Was the 1948 fall of Hyderabad peaceful and necessary?

The five days, and what the casualty tables actually rest on

Indian columns crossed into Hyderabad before dawn on 13 September 1948. The Hyderabad State Forces surrendered on the 17th; Major-General J. N. Chaudhuri became Military Governor, and Delhi decided that the Nizam could retain his position as Rajpramukh, 'though law-making and enforcement power rested with the Military Governor' ⁵. Delhi never called it a war. 'Police action' was a legal and diplomatic choice as much as a description, and it had a specific job to do: by cablegram dated 21 August 1948 Hyderabad brought a complaint to the Security Council under Article 35(2), stating that a grave dispute had arisen with India and that 'Hyderabad, a State not a Member of the United Nations, accepts for the purposes of the dispute the obligations of pacific settlement provided in the Charter of the United Nations' ⁹. Calling the operation policing rather than war asserted that Hyderabad was an internal jurisdiction and not a belligerent. That the label was a euphemism rather than a description is not only a journalistic or partisan reading. Saksena's footnote in the Journal of the History of International Law states that 'Quite duplicitously, official Indian histories have always referred to the invasion as a "police action" rather than an invasion', and names as its instance S. N. Prasad's Ministry of Defence history of 1972 — the same official volume to which the campaign casualty table is routinely credited and which this page lists in further reading ²⁸. Both things are true at once, and the page says both: the label did real diplomatic work at the Security Council, and it is treated in peer-reviewed scholarship as duplicitous.

The campaign figures are worth stating carefully. The set most often quoted — 807 Hyderabad State Forces killed and 1,647 captured, 1,373 Razakars killed and 1,911 captured — is routinely credited to S. N. Prasad's official Ministry of Defence history of 1972. That is a real, catalogued book (Historical Section, Ministry of Defence, New Delhi, 1972) which this page could not consult; the attribution is therefore unverified here. What can be checked is that Purushotham cites Prasad elsewhere but not for the casualty table, and credits the table instead to Lucien Benichou's 'From Autocracy to Integration', p. 243: 'The official casualty statistics were: 1,373 Razakars killed, 42 wounded, and 1,911 captured; Hyderabad Army: 807 killed, 64 wounded, and 1,647 captured; an additional 43 Hyderabadi combatants were killed. On the Indian side only ten were killed.' ⁴ The figures quoted here are Benichou's via Purushotham, and are labelled as such.

A materially different set circulates in Indian journalism, attributed there to S. N. Joshi's 'The Police Action Against Hyderabad' rather than to Prasad: the Indian Army at 42 killed, 97 wounded and 24 missing; the Hyderabad Army at 490 killed; the Razakars at 102 killed and 3,364 captured ²³. The Hyderabad State Forces figure therefore varies by a factor of about 1.6 and the Indian Army death figure by a factor of four across sources that all present themselves as official — and the second set's Razakar deaths are an order of magnitude below the first's. Summed, one set gives about 2,223 combatant dead and the other about 634. A third set, from a third-party diplomatic observer rather than from any Indian official history, is set out below: the UK High Commissioner's representative in Hyderabad reported 1,400 Hyderabad forces killed and 700 wounded, without separating State Forces from Razakars ³¹. What survives that variance is the shape rather than the number: combatant deaths in the hundreds or low thousands, Indian military deaths in the tens. On any of them the five days were cheap.

The problem is what the claim attaches to. Under the narrowest reading — the operation, 13 to 17 September, soldiers only — the claim is true, and this page says so plainly rather than defining the difficulty away. The BBC's own 2013 article concedes it: 'the Nizam's forces were defeated after just a few days without any significant loss of civilian lives' ¹¹. But as circulated the claim carries 'cost very little blood', which is a claim about the event and not about the order of battle, and it is answered by a different body of evidence — evidence the Government of India collected itself, from a mission it paid for, and then declined to act on.

What the goodwill mission was, and what it wrote

From October 1948 reports of large-scale killing and looting began reaching Nehru. On 14 November 1948 he wrote a note to the Ministry of States: 'The impression I have gathered from these talks is that while our army is generally believed to have functioned well and to have protected the people, there is little doubt that a very large number of outbreaks took place in the small towns and villages resulting in the massacre of possibly some thousands of Muslims by Hindus, as well as a great deal of looting, etc... This information is contrary to what I had believed and I should like it to be verified through our military and civil authorities in Hyderabad. We must know the truth, or else we shall be caught saying things which are proved to be false later.' ⁶

The mission that resulted was small and short. Its own confidential report to the Prime Minister and the States Minister opens: 'The delegation consisting of Pandit Sundarlal, Kazi Abdul Ghaffar and Moulana Abdulla Misri arrived at Hyderabad on the 29th of November and returned to Delhi on the 21st of December 1948. During this period we toured through 9 out of the 16 districts of the state, visiting 7 district headquarters, 21 towns and 23 important villages. In addition we interviewed over 500 people from 109 such villages as we did not visit.' ¹ That is a tour of just over three weeks. The frequently repeated start date of 19 November comes from transcriptions of an earlier draft in which the numeral is struck through and illegible ². The composition is also not settled: the report names three delegates, while the editor of the Selected Works, quoted by Noorani, records 'A four-man goodwill mission, consisting of Kazi Abdul Ghaffar, Pandit Sundarlal, Moulana Abdulla Misri and Furrukh Sayer Shakeri' ⁶ — and the bundle's own notes state that Shakeri and Mr P. P. Ambulker 'accompanied the mission as its Secretaries' ³, which reconciles the two if Shakeri is counted as staff. The archival title Purushotham gives is 'Communal Frenzy in Hyderabad, 1948. Reports of the Goodwill Mission which Visited Hyderabad to Find out the Damages Done in Hyderabad Owing to the Communal Frenzy and the Possible Remedies for Overriding Future Outbreaks', Pandit Sunder Lal Papers, file 2, 'Nehru Museum and Memorial Library' [sic] ⁴; the familiar 'Report on the Post-Operation Polo Massacres' is Khalidi's 1988 heading, not the file's.

