The government of British India

As background information, the system of government in India at the beginning of the twentieth century should be outlined. The Indian empire comprised the territories under direct British rule and the princely states under indirect rule. The head of the imperial government in Calcutta was the viceroy acting as the governor-general in Council. This executive council had five civilian members, which increased to six in 1904, when the overworked home member's extra responsibility for the Revenue and Agriculture department was transferred to a new member. Each department's bureaucracy was headed by its secretary. Excluding Burma, British India was divided into seven provinces, each with its own government: Madras, Bombay, Bengal, the North-Western Provinces and Oudh (later United Provinces), Punjab, Assam, and the Central Provinces. The government in London appointed the viceroy, and its secretary of state for India had ultimate responsibility for Indian affairs. Lord Curzon was the viceroy from 1899 to 1905 and Lord Minto from 1905 to 1910.

The seven provinces were subdivided into approximately 250 districts in total. District officers, the key administrative personnel of British India, were responsible for all local affairs. All senior district officers were members of the Indian Civil Service, the ICS, which made up an elite group of about one thousand civil servants, known in their day simply as ‘civilians’. On the eve of the First World War, around 95 per cent of civilians were British, although the proportion of Indians rose later. New Indian Civil Service recruits almost always began their careers in rural districts, typically learning about land settlements (registering proprietary rights, surveying fields, and assessing revenue rates), and dealing with minor court cases as subordinate magistrates. Junior officers were normally promoted fairly quickly, usually within the same or another district, but sometimes moved to departmental secretariats in their own provincial government or the imperial government. At all times, however, most civilians were district officers at various levels of seniority. A minority eventually became senior officials in a provincial government and a few rose to the very top to become departmental secretaries in the imperial government, members of the Viceroy's Council, or lieutenant-governors or chief commissioners in charge of provinces.

Denzil Ibbetson and Herbert Risley

The great majority of ethnographers and anthropologists in British India before the mid-1920s belonged to the Indian Civil Service. The official anthropology they produced was published in census reports, ‘tribes and castes’ handbooks, gazetteers, and other related literature, as well as books and articles written by individuals. Official publications often duplicated each other, but the amount of material they cumulatively contain is immense.

In Indian anthropology's formative years after the decennial censuses began, Denzil Charles Jelf Ibbetson and Herbert Hope Risley were the two leading figures. They made the most significant contributions to the ethnography and theory of caste, and to the promotion of anthropological enquiry. Certainly, though, many other men contributed as well, notably William Crooke (1848–1923), a rural district officer in the North-Western Provinces from 1871 to 1896. Crooke collected an impressive amount of ethnographic data and wrote prolifically on anthropology and folklore, but he had less theoretical influence than Ibbetson or Risley.

Ibbetson and Risley, both ambitious men, were two of the pre-eminent Indian Civil Service officers of their generation and ended their careers with knighthoods. Ibbetson, born in 1847, started his Indian Civil Service career in Punjab in 1870. He worked as a settlement officer in Karnal District (north of Delhi) from 1872 until 1880, and then as superintendent of the 1881 census of Punjab. The intellectual brilliance of both his settlement report and census report was quickly recognized. He was next promoted to more senior positions in the Punjab government, before becoming secretary of the Revenue and Agriculture department in Calcutta in 1894, and then chief commissioner of the Central Provinces in 1898. In 1902, Ibbetson took his seat on Curzon's Council as home and revenue member; after 1904, he remained in charge of revenue, but A. T. Arundel became home member. Ibbetson left the Council for six months in 1905 to become acting lieutenant-governor of Punjab and in 1907 he was confirmed in the post, the highest that a civilian could attain. But soon after Ibbetson fell ill and resigned from the Indian Civil Service in January 1908; he died in London a month later.

Risley was born in 1851 and arrived in Bengal in 1873. In his career, Risley was transferred frequently. He began as a junior district officer in rural Midnapur (west of Calcutta) and, during the next ten years, he held more senior posts in different rural districts and also on the secretariat staff of both the Bengal and Indian governments in Calcutta. In 1885, in response to a proposal for research on ‘the little known subject of caste’ by the commissioner in charge of the 1881 census, the Bengal government deputed Risley to carry out an ‘exhaustive enquiry into the castes and occupations’ in the province for three years.Footnote ¹² Various other posts followed, before Risley became the 1901 census commissioner; in 1901, he was also made honorary director of ethnography for India. In 1902, he was promoted to secretary of the Home department, a post he held under both Curzon and Minto, and he ended his career as the temporary home member. Risley returned to London in 1910 to become permanent secretary at the India Office; he also served as president of the Royal Anthropological Institute. Risley died in September 1911.

In his preface to the 1881 Punjab census report, Ibbetson highlighted ‘the extraordinary interest’ of Indian ethnographic material, which could be of ‘immense value to students of sociology’. (Ibbetson probably followed Herbert Spencer in using the term ‘sociology’, but Risley preferred ‘ethnology’; ‘anthropology’ became commoner by the end of the century.) ‘Our ignorance of the customs and beliefs of the Indian people,’ Ibbetson continued, ‘deprive[s] European science of material which it greatly needs’ and ‘also involves a distinct loss of administrative power to ourselves’.Footnote ¹³

When Ibbetson, like Risley and many other administrator-scholars, insisted that ethnography and anthropology (sociology, ethnology) were valuable politically, as well as academically, they probably had more in mind than practical improvement in government and administration. Official thinking in late nineteenth-century India, particularly among civilians in Punjab, including Ibbetson, was permeated by a conservatism owing much to Henry Maine's influence.Footnote ¹⁴ Nevertheless, the liberal idea that imperial rule could bring progress to India never vanished and it depended on the premise that the British possessed modern, scientific knowledge that their Indian subjects lacked. In 1875, Maine reviewed the achievements of comparative philology and mythology, and hoped for similar ones in comparative jurisprudence.Footnote ¹⁵ He grandly concluded his lecture by declaring that the ‘intellectual excellencies which we regard as characteristic of the great progressive races of the world’, which had originated in ancient Greece, had now been inherited by the British, who were duty bound to pass them on to India.Footnote ¹⁶ Risley soon read the lecture in India, and he and Ibbetson, who had been inspired by Maine as young men, were convinced that civilians should possess an intellectual understanding of India, not just bureaucratic competence. They also clearly believed that advances in comparative ethnology and sociology, paralleling those described by Maine, could be made by collecting data in India.Footnote ¹⁷ Thus, even if the intricacies of, say, the Bengali or Punjabi caste systems, which they would soon study, were not directly relevant to governance, their investigation would augment the scientific knowledge, including about India itself, that distinguished the progressive British from their Indian subjects and justified their enlightened imperial rule.Footnote ¹⁸ It is partly because they were ardent imperialists that Ibbetson and Risley were also committed to scholarly enquiry into ‘traditional’ Indian society.

Ibbetson's ethnographic and anthropological work

Most Indian land settlement reports were confined to strictly relevant matters. Considerable ethnographic information was included, though, in a few Punjabi reports, one of them by Ibbetson, who spent eight years on the revision of the settlement in Karnal District. Ibbetson's report, published in 1883, contains a clear description of crucial features of the social structure—in particular, the local organization of tribes, castes, and villages—which was informed by his knowledge of contemporary anthropological theory.Footnote ¹⁹ These qualities make Ibbetson's settlement report stand out, even from similar Punjabi reports, and in many respects it actually resembles a modern, functionalist ethnography of a local society, albeit at a district rather than village level.

