How to read this part
This is the twelfth part of a fourteen-part history of Punjab. It runs from 1915 to the summer of 1947 — from a failed rebellion by men who had come home from America, to the day it was agreed that Punjab would be cut in two.
It is the shortest stretch of time any part of this series has covered. It is also the most crowded. In thirty-two years Punjab produced a rising, a massacre, a mass religious reform movement, the most famous young revolutionary in modern Indian history, a form of politics nobody else in India managed to build, and then the total collapse of that politics. Almost everything people argue about today, when they argue about Punjab, was set down in these years.
Three rules hold for this part, as they have for every part before it.
Plain language. Short sentences. Ordinary words. Every hard word explained the moment it appears, in a Word Box. Nothing is left out to make the writing simpler. The ideas stay whole; only the wording gets easier.
Evidence first. Where a claim comes from a document, the document is named. Where the evidence is thin, this is said out loud. This part is unusual in one way: much of what we know comes from the government's own paperwork. A government describing why it did something is a particular kind of source, and it is treated as one.
Competing views. Where historians disagree, both sides get their strongest case, in an Argument Box. Where the disagreement has not been settled, it is left unsettled. You are not being steered to a conclusion.
There is one more box you will meet often here, more often than in any earlier part. It is called The Hidden Assumption. It stops the story and asks what a claim is quietly taking for granted. This period generates more of those than any other, because so much of what was said at the time — by the British, by every Indian party, and by people writing since — rests on an assumption that was never stated and never tested.
A note on names. Almost every place in this part is now in Pakistan or in India, and some have changed names since. Names are given as they were used at the time, with the modern name in brackets where it helps.
A note on the ending. This part stops in the last week of June 1947, with the decision made and the border not yet drawn. What happened next — the line itself, and the killing and the walking that came with it — is Part 13. That is not squeamishness. It is that the decision and the disaster are two different subjects, and running them together makes it harder to see either one clearly.
Thirty-two years at a glance
Come back to this page whenever you lose your footing. Everything here is explained in full later.
1The Men Who Sailed Back
A rebellion planned in Oregon, printed in San Francisco, and broken in a Punjab village by one man with a notebook.
In the years just before the First World War, several thousand Punjabis were living on the west coast of North America. Most were Sikhs. Most were from the central districts — Jalandhar, Hoshiarpur, Amritsar, Ludhiana. Many had been soldiers. They worked in the lumber mills of British Columbia, on the railways, and in the fruit orchards of California.
They had gone for money, and they were earning it. A mill hand in Vancouver could make in a month what a farmhand in Doaba made in six. Land at home was being bought with wages sent from Oregon.
But they had also gone into something they had not expected. They were treated as a race problem. Canada and the United States both moved to shut them out. There were riots against them in Bellingham and in Vancouver. They were told, in law and in the street, that they were not wanted.
What made this unbearable was a comparison. These men were British subjects. Many had worn the King's uniform. They had been told all their lives that the Empire was a family of peoples under one Crown. Then they arrived in another part of that same Empire and were treated as an infestation. And when they complained, the British government did nothing, because doing something would have annoyed a white settler colony that mattered more.
That is the exact point at which a loyal subject becomes a rebel. Not when he is oppressed — he had been oppressed at home too. When he notices that the story he was told about the Empire is not true.
A newspaper called Mutiny
In 1913 a group of these men founded a party in Astoria, Oregon. Sohan Singh Bhakna, a former soldier from Amritsar district, was its first president. Har Dayal, a Delhi-born intellectual who had been at Oxford and had abandoned it, became its secretary and its main writer.
They called it the Ghadar Party, after its newspaper.
Ghadar (also written Ghadr) is Urdu for mutiny, or revolt. The word was not chosen by accident. The British called the great rising of 1857 "the Mutiny", and used the word as an insult. The party picked up the insult and wore it. Calling your newspaper Mutiny in 1913 was a way of saying: yes, that is exactly what we are.
The first issue was printed in San Francisco on 1 November 1913, from a house the members called the Yugantar Ashram. It came out in Urdu and Gurmukhi, and later in other languages. It was posted, carried, smuggled and read aloud to men who could not read, in Hong Kong, Shanghai, Manila, Singapore, Rangoon, Fiji, Panama, and across Punjab.
Its message was short enough to repeat from memory. India was poor because it was ruled. The British stayed because Indians served them — as soldiers, as police, as clerks. If those Indians stopped serving, the whole thing would fall over in a week. Religion was not the point. Reform was not the point. The point was to throw them out, by force, and quickly.
The paper carried the names of the men hanged in earlier risings. It printed the pay of a British officer next to the pay of an Indian sepoy. It ran a standing feature listing what the Empire had taken and what it had given back.
We have the newspaper. Complete runs of Ghadar survive in libraries in India and North America, and the British collected it obsessively themselves — intelligence officers translated it and pasted it into files that are now in the archives in London and Delhi.
This is an unusual luxury. For most rebel movements in history we only have what the state said about them. Here we have what the rebels said about themselves, in their own words, week by week, before they knew how it would end.
The ship
In the spring of 1914 a Singapore-based businessman named Gurdit Singh chartered a Japanese steamship, the Komagata Maru, and filled it with passengers bound for Canada. There were 376 of them. Most were Sikhs.
The plan was legal in spirit and pointed in intent. Canada had a rule that immigrants must arrive by continuous journey from their country of origin. There was no continuous shipping route from India to Canada, which was exactly why the rule existed. Gurdit Singh chartered a ship to make one.
The ship reached Vancouver harbour on 23 May 1914. It was not allowed to land. It sat in the harbour for two months while lawyers argued and food and water ran low. Twenty-four passengers were admitted. The rest were sent back.
The ship returned to India and anchored at Budge Budge, near Calcutta, on 29 September 1914. The government intended to put the passengers straight onto a train to Punjab, where they could be watched. The passengers refused. There was a confrontation, then firing. Around nineteen passengers were killed.
Two months in a harbour, in sight of a city, not allowed to step onto the dock. Long enough for a baby to be born on board. Long enough for the drinking water to become the main thing anyone talked about.
The distance from the ship to the shore was a few hundred metres — less than a five-minute walk, if walking had been allowed. That gap is the whole story of the voyage.
The Komagata Maru did not create Ghadar. It was already three years old in spirit and one year old in print. What the voyage did was turn an argument into an experience. Thousands of people in Punjab now knew somebody who had been on that ship, or had been at the dock, or had been shot at. The newspaper's claim — that the Empire's promise was a lie — had been tested in public and had failed in public.
Coming home to start a war
Then the war began, and Ghadar decided that its moment had arrived. The reasoning was simple. Britain was fighting in Europe. The best British troops had left India. The Indian Army was full of Punjabis. If the Punjabi regiments could be turned, India would be lost to Britain in a month.
So the party told its members to go home and do it. Several thousand did. They sailed from San Francisco, from Manila, from Hong Kong, through the second half of 1914 and into 1915. Many were arrested at the docks in Calcutta the moment they landed, because the government had read the newspaper too and had a list.
The rest reached Punjab and set about organising a rising. They contacted regiments at Ferozepur, Lahore, Rawalpindi and Meerut. They collected weapons. They robbed to fund themselves. They made bombs, badly. The date was fixed for 21 February 1915.
The plan failed completely. A police informer named Kirpal Singh had got inside the organisation. The leaders learned they were being watched and moved the rising forward to 19 February. It made no difference. The government knew about that too. Regiments were disarmed before the day. Arrests began immediately.
On the face of it the answer is obvious. It planned a rebellion; the rebellion did not happen; its members were hanged. But historians have disagreed sharply about what that means.
It achieved none of its aims. Not one regiment mutinied. Not one district rose. The organisation was penetrated by a single informer, which suggests it had no serious security. Its leaders misjudged Punjab completely: they assumed that villages full of serving soldiers' families were ready to revolt, when in fact those families were doing well out of the army. Recruitment in Punjab went up during the war, not down. Judged as a plan to remove the British in 1915, it was a fantasy.
