Before Part 10
This is the hardest part of this work to write, and the reason is simple to state and difficult to execute.
This decade contains two large atrocities and a great many smaller ones. Almost everyone who writes about it uses one to excuse the other. Indian nationalist accounts describe the bus massacres and the Kanishka bombing at length and treat November 1984 as an unfortunate outbreak of public anger. Sikh nationalist accounts describe November 1984 and the fake encounters at length and treat the killing of Hindu bus passengers as either exaggerated, or provoked, or the work of agents provocateurs.
I am going to do neither. The rule for this part is the one I set in Part 3, when writing about Timur's order at Loni, and I want it restated here because this is where it costs something:
Every atrocity in this part is stated with the same bluntness, regardless of who committed it, and no atrocity is used to explain away another.
So: Chapter 2 says that voter lists were used to find Sikh homes in Delhi and that the police stood by. Chapter 4 says that militants stopped buses, separated passengers by religion, and shot the Hindus — and that a bomb planted by a Canadian Sikh organisation killed 329 people including 82 children. Chapter 5 says that the Punjab Police ran a bounty system that produced extrajudicial killings on a scale that has never been counted, and that the man who investigated it was abducted and murdered by police officers who were later convicted of it.
All of those are in here at full strength. If you find yourself reading quickly through one of those chapters to get to the one that suits you, that is the thing this part exists to interrupt.
Two other things.
There is one finding in Chapter 5 that settles a question left open in Part 9. In 1988 the Indian state faced the same problem at the same complex and solved it a completely different way — a siege rather than an assault, with the media present and the building largely intact. That is not my opinion about what was possible in 1984. It is what the same state did four years later.
And I have not used the highest available casualty figures. One estimate for the decade runs to 250,000. I have flagged it as an outlier rather than repeated it, for the same reason I did not use the highest Partition figures in Part 7. The verified numbers are appalling enough and they have the advantage of being defensible.
— L.S.Eleven years
Part 9 ended on 6 June 1984 in the ruins of the Akal Takht. This part covers what came next: an assassination, an organised pogrom, an eleven-year insurgency, a counter-insurgency that broke the law comprehensively, and a peace nobody signed.
The shape of it
| When | What |
|---|---|
| 31 Oct 1984 | Indira Gandhi assassinated by two of her own Sikh bodyguards. |
| 1–4 Nov 1984 | Organised killing of Sikhs across India, worst in Delhi. Around 2,733 dead in Delhi by the official commission's own figure. |
| 1985–1992 | Insurgency. Militant groups multiply. Bus massacres. Sikh moderates killed. The Kanishka bombing kills 329. |
| 1988 | Operation Black Thunder — the same complex, handled entirely differently, and successfully. |
| 1991–1993 | Counter-insurgency under KPS Gill defeats the militancy. Fake encounters, disappearances, and a bounty system. |
| 1995 | Jaswant Singh Khalra documents thousands of secret cremations, and is abducted and murdered by police. Chief Minister Beant Singh assassinated. The decade ends. |
The counting problem
Before any of it, a warning about numbers. There is no agreed figure for anything in this decade — not the pogrom, not the insurgency, not the counter-insurgency.
Official Indian government data for the conflict gives roughly: 1,714 security personnel, around 1,700 soldiers, 7,946 militants, and 11,690 non-combatants — around twenty-three thousand in total.
But the "militant" category is precisely the one that cannot be trusted, because a man killed in a staged encounter is recorded as a militant. That is the whole point of a staged encounter. Chapter 5 shows why the number of people who disappeared without ever appearing in any category at all may be very large.
131 October 1984
She was killed by the men assigned to protect her, in her own garden, five months after Operation Blue Star. Nobody has ever disputed the causal connection.
What happened
At around 9.20 on the morning of 31 October 1984, Indira Gandhi walked from her residence at 1 Safdarjung Road toward her office at 1 Akbar Road for a television interview.
Two of her security guards, Beant Singh and Satwant Singh, opened fire on her.
Beant Singh was killed shortly afterwards. Satwant Singh was wounded, survived, and was tried along with Kehar Singh; both were convicted and hanged in January 1989.
The detail that deserves recording
DOCUMENTED: After Operation Blue Star, security advice was given that Sikh personnel should be removed from Indira Gandhi's personal protection detail. She declined. Accounts from those around her record her taking the view that removing Sikhs from her guard would contradict everything she claimed to stand for on secularism.