Notwithstanding the quality of his Karnal report, Ibbetson's voluminous 1881 census report on Punjab had the greater impact. It includes a substantial chapter on religion, but one-third consists of a chapter on ‘races, castes and tribes’, later republished separately, which became a foundational text for the anthropology of British India.Footnote ²⁰ In effect, Ibbetson had written a tribes and castes handbook before Risley wrote his, formally the first, for Bengal; H. A. Rose's later Punjab handbook heavily depended on Ibbetson's text.Footnote ²¹ Ibbetson's census report drew on his knowledge of Karnal, but he acknowledged its limitations for the rest of Punjab, and the bulk of his information came from other publications, mostly about particular districts.

Introducing the chapter on religion, Ibbetson insisted that the numerous books on the subject ‘fail utterly and entirely’ in describing the religions ‘as actually practised’ by ordinary villagers.Footnote ²² Classifying the people by religion was extremely difficult and defining ‘Hindu’ was impossible. ‘Creed,’ he wrote, is ‘rather a social than a religious institution.’ Footnote ²³ Ordinary people's religious practices were typically lax and eclectic, and very similar among Hindus, Sikhs, and Muslims. Matching observations would recur in census reports and other official literature for every province of India until the end of colonial rule, even though Risley and other British spokesmen, as we shall see, later declared that Hindus and Muslims were separate communities.

Before turning to Ibbetson's chapter on castes and tribes, I must explain that the British established a legal system based mainly on customary law in Punjab. This depended on information collected from people in different social groups, primarily the province's main ‘tribes’, by asking them about customs relating to family relationships, marriage, inheritance, and so on. C. L. Tupper's Punjab Customary Law, published in 1881, is a compilation of relevant documents which includes his own ideas about Punjabi society.Footnote ²⁴ Tupper argued that kinship, specifically the agnatic kinship of tribes and clans, was the basic principle underlying the ownership and distribution of property in Punjab, not only in the fully tribal regions in the west of the province, but also within the village communities prevalent in the east. For Tupper, the tribe—supplemented by the clan and village—was the fundamental unit of Punjabi society and ‘what is commonly called caste’ is ‘more properly described as tribal connection’.Footnote ²⁵ All the reports on customary law in different districts were organized as handbooks of ‘tribal’ customs, even for groups such as Brahmans.

Ibbetson respectfully cited Tupper, but did not follow his lead in treating ‘caste’ as misnamed ‘tribe’.Footnote ²⁶ Instead, Ibbetson identified different types of Punjabi caste: ‘true Brahminical’ castes; ‘tribal’ castes, especially among Muslims in the western frontier region; ‘occupational’ castes; Muslim ‘trades-guild’ castes; ascetic orders; and other small and intermediate types.Footnote ²⁷ To organize the tables showing geographical distribution, as well as his ethnographic description, Ibbetson divided the castes and tribes into ‘three great groups’. The first is the ‘landowning and agricultural group’, half of Punjab's population and ‘even more important socially, administratively, and politically than it is numerically’, a comment reflecting the great importance Ibbetson and other officials gave to this group as the bulwark of British power. The first group are also identified as ‘the agricultural castes’, subdivided into six sections labelled ‘races’ (for example, the frontier Pathans or the Jats) or ‘agricultural’ and ‘dominant tribes’.Footnote ²⁸ The second group comprises ‘the priestly, ascetic, professional, and mercantile castes’, which are generally described as ‘castes’, except for the Muslim ascetic orders.Footnote ²⁹ The third group of ‘menial castes’—the ‘lower strata of society’ denoted by no consistent terminology—includes ‘the vagrant, criminal, and gypsy tribes, the village menials, and the industrial classes’.Footnote ³⁰ Ibbetson acknowledged the confusion in how census officials enumerated castes and tribes, and his own use of terms was loose, although he also insisted that ‘it is almost impossible to define a caste and difficult to define a tribe’ as discrete, bounded groups.Footnote ³¹ These problems were partly conceptual, but also stemmed from Punjab's enormous diversity, from the predominantly Muslim west to the Hindu east, and from the agricultural plains to the hilly tracts. Throughout, Ibbetson generally used ‘race’ as a synonym for ‘people’ and gave it no specific ethnological meaning, unless he was explaining, for instance, why the probable racial ancestry of Jats and Rajputs was irrelevant to extant social distinctions.Footnote ³²

The 1881 census was an improvement on its unsatisfactory predecessor of 1871–2, although the all-India classification of castes and tribes remained highly problematic. Among the 1881 provincial reports, however, Ibbetson's was also outstanding because, like Athelstane Baines's on Bombay, it set out a new theory of caste.Footnote ³³ Ibbetson and Baines worked independently, as far as I know, but both proposed schemes to classify castes according to occupation, which were linked to a theory of caste predicated on evolutionism, then the dominant paradigm of social theory. In the background of their arguments, but of small importance to them, was the so-called Aryan invasion thesis. It was widely believed in the nineteenth century that speakers of Aryan or Indo-European languages belonged to a common Aryan race and that, in ancient times, fair-skinned Aryan tribesmen from central Asia had invaded or migrated into India to displace or subjugate the dark-skinned, indigenous peoples, predominantly Dravidian, who became their hierarchical inferiors.

According to evolutionist theory, the division of labour becomes more complex as society develops. Thus Ibbetson wrote that primitive societies are divided into tribal groups and no diversity of occupation exists. In more advanced societies, occupations become differentiated and tribal groups almost disappear; in Europe they were replaced by guilds held together by ‘common occupation’, not ‘common blood’. Yet, according to Ibbetson—who probably drew on Maine and Spencer, but did not name his ‘modern authorities’ on social theory—the guild, like the tribe, was originally based upon common descent, so that ‘the fundamental idea’ underlying the institution was ‘the hereditary nature of occupation’.Footnote ³⁴ Furthermore, he continued: ‘The whole basis of diversity of caste is diversity of occupation.’Footnote ³⁵ ‘Occupation’ or ‘function’ was defined broadly to include political dominance and subjection—‘ruling’ and ‘being ruled’—as well as jobs and means of livelihood. Consequently, especially among the agrarian population engaged in similar work, ‘social standing, which is all that caste means, depends very largely upon political importance, whether present or belonging to the recent past’.Footnote ³⁶ The difference between social evolution in India and Europe, Ibbetson stressed, was ‘a difference of degree rather than of kind’, brought about by the exaltation of the Brahman priesthood, which turned into a hereditary group. The Brahmans then gave religious sanction to the hereditary principle, which became generalized throughout Hindu society.Footnote ³⁷ In short, therefore, occupational groups had become closed, hereditary, ranked castes in India, instead of open guilds as in Europe. In Punjab, the picture was complicated, as already implied, by the tribal and trades-guild types of caste, which Ibbetson explained as the outcome of Muslim influence on the pre-existing Hindu system.Footnote ³⁸

Ibbetson and Baines worked out their evolutionary arguments slightly differently, but agreed on the main points.Footnote ³⁹ They also set out caste classification schemes defined by traditional, hereditary occupation or function that were consistent with their theories. Baines, unlike Ibbetson, wrote concisely without much ethnographic detail. After the mid-1880s, presumably because his official duties became much heavier, Ibbetson did hardly any more anthropological work. Baines, however, was appointed commissioner for the 1891 census. In his general report, he succinctly refined the occupational theory and produced a classification scheme covering all of India for castes and tribes, and non-Hindu minorities. For Hindu castes, the principle of classification was ‘on a functional basis, with as much regard for position in the popular estimation [i.e., rank] as possible’. Baines admitted that his ‘grouping and arrangement is imperfect’, but it seemed to be the only ‘practicable’ way ‘to include all the different systems of the country in a single return’.Footnote ⁴⁰

Two features of the occupational theory should be highlighted. First, the theory explains caste by comparing India with Europe and identifying it as a form of the division of labour within a society that was evolutionarily relatively advanced, albeit unevenly. Secondly, religion is not a determining factor, except in explaining the Brahman priesthood's superiority and influence, and the occupational groups within the classification schemes include castes and communities belonging to different religions.