Movements are not only judged by whether their plan works. Ghadar put a fully worked-out anti-colonial politics into circulation in Punjab, in Punjabi, for the first time — one that was secular, republican and unbothered by caste or creed. It supplied the language and the martyrs that later movements used. Bhagat Singh's circle read Ghadar literature and held Kartar Singh Sarabha's photograph. It also frightened the government into the repression that produced 1919, which produced everything after that.
Where this leaves us
Both are true and they are not really in conflict. Ghadar failed as an operation and succeeded as a transmission. The honest way to say it is that it lost its war and won its argument, but only after everyone who had made the argument was dead.
The trials
What followed was a set of mass trials at Lahore, held under emergency wartime law by a special tribunal of three judges. There was no jury. There was no appeal.
In the first Lahore Conspiracy Case, 291 men were tried. Forty-two were sentenced to death. More than a hundred were transported for life to the penal settlement in the Andaman Islands. Further cases followed through the war.
Transportation for life meant being shipped to a prison colony overseas and never coming back. For Indian political prisoners it usually meant the Cellular Jail in the Andaman Islands, in the Bay of Bengal — a purpose-built prison where men were kept in single cells and worked at oil mills. Prisoners called it Kala Pani, the black water.
The best remembered of the condemned was Kartar Singh Sarabha, from a village near Ludhiana. He had gone to California as a student. He came back to organise a rebellion. He was hanged at Lahore Central Jail on 16 November 1915. He was nineteen years old.
Ghadar's whole strategy rested on one unexamined idea: that Punjabi soldiers served the British only because they had not yet been told the truth. Tell them, and they would turn.
This assumed that the soldiers' loyalty was ignorance. It was not. For a peasant family in Ludhiana or Rawalpindi, army service meant cash, a pension, land grants in the new canal districts, and status at home. The soldiers had looked at the deal and taken it — not because they loved the Empire, but because it paid.
The same hidden assumption comes back again and again in this part, and it is worth naming now, because it will keep returning: the belief that people who do not join your movement have simply not understood it yet. Ghadar believed it in 1915. Congress believed it about Punjab's Muslim peasants in 1937. The Unionists believed it about their own voters in 1946. Every one of them was wrong in the same way.
Ghadar was made abroad, by migrants who discovered that being a British subject meant nothing once they left India. It brought back a politics that was secular, republican and violent, and it failed at once because it misread its own homeland.
It matters for what it handed on: a vocabulary, a set of martyrs, and the proof that Punjabis would organise against the Raj. The government drew a conclusion too — that Punjab was dangerous and should be governed with a hard hand. That conclusion led directly to 1919.
2The War and What It Cost
Punjab gave the Empire more soldiers than any other province, and was handed a bill instead of a reward.
While a few thousand Ghadar men were trying to start a mutiny, several hundred thousand other Punjabis were doing the opposite. They were enlisting.
Punjab was already the recruiting heart of the Indian Army before 1914. Part 11 explained why: after 1857 the British had rebuilt the army around a theory that certain communities were natural fighters, and had concentrated recruitment in the north-west. The war turned that arrangement into an industry.
Punjab held roughly one in every thirteen people in British India. It supplied roughly half of all the combat soldiers India sent to the First World War — somewhere around 360,000 men.
Put it another way. Imagine a country where one state has eight per cent of the people and provides fifty per cent of the soldiers. Now imagine that happening for four years without a break. That was Punjab between 1914 and 1918.
In some districts of the Rawalpindi and Jhelum area, so many men of fighting age were gone that the harvest was being brought in by women, boys and the old.
How the men were found
The Lieutenant-Governor of Punjab through almost the whole war was Sir Michael O'Dwyer, who held the post from 1913 to 1919. He is the second most important British name in this part, and in some ways the more important of the two.
O'Dwyer believed two things firmly. The first was that Punjab's villages were sound and its towns were poisonous — that the trouble in India came from educated city men, lawyers and journalists, and that the peasant was loyal if left alone. The second was that his job was to deliver soldiers.
He delivered them by pushing the demand down through the existing machinery of rural government. District officers were given targets. They passed the targets to tehsildars. Tehsildars passed them to the men who actually ran the villages — the zaildars, the lambardars, the sufedposhes — who held their positions and their land grants at the government's pleasure.
A lambardar was the government's man in a village: he collected the land revenue and took a cut of it. A zaildar was the same idea one level up, over a group of villages. Both were appointed, both were paid in status and land as much as in cash, and both could be removed.
This matters because it explains the recruiting system. The government did not knock on doors. It leaned on the men who already had power in the village, and let them do the knocking.
Formally, enlistment was voluntary. In practice a lambardar told to produce four men produced four men. There were villages where families were told their land claim would suffer, or their well would be reassessed, or their son's case in the revenue court would go badly. There were also villages where the recruiter did not need to say anything at all, because the pay was good and the alternative was a small share of a small holding.
Nobody was conscripted by law. Enlistment carried real rewards: regular cash pay sent home, a pension, and after the war, grants of land in the canal colonies. Families competed for places in favoured regiments. Military service was already a respected career in these districts, with fathers and uncles who had served. Men walked to recruiting depots on their own feet.
The government set district quotas and judged its officers on meeting them. When a state sets a target and hands enforcement to men who control your land, your water and your court cases, the word "voluntary" is doing very little work. Contemporary reports describe recruiting parties camping in villages until numbers were produced. Some districts saw open resistance to recruiters.
Where this leaves us
The honest answer is that both descriptions fit different men, and there is no way to count how many fall in each group. What can be said is that the choice was made inside a structure the family did not control. That is not conscription, and it is not free choice either. It is the ordinary way empires get what they want without admitting they are taking it.
What the province got back
Punjab was told, repeatedly and in public, that this effort would be remembered. Then the war ended.
What arrived instead was a long list of troubles, most of them arriving at once.
Prices had risen sharply during the war and did not come down. Grain, cloth and salt all cost more. Soldiers' remittances had propped up village incomes; when the army began demobilising, that money stopped.
The land revenue demand continued and in places was revised upward. War loans had been raised in the countryside, sometimes with the same gentle pressure that had produced the recruits.
Then, in the autumn of 1918, influenza arrived.
The influenza pandemic of 1918–19 killed somewhere between twelve and eighteen million people in India. Punjab was among the worst-affected provinces. Estimates for the province run to several hundred thousand dead in a matter of months.
Set that against the province's war dead, which were in the tens of thousands. The disease killed several times more Punjabis in one autumn than four years of fighting had. Villages that had sent men to France and Mesopotamia and got most of them back then buried people at home, faster than they could be counted.
And the men who had served came home to find that nothing about their standing had changed. They had been treated in the army as trusted professionals. They came back to a province where an Indian still stepped off the pavement.
The promise that was not kept
There was also a political promise. In August 1917 the Secretary of State for India, Edwin Montagu, had announced in the House of Commons that British policy was the gradual development of self-governing institutions in India. It was the first time the goal had been stated so plainly.
What that produced, in 1919, was the Government of India Act — a scheme usually called dyarchy.
Dyarchy means "rule by two". Under the 1919 reforms, the subjects a provincial government dealt with were split in half. Elected Indian ministers got one set — education, health, local government, agriculture. The British governor kept the other set, and the other set was the one that mattered: police, prisons, land revenue, and finance.
So Indians were given ministries with responsibility and no money, and no control over the men with guns. It was designed to teach responsible government. What it mostly taught was that responsibility without power is a trap.
So the offer on the table at the end of 1918 was: a limited share in government, arriving at some unspecified future date, in exchange for four years of blood, money and grain.
And then, before the reforms were even in place, the government did something else. It decided to keep its wartime emergency powers.
Michael O'Dwyer's entire method rested on a belief he never doubted: that the Punjabi countryside was loyal, and that disloyalty was an urban infection carried by educated men.
The assumption had a practical use. It justified governing through village bosses, ignoring the press, and treating lawyers and students as the only threat. It also flattered the officials who held it, because it meant the people they actually dealt with day to day liked them.
It was wrong in a specific way. The countryside was not loyal; it was compliant, which is different. It complied because complying paid, and because the men who could hurt it were on the government's side. Once the paying stopped — high prices, no reward, dead relatives, a revenue demand that did not soften — the compliance had nothing holding it up.