The mission was careful to say what it was not: 'We clarified our position, whenever opportunity presented itself saying that ours was not a Commission of investigation or Inquiry into events preceeding or following the police Action and that ours was merely a goodwill mission charged with the task of restoring better communal relations. All the same, we feel it our duty to bring to your notice what we saw and gathered in our tourings' ¹.

It was also not a document written to indict India, and quoting only its India-adverse passages misrepresents it. Its account of the origins of the conflict lays responsibility on the Nizam's state. On the Razakars: 'The police and the administration, manned mostly by the Muslims, either connived at these atrocities or willingly co-operated with the Razakars in all these lawless activities because they believed that the Razakars alone could save them from the Hindu raj and as such had the support of the Nizam's Government.' On the Ittehad-ul-Muslimin: originally 'a religious & social reform movement', it 'became a mainly political movement, its main objective being opposing the Hindu demand for democratic government. The Razakars were the military wing of the Itte. M.' On demographic engineering: alongside the Razakars 'a new movement was started by the Muslims, which gained a momentum rapidly. It consisted of systemised efforts to induce Muslims from the neighbouring Indian Provinces to migrate to the State and settle there so that in course of time, the small Muslim minority might be converted - if not into a majority - at least into a formidable minority.' And on the arrival of the Indian Army: 'Thousands of Hindus who had in the regime immediately preceding left their hearths & homes either on account of Razakar atrocities or through sheer panic, and had taken refuge in adjoining Indian territories came back to their homes in the state, as soon as they heard of the achievements of the Indian Army. Thousands of others who had continued to live inside the state but had all along been feeling the weight of the Razakar regime welcomed the Indian Army as their saviours and heaved a sigh of relief.' ³ These are the mission's own words and they are evidence on the necessity limb.

Its findings on the killing, in the fair copy: 'Hyderabad State has 16 districts, comprising nearly 22,000 villages. Out of them only three districts remained practically, though not wholly, free of communal trouble which effected the state first during the activities of the Razakars and then during the reprisals that followed the collapse of that organization.' In Osmanabad, Gulbarga, Bidar and Nanded 'the number of people killed during and after the police action was not less, if not more than 18,000. In the other four districts viz. Aurangabad, Bir, Nalgunda and Medak those who lost their lives numbered at least 5 thousand.' And then the sentence that has done all the work since: 'We can say at a very conservative estimate that in the whole state at least 27 thousand to 40 thousand people lost their lives during and after the police action.' ¹ Two scope facts govern that sentence and they pull in opposite directions. It says 'during and after', so it is not a count of post-surrender killing alone. And the notes say: 'When we talk of killed we do not include those who died fighting, but only those murdered in cold blood.' ³ The band excludes combat deaths on both sides. A reader who takes one of these caveats must take the other.

On Latur, Kasim Razvi's home town: 'the killing continued for over twenty days. Out of a population of about ten thousand Muslims there we found barely three thousand still in the town. Over a thousand had been killed and the rest had run away with little else besides their lives and completely ruined financially.' ¹ And on the conduct of the incoming forces: 'Duty also compels us to add that we had absolutely unimpeachable evidence to the effect that there were instances in which men belonging to the Indian Army and also to the local police took part in looting and even other crimes. During our tour we gathered, at not a few places, that soldiers encouraged, persuaded and in a few cases even compelled the Hindu mob to loot Muslim shops and houses.' It adds: 'Complaints of molestation and abduction of girls, against sikh soldiers particularly, were by no means rare.' ¹ In the longer notes the language is harder still: 'At a number of places members of the armed forces brought out Muslim adult males from villages and towns and massacred them in cold blood. At many places they encouraged and even persuaded the Hindu mobs to loot Muslim shops and houses. At some places they even joined in the looting.' ³ The notes also record the disarmament asymmetry: the Army ordered surrender of arms and 'in practice while all arms were taken away from the Muslims, sometimes with the help of the Hindu population, the Hindus from whom the Indian Military had little to fear were left in possession of their arms', and at many places arms taken from Muslims 'were taken either by the Hindus with the connivance of the I.A. or were distributed to the Hindus by the I.A. itself' ³.

The same document exculpates the army as an institution — 'the Indian Army and its officers in Hyderabad generally maintained a high standard of discipline and sense of duty. In General Choudhri we found a man without any tinge of communal prejudice a firm disciplinarian and thorough gentleman' ¹ — and turns on the State Congress: 'In a large number of cases Congress people soon after the police action took the law into their own hands almost as a parallel administration, and not always to their credit', including protection money of which, at one place, 'out of every rupee thus received -/4/- annas went to the man realising the amount, -/4/- to the local police and the remaining -/8/- to the Congress' ³. It also records rescue: 'at places Hindu weavers defended Muslim weavers against Hindu and protected them often at a very heavy cost (including loss of life) to themselves. Many Hindus helped in the recovery of abducted Muslim women.' ¹

