1The Last Part
Fifteen parts of invasions, and the worst thing that ever happened to the Punjab was done by Punjabis, to each other, in about ten weeks.
This book has described armies coming down the same road for five thousand years.
Persians, Greeks, Kushans, Huns, Arabs, Turks, Mongols, Afghans, Mughals, Persians again, Afghans again,
and finally an army that came up the Indus from the sea. Every one of them killed people here. None of them
killed as many as died in the Punjab between August and November 1947, and none of them had to be
invited.
That is the subject of this part, and it is why the last chapter of a book about wars is not about a battle.
Four things to hold on to
- It was not inevitable. For roughly twenty years the Punjab was governed by a cross-communal coalition of Muslims, Hindus and Sikhs, elected, functioning, and built on the interests of cultivators against townsmen. Anybody who tells you the province was always going to tear itself apart has to explain the Unionist Party, and most accounts simply leave it out.
- The killing started before partition was announced. The massacres in the Rawalpindi districts happened in March 1947 — three months before the plan, five before the border.
- The border was drawn in five weeks by a man who had never been to India, and it was published two days after independence, so that the government that drew it would not have to police the result.
- And it did not end in 1947. There were wars on this soil in 1965 and 1971, the province was divided again in 1966, and there was a decade of killing in the 1980s and 1990s. This part covers those too, because the book promised every war and this is the last chance to keep the promise.
Two subjects here are inside living memory and are actively fought over in the politics of three countries. This book’s
method does not change for them, but it needs stating explicitly.
All three communities killed and all three were killed. There is no version of the Punjab in 1947 in which one
group is only a victim, and this part will not write one. It will name specific events, specific places and — where they
are known — specific people, and it will not attribute anything to a religion.
The violence will be stated and not described. The same rule this book used for 1606, 1704, 1762 and 1919 applies
here with more force, not less, because there are people alive whose families are in it.
And on 1984 and after, this book will report and not adjudicate. Those events are recent, the facts are contested,
inquiries are still argued about, and reasonable people who agree on the evidence disagree about the meaning. This
part sets out what is documented, notes where the figures and the interpretations diverge, and stops there. A
history written in 2020 has no business handing down a verdict on 1984.
2The Akali Morchas, 1920–1925 CERTAIN
A five-year campaign to take the gurdwaras back from hereditary custodians — fought almost entirely by people refusing to hit anybody — and it won.
When: 1920–1925.
Where: Nankana Sahib, Guru ka Bagh near Amritsar, and Jaito in Nabha state.
Who: The Shiromani Gurdwara Prabandhak Committee and the Shiromani Akali Dal, both formed in late 1920.
The evidence: Government records, court proceedings, contemporary press. CERTAIN.
What it was about
Over the nineteenth century, control of most major Sikh shrines had passed to hereditary managers — mahants —
who held the endowments as personal property, and who were, in many cases, backed by the British
administration because a compliant custodian was useful. The campaign was to transfer the shrines to
elected community control.
What happened
1. Nankana Sahib, 20 February 1921. The birthplace of Guru Nanak. A body of reformers entered the shrine
- and about a hundred and thirty of them were killed by men in the mahant’s employ, and their bodies burned. The mahant was tried and convicted. It is the event that turned the movement from a reform campaign into a political one.
- Guru ka Bagh, 1922. Akalis went in daily batches to cut wood on disputed land, having taken a vow not to resist. The police beat them, batch after batch, for weeks. Something like five thousand six hundred were arrested.
- Jaito, 1923–1925. A prolonged campaign in the princely state of Nabha, in which a column of five hundred was fired on in February 1924 and around twenty were killed. 4. And the government gave way. The Sikh Gurdwaras Act of 1925 transferred management of the shrines to an elected committee — the SGPC — which still holds them.
The Guru ka Bagh campaign was watched by Charles Freer Andrews, an English clergyman and friend of Gandhi and
Tagore, and his account of it did more damage to the government’s position than anything the Akalis said.
He described men walking forward in batches, in silence, to be beaten to the ground by police with staves, and not
raising a hand or a word — and then the next batch walking forward.
He wrote that he had never seen anything like it, that he found it almost unbearable to watch, and that what he was
looking at was a new kind of heroism.
The account was published, it was read in England, and the beatings stopped.
This matters for the whole shape of Part Sixteen. The Akali movement of the 1920s was the largest sustained non-
violent campaign in India before Gandhi’s salt march, it was run by a community with a considerable military
tradition, and it worked — by legislation, in five years, against the government that had ordered the firing at
Jallianwala Bagh three years earlier.
Some thirty thousand people went to prison. Around four hundred died. The gurdwaras changed hands.
- The objective was specific and legal. Control of named institutions under a statute — winnable, unlike a demand for independence.
- Non-violence was tactically correct. A community famous for fighting, declining to fight, was impossible to answer with force in public.
- It was well organised. The SGPC and the Akali Dal gave it a structure, a treasury and a leadership, which is what the morchas of previous centuries never had.
- And 1919 had made the government cautious about being seen to shoot Punjabis.
- It created a permanent religious-political machine. The SGPC controls large endowments and an electorate; the Akali Dal became a party. Punjabi politics after 1925 has a standing Sikh political institution in it, which it did not before.
- It sharpened the boundaries. A campaign to define Sikh institutions required defining who counted as a Sikh, at a moment when every community in the province was being counted and listed.
- And it aligned the Akalis with Congress, which mattered enormously in 1947.
Put the two things the Punjab did in the 1920s side by side, because they happened in the same districts,
among the same families, at the same time.
At Guru ka Bagh in 1922, thousands of men walked forward in batches to be beaten unconscious without
raising a hand, on a vow, for months, until the government stopped.
At Lahore in 1928, three young men shot a police officer dead in the street in revenge for the death of a
nationalist leader, and one of them was hanged at twenty-three and became the most famous revolutionary
in modern India.
Both were reactions to the same thing. Both came out of the same province in the same decade. Both
worked, in their own terms — the Akalis got the Gurdwaras Act, and Bhagat Singh’s execution did more
to radicalise Indian politics than any single event of the 1930s.
This book has watched the Punjab argue with itself about method for a very long time. Part Nine
described a community that spent a century refusing to arm and then armed. Part Eleven
described a confederacy that survived by refusing battle. Part Twelve described one that could
not stop fighting itself.
The 1920s is the same argument, conducted between people who knew each other.
And it is worth noting which of the two the province chose to institutionalise. The SGPC and the Akali Dal
are still there. The Hindustan Socialist Republican Association lasted four years.
3Bhagat Singh, 1928–1931 CERTAIN
He was twenty-three, he was an atheist and a socialist, his organisation was deliberately non-communal — and in a province being pulled apart by religion he is the one figure everybody still claims.
When: 1928–1931.
Where: Lahore, and the Central Legislative Assembly at Delhi.
Who: Bhagat Singh, Sukhdev Thapar, Shivaram Rajguru and the Hindustan Socialist Republican Association.
The evidence: Trial records, his own writings, contemporary press. CERTAIN.
What happened
1. He was born in 1907, in the year of the canal colony agitation, into the family of Part Fifteen — his uncle Ajit
- Singh was deported to Mandalay that May.
- In October 1928 the Simon Commission came to Lahore and was met by a protest. Lala Lajpat Rai — the other leader of 1907, by then the outstanding nationalist of northern India — was baton-charged and died a few weeks later.
- In December 1928 Bhagat Singh and Rajguru shot a police officer, John Saunders, in the street at Lahore. They had intended to kill the superintendent who had ordered the charge and killed the wrong man.
- In April 1929 he and Batukeshwar Dutt threw two bombs into the Central Legislative Assembly at Delhi. The bombs were designed not to hurt anybody and did not. They stayed where they were, scattered leaflets and waited to be arrested — because the object was a trial, and a platform.
- The leaflet carried the line the whole thing is remembered by: that it takes a loud noise to make the deaf hear.
- In jail he led a hunger strike for political-prisoner status. Jatindranath Das died on the sixty-third day of it in September 1929. 7. He was tried by a special tribunal with no appeal and hanged at Lahore Central Jail on 23 March 1931, with Rajguru and Sukhdev. He was twenty-three.
Bhagat Singh is claimed today by a remarkable number of movements with incompatible programmes, and the simplest
corrective is to read what he wrote, because he wrote a great deal in two years in a condemned cell.