O'Dwyer could not see this coming, because his theory said it could not happen. In April 1919 it happened.
Punjab carried the Indian Army through the First World War. Around half of India's combat soldiers came from one province with a small fraction of India's people, extracted through a rural machine that made refusal expensive.
The province ended the war with high prices, a heavy revenue demand, hundreds of thousands dead of influenza, demobilised soldiers with nothing to come back to, and a political reform that gave elected Indians everything except power.
That is the ground on which 1919 happened. Jallianwala Bagh is usually told as a story about one general on one afternoon. It is better understood as what happens when a province is squeezed for four years and then sent a bill.
3Rowlatt and the Road to the Bagh
A law nobody in India wanted, a protest nobody had tried before, and six days in Amritsar.
During the war the government had ruled by emergency powers. It could arrest without a warrant, hold without trial, and try political cases before judges with no jury and no appeal. This was justified as a wartime necessity. The war ended in November 1918.
In March 1919 the government passed a law that kept those powers in peacetime. Its formal name was the Anarchical and Revolutionary Crimes Act. Everyone called it the Rowlatt Act, after the English judge whose committee had recommended it.
It allowed trial of political cases by three judges without a jury and with no right of appeal. It allowed suspects to be held without charge. Every single elected Indian member of the Imperial Legislative Council voted against it. It was passed anyway, by the official majority.
The Imperial Legislative Council had elected Indian members and appointed British officials. The officials were always more numerous, and voted as instructed. So Indian members could debate, amend, protest and vote — and lose every time, by design.
This is why the Rowlatt Act mattered out of proportion to its contents. It demonstrated, in public, that the Indian vote in Indian government was decoration.
A slogan went round the country in three words: na dalil, na vakil, na appeal — no argument, no lawyer, no appeal.
Gandhi's new weapon
Mohandas Gandhi had returned from South Africa in 1915 and had so far worked on local causes. The Rowlatt Act was the issue on which he went national. He called for a hartal — a general suspension of business — on 6 April 1919, followed by fasting, prayer and non-violent defiance of the law.
A hartal is a shutdown: shops close, work stops, streets empty. It is a strike by a whole population rather than by one workforce.
Satyagraha is Gandhi's word for his method. It joins satya (truth) and agraha (holding firmly). It means refusing to obey an unjust law openly, accepting the punishment without resisting, and refusing to injure the person enforcing it. A person doing this is a satyagrahi.
The hartal was very widely observed. In most of India it passed off peacefully. In Punjab it did not, and the reason was the ground described in the last chapter: a province wound tight by four years of extraction, disease and disappointment, governed by a man who believed none of that was happening.
Amritsar, 6 to 12 April
Amritsar observed the hartal completely on 6 April. What was striking to observers, and alarming to officials, was that Hindus and Muslims took part together. There were public meetings where Hindu speakers were invited to address Muslim gatherings and where members of both communities drank from the same cups, in deliberate public breach of custom.
Two local leaders became the focus. Dr Satyapal, a Hindu doctor, and Dr Saifuddin Kitchlew, a Muslim lawyer, had been organising jointly. To the administration this joint organising was the single most threatening fact in Punjab.
On 9 April, the festival of Ram Naumi was celebrated in Amritsar with Muslims joining Hindu processions.
On the morning of 10 April, Satyapal and Kitchlew were summoned to the Deputy Commissioner's house, arrested, and driven out of the district to Dharamsala.
A large crowd formed and marched towards the Civil Lines — the European quarter — to petition for their release. It was stopped at a railway bridge and fired on. Several people were killed.
The crowd went back into the city and the city broke. Banks were attacked and five Englishmen were killed. The telegraph office and a church were burned. An English schoolteacher, Marcella Sherwood, was pulled off her bicycle in a narrow lane and beaten nearly to death; she was rescued and hidden by Indian neighbours.
By 11 April the violence had spent itself and the city was quiet. On 11 April, Brigadier-General Reginald Dyer arrived from Jalandhar and took military command of Amritsar.
The week of 6–13 April 1919 is documented in unusual detail, because it was investigated twice by bodies that wanted opposite conclusions.
The government appointed the Disorders Inquiry Committee under Lord Hunter, which took sworn evidence from officials, including Dyer himself. The Indian National Congress ran its own inquiry, which took evidence from hundreds of Indian witnesses the official body did not call.
Where the two disagree, they disagree about numbers and about motive. Where they agree — on the sequence of events, on who did what, and on the orders given — we can be confident, because agreement between two hostile inquiries is the strongest kind of evidence there is.
The order
On the morning of 13 April, Dyer had a proclamation read out at points around Amritsar. It banned public gatherings and imposed a curfew. It was announced by drum in perhaps a dozen or so places in a large walled city.
13 April 1919 was Baisakhi.
Baisakhi is the spring harvest festival and, for Sikhs, the anniversary of the founding of the Khalsa. It is the biggest day in the rural Punjabi year. On that day Amritsar was full of villagers who had come in for the fair and to visit the Golden Temple.
Thousands of the people in the city on 13 April had walked in that morning from villages outside. They had not heard the proclamation, because it had been read out in a city they were not yet in.
Everything the administration did in that week rested on an assumption stated openly by O'Dwyer and never questioned: that Hindu–Muslim cooperation in Amritsar was artificial, the work of a few agitators, and would collapse if those agitators were removed.
That assumption is why Satyapal and Kitchlew were arrested. Remove the organisers and the unnatural alliance dissolves.
It produced the opposite. The arrests gave the two communities a shared grievance and a shared march. The administration's model of Punjab told it that pulling out two men would calm the city. The model was wrong, and the city was still burning two days later.
Notice how durable this assumption is. Twenty-eight years later, the same claim — that unity between communities is artificial and division is natural — would be used to argue for partition. It is the same assumption, pointing the other way.
The Rowlatt Act kept wartime emergency powers in peacetime, over the votes of every elected Indian. It gave Gandhi his first all-India cause and gave Punjab a spark in a province already packed with fuel.
In Amritsar the arrest of two leaders — one Hindu, one Muslim, working together — produced a march, then firing, then a riot in which five Englishmen were killed. By 12 April it was over. On 13 April, with the city already quiet and full of villagers who had come for Baisakhi, a British general arrived with an order banning gatherings and a determination to enforce it.
4Jallianwala Bagh
Ten minutes in a walled garden, and an argument about numbers that has lasted a century.
The Jallianwala Bagh was not a garden in any pleasant sense. It was a piece of waste ground of roughly six or seven acres inside the old city of Amritsar, close to the Golden Temple, used for meetings and fairs. Houses backed onto it on all sides. There was a well in it.
The crucial fact about it is the way in and out. It was enclosed by walls and the backs of buildings. The main entrance was a narrow lane, wide enough for a few people abreast.
Picture a space about the size of four or five football pitches, with a solid wall of buildings all the way round, and one usable door — a passage narrower than a single-lane road.
A space like that is safe when nothing is happening and lethal the moment something is. Everyone inside has to leave through a gap that only a handful can pass at a time.
On the afternoon of 13 April a public meeting was held there in defiance of the ban. Estimates of the crowd vary widely; contemporary figures run from around 6,000 to 20,000 or more. It contained political listeners, curious onlookers, and a large number of villagers in for the Baisakhi fair who were resting in the shade with their families.
Dyer learned of the meeting and marched to the Bagh with around fifty riflemen, drawn from Gurkha and Baluchi units, and a further group of soldiers armed only with kukris. He also brought two armoured cars mounted with machine guns. The cars could not get through the entrance lane and were left outside.
He entered by the narrow lane, positioned his riflemen on the raised ground just inside it, and gave the order to fire. No warning was given to disperse.
The firing went on for about ten minutes. It stopped when the ammunition was nearly exhausted. Roughly 1,650 rounds were fired.
1,650 rounds in ten minutes is about three shots every second, without pause, for the length of a song.
The troops were not firing into the air. Dyer stated afterwards that he had directed fire towards the places where the crowd was thickest, including towards the exits as people tried to reach them.
