Westernisation”
Part One said this word bundles four claims and that the last one does the real work. We can now test it against eight chapters of evidence. The result is that both camps are holding foreign objects and calling them their own — and then the whole argument turns out to be invalid anyway, including mine.
9.1 — The audit
Put every practice discussed in this part into a table and ask a single question of each: where did it actually come from?
| Practice | Where it actually comes from |
|---|---|
| Caste endogamy | Indigenous. Boundaries close somewhere between two and four thousand years ago and hold since. |
| Enforced widowhood | Indigenous and ancient. A solution to the surplus-woman problem in a closed circle. |
| Child marriage | Indigenous and ancient. The supply side of the same arithmetic. |
| Dowry | Indigenous in origin — and its modern scale and spread are a twentieth-century development, largely after it was banned. |
| Seclusion norms | Indigenous, pre-colonial, and spread downwards through status imitation rather than being imposed from above. |
| A national ranked caste list | Colonial. A product of the census from the 1870s, which turned local understandings into a national scoreboard groups could compete on. |
| “Obscenity” as a legal category | English. Statutory architecture from 1860, and a judicial test from 1868 that England dropped in 1959 and India adopted in 1965. |
| Criminalising certain sexual acts | English, with a lineage back to a sixteenth-century statute. In force in India for 156 years. |
| A husband cannot rape his wife | English, seventeenth-century doctrine, imported 1860. Abolished in England in 1991. Still Indian law. |
| Medical surveillance of “prostitutes” | English, imported in the 1860s for the health of soldiers. |
| The vocabulary of respectability | Victorian. The good woman and the fallen one, decency, vulgarity — naturalised so completely that it now reads as indigenous. |
| Woman as bearer of national culture | Nationalist, late nineteenth and early twentieth century. A response to colonial humiliation, not an ancient idea. |
| Treating criticism as foreign attack | Assembled 1891, hardened 1927. |
| Abolishing sati; legalising widow remarriage | Colonial statutes, driven by Indian campaigns. Foreign instrument, Indian argument. |
| A woman’s life is hers to direct | As a political claim in its modern form: recent, and substantially of European origin. Which is exactly as irrelevant as everything else in this column — see §9.3. |
9.2 — What the table shows
Read the column and the neat division collapses.
The conservative is holding foreign objects. The legal architecture within which he prosecutes vulgarity is English. The vocabulary of respectability he uses is Victorian. The scoreboard his ancestors competed on was built in a colonial census office. The doctrine that a husband cannot rape his wife was written by an English judge and thrown out by England thirty-five years ago. A great deal of what he is defending as Indian tradition arrived by ship.
And the liberal is holding foreign objects too. The claim that a person’s life belongs to them, that consent is the relevant test, that a family has no standing to direct an adult’s choices — as a political demand, in the form now made, this has a traceable European intellectual history. Pretending it is simply what all Indians always wanted is not honest, and Indians hearing it know it is not.
So the argument as normally conducted — this is ours, that is theirs — cannot be settled, because both sides are composites and neither can produce a clean pedigree.
And there is a further complication that the table makes visible. The colonial contribution ran in both directions at once. The same apparatus that imported obscenity law abolished widow-burning. The same officials whose census tightened caste also raised the age of consent. Anybody who wants a single verdict on what colonialism did to Indian women has to explain both columns, and nobody who offers a single verdict ever does.
9.3 — And now the part where I have to indict myself
Everything above is accurate and most of it is invalid as an argument, and I have been using it anyway.
This entire chapter — and Chapter Five, and the argument it is answering — assumes that where a practice came from determines whether it is legitimate.
That is the genetic fallacy, named in Part One, Chapter One: judging a claim by its parentage rather than by whether it is true. I named it in the first chapter of the series and then built four chapters on it.
Consider what Chapter Five actually establishes. It shows that a body of Indian sexual conservatism has an English origin. What follows about whether those rules are good? Nothing. A rule is not made bad by having been drafted in London, and it is not made good by having been written in Sanskrit. If restricting a woman’s clothing is wrong, it is wrong whether Manu wrote it or Macaulay did. If it is right, the same.
So why did I write Chapter Five? Because it is devastating — it takes the strongest rhetorical weapon out of the conservative’s hand by showing that his tradition has a foreign birth certificate. It wins arguments. It does not establish anything.
And there is a reader-side version of this, which is the more useful half. If Chapter Five was the most satisfying thing you have read in this series, notice what the satisfaction was made of. It was not the pleasure of learning something about whether the rules are justified. It was the pleasure of watching an opponent lose a claim. Part One said to notice which findings you hope are true, because that marks where you are least reliable. This is one of those places, and I built it deliberately.