Beant Singh had served on her staff for around a decade, and by several accounts she knew him personally and had specifically kept him.
I include this because it is true, because it is genuinely poignant, and because it is one of the few moments in this decade where a decision was made for a reason that was not cynical.
The causal chain
This is not a contested point in any account I have read, from any position: Operation Blue Star caused this assassination. The chain runs from the decision to assault the Akal Takht in June to two men deciding in October that the person who ordered it should not live.
What is contested is everything that happened next.
2November 1984
The word used matters, so let us settle it with evidence. A riot is spontaneous, disorganised and two-sided. What happened in Delhi between 1 and 4 November 1984 was none of those things.
Six features, and what each one requires
Here is the evidence, item by item, and what each item implies about organisation.
| What is documented | What it requires |
|---|---|
| Voter lists and ration card records were used to identify Sikh homes and businesses. Mobs arrived at specific addresses in mixed neighbourhoods and passed over the houses next door. | Access to official electoral and civil records, and time to compile lists from them. Ordinary crowds do not have this. |
| Mobs were transported to locations, in some documented cases by bus and truck. | Vehicles, organisation, and someone directing them. |
| The method was consistent across widely separated localities — Trilokpuri, Sultanpuri, Mangolpuri, Palam Colony — including the specific practice of burning victims with kerosene and tyres. | A shared instruction. Spontaneous crowds in separated neighbourhoods do not independently converge on identical methods. |
| Kerosene and tyres were available in quantity, in a period when kerosene was a rationed commodity. | Supply, and therefore supply chains. |
| The police did not intervene. The Nanavati Commission found police investigation "absolutely casual, perfunctory and faulty." In numerous documented cases police disarmed Sikhs who tried to defend themselves. | At minimum, instructions not to act. |
| The army was not deployed for roughly three days, in the national capital, with troops available. | A decision at a senior level. |
The figures
Justice Nanavati Commission of Inquiry (report published February 2005): 2,733 killed in Delhi. This is the Government of India's own commission and its own number.
Nationally: commonly cited totals run to around 3,000–3,350, with killings also in Kanpur, Bokaro and elsewhere.
Sikh organisations and some independent researchers give substantially higher figures, in some cases above 8,000 nationally.
I use the Nanavati figure as the floor because it is the state's own, and note that the range above it is contested. As with Blue Star, the uncertainty exists because no thorough contemporaneous count was made.
What the state's own commission found
The Nanavati Commission was appointed by the Ministry of Home Affairs in May 2000 and reported in February 2005. Its findings, in its own words:
- H. K. L. Bhagat, Congress Member of Parliament for East Delhi, along with local Congress figures Rampal Saroj and Dr Ashok of Trilokpuri, had taken "active part" in the attacks.
- On Sajjan Kumar: there was "credible evidence" that he was "probably involved" as alleged by witnesses, and the Commission recommended examining cases in which witnesses had named him but no charge sheets had been filed.
- Police investigation had been "absolutely casual, perfunctory and faulty."
Note what the Commission then did with its own finding on Bhagat: it recommended no further action, on the ground that he had already been acquitted in criminal cases — cases which its own report describes as having been investigated in a manner that was casual, perfunctory and faulty.
The impunity, measured
| Measure | Result |
|---|---|
| Trial court cases analysed by the Nanavati Commission | 126 |
| Of those, convictions | 8 |
| Of those, acquittals | 118 |
| Commissions and committees appointed over four decades | Around ten, including Marwah, Misra, Kapur-Mittal, Jain-Aggarwal, Ahuja, Dhillon, Narula, Nanavati, and a later Special Investigation Team |
| Sajjan Kumar | Convicted by the Delhi High Court and sentenced to life imprisonment in December 2018 — thirty-four years afterwards |
| Senior political organisers who have served a full sentence for organising | None |
The apology, and the sentence that did not help
In the aftermath, Rajiv Gandhi — who had become Prime Minister on his mother's death — said publicly that when a big tree falls, the earth shakes.
Whatever was intended, what it conveyed was that the killing was a natural phenomenon rather than a crime, and it has been quoted against Congress ever since.
A formal apology came from Prime Minister Manmohan Singh in Parliament in August 2005 — twenty-one years later.
The failure in November 1984 was not an inability to protect citizens. Delhi had police. Delhi had the army available. The state was not overwhelmed.
It was an unwillingness, sustained across three days, in the capital city, followed by four decades of investigations producing eight convictions out of a hundred and twenty-six cases.
