In 1858 the Governor of Bombay Wrote That 'Divide et Impera' Should Be Britain's Motto in India. By 1909 the Census Had Counted 2,378 Castes and Tribes and Council Seats Were Reserved by Religion

On 14 May 1858 the Governor of Bombay, Lord Elphinstone, signed a minute on how the army should be rebuilt after the rising of the year before. It was later printed in the appendix to the evidence taken by the Peel Commission, the royal commission appointed that summer to decide the future of India’s army. “Divide et impera was the old Roman motto,” Elphinstone wrote, “and it should be ours.”

“The safety of the great iron steamers … is greatly increased by building them in compartments,” he wrote, and he proposed to build the native army the same way, using “those diversities of race and language which we find ready to our hands.”

The commission recommended the opposite, and over the next seventy years the officials who sorted Indians into regiments, castes and electorates mostly gave other reasons. The authors of the constitutional report of 1918 condemned separate electorates in one paragraph and kept them in the next. Our shorter piece on how colonial rule designed India’s divide gives the outline, and the piece on British policy and Jinnah covers the Simla deputation of 1906. This one goes through the papers of three systems, the recruiting lists, the census tables and the electoral regulations, and sets out where historians disagree about who wanted what.

The commission that recommended mixing

The Peel Commission sat in London under General Jonathan Peel, the Secretary of State for War. It heard 47 witnesses between August and December 1858, and the historian David Omissi notes that no Indians were asked to take part. Witness after witness was asked whether the Bengal native army should be composed of different nationalities and castes, and in what proportion. Colonel Becher answered, “I think that it should be a mixed body, without reference to caste in any way.” Sir Charles Trevelyan wanted native troops kept as “light-armed auxiliaries” to the European force and dressed in their own fashion, which would “preserve in a much greater degree the distinction between the different nationalities and races, which is a point very necessary to be attended to for our security.”

A committee of officers in the Punjab under Sir John Lawrence argued the other way. Its replies reached the appendix in a summary by Lieutenant-Colonel Henry Durand, who was on special duty with the Governor-General, Lord Canning. The committee said that “our main object is to make that army safe; and next to the grand counterpoise of a sufficient European force, comes the counterpoise of natives against natives.” It had weighed mixing every race in every regiment and rejected it, because different races mixed together “do not long preserve their distinctiveness.”

Canning’s own memorandum gave the figures. The Bengal army, counting its militarily organised police, came to about 130,000 men, of whom “about 75,000 are Punjabees,” and he thought probably 23,000 of those were Sikhs. “The difference, therefore, between our position in 1858 and our position in 1857,” he wrote, “is, that there is now a larger native force, and that the bulk of it, instead of being drawn from Oude, is drawn from the Punjab.” Oude is Awadh, the main recruiting ground of the old Bengal army. The commission itself said only, in the words Omissi quotes from its report of March 1859, that the native army “should be composed of different nationalities and castes, mixed promiscuously through each regiment.”

A Bengal army rebuilt from the Punjab

Omissi puts the Bengal army at about 120,000 men in 1856, with about three-quarters of its native infantry, the sepoys, high-caste men from Awadh. After the rising the British kept the regiments that had not mutinied, added enough of the Punjabi and low-caste units raised in 1857 and 1858 to make up the numbers, and disbanded the rest. Lord Ripon, the Viceroy, said in 1881 that the result was “purely accidental” and “a mere fortuitous congeries of regiments, raised in haste, brought together in haste, reduced in haste.”

Omissi agrees for the 1860s. Decisions were improvised in an emergency, he writes, and divide and rule was invoked afterwards to explain the shape the army had taken. Old hostilities were also real. A British major-general noted: “It was not because they loved us, but because they hated Hindustan that the Sikhs have flocked to our standards.”

Roberts and the regiments he disbanded

The policy changed in the 1880s. Lord Roberts, Commander-in-Chief in India from 1885 to 1893, wrote in his memoir of 1897 that the first step was “to substitute men of the more warlike and hardy races for the Hindustani sepoys of Bengal, the Tamils and Telagus of Madras, and the so-called Mahrattas of Bombay.” The obstacle, he said, was the old theory that an equilibrium had to be kept between the armies of the three presidencies, the British provinces of Bengal, Madras and Bombay. He held that “no comparison can be made between the martial value of a regiment recruited amongst the Gurkhas of Nepal or the warlike races of northern India, and of one recruited from the effeminate peoples of the south.” He disbanded companies and regiments “composed of doubtful material,” took on men of “well-known fighting castes,” and formed class regiments, which he judged “more congenial to the men and more conducive to esprit de corps.”