The general form: using an invalid argument because it is effective, in a document committed to valid ones. I am not going to remove the chapter — the history in it is true and worth knowing. I am going to say plainly that it does not do the work it appears to do.
9.4 — What the word actually does
So if origin does not settle legitimacy, what is the word for?
Part One identified four things bundled inside it: a description, a causal claim, a judgement of decline, and a claim about belonging. Eight chapters later we can be specific about each.
The description is often accurate. Some Indian women are behaving in ways more common elsewhere.
The causal claim — that they do so because of foreign influence — is usually wrong, and this part shows why. A woman taking paid work in a city is responding to urbanisation, education and her family’s need for income. Those are Indian conditions. Chapter Three showed that the reverse movement, women leaving work, is what a family buys when it can afford it. Neither is imitation of anybody.
The judgement of decline is a value, and Part One, Chapter Four says a value has to be argued rather than smuggled inside a noun.
The claim about belonging is the one doing the work, and now we know exactly where it was manufactured. It was assembled in 1891, when opposing a reform became an act of national self-assertion. It hardened in 1927, when accurate criticism arrived inside a hostile book. And it sits on the settlement in Chapter Seven, which made a woman’s conduct the frontier of a civilisation.
That is the whole apparatus. When the word is used, all four fire at once, and the fourth is the one that lands — because it does not argue with a woman, it relocates her. She is not wrong. She is foreign. And a country that spent two centuries being told it was inferior by foreigners has a well-developed set of responses to that.
Audit every practice in this part and the clean division collapses. The conservative is holding foreign objects — the obscenity law, the vocabulary of respectability, the colonial caste scoreboard, and a marital rape doctrine written by an English judge and abolished in England in 1991.
And the liberal is holding foreign objects too. The claim that a person’s life is theirs to direct, in the political form now made, has a traceable European history. Pretending otherwise is not honest and nobody is fooled.
The colonial contribution ran in both directions at once: the apparatus that imported obscenity law also abolished sati and legalised widow remarriage.
And then the whole argument is invalid. Where a rule came from settles nothing about whether it is good. If restricting a woman’s clothing is wrong, it is wrong whether Manu wrote it or Macaulay did. I wrote Chapter Five anyway, because it wins arguments — and if it was the most satisfying thing you have read here, that satisfaction was the pleasure of watching someone lose a claim, not of learning whether the rule is justified.
The word does not argue with a woman. It relocates her. She is not wrong, she is foreign — and a country that spent two centuries being called inferior by foreigners has a very well-developed set of responses to that.
10An Honest List Of What We Do Not Know
Two lists, no hedging in either. What is genuinely unknown about the Indian machine, with the reason it is unknown. And what is solid enough to carry into the rest of the series.
10.1 — Genuinely unknown
This part rests on a mix of genetics, survey data, legal records and historical argument. Each has different failure modes and they do not cancel out.
Whether endogamy is the whole mechanism or one of several. Ambedkar’s account predicts three things correctly, which is real support. It does not follow that nothing else contributes. Ritual purity, property law, regional politics and simple inherited habit all plausibly carry some of the load, and nobody has measured their relative weights. A functional explanation that fits is not the same as a functional explanation that is complete.
When exactly the boundaries closed, and why then. The genetic range of two to four thousand years is wide, different groups closed at different times, and the estimates depend on modelling assumptions. What was happening in the subcontinent at the moment of closure — and whether it was one process or many — is not settled.
What India would look like without colonial rule. This is unanswerable and it is worth saying so, because both camps answer it constantly. Whether reform would have come faster from within, or slower, or not at all, cannot be known. The comparison that would help most — princely states, which the British ruled indirectly, against directly administered provinces — has not been systematically exploited on these measures.
Whether status imitation is still the main driver of women leaving work. Chapter Three’s account fits the pattern well. So do other explanations: the kinds of jobs available, safety in transport and workplaces, childcare, and the way official statistics count unpaid family labour. Which of these dominates is a live argument among economists and Part Twelve will not resolve it either.
What women thought, at every stage. The same hole as in Part Two, in a sharper form here. The reform debates of the nineteenth century were conducted overwhelmingly by men, on both sides, about women. Where women’s own accounts survive they are late, often elite, and shaped by whichever camp was publishing them. Chapter Three’s finding — that communities adopted restrictions voluntarily — is drawn from records of decisions in which women’s participation is largely invisible. That is a serious limit on how far that finding can be pushed.
Whether the reflex from 1927 can be disarmed. It is real, it was earned, and it now fires at correct criticism from Indians. Nobody knows what dissolves it, and this part offers no method.
10.2 — Solid
That around one Indian marriage in twenty crosses a caste line, and that this has barely moved. Multiple large surveys, decades apart, compatible answers.