That distinction is the single most important political fact of this decade, and Chapter 3 is about what it did.
3What November did
Before November 1984, Khalistan was a marginal proposition inside Punjab. After it, the movement had a mass base. This chapter explains the mechanism, and states the complication that Sikh nationalist accounts leave out.
The argument that changed status
The core claim of Sikh separatism had always been: Sikhs are not safe as a minority in India.
Before November 1984 that was a claim — arguable, contested, and in the view of most Punjabi Sikhs at the time, not obviously true. Punjab was the richest state in India. Sikhs were heavily over-represented in the army and the civil services. A Sikh had been President of India. The claim did not match most people's daily experience.
After November 1984, for a very large number of Sikhs, it stopped being a claim and became a demonstrated proposition — demonstrated not by argument but by three days in the national capital during which the state did not act.
That is the mechanism. It is not complicated, and no amount of subsequent argument has undone it.
The complication, stated honestly
Two facts belong here that Sikh nationalist accounts tend to leave out, and I am including them because Rule 1 does not have exceptions for inconvenience.
First: the violence was worst outside Punjab. The killing was concentrated in Delhi, and occurred in Kanpur, Bokaro and other cities. Inside Punjab, where Sikhs were the majority, there was no comparable pogrom.
Second: there were many documented cases of Hindus sheltering Sikhs — hiding neighbours, lying to mobs, moving families to safety at real personal risk. This happened in Delhi and elsewhere and is recorded in survivor testimony. The mobs were organised; the general Hindu population of India was not in them.
Neither of those facts reduces what was done. But they matter, because the militant movement's subsequent framing — that Hindus as such were the enemy of Sikhs as such — is contradicted by the conduct of a great many ordinary Hindus in the very week that is used to justify it.
Recruitment
DOCUMENTED, and attested from within the movement: recruitment into militant groups surged after November 1984. Scholars who interviewed militants — Cynthia Mahmood and Joyce Pettigrew both did extended fieldwork — record participants naming November 1984 as their reason for joining more often than Blue Star, and far more often than any political demand.
Congress promoted and protected Bhindranwale for electoral advantage (Part 9, Chapter 2).
Bhindranwale fortified the Akal Takht (Part 9, Chapter 3).
The state assaulted it, with tanks, on a martyrdom anniversary, with pilgrims inside (Part 9, Chapter 5).
Two of her guards killed the Prime Minister.
An organised pogrom killed at least 2,733 Sikhs in Delhi while the state stood back for three days.
And the insurgency that followed was staffed by people whose stated reason for joining was that last event.
Every link in that chain involves a decision by an identifiable person or institution. None of it was inevitable, and at four separate points a different decision was available.
4What the militants did
Stated with the same bluntness as Chapter 2. This chapter is where a great deal of Sikh nationalist writing goes quiet, and where a great deal of Indian nationalist writing gets the emphasis wrong in a way that matters.
The groups
| Group | Notes |
|---|---|
| Khalistan Commando Force (KCF) | Formed April 1986; the first major organised group. Assassinated General A. S. Vaidya — Army Chief at the time of Blue Star — in Pune in 1986. Later responsible for the assassination of Chief Minister Beant Singh in 1995. |
| Khalistan Liberation Force (KLF) | |
| Babbar Khalsa International (BKI) | Led in Canada by Talwinder Singh Parmar. Responsible for the Kanishka bombing. |
| Bhindranwale Tiger Force of Khalistan (BTFK) | |
| Khalistan Zindabad Force and others | By the late 1980s there were dozens of groups and factions, frequently in conflict with one another. |
The bus massacres
DOCUMENTED, repeatedly, across years: the method was to stop a bus on a rural road, order the passengers out, separate them by religion — usually by name, appearance or by asking them directly — and shoot the Hindus.
These occurred through the late 1980s and into the 1990s. In July 1987, in two attacks on the same day at Lalru and near Fatehabad, some seventy to eighty bus passengers were killed in this manner. There were many other incidents, some smaller, over the following years.
I am not going to describe these further. What matters analytically is the method: the victims were selected by religion and killed for that reason, having been unarmed civilians on public transport. By any definition available in any tradition, including the one in Part 4 of this work, they were non-combatants.
The killing of Sikhs — and this is what gets dropped
Here is the fact most likely to be missing from whichever account you last read.
A very large share of the people killed by Sikh militants were Sikhs.
- Sikh police officers, in very large numbers. The Punjab Police, which fought this insurgency, was overwhelmingly Sikh. Militants killed them and, in many cases, their families.