Risley's ethnographic and anthropological work

After Risley was appointed to enquire into Bengali castes in 1885, he discussed his task in Lahore with Ibbetson and John Nesfield, an educational officer in Oudh and author of another version of the occupational theory.Footnote ⁴¹ Risley began work in 1886 and his handbook, The Tribes and Castes of Bengal, was published in two ‘anthropometric’ and two ‘ethnographic’ volumes in 1891–2.Footnote ⁴²

Risley wanted to investigate ‘social precedence’ or hierarchical ranking and to develop a racial theory about the origins of caste. The most useful tool, Risley believed, was anthropometry, the science and technique of bodily measurements, which could both determine the racial composition of the Indian population and prove scientifically that a correspondence existed between social status and physical features. To this end, his Bengal handbook's anthropometric volumes contain extensive numerical data on different caste samples. Risley—who respectfully cited Ibbetson and Nesfield, but not Baines—argued that the occupational theory was incorrect, because the hierarchical caste system was ultimately based on distinctions of race, sustained over time by endogamous marriage. Risley's racial theory, published in an article in 1891 and in the Bengal handbook, was amplified, but not essentially modified, in both the 1901 census report and his book, The People of India, first published in 1908 and reissued under Crooke's editorship in 1915.Footnote ⁴³

Risley's racial theory and use of anthropometry have preoccupied his modern critics, who commonly cite his notorious ‘working hypothesis’, later hardened into ‘a law of the caste organization’, that, at least in some regions, ‘the social status of the members of a particular group varies in inverse ratio to the mean relative width of their noses’.Footnote ⁴⁴ Read in context, Risley's claim about the ‘nasal index’ looks somewhat less dogmatic and ridiculous. All the same, the ‘science’ of anthropometry is now rightly discredited. More importantly, theories of racial evolutionism, which postulate the superiority of civilized races over primitive races existing in barbarism, are universally condemned today. In this light, Risley's entire approach looks indefensible and I shall not try to defend it.

Nonetheless, latter-day criticism of Risley's racial theory of caste is rather misleading. First, the Aryan invasion thesis, and the notion that the Aryan race was more evolutionarily advanced than the Dravidian race, had been axiomatic for virtually all scholars of India for several decades.Footnote ⁴⁵ Occupational theorists, such as Ibbetson and Baines, also assumed that different castes were originally formed from various racial ‘stocks’, even though, unlike Risley, they did not believe that racial distinctions mainly determined the caste system's evolution in more recent centuries.

Secondly, it is salient that both Europeans and high-status Indians were ‘Aryans’, although Maine, in writing about ‘ancient society’, distinguished between the progressive Aryan civilization of the West and its static or degenerate counterpart in the East. By the 1880s, Maine's theory of evolution from ancient to modern society had been comprehensively criticized by anthropologists such as J. F. McLennan and E. B. Tylor, whose work was familiar to Ibbetson and Risley, although both men probably presumed, like Maine, that a real distinction existed between European and Indian Aryans.Footnote ⁴⁶ What matters far more, though, is that all Victorian evolutionist theories legitimated European supremacy and imperialism in one way or another, and Risley's theory of caste did so no more potently than its rivals.

The third point, which matters for evaluating British official anthropology as a whole, is that the majority of Risley's colleagues thought his racial theory was wrong, even though some academic anthropologists still took it seriously in later decades. Empirically, the argument that caste divisions are fundamentally racial depends on evidence that different caste populations are physically distinct. Before Risley had published anything, Nesfield was convinced that Aryan and aboriginal blood were so thoroughly mixed that there was ‘unity of the Indian race’.Footnote ⁴⁷ Ibbetson, too, was sceptical that putative racial ancestries were sociologically significant in Punjab, although later he deferred to Risley's expertise and expected anthropometric measurements to provide valuable data.Footnote ⁴⁸ After The Tribes and Castes of Bengal appeared, C. J. O'Donnell, superintendent of the 1891 Bengal census, quickly insisted that Risley's nasal index data failed to support his conclusions, bluntly dismissed the ‘race theory of caste [which] is found to have practically no statistical support’, and concluded that, between the Afghan frontier and the Bay of Bengal, all peoples were ‘so absolutely intermingled in blood’ that castes were inseparable along racial lines.Footnote ⁴⁹ Crooke was more sympathetic to Risley's endeavour than O'Donnell and did endorse anthropometry, but it was precisely the anthropometric data from the North-Western Provinces that made ‘the theory of the ethnological basis of caste’ largely untenable for Crooke, so that he cautiously supported the occupational theory.Footnote ⁵⁰ Later, when anthropometry itself was losing credibility, Crooke opposed the racial theory more unequivocally and listed a series of objections to it in the second edition of Risley's The People of India.Footnote ⁵¹ Edgar Thurston, whose tribes and castes handbook for Madras appeared in 1909, produced anthropometric data which, he believed, supported the racial theory, but he was an exception.Footnote ⁵² Other critics included Baines, who was dubious in 1893 and flatly rejected Risley's premise about racially distinct castes based on ‘purity of descent’ in 1912.Footnote ⁵³ E. A. Gait, who had become the 1901 census commissioner after Risley's promotion, was also the 1911 commissioner and carefully avoided judgement on his former colleague's theory, but he was more sceptical a little later.Footnote ⁵⁴ Anthropometry and the racial theory were simply dismissed by the 1911 United Provinces and Bengal superintendents, E. A. H. Blunt and L. S. S. O'Malley respectively, who later wrote valuable books on caste.Footnote ⁵⁵ Writers on colonial anthropology have generally failed to make it clear that Risley's contemporary critics were in the majority, so that Dirks’ claim that the racial theory of caste was ‘clearly fundamental to the definition of the ethnographic project in turn-of-the-century colonial India’ is an overstatement.Footnote ⁵⁶

Let me now turn to ‘social precedence’, the second component of Risley's theory of caste. It can easily be separated from the racial theory, as it was by most of Risley's colleagues, and in the long run it mattered far more for the study of caste. In 1886, Risley circulated a list of the main Bengali castes ‘arranged in order of special precedence, as stated by different authorities whom I have consulted’ to numerous British and Indian officials in the province, although the circular actually reproduced only one Brahman pandit's list. Correspondents were asked to respond to the list, particularly noting castes from which Brahmans could take water, which were entitled to village barbers’ services, and which could enter the courtyards of great temples. Many replies, some brief and to the point, and some lengthy and passionate, were received by Risley.Footnote ⁵⁷ He used them, together with a lot of other material, to write his Bengal handbook, which was organized as an alphabetical glossary, providing data on each group's social status and caste rank, as well as other features.

When Risley was director of ethnography from 1901 until his retirement, his principal task was to organize an ethnographic survey, so that a full set of tribes and castes handbooks could be published. The second one, Crooke's The Tribes and Castes of the North-Western Provinces and Oudh, had already come out in 1896; companion volumes for every province in British India (except Assam) were completed by the early 1920s, plus a few for princely states later, thus finalizing the classificatory, alphabetical objectification of Indian society. By then, however, the handbooks were being criticized, partly because their glossarial format precluded comprehension of caste as a system. For Assam, monographs on separate tribes continued to appear until the late 1930s, but plans for a handbook came to nothing.