That the boundaries are visible in DNA, closed within a datable window, and carry a medical cost. Independent teams, different samples, compatible findings. The founder effects in many groups exceed those in populations studied specifically for their inherited disease burden.
That Ambedkar’s mechanism claim makes three correct predictions. The customs were strictest where endogamy was strictest, the marriage rule outlived the occupational system, and the genetic record shows boundaries closing and holding. His account of caste’s origins is superseded and should be treated separately.
That restriction spread downwards through status imitation, and that colonial censuses accelerated it. Caste associations petitioned for higher ranking and submitted their customs as evidence. Some formally prohibited practices they had permitted. This is documented in the administrative record.
That Indian practice regarding women varied enormously by region, century, caste and class. Khajuraho and the Manusmriti are separated by a thousand years and belong to the same civilisation. That variation is fatal to the sentence “this is how it has always been” and requires no claim that the past was better.
That a specific layer of Indian sexual conservatism has a dateable English origin. Statutes, judgments and case citations, all published. The obscenity test England dropped in 1959 and India adopted in 1965 is the clearest instance.
That the marital rape exemption is English in origin, was abolished in England in 1991, survived India’s 2023 replacement of the colonial penal code, and remains in force. Every element is a matter of public record.
That the women’s question largely left Indian nationalist politics after roughly eighty years of intense public argument. This is observable in the record of what was debated, and Chapter Seven’s explanation of why is the best available.
That Mother India was propaganda and that its facts were largely accurate. Both are documented, and the combination is the finding.
10.3 — Who wrote the record
One closing observation, following the pattern of the previous parts.
Notice what this part could be built from. Genetic samples. Census returns. Statutes and law reports. Caste association resolutions. Newspaper controversies. Legislative debates. Temple sculpture commissioned by kings.
Every single one of those was produced by somebody with power. States counted. Courts recorded. Associations petitioned. Kings commissioned. Newspapers were owned. The archive of the Indian machine is, almost in its entirety, the archive of the people operating it.
There is one exception in this part and it is worth naming. Ambedkar was writing from inside the category the machine was built to exclude. That is not a small thing. It may be the reason he saw the mechanism when others who had studied it far longer did not — and it is a reasonable working hypothesis that the rest of what is missing from this archive would have been visible to people in a similar position, had anybody thought to write it down.
Which is a way of saying: the thinnest part of the record is the testimony of the people the system was built to act upon, and this is the part of the story that would most change what we think we know.
Genuinely unknown: whether endogamy is the whole mechanism or one of several. Exactly when the boundaries closed and why then. What India would look like without colonial rule — unanswerable, and both camps answer it constantly. Whether status imitation still drives women out of work. What women themselves thought at any stage. Whether the 1927 reflex can be disarmed.
Solid: one marriage in twenty crosses a caste line and has for decades. The boundaries are readable in DNA, datable, and carry a medical cost to children. Ambedkar’s mechanism claim predicts three separate things correctly — while his origin claim is superseded. Restriction spread downwards through imitation, accelerated by a colonial scoreboard. Indian practice varied enormously across region and century. A dateable layer of Indian conservatism is English, including a marital rape doctrine England abolished in 1991 and India kept through a 2023 rewrite meant to shed colonial legacies. The women’s question left nationalist politics after eighty years of argument. Mother India was propaganda and its facts were largely true.
And notice who made this archive. Genetic samples, censuses, statutes, law reports, petitions, legislative debates, temples commissioned by kings — every source was produced by someone with power. It is almost entirely the archive of the people operating the machine. The one exception in this part is Ambedkar, who was writing from inside the category it was built to exclude, and who saw the mechanism that people who had studied it far longer did not.
Sources & further reading — Part 3
Timeline
From the closing of the marriage circles to a doctrine written by an English judge and still in force. Early dates are ranges rather than years, and the uncertainty is genuine.