An independent chain of custody for the army's conduct

Everything this page says about the conduct of the incoming troops has rested on one bundle — the fair copy and the notes — read through an anonymous archive.org scan whose provenance the page itself flags as unresolved. Two months after the surrender the same finding reached Washington by a route with nothing to do with Sunderlal, the NMML or that scan. On 29 November 1948 the US Chargé cabled a summary of a secret report by the United Kingdom High Commissioner's representative in Hyderabad, 'based on personal observation and conversations with top Indian and Hyderabad officials and UK businessmen': 'According report police action casualties Hyderabad forces were 1400 killed, 700 wounded. Initial entry Indian troops was followed by pillage, arson and murder by Indian forces, especially Sikhs. Disciplinary action soon brought troops under control. Military government is popular with public and is functioning efficiently. Major General Chaudhuri has gained high prestige for impartial handling of all disruptive elements.' ³¹

British observation, American transmission, US State Department publication: a chain of custody wholly independent of the goodwill mission, reaching the mission's conclusion about the conduct of the incoming troops — down to the Sikh-specific detail the fair copy records — within ten weeks. What it removes is a weakness. Every criticism of the mission (a three-week tour, no appended evidence, three numeric states, an unattested scan) previously landed on the whole army-conduct finding, because there was nothing else to land on. There is now something else, and it is a diplomatic report by a third power written for its own government.

The same telegram has to be reported at the same prominence against the counter-narrative. It says the troops were brought under control by discipline, that the military government was popular and efficient, and it praises the Military Governor in almost the terms the mission used of him. It also supplies a third campaign casualty figure — 1,400 Hyderabad forces killed and 700 wounded — which does not map cleanly onto either of the sets given above, because it does not separate the State Forces from the Razakars. The campaign numbers are now attested in three mutually inconsistent official or semi-official states rather than two, and the honest summary of them is unchanged: hundreds to low thousands of combatant dead, Indian military deaths in the tens.

The mission's own arithmetic: three numeric states, and a district ladder

An earlier version of this page said the mission 'declined to set out its evidence' and rested its figure on 'extrapolation from testimony without enumeration'. That was wrong, and the correction cuts against the sceptical reading this page had been inclined to. The bundle's working notes contain a district-by-district derivation, and it is the most methodologically informative thing in the whole file.

Osmanabad first: 'In the town of Osmanabad which is the H.Q. of the district a very reliable gentleman who was 6 months in jail in the previous regime on account of his anti-Razakar activities, at our instance went from house to house and after due inquiry prepared for us a list of persons who had been killed in Osmanabad proper. Rumour runs very high but the list prepared by this gentleman which we are convinced is accurate mentions 725 as murdered and another 190 as missing, many of whom are believed to have been murdered.' For Latur, in the same district, 'Some witnesses told us that the number of Muslims murdered in Latur was somewhere between 2000 and 2500. After examining all the evidence before us we have come to the conclusion that the total killed in Latur proper or in its immediate neighbourhood could in no case have been less than 1600.' For Morum village, the mission recorded both versions — 'according to a Hindu Congress worker the number of dead was nearly 400 and according to the Muslim version nearly 1000' — and a marginal annotation states the rule it applied: 'Hindu and Congress versions have been considered in arriving at our estimated figures.' Its district total: 'somewhere between 5,500 and 10,000.' ³

The ladder continues: Gulbarga 'between 5000 and 8000'; Bidar 'at least as ill if not worse than Gulbarga'; Nanded 'somewhere between 2000 and 4000'; Aurangabad 'somewhere between 2000 to 3000'; Bir 'somewhere between 1000 & 1500'; Nalgonda 'we may safely put the total number of killed as 1000 at least'; Medak, where the mission visited five roadside villages and estimated 30 dead in them and 45 at other places including Medak town, 'can safely be put at 1000' on the reasoning that of nearly 1,500 villages in the district 'very few if any remained unaffected'. It then gives an eight-district subtotal and a statewide statement: the total 'may well have been somewhere between 30,000 & 40,000' and 'could in no way have been less than 25,000' ³. That is a floor in the mission's own hand, and it is not the fair copy's floor.

So there are three numeric states in the bundle, not two, and they do not line up the way the sceptical reading assumes. The draft covering letter gives '25 to 30 thousand' statewide and 'between 3,500 to 4,000' for the second tier of four districts ². The fair copy gives 'at least 27 thousand to 40 thousand' and 'at least 5 thousand' ¹. The notes give 30,000–40,000 with a 25,000 floor ³. Now do the arithmetic on the notes' own ladder: the second-tier districts (Aurangabad 2,000–3,000, Bir 1,000–1,500, Nalgonda 1,000+, Medak 1,000) sum to roughly 5,000–6,500. That matches the fair copy's 'at least 5 thousand' and not the draft's '3,500 to 4,000'. On the evidence in the file, the upward revision between draft and fair copy looks like the covering letter being brought into line with the district workings, not a number being inflated. This page had been treating the variance as damaging to the report; on inspection it is not, and saying so costs the sceptical case something.

What the mission did withhold is witness statements and the underlying lists, which it offered to supply and apparently never did. The right criticism is that the district estimates for six of the eight districts are the mission's judgement on evidence it did not print, not that no derivation exists. And the ranges are wide — Osmanabad alone spans 5,500 to 10,000 — which is a reason to treat the band as an order of magnitude rather than a count, and the mission said as much by giving a band.

One caveat about this section's source text. The notes are legible in the archive.org scan but were read here through its OCR derivative, which is imperfect; the eight-district subtotal in particular is garbled in the OCR and is not quoted. Individual passages above were cross-checked against the independent quotations of the same lines in the BBC's 2013 report, Scroll.in and The India Forum, which agree. The typescript itself should be collated by anyone relying on the district figures.

