Encyclopédie internationale
des histoires de l’anthropologie

Colonial Ethnography and Theories of Caste in Late-Nineteenth-Century India

Chris Fuller

London School of Economics

Pour citer cet article

Fuller, Chris, 2023. “Colonial Ethnography and Theories of Caste in Late-Nineteenth-Century India”, in Bérose - Encyclopédie internationale des histoires de l'anthropologie, Paris.

URL Bérose : article2793.html

Télécharger en pdf

Résumé : Aucun des anthropologues coloniaux du Raj britannique n’a été largement lu par les anthropologues métropolitains, mais leurs théories rivales de la caste ont influencé de manière cruciale les premiers sociologues qui ont écrit sur la caste, notamment Célestin Bouglé et Max Weber. Même si ces derniers l’ont critiquée, la théorie des professions mise en avant dans l’Inde de la fin du XIXe siècle a indirectement mais significativement influencé les spécialistes contemporains de la société sud-asiatique. Cet article explore les écrits théoriques et ethnographiques sur les castes de trois éminents anthropologues coloniaux de l’Inde de la fin du XIXe siècle : Sir Denzil Ibbetson (1847-1908) et Sir Athelstane Baines (1847-1925), qui étaient tous deux membres de la fonction publique indienne, et John Nesfield (1836-1919), qui appartenait au service éducatif. En 1883, Ibbetson a rédigé un rapport sur le règlement des revenus fonciers dans le district rural de Karnal, au Pendjab, qui comprenait une vaste documentation ethnographique ainsi qu’une description proto-fonctionnaliste de la communauté villageoise locale. Il a également rédigé le rapport de surintendant sur le recensement de 1881 au Pendjab, dont le chapitre sur les "races, castes et tribus" est particulièrement remarquable. Baines était le surintendant du recensement de 1881 de la province de Bombay, puis le commissaire chargé de l’ensemble du recensement de l’Inde de 1891. Utilisant les données du recensement de 1881, Nesfield a écrit un livre sur les castes dans les provinces du Nord-Ouest et l’Oudh (l’Uttar Pradesh moderne). Dans ces ouvrages, Ibbetson, Baines et Nesfield présentaient des versions légèrement différentes d’une théorie professionnelle des castes, dans laquelle ils classaient les castes principalement par profession héréditaire ou fonction politico-économique. Ils ont également expliqué le système de castes “fermé" (par opposition au système de classes “ouvert” en Europe) comme le résultat de l’évolution de la division du travail spécifique à l’Inde. La théorie professionnelle des castes a été particulièrement critiquée par Herbert Risley (1851-1911), qui soutenait que le système trouvait son origine dans l’inégalité raciale et que sa caractéristique principale était la "préséance sociale" ou le classement hiérarchique, plutôt que la différenciation professionnelle. En examinant cette littérature largement oubliée, cet article ajoute une profondeur historique à un débat fondamental de l’histoire de l’anthropologie et de la sociologie.

The systematic anthropology of British India developed alongside the decennial censuses, which started in 1871–2, and its declared purpose was always both ‘scientific’ and ‘administrative’ : to contribute to modern, European scientific knowledge and also to strengthen and improve British rule. Various labels have been adopted in the literature for colonial anthropologists in India, including ‘official anthropologists’, a term that usefully indicates both their status as officials and the fact that their work – ‘official anthropology’ – was mostly undertaken on behalf of the government. From the middle of the nineteenth century until the First World War, official anthropologists had a virtual monopoly in the field, because very few Indians and very few academics carried out ethnographic or anthropological research in India. The majority of them belonged to the Indian Civil Service (ICS), the elite administrative corps of the British Raj, whose members were known as ‘civilians’, and the remainder were members of other government services or army officers. Sir Denzil Ibbetson (1847–1908), who became one of the British Raj’s most powerful civil servants and policymakers, was one of the leading figures in official anthropology’s formative years alongside his colleague Sir Herbert Risley (1851–1911). [1] Valuable contributions were also made by their near contemporaries Sir Athelstane Baines (1847–1925) and John Nesfield (1836–1919).