He was a socialist. The organisation was renamed the Hindustan Socialist Republican Association deliberately. He
argued that replacing British rulers with Indian ones would change nothing for a peasant or a millworker.
He was an atheist, and wrote an essay in 1930 called Why I Am an Atheist explaining it, in prison, awaiting execution,
when a profession of faith would have been the easiest thing in the world.
And he attacked communalism directly. He wrote about communal riots with contempt for everybody involved,
argued that religious leaders and politicians were manufacturing division for their own purposes, and held that the
cure was economic and educational rather than a better bargain between communities.
He was writing that in the Punjab, in 1930, seventeen years before the province divided.
He was also twenty-three, and he had shot a policeman by mistake. This book is not going to sand that off. He
believed political violence was justified and he acted on it, and the man he killed had done nothing.
Both things are in the record and both belong on the page.
- He wanted the trial and they gave it to him. The Assembly bomb was designed to produce a courtroom, and the special tribunal produced a martyr instead of a conviction.
- The timing was catastrophic. He was hanged days before the Karachi Congress, and the anger swamped everything else on the agenda.
- He was young, articulate and photogenic in a decade when Indian newspapers were reaching a mass readership for the first time.
- And they hanged him in Lahore, in the province that had already given the movement Jallianwala Bagh.
- The HSRA was tiny and was destroyed within a few years by arrests.
- Individual violence has no theory of the next step. He knew it; his prison writings say so.
- And the province went the other way. The politics that determined 1947 in the Punjab was communal arithmetic, and his was its explicit opposite.
Part Fifteen ended one section with a nineteen-year-old hanged at Lahore Central Jail in November 1915 for
the Ghadar conspiracy.
Kartar Singh Sarabha’s photograph was in Bhagat Singh’s pocket, and he is recorded as having
carried it for years and as having named him as the person he most wanted to be like.
Sixteen years later Bhagat Singh was hanged in the same jail.
Now trace the whole line, because it is astonishingly short and it runs through about three districts.
1907: the canal colonies agitate; Ajit Singh is deported; his nephew is born. 1915: Sarabha is hanged at
nineteen. 1919: Jallianwala Bagh; a boy from Sunam is in Amritsar. 1928: Lajpat Rai, the other leader of
1907, dies after a baton charge. 1931: Bhagat Singh is hanged at twenty-three. 1940: the boy from Sunam
shoots Michael O’Dwyer in London and gives his name as Ram Mohammad Singh Azad.
Thirty-three years. One province. Several of the people involved knew each other or were related to
each other.
This book has said repeatedly that the Punjab is not a large place. What it produced in those three
decades was a political tradition that was young, secular, violent, extremely brave and almost
entirely unsuccessful — and which, sixteen years after Bhagat Singh’s execution, could do nothing at all
to stop what happened.
4The Unionists — The Twenty Years That Worked, 1923–1942 CERTAIN
The most important thing in this part is a political party that most histories of partition do not mention, because it is inconvenient for every version of the story.
The Punjab was governed across religion, by election, for about WHAT HAPPENED
twenty years
- Founded 1923; in office through the 1920s and 1930s; effectively finished by 1946.
- The Punjab Legislative Council and Assembly at Lahore.
- Mian Fazl-i-Husain, Chaudhry Chhotu Ram, Sir Sikandar Hayat Khan and Sir Khizar Hayat Tiwana.
- Assembly records, election returns, legislation. CERTAIN.
What the Unionist Party was
1. It was a party of agriculturalists, and it cut straight across religion. Its founders were a Muslim from
- Punjab, a Hindu Jat from Rohtak, and a Muslim from the Salt Range, and its base was the landholding and cultivating classes of every community.
- Its enemy was the town. The moneylender, the urban trader, the professional politician — and this is where it connects directly to Part Fifteen. The Unionist Party is the political expression of the “agricultural tribes” of the Land Alienation Act of 1900.
- It legislated for its base. Debt relief, restrictions on foreclosure, regulation of the grain markets, and reform of the moneylending laws — chiefly the work of Chhotu Ram, who is still remembered across the Haryana and eastern Punjab countryside for it.
- And it governed. It won the 1937 election under the new provincial constitution, and Sikandar Hayat Khan ran a Punjab ministry containing Muslims, Hindus and Sikhs for five years.
Because it destroys the central claim on which every simple account of 1947 rests: that Hindus, Muslims and Sikhs in
the Punjab could not live together and were always going to divide.
For twenty years they governed together, by election, in the most communally mixed province in India.
They did it not because they loved each other but because the Unionist coalition organised politics around a real
interest that cut across religion — land, debt, water and prices. A Muslim Rajput landholder in Jhelum and a Hindu
Jat in Rohtak and a Sikh Jat in Ludhiana had more in common with each other than any of them had with the traders
in Lahore.
That is not sentimental; it is the ordinary way politics works when the categories that are available to organise it are
economic.
Part Fifteen argued that the British converted community into legal categories with things attached. The
Unionist Party is the evidence that it did not have to be the only available category — and its collapse is the
evidence for how quickly the alternative disappears when the structure holding it up gives way.
Between 1942 and 1947 that structure gave way completely, and the next section is how.
- The interest was real. Debt and land are not symbolic questions and they affected every rural family in the province.
- Its leaders were serious men with a programme, and they delivered legislation people could see.
- The British backed it. A stable, loyal, rural, non-Congress ministry in the recruiting province was exactly what Delhi wanted.
- And the province was genuinely mixed, so no community could govern it alone. Roughly half Muslim, a third Hindu, an eighth Sikh — the arithmetic forced coalition.
- It rested on individuals. Fazl-i-Husain, Sikandar and Chhotu Ram were the party, and all three were dead by 1945.
- It had no answer to a national question. Unionism was about the Punjab’s fields; the argument after 1940 was about India’s constitution.
- It was a party of landlords as well as peasants, and its social base narrowed as the electorate widened.
- And it depended on the British being there. Once they were leaving, the only question was who would inherit — which is a communal question, and Unionism had nothing to say about it.
Political history is not usually decided by actuarial accident, and the Punjab between 1936 and 1945 is the
exception that this book will offer.
Mian Fazl-i-Husain, the founder of the Unionist Party and the most capable Muslim politician in
northern India, died in July 1936, aged fifty-nine — months before the first election under the new
constitution.
Sir Sikandar Hayat Khan, who won that election and ran a cross-communal Punjab ministry for five
years, and who had the standing to negotiate with Jinnah as an equal rather than a subordinate, died on 26
December 1942, of a heart attack, aged fifty.
Chaudhry Chhotu Ram, the Hindu Jat who was the other pillar of the coalition and the most popular
rural politician in the province, died in January 1945, aged sixty-three.
Nine years, three deaths, all natural, and the Unionist Party never recovered from any of them.
What was left was Khizar Hayat Tiwana — an honourable man without his predecessors’ authority —
holding a coalition together against a national movement with money, organisation and a simple message,
in a province where every institution had been teaching people to think of themselves in communal
categories for eighty years.
This book does not deal in counterfactuals and it is going to break the rule once. A Punjab in which
Sikandar Hayat Khan was alive and in office in 1946 is not obviously a Punjab that divides the way this one
did.
He died at fifty, in his bed, of a heart attack, four and a half years before it happened.
5The Collapse, 1942–March 1947 CERTAIN
In four years the Punjab went from a working cross-communal government to no government at all, and the sequence is worth following closely because each step is small.
When: December 1942 to 5 March 1947.
Where: The Punjab Assembly at Lahore.
Who: Khizar Hayat Tiwana, the Muslim League, Congress and the Akalis.
The evidence: Election returns, Assembly records, Viceregal correspondence. CERTAIN.
The sequence
1. March 1940: the Lahore Resolution. The Muslim League, meeting at Lahore, resolved that the areas in which
- Muslims were numerically in a majority should be grouped into independent states. It was passed in the Punjab, which was the province it would divide.
- December 1942: Sikandar dies and Khizar Hayat Tiwana becomes Premier.
- 1944: Jinnah expels Khizar from the Muslim League for refusing to convert the Unionist coalition into a League ministry. The League’s argument was that a Muslim Premier could not govern in coalition with Hindus and Sikhs; Khizar’s was that in the Punjab there was no other way to govern.
- February 1946: the election. The Muslim League won seventy-three seats and became the largest single party. Congress took fifty-one, the Akalis twenty-two, the Unionists were reduced to about twenty. Nobody had a majority.