Because Gait, who was also the Bengal superintendent, replaced Risley as the 1901 census commissioner, he wrote most of the general report, although Risley completed the chapter on caste, tribe, and race (which formed the core of The People of India), and G. A. Grierson wrote the one on language. At the 1901 census, Risley was determined to collect data on caste ranking, which he expected to be useful for the ethnographic survey as well. Risley criticized the caste classification scheme in Baines's 1891 census as ‘a patch-work classification in which occupation predominates’, but varied by other criteria, whereas a ‘classification by social precedence’, he believed, would be clearly superior.Footnote ⁵⁸

Risley instructed all superintendents to investigate social precedence and they each devised schemes suitable for their own provinces. Gait explained that in Bengal five ‘well recognised tests of social position’ were used for Hindu castes: whether good or degraded Brahmans serve as priests; rules about taking food and water; whether touch is polluting; ceremonial observances (for example, whether widow remarriage is permitted); and availability of barbers’ and dhobis’ services. Committees of ‘native gentlemen’ were also consulted about the accuracy of lists of ranked castes. The outcome for Bengal proper (excluding Bihar and Orissa) was a set of seven categories, from Brahmans at the top to ‘unclean feeders’ (not yet called ‘untouchables’) at the bottom. But none of this was straightforward, because trying to sort out social precedence caused an ‘extraordinary amount of ill-feeling and jealousy’, and some local committees could not reach agreement.Footnote ⁵⁹ In other regions, the tests were not the same as in Bengal and nor were the lists. In the North-Western Provinces twelve ‘social precedence groups’ among Hindus were drawn up; in Madras, eleven; and in the Central Provinces, nine. In Bombay, no proper list was devised, partly because the superintendent did not want to provoke discontent over ranking. Similarly, in Baroda princely state there were eight groups, but in Travancore no list was made. In Punjab, Rose insisted that, except for a division into ‘clean’ and ‘unclean’ among both Hindus and Muslims, castes could not be classified by social precedence, so they were listed alphabetically within seven broad categories.Footnote ⁶⁰

Risley declared that the provincial reports on social precedence ‘added very greatly to our knowledge of an obscure and intricate subject’.Footnote ⁶¹ Actually, they showed that classifying castes by rank generated categories scarcely more uniform than earlier occupational ones. The countless petitions sent to census officials by educated Indians and caste associations about their own status typically appealed to Brahmanical textual authority, but they simultaneously showed that little consensus about rank really existed. Risley saw the petitions as evidence that he was right to investigate social precedence, whereas many of his colleagues and successors thought the whole exercise produced more heat than light.

Nonetheless, Risley's focus on social precedence materially altered the understanding of caste. His criticism of the occupational theory was telling, even if his own racial theory was unconvincing, and hierarchical rank has been identified as more crucial than hereditary occupation in most subsequent anthropological writing on caste. Significantly, though, social precedence, as conceptualized by Risley, prioritized high-caste or Brahmanical criteria and was, first and foremost, a property of the Hindu caste system, especially as it existed in Bengal. Risley paid little attention to ‘primitive’ tribes, or to social precedence among Muslims.Footnote ⁶² In his work, therefore, despite the large Bengali Muslim population, non-Hindu concepts of rank became entirely secondary, unlike those for the occupational theorists, particularly Ibbetson describing Punjab, where status attribution was also more fluid than in Bengal. And because the caste system was now seen as primarily Hindu or Brahmanical, it followed that it was special to Indian society conceptualized as pre-eminently Hindu. Indeed, the anthropological notion of caste as distinctly, or even uniquely, Hindu-cum-Indian can be traced to Risley's emphasis on social precedence and the influence it subsequently exerted.

On the other hand, Risley himself probably did not see the principle and practice of caste hierarchy as particularly strange. Maybe it is coincidental, but Risley's ‘social precedence’ expression brings to mind the Warrant of Precedence in India, which (in the 1898 version) listed 78 separate status levels from the viceroy down to third-class civil officers.Footnote ⁶³ Victorian Britain was a society deeply divided by social class and the ‘subtleties of the British class system . . . became elaborately codified in the Warrant of Precedence, which was designed as an infallible guide to hierarchy in India’.Footnote ⁶⁴ Ibbetson wrote that ‘caste in India [is] what we call position or rank in England’, whereas Risley never did.Footnote ⁶⁵ Yet class hierarchy in Britain, as amplified in the Raj, was almost certainly taken for granted as a fact of life by Ibbetson, Risley, and their civilian colleagues, while caste hierarchy was seen as correspondingly normal in Indian society. Caste was peculiar in its extremes of Brahman pretension and the ‘unclean’ castes’ degradation, but in most respects it was familiar enough.

In the 1911 census report, Gait declared that monographs on important castes and tribes would be valuable, but for ‘a general description of them . . . comparatively little remains to be done’, so his census had not collected much new information on individual castes.Footnote ⁶⁶ Even if Gait overstated his case, the ‘little known subject of caste’, as the 1881 commissioner labelled it, was certainly not little known by 1911. It should be stressed that the debates about caste that engaged official anthropologists were closely related to intellectual questions current in Victorian ethnology, anthropology, and social theory. The same was true for religion or marriage, for example, and topics irrelevant to government—such as totemism and exogamy, which then preoccupied metropolitan scholars—were often discussed at length.Footnote ⁶⁷ Scientific knowledge, as I mentioned earlier, could ostensibly legitimate imperial rule, but a lot of official anthropology was just purely academic. Yet by 1911, perhaps inevitably, the pioneering sense of ethnographic and intellectual discovery that had particularly animated Ibbetson and Risley, as well as Crooke (who retired to England in 1896), was on the wane and an anthropological era in British India was ending. By 1902, Ibbetson and Risley were working in Curzon's Home Department and within a decade they had both died.

Land policy in Punjab

In the rest of this article, I examine Ibbetson's and Risley's work as government officials. Land policy in Punjab in the late nineteenth century illustrates an exceptionally close link between anthropology and imperial rule. The ‘Punjab school’ of administration was committed to an ideal of paternalist government based on direct and personal rule, with local customary law forming the basis of the judicial system. In the 1870s, Tupper argued forcefully that the British government should work to uphold the tribe, clan, and village community, which were the fundamental institutions of Punjabi society: particularly through the tribe and clan, too, the government ‘can gain its firmest hold on the inclinations and motives of the people . . . [who] can be led by their own leaders’.Footnote ⁶⁸ Tupper, who insisted that good administration could both improve and preserve the tribal system, believed that in Punjab, ‘the change to individualism’ would not be ‘a change for the better, compared with what we have now’, so that the government should retard the growth of individualism without preventing all progress.Footnote ⁶⁹ No civilian argued more articulately than Tupper (who depended heavily on Maine for theoretical inspiration) against the introduction of alien law and political economy into ‘traditional’ Punjabi society.

In the 1870s and 1880s, a new round of land settlements was undertaken in Punjab, including the one by Ibbetson. At this time, concern about peasant indebtedness was growing in several regions of India, including Punjab, where S. S. Thorburn, another settlement officer, raised the alarm about excessive land transfers from cultivators to moneylenders in the northwestern Bannu District. Peasant farmers were becoming landless sharecroppers or labourers; if this problem were not dealt with, disaffection among cultivators could threaten the security of British rule. Senior Punjabi officials, however, disagreed with Tupper's arguments, dismissed Thorburn's warnings, and opposed any interference in private property rights and the free market.Footnote ⁷⁰

In the late 1880s, land alienation appeared to increase sharply and agrarian unrest spread. Controversy over state intervention persisted in Punjab and the imperial government in Calcutta established an enquiry commission with Ibbetson as the Punjab government's representative. He then became secretary of the Revenue and Agriculture department and formulated proposals to deal with land alienation in 1895. In the end, new legislation was considered only for Punjab, but nothing had been finalized before Curzon became viceroy in 1899. Curzon gave his full support to legislation, partly on Ibbetson's advice, and the Punjab Alienation of Land Act became law in 1900. The Act's most vital clause stated that a permanent alienation of land could be made only if either the alienor was not a member of an ‘agricultural tribe’, or both alienor and alienee were members of the same tribe or of tribes in the same group. In other words, land belonging to farmers from an agricultural tribe—as identified by the Punjab government—could no longer be transferred to outsiders, which primarily meant urban moneylenders from non-agricultural communities.