People trying to escape crowded the walls and the lane. Some jumped into the well. The troops then withdrew. No medical assistance was organised, and a curfew was in force that night, which meant families could not go out to look for the wounded until the following day.
The numbers
There is no agreed figure, and there never has been. It is worth understanding why.
The Hunter Committee accepted 379 dead and roughly 1,200 wounded, compiled afterwards by the administration from bodies identified and claims made. Its supporters say it is the only count based on named individuals rather than estimate.
The Congress inquiry concluded the true number was above a thousand. Its case: the official count depended on families coming forward, and under martial law that meant identifying your household as having defied a ban. Many of the dead were villagers whose families were not in the city to claim them.
Some historians point to a count by the Sewa Samiti, a local relief body, which compiled several hundred names, and argue for a figure in the high hundreds — too low for the official number, short of the round thousand.
Where this leaves us
The honest position is that we do not know, and the uncertainty is itself evidence. A state that kills people in a walled space and then does not count them is telling us how it regarded the dead.
Note also what nobody disputes: 1,650 rounds fired into an enclosed crowd at close range, without warning, with the exit blocked.
Martial law
Martial law was formally declared in Amritsar, Lahore and other districts on 15 April, and lasted into June. Under it came public floggings, arrests, aeroplane bombing of a crowd at Gujranwala, and orders designed less to keep order than to humiliate.
The most notorious was the crawling order. In the lane where Marcella Sherwood had been attacked, Dyer ordered that any Indian passing through must do so on hands and knees. It was enforced for about a week. People whose homes were on that lane could not leave except on their bellies.
Martial law means the army replaces the ordinary courts and administration. Military officers issue orders that have the force of law and try civilians in summary courts. There is no appeal to the normal legal system while it lasts.
Afterwards
The Hunter Committee reported in 1920. Its majority found that Dyer had acted beyond what the situation required, and criticised his belief that he was producing a moral effect across Punjab rather than dispersing a crowd. Dyer was relieved of his command and retired.
What happened next did more damage than the report. In Britain, the House of Lords passed a motion sympathetic to Dyer. A newspaper fund raised a very large sum for him from public subscription.
The fund raised for Dyer came to around £26,000. In the money of the time that was a fortune — many times the lifetime earnings of a British working man, and an unimaginable sum to an Indian farm labourer.
The message received in India was not that Britain had punished Dyer. It was that Britain had paid him.
Michael O'Dwyer, who had defended Dyer's action publicly and continued to defend it for the rest of his life, was shot dead in London on 13 March 1940 by Udham Singh, a Punjabi who had been present in Amritsar as a young man in 1919. Udham Singh was hanged in July 1940.
Dyer told the Hunter Committee something remarkable. He said he had not been trying only to disperse that crowd. He had been trying to produce an effect on Punjab as a whole — to make an impression that would prevent trouble everywhere.
Underneath that lies an assumption about what the Raj was. If your job is to police citizens, you use the least force needed on the people in front of you. If your job is to hold down a conquered population, you use whatever force will frighten everyone else. Dyer was not confused about his job. He answered honestly, and his honesty is what damaged him.
The deeper assumption is one the Empire could never say aloud: that British rule in India rested finally on fear, not consent. Dyer said it aloud. That is why officials found him embarrassing rather than incomprehensible.
On 13 April 1919 troops fired about 1,650 rounds into an enclosed crowd in Amritsar with no warning and the only exit blocked. The official death toll was 379; Congress said over a thousand; nobody counted properly, and that failure is part of the story.
What followed mattered as much: two months of martial law, the crawling order, and then a public fund in Britain for the man responsible.
1919 did not make Punjab nationalist overnight. What it did was destroy the idea that the Raj could be reasoned with as a lawful government. Every movement in the rest of this part begins from that.
5The Akali Decade
A campaign to reform temples that accidentally built the most disciplined political machine in Punjab.
The movement that followed 1919 in Punjab was not, at first, a political movement at all. It was an argument about who owned the gurdwaras.
Over the eighteenth and nineteenth centuries the great Sikh shrines had come under the control of hereditary managers called mahants. Part 10 described how this happened: land grants, endowments and a long period in which nobody had the authority to supervise them. By 1920 many mahants treated the shrines as family property, took the offerings as personal income, and in some cases had drifted far from Sikh practice — keeping idols in the precincts, barring people of low caste, and living in ways that scandalised ordinary worshippers.
A gurdwara is a Sikh place of worship. The word means "door of the Guru".
A mahant was the hereditary custodian of a shrine and its endowment. The position passed down within a family. Crucially, the British legal system recognised mahants as the lawful proprietors, which meant that removing one was not a religious dispute but a challenge to a property right the government was obliged to defend.
That last point decides everything that follows. A reform movement against the mahants was, in law, a movement against the government. The reformers did not choose that fight. The property law chose it for them.
Two organisations in one month
In November 1920 a representative body was formed to manage the shrines: the Shiromani Gurdwara Parbandhak Committee, or SGPC. In December 1920 a second body was formed to supply the volunteers who would take the shrines: the Shiromani Akali Dal.
Akali means a follower of Akal, the Timeless One — God. The name was borrowed from an older warrior tradition.
A jatha is a band or a squad — a group of volunteers sent to do something together.
A morcha is a front, in the military sense; in Punjabi political usage it means a sustained campaign of confrontation over one issue.
The method was Gandhian in form and Sikh in discipline. A jatha would march to a shrine, unarmed, take an oath not to strike back, and present itself. It would be beaten, arrested, or both. The next jatha would arrive the following day. Then the next.
Nankana Sahib
On 20 February 1921 a jatha led by Bhai Lachhman Singh entered the gurdwara at Nankana Sahib — the birthplace of Guru Nanak. The mahant, Narain Das, had hired armed men in anticipation. The gates were shut and the jatha was killed inside the compound. Contemporary accounts put the dead at around 130. Bodies were burned.
Around 130 unarmed men, who had taken an oath not to resist, killed in the compound of the shrine marking the birthplace of the founder of their faith.
If you want to understand why the Gurdwara Reform movement became unstoppable within weeks, that is the whole explanation. No amount of organising could have produced what one morning at Nankana Sahib produced.
Two further campaigns defined the decade. At Guru ka Bagh, near Amritsar, in 1922, jathas were beaten systematically by police over several months while crowds watched and Indian and foreign observers reported it. At Jaito, in the princely state of Nabha, a morcha begun in 1923 ran for years, and a jatha marching there was fired on in February 1924.
The Akali morchas are among the best-documented protest campaigns of the period, for a reason the movement engineered on purpose.
The SGPC kept lists — of jathas sent, of men arrested, of men injured, of money collected. It invited journalists. At Guru ka Bagh the beatings were witnessed by reporters and by the English clergyman C. F. Andrews, whose published account caused an uproar.
This was strategy, not luck. A movement whose entire method is being beaten without hitting back needs the beating to be seen. The record exists because being on the record was the weapon.
The Act, and what it built
The government gave way. The Sikh Gurdwaras Act was passed in 1925. It transferred control of the historic shrines to an elected committee — the SGPC — chosen by a register of Sikh voters.
The movement had won what it asked for. It had also, without quite intending to, created something new.
The SGPC controlled the great shrines. The shrines had enormous incomes from offerings. The SGPC was filled by elections on a Sikh electoral roll. And standing beside it was the Akali Dal, a party with a proven ability to put thousands of disciplined volunteers on the road at short notice.
It recovered the central institutions of the faith from hereditary owners and put them under community control, which had been the demand of reformers for fifty years. It did so with almost no violence from its own side, and it won. It gave Punjabi Sikhs a mass political education, a national reputation, and an organisation that could act. Nothing else in Punjab in the 1920s achieved as much by peaceful means.
It fused religious authority and political power in one machine, and tied both to an electorate defined by religion. From 1925 onward, the way to lead Sikhs politically ran through control of their shrines and their money. That made Sikh politics intensely factional, and it locked Sikh political identity into a religious frame at the exact moment when Punjab needed cross-community politics most. Some historians argue this made the Akali Dal structurally unable to join the Unionist bargain described in Chapter Seven.