Every figure tested by the same rule

This section states the standard and then applies it to both sides, because the temptation on a question like this is to run demolition arithmetic on the numbers one dislikes and let the others pass. The rule used here is: a death figure is only as good as the enumeration behind it, and any figure must survive a ratio test against the population it claims to have killed. The same rule is applied to figures for Muslims killed in the reprisals and to figures for Hindus killed by the Razakars, and where a figure fails the test this page says so regardless of which side it helps.

The population base first, because everything else depends on it. Hyderabad State was, on the account of the 2024 anthology's reviewer, 'a state where 88% of the population was of the Hindu faith' ¹³, leaving Muslims at roughly 12 per cent. The state's total population at the 1941 census is generally given as about 16.3 million, a figure this page could not check against the census volumes themselves and which is therefore flagged in the open questions. On that base the Muslim population is of the order of 1.9 to 2 million, of whom perhaps a million were male.

Run the test. Chaudhuri's 'about 2,000' is roughly 0.1 per cent of Hyderabad's Muslims. The mission's floor of 25,000 is about 1.3 per cent and its ceiling of 40,000 about 2 per cent. Smith's reported 50,000 is about 2.5 per cent. Khalidi's 200,000 is about 10 per cent. And Margrit Pernau's 'one-tenth to one-fifth of the male Muslim population' ⁸ implies roughly 100,000 to 200,000 dead — of the same order as Khalidi's figure and three to eight times the mission's band. An earlier version of this page waved Pernau off as 'a proportion rather than a count' and left it there while dismantling Smith. That was an asymmetry and it is corrected here: stated as an absolute, Pernau's estimate is irreconcilable with the mission's own arithmetic, and the two cannot both be right. Nothing in the file supports a figure at that level.

Now the same test on the other side. No published figure for deaths caused by the Razakars and the Nizam's forces survives it either, because none carries an enumeration at all. Munshi's '260 events of brutality' and the 1948 White Paper's '53 events' are incident indices without totals ¹⁶; the Gorata figure of 'over 200' is a single-village report ¹⁷; the figures of 880 murders and about 1,125 rapes that circulate in the Indian press could not be traced to a primary record for this page. Third-party corroboration that Razakar violence was real and serious now exists with no Indian government provenance — General Bucher told the US Chargé on 20 September 1948 that 'American missionaries in Bidar Hyderabad state had some harrowing stories to tell' ³⁷ — but stories are not a count, and the addition of that document does not move this side of the ledger by one death. The mission's notes contain one arresting datum on this side, though the OCR of the line is imperfect and it should be checked against the typescript: that the State Congress movement 'was admittedly a violent movement. They used arms and ammunition, explosives and dynamite', and that its own secretary's official report admitted over 2,000 deaths resulting from it ³. Symmetry here means an uncomfortable result for both sides: the anti-Muslim killing of 1948 has a documented derivation with a floor, and the Razakar killing that preceded it does not — which is a fact about who did the counting, not about who did the killing, and the proportionality claim that Chaudhuri, Patel and Munshi all made cannot be assessed until somebody counts.

The number moved between drafts, and no estimate is independent of the others

There is no reliable count of the dead and there will not be one. What exists is a set of estimates with very different bases, and one of them is not even internally stable.

Wilfred Cantwell Smith's much-cited 50,000 is not an independent count and Smith did not claim it was. Writing in the Middle East Journal in 1950 he said: 'The only careful report on what happened in this period was made a few months later by investigators — including a Congress Muslim and a sympathetic and admired Hindu — commissioned by the Indian Government to study the situation. The report was submitted but has not been published; presumably it makes unpleasant reading. It is widely held that the figure mentioned therein for the number of Muslims massacred is 50,000. Other estimates by responsible observers run as high as 200,000, and by some of the Muslims themselves still higher.' ⁷ That single passage is the origin of both the 50,000 that scholars quote as conservative and the 200,000 that circulates on the other side. Smith was reporting rumours about a report he had not seen, and the rumoured 50,000 does not match what the report actually says. Citing Smith and the mission as two corroborating sources double-counts one archive — a point Sherman makes explicitly when she gives 50,000 and notes that 'Smith bases this figure on a report commissioned but never published by the Indian Government' ⁵. The chain outlives its origin: Saksena's note 70 in a peer-reviewed 2020 article still repeats the range 'from 50,000 to 200,000', which is Smith's rumour at three removes ²⁵.

The higher figures have no separate base that anyone has produced. Omar Khalidi, who first printed fragments of the report, elsewhere concluded that 'nearly 200,000 Muslims were butchered in an orgy of violence directed by the Congress Party gangs'; the reviewer's own comment is that Khalidi 'uses certain unidentified sources to arrive at the far higher estimate' ¹³. Qutubuddin Aziz's 'The Murder of a State' (1993) carries a comparable figure; Sherman's note is exact about the problem: 'Aziz mentions Sunderlal and Abdulghaffar's report in connection with this figure (p.199-202), but the version of their report cited in this article does not calculate or estimate the total number killed.' ^{5 19} Pernau's estimate is discussed above and does not survive the ratio test ⁸.

Against all of these stands the government's own number, which is startlingly low and which nobody defending the peaceful reading tends to quote either, because it concedes the substance. Military Governor Chaudhuri, having admitted that 'immediately after the military occupation… a large majority of Hindus in all districts of Hyderabad thought that Hindu Raj had come into being' and that 'as a result of this retaliatory action by Hindus, some Muslims in certain districts suffered', concluded: 'It is estimated that about 2,000 Muslims may have been massacred.' ¹⁵ At the June 1949 amnesty the administration's own tally of 'cases arising out of retaliation of Police action' stood at 875, including 114 murders, involving 3,031 accused ⁴. Two thousand dead is not peaceful either.