- Khizar formed a coalition of Unionists, Congress and Akalis and governed with it.
- January 1947: the League launched a civil disobedience campaign against that ministry — mass agitation, arrests, and pressure that the government could not sustain.
- 2 March 1947: Khizar resigned. The League could not command a majority either. On 5 March the Governor took the province into his own hands.
- And on 3 and 4 March there were demonstrations at Lahore. Outside the Assembly, Master Tara Singh, the leading Akali politician, is reported to have drawn his kirpan and denounced Pakistan from the steps.
The incident on 3 March 1947 is cited constantly and it is worth being careful about it, because it is used to assign
blame for everything that followed.
What is reported: that Master Tara Singh, at the head of a Sikh demonstration outside the Punjab Assembly, drew his
kirpan and made a speech against Pakistan, and that slogans were raised on both sides.
What is claimed: that this was the spark for the massacres in the Rawalpindi districts that began three days later.
This book’s view is that the claim does not survive contact with the geography and the timing. Rawalpindi is two
hundred and fifty miles from Lahore, the attacks began within days, and they were organised on a scale — across
four districts, against hundreds of villages — that is not assembled in seventy-two hours by a speech.
It is entirely possible that the demonstrations at Lahore were used as a justification. That is not the same as being a
cause.
What actually happened on 2 March was the disappearance of the only institution capable of restraining
anybody, in a province where every political actor now knew that a border was coming and that the land under their
feet was about to be allocated.
The killing began four days after the government stopped existing. That is the sequence, and it is the more
important fact.
- It had a simple proposition in a moment when everybody wanted to know what would happen when the British left, and Unionism had no answer to that question.
- The 1946 election was fought on national issues and the Unionists had spent twenty years being a provincial party.
- The war had transformed the province economically — inflation, requisitioning, and a much larger and more politicised electorate.
- And their opponents were dead. The three men who could have held the coalition together were all in their graves by January 1945.
- The Punjab had no majority community. Roughly half Muslim, a third Hindu, an eighth Sikh — which meant that any settlement would leave enormous minorities on both sides.
- The Sikhs were the smallest and had the most to lose. Their shrines, their canal colony lands and their population were on both sides of any line that could be drawn, and no arrangement gave them a province.
- The province had no ministry from 5 March 1947 onwards. Governor’s rule is an administration, not a government, and it has no political authority over anybody.
- And the police and services were themselves communally divided, which became fatal in August.
Sir Khizar Hayat Khan Tiwana is the last Premier of undivided Punjab and he is remembered by almost
nobody, which is unjust.
He was a Muslim landholder from the Salt Range, a Unionist, and he spent his premiership arguing a
position that turned out to be unsellable: that the Punjab was too mixed to be governed by any one
community, and that a coalition was not a compromise but the only arrangement that fitted the
facts.
For that he was expelled from the Muslim League by Jinnah in 1944, attacked as a traitor to his community,
subjected to a mass agitation in 1947, and forced out of office on 2 March.
He was not a great man and he was not a match for what he was up against. What he was is the last
politician in the Punjab whose programme was that the province should stay whole, and he was
beaten by a proposition that was clearer, better organised and better funded than his.
He left politics after partition and lived quietly, in Pakistan and then abroad, until 1975.
This book has recorded a great many defeats. The particular kind Khizar Tiwana suffered is worth
naming because it recurs everywhere: the moderate holding a complicated position in a moment when
everybody else has a simple one, who is destroyed not by an argument but by a schedule.
His ministry fell on 2 March 1947. The Punjab was divided one hundred and sixty-eight days later.
6Rawalpindi, March 1947 CERTAIN
The killing in the Punjab did not begin in August with the border. It began in March, in the north-western districts, three months before partition was announced.
When: 6–13 March 1947, principally.
Where: The Rawalpindi, Attock, Jhelum and Campbellpur districts — and separately at Multan.
Who: Organised bands attacking Sikh and Hindu villages, in districts with very small non-Muslim minorities.
The evidence: Government reports, contemporary press, and the refugee accounts collected afterwards.
What happened
1. Within days of the government’s collapse, attacks began on Sikh and Hindu villages across a wide area
- of the north-western Punjab — an area where those communities were a small minority scattered among a large Muslim majority.
- They were organised. The scale, the coordination between villages and the speed indicate planning, and the official reports at the time said so.
- The estimates of the dead run from about two thousand to seven thousand. Nobody counted.
- At the village of Thoa Khalsa, on 12 March, a large number of women died in a well rather than be taken. The figure usually given is around ninety. This book will state that and go no further.
- Tens of thousands fled east — the first refugee movement of the partition, five months before there was a border to cross.
- And the news travelled. Refugees arriving in the central and eastern Punjab in March and April 1947 brought accounts of what had happened, and those accounts are a direct cause of what was done in retaliation in the east from August onwards.
Most accounts of the Punjab in 1947 begin in August, with the announcement of the boundary, and treat the violence as
the consequence of the line. That is the wrong order and it matters.
The massacres in the Rawalpindi districts happened in March. The partition plan was announced on 3 June. The
boundary award was published on 17 August.
So the first large-scale organised killing in the Punjab took place before anybody knew where the border would
run, before Radcliffe had arrived in India, and before partition was government policy.
What it establishes is that the sequence was not border-then-violence. It was: government collapses in March;
organised violence follows within days; a refugee movement carries the news across the province through the spring;
and by August, when the line is finally published, every community in the Punjab has spent five months preparing.
This does not settle the question of who was responsible for 1947, which is not a question this book can settle. It does
dispose of the idea that the killing was a spontaneous reaction to a map.
And it disposes of one other thing. Retaliation in the east from August onwards was retaliation for something
specific that had already happened, which explains it and does not excuse a single one of the people who carried it
out.
- Isolated minorities. The Sikh and Hindu populations of those districts were small, scattered, and often prosperous — traders, moneylenders and landholders among a Muslim rural majority.
- No government. The ministry fell on 2 March and the attacks began on the 6th.
- The League’s agitation had mobilised the countryside for two months and the organisation existed.
- And the districts were full of demobilised soldiers — this was the heart of the recruiting country of Part Fifteen, and the Second World War had ended eighteen months earlier.
- The first refugee movement, which carried the news east and made the province’s Sikhs and Hindus certain that they could not remain in the west.
- Organisation on all sides. Bands, arms and plans were assembled across the province from the spring onwards.
- The Sikh political leadership’s position hardened from bargaining over a settlement to preparing for a division.
- And it removed any prospect of a peaceful transfer. By June there was nobody left in the province who believed one was possible.
There is a fact about the March massacres that is almost never stated and that this book has spent an entire
previous part setting up.
Rawalpindi, Attock, Jhelum and Campbellpur were the four most heavily recruited districts in the
British Indian Army.
Part Fifteen described the machinery: the martial-races doctrine, the recruiting handbooks, the land grants
tied to service, the district quotas. These are the districts it was pointed at. They supplied a very large
share of the Punjabi Muslim regiments that fought in France, Mesopotamia, North Africa, Italy and Burma
across two world wars.
The Second World War had ended in August 1945. By March 1947 those districts were full of men who
had been demobilised within the previous eighteen months — trained, experienced, organised, and
in many cases with weapons.
The same is true on the other side of the province. The Sikh Jat districts of the Majha and the Doaba had
been recruited just as hard, and the bands that operated in East Punjab from August onwards contained
large numbers of demobilised soldiers too.
The British had spent ninety years turning the Punjab into the most militarised society in Asia,
and then left it in 1947 with no government, a border it did not know about, and about half a
million recently discharged veterans in it.
Part Fifteen’s argument was that the province was made into an army and then shot by it. This is the
same sentence with the other ending: it was made into an army, and in 1947 it was turned loose
on itself.
7The Line — Radcliffe, July–August 1947 CERTAIN
A barrister who had never been east of Europe was given five weeks, out-of-date maps and no time in the districts, and told to divide the most integrated province in India.
When: Arrived 8 July 1947; award signed 12 August; published 17 August — two days after independence.
Where: Delhi, in a bungalow, from maps and census returns.
Who: Sir Cyril Radcliffe, chairman of the Punjab and Bengal Boundary Commissions.
The evidence: The award itself, the commission papers, and the correspondence. CERTAIN.