Where this leaves us
The disagreement is less about facts than about what you are comparing with. Against the alternative of leaving the shrines with the mahants, the movement was plainly a success. Against an imagined Punjab where politics was organised by class or region rather than religion, it was a step away. The second comparison is with something that never existed, which is worth remembering before treating it as a verdict.
Both the reformers and the government assumed the question "who is a Sikh?" had a clear answer, because the Act required one. To hold elections you need a roll. To make a roll you need a definition.
In 1925 that was not obvious at all. Punjabi religious practice was full of households that visited gurdwaras, shrines of Muslim saints and Hindu temples in the same week, and families where one son kept his hair and another did not.
The Act had to draw a hard line through that soft ground. So did the census. So did the separate electorates described in Chapter Seven. The line was necessary for administration, and once drawn it began to feel like a description of reality rather than a decision made for convenience. Much of the politics of the 1940s runs on lines like this one.
The Gurdwara Reform movement set out to take the Sikh shrines from hereditary managers. Because the law protected those managers, the fight became a fight with the government.
It was won by sending unarmed volunteers to be beaten, in public, repeatedly, until the government gave way in 1925. The killings at Nankana Sahib in 1921 turned it into a mass movement.
The lasting result was not only the Act. It was the SGPC and the Akali Dal — money, shrines, an electoral roll and a volunteer army, all organised around religion. That combination is the third force in every political calculation in Punjab from 1925 to 1947.
6Bhagat Singh
A young man who threw a bomb designed not to kill, and used his trial as a printing press.
Bhagat Singh was born in 1907 in Banga, in Lyallpur district, into a family already in politics — an uncle had been imprisoned, his father had been jailed. He was eight when Kartar Singh Sarabha was hanged. He grew up with Ghadar as family history rather than distant news.
He was twelve at the time of Jallianwala Bagh. He joined the non-cooperation movement as a schoolboy, and left it, like many others, when Gandhi called the movement off in 1922 after violence at Chauri Chaura. That withdrawal is important. For a certain kind of young man in Punjab it settled the question of whether Congress methods could work.
The organisation
By the late 1920s he was part of a small revolutionary group that in 1928 renamed itself the Hindustan Socialist Republican Association. The middle word was added deliberately.
Republican here means: no king, no viceroy, no princes. Power comes from the people and from nowhere else.
Socialist in the HSRA's usage meant that removing the British was not enough — the land and the factories had to change hands too, or an Indian rich man would simply replace an English one.
Adding that word was a direct criticism of the Congress. It said: your independence would leave the poor exactly where they are.
In October 1928 the Simon Commission — a body sent from Britain to review Indian government, containing no Indians at all — reached Lahore. A protest against it was led by the veteran nationalist Lala Lajpat Rai. Police broke it up with lathis, and Lajpat Rai was struck. He died on 17 November.
The HSRA decided on revenge against the police superintendent held responsible, James Scott. On 17 December 1928 they shot the wrong man: J. P. Saunders, a junior officer.
This is worth stating plainly rather than smoothing over. The most famous action in the group's history before the Assembly bomb was a killing carried out by mistake.
The bomb that was meant to fail
On 8 April 1929, Bhagat Singh and Batukeshwar Dutt entered the visitors' gallery of the Central Legislative Assembly in Delhi and threw two bombs onto the empty benches below, along with printed leaflets. The bombs were low-powered and thrown away from where members were sitting. A few people received minor injuries.
The two men did not run. They stayed, shouting slogans, and waited to be arrested.
The leaflet contained the line the action was built around: that it takes a loud noise to make the deaf hear.
Almost every argument about Bhagat Singh assumes the Assembly bomb was an act of violence that failed to kill. Read it that way and it looks incompetent.
It was not designed to kill. Weak bombs, thrown at empty benches, by men who stayed to be caught. It was designed to produce a trial.
The assumption worth naming is a general one: that political violence is always aimed at the body it is thrown at. Sometimes the target is the audience. Bhagat Singh wanted a courtroom, a newspaper report, and a stage — and he got all three, at the price of his life. Whether that is admirable or appalling is a separate question. It is not, however, a failure.
The trial
He was tried for the Assembly bomb, and then separately for the Saunders killing in the Lahore Conspiracy Case. He used both proceedings as a platform, making statements that were reported across India.
In prison he and his comrades demanded to be treated as political prisoners rather than common criminals, and went on hunger strike. The strike lasted months. Jatindra Nath Das, a fellow prisoner, died on 13 September 1929 on the sixty-third day of his fast. His funeral procession drew enormous crowds.
In October 1930, in jail, Bhagat Singh wrote the essay he is least often quoted on and which he thought most important: an explanation of why he had stopped believing in God. He argued that a man facing execution has every reason to comfort himself with faith, and that refusing that comfort was the test of whether his beliefs were real.
Bhagat Singh, Sukhdev Thapar and Shivaram Rajguru were hanged in Lahore Central Jail on 23 March 1931. He was twenty-three.
He was twenty-three at his death, and twenty-one when he threw the bomb. Jatindra Nath Das was twenty-four when he died. Kartar Singh Sarabha, whose photograph Bhagat Singh carried, had been nineteen.
These were not seasoned commanders. They were students, at the age of a final-year undergraduate.
He belonged to an organisation that shot a police officer dead in the street, and shot the wrong one. He threw explosives into a legislature. He accepted that killing was a legitimate political tool. Judged by the standard applied to armed groups anywhere, that is what the word describes, and applying a gentler word to people we happen to admire is not analysis.
Terrorism, as normally defined, aims at civilians to spread fear. The HSRA targeted state officials, and the one action everyone remembers was engineered to injure nobody. He surrendered rather than escaping, which no clandestine terrorist does, and he used the courtroom rather than the bomb to make his case. Judged by conduct rather than category, he was a revolutionary who used a small amount of violence for publicity and then argued in public.
Where this leaves us
The dispute is partly about the word rather than the man. The facts are not seriously contested: one killing by mistake, one bomb designed not to kill, a surrender, and two years of argument from a cell.
A fairer question is whether the strategy worked. It did not remove the British. It did create a figure that every later Indian movement, from the Congress left to Punjabi communists to modern nationalists, has claimed as its own — which suggests he was right about where the real power lay.
We have a great deal in his own hand: jail notebooks with his reading lists, letters, the court statements, and the 1930 essay on atheism. We also have the full apparatus of a state trying to convict him, which produced transcripts and evidence files.
One warning. Because he became a symbol, a large quantity of material is attributed to him that he did not write. Slogans and quotations circulate under his name with no source at all. The safe rule is to trust the jail writings and the court record, and to be sceptical of anything that sounds like it was written for a poster.
Bhagat Singh's politics were republican and socialist, and were aimed as much at the Indian rich as at the British. He believed Congress would win a freedom that changed nothing for the poor.
His central action was a bomb designed not to kill, followed by a voluntary arrest, because he wanted a trial and a hearing rather than a body count. He got both, and was hanged at twenty-three.
He is the point where Punjab's revolutionary tradition — Ghadar, 1919, the beatings of the 1920s — turns into something every later Indian politics wants to inherit. Notice that this is also the last moment in this part when Punjabi politics is organised around something other than religious community.
7The Unionist Bargain
For twenty years Punjab ran on a party that put Muslim, Hindu and Sikh landholders on the same side. Nowhere else in India managed it.
Here is the fact that makes Punjab different from every other Indian province between the wars, and it is the fact most often left out.
Punjab was not governed by the Congress. It was not governed by the Muslim League. From the 1920s until 1947 it was governed by a party built on the idea that the important division in Punjab was not between religions at all — it was between the countryside and the town.
Who was on which side
The Unionist Party was founded in 1923. Its architects were Mian Fazl-i-Husain, a Muslim lawyer from Gurdaspur, and Chaudhri Chhotu Ram, a Hindu Jat from Rohtak. Later its outstanding leader was Sir Sikandar Hyat Khan, from a large landholding family in Attock.