The most useful figures may be the administrative ones, because they were produced for spending money rather than for argument. A 1949 Hyderabad government publication put the number of Muslim women widowed after the police action at 'at least' 3,375; the government-appointed Rehabilitation Committee reported in 1951 that in 47 villages of Osmanabad district alone there were approximately 2,500 widows and 6,000 orphans, and recommended planning on approximately 10,000 widows and 25,000 orphans for that district ¹⁵. Widows and orphans are a shadow of a death toll, but they are a shadow cast by an administration counting its own liabilities.

Not a sixty-five-year secret: the access history, stated precisely

The strongest version of the suppression claim is that the report was buried until the BBC unearthed it in 2013. It was not, and the inaccuracy matters, because the accurate version is more damaging.

What happened first was repudiation, not concealment. Kazi Ghaffar sent Patel a copy on 3 January 1949, after Patel had refused to meet the delegation ¹⁵. Patel's reply of 4 January denied the mission had any standing: 'I notice that in your report you mentioned that you were asked by the Government of India to proceed to Hyderabad State on a goodwill mission. At least I am not aware of any such mission having been entrusted to you by the Government of India. As far as I know, you wanted to go there and it was arranged that you should go there at Government expense.' He went on: 'the detailed inquiries which have been made by the local administration over a fairly long period as opposed to the roving enquiries which you have made during such a short period show that your estimate and your appreciation of the position lack balance and proportion.' ^{6 4} Noorani's comment is that the assertions 'were simply untrue', since nobody could have toured the state without official approval and the team went at government expense ⁶.

After that the document did not vanish; it simply had no official existence. It was never published, never entered into any government account of September 1948, and produced no prosecutions of soldiers. But the Sunderlal papers were donated to the Nehru Memorial Museum and Library by his heirs in 1968 ¹². Khalidi printed fragments in 1988 — fragments Noorani judged unreliable, noting that Khalidi had been told the document was partly in Urdu when 'the entire document is in English' ⁶. Noorani obtained the full text from the NMML and published it with a companion essay in Frontline in March 2001, where it has been online ever since ⁶. Sherman cited it in the Indian Economic and Social History Review in 2007 ⁵. Mike Thomson's BBC report of 24 September 2013 did what none of these had done — it reached a mass audience — and it said plainly that 'The Sunderlal report, although unknown to many, is now open for viewing at the Nehru Memorial Museum and Library in New Delhi' ¹¹.

The report's standing was publicly challenged after that coverage, most fully by Tejasvi Surya in an IndiaFacts essay of 6 November 2013, on several grounds. (a) The mission was a goodwill delegation and not a judicial commission. This is correct, and it is the mission's own statement about itself ^{1 21}. (b) Contemporary newspapers carried nothing on this scale. Weaker than it looks but not empty: Purushotham, hunting for corroboration of the Gorata killings, found only a Times of India report of 19 May 1948 quoting a Hyderabad government communiqué denying them, and observed that an event of that scale should have reached the nationalist press ¹⁷ — a press that failed to report Razakar massacres is not a press whose silence proves anything about reprisals. The standard has to be the same in both directions and here it defeats the objection; and the objection is now weakened from a further direction that has nothing to do with the press or the mission, since a secret UK High Commission report summarised to Washington on 29 November 1948 independently recorded 'pillage, arson and murder by Indian forces, especially Sikhs' ³¹. (c) A Right to Information query, filed by Capt. Pandu Ranga Reddy, was told the museum did not hold a copy. That is a fact about one query, against Simeon's account of the 1968 donation and Noorani's account of obtaining the text from the same library ^{12 6}. (d) The most-quoted atrocity imagery — wells full of corpses, charred bones — is not in the confidential notes. This is the essay's most specific and most checkable allegation, and it is false. The notes read: 'In many places we were shown wells or Bawaries still full of corpses that were rotting. In one such we counted 11 bodies which included that of a woman with a small child sticking to her breasts... We saw remnants of corpses lying in ditches. At several places the bodies had been burnt and we would see the charred bones and skulls still lying there.' ³ The BBC quotes the same passage and attributes it correctly to the confidential notes ¹¹. (e) The 200,000 figure is demographically implausible. This is sound, and this page reaches the same conclusion by the same route above — which is why the objection is reported here rather than dismissed.

The more serious methodological objections come from inside the scholarship. The mission's own disclaimer and its refusal to append its evidence ¹; the width of its district ranges ³; the movement of the figure between drafts ²; and the textual state of the document itself. Sherman's remark that 'the original was suppressed and only scraps of it remain' is often quoted as an independent archival finding ⁵; it is not. Her footnote to that sentence points to Khalidi's 1988 extracts, so she is describing what Khalidi printed. Separately she cites the report at page 100 for a sentence that does not appear anywhere in the bundle used here — 'In general the attitude of the military officers was good but the soldiers showed bigotry and hatred' — which means a text paginated to at least a hundred pages was in front of her in 2007. At least four textual states are therefore attested: the draft covering letter, the fair copy, the separately paginated notes, and Sherman's hundred-page text. The bundle relied on here runs to about nineteen images. This page does not have the whole document either, and says so.

The remedy the Standstill Agreement itself provided, and which nobody used

One thing the necessity case has to be stated against, and this page previously omitted it. The Standstill Agreement is invoked on both limbs — India's accusations under it are set out above — but the page never said what the Agreement provided for when it broke down. On Saksena's account the agreement of November 1947 'retained pre-independence arrangements on matters of common concern (defence, external affairs, and communications), precluded India from sending troops to the state, and provided for arbitration in the case of disputes' ²⁵; and it was 'After India refused Hyderabad's request to refer the dispute to arbitration' that the Nizam's advisers, Walter Monckton and Hersch Lauterpacht among them, 'decided to approach the United Nations' ²⁶.