How it was done
1. He had never been to India. He was a distinguished London barrister, chosen precisely because he knew nothing about the subcontinent and therefore could not be accused of favouring anybody.
2. He arrived on 8 July 1947 and the boundary had to be ready for independence on 15 August.
3. He chaired two commissions, Punjab and Bengal, each with four Indian judges — two nominated by the
- Muslim League, two by Congress. Every commission deadlocked on every question, which meant that in practice Radcliffe decided everything personally.
- His terms of reference were to divide on the basis of contiguous majority areas of Muslims and non-Muslims, “taking into account other factors” — a phrase nobody defined and which he was left to interpret.
- He worked from census returns of 1941 and maps, in Delhi, in the hot weather. He made almost no visits to the ground.
- The award was signed on 12 August.
- And Mountbatten held it back. It was not published until 17 August, two days after both countries had become independent — so that the celebrations would not be spoiled, and so that the British would not be the government in charge when the consequences arrived. 8. Radcliffe burned his working papers, refused his fee, and left India. He never returned.
The Punjab in 1941 was roughly half Muslim, a third Hindu and an eighth Sikh, and there was no line anybody could
draw that would have separated them, because they lived in the same villages.
Lahore went to Pakistan. It had been the capital of the Punjab since Part Five of this book, it was about a third Hindu
and Sikh in 1941, and within a year it had almost none.
Amritsar went to India — twenty-five miles away, with the Harmandir Sahib in it, and it lost almost its entire
Muslim population.
The Sikhs were split down the middle by any line. Nankana Sahib, the birthplace of Guru Nanak, is in Pakistan.
Kartarpur, where he died, is in Pakistan. Panja Sahib is in Pakistan. Amritsar and Anandpur are in India. There was no
boundary that could have kept the community’s population and its shrines on the same side, and everybody
involved knew it in advance.
Gurdaspur district is the most argued-over decision. Overall it had a small Muslim majority; Radcliffe awarded
most of it to India, which — among other things — gave India its only road access to Kashmir. Whether that was the
reason, or a consequence, has been disputed for seventy-five years and cannot be settled from the surviving papers.
And the canal system was cut. Part Fifteen described the largest irrigated tract on earth, built as one system. The line
ran through it, leaving headworks on one side and the fields they watered on the other, which produced a water dispute
that took until 1960 to settle by treaty and is not settled now.
- Speed was the policy. Mountbatten had brought the date of independence forward by ten months, and everything else was fitted to it.
- An outsider was thought to be impartial, and the commissions’ deadlock was foreseen — which meant an outsider deciding was the design, not an accident.
- Nobody wanted to own the decision. Congress and the League both preferred an unaccountable arbitrator to a compromise either would have to defend.
- And the delay in publication was deliberate and is admitted in the papers.
- Nobody knew which country they were in on Independence Day. Districts celebrated the wrong flag; officials had no orders; families made decisions in the dark.
- The administration could not prepare because the administrators did not know the line either.
- The delay handed the initiative to whoever was organised, and in August 1947 that meant armed bands.
- And the line was drawn from a census six years old in a province that had been moving people west for fifty years.
In 1966, nineteen years afterwards, W. H. Auden wrote a short poem about Cyril Radcliffe called Partition.
It is one of the coldest things in English about the twentieth century, and it is the only piece of literature
this book will recommend.
It describes a lawyer sent out from London to divide a country he had never seen, given a few weeks and no
time to inspect the ground; shut in a lonely mansion with police at the gate; working from maps and out-
of-date census returns because contact with either side would have compromised him. The poem’s phrase
for his qualification is that he was unbiased — followed immediately by “alas.”
It ends with him going back to England as fast as he could, because — Auden writes — he did not want to
be shot at, and because he had already forgotten the case, as a good lawyer must.
The poem is unfair to Radcliffe personally and completely fair about the thing that was done. He
did not want the job, he took it because he was asked, and the papers suggest he understood better than
anybody that it could not be done properly in the time.
He refused his fee. He destroyed his notes so that nobody could reconstruct his reasoning. And he never
spoke publicly about it in any detail for the rest of his life.
Auden’s last line is the accusation that matters, and it is not aimed at Radcliffe. It is aimed at the
arrangement that put a man in that room in the first place — a government leaving in a hurry, two
movements that wanted a decision they did not have to defend, and a deadline that everybody agreed was
more important than the ground.
Between ten and fifteen million people moved because of what came out of that bungalow, and nobody
knows how many died.
8The Boundary Force, August 1947 CERTAIN
Fifty thousand troops were given twelve districts, fourteen million people and one month. It is the most damning administrative fact in this book.
When: Formed 1 August 1947; disbanded 1 September 1947.
Where: Twelve districts of the central Punjab, on both sides of the coming line.
Who: Major-General T. W. Rees, with Indian and Pakistani officers as advisers.
The evidence: Its own war diary and the official reports.
The arithmetic
1. The area assigned to it was about seventeen thousand nine hundred square miles.
2. The population in that area was about fourteen and a half million people.
3. The force was nominally around fifty thousand men, of whom the effective fighting strength was
- considerably less.
- That is roughly one soldier for every three hundred people, spread across an area the size of Denmark, in a province where organised bands were operating in every district.
- Its own commander said it was inadequate before it started.
- It was disbanded on 1 September, at the point when the killing and the migration were accelerating, because both new governments preferred to handle their own side.
There was a second difficulty with the Boundary Force and it is the more important one.
It was made up of the units of the old Indian Army, which were themselves being divided. Every soldier in it
knew which country he was going to, and in many cases knew that his own family was on the wrong side of the line
and moving.
The official reports record what followed: units that could not be relied on in particular districts; officers who
had to be moved; and, in a number of documented cases, soldiers who took part in the violence rather than
preventing it. Not most of them, and not everywhere — but enough that commanders had to plan around it.
This is not a scandal about the Indian Army. It is an unavoidable consequence of asking an army that is being
split by religion to police a province being split by religion, in the same eight weeks. No force in the world
would have held together under it.
And it produced the single fact that this section exists to state. In August 1947, the largest forced migration in
human history took place across a province in which there was no government, no police force anybody
trusted, no published border for the first two days, and one soldier for every three hundred people.
That was foreseeable in June and it was foreseen. It happened anyway because independence had been scheduled
for 15 August and everything else was fitted around the date.
- The numbers were never close to adequate and everybody involved said so at the time.
- It had no political authority behind it, because from 5 March the Punjab had no ministry and from 15 August it had two new governments with other priorities.
- Its own units were dividing.
- And it was given one month for a movement of people that took four.
- The date could have been later. Independence had been brought forward by ten months; the schedule was a choice.
- The award could have been published in July, giving administrators, police and families weeks rather than hours.
- The force could have been much larger, and troops from outside the Punjab could have been used.
- This book is not going to pretend any of that would certainly have prevented what happened. It states them because they were all proposed at the time and all rejected.
The Punjab Boundary Force existed from 1 August to 1 September 1947.
In the whole of this book, across five thousand years, this province has been garrisoned by Mauryan
viceroys, Kushan satraps, Ghaznavid governors, Delhi’s frontier commands, Mughal subahdars, Sikh sardars
and British district officers. Some of them were competent and most of them were adequate to keep
a road open and a market safe.
In August 1947, at the single most dangerous moment in the province’s history, it had one composite force
under a British general for thirty-one days, and then nothing.
Part Four of this book opened a complaint that ran for eleven parts: nobody garrisons the door. Part
Thirteen answered it, once, at Jamrud in 1836.
This is the last entry in the same ledger and it is the worst. The door was not the problem in 1947.
There was no invasion; nothing came down the passes; no foreign army crossed the Indus.
The province was left with fourteen million frightened people, no government, no published border and no
adequate force — and the thing it needed protecting from was itself.
9The Crossing, August–November 1947 CERTAIN
Between ten and fifteen million people moved. Nobody knows how many died. It is the largest forced migration in recorded history and most of it happened here, in about ten weeks.
When: August to November 1947, at its worst in September.
Where: Both Punjabs, and the roads and railways between them.
Who: The population of the province.
The numbers: Roughly 4.5 million Hindus and Sikhs moving east; roughly 6 million Muslims moving west. Deaths: estimates range from two hundred thousand to two million.
What the movement was
1. It was not planned and it was not expected. Both governments and both political movements had assumed that minorities would stay where they were. There was no arrangement for a transfer of population because
- nobody had intended one.