Its argument ran like this. The Punjabi peasant — Muslim in the west, Sikh in the centre, Hindu Jat in the south-east — had the same problems everywhere: debt, the moneylender, the price of grain, water rates and land revenue. The moneylender and the grain dealer were usually urban and usually Hindu; the peasant was Muslim in one district and Sikh in the next.
So a party of cultivators, of all three religions, against the towns and the moneylenders.
A law of 1900, the Punjab Alienation of Land Act, had divided Punjabis into "agricultural tribes" and everyone else, and barred land from passing from the first group to the second. It was designed to stop peasants losing land to moneylenders.
Its side effect was to create a legal category cutting straight across religion. A Muslim Rajput, a Sikh Jat and a Hindu Jat were all in it; an urban Hindu trader and an urban Muslim shopkeeper were both outside it.
The Unionist Party is, in effect, that legal category turning into a political one. This is what Part 11 meant when it said British land law would shape Punjabi politics for fifty years.
Why it worked
Two further things made the Unionists strong.
The first was the franchise. Under the reforms of 1919 and then the Government of India Act of 1935, the vote was tied to property and to land revenue payment. That weighted the electorate heavily towards landholders and rural districts, and against the towns where Congress was strong.
The second was that the party had something to deliver. Chhotu Ram in particular pushed through laws restricting moneylenders, regulating agricultural markets, and relieving debt. A Unionist MLA could go back to his constituency and point at something.
In the elections of 1937 — the first under the new constitution, and the first real test — the Unionists took about 95 of the 175 seats in the Punjab Assembly. The Congress won 18. The Muslim League won two.
Two seats. In the province that would later be the whole point of the demand for Pakistan, in a province with a Muslim majority, the Muslim League won two seats out of 175 in 1937.
Hold that number in your head. It is the single best measure of how far Punjab travelled in nine years. In 1937 the League was nothing here. In 1946 it was the largest party in the province.
Sikandar Hyat Khan became Premier of Punjab and governed with a cabinet containing Muslims, Hindus and a Sikh.
Election returns for 1937 and 1946 survive in full — seats, parties, and in many constituencies the vote counts. Assembly debates were recorded and printed. Unionist ministers left correspondence, and the governors sent regular confidential reports to the Viceroy which have since been published.
One limit is worth stating. The electorate was small — a fraction of adults, weighted to property. So the returns tell us what enfranchised Punjab wanted, not what Punjab wanted. When people say the 1946 result proved what Punjabi Muslims believed, that is worth remembering.
The weakness inside it
The Unionist bargain had a flaw that was invisible while things went well.
It was a coalition of landholders. It asked its voters to think of themselves as cultivators first and Muslims, Hindus or Sikhs second. That works while the questions on the table are debt, water and grain prices — questions where the cultivator identity is the useful one.
It stops working the moment the question on the table is: who will rule when the British leave? There is no agricultural answer to that question.
From 1909 onward, Muslims voted in separate electorates: Muslim voters chose Muslim candidates for seats reserved for Muslims. Later, seats were also reserved for Sikhs.
This meant a Muslim politician in Punjab never needed a single Hindu or Sikh vote, and could not gain one. The Unionists held their alliance together at the level of the Assembly, between elected members. At the level of the voter, the system was already communal.
That gap — cross-community at the top, single-community at the bottom — is the crack the Muslim League would later drive a wedge into.
They governed a province of nearly thirty million people for a decade without communal breakdown, at a time when politics elsewhere in India was hardening. They passed real laws that helped real farmers. They showed that Punjabis of three religions could sit in one cabinet. If that had held, partition in Punjab would have had no political vehicle.
They were a coalition of the propertied, elected on a narrow franchise, defending an order that suited landlords. They did nothing to build a shared Punjabi identity below the level of the landholding classes, and they depended on the British to keep the constitutional questions off the table. Their unity was an arrangement among elites with common interests, not a movement. Arrangements like that do not survive the question of who inherits the state.
Where this leaves us
What is not in doubt is that it worked for twenty years and then collapsed very fast. The disagreement is about whether it collapsed because of outside pressure — the League, the war, the end of British power — or because it was always shallower than it looked.
Chapter Nine gives the evidence for both readings. Notice as you read it that a thing can be genuinely valuable and structurally fragile at the same time. The Unionist bargain was not a fraud. It was simply built for a set of questions that stopped being the questions.
Punjab between the wars was governed by the Unionist Party — a cross-community alliance of Muslim, Hindu and Sikh landholders, organised around rural interests against the towns and the moneylenders.
In 1937 the Unionists won about 95 of 175 seats; the Muslim League won two. That number is the baseline for everything in the next two chapters.
The bargain worked because the questions were agricultural. It had no answer to the question of who would inherit power from the British, and the electoral system beneath it was already divided by religion.
8Three Communities, Two Futures
Why the demand for Pakistan was announced in Lahore, and why Punjab was the one place it could not be granted cleanly.
On 23 March 1940, the All-India Muslim League met at Minto Park in Lahore and passed the resolution that history remembers as the demand for Pakistan.
It did not use the word. It said that areas where Muslims were in a majority, in the north-west and the east of India, should be grouped to form independent states. It was vague on purpose. Vagueness kept everyone on board.
It is worth pausing on the location. The League passed its founding demand in the capital of a province where, three years earlier, it had won two seats.
The arithmetic that made Punjab impossible
Punjab's population in the 1941 census broke down roughly as follows: Muslims a little over half, Hindus around a third, Sikhs somewhat over a tenth.
Round the 1941 figures for undivided Punjab: about 53 in every 100 Punjabis were Muslim, about 30 were Hindu, and about 15 were Sikh.
Now notice two things about those numbers. First, the Muslim majority was real but narrow — nothing like the four-fifths of Sind or the North-West Frontier. Second, that majority was concentrated in the west. In the central districts around Amritsar and Lahore, the three communities were mixed together, village by village.
This is why Punjab could not be handed to either side whole. A province that is 53 per cent one thing has a 47 per cent problem, and in Punjab that 47 per cent lived in the districts with the best land, the biggest cities and the holiest shrines.
The Sikh position
Which brings us to the community with the weakest hand and the strongest attachment to the ground.
Sikhs were about a seventh of Punjab. They were a majority in no district. But they held a share of the land, of the canal colonies and of the army far above their numbers, and the shrines central to their faith — Nankana Sahib, and much else — lay in the western districts where Muslims were the clear majority.
So every available future was bad. A united India left them a small minority in a large country. A Pakistan containing all of Punjab left them a small minority under a Muslim government, with their best land and holiest sites inside it. A divided Punjab split the community itself and left the birthplace of Guru Nanak on the far side of a border.
The Akali Dal, led from the 1930s by Master Tara Singh, tried several answers. The best known was the Azad Punjab scheme of 1943 — a proposal to redraw Punjab's boundaries so that no community held a majority, forcing all three to share. It attracted little support outside Sikh circles, because it asked Muslims to give up a majority they already had.
Every proposal on the table by 1940 — Congress's, the League's, and the Akalis' — assumed that the right unit of politics was the religious community, and that the question was how to divide power between such communities.
Once everyone accepts that, the argument is only about percentages and boundaries. And an argument about percentages has a logic of its own: it rewards whoever can show the largest, most unified bloc, and it punishes anyone whose supporters are mixed in with someone else's.
The Unionists were the last significant party operating on a different assumption. Their decline is not only a political defeat. It is the moment Punjab loses its only vehicle for a politics not counted by religion — after which there is no longer any way to be counted except as a Muslim, a Hindu or a Sikh.
Sikandar's deal and its price
Sikandar Hyat Khan faced a specific problem. He led a cross-community party, but he was a Muslim leader in a Muslim-majority province at a moment when Jinnah was claiming to speak for all Indian Muslims. If Sikandar refused that claim outright, he risked being painted as a British stooge who had sold out his community.
His answer was the Sikandar–Jinnah Pact of October 1937. Unionist Muslim members would join the Muslim League at the all-India level, while the Unionist Party continued to govern Punjab unchanged.
On paper it cost nothing. In practice it was a slow disaster.