| When | What happened |
|---|---|
| c. 2000–4000 BP | Widespread population mixing across the subcontinent ends. Groups begin marrying only inside themselves and continue doing so. The boundary is still readable in DNA today, along with the founder effects it produced. |
| c. 200 BCE–200 CE | Compilation of the Manusmriti, the text most often cited for restrictions on women. |
| c. 300 BCE–300 CE | Tamil Sangam poetry, including an established genre treating clandestine pre-marital love as a standard poetic situation. |
| c. 2nd–4th c. CE | The Kamasutra. A systematic treatise on desire, addressed to a wealthy urban man, in the same civilisation and roughly the same era as the severe legal texts. |
| c. 950–1050 CE | The Khajuraho temples are built, carrying explicit sculpture in public view — a thousand years after the Manusmriti and a thousand before the present. |
| 1829 | Sati abolished, after campaigning by Indian reformers using a colonial state as the instrument. |
| 1856 | Widow remarriage legalised, following years of scriptural argument by an Indian scholar. |
| 1860 | The Indian Penal Code is enacted, drafted in the 1830s by a commission headed by Macaulay. It carries in an English prohibition on certain sexual acts and an English doctrine that a husband cannot rape his wife. |
| 1860s | Legislation modelled on British statutes provides for registration and compulsory medical examination of women near military stations. Nobody proposes examining the soldiers. |
| 1868 | An English court lays down a very broad obscenity test. |
| 1870s | Decennial censuses begin recording and ranking caste. Associations form and petition for higher standing, submitting their customs as evidence — some formally banning widow remarriage to improve their claim. |
| 1890 | Phulmonee Dasi, about ten, dies of internal injuries in Calcutta. Her husband cannot be convicted of rape because she was above the age of ten. He receives twelve months. |
| 1891 | The age below which the marital exemption does not apply is raised from ten to twelve — and produces the largest political mobilisation India has yet seen on a social question, conducted mainly as an argument about sovereignty. |
| 1890s | The campaign against temple dedication and professional dancing women begins, an alliance of missionary organisations and Indian reformers, using a Victorian vocabulary of respectability. |
| 1916 | Ambedkar, aged twenty-five, reads a paper in New York setting out why a closed marriage circle produces sati, enforced widowhood and child marriage as one solution to one arithmetic problem. |
| 1927 | Mother India is published: propaganda with a documented political purpose, and largely accurate in its facts. Dozens of Indian replies follow. |
| 1929 | Indian legislators raise the minimum age of marriage, partly to demonstrate that Indians could reform themselves without being told. |
| 1947 | Independence. Temple dedication is prohibited in Madras, in a campaign led in part by an Indian woman legislator from a devadasi lineage. |
| 1959 | England abandons the 1868 obscenity test by statute. |
| 1965 | India adopts it. It remains the operative Indian test for roughly half a century. |
| 1991 | England abolishes the doctrine that a wife cannot refuse her husband. |
| 2018 | The Supreme Court of India reads down the colonial provision on so-called unnatural offences, 156 years after it came into force. |
| 2023 | India replaces the colonial penal code, expressly to shed colonial legacies. The marital rape exemption is carried across. Petitions challenging it remain pending. |
| ↓ | Part Four starts here — and leaves history entirely. It asks how different men and women actually are, measures every difference, and reports the sizes. |
Glossary
Every hard word used in this part, in plain English. Each was explained where it first appeared; this page is a convenience.
| Term | Plain meaning |
|---|---|
| Bahir and ghar | The outer world and the home. The nationalist settlement gave each different rules: learn from the West outside, keep the inside uncontaminated. Women lived on the inside. |
| Devadasi | A woman or girl dedicated to a temple. In some periods the arrangement carried property, literacy and a public role; by the nineteenth century it had largely become a route by which poor girls were dedicated and used. |
| Endogamy | The rule that you must marry inside your own group. The subject of this entire part. |
| Enforced widowhood | The regime applied to widows in strict communities — shaved head, white cloth, no ornaments, restricted food, exclusion from celebrations. Every element removes a signal of sexual availability. It is a uniform, not a punishment. |
| Exogamy | Marrying outside a specified group. The opposite of endogamy. Indian kinship typically combines both: marry outside your immediate lineage, inside your caste. |
| Founder effect | The genetic signature of a population descended from a small number of ancestors. It concentrates rare inherited conditions, and the cost falls on children. |
| Genetic fallacy | Judging a claim by where it came from rather than whether it is true. Named in Part One and committed deliberately in Chapter Five of this part, as Chapter Nine admits. |
| Hicklin-type test | The broad Victorian English standard for obscenity, permitting a work to be judged by isolated passages. Abandoned in England in 1959, adopted in India in 1965. |
| Jati | The actual group people belong to and marry within. Thousands exist, usually regional. This is what Indians mean by caste in ordinary speech. |
| Marital rape exemption | The rule that intercourse by a husband with his wife is not rape. Written by a seventeenth-century English judge, imported into Indian law in 1860, abolished in England in 1991, retained in India’s 2023 criminal code. |
| Sanskritisation | A lower-status group raising its position by adopting the customs of a higher one — giving up meat and alcohol, and restricting its women. Named in 1952. |
| Sati | The burning of a widow on her husband’s pyre. Never universal, concentrated in particular regions and groups, abolished in 1829. In Ambedkar’s reading, one solution to the surplus-woman problem. |
| Surplus man / surplus woman | Ambedkar’s terms. A person left without a partner by a death, inside a closed marriage circle, who must be dealt with or the boundary leaks. |
| Varna | The fourfold classical scheme in ancient texts. Four categories for a country containing thousands of marrying groups — which is why arguments that slide between varna and jati go nowhere. |