Two provenance cautions travel with that sentence and the page states them rather than borrowing the authority. Saksena is peer-reviewed and open access, but her footnote for the Agreement's terms (n. 60) is to A. G. Noorani, 'The Destruction of Hyderabad', p. 173 — the same advocate this page discounts elsewhere — and neither she nor this page has read the Agreement's text. On the terms, therefore, an A-tier author is reporting a B-tier reading.

The refusal itself is attested from three directions, and one of them is disinterested. Hyderabad's Agent-General told the US Chargé on 21 August 1948 that 'GOI refusal to arbitrate any of the points in dispute made it necessary for Hyderabad to appeal to UN' — an interested statement by a party ³². Pakistan told the Security Council on 19 May 1949 that 'the Nizam had offered to arbitrate any differences arising over the working of the Standstill Agreement concluded on 29 November 1947' — equally interested ⁴³. And on 24 August 1948 the Chargé recorded his own judgement, which is not: 'Embassy does not believe GOI would under any circumstances consent to arbitration, so point raised by Razvi is not a live issue.' ³³ (The immediate question was whether the United States might arbitrate; the sentence is nonetheless general, and it is the Embassy's own view rather than a party's.)

Two consequences follow, and they are smaller than they look. An agreement whose own dispute-resolution mechanism the aggrieved party declined to use is weaker evidence of the other party's bad faith than the phrase 'the Standstill Agreement's breakdown' implies. And if the Agreement barred Indian troops from the state, the September operation was on its face a breach of it too. Neither shows the disorder was fabricated, and neither touches the Nanaj ambush or the loss of control over the Razakars. What they mean is that the Agreement cannot be entered on one side of the ledger only, which is what this page had been doing.

The date is not explained by the disorder

Hyderabad's complaint reached the Security Council by cablegram of 21 August 1948, and the Council admitted it to the agenda at the 357th meeting in Paris on 16 September, 'after several representatives had made the reservation that this action did not prejudge the Council's competence or any of the merits of the case' ^{9 39}. Indian columns crossed the frontier on 13 September. Saksena states the sequence flatly: 'On 13 September, before the complaint was put on the Security Council's agenda, Indian troops invaded Hyderabad.' ²⁷ The page already carried the downstream half of this — Washington's expectation, four days after the surrender, that India would 'be able to present SC and world on Sept 20 with fait accompli' ¹⁰ — without the upstream half.

The argument that the United Nations calendar rather than any change of conditions inside Hyderabad set the date is Sunil Purushotham's, developed in 'From Raj to Republic' (Stanford, 2021), chapter 1. That chapter could not be opened for this page — Stanford University Press's site refuses automated requests — so the argument is attributed to him but not quoted ⁴⁷. It does not depend on him: the dates are in the Council's own record.

Nor does it show the disorder was manufactured, or the necessity case empty. It shows that the disorder does not date the operation. Alongside it stands the US Chargé's report of 12 September that the Cabinet Defence Committee would take the final decision that day, and that the 'Decision will be political not military as army is convinced its ability carry out campaign successfully but has pointed out certain risks to government, particularly possible intensification military pressure from GOP in Kashmir' ³⁴.

The plebiscite that everyone proposed and nobody held

The counterfactual has more evidence in it than this page previously showed, and it is contested at every joint. A plebiscite was proposed at four separate moments and never held.

India told the Council at its 360th meeting on 28 September 1948 that 'as early as August 1947, the Indian Government had suggested a plebiscite on the issue of Hyderabad's accession, but the Hyderabad Government had rejected that proposal' ⁴¹. Pakistan told the Council at its 426th meeting on 24 May 1949 that in the June 1948 negotiations India had put three alternatives — accession, immediate responsible government on lines determined by India, or 'a general plebiscite under neutral observation' — and that Hyderabad accepted the third: 'The third alternative was accepted provided the plebiscite was held under the supervision of an independent international body such as the United Nations or the International Court of Justice.' ⁴⁴ That is Pakistan's account, given in advocacy against India, of a negotiation to which Pakistan was not a party, and the Council never adjudicated it; it is reported here at exactly the discount this page applies to the 1948 White Paper, and the underlying drafts have not been read by anyone on this page. Hyderabad's own representative offered a UN-supervised plebiscite at the 357th meeting on 16 September, three days after the invasion began and, on India's account, having refused one in 1947 ³⁹. And on 14 September the State Department instructed its Embassy that India 'might wish consider as first step in this direction making immediate announcement that it would invite UN to undertake, when conditions permit, supervision of free plebiscite Hyderabad' ³⁵; V. P. Menon's response on 17 September was that India 'would never on principle permit UN interference, but GOI might on own initiative invite UN observer plebiscite when held' ³⁶. The Embassy's own conclusion a week later was that 'the question of a plebiscite in Hyderabad is immaterial since the result would be a foregone conclusion' ³⁰.

A settlement mechanism existed, was proposed by every party at some moment, and was refused by every party at some moment. That establishes that the counterfactual had a shape. It does not cost it, and the page's tenth open question stands untouched.

Necessity: the best case on each side

None of the above touches the second limb of the claim, and it is the limb where the scholarship genuinely divides. What follows sets out each case at comparable length.