It conceded, in public, that the League spoke for Muslims in national matters. It gave the League a foothold and a membership list in a province where it had none. And it told Punjabi Muslim voters that being Unionist and being League were compatible — so that when the two later came into conflict, choosing the League did not feel like a betrayal.
Punjab had a functioning cross-community government until 1947. Its Muslim leadership was hostile to dividing the province, and Sikandar said publicly that he did not want a Pakistan that meant Punjab's break-up. The 1937 result shows the League had almost no organic support here. On this reading, partition was imposed on Punjab by all-India politics, a war, and a British decision to leave in a hurry — not generated by Punjab itself.
The cross-community arrangement existed only among elected members and landholding elites. Beneath it, voters had been sorted into separate electorates for forty years and had never once cast a vote across community lines. The Sikh position was unresolvable by any scheme anyone proposed. And the Unionists depended on the British staying — their whole model assumed someone else would answer the sovereignty question. On this reading the collapse was built in, and 1946 merely revealed it.
Where this leaves us
Historians remain genuinely divided, and the split tends to follow a wider disagreement about whether partition was long-prepared or a late accident.
What can be said with confidence is narrower and still important. Punjab was the province with the most to lose from partition and the least communal politics for most of this period. Whatever explains 1947, it cannot be that Punjabis had always been divided, because for twenty years they demonstrably were not.
The Lahore Resolution of 1940 demanded that Muslim-majority areas be grouped into independent states. It was passed in a province where the League had won two seats in 1937.
Punjab was roughly 53 per cent Muslim, 30 per cent Hindu and 15 per cent Sikh, with the communities most tangled together in the richest central districts. That arithmetic made a clean transfer of the whole province impossible.
Sikhs had no majority anywhere and the most to lose from every option, and never found a scheme anyone else would accept. Sikandar's 1937 pact with Jinnah looked cheap and quietly handed the League the ground it later stood on.
9The Ground Gives Way
A second war, a second recruiting drive, and five years in which the Unionist bargain came apart.
In September 1939 the Viceroy declared that India was at war with Germany. He did not consult a single Indian legislature. The Congress ministries in eight provinces resigned in protest.
The Unionists did not resign. Punjab stayed in office, and went back to doing what it had done in 1914.
The sword arm again
Punjab supplied soldiers on an even larger scale than in the first war. Recruitment ran through the same rural machinery, and the money and the land grants flowed the same way. For the Unionist leadership this was not merely loyalty. Wartime spending, procurement of grain at guaranteed prices and soldiers' remittances all flowed into the countryside they represented.
Punjab again provided a share of the Indian Army far beyond its share of India's people. Hundreds of thousands of Punjabis served in the Second World War.
Consider what that means for the politics of 1946 and 1947. In villages across the province there were now men of fighting age who had been trained to use modern weapons, had served in organised units, and had recently come home. When violence arrived, it did not arrive among people who did not know how to fight.
The war also produced a rival claim on those same soldiers. Indian prisoners captured by the Japanese were recruited into the Indian National Army, which fought alongside Japan under Subhas Chandra Bose. Punjabis were prominent in it. When INA officers were put on public trial in Delhi in 1945, the trials became a nationalist cause across India, including in Punjab.
What changed underneath
Four things eroded the Unionist position between 1939 and 1945.
The Congress vacated the field. By resigning in 1939 and then launching the Quit India movement in 1942, the Congress removed itself from provincial government and got its leadership imprisoned. The Muslim League, which did neither, spent the war years organising freely while its main rival was in jail.
The League got a machine. It built branches, recruited students from Aligarh and Lahore, and worked through rural networks of pirs — the hereditary custodians of Muslim shrines, who had influence over large numbers of followers in western Punjab.
A pir is a Muslim spiritual guide. In rural western Punjab, pirs of established shrines held hereditary followings, and many were also substantial landholders.
They matter here because they were the one network that could reach Muslim villagers directly, in the language of religious obligation. When significant numbers of pirs came out for the League, the League acquired in a few years what the Unionists had built over a generation.
Sikandar died. He died in December 1942, aged fifty. He was the one Punjabi Muslim politician with the personal standing to hold the cross-community line and manage Jinnah at the same time. He was succeeded by Khizr Hyat Khan Tiwana, a capable man with less authority in his own community.
The argument changed shape. By 1944 the question in Indian politics was no longer what a provincial government should do about debt or water rates. It was what would happen when the British left. The Unionist Party had no position on that, because having one would have split it.
1946
Elections were held in the winter of 1945–46. The League campaigned in Punjab on one question: are you a Muslim, and do you want a Muslim homeland?
The result transformed the province.
| Party | Seats (of 175) |
|---|---|
| Muslim League | 73 |
| Congress | 51 |
| Akali Dal (Panthic) | 22 |
| Unionist Party | 20 |
The Unionists went from about 95 seats to 20 in nine years. The League went from 2 to 73.
That is not a swing. That is one political world being replaced by another. And notice the shape of the new Assembly: three blocs, each drawn from one community, and no bloc with a majority.
The League was the largest party and had no majority. So a coalition was formed without it — the remaining Unionists, the Congress and the Akalis — with Khizr Hyat Tiwana as Premier.
This was constitutionally proper and politically explosive. Punjabi Muslims had voted overwhelmingly for the League and were being governed by a ministry the League was excluded from, led by a Muslim the League regarded as a traitor.
The five months
Through late 1946 the League ran an agitation against the Khizr ministry. On 2 March 1947 Khizr resigned. No alternative government could be formed. The Governor took direct control.
The next day, Master Tara Singh appeared outside the Punjab Assembly and made a public declaration against Pakistan. Within days the violence began — not in the mixed central districts, but in the north-west, in the Rawalpindi division, where Sikhs and Hindus were a small scattered minority among a large Muslim majority.
Thousands were killed in March 1947 in the Rawalpindi, Attock and Jhelum districts. Villages were destroyed. Survivors fled east.
For March 1947 we have official casualty returns, police reports, and the fortnightly reports the Governor of Punjab sent to the Viceroy, most of which have been published. We also have refugee testimony collected at the time by relief committees, and a large body of oral history recorded much later.
These sources do not agree on numbers, and the official figures are certainly low, because administration in the affected districts had partly stopped functioning. What they agree on is the pattern: organised attacks on minority villages, not spontaneous rioting between neighbours of equal strength.
The March violence matters for a reason often missed. It happened before partition was announced. It is not a consequence of the border. It is part of what produced it.
It is very common to read that Punjab's communities "turned on each other" in 1947, as though a shared temper snapped.
That phrasing hides an assumption: that violence between communities is something that wells up from below, from ordinary people who secretly hated each other all along.
The evidence from March 1947 does not support it. The attacks were organised, often by armed bands, often involving men with military training, in districts where one community heavily outnumbered another. They followed a political collapse at the top — the fall of a ministry and the failure to form another — rather than preceding it.
This does not make ordinary Punjabis innocent; many took part. It does mean the sequence runs downward, from the failure of politics to the killing, and not upward.
Punjab fought the Second World War for the Empire as it had the first, which kept money flowing to the countryside and put hundreds of thousands of trained men back into its villages.
The Unionists collapsed between 1939 and 1946 because Congress left the field, the League built a rural machine through the pirs, Sikandar died, and the central question changed from agriculture to sovereignty — the one question the Unionist bargain could not answer.
In 1946 the League took 73 seats and the Unionists 20, and a coalition governed without the party most Muslims had voted for. It fell on 2 March 1947. Mass killing began in the Rawalpindi division within days — three months before partition was announced.
10The Decision
Five weeks in the summer of 1947, in which Punjab agreed to divide itself and then waited to be told where.
By the spring of 1947 three things were fixed. Britain was leaving, and soon. The Muslim League would not accept a united India. And the Congress and the Akalis would not accept a united Punjab inside Pakistan.
Once those three positions were on the table, the outcome followed almost mechanically. If there was to be a Pakistan, and if Punjab's non-Muslims would not go into it, then Punjab had to be cut.