Ibbetson, of course, had described the ‘landowning and agricultural group’ of castes and tribes as the most important component of Punjabi society in his 1881 census report. A few years later, he was fulminating about land alienation. Thus in 1889, Ibbetson wrote that no problem existed if alienations were voluntary and took place within the community, but if a peasant's loss of land was traced to a court decision (often when a moneylender took legal action over unpaid debts), ‘we [the British] shall bear the whole odium’. Our ‘unjust’ laws force peasants to part with their land, our civil courts ‘stink in the nostrils of the peasantry’, and we must change laws, especially about land ownership, which are ‘utterly unsuited to the circumstances, necessities, habits and ideas of the people’. ‘Righteous personal Government is what the people understand and like’ and, not waiting until they hate us, we must try ‘governing them after their own fashion’ with ‘honesty of purpose and impartiality’.Footnote ⁷¹

This note's rhetorical force derives from Ibbetson's first-hand experience among the peasantry, as well as his implied reference to the prevalent colonial myth of ideal, self-sufficient, peasant cultivators, who were best represented by the Jats, ‘pre-eminently the agricultural caste’ of eastern Punjab and a ‘fine stalwart race’; generalizing in the same vein, Crooke described the Jat as ‘the typical yeoman, the finest farmer in Northern India’.Footnote ⁷² Ibbetson's intellectual argument, however, closely resembled Tupper's. He was adamant that private property in land and its associated legal system had seriously damaged traditional society; moreover, the security of imperial rule would be imperilled if remedial action were not taken. Land alienation had to be controlled, but governing the people with the paternalistic authoritarianism they understood and respected was also vital. Both economically and politically, in other words, the British had to cleave closer to the norms of traditional, patriarchal, agrarian Punjabi society. Clive Dewey describes Ibbetson and other like-minded officials, such as Thorburn and Tupper, as ‘the official anthropologists’ of Punjab, whose thinking was influenced by both Maine and their local ethnographic knowledge. Their arguments finally ended the influence of classical political economy on Punjab land policy and nothing shows that better than the Alienation of Land Act, ‘a particularly brutal breach with laissez-faire’.Footnote ⁷³

The situation in Punjab was exceptional in many respects. Furthermore, the British could never restore ‘traditional’ society, so that they changed more than they preserved in the ‘imperial state’ they created.Footnote ⁷⁴ Nonetheless, the Alienation Act gave legal shape to an idealization of traditional Punjabi society, whose landowning and agricultural communities defined its core. In making that happen, Thorburn, Tupper, and others were instrumental, but no high-ranking Indian Civil Service officer did more to formulate policy and legislation according to his ethnographic understanding of Indian society than Ibbetson. Significantly, though, in this Punjabi context, the agrarian class structure mattered far more than the caste system Ibbetson had described and theorized about at such length.

The partition of Bengal

One of Curzon's most controversial decisions as viceroy was to partition Bengal. Partition was eventually completed in 1905, but it was strongly opposed by the Bengali middle class and, because it gave a decisive impetus to the Indian nationalist movement, it had huge consequences for the entire subcontinent. Curzon's viceroyalty, to cite a standard historical judgment, ‘was the last blaze of the old self-confident benevolent imperialism. . . . From 1905 India entered into a new era.’Footnote ⁷⁵

Ibbetson worked closely with Curzon, fully shared his imperialist outlook and admired him greatly; Curzon in turn regarded Ibbetson as a rare colleague on whom he could rely.Footnote ⁷⁶ In late 1901, Curzon asked Ibbetson about various appointments and he suggested Risley as a good choice for secretary in his Home department: ‘Risley is clever enough for anything—almost too clever, they say . . . For some reason or other, he is not trusted; but whether there is anything in this than his manner, I do not know . . . I know him intimately, and should feel no hesitation about appointing him.’Footnote ⁷⁷ Risley was widely regarded as both clever and untrustworthy, but he did gain the confidence of both Curzon and Minto, the next viceroy. Risley was responsible for drafting the proposals for both the partition of Bengal and the Morley-Minto reforms (which I discuss below), and the documents he wrote allow us to examine how far his anthropological understanding influenced the imperial government's policy making.

Before partition, Bengal was a vast province which also included Bihar and Orissa. The proposal to divide it was originally part of a larger plan to adjust the boundaries of Madras, the Central Provinces, Bombay, and Punjab, as well as Bengal and Assam. This plan, to quote a famous complaint from Curzon in 1902, had been going ‘round and round’ the bureaucratic system inconclusively for a long time.Footnote ⁷⁸ In early 1903, Ibbetson and Risley gave their views about transferring parts of Berar (Central Provinces) to Bombay, adjusting the boundary between Sind (part of Bombay) and Punjab, and bringing all Oriya (Uriya) speakers within one administrative unit. Ibbetson favoured Berar's addition to Bombay because their land revenue systems were so similar, but Curzon had already rejected the proposal on the grounds that it would strengthen the Marathas (by which he really meant the Maratha Brahmans), the ‘most able’ and ‘most dangerous’ opponents of British rule. On the second issue, citing similar land revenue and legal systems, Ibbetson argued that the boundary between Sind and Punjab was in the wrong place; furthermore, it corresponded to no ethnological or racial division, as Risley—‘the highest living authority on the subject’, in Ibbetson's words—had confirmed. Curzon does not appear to have dissented, but the Sind-Punjab boundary was eventually left unchanged anyway. On the Oriya question, Ibbetson emphasized an ethnological distinction between Oriya plains people and the hill tribes of Dravidian origin. Risley argued against him, and persuasively to Curzon, that the racial division was less salient than linguistic unity among Oriya speakers, who should be united in Bengal, instead of being divided (as they were) between Bengal, the Central Provinces, and Madras.Footnote ⁷⁹ Curzon presumably recognized Ibbetson's and Risley's expertise on anthropological matters, but their advice on these three issues shows that they deployed ethnographic and related evidence variably and even inconsistently, and that becomes clearer in relation to the bigger problem of Bengal itself.

The plan ‘for the reconstitution of the provinces of Bengal and Assam’ was formally announced in December 1903 in Risley's letter to the government of Bengal.Footnote ⁸⁰ The plan included proposals to unite Oriya speakers within Bengal for the reasons just outlined, and to transfer part of the mainly tribal Chota Nagpur area to the Central Provinces, primarily because the Bengali administrative system was ‘too highly developed, too legalized, and too impersonal for backward tracts and primitive hill-tribes’. The most important proposal, of course, was to incorporate eastern Bengal into a new province eventually called Eastern Bengal and Assam, and to create a new, smaller province of Bengal comprising western Bengal and Bihar and Orissa.

