Kashmir 1947: A 14-Month War and a Vote That Never Happened

On the morning of 27 October 1947, three Dakota transport planes lifted off from Palam airfield outside Delhi with an unusual instruction: circle Srinagar before you land, and check whether the runway is still in friendly hands.

Nobody in Delhi knew. Armed men had crossed into the Kashmir Valley from the Pakistani side five days earlier and were fifty kilometres off. The first Indian soldiers touched down at half past nine. Their commanding officer, Lieutenant Colonel Dewan Ranjit Rai, was killed within days.

Most Indian accounts of the first India-Pakistan war start with that flight. Most Pakistani accounts start two months earlier, in a hill district called Poonch. The gap between them is why the war never properly ended.

A kingdom with no good options

When the British left in August 1947 they did not only cut British India in two. They also released around 560 princely states, kingdoms that had run their own internal affairs under British supervision, and told each ruler to pick a side: India, Pakistan, or, with heavy caveats, neither.

Jammu and Kashmir was the largest and the most awkward. The 1941 census counted a population roughly 77 per cent Muslim, and the Kashmir Valley was over ninety per cent Muslim. The ruler, Maharaja Hari Singh, was a Hindu of the Dogra dynasty, which had held the state since 1846.

He did not want to choose. On 12 August 1947 his government offered both new countries a standstill agreement, a holding arrangement to keep post, telegraph and trade running while he made up his mind. Pakistan accepted three days later. India asked for talks first, and never signed one.

He was still thinking in October. By then he had lost the argument on the ground.

The rebellion that came first

A campaign against the Maharaja’s taxes was running in Poonch by June 1947. By late August it had become an armed revolt.

The Australian historian Christopher Snedden, who has done more than anyone to reconstruct this, points to a trigger. Poonch was full of men who had fought for Britain in the Second World War. After tens of thousands of them gathered at Rawalakot in April, a nervous administration pressed local Muslims to hand in their weapons in July, and passed the guns to Hindus and Sikhs.

These were soldiers. Within six weeks the rebels held most of the district, and by September some 50,000 men were under arms as an “Azad”, or free, army. The state’s answer was heavy: Snedden’s sources describe troops firing on crowds, burning villages and imposing local martial law, with shoot-on-sight orders apparently issued on 2 September.

On 22 September, a month before any tribesmen crossed the border, the Maharaja’s own senior commander reported that he was losing control of large parts of the state.

Snedden’s conclusion, that the dispute began here rather than with the invasion, is contested; Indian reviewers argue the trouble was local and containable until outsiders arrived. The sequence itself is not in doubt.

Jammu, autumn 1947

Further south, something worse was happening.

Through the autumn of 1947, communal killing swept Jammu province in both directions. Muslims were killed in and around Jammu city; Hindus and Sikhs were killed in the western districts. Hundreds of thousands fled.

The numbers should make anyone cautious. Snedden’s range for Muslims killed runs from 70,000 to 237,000, the upper figure from The Times of London on 10 August 1948. Horace Alexander, in The Spectator that January, put it near 200,000 and said the killings had “the tacit consent of State authority”. Other scholars work with a much lower band, 20,000 to 100,000. Estimates of Hindus and Sikhs killed sit around 20,000 to 40,000.

None of these figures rests on a count of bodies. The historian Ian Copland has argued the violence against Jammu’s Muslims was partly revenge for the Poonch rebellion. The outcome is not in dispute: a province that had been Muslim-majority overall stopped being one, and never became one again.

The lorries on the Jhelum road

On 22 October 1947, Pashtun tribal militias from Pakistan’s North West Frontier crossed into Kashmir and drove down the Jhelum Valley road in lorries, heading for Srinagar.

They reached Baramulla on 26 October and sacked it. A party stormed the Catholic mission at St Joseph’s, killing a British colonel, his pregnant wife and a nun among others. Townspeople afterwards spoke of around 3,000 dead; the number was never verified, and Andrew Whitehead, who wrote the book-length study of the raid, treats it as uncertain. What the raiders did do, by stopping to loot, was lose two days they could not spare.

How much of this was Pakistan’s doing is the oldest argument here.

Documented: Major General Akbar Khan of the Pakistan Army wrote a plan called “Armed Revolt inside Kashmir” and later commanded under the name “General Tariq”, as he admitted in a memoir later banned in Pakistan. On 27 October Jinnah ordered his acting army chief, General Sir Douglas Gracey, to send in regular troops; Gracey refused, and Field Marshal Auchinleck got the order withdrawn.