Omissi’s reading is that the immediate cause was Russia. Its advance in Central Asia made war on the north-west frontier look likely, and Roberts judged that danger greater than a mutiny. In Omissi’s table of infantry battalions by region of recruitment, 28 of 131 came from the Punjab and the North-West Frontier in 1862, and 40 from Madras. By 1914 it was 57 of 121 from the Punjab and the Frontier and 11 from Madras, figures he calls approximate.

He also separates the evidence from the prejudice. Some Madras regiments did badly in the Third Burma War, but Madras units had seen almost no fighting since 1857, and Omissi concludes that “the historian must seek social, not biological, explanations.” The theory itself, he writes, was “not a pure figment of the colonial imagination”, since Sikhs, Rajputs and Gurkhas had martial traditions of their own, and it served as “a useful mobilizing myth” for officers who needed to trust their men.

Risley’s 2,378 castes

The census was a second system, and Herbert Risley, the Census Commissioner of 1901, wrote its scheme. In his chapter of the report he called the scheme of 1891 “a patchwork classification in which occupation predominates” and proposed “classification by social precedence as recognised by native public opinion.” The tests were concrete. Would Brahmans take water from a caste, would Brahmans of high standing serve it, could its members use the village well, did they have to leave the road when a high-caste man approached? Table XIII, the alphabetical list of every group returned, “includes 2,378 main castes and tribes and 43 races or nationalities,” Risley wrote.

He also wanted a physical basis for rank. In The People of India (1908) he wrote that in the parts of India “where there is an appreciable strain of Dravidian blood it is scarcely a paradox to lay down, as a law of the caste organization, that the social status of the members of a particular group varies in inverse ratio to the mean relative width of their noses.” His example was Bihar and the United Provinces, where he ran the scale from the casteless tribes through Musahars and Chamars up to the Khatris and Babhans.

Colleagues did not accept it. In the report on the census of 1911, E. A. Gait noted that Risley’s conclusions “have been called in question by Crooke in the United Provinces, Enthoven in Bombay, and Thurston in Madras.” Gait added that whole communities which raised their social rank were “disturbing the correlation between caste and status which Risley alleged to exist.”

The Khatris of Bareilly write back

Risley counted the petitions as proof that his method worked. “The best evidence of the general success of the experiment,” he wrote, “is the great number of petitions and memorials to which it gave rise.” The most elaborate came from the Khatris of the Punjab and the United Provinces, a trading caste. One provincial superintendent had provisionally classed them as Vaisyas, the merchant rank, although the model table Risley circulated put them with the Rajputs. A meeting of protest was held at Bareilly, and a great array of authorities was marshalled to show that the Khatris descended from the Kshatriyas, the warrior rank. Risley compared it to modern Greeks claiming descent from Achilles, then ordered the Khatris classed as Kshatriyas. “The decision gave general satisfaction,” he wrote.

Gait found the same pressure ten years on. “At every census,” he wrote, “numerous castes come forward claiming to be ranked in one or other of the four main classes,” meaning to raise their status in the hierarchy of caste. The anthropologist Bernard Cohn read that as a puzzle of cause and effect. The ranks printed in census reports, he wrote, “became the equivalent of traditional copper-plate grants declaring the status, rank and privileges” of a caste, but caste associations, new marriage networks and caste hostels at colleges were at work in the same years, and the answer “is not a simple one.” Susan Bayly is firmer against the idea of an imposed invention. It was “certainly not a case,” she writes, “of fabricated ethnographic findings being somehow imposed on Indians.” Indians did have reason to take the listings seriously, she adds, when officials used them to decide which communities were “manly” enough to supply recruits.

How much of this reached policy is also disputed. C. J. Fuller argues that the anthropological knowledge Risley and Denzil Ibbetson produced about caste had limited influence on the partition of Bengal and the Morley-Minto reforms. By 1931 the Census Commissioner, J. H. Hutton, was recording that the Hindu Mahasabha had campaigned for as many people as possible to be returned as Hindus, that attempts had been made to minimise the numbers of the Depressed Classes, and that depressed-class organisations had claimed castes “almost certainly not in that category.”

Five seats in 1909

The third system was written in regulations. The Government of India’s notification of 15 November 1909, issued under the Indian Councils Act, set the Imperial Legislative Council at about 60 additional members, of whom at least 25 were elected. Five of those 25 were elected by Muslim electorates, one each for Madras, Bombay, Bengal, the United Provinces and Eastern Bengal and Assam. Six more were elected by landholders and two by chambers of commerce.

The qualifications were steep. In the United Provinces a Muslim could vote for the Imperial Council seat only if he owned land paying 10,000 rupees a year in revenue, paid income-tax on 10,000 rupees of income, sat on the provincial council, or met one of several other tests, among them being a Fellow of the University of Allahabad or a trustee of the Aligarh college. The Simla deputation of 1906 and Jinnah’s part in the Lucknow Pact of 1916, which accepted separate electorates for Muslims, are told elsewhere here.