The logic is worth stating in its own words, because it is the same logic that had been used to demand Pakistan in the first place. The League argued that Muslims could not live as a permanent minority under a Hindu majority. The Congress and the Akalis replied: then neither can Hindus and Sikhs live as a permanent minority under a Muslim majority in Punjab. There was no answer to that which did not divide the province.
3 June
Lord Mountbatten, the last Viceroy, announced the plan on 3 June 1947. Britain would transfer power in August — far earlier than anyone had expected. India would be divided. And Punjab and Bengal, the two provinces with large mixed populations, would each be divided too, if their legislators voted for it.
The voting mechanism was ingenious and revealing. The members of the Punjab Legislative Assembly would meet in two separate groups: those from Muslim-majority districts, and those from the rest. If either group voted for partition, Punjab would be partitioned.
Look closely at that rule. It does not ask whether Punjab wants to be divided. It asks each half separately, and division wins if either says yes.
Built into the procedure is an assumption that had by then become invisible: that Punjab was not one political community whose collective will mattered, but two communities sharing an administrative box.
A different rule — a simple majority of the whole Assembly, or a referendum of all Punjabis — would very likely have produced a different answer, because the Muslim majority of the province did not want Punjab divided. The rule chosen guaranteed division. It was not neutral machinery. It was the answer, written as a procedure.
The Assembly met on 23 June 1947. The two sections voted separately. The section from the non-Muslim-majority areas voted for partition. That was sufficient. Punjab would be divided.
The man with the pencil
A Boundary Commission was appointed to decide where the line went. Its chairman was Sir Cyril Radcliffe, an English barrister.
He had never been to India. He arrived on 8 July 1947.
Radcliffe landed in India in the second week of July. Power was to be transferred on 14 and 15 August.
That gave roughly five weeks to divide two provinces containing tens of millions of people, in a country he had never seen, during the monsoon, using census returns and maps.
To put the scale in ordinary terms: he was drawing a border through farmland, canal systems, railway lines, city suburbs and villages where the two communities lived on the same street — at a rate of a district every day or so.
The Punjab Boundary Commission had four other members: two judges nominated by the Muslim League and two by the Congress. They deadlocked on every significant question, exactly as anyone could have predicted, which left the decision to Radcliffe alone.
His instructions were to divide the province on the basis of contiguous majority areas of Muslims and non-Muslims, while also taking account of "other factors" — which were not defined.
The rule meant: group together neighbouring areas where one community is the majority, and put them on that community's side.
It sounds objective. It is not, for two reasons. First, majority in what unit? Counting by district, by tehsil or by village gives different borders. Second, the rule says nothing about canals, roads, railways or which side a city's water supply comes from — and those were exactly what the undefined "other factors" had to cover.
Critics point to specific decisions, above all the award of the Muslim-majority tehsils of Ferozepur and Gurdaspur to India, which gave India a road route towards Kashmir. Papers published later suggested the Viceroy discussed boundary questions with Radcliffe. On this view the line was politically shaped and then presented as a technical judgment.
Defenders argue there was no fair line available. Any border through central Punjab would strand millions. Radcliffe was given contradictory instructions, an impossible deadline and a deadlocked commission. Both sides denounced the award, which is at least consistent with it not favouring either. The disputed decisions can be defended on the "other factors" the terms of reference required him to weigh.
Where this leaves us
The specific allegations remain contested and some of the evidence is indirect. What is not contested is the process: five weeks, no local knowledge, a deadlocked commission, undefined criteria, and a single foreign lawyer with final say.
It is worth separating two questions that usually get merged. Whether Radcliffe was leaned on is disputed. Whether the procedure was a reasonable way to divide a province of thirty million people is not really in dispute at all.
The silence
Radcliffe finished his award before independence. It was not published.
The decision was taken to withhold it until after the transfer of power, so that the new governments would be responsible for what followed. It was made public on 17 August 1947 — two days after independence.
So on 14 and 15 August 1947, millions of people in central Punjab celebrated, or dreaded, an independence whose border they had not been told. Villages did not know which country they were in. Officials did not know which government they served. Families deciding whether to leave were deciding without the one piece of information that mattered.
The June–August 1947 period is documented from the British side in extraordinary detail. The Viceroy's staff kept minutes of meetings, and the transfer-of-power papers were later published in full, in twelve volumes.
This creates a trap worth naming. Because the British record is so rich and so accessible, it is easy to write the story of 1947 as a story of decisions made in Delhi drawing rooms — which is exactly how the people in those rooms saw it.
What survives far less well is what was happening in a village in Gurdaspur that August. That evidence is thinner, later, and mostly oral. Part 13 leans on it heavily, and it will say clearly where it is on soft ground.
Everything in the summer of 1947 assumed one thing: that the population would stay where it was, and that a border would simply place people under a new government.
Nobody planned for mass movement. There was no serious scheme to move people, protect them in transit, or receive them. A single military force, the Punjab Boundary Force, was created for the whole of central Punjab, and it was small.
The assumption was not stupid — it was how borders had usually behaved. It was simply not true of this one, and by March 1947 there was already evidence that it was not true, because people were already fleeing Rawalpindi.
That gap between what was assumed and what was already happening is where Part 13 begins.
Once Britain agreed to Pakistan and Punjab's Hindus and Sikhs refused to enter it, dividing the province followed automatically. The 3 June plan set the mechanism, and on 23 June the Punjab Assembly, voting in two separate sections, chose partition.
The line was drawn by a British lawyer who had never been to India, in about five weeks, with a deadlocked commission and undefined criteria — and it was kept secret until two days after independence.
The province that had held a cross-community government together for twenty years was cut apart in five months. Nobody planned for anyone to move.
Sources & further reading — Part 12
Glossary
Every term explained in this part, gathered in one place.
- Agriculturist tribes A legal category created in 1900, listing communities allowed to buy agricultural land in Punjab. It cut across religion and became the basis of Unionist politics.
- Akali Literally a follower of Akal, the Timeless One. Used from 1920 for the volunteers, and then the party, of the Gurdwara Reform movement.
- Contiguous majority areas The main rule given to the Boundary Commission in 1947: group neighbouring areas by which community held the local majority. It depended entirely on which unit you counted in.
- Dyarchy "Rule by two." The 1919 arrangement splitting provincial government between elected Indian ministers, who got the minor subjects, and the British governor, who kept police, revenue and finance.
- Ghadar Urdu for mutiny or revolt. The name of the newspaper founded in San Francisco in 1913 and of the party built around it.
- Gurdwara A Sikh place of worship; literally "door of the Guru".
- Hartal A general shutdown of shops and work, used as a protest by a whole population rather than one workforce.
- Jatha A band or squad of volunteers, sent together to carry out an action. The basic unit of the Akali morchas.
- Lambardar The government's appointed man in a village, responsible for collecting land revenue and paid a share of it. The bottom rung of the machinery that produced army recruits.
- Mahant The hereditary custodian of a shrine and its endowment. British law recognised mahants as proprietors, which turned the Gurdwara Reform movement into a fight with the state.
- Martial law Rule by the army in place of the ordinary courts, with military officers issuing binding orders and trying civilians summarily. Imposed on central Punjab from April 1919.
- Morcha A front, in the military sense; in Punjabi political usage, a sustained campaign of confrontation over a single issue.
- Official majority The arrangement by which appointed British officials always outnumbered elected Indians in the central legislature, so Indian members could never win a vote.
- Pir A Muslim spiritual guide. In rural western Punjab, pirs of established shrines held hereditary followings and often large landholdings, and their support was decisive for the Muslim League in the 1940s.
- Satyagraha Gandhi's method: openly disobeying an unjust law, accepting the punishment, and refusing to injure the person enforcing it. One who does this is a satyagrahi.
- Separate electorates A system under which Muslim voters chose Muslim candidates for reserved Muslim seats, and later Sikhs likewise. It meant no Punjabi politician ever needed a vote from another community.
- Transportation A sentence of removal to an overseas prison colony, in practice the Cellular Jail in the Andaman Islands, known to prisoners as Kala Pani, the black water.
- Zaildar A government-appointed official over a group of villages, one level above the lambardar. Part of the same rural machinery.