- It went in both directions at once, along the same roads, in columns that passed each other.
- The kafilas — the foot columns — were enormous. The largest recorded was around four hundred thousand people and was some tens of miles long; it took days to pass a given point.
- The trains were the worst of it. Refugee specials were stopped and attacked on both sides, and trains arrived at Lahore and at Amritsar with everybody aboard dead. This book will state that and not elaborate.
- The monsoon was on, the rivers were high, cholera broke out in the camps, and a very large number of the dead died of disease and exhaustion rather than violence.
- And by November it was substantially over. Lahore, which had been about a third Hindu and Sikh, had almost none. Amritsar, Jalandhar and Ludhiana had lost almost their entire Muslim populations. Two of the most mixed cities in Asia became mono-religious inside four months.
This book has interrogated casualty figures in every part since the first, and this is the one that cannot be interrogated,
because there is nothing to interrogate.
The contemporary official figure was around two hundred thousand. It was produced by two new governments
with every reason to minimise, in a province where the administration had collapsed.
Figures of one million and two million are widely cited. They are extrapolations, not counts.
The most careful modern work, using census comparisons to identify people who are missing rather than
counting bodies, generally arrives somewhere between five hundred thousand and one million, with the
Punjab accounting for the overwhelming majority.
The reason the range is so wide is worth stating plainly. There were no registrars, no police reports, no hospital
records and no bodies to count. People died in fields, in canals, on trains, in columns, and in villages that ceased to
exist. Whole families disappeared, so there was nobody to report them missing. The census that would have caught it
was not taken until 1951.
What can be said with confidence is the shape: that it was in the hundreds of thousands at least, that it happened
in about ten weeks, that it was concentrated in one province, and that all three communities were both perpetrators
and victims on a very large scale.
- Nobody had planned for a migration because nobody had intended one, so there were no camps, no transport, no rations and no route management.
- There was no government in the province from March and no adequate force in August.
- Both populations were moving simultaneously in opposite directions, which meant that every column met the evidence of what had happened to the other side.
- And retaliation was the organising logic. March in the west produced August in the east, which produced September in the west, and each round was justified by the last.
- The mixed Punjab. A province that had been shared for a thousand years became two provinces that were not.
- The canal colonies. Part Fifteen’s great construction, settled from the east for fifty years, emptied of its eastern settlers in three months.
- The shrines. Nankana Sahib, Kartarpur and Panja Sahib on one side; Amritsar and Anandpur on the other.
- And a very large number of people who are not counted anywhere, which is why this section exists.
Everything in this section is about what people did to each other, and it would be a false history that
stopped there, so here is the other half of the record.
The best-documented instances of rescue in 1947 are all of the same kind: individuals hiding their
neighbours, escorting them to a column, standing in a doorway, or driving them to a border they could have
delivered them to a mob instead.
They are recorded in the thousands, in every district, in all three directions. Muslim families hid Sikhs. Sikh
villages escorted their Muslim neighbours to the camps. Hindu shopkeepers hid Muslim employees.
Almost none of these people are named anywhere, because they did it once, for people they knew,
and then had the rest of their lives to get on with.
And there is one place that has to be mentioned, for the third time in this book.
Malerkotla came through 1947 without the massacres that consumed every district around it. Its
Muslim population was not driven out and did not leave; it is still there. The town is now a Muslim-
majority town in Indian Punjab, which nowhere else in the eastern Punjab is.
The explanation offered locally, then and since, is a protest made in a courtroom at Sirhind in
December 1704 by a nawab who objected to the execution of two children — the haah da naara of Part
Eleven — and the promise that the Sikhs afterwards kept for two hundred and forty-three years.
Historians point out, correctly, that Malerkotla also had specific local factors: an active administration,
leaders on all sides who cooperated, and no large refugee influx.
Both explanations can be true. And in a province where several hundred thousand people were killed by
their neighbours in ten weeks, one town where it did not happen is worth more attention than it usually
gets — because it establishes that it was possible.
10The Women CERTAIN
Somewhere near a hundred thousand women were abducted in the Punjab in 1947, and what the two governments did about it afterwards was almost as bad.
When: 1947, and the recovery operations from 1947 to 1956.
Where: Both Punjabs.
Who: Women of every community, and the two governments that afterwards decided what should happen to them.
The evidence: The recovery operation records, the legislation, and the oral histories collected from the 1980s onwards.
What is documented
1. The commonly cited figure is between seventy-five thousand and a hundred thousand women
- abducted across the partitioned territories, the great majority in the Punjab. It is an estimate.
- It happened in all directions. The recovery operations that followed were set up to recover Muslim women from India and Hindu and Sikh women from Pakistan, which is itself the record of what had been done on both sides.
- Women were also killed by their own families and communities to prevent their capture, and in some cases killed themselves. Thoa Khalsa in March 1947 is the best known instance and it was not the only one.
- This book will state those facts and will not describe any of it. The reasons are the same ones it gave in Part Nine and Part Fifteen, and they apply here with more force than anywhere else in these sixteen parts.
What happened next is less known and it needs to be on the record.
In December 1947 the two governments agreed to recover and return abducted women, and India passed the
Abducted Persons (Recovery and Restoration) Act in 1949 to enforce it. Social workers and police crossed the
border for years, locating women and bringing them back.
The operation continued into the 1950s and recovered tens of thousands of women. India returned something
over twenty thousand Muslim women; Pakistan returned something over eight thousand Hindu and Sikh women.
And it was carried out regardless of what the women themselves wanted.
By the time many were located, years had passed. Some had been married, had children, and had made lives. Under the
Act a woman’s own statement of her wishes had no legal force — she could be removed and repatriated by the
state, and in many cases her children, born in the country she was taken from, could not go with her.
Some were then rejected by the families they were returned to, on grounds of honour, and ended their lives in
government ashrams.
Two governments that had failed to protect these women in 1947 spent the next decade moving them across
the border a second time, without asking them, in the name of national honour — and the record of it exists
because the officials involved kept files.
- Shame, imposed on the victims rather than on anybody else, in every community.
- National narratives had no place for it. Both new states were building founding stories, and this fitted neither.
- The families would not speak and in many cases had themselves been the ones who killed or rejected.
- And the history was written from government records, which record the recovery operation and not the women.
- Oral history, beginning in the 1980s. Researchers went and asked, forty years late, and found that a very large number of families had a story nobody had told.
- The work of Indian and Pakistani women historians in the 1990s established this as a subject at all.
- And the recovery operation’s own files, which document what was done to the women afterwards in the state’s own words.
Sixteen parts of this book have asked the same question about every source: who is writing, and what do
they need you to believe?
The answer has almost always been an outsider — a Greek four centuries late, a Persian secretary counting
for his employer, a court historian establishing a reputation, a British officer explaining a casualty return.
For 1947 the sources are different and the silence is the same. There are government reports, refugee
registrations, relief committee minutes, newspapers and a very large official literature about the transfer of
populations. There are memoirs by politicians and by administrators.
And for about forty years there was almost nothing at all about what happened to women, in a province
where somewhere near a hundred thousand of them were abducted — because the people who knew
were the families, and the families had decided not to say.
It was broken by historians, mostly women, mostly Indian and Pakistani, who did the obvious thing that
nobody had done: they went to the villages and asked the old people, and wrote down what they
were told.
What came back was not one story. It was tens of thousands of them, in every district, on both sides, and it
changed what partition means as a historical subject.
This book has complained since Part One that the Punjab did not keep its own record. Here is the
one case where the record existed the whole time, in the memory of ordinary people, and was not collected
for two generations because nobody with a notebook thought to knock on the door.
111965 — Asal Uttar and Chawinda CERTAIN
Eighteen years after the line was drawn, two armies fought the largest tank battles since the Second World War across it — on ground this book has been describing since Part One.
When: 6–23 September 1965.
Where: Khem Karan and Asal Uttar; the Ichhogil canal outside Lahore; and Chawinda in the Sialkot sector.
Who: The Indian and Pakistani armies — both descended from the same army, with regiments that had shared a mess in 1946.
The evidence: Official histories on both sides, which disagree about who won and agree about almost everything else.
What happened
1. The war began in Kashmir and India widened it by advancing across the international border towards Lahore on 6 September.