- Sikh journalists who reported critically.
- Sikh politicians of the Akali Dal and other parties who negotiated or dissented.
- Sikh religious figures who did not endorse the movement.
- And ordinary Sikh villagers who refused to shelter, feed or fund them.
The clearest single case: Harchand Singh Longowal, the Akali Dal president who had led the Dharam Yudh Morcha, signed the Rajiv–Longowal Accord on 24 July 1985 — the agreement that promised Chandigarh to Punjab by January 1986.
He was shot dead on 20 August 1985, less than a month later, inside a gurdwara, by Sikh militants.
The Kanishka bombing
On 23 June 1985, a bomb placed in luggage loaded in Canada exploded aboard Air India Flight 182 — a Boeing 747 named Emperor Kanishka — off the coast of Ireland.
| Item | Detail |
|---|---|
| Killed | 329 — 307 passengers and 22 crew |
| Children among the dead | 82 |
| Nationality of most victims | Canadian, the majority of Indian origin |
| Same day | A second bomb from the same operation exploded at Narita airport in Japan, killing two baggage handlers |
| Identified as mastermind | Talwinder Singh Parmar, leader of Babbar Khalsa in Canada |
| Convictions | Inderjit Singh Reyat is the only person convicted — manslaughter, and subsequently perjury. Two others were acquitted in 2005. |
| Significance | Until 11 September 2001, the deadliest act of aviation terrorism in history |
The Canadian public inquiry led by Justice John Major, reporting in 2010, found what it described as a cascading series of errors by Canadian agencies before and after the bombing.
Three things about this need saying without qualification. The victims were overwhelmingly Canadian civilians. Eighty-two of them were children. And it was planned and executed in Canada, by a Canadian organisation, against people who had no connection whatsoever to anything that had happened in Punjab.
The degeneration
This is documented from within, including by scholars sympathetic to the movement's origins.
DOCUMENTED: by the late 1980s and into the 1990s, many militant groups had substantially degenerated into criminal operations. Extortion of farmers and businesses under the name of "taxes." Kidnapping for ransom. Rape, documented by Punjabi villagers and recorded in the fieldwork of researchers including Joyce Pettigrew and Cynthia Mahmood, both of whom took the movement's political origins seriously and interviewed its members.
Groups also issued and enforced social diktats at gunpoint in villages: dress codes for women, bans on meat and alcohol, rules on weddings and dowry.
Chapter 6 shows why this mattered more than any counter-insurgency operation.
Pakistan and the ISI
DOCUMENTED: Pakistan's Inter-Services Intelligence established a dedicated cell for Punjab from the early-to-mid 1980s, provided arms, funding and sanctuary, and ran training camps — locations including Lahore and Karachi are named in multiple accounts. Support went to Babbar Khalsa, the Khalistan Commando Force, the Khalistan Liberation Force and the Bhindranwale Tiger Force among others.
The framing that is wrong: "the ISI created the Khalistan movement." It did not. The grievances documented in Part 8, the manipulation documented in Part 9, the Karnal acquittals, Blue Star and November 1984 were all entirely domestic in origin, and every one of them happened without Pakistani involvement.
The framing that is right: the ISI took an existing insurgency with genuine domestic causes and supplied it with weapons, money, training and safe territory — which dramatically extended its lifespan and lethality, and shifted it further toward methods and targets that had no relationship to Punjabi grievances at all.
Both halves matter. An account that omits the ISI is incomplete. An account that blames the ISI for everything is a way of avoiding Parts 8 and 9 entirely.
5What the state did
Same bluntness. This chapter contains the finding that settles a question I left open in Part 9, and the story of a bank manager who worked out how to count the dead using firewood receipts.
1988: Operation Black Thunder — and what it proves
In May 1988, militants had again occupied the Golden Temple complex. The state faced, in substance, the same problem it had faced in June 1984: armed men in the holiest site in Sikhism.
This time it did something different.
| Element | Operation Blue Star, June 1984 | Operation Black Thunder, May 1988 |
|---|---|---|
| Method | Frontal military assault | Siege. Water, power and supplies cut; snipers; wait. |
| Force used | Army, including tanks firing into the Akal Takht | National Security Guard and Punjab Police under KPS Gill. No armour. |
| Media | Total blackout; journalists expelled | Journalists allowed to observe from nearby rooftops |
| Damage to the complex | Akal Takht destroyed; Reference Library burned | Minimal |
| Outcome | Casualty range from ~500 to several thousand, disputed | Militants surrendered. Around forty deaths. |
In Part 9 I gave the strongest case for Blue Star — that a state which cannot enter an armed fortress in its own territory has ceased to govern — and said the operation may have been necessary while its method was not defensible.