The most significant of these contributions and the one with the greatest influence on the anthropological understanding of India in later years was the occupational theory of caste, which Ibbetson, Baines and Nesfield proposed in slightly different versions in the late nineteenth century. In this article, I discuss this theory in most detail, but before doing so I will look at Ibbetson’s ethnographically innovative report on the land revenue settlement he carried out at the beginning of his ICS career.

Biographical Sketches

Denzil Charles Jelf Ibbetson was born in Gainsborough, a market town in Lincolnshire, on 30 August 1847. His father was a civil engineer until he was ordained in the Church of England and emigrated to Australia, where he served as a vicar in Adelaide in 1861–71. Denzil Ibbetson was educated at a grammar school in Adelaide before entering St John’s College, Cambridge, in 1865. At Cambridge, he read mathematics and was awarded his degree in 1869, but he also studied for the ICS competitive entrance examination, in which he was successful in 1868. In 1870, he passed the final examination, joined the ICS and arrived in the Punjab in December to begin his career as a junior district officer. From 1871 until 1879, he was a settlement officer in Karnal, a predominantly rural district in eastern Punjab (approximately 80 miles north of Delhi), and in 1880 he was appointed as the superintendent for the 1881 census of the Punjab province. He completed both the census report and the delayed Karnal settlement report in 1883. One year earlier, he wrote a Memorandum on Ethnological Inquiry in the Panjab (1882), which provided guidance on collecting reliable data on castes and tribes, religion, kinship, occupations, and sundry other topics. [2] The two reports, which were quickly acclaimed for their intellectual quality and scientific interest, made Ibbetson’s name as a pioneering anthropologist and ethnographer of British India. In his office in Lahore, the capital of the Punjab, he and Nesfield advised Risley in 1885 when he was starting his ethnographic inquiry to collect the data published in Risley’s The Tribes and Castes of Bengal (1891).

Ibbetson did relatively little anthropological work after 1883 as he rose through the ranks in the provincial government of the Punjab and then the central government of India. In 1902, the viceroy, Lord Curzon (1859–1925), selected him as the Home and Revenue member of his executive council, with Risley – the Home department’s secretary – as his subordinate. In 1905, Ibbetson temporarily acted as the lieutenant-governor of the Punjab and, in 1907, he was confirmed in this post, the highest he could attain as a civilian. But he soon fell seriously ill, resigned from the ICS and died in London on 21 February 1908. Ibbetson, who opposed any concessions to the Indian National Congress when he was a council member, tried to repress and crush the movement in the Punjab, which caused even Lord Minto (1845–1914), Curzon’s conservative successor as viceroy, to criticise him as excessively reactionary. Ibbetson’s imperialist politics are not discussed in this article, but it is important that his anthropological perspective and political outlook were both strongly shaped by his Punjabi experience.

Jervoise Athelstane Baines was born on 17 October 1847, the son of an Anglican vicar in Yalding, a village in Kent. He was educated at Rugby, one of England’s oldest public schools, and Trinity College, Cambridge, which he entered in 1866. He read mathematics and was awarded his degree in 1869. He was successful in the ICS entrance examination in 1868 and passed the final examination in 1870 ; he then joined the ICS, arrived in the Bombay Presidency in November and was first appointed as an assistant to the Collector of Surat. Baines held a wide range of posts in his career, though he never reached high office in the ICS, but he was the 1881 census superintendent for Bombay and the commissioner in charge of the 1891 census of India. After retiring to England in 1895, he was a local council member, first in London and then Oxfordshire ; he was also active in the Royal Statistical Society and served as its president in 1909–10. Baines died on 26 November 1925.

John Collinson Nesfield, who was born on 14 August 1836, was also an Anglican clergyman’s son. He was educated at Highgate grammar school in London and Merton College, Oxford, where he was awarded a degree in 1860. Nesfield (who was not a member of the ICS) began his career as an educationalist in India in 1867 and spent most of it in the North-Western Provinces and Oudh (modern Uttar Pradesh). He developed a serious interest in Indian anthropology and published several ethnographic papers and a book on the caste system, which I discuss below. [3] He also helped Risley to collect anthropometric data for his tribes and castes inquiry. Nesfield retired from India in 1894 and died on 28 June 1919.