2. Indian troops reached the Ichhogil canal — the water obstacle covering Lahore, a few miles from the city
- — and could not get past it.
- Pakistan counter-attacked at Khem Karan with its 1st Armoured Division, equipped with American Patton tanks, aiming to cut the road to Amritsar.
- At Asal Uttar, from 8 to 10 September, that attack was destroyed. The defenders flooded the sugarcane fields, let the armour come into a horseshoe of prepared positions, and shot it apart at close range. Something approaching a hundred tanks were destroyed or captured. The site was nicknamed Patton Nagar — the town of Pattons.
- Company Quartermaster Havildar Abdul Hamid of the 4th Grenadiers destroyed several tanks with a jeep-mounted recoilless rifle before being killed. He was awarded the Param Vir Chakra.
- At Chawinda, in the Sialkot sector, the Indian armoured advance was fought to a standstill in what was, by the number of tanks engaged, one of the largest armoured battles since 1945.
- A ceasefire came into force on 23 September and the Tashkent Declaration followed in January 1966.
Look at where these battles were fought, because a reader of this book will recognise every one of them.
Khem Karan and Asal Uttar are in the Bari Doab between the Beas and the Sutlej — the country where Guru
Hargobind fought in Part Nine and where the misls divided the districts in Part Twelve.
The Ichhogil canal is a defensive line dug in the 1950s to cover Lahore from the east, on exactly the approach used by
every army that has ever taken the city — Mahmud, Ghor, Timur’s officers, Babur, Nadir Shah, Abdali and Ranjit Singh.
Chawinda is in the Sialkot country, in the Rechna Doab, thirty miles from the Chenab.
And the armies that fought there were, eighteen years earlier, one army — the Indian Army of Part Fifteen, recruited
from these districts under the martial-races policy, split in August 1947 along with everything else.
A great many of the men in both armies in September 1965 were Punjabis. Some of the regiments on both sides
had fought alongside each other in North Africa and Italy twenty years before.
Part One of this book began with a battle on the Ravi described in a hymn. The last conventional battles fought on
Punjabi soil were fought two thousand years later, over the same doabs, by two halves of the same army.
- Prepared ground. Flooded fields, sugarcane for concealment, and a horseshoe of positions — the same principle as an entrenched camp at Panipat.
- Armour without infantry in close country is at the mercy of anybody who can get within a hundred yards.
- The defenders let them come. The oldest winning move in this book: make the other side attack a position you chose.
- Neither side could break through in the Punjab, which is flat, irrigated, canal-cut country that heavily favours the defence.
- Both were fighting on external supply and both ran short.
- International pressure ended it before either could force a decision.
- And the underlying question was in Kashmir, where nothing changed at all.
The Ichhogil canal — properly the BRB, the Bambanwala–Ravi–Bedian canal — runs north to south a few
miles east of Lahore.
It was dug in the 1950s and it is an irrigation canal, and it is also a tank obstacle. It was designed as both,
deliberately, by a state that knew exactly which direction an army would come from because it
had five thousand years of evidence.
In September 1965 it did its job. Indian armour reached it and could not cross in strength, and Lahore was
not taken.
Now stand back and look at what that means in the context of this book.
Part Fifteen described the canal system as the one thing in these sixteen parts that was built
rather than destroyed — the largest irrigated tract on earth, made to grow wheat.
Eighteen years after partition, a section of it had been re-engineered as a fortification, because the province
had been divided and the two halves were at war.
The Punjab’s rivers have been its defining feature since the first page of this book. Alexander
crossed the Jhelum in a storm. Babur crossed the Indus at Attock. The Khalsa army died against the Sutlej
at Sobraon. A family was cut in half by the Sarsa in flood in 1704.
And in 1965 the water that had been brought to the western Punjab to end its famines was standing in a
ditch, stopping tanks.
121966 and 1971 — The Province Divided Again CERTAIN
Nineteen years after being cut in half, the Indian Punjab was cut into three — and five years after that, the armies were back on the Shakargarh plain.
When: Reorganisation on 1 November 1966; the war 3–16 December 1971.
Where: The Indian Punjab; and the Shakargarh salient on the Basantar river.
Who: The Government of India; and then the Indian and Pakistani armies again.
The evidence: The Punjab Reorganisation Act, and the official histories of the war. CERTAIN.
The second division
1. From the 1950s there was a movement for a Punjabi-speaking state — the Punjabi Suba — within India,
- argued for on linguistic grounds and opposed on the ground that it was a religious demand in linguistic clothing.
- The Punjab Reorganisation Act took effect on 1 November 1966. The Hindi-speaking southern districts became the new state of Haryana; the hill districts went to Himachal Pradesh; and Chandigarh, a capital designed and built after 1947 to replace Lahore, became a Union Territory serving as the capital of both new states.
- The arithmetic is worth stating. The Punjab of 1946 was around a quarter of a million square kilometres. East Punjab in 1947 was about a hundred thousand. After 1966 the state of Punjab was about fifty thousand — roughly a fifth of what the province had been twenty years earlier.
- And the settlement left disputes running. Chandigarh’s status and the division of river waters were not resolved, and are not resolved now.
1971
1. The war of December 1971 was decided in the east, in what became Bangladesh, and the western front was a holding operation.
2. The heaviest fighting on the western front was in the Shakargarh salient, on the Punjab–Jammu border
- — armoured and infantry battles across the Basantar river in mid-December.
- Second Lieutenant Arun Khetarpal of the 17 Poona Horse, aged twenty-one, was killed at Basantar after refusing to withdraw a damaged tank and destroying several of the enemy’s. He was awarded the Param Vir Chakra posthumously.
- The war ended on 16 December with the surrender in Dhaka, and the Simla Agreement followed in July 1972.
It is worth pausing on what has happened to this province as a unit of territory, because it is unusual.
The Punjab described in Parts One to Fourteen of this book ran from the Sutlej to the Indus and from Multan
to the hills. Ranjit Singh’s kingdom took in Kashmir, the Derajat, Hazara and Peshawar. The British province of 1901
ran from Delhi’s edge to the Afghan frontier.
In 1947 it was cut in two along a line drawn in five weeks.
In 1966 the Indian half was cut into three.
So the entity that this book has been following since the Rig Veda now exists as: a large province of Pakistan; a small
state of India; two other Indian states carved out of it; a shared capital that belongs to neither; and a diaspora.
The word “Punjab” means five rivers. The Indian state of that name has two of them running through it and a share
of a third. The other rivers are in Pakistan.
That is not a complaint and it is not a political position. It is a statement about what happened to a place, and this
book began by defining that place geographically, so it has to end by noting that the definition no longer
works.
- The decision was elsewhere. The Punjab front was fought to hold ground, not to win a war.
- Both sides had learned from 1965 about armour in canal country and fought accordingly.
- And it was short. Thirteen days, against twenty-two in 1965.
- Chandigarh, promised to Punjab repeatedly and still a Union Territory.
- The river waters, disputed between Punjab, Haryana and Rajasthan ever since.
- Punjabi-speaking areas left outside the state, which remained a grievance.
- And a settlement that satisfied nobody, which fed directly into the politics of the following decade.
When the Punjab was divided in 1947, the Indian half lost its capital. Lahore had been the capital of the
Punjab, on and off, since Part Five of this book — through Ghaznavids, Ghurids, Mughals, Sikhs and the
British, for the better part of a thousand years.
So India built a new one.
Chandigarh was laid out from nothing in the 1950s, on farmland at the foot of the Siwaliks, and the
government hired Le Corbusier to plan it. It is the only city in India designed as a whole by a major
modernist architect, it is laid out on a grid of numbered sectors, and it has a Capitol Complex of concrete
buildings that architectural historians take very seriously.
It was built to be the capital of the Indian Punjab, and it was completed in time for the state to be divided
again in 1966 — after which it became the capital of two states and belonged to neither.
Part Fifteen ended a section on Lyallpur by observing that the canal colonies were the one thing
in these sixteen parts that was built rather than destroyed.
Here is the second one, and it is the same story with the ending changed. A city built from nothing on
empty ground, by a government, to replace something that had been lost — and then argued over
from the day it opened.
Le Corbusier’s plan included a monument that was never finished, meant to commemorate the partition.
The city is full of roundabouts and rose gardens and works extremely well, and the thing it was
built because of has no memorial in it.
131984 and After CERTAIN
This section reports and does not judge, and it will say why.