I was relying on argument there. I do not need to now.
Four years later, the same state, at the same complex, facing the same problem, used a siege instead of an assault, let the press watch, left the building standing, and the militants surrendered.
That is not my speculation about what might have been possible in 1984. It is a documented operation. The method used in June 1984 — tanks, assault, blackout, on a martyrdom anniversary with pilgrims inside — was a choice, and Operation Black Thunder is the proof.
The counter-insurgency, 1991–1993
K. P. S. Gill, Director General of Punjab Police, ran the campaign that defeated the militancy. That it worked is not in dispute, and pretending otherwise would be dishonest.
How it worked is also documented:
- A bounty system. Cash rewards and out-of-turn promotions for police officers who killed militants. This created a direct, standing financial incentive to kill rather than arrest, and to classify the dead as militants.
- Fake encounters. Staged killings recorded as gun battles. The practice is documented extensively by Indian and international human rights organisations and is not seriously denied.
- Disappearances. People taken into custody who never reappeared, in any court, prison record or morgue.
- Targeting families. Relatives of suspects detained and tortured to force surrenders.
- Legal cover. TADA — the Terrorist and Disruptive Activities (Prevention) Act — and the Punjab Disturbed Areas Act, which together produced extremely broad powers of detention and substantial protection from prosecution.
Jaswant Singh Khalra
This is the centre of the chapter.
Jaswant Singh Khalra (1952–1995) was a bank director in Amritsar who became a human rights investigator. He was trying to establish what had happened to people who had disappeared, and he found a method that nobody had thought of.
Cremation grounds have to buy firewood, and they have to keep records of it.
Khalra and his colleague Jaspal Singh Dhillon obtained municipal crematorium records and firewood purchase receipts from three cremation sites in Amritsar district, and cross-referenced them against police records of bodies received.
On 16 January 1995 they published their findings in a press note titled Disappeared: Cremation Grounds.
| Finding | Figure |
|---|---|
| Secret cremations documented in Amritsar district alone, of bodies recorded as "unidentified" or "unclaimed" | More than 2,000 |
| Estimate for Punjab as a whole, extrapolated by his team | At least 25,000 |
These were people picked up by police, never charged, never produced in court, and disposed of without their families being told.
What happened to him
On 6 September 1995, Jaswant Singh Khalra was abducted from outside his home in Amritsar.
He was murdered in police custody.
DOCUMENTED, AND ESTABLISHED BY AN INDIAN COURT: six Punjab Police officials were convicted of his abduction and murder and sentenced to life imprisonment.
Because it removes the argument.
Anyone who wants to dismiss the disappearances as separatist propaganda has to explain why the Indian judicial system convicted six serving police officers of abducting and murdering the man who was documenting them.
The state's own courts established that police killed the investigator. A subsequent CBI investigation confirmed thousands of illegal cremations in the Amritsar-district crematoria he had examined.
The scale beyond Amritsar has never been fully established, because no comprehensive independent investigation has been permitted across Punjab. Khalra's 25,000 figure is an extrapolation from one district and should be treated as an estimate, not a count. That it remains an estimate is, once again, a consequence of a decision not to count.
The conflict's casualty figures
Official Indian government data for the conflict: 1,714 security personnel, around 1,700 soldiers, 7,946 militants, and 11,690 non-combatants. Roughly 23,000 total.
The problem with that table is the militant column. In a system with a bounty on militants killed and a documented practice of staged encounters, "militant" is a classification made by the person who did the killing, with a financial interest in the classification.
Higher estimates exist. One figure of up to 250,000 circulates; I regard it as an outlier that I cannot substantiate and I am not going to use it, on the same principle by which I declined the highest Partition figures in Part 7.
What can be said defensibly: the official figure of roughly 23,000 is a floor, not a total, and the true number is unknown by design.
6How it ended
Six reasons, and the one most people leave out is probably the most important.
The end, by date
1992: state elections were held under an Akali boycott. Turnout was around 24%. A Congress government took office under Beant Singh — a government with thin legitimacy, but a civilian government nonetheless.
1993: the insurgency was substantially over as a military proposition.