Risley's letter quickly provoked an outcry in Bengali newspapers. He was vilified personally as Curzon's ‘buffoon or jester’ and mocked for ‘making people laugh at his ignorance’, and his arguments for partition were denounced, typically as ‘astonishing in their queerness’, or ‘worthless and ludicrous’.Footnote ⁸¹ The abuse stuck and Risley's unpopularity in Bengal never died away. In early 1904, Bengali associations sent more than 20 memorials to Risley or Andrew Fraser, the lieutenant-governor of Bengal, objecting to the reconstitution.Footnote ⁸² Many objections, in the newspapers and memorials, were very similar. Some were about potential damage to the economy and the legal and educational systems, but others were about the effect on Bengali society and culture. One complaint was that the reasons given for uniting Oriya speakers were ignored by the plan to split Bengali speakers, so that Risley was told he was not invoking ‘ethnological, geographical, and linguistic considerations’ consistently.Footnote ⁸³ An insistent argument about language was that ‘the Bengali language, the most perfect and developed in India, instead of developing further, instead of receiving further vitality, will be stunted’ by partition. There were also protests that intermarriages between eastern and western families would become impossible for Kulin Brahmans and Kayasthas, because the East Bengalis would be ‘denationalised’ and deemed socially inferior after being attached to Assam, ‘a backward and primitive province’, and ‘merged into a people placed on a much lower plane of civilisation’.Footnote ⁸⁴ A memorial from Dacca asserted that territory east of the river Meghna was ‘looked down upon as the land of the aborigines’; if East Bengalis were incorporated into it, it would be ‘degrading’ in Hindu eyes and they would eventually be cut off from their ‘kith and kin’ in West Bengal. Europeans, it was said, only imperfectly understood such arguments about ‘social disability’.Footnote ⁸⁵ The rulers’ misunderstanding was also emphasized in a memorial from Calcutta, which insisted that it was not ‘idle sentimentalism’ to say that Bengalis would ‘suffer nationally, linguistically, and socially’ by partition.Footnote ⁸⁶

According to Sumit Sarkar, the final form of the partition plan was mainly Risley's and Fraser's work.Footnote ⁸⁷ Their responses to the memorials were summed up in the detailed dispatch sent by Curzon and his Council to St John Brodrick, the secretary of state, in February 1905.Footnote ⁸⁸ The logical complaint about inconsistency between the proposals for Oriya and Bengali speakers was simply ignored. The other arguments, which repeatedly asserted Bengali superiority, were answered. On language, the dispatch quoted an earlier letter giving Fraser's views.Footnote ⁸⁹ It insisted that an administrative division could not affect Bengali language and literature, and Risley added that because Bengali would be the dominant language in the new eastern province, it could not deteriorate. The claim that partition could affect intermarriage was pronounced inconsistent with relevant ethnographic evidence, and the suggestion that Bengalis could be degraded and denationalized in the new province was baseless: ‘Mere administrative division does not produce social division, any more than administrative unity produces social union.’ This conclusion, quoted from the earlier letter and emphasized in the dispatch, was undoubtedly lent credibility by Risley's anthropological knowledge of Bengal, but none of his responses really recognized that partition was ‘a matter of sentiment that deeply moved all sections of educated society in Bengal’.Footnote ⁹⁰ Risley, though, did not contest the claim, to which he almost certainly subscribed, that the Bengalis were a more advanced race than the Assamese.

In the 1905 dispatch, most weight was given to positive arguments for partition. The boundary dividing Bengal, it was claimed, would in fact ‘correspond with certain physical, linguistic and ethnological characteristics’, the most crucial being that ‘the typical Muhammadans of Bengal’ would be concentrated within the new eastern province and ‘the city of Dacca [would become] the natural capital of such an aggregation’. When partition was first mooted, Muslims were not mentioned, but in September 1904 Risley quoted population figures and specifically noted the strong Muslim presence in the new province and Dacca's future status as its capital.Footnote ⁹¹ From the outset, though, the main political aim of the British was to weaken the educated, urban, middle-class nationalists and their supporters, who predominantly belonged to the Bengali Hindu Brahman, Kayastha, and Baidya high-status ‘gentry’ or bhadralok in Calcutta and Dacca. Promoting a Muslim interest in the east became important primarily because it could assist that aim. In a memorandum written at Ibbetson's request in early 1904, Risley stressed that nationalist agitation against partition was ‘fomented by the literate class’, especially those who knew English. The literate class, as its own members realized, had ‘no real community of interest’ with the bulk of the population, so that their agitation was designed to try to create a new sense of unity among Bengalis. But partition would break that unity and that is ‘one of the great merits of the scheme’, for a united Bengal would be a powerful force. Ibbetson, who generally agreed with Risley, also stressed that it is ‘most desirable to diminish’ Bengali preponderance in provincial politics.Footnote ⁹² Bengali politicians and agitators, especially in Calcutta, were frequently castigated as ‘wire-pullers’ working nefariously behind the scenes, and the 1905 dispatch mentioned the need to weaken them. The memorialists’ complaints about lost national unity were dismissed as spurious, because ‘national unity . . . is a genuine growth rooted in the common bond of race, religion, and language, and not the artificial product of professional agitation’, as now seen in Bengal.Footnote ⁹³

This argument about ‘genuine’ national unity partly foreshadows Risley's discussion of nationalism in The People of India, where he compared India with Europe and modern Japan.Footnote ⁹⁴ As Dirks suggests, Risley surely had partition and its consequences in mind when he wrote his book's chapter on ‘caste and nationality’ in 1907–8, although its thesis was anticipated by an article in 1890.Footnote ⁹⁵ Interestingly, though, the 1905 dispatch never referred to caste, whereas in his book Risley called it ‘the governing principle of Indian society’ and argued that division into multiple, separate, endogamous units strongly inhibited the growth of nationalism, even if did not entirely prevent it.Footnote ⁹⁶ Indeed caste, the very topic that so preoccupied Risley as a scholar, is not mentioned in the documents he prepared, except in specific instances, such as claims about Brahman and Kayastha intermarriage. In 1906, after partition, Risley thought the ‘spread of education’ among Bengalis, including the lower castes, justified the attempt to undermine the longstanding ‘Brahman and Kayasth monopoly’ over government jobs.Footnote ⁹⁷ But discussion of caste differences between the high-caste bhadralok and the lower-status majority of the population, or a suggestion that they could be exploited to ‘divide and rule’—perhaps like Brahmans versus non-Brahmans in Madras and Bombay—were singularly absent from the Curzon government's deliberations about partitioning Bengal.

Concerning Muslims, Risley wrote in his book that ‘Islam is a force of the volcanic sort, a burning and integrating force, which, under favourable conditions, may even make a nation’; a similar comment appeared in his earlier article.Footnote ⁹⁸ Maybe this notion of political Islam coloured Risley's perception of the Muslims’ future in East Bengal; Curzon indeed proclaimed that in the new province they would gain a unity missing since the period of Muslim rule.Footnote ⁹⁹ Probably just as influential, though, was the general British assumption that Muslims were preferable to ‘talkative and audacious Bengali babus’, to quote Sarkar, reinforced by a desire for ‘an anti-babu extravaganza’, to quote Gordon Johnson.Footnote ¹⁰⁰ In any case, though, the information about Muslims in Gait's 1901 Bengal census report, which Risley summarized in his book, certainly did not validate Curzon's belief.Footnote ¹⁰¹ Among East Bengali Muslims, there was a high-status minority, which Curzon looked to. But the ‘typical Muhammadans’ were descendants of Hindu low-caste converts attracted to Islam by the opportunity to escape social degradation, who—according to Risley himself a few years later—‘do not count for much’, politically or historically.Footnote ¹⁰² Throughout Bengal, reform movements were campaigning to purify Islam, which was ‘deeply infected by Hindu superstitions’ among lower-class, uneducated Muslims, but Gait's report made no mention of any political strands in these movements, ‘volcanic’ or otherwise, which might have generated a new unity among high- and low-status Muslims.Footnote ¹⁰³ In the final analysis, therefore, none of the arguments advanced by Risley, or Curzon and Fraser, about prospective Muslim unity was ever based on ethnographic evidence or anthropological insight.