Paragraph 229 and paragraph 231

Edwin Montagu, the Secretary of State for India, and Lord Chelmsford, the Viceroy, signed the constitutional report of 1918 and spent several paragraphs against communal electorates. “Division by creeds and classes,” paragraph 229 says, “means the creation of political camps organised against each other, and teaches men to think as partisans and not as citizens.” It goes on: “The British Government is often accused of dividing men in order to govern them. But if it unnecessarily divides them at the very moment when it professes to start them on the road to governing themselves, it will find it difficult to meet the charge of being hypocritical or short sighted.”

Paragraph 231 then keeps the Muslim electorates of 1909, because Hindu acquiescence was “embodied in the present agreement between the political leaders of the two communities” and Muslims regarded the arrangement as settled. “Much as we regret the necessity,” the authors wrote, the system “must be maintained until conditions alter, even at the price of slower progress towards the realization of a common citizenship.” In paragraph 232 they refused to extend the system to the other communities that had asked, with one exception. “The Sikhs in the Punjab are a distinct and important people: they supply a gallant and valuable element to the Indian Army; but they are everywhere in a minority, and experience has shown that they go virtually unrepresented. To the Sikhs therefore, and to them alone, we propose to extend the system already adopted in the case of the Muhammadans.”

The Simon Commission described the result in 1930. Voters in general constituencies were split into Muslim and non-Muslim lists, and in the provinces where Muslims were a minority their representation was “considerably in excess of its population ratio,” because the franchise committee took the Lucknow Pact as its guide. Sikhs were 11.1 per cent of the population of the Punjab, 24.1 per cent of its voters and held 17.9 per cent of the communal seats. Prosperous Sikh farmers were numerous, the Commission explained, and Sikhs probably formed a higher share of those who won the vote as ex-soldiers.

Two votes for the Depressed Classes, 1932

On 11 March 1932 Gandhi, then a prisoner, wrote to the Secretary of State, Sir Samuel Hoare, that if the government created separate electorates for the Depressed Classes, as the Scheduled Castes were then called, “I must fast unto death.” Separate electorates, he said, were “neither a penance nor any remedy for the crushing degradation they have groaned under.” The Prime Minister’s decision on communal representation came in August. Its ninth paragraph gave the Depressed Classes a vote in a general constituency and a second vote in a special constituency of their own, which would return a fixed number of members, and it intended those special constituencies to end after 20 years.

On 18 August, writing from Yeravda jail, Gandhi told Ramsay MacDonald that he had “to resist your decision with my life,” and that the fast would begin at noon on 20 September. B. R. Ambedkar, who had represented the Depressed Classes at the Round Table Conference, and Hindu leaders agreed the Poona Pact, which was signed on 24 September, and Gandhi accepted its terms. It dropped the separate electorate and reserved 148 seats in the provincial councils for the Depressed Classes out of the general seats. The Depressed Classes on the general roll would elect a panel of four candidates for each reserved seat, and that primary election was to end after ten years.

Ambedkar later set out his objection in a book that prints the documents. The decision had given the Depressed Classes 78 seats, he wrote, and the Pact gave 148, but the decision had also given them a double vote, and “the second vote given by the Communal Award was a priceless privilege.” The extra seats, he wrote, were “no recompense for the loss of separate electorate and the double vote.”

Where the three lists met

Cohn observes that the census, the army and the electoral question developed together. While the census was collecting information on caste, he writes, “the army was beginning to be reorganized on assumptions about the nature of ‘martial races’.” The papers show the lists overlapping. Montagu and Chelmsford named the Sikhs’ place in the army when they made the case for a Sikh electorate, and the Simon Commission linked Sikh votes to ex-soldiers.

In March 1921 the Legislative Assembly took up a resolution moved by Sir P. S. Sivaswamy Aiyer, a Madras moderate, asking that Indians be admitted freely to King’s commissions. The Government of India amended it so that the commissions granted to a community would reflect the number of its men serving in the ranks. Omissi writes that the amendment passed by one vote, with the strong support of Punjabi Sikhs and Muslims.

The Simon Commission gave the British defence of the recruiting pattern in 1930. Races which furnish the best sepoys, it said, “are emphatically not those which exhibit the greatest accomplishments of mind in an examination,” and the Punjab supplied 54 per cent of all combatant troops in the Indian Army. It was aware of “the suggestion” that this was “a deliberate policy adopted by the Army authorities for some sinister purpose,” and answered with the figures of the Great War, “when it cannot be suggested that any discouragement was offered to recruitment in any area.” Omissi’s table shows the army of 1914, the year the war began, already drawing 57 of its 121 infantry battalions from the Punjab and the Frontier and 11 from Madras. The Commission’s figures for the war were these: “Bengal, with a population of 45 millions, provided 7,000 combatant recruits; the Punjab, with a population of 20 millions, provided 349,000 such recruits.”