31 August 1995: Chief Minister Beant Singh was killed by a suicide bomber outside the Punjab Civil Secretariat in Chandigarh. The bomber, Dilawar Singh, was a serving Punjab Police officer acting for Babbar Khalsa. Seventeen others died. This is generally taken as the last major act of the insurgency.
Why it ended — six factors
1. The police methods worked. Brutally, illegally, and effectively. Denying this would be dishonest, and it is the uncomfortable finding for anyone who believes that atrocity is always also strategically counter-productive. Here it was not.
2. The militants lost the population — and this is the big one.
By the early 1990s, the movement had spent years extorting farmers, kidnapping for ransom, committing rape, issuing social diktats at gunpoint, and killing Sikhs who disagreed. Rural Punjab, which had sheltered and fed the movement after 1984, stopped doing so.
An insurgency in a rural society survives on village support. When the village stops feeding you, stops hiding you, and starts telling the police where you sleep, you are finished — regardless of how many weapons the ISI supplies.
3. Fragmentation. Dozens of groups and factions, competing for territory and extortion revenue, frequently killing each other.
4. No political programme. In eleven years the movement never produced a governing proposal, a constitution, an economic plan, or a credible political leadership capable of negotiating. It had a demand and it had guns, and it never built the third thing.
5. Exhaustion. A decade.
6. The return of ordinary politics. In 1997 an Akali Dal–BJP coalition won the Punjab election. A Sikh regional party governing in alliance with a Hindu nationalist party, two years after Beant Singh's assassination.
That last fact is worth sitting with. Whatever the militant movement believed about the permanent and irreconcilable enmity of Sikhs and Hindus, Punjabi voters demonstrated a different view within two years of the shooting stopping.
7The accounting
Two ledgers, stated in the same language. Then the asymmetries that actually matter — because there are asymmetries, and they run in both directions.
| The militants | The state |
|---|---|
| Stopped buses, separated passengers by religion, and shot unarmed Hindu civilians — repeatedly, over years. | Allowed an organised pogrom to kill at least 2,733 Sikhs in Delhi over three days without deploying available troops. |
| Bombed a civilian airliner, killing 329 people including 82 children, almost all Canadian civilians with no connection to Punjab. | Assaulted the holiest site in Sikhism with tanks on a martyrdom anniversary, with pilgrims inside, and burned an irreplaceable manuscript library. |
| Killed large numbers of Sikhs — police, journalists, politicians, clergy and villagers — including the Akali leader who signed a peace accord, shot in a gurdwara four weeks later. | Ran a bounty system that incentivised killing over arrest, staged encounters, and disappeared people who were then cremated as unidentified. |
| Degenerated into extortion, kidnapping and rape, and enforced social rules on villagers at gunpoint. | Abducted and murdered the human rights investigator who documented the cremations — a fact established by the conviction of six police officers. |
| Took arms, money, training and sanctuary from a hostile foreign intelligence service. | Produced eight convictions from a hundred and twenty-six cases over four decades, and has permitted no independent inquiry into Blue Star in forty-two years. |
The asymmetries that matter
Having set the ledgers side by side, I do not think the correct conclusion is that both sides were simply equivalent. There are three real asymmetries and they do not all run the same way.
1. A state is held to a higher standard, and that is not bias. It is the definition of a state. A government holds a monopoly on legitimate force precisely on the condition that it exercises it under law. An insurgency that murders civilians is committing a crime. A state that murders civilians is also destroying the thing that distinguishes it from an insurgency. That is why "but they did it too" is not available to a government.
2. The militants killed people specifically for their religion, as a method, at scale. The bus massacres were not excesses. They were a technique, repeated deliberately over years, and its entire logic was that being Hindu was sufficient reason to be shot. The Kanishka bombing killed 329 people who had no involvement in anything at all. There is no framework — legal, moral, or in the Sikh tradition documented in Part 4 — under which that is defensible.
3. And the impunity is symmetrical in a way that damns everybody. Eight convictions out of a hundred and twenty-six for a pogrom that killed thousands. Effectively no accountability for the fake encounters. No independent inquiry into Blue Star. One conviction, for manslaughter, for a bombing that killed 329 people. Whichever community you belong to, the institutions available to you failed to deliver justice for what was done to you.