The case for necessity, stated at its strongest and drawn mostly from a scholar unsympathetic to Delhi's official account. Under the Standstill Agreement India charged Hyderabad with 'divesting itself of its Indian securities, banning the Indian currency, halting the export of ground nuts, organising illegal gun-running from Pakistan, and inviting new recruits to its army and to its irregular forces, the Razakars' — read in Delhi as a 'comprehensive plan to break up the economic cohesion of India' ⁵. The Razakars raided 'trains and villages in Madras, the Central Provinces (CP) and Bombay', and in July 1948 killed six Indian Army soldiers in an ambush near the Indian enclave of Nanaj; Delhi concluded that the Nizam 'had lost control over his own territory' ⁵. Negotiations collapsed in June 1948 ¹⁴. Refugee movement made the state an Indian domestic problem: Purushotham gives 'an estimated 5,00,000 "non-Muslims"' leaving Hyderabad and 'another 7,00,000 Indian Muslims' entering it in the thirteen months after August 1947 ¹⁴, and Sherman notes that 'Some estimates put the number of refugees at 40,000 in CP alone' ⁵. Nehru regarded the Razakar problem as the one before which all other issues were 'relatively secondary' ⁵. Internally, a state 88 per cent Hindu had no representative institutions and no path to them: the mission itself recorded that the Ittehad's 'main objective' had become 'opposing the Hindu demand for democratic government', that the Razakars were its military wing, and that the Nizam's own police and administration 'either connived at these atrocities or willingly co-operated with the Razakars' ³. And the operation had substantial Indian Muslim support at the time, with no communal disturbance in India while it ran ⁵. The strongest single item on this side is not Indian at all: the US Embassy's despatch of 24 September 1948, that the Nizam's government had lost control of civil administration in districts adjoining Madras and that entering Hyderabad 'was the only possible remedy for an impossible situation' ³⁰. That is a real case, and it does not depend on the White Paper.

The case against, stated at its strongest. The outcome was treated as settled before the grievances arose: as early as June 1947 Nehru warned that he would 'encourage rebellion in all states that go against us', and Patel's view was that 'Hyderabad is, as it were, situated in India's belly. How can the belly breathe if it is cut off from the main body?' ⁵ The Standstill Agreement provided for arbitration and Hyderabad asked for it; India refused, and the US Embassy privately expected that it always would ^{25 26 33}. The operation was launched on 13 September, three days before the Security Council took Hyderabad's complaint onto its agenda ^{27 39}. The disorder Delhi invoked was in part manufactured, and the page's own primary source says so: the mission recorded that the State Congress, 'suppressed inside... organised their centres in places within the Indian border outside the state, and from there attacked Octroi posts and Police Chowkies inside the State', and that the movement 'was admittedly a violent movement. They used arms and ammunition, explosives and dynamite' ³. Purushotham's archival work covers the same ground — his notes cite a 'List of Camps of the Hyderabad State Congress' and a letter from Munshi to Patel of 21 May 1948 ⁴ — and his EPW review gives the specifics (79 camps, at least 1,600 men, 3,000 State Congress cadets serving as guides, and Munshi writing that the 'border incidents only provide justification' for intervention). Those specifics could not be confirmed at any location this page could open and are reported as his, unverified here ¹⁴. Sherman's own formulation of the same point is more guarded and is quoted as she wrote it: 'To a limited extent, Congress volunteers engaged in satyagraha had contributed to the internal disorder by disrupting courts, filling jails, and engaging in sabotage.' ⁵ A law-and-order case built partly on disorder one has helped generate is not thereby false, but it is not the innocent case the official account presents.

The international limb collapsed under occupation rather than on its merits alone. No state recognised Hyderabad's independence and the Security Council never granted it relief — but Hyderabad's complaint was admitted to the Council's agenda at its 357th meeting in Paris on 16 September 1948, where its representative 'urged that the situation demanded immediate action by the Security Council, not only under Chapter VI of the Charter, but also under Articles 39 and 40'. What was argued there is recoverable from the Yearbook of the United Nations, which summarises the meetings the Repertoire only indexes, and it does not support the sharper version of this charge. India's representative 'maintained that Hyderabad was not competent to bring any question before the Security Council, that it was not a State and was not independent'; but the same summary records that, 'Without going into the merits of the case, he indicated the nature of the reign of terror by private armies in Hyderabad, and other events which, he said, had obliged the Government of India to take action' ³⁹. India's primary and repeated position was that the forum had no competence — at the 425th meeting on 19 May 1949 B. N. Rau held that Hyderabad 'had never been a State in the sense of international law and could never be one in the future if India was to live', and pressed for the item's removal as wholly within India's domestic jurisdiction ⁴² — but it is not true that India declined to state its grounds. It stated them, before a body whose right to weigh them it denied. Hyderabad's own counsel privately expected the question to be settled that way: Lauterpacht wrote that 'if there is a disposition on the part of the members of the Security Council or of the General Assembly to assist Hyderabad and to restrain India, then they will have no difficulty in finding that we are a state. If the tendency is the other way, then they may shield themselves behind the finding that we are not a state.' ⁴⁶ On the legal limb neither party expected law to do the work. By cablegram of 22 September the Nizam informed the Secretary-General that he had withdrawn the complaint and that his delegation 'had ceased to have any authority to represent him or his State'. At the 359th and 360th meetings the Council debated three questions, of which the second was 'whether the withdrawal of the case by the Nizam of Hyderabad had been made voluntarily or under duress' ⁹. Washington's reading, four days after the surrender, was that India would 'be able to present SC and world on Sept 20 with fait accompli', and the State Department instructed its delegation to resist dropping the item because that 'would create dangerous precedent by encouraging use of force by aggressive states' ¹⁰.