Contested: Indian sources describe a Pakistan Army plan codenamed “Operation Gulmarg”, set in motion from about 20 August 1947. Pakistani sources deny any such plan existed.

Settled in the end: Pakistani regular units were inside Kashmir by May 1948, and Pakistan acknowledged them to the UN commission that July, having denied it before.

A signature, an airlift, and an argument about the order

Hari Singh asked India for troops. India said yes on one condition: he must sign an Instrument of Accession, the legal document by which a princely state joined one of the two new countries, handing over defence, foreign affairs and communications.

The official Indian date for that signature is 26 October 1947. Lord Mountbatten’s acceptance the next day carried a promise: once law and order were restored and the invaders cleared out, “the question of the State’s accession should be settled by a reference to the people”.

Here the historians split, and it matters: was India an invited ally, or did it write the invitation afterwards? Alastair Lamb argued the Maharaja could not have signed on the 26th, since he was on the road to Jammu with no recorded contact with Indian officials, and that he most likely signed on the 27th, after the troops had landed. Srinath Raghavan credits Lamb with the finding. Prem Shankar Jha argued the other way, that Hari Singh signed on the 25th, before leaving Srinagar. An American cable from Karachi on 28 October reported that the Maharaja had “signed on October 27”.

Lamb later claimed he never signed at all, which is generally treated as a step too far. But the narrow question, whether the signature came before the airlift, has no agreed answer.

Fourteen months in the snow

The war was fought in terrible country, mostly by small numbers of men. Poonch town was cut off from November 1947 for about a year, supplied by air after besieged civilians built an airstrip by hand. In Gilgit, the Gilgit Scouts arrested the Maharaja’s governor and raised the Pakistani flag on 1 November 1947. At Skardu, roughly 300 state troops held out through snow and hunger until 14 August 1948. In November 1948 Indian engineers winched dismantled light tanks up a mule track to take the Zoji La pass at around 11,500 feet, then the highest altitude armour had been used in combat. By the end of 1948 both armies were spent, and the argument had moved to New York.

The vote that was promised

India had taken the matter to the UN Security Council in January 1948, and Pakistan filed its own complaint. The Council set up a commission and on 21 April 1948 passed Resolution 47, recommending steps to stop the fighting and create conditions for a free and impartial plebiscite: a direct vote in which Kashmiris would choose which country to belong to.

The order was the problem. Pakistan was to withdraw the tribesmen and its own nationals; India was then to cut its forces to the minimum needed for law and order; only then would a UN-appointed administrator run the vote. The commission’s resolution of 13 August 1948 restated the sequence in firmer language.

Neither step happened.

India’s reading: the sequence is binding, it begins with Pakistan, and Pakistan never completed the first step, so nothing after it fell due. Pakistan’s reading: no vote could be free while Indian troops remained, so withdrawing first was surrender, and the plebiscite was the substance of the deal, not the order of operations.

Both readings hold together on their own terms. Neither side ever agreed what “minimum forces” meant, or what to do about the Azad Kashmir forces. The vote has never been held.

A line, not a border

The ceasefire took effect one minute before midnight on 1 January 1949. UN observers arrived three weeks later. On 27 July 1949, officers from both sides met in Karachi and drew a ceasefire line, marking where each army stood when the shooting stopped, with no claim to be a real frontier.

India held roughly two thirds of the old princely state: the Kashmir Valley, most of Jammu, and Ladakh. Pakistan held roughly one third: the western districts that became Azad Jammu and Kashmir, plus Gilgit and Baltistan. In 1972 the line was adjusted and renamed the Line of Control. It is still there.

The military death toll is small and unreliable. Indian official records give 1,103 soldiers killed, other Indian statements 1,500. Pakistan’s own count is 2,633 troops and tribesmen killed; Indian and outside estimates of the Pakistani-side dead run to about 6,000. Nobody counted the irregulars properly, and nobody counted the civilians at all.

That is the shape of it: a ruler who would not choose, subjects who chose without him, two countries ten weeks old, and a war each could plausibly say it had not started. The valley has had harder years since, and every one runs back through 1947.

Fourteen months of fighting produced a line on a map and a promise on paper. The line has outlasted almost everyone who drew it. The promise is still waiting.