Just as importantly, the same was true of the arguments about combating the rising middle class. In British India, anthropology was axiomatically about ‘traditional’ rural society and its castes and tribes, so that the educated, urban middle class was completely unstudied; I return to this significant lacuna in the conclusion. Hence the very group that challenged imperial rule in Bengal during partition and would soon be pivotal to legislative reform was excluded from the corpus of colonial anthropological knowledge developed by Ibbetson, Risley, and their Indian Civil Service colleagues.

The Morley-Minto reforms

After Minto replaced Curzon as viceroy in November 1905, John Morley, secretary of state for India, decided that the very limited provisions for electing Indians to legislative councils in place since 1892 had to be reformed. The Indian Councils reforms, better known as the Morley-Minto reforms, became law in 1909. In all the consultations about them, the focus was on the educated, urban middle class and the need to make concessions to ‘moderate’ Congressmen in the nationalist movement.

In 1906, Risley desperately wanted to replace Arundel as home member on the Viceroy's Council, which would improve his chances of crowning his Indian Civil Service career as lieutenant-governor of Bengal. But Morley blocked Risley's promotion, although Curzon and, more relevantly, Minto supported it, with misgivings. The final decision was ‘a terrible blow’ to Risley, according to James Dunlop Smith, Minto's private secretary, who had to write to him. The story of Risley's non-promotion is murky, but plainly too many people in high places, including even Ibbetson, now distrusted him.Footnote ¹⁰⁴

The outcome was that Risley stayed on as secretary in the Home department throughout the process of working out the reforms between August 1906 and November 1909. Minto, who delegated most of the work to Risley, told Morley in 1909 that he could not have managed without him and still depended on him to complete the work, because he was ‘conversant with the story from the beginning’.Footnote ¹⁰⁵ Morley periodically became exasperated by Minto's slowness, which he sometimes blamed on Risley's bureaucratic fussiness, and the secretary of state's sense of urgency was definitely not shared by the viceroy and his government.Footnote ¹⁰⁶ In his comparison between Risley and G. K. Gokhale, the moderate Congressman with whom Morley and Minto exchanged views, Dietmar Rothermund argues that Risley was not just working sluggishly; more importantly, he ensured that the reforms reflected his conservative political views, so that in the end they ‘turned into a victory of the bureaucracy’, which particularly ‘loaded the dice against the educated class and made sure that they would not “have things all their own way”’, in Risley's own words.Footnote ¹⁰⁷

Rothermund is right about Risley's antipathy towards the pro-nationalist middle class. Thus, for example, even in 1898, Risley had publicly denounced this ‘soi-disant democratic section of a society which, from top to bottom, is essentially undemocratic’.Footnote ¹⁰⁸ Yet Rothermund also overstates Risley's conservatism and the uniformity of opinion among the British. They saw India, he writes, ‘as a conglomeration of castes, classes, and tribes’, which political representation should ideally reflect, so that Indian society could be properly protected and even restored. These objectives would be undermined by any enhancement of the middle class's constitutional position, because it was ‘socially unrepresentative and politically undesirable’. Hence ‘the natural weapons of bureaucratic rule’ had to be used to block over-enthusiastic reform schemes.Footnote ¹⁰⁹ Within the imperial government, however, there were real disagreements about reform, and although Risley's opinions are not always transparent because he was a bureaucrat, not an executive member of government, he appears to have shared Minto's opinions, which differed from Ibbetson's.

The enquiry committee set up in August 1906, which Arundel chaired, had to consider appointing an Indian as a departmental member of the Viceroy's Council, as well as changing the membership of the legislative councils of both the viceroy and the provincial governments; appointing an Indian member was extremely controversial, but the legislative councils proposals are more important for my discussion. Ibbetson was a member of the committee and Risley was its secretary. For the viceroy's legislative council, the committee proposed a substantial majority of nominated members and a minority of elected members, who would include one representative of commercial interests, seven of non-officials from provincial councils, seven of nobles and great landowners, and two of Muslims; among non-officials nominated to represent minorities, at least two should be Muslims. A similar arrangement was proposed for the provincial councils. Special dispensation for Muslims was needed because Hindus would inevitably be predominant among the rest of the elected members. Although all its members signed the Arundel committee's report, Ibbetson's own views were recorded. Ibbetson stressed the importance of ‘organising and enlisting on the side of Government the conservative elements of native society, as a counterpoise to the advanced party’, which could be achieved by selecting council members from ‘the native aristocracy and great landowners’. This class ruled over the peasantry, whom Ibbetson had sought to protect in Punjab through the Alienation of Land Act, and he was determined to safeguard the position of the whole agrarian population, which was endangered by both urban agitators and moneylenders. The ‘advanced party’ of educated people with nationalist aspirations was the fruit of British educational policy and such people deserved some ‘sympathy and recognition’, according to Ibbetson. But they were a tiny minority who ‘represent nobody but themselves’, whereas the ‘conservative elements’ were the real leaders of the ‘inarticulate masses’, whose contentment was vital for imperial rule. Moreover, the nobles and landowners favoured strong government and were contemptuous of the British compromising with ‘political agitation’, so that only the most minimal concessions should be made.Footnote ¹¹⁰

In March 1907, Risley wrote a dispatch discussing the report for Morley. It expressed concern about the educated classes’ antagonism towards the British, and advocated increased Indian participation in government to encourage ‘moderate and reasonable’ Indians and weaken extremists. The middle class was entitled to representation in proportion to both its numbers and growing influence, but not ‘a virtual monopoly of the power exercised by the Councils’. The best way to counterbalance the middle class would be separate electorates for ‘the landed and monied classes’ and for Muslims, although only appointed officials could represent the mass of ordinary cultivators.Footnote ¹¹¹ Since nobody in the government was advocating extensive reforms, let alone Indian self-rule, it may look like splitting hairs to contrast different positions. But the differences mattered at the time. Ibbetson was more dismissive about the middle class and nationalists, and more insistent on the need for support from aristocrats and landlords, than were Minto and other enquiry committee members, whose views prevailed. Certainly, Minto's government wanted to constrain the pro-nationalist middle class, but it gave their views more weight than either Ibbetson wanted or Curzon's government had done during the partition of Bengal.