When: June 1984; October–November 1984; and roughly 1984 to 1995.
Where: Amritsar, Delhi and other cities, and the Punjab countryside.
Who: The Indian state, armed militant groups, and — as always — mostly civilians.
The evidence: Government white papers, commissions of inquiry, court records, human rights investigations and journalism. Abundant, and contested.
What is documented
1. From the late 1970s, a political dispute over the unresolved questions of 1966 — Chandigarh, river waters, the terms of the state’s autonomy — combined with a religious revivalist movement and, from about
- 1981, an armed campaign.
- In June 1984 the Indian Army carried out Operation Blue Star, an assault on the Harmandir Sahib complex at Amritsar to remove an armed group that had fortified it. The Akal Takht was very heavily damaged. The operation coincided with a martyrdom anniversary and the complex was full of pilgrims. Official figures give several hundred civilian and militant dead and eighty-three soldiers; other estimates are considerably higher.
- On 31 October 1984 the Prime Minister, Indira Gandhi, was assassinated by two of her own bodyguards.
- Over the following days, organised attacks on Sikhs took place in Delhi and other cities. Official figures put the Delhi dead at around two thousand seven hundred and the national total at over three thousand; other estimates are higher. Multiple inquiries over three decades established that the violence was organised rather than spontaneous. Convictions have been very few.
- The armed campaign in the Punjab ran until about 1995. Official figures for the whole period give upwards of twenty thousand dead — militants, security personnel and, in the large majority, civilians. 6. Civilians were killed by militants, including in bus and market attacks and in targeted killings of Hindus, of Sikhs regarded as opponents, and of journalists.
- Civilians were also killed by the state. Extrajudicial killings and disappearances during the counter-insurgency have been documented by Indian courts, the National Human Rights Commission and independent investigators, including the inquiry into unidentified cremations in the Amritsar districts.
- It ended in the mid-1990s. The state’s Chief Minister was assassinated in 1995; the militancy collapsed; and the Punjab has been at peace since.
Every other section in these sixteen parts ends with a judgement. This one does not, and the reasons should be stated
rather than left as a gap.
It is recent. People who were in the Golden Temple in June 1984, and people whose families were killed in Delhi in
November 1984, and people who lost relatives on both sides of the counter-insurgency, are alive. A book written this
close to events is not writing history; it is joining an argument.
The facts are genuinely contested. Casualty figures for Blue Star differ by a factor of several between the
government’s account and others. The Delhi killings have been investigated at least ten times with different findings.
The counter-insurgency records are incomplete by design.
And it is live politics in a country of a billion and a half people. This book has been willing to say hard things
about Ashoka, Timur, Aurangzeb, Ranjit Singh and the British government, all of whom are dead. It is not going to use
the authority it has built over fifteen parts to hand down a verdict on a dispute that is still being litigated in courts and
fought over in elections.
What it will say is the thing that is not in dispute. Between 1984 and 1995, in a state of about twenty million people,
somewhere upwards of twenty thousand were killed. The great majority were civilians. They were killed by armed
groups and they were killed by the state. And the Punjab, which had already been divided twice in forty years, spent a
decade of the twentieth century burying its own again.
That is the account. The verdict belongs to somebody writing in fifty years, with the files open.
- The political questions of 1966 were left unresolved for twenty years and that mattered.
- The violence was overwhelmingly borne by civilians on all sides.
- The Delhi killings of November 1984 were organised. Successive inquiries have found so.
- And accountability has been very limited, for the state and for the militants alike.
- The casualty figures, at almost every point.
- Whether Blue Star was avoidable and whether alternatives were seriously attempted.
- The scale and authorisation of extrajudicial killings during the counter-insurgency.
- And the causes, about which serious historians disagree completely.
There is one thing in the record of 1984 that belongs unambiguously in this book, because this book has
described it three times already.
In April 1757, Afghan troops under Jahan Khan demolished the Harmandir Sahib and filled the
tank. Part Eleven.
In April 1762, Ahmad Shah Abdali blew it up again and filled the tank again. Part Eleven.
In December 1764, thirty men who would not leave the Akal Bunga were killed there, and a hostile
qazi riding with the invading army wrote down what he thought of them. Part Eleven.
In June 1984 the Akal Takht was very heavily damaged in a military operation, and was rebuilt.
Part Eleven made an argument about the first three, and it is the same argument now. The reason the
shrine has been attacked so often is that it is the one fixed address of a community that has
otherwise been extremely difficult to find — and the reason the attacks have never accomplished what
they were for is that the building is where people meet, not why.
The tank was cleared and refilled in 1757 by people with a price on their heads. It was cleared and refilled
again in 1762 by the survivors of the worst day in their history. The Akal Takht was rebuilt after 1984, twice
— once by the government and then again by the community, which demolished the government’s version
and did it itself.
Four times in two hundred and thirty years, and the water is still in the same place.
14The Last Chapter CERTAIN
Sixteen parts, about five thousand years, and it is time to add it up.
The accounting
Start with the plainest number in the whole book.
In the last one thousand one hundred and fifty years, the Punjab has governed itself for about eighty-
four of them.
The Hindu Shahis lost Kabul in 870 and the plain within a century and a half. From then until 1765 the province
was administered from Ghazna, Ghor, Delhi, Agra, Kandahar, Kabul, Isfahan and Lahore-under-orders — eight
hundred and ninety-five years. Then the Dal Khalsa took Lahore in 1765 and struck coin there, and for eighty-four
years the Punjab was ruled by people born in it. In 1849 it became a British province. In 1947 it became two
provinces of two other countries.
Eighty-four years out of eleven hundred and fifty. That is the arithmetic this book has been walking towards
since Part Five, and it is the single most important fact in it.
Why
Not because of any quality of the people who live here, and this book has resisted that explanation in every part. It
is geography, and it is not complicated.
Everything that has ever entered India by land has come through four or five passes in the north-west, and
everything that has come through those passes has arrived on this plain. The Punjab is the ground between the
door and the house. It is rich, it is flat, it is impossible to defend, and it lies across the only road.
Nothing else in these sixteen parts explains as much.
Coalitions lose. Ten kings on the Ravi. The Malavas and Kshudrakas, who combined and could not agree on a
commander. Nine republics submitting one at a time. Anandapala’s confederacy, dissolved in ten minutes. Prithviraj at
Tarain. The hill rajas, changing sides every three years. Small powers here do not federate; they bargain
individually and are eaten in sequence. The one time it was solved — the Sarbat Khalsa between 1748 and 1767 — it
was solved by an enemy, and it came apart when he stopped coming.
Armies die with their commanders. Porus’s elephants. Anandapala’s elephant. Ghor at Tarain. Qutlugh Khwaja at
Kili. Hemu’s arrow at Panipat. Wazir Khan at Chappar Chiri. Vishwasrao at Panipat again. Seven armies, one failure
mode — authority living in one visible body. The armies that survive it are institutions.
Blockades work and assaults do not. Multan’s water in 712, Mankot in 1557, Anandpur in 1704, Gurdas Nangal in
1715, Panipat in 1760. A besieger eats from a province and the besieged eats from a store, and the store always runs out
first.
The hills preserve you and imprison you. You cannot be finished there and you cannot rule from there.
Somebody always opens the gate. Ambhi invited Alexander. Daulat Khan invited Babur. Both lost everything. The one
time the invitation went to a local — Lahore in 1799 — it produced the only independent Punjabi state in nine
centuries.
And the winner is whoever gets written about. Which is the rule this book cares about most.
The record
For eleven of these sixteen parts, the Punjab does not speak.
Its history reaches us through Greeks writing four centuries late, Chinese pilgrims describing ruins, Persian
secretaries counting the dead for their employers, victory pillars carved by winners in distant provinces, a
Kashmiri poet six hundred years after the event, and Delhi court historians who were paid by the month.
Then, in Part Seven, somebody in the Punjab watches an army sack a town and writes it down in
Punjabi. In 1604 it is put in a book, with an eyewitness account of an invasion in it, and the book is installed in a
shrine and never lost. In 1688 a commander writes an account of his own battle. In 1705 a man who has lost
everything writes to the emperor who took it, in the emperor’s own language, and tells him he has lost.
And after that it never stops. A couplet about a governor and a sickle. Two men who set up a toll post on the Grand
Trunk Road specifically so that the government would have to admit they existed. A three-word refrain from a
meeting at Lyallpur that is still sung at farmers’ demonstrations.