Who won
Nobody.
| Claimed objective | Achieved? |
|---|---|
| Khalistan | No. Never came close. No territory ever held, no government ever formed, no state ever recognised. |
| Chandigarh to Punjab | No. Still a Union Territory in 2026. |
| River water reallocation | No. The 1976 allocation stands. |
| Article 25 amendment | No. |
| The Anandpur Sahib Resolution | Not implemented. |
| Justice for November 1984 | Eight convictions from 126 cases. One senior figure convicted, in 2018. |
| Justice for the disappearances | Six convictions, for the murder of the investigator. Essentially none for the killings he documented. |
| Punjab's economy and society | A decade of capital flight, halted investment, a lost generation, and the beginnings of the crisis in Part 13. |
8Scorecard: what Part 10 proves
FALSE Voter lists and ration records used to identify Sikh homes in mixed neighbourhoods. Mobs transported. Identical methods across widely separated localities. Rationed kerosene supplied in quantity. Police non-intervention, including disarming Sikhs who tried to defend themselves. The army not deployed in the capital for around three days.
Those are not features of a spontaneous crowd. They are features of an organisation. The word is pogrom.
NO — AND THE STATE'S OWN COMMISSION SAYS SO The Nanavati Commission found police investigation "absolutely casual, perfunctory and faulty," named H. K. L. Bhagat and local Congress figures as having taken "active part," and found "credible evidence" that Sajjan Kumar was "probably involved."
Of 126 trial court cases it analysed: 8 convictions, 118 acquittals. Around ten commissions across four decades. Sajjan Kumar convicted in 2018, thirty-four years later. No senior organiser has served a full sentence. A formal apology came in 2005, twenty-one years afterwards.
CONTRADICTED BY THEIR OWN RECORD A very large share of the people they killed were Sikhs — police, journalists, politicians, clergy, and villagers who would not fund them. Harchand Singh Longowal, who signed the accord promising Chandigarh to Punjab, was shot in a gurdwara four weeks later.
They bombed a civilian airliner and killed 329 people, 82 of them children, almost all Canadian civilians. They extorted Punjabi farmers, kidnapped for ransom, committed rape, and enforced dress codes on Punjabi women at gunpoint.
Whatever that was, it was not a defence of Sikh rights, and it fails every test set out in Part 4 from the tradition's own sources.
REFUTED — BY THE INDIAN STATE ITSELF Operation Black Thunder, May 1988. Same complex. Same problem. A siege instead of an assault. No tanks. Journalists permitted to watch. The building left standing. The militants surrendered. Around forty deaths.
Four years apart, two operations, two methods, two outcomes. The 1984 method was a choice, and this is not an inference — it is a documented comparison.
REFUTED IN AN INDIAN COURT Jaswant Singh Khalra used municipal crematorium and firewood purchase records to document over 2,000 secret cremations in Amritsar district alone, of bodies logged as unidentified.
He was abducted outside his home on 6 September 1995 and murdered in police custody. Six Punjab Police officials were convicted and sentenced to life imprisonment for it. A subsequent CBI investigation confirmed thousands of illegal cremations at the sites he examined.
His statewide estimate of 25,000 is an extrapolation from one district and should be treated as an estimate. That it has never been replaced with a count is the direct result of no comprehensive independent investigation being permitted.
FALSE AS STATED; TRUE WHEN AMENDED The ISI established a Punjab cell, ran training camps, and supplied arms, money and sanctuary to Babbar Khalsa, the KCF, the KLF and others from the early-to-mid 1980s. That is documented and significant.
But every cause of the movement was domestic: the broken Chandigarh commitments, the water allocation, the Karnal acquittals, the Congress cultivation of Bhindranwale, Blue Star, and November 1984. Not one of those required Pakistan.
The ISI did not create the fire. It poured petrol on it for a decade, which made it burn far longer and kill far more. Blaming Pakistan for everything is a way of not reading Parts 8 and 9.
NOBODY. PUNJAB LOST. No Khalistan. No Chandigarh — still a Union Territory. No change to the water allocation. No Article 25 amendment. No implementation of the Anandpur Sahib Resolution. Eight convictions for a pogrom that killed thousands. Effectively none for the disappearances. One manslaughter conviction for 329 dead.
And a decade of capital flight, halted investment and a lost generation, which is where Part 13 has to begin.
Timeline of Part 10
What happens in Part 11
Part 11, Khalistan: The Idea and the Network, follows the movement from the ruins of 1995 to 2026 — and it is largely a story that takes place outside Punjab.