The counterfactual, which the necessity claim requires and which nobody has costed. 'Necessary' is comparative: it asks what non-intervention would have cost. The materials for the question exist in the sources used here and the answer does not. On one side of the ledger, a state whose own police connived at militia violence ³, cross-border raiding into three Indian provinces and the Nanaj ambush ⁵, and a communist-led peasant insurrection in Telangana that had already 'crippled' the state ⁵ and would take the Indian Army three further years to suppress. On the other, the negotiated alternatives that were live until June 1948 ¹⁴, the arbitration the Standstill Agreement itself provided for and India declined ^{26 33}, the plebiscite proposed and refused by every party in turn ^{41 39 35 36}, and the fact that the Nizam accepted accession after the surrender and was retained as Rajpramukh ⁵ — which shows a settlement of that shape was available to be reached. What either trajectory would plausibly have cost in lives has not been modelled by anyone this page could read, and until it is, 'necessary' remains an assertion on both sides rather than a finding.

The occupation that followed is where the claim is most easily tested, because the administrative record survives where the burial record does not. By the authorities' own figures the military and police detained over 13,000 Muslims plus several hundred Arabs and Pathans, jailed several thousand Hindus for reprisals, and imprisoned an estimated 17,550 people in all; in the six months after the surrender the government 'released over 11,000 Muslims without trial because no incriminating evidence against them existed', and deported some 2,000 Arabs to Aden and a similar number of Pathans elsewhere in India ⁵. Those numbers do not reconcile — several hundred Arabs and Pathans detained, some four thousand removed — and Sherman's sources leave the discrepancy standing. A Special Courts Order relaxed evidentiary standards, blocked delay by hunger strike, and initially provided no right of appeal; appeals to the High Court for major offences were restored only in October 1949 ⁵. Over a hundred officers were dismissed, including the Chief Secretary, the Director General of Police, 16 Collectors, 16 Superintendents of Police and almost 100 tehsildars, and, in Sherman's words, 'They adopted a policy of not hiring new Muslims in the services. The civilian administration under Vellodi continued this policy.' ⁵ The mission had already noticed the beginning of it: 'Many Hindus lower in service had been appointed to higher posts by supersession over a number of Muslims higher up in service. Many Muslims had been suspended on mere suspicion.' ³

Nor did the violence stop with the Razakars. The army that entered to protect Hindu villagers from the Razakars spent the next three years fighting the communist-led peasant movement in Telangana; Nehru instructed Chaudhuri that the fighters there should be called not communists but terrorists, on the ground that 'too much talk of communists confuses the issue because communists in other countries function differently' ⁵. Behind that campaign lay the jagirdari system, which governed 43 per cent of the land in the state ⁵. Figures for the number of villages from which landlords were driven circulate widely and derive from the CPI literature — Noorani cites P. Sundarayya's 'Telengana People's Struggle and its Lessons' (1972) at pages 88–89 for the 'untold miseries' inflicted on 'the ordinary Muslim people' ⁶ — but no such count was traceable to a source this page could open, and the figure of 3,000 villages that appeared in an earlier version of this page has been withdrawn. Whatever else 1948 was, it was not a restoration of order followed by quiet.

One more thing has to be argued rather than assumed: whether the operation caused what followed. The mission attributes the reprisals to Hindu mobs, returning refugees, an outside Hindu organisation, communists and State Congress cadres — agents the Indian Army was in part deployed against. The causal link is nonetheless in the mission's own text. The infiltrators arrived 'in the wake of the Indian Army' ¹. The returning refugees came back 'as soon as they heard of the achievements of the Indian Army', 'more or less saturated with the communal venom' ³. The Army disarmed Muslims and left Hindus armed, and in places redistributed the confiscated weapons to Hindus ³. Soldiers 'encouraged, persuaded and in a few cases even compelled the Hindu mob' ¹. And the timing is not separable: the fair copy counts the dead 'during and after the police action', and the Latur killing ran twenty days from the collapse ^{1 3}. An operation that removes a state's coercive apparatus in five days and installs nothing in its place for weeks owns the resulting vacuum. That is an argument about responsibility, not identity, and it is the argument on which the wider reading of 'peaceful' fails.

Where historians divide is on what question this material answers. For Purushotham the axis of 'necessity' is the wrong one: the violence was 'an important engine of state formation in India in the period following independence', and the incorporation of Hyderabad's Muslims into the postcolonial order was mediated through it ⁴. For Noorani a negotiated settlement remained available and the responsibility is Patel's; his book, which reproduces the report in full, concludes that the police action 'was much more than conquest of a rebel state. It was an annihilation of a certain way of life, the uprooting of a people, and the sweeping away of a culture, swiftly and almost completely. None of this was necessary in the interest of India's security or national integrity.' ^{15 22} For Saksena and Pillai the operation was an invasion and the legal argument instrumental on both sides ^{25 46}. For Sherman the interesting fact is that the early Indian state in Hyderabad was neither monolithic nor in control, which is why the arrests were indiscriminate and the prosecutions few ⁵. And for the contributors to the 2024 anthology '1948 Hyderabad Readings' the polarisation between accession and independence 'was not reflective of the Hindu-Muslim divide' at all — Muslims held two-thirds of administrative posts in a state 88 per cent Hindu, over 80 per cent of rentier-landowners were Hindus, and village office was almost exclusively Hindu ¹³, while the Depressed Classes Association and Depressed Classes Conference had sided with the Nizam against accession ⁵. None of these positions rescues the word 'peaceful' as applied to the transfer of power. Several of them leave 'necessary' standing as a real question.