Almost every word the ordinary people of this province left behind rhymes, is short, was made by
somebody who owned nothing, and is about not being pushed off the land.
Guru Nanak died in 1539 at Kartarpur, a village he had founded on the bank of the Ravi, where he had
spent the last eighteen years of his life farming.
Part Seven described what he had seen: an army coming down the road in 1521, a town sacked, and a man
writing about it in the language of the people it happened to — the first Punjabi account of any event in
these sixteen parts.
In 1947 the line ran between Kartarpur and the country most of his followers ended up in. The shrine is
about four and a half kilometres from the border.
For seventy-two years, pilgrims stood on a raised platform on the Indian side with binoculars, provided by
the border authorities, and looked at it.
In November 2019 the two governments opened a corridor. Indian pilgrims can now cross the border
to Kartarpur without a visa, spend the day, and come back. It required an agreement between two states
that agree on almost nothing, and it was built by both of them, and it works.
It is a small thing. It is four and a half kilometres, one shrine, a few thousand people a day, and it has been
suspended more than once when relations were bad.
But this book has spent sixteen parts describing a place defined by a line it did not draw — the
passes, the frontier, the Sutlej treaty of 1809, the Durand Line of 1893, and the line of 1947 that was drawn in
five weeks by a man who burned his notes and never came back.
Every one of those lines was drawn to keep people out, and every one of them was drawn by
somebody who did not live here.
In 2019, for the first time in the whole of this history, a border in the Punjab was made easier to cross rather
than harder — by the two governments that own it, on purpose, so that people could go and stand
in a field where a man had once written down what an invasion looked like.
That is where this book ends. Not with a victory, because the Punjab does not get many, and not with a
lesson, because five thousand years of this ought to have taught somebody something by now and evidently
has not.
It ends with a gate that opens.
15The Whole Chapter on One Page
Five things this book found
1. Geography did most of it. The Punjab is the ground between the passes and the plain of India. Everything
- that has ever come by land has come through here, and a rich flat province lying across the only road is going to be fought over for as long as there are people to fight. No explanation involving the character of Punjabis is needed and none has been offered.
- Eighty-four years out of eleven hundred and fifty. From the Shahis losing Kabul in 870 to today, this province has governed itself for the eighty-four years between the Dal Khalsa taking Lahore in 1765 and the annexation in 1849. Everything else was administered from somewhere else.
- The institution beats the man, and the Punjab worked this out and then forgot it. Seven armies in this book died because authority lived in one visible body. In 1699 somebody deliberately engineered that out — the order above the leader, the book above the order — and it survived a bounty on every head, two ghallugharas and the demolition of its shrine twice. Then in 1799 a maharaja put the single point of failure back, and in 1839 he died in his bed and it took ten years to lose everything.
- The categories were made and then cut along. Between 1857 and 1919, community in the Punjab was converted from something people were into legal boxes with land rights, recruitment quotas and electoral rolls attached. The Unionist Party proves it was not the only category available. Its collapse in the 1940s proves how fast the alternatives disappear.
- And the worst thing that ever happened here was not an invasion. Five thousand years of Persians, Greeks, Huns, Turks, Mongols, Afghans and Europeans, and the highest death toll in these sixteen parts belongs to about ten weeks in 1947 when nobody crossed the frontier at all.
That is sixteen parts, and the promise made on the cover of the first one was every war fought on the soil of Punjab, told
in plain words, from the very beginning.
It has not been kept perfectly. Wars were left out — some because there is no evidence, some because they were small,
and some because a book has to end. Whole centuries have gone past in a paragraph. Every casualty figure in it should
be read as an argument rather than a count.
And it has been written from a particular place. Somebody from the western Punjab would put the emphasis
elsewhere; somebody writing in Urdu would find different things worth a chapter; somebody from Multan or Peshawar
or Hisar would object to how much of it is about the doabs.
The one commitment this book has tried to keep throughout is the one it made in Part One and repeated in every part
since: to say who is writing, and what they needed you to believe.
So, finally, about the person writing this one. It was written by a Punjabi, for people who do not know this history
and have every right to, in the language most likely to reach them. It has been harder on the Punjab than most
Punjabis will like — about Banda’s sack of Samana, about Ranjit Singh and his mother-in-law, about 1857, about what
all three communities did in 1947 — and that was deliberate, because the alternative is a monument.
A history that only records a place at its best is not a history. It is an advertisement.
Thank you for reading it.
16Words Used in This Part
A campaign or agitation, usually sustained and organised. The Akali morchas of 1920–25 were mass non-violent campaigns for control of the gurdwaras.
The Shiromani Gurdwara Prabandhak Committee, formed in November 1920 and given statutory control of the historic Sikh shrines by the Sikh Gurdwaras Act of 1925. Still the most powerful Sikh institution.
A hereditary custodian of a shrine. The transfer of the gurdwaras from mahants to elected committees is what the morchas were about.
The cross-communal party of Punjabi agriculturalists founded in 1923, which governed the province for about twenty years — Muslims, Hindus and Sikhs together, on economic rather than religious lines.
The Muslim League resolution of March 1940 calling for the Muslim-majority areas to be grouped into independent states. Passed in Lahore, in the province it would divide.
Direct administration of a province by its governor when no ministry can be formed. Imposed on the Punjab on 5 March 1947 and never lifted before partition.
The boundary decision for Punjab and Bengal, signed 12 August 1947 and published on 17 August — two days after independence.
The composite military force of about 50,000 men responsible for 17,900 square miles and 14.5 million people from 1 August to 1 September 1947.
A refugee foot column. The largest recorded in 1947 held around four hundred thousand people.
The Indian legislation of 1949 under which women were recovered and returned across the border, regardless of their own stated wishes and often without their children.
The canal east of Lahore, dug in the 1950s to serve as both an irrigation work and a tank obstacle. It stopped the Indian advance in September 1965.
India’s highest award for valour. Awarded posthumously to Abdul Hamid for Asal Uttar in 1965 and to Arun Khetarpal for Basantar in 1971.
The movement for a Punjabi-speaking state within India, which produced the reorganisation of 1 November 1966.
The visa-free crossing opened in November 2019 to the shrine where Guru Nanak died in 1539, four and a half kilometres inside Pakistan.
Sources used for this part
- Government of India and Punjab records, 1919–1947, including the Boundary Commission papers and the Punjab Boundary Force
- war diary
- Punjab Legislative Assembly records and election returns, 1937 and 1946
- C. F. Andrews's published account of Guru ka Bagh, 1922
- Trial records of the Lahore Conspiracy Case (1929–31) and Bhagat Singh’s own writings, including Why I Am an Atheist
- The Transfer of Power documentary series and the Mountbatten and Viceregal papers
- Refugee registration records, relief committee reports, and the files of the Central Recovery Operation
- The oral history work on partition of Urvashi Butalia, Ritu Menon and Kamla Bhasin, and later collection projects on both
- sides of the border
- Official histories of the 1965 and 1971 wars, Indian and Pakistani
- Government white papers, commission of inquiry reports, court judgments and human rights investigations relating to 1984 and
- after
- W. H. Auden, Partition (1966)
The open questions here include: the death toll of 1947, which cannot be established and where the credible range
spans a factor of four; the number of women abducted, which is an estimate built on the recovery operation’s own
figures; the reasoning behind the Gurdaspur award, which Radcliffe destroyed the papers for; the responsibility for the
Rawalpindi massacres, which was never investigated by any authority; and almost every figure relating to 1984 and
after.
Four things about method, and they are the last in this book.
First, the violence of 1947 has been attributed to organisers and to events, never to communities. All three
killed and all three were killed, in different districts at different times, and any account that reads otherwise is a
political document rather than a history.
Second, the violence has been stated and not described. This part contains no account of what was done to
anybody. That is a deliberate editorial choice, applied consistently since Part Nine, and it is stronger here because the
people involved have grandchildren.
Third, on 1984 and after, this book has reported and refused to judge, for the reasons given in that section. It is
the only place in sixteen parts where it has declined to reach a conclusion, and that refusal is itself the position.
Fourth and last: every figure in this book is an argument. Sixteen parts of interrogating other people’s numbers,
and the honest close is to say that these ones deserve the same treatment. Where this book has been wrong, it hopes
somebody will say so, in print, with the sources.