It traces the idea from the 1940 pamphlet that Master Tara Singh denounced (Part 7), through Jagjit Singh Chohan's declarations from London in the 1970s and 1980s, to the present network: Sikhs for Justice and Gurpatwant Singh Pannun, the Referendum 2020 exercise, the diaspora funding structure, and the political conditions in Canada, Britain and the United States that sustain it.
It covers the killing of Hardeep Singh Nijjar in British Columbia in June 2023 and the diplomatic crisis that followed, the alleged plot against Pannun in New York, and what those episodes revealed about both the movement and the Indian state.
And it asks the question this part makes unavoidable: if the movement was destroyed inside Punjab by 1995, and Punjabi voters returned an Akali–BJP coalition in 1997, what exactly is the thing that still exists — and where does its support actually come from?
The Ghadar template from Part 6 becomes the analytical frame. The electoral evidence from inside Punjab becomes the test.
Sources & further reading — Part 10
Official reports and primary documents
- Report of the Justice Nanavati Commission of Inquiry (1984 Anti-Sikh Riots), Volumes I and II, February 2005 — available from the Ministry of Home Affairs. The state's own findings on H. K. L. Bhagat, Sajjan Kumar and the quality of police investigation.
- Earlier inquiries: the Marwah, Misra, Kapur-Mittal, Jain-Aggarwal, Ahuja, Dhillon and Narula committees.
- Air India Flight 182: A Canadian Tragedy — Report of the Commission of Inquiry under Justice John C. Major, 2010. Available from Public Safety Canada.
- Khalra, Jaswant Singh and Dhillon, Jaspal Singh, Disappeared: Cremation Grounds, press note, 16 January 1995.
- CBI reports on illegal cremations in Amritsar district, filed pursuant to Supreme Court direction.
- Rajiv–Longowal Accord, 24 July 1985.
Human rights documentation
- Human Rights Watch, India: No Justice for 1984 Anti-Sikh Bloodshed (2014), and its earlier Punjab reporting.
- Ensaaf and the Punjab Documentation and Advocacy Project — extensive documentation of disappearances and cremations, including the Punjab Disappeared archive.
- People's Union for Civil Liberties (PUCL) and People's Union for Democratic Rights (PUDR), Who Are the Guilty? (November 1984) — the earliest independent investigation into the Delhi pogrom, produced within weeks.
- Amnesty International reporting on Punjab in the 1990s.
Scholarship and journalism
- Joyce Pettigrew, The Sikhs of the Punjab: Unheard Voices of State and Guerrilla Violence — fieldwork-based, and unusual in documenting both.
- Cynthia Keppley Mahmood, Fighting for Faith and Nation: Dialogues with Sikh Militants.
- Uma Chakravarti and Nandita Haksar, The Delhi Riots: Three Days in the Life of a Nation.
- Manoj Mitta and H. S. Phoolka, When a Tree Shook Delhi — the most detailed account of the 1984 pogrom and the legal aftermath.
- Jaskaran Kaur, Twenty Years of Impunity.
- K. P. S. Gill, Punjab: The Knights of Falsehood — the counter-insurgency commander's own account.
- Kim Bolan, Loss of Faith: How the Air-India Bombers Got Away with Murder — the definitive journalistic account of the Kanishka case.
- Ramachandra Guha, India After Gandhi, chapters on Punjab and on 1984.
Reference sources used for dates and figures
- Wikipedia, Jaswant Singh Khalra, Kanwar Pal Singh Gill, Beant Singh, Dilawar Singh, Talwinder Singh Parmar, Khalistan Commando Force, Khalistan Liberation Force; Britannica, "Air India Flight 182 disaster."
- The Tribune and Deccan Herald retrospectives, "1984 anti-Sikh riots: 40 years on, major highlights and where cases stand."
- ThePrint, "Jaswant Singh Khalra's long hunt for truth of Punjab disappearances"; the Sikh Coalition's biographical materials on Khalra.
- Sciences Po Mass Violence and Resistance research network, "The Anti-Sikh Pogrom of October 31 to November 4, 1984."
Two numbers circulate widely for this decade that I have not adopted. One puts total conflict deaths as high as 250,000. Another puts the November 1984 national toll above 8,000.
Neither is impossible; neither is something I can substantiate from sources I would rely on. I have used the state's own figures as floors — 2,733 for Delhi, roughly 23,000 for the conflict — and said explicitly that they are floors rather than totals, and why.
This is the same discipline I applied to the Vadda Ghallughara in Part 5 and to Partition in Part 7. A defensible number that is appalling is more useful than a larger one that can be dismissed.