Conceivably, Risley's true opinions were more conservative than Minto's, so that they were not reflected in the documents he wrote and the viceroy signed. But that seems unlikely. First, Morley and Minto often discussed Risley in their correspondence, but neither of them—despite Morley's complaints about his slowness—ever suggested that he was intriguing to prevent reform, whereas even Minto thought Ibbetson was an utter reactionary, ‘opposed to what he looks upon as unfortunate innovations’.Footnote ¹¹² Minto once told Morley that Risley ‘is broad-minded as to future possibilities’ and, later, that he was ‘thoroughly absorbed in our reforms’.Footnote ¹¹³ When the task was nearly finished in September 1908, Minto thanked Risley for all his work on their ‘great dispatch’ to Morley; Risley replied by saying that he hoped for and believed in the success of the reforms, whose ‘importance . . . is impossible to overrate’, because they will renew ‘our dealings with the people of India’ and transform ‘latent hostility into cordial fellow-service’.Footnote ¹¹⁴

Secondly, how Minto and Risley dealt with Morley's replies to the dispatches from India is telling. In London, the reforms scheme was discussed by the Council of India, which advised the secretary of state. One of its senior members was W. Lee-Warner, a former official in Bombay and an implacable opponent of Congress. During discussion of the March 1907 dispatch, Lee-Warner proposed an alternative scheme for electing council members, ‘based upon the broad principle that Indian society lives, thinks, and acts according to castes, races, and religions’. These ‘natural compartments’ should be the foundations of a new system and Lee-Warner advocated indirect election from local governments, whose representatives would be chosen by all larger castes and religious groups in their neighbourhoods.Footnote ¹¹⁵ A majority of Morley's council recommended this scheme to him. When Morley replied to the dispatch, he referred to the problem of representing ‘several classes’ of race, caste, and religion, and suggested a scheme closely resembling Lee-Warner's.Footnote ¹¹⁶

The government in Calcutta then began to consult the provincial governments about the reforms. In October 1908, after Risley's slowness had infuriated Morley, the ‘great dispatch’ was sent to London. It began by specifically mentioning ‘the separation of classes, castes, races, and communities’ that is ‘so marked’ in India, where little national unity had yet developed. It also explained that the educated classes continued to demand open, territorial constituencies in which they could expect to win elections, whereas the government—to safeguard other groups and interests—wanted separate electorates, especially for landholders and Muslims. The dispatch included revised proposals for council representation and, more emphatically than before, it described the Muslims as ‘an absolutely separate community’ from the Hindus. It concluded that the new proposals would be liberal enough for ‘advanced’ Indians, but would also enlist the support of ‘more conservative elements’.Footnote ¹¹⁷ Missing from the dispatch was any scheme of the type recommended by Lee-Warner and then Morley.

When the Council of India examined these revised proposals, Lee-Warner actually supported them. But Morley was still worried about how to define electorates and, as a compromise between territorial and separate ones, he suggested mixed electoral colleges, composed of elected representatives from different communities.Footnote ¹¹⁸ Minto rejected mixed electoral colleges, however, quoting a hypothetical example worked out by Risley to show how Congressmen, who were predominantly Hindu, could manipulate a college election at the expense of both Muslims and landholders.Footnote ¹¹⁹ In December 1908, Morley announced the plans for reform in parliament. The details were then discussed before the Indian Councils Act was published in May 1909. Proposals for the Act's regulations were sent to London during the following months. The regulations published in November defined the membership of the legislative councils of the viceroy and the provinces, making ‘special provision . . . for the representation of the professional classes, the landholders, the Mahomedans, European commerce and Indian commerce’. The professional classes—the educated middle class—would elect representatives from territorial constituencies and the other groups from their own special electorates. Electoral colleges were not part of the system, although some councils’ members would be delegates from elected district boards and municipalities.Footnote ¹²⁰

Included in a note about the regulations was an appendix on ‘Muhammadan representation’, a shorter version of Risley's earlier note reviewing Minto's and Morley's previous ‘pledges’ to the Muslims, as well as the claims advanced by Muslim organizations and their supporters. The key point in the pledges was that Muslims should be represented ‘in excess of’ their numerical strength to reflect, in addition, their political importance and need for protection as a minority. But some Muslim leaders and their parliamentary allies in London were pressing for further concessions, which could tie the Calcutta government's hands and alienate Hindu opinion. Both notes emphasized that the government should stand firm. They also contained the results of Risley's efforts to convert the pledges into workable regulations, which required complicated arrangements for elected and nominated council seats. Risley sought to devise regulations that would guarantee enough seats to Muslims to satisfy their ‘just claims’, without making the ‘excess’ unacceptably large to moderate Hindus. He contended in his earlier note that the pledges were already met in Bombay and Punjab, where the majority of enfranchised landholders were Muslims, and that in Eastern Bengal and Assam most Muslims actually lacked political importance, although their position was endangered by Hindu dominance among landholders and in municipalities. Risley slightly changed his conclusions in his later note, but stressed consistently that Muslims could not claim special representation where they were in the majority; moreover, they were ‘protected as an important minority, not as a religious community’, so that no ‘excess representation’ was justified in Eastern Bengal and Assam.Footnote ¹²¹

In different ways, Rothermund, M. N. Das, and Stanley Wolpert argue that Minto, aided by Risley, stifled Morley's hopes of more radical change.Footnote ¹²² Das and Wolpert, as well as Rajat Kanta Ray, also argue that Minto, despite Morley's misgivings, successfully built a pro-Muslim bias into the reforms, with fateful future consequences for Hindu-Muslim communal relations.Footnote ¹²³ Hermann Kulke and Rothermund contend, too, that Risley, countering Morley, ‘insisted on a representation of communities and interests in keeping with the structure of Indian society as he saw it’ and also ‘brilliantly argued the case for separate electorates for Muslims’.Footnote ¹²⁴ After referring to Risley's role in the partition of Bengal (citing Sarkar), Dirks relies on Kulke and Rothermund to conclude that Risley, in the Morley-Minto reforms process, ‘used his ethnological view of India to make one of the most influential, and deadly, decisions of Britain's colonial era’, which ‘left a bloody legacy for South Asia that continues to exact a mounting toll’.Footnote ¹²⁵

The evidence, as I read it, scarcely supports these arguments in relation to Risley. Morley was a Liberal politician, but, despite wanting the reforms quickly, he was not always radical on their substance. Furthermore, it was Morley, not Minto and Risley, who found Lee-Warner's suggestions persuasive.Footnote ¹²⁶ As Lee-Warner must have known, his contention that Indian society was made up of castes, races, and religions could almost have been quoted from Risley, although his idea of local governments differed from Risley's, which favoured village-based government to control caste separatism.Footnote ¹²⁷ Yet Lee-Warner's proposal was quickly rejected by Minto, who told Morley that although his idea of ‘reforming upwards’ from local government was a good idea in principle, ‘the pressure we have to deal with does not emanate from below, but from the ambitions of the educated class above’.Footnote ¹²⁸ A phrase about separate classes, castes, races, and communities was inserted in the second dispatch in 1908; significantly, however, no document drafted by Risley at Minto's behest ever contained suggestions for caste-based constituencies or local governments that would undermine the proposals for middle-class representation.

On the Muslim question, Risley's role and responsibility are less obvious than Kulke and Rothermund, and Dirks suppose. Separate Muslim electorates were already being favourably considered in September–October 1906, when Minto gave his crucial pledge about the Muslims’ ‘political importance’.Footnote ¹²⁹ Risley took no part in those discussions. In 1909, Risley argued against further concessions to Muslims and in his bureaucratic reasoning about ‘Muhammadan representation’, when drafting the councils’ regulations, he strove to ensure that Muslims would have an ‘excess’ representation that was not too excessive. It is true that the second dispatch to Morley in October 1908 described Muslims as ‘an absolutely separate community distinct by marriage, food, and custom, and claiming in many cases to belong to a different race from the Hindus’.Footnote ¹³⁰ This was a highly tendentious statement, of course, which could readily be used to reiterate the rightness of communal representation—as it was, for example, by the strongly pro-Muslim Dunlop Smith when he advised Minto how to criticize the Hindu Sabha, after flattering the Muslim League, in Lahore in 1909.Footnote ¹³¹ Maybe, therefore, Risley's description of the Muslim ‘community’ makes him guilty as charged by Kulke and Rothermund, and Dirks, but it is highly debatable, if only because so many other men were involved in policy making. On the other hand, Risley—unlike Morley—must have known that the description of Muslims, even if it strengthened the case for separate electorates, was mostly false. For better or worse, therefore, the one thing that Risley did not employ in writing the dispatch was his anthropological knowledge about India.