The Karnal Settlement Report

In Ibbetson’s time, the Punjab province, which was partitioned between India and Pakistan when British rule ended in 1947, stretched from the North-West Frontier and the border with Afghanistan in the west to Delhi and the river Yamuna in the south and east ; Lahore was roughly in the province’s centre. In the west, the population was predominantly Muslim and, in the east, predominantly Hindu with a significant Sikh minority. The Punjab was a ‘non-regulation’ province whose laws governing the family, marriage, inheritance and related issues, as well as religious matters, were in principle based on customary personal law, which varied among different communities, whereas in ‘regulation’ provinces, such as Bengal, uniform Anglo-Indian law mostly prevailed. The real differences between the two systems are easily exaggerated ; nonetheless, a district officer in the Punjab ideally administered the local people more personally and flexibly than his counterpart in a regulation province, who sought to apply rules and regulations consistently to all. In short, traditional patrimonialism, rather than bureaucratic rationality, was the ideal norm in the Punjab. Ibbetson had his own formative experience as a young civilian during eight years as a settlement officer in rural Karnal, where he became well acquainted with the local people, especially landlords and peasant farmers. Throughout the late-nineteenth-century Punjab, many landowners were heavily indebted to urban moneylenders and often forfeited their land when taken to court over unpaid debts. Ibbetson, who strongly sympathised with the landowners, successfully argued that the government should defend this group, who tended to be conservative, pro-British ‘loyalists’, not the moneylenders or their urban lawyers, who commonly supported the Congress. Ibbetson, in sum, was predisposed to see Indian society, both anthropologically and politically, through the eyes of rural Punjabis, especially landowners, whereas Risley, for example, contrastingly saw it through those of the bhadralok, the high-caste, educated, urban middle class of Bengal.

Throughout British India, agrarian regions were ‘settled’ soon after they came under the colonial government’s control so that the land revenue could be collected. In Bengal, the Permanent Settlement was imposed in 1793, but in other provinces in later years, settlements were normally revised periodically, mainly to update earlier records and to take account of both material changes (such as improved irrigation systems) and alterations in the assessment rate fixed by the government. It was standard practice at a settlement to survey each village within a district to make both a field map and a register listing each field’s owners and tenants, which included information about soil quality, crops grown and other salient details. A second register of all landowners was also drawn up, which recorded their often complicated proprietary rights in plots of land, as well as their revenue liabilities to the government. Making the maps and the registers were usually difficult, time-consuming tasks, especially in villages with many separate fields ; often, too, local landowners were in dispute with each other and ready to challenge a settlement officer’s decisions. Ibbetson’s task in Karnal was to revise the settlement done in part of the district in the 1840–50s and it took him eight years, far longer than usual, partly because he wanted to collect more information than strictly necessary. The Punjab government mildly criticised Ibbetson’s slowness, but largely absolved him of blame. Moreover, it declared that the report was ‘one of unusual interest and excellence’ and its ‘chapters on the social customs and habits of the people [were] particularly full and interesting’. [4] Official anthropologists at the time, as well as their academic successors, have reached the same conclusion, but it is a long report and I will mention only a few examples from Chapters 6 on ‘Tribal Organization of the People’ and 7 on ‘The Village Community’. Ibbetson sometimes used the terms ‘tribe’ and ‘caste’ indifferently in his Karnal report, but this rarely caused confusion or detracted from its ethnographic and theoretical quality.

In the section on ‘local organization of tribes’ in Chapter 6, to give one illustrative example, Ibbetson stated that tribes were primarily divided into thapas (or thambas). When a tribal group possessing a tract of land divided, one part of the group established a new village. After repeated divisions, the tract contained numerous different villages that had sprung from the original, parental one. ‘The group of villages so bound together by common descent form a thapa’, he explained, and the original village was ‘still acknowledged as the head’. Thus for instance, when a headman died, his heir was installed at an assembly of all these villages and had the original village’s turban tied on his head first ; similarly, when the villages’ Brahmans were fed, those from the original village ate first and received double fees. Brahmans were generally regarded as grasping and Ibbetson was told in one village that its inhabitants changed their thapa because they had too many Brahmans to feed, but people in the original village said that was impossible, because ‘a son may forget his sonship ; but not a mother her motherhood’. Importantly, too, a thapa could include ‘strangers’, so that, for example, a man without sons could settle his son-in-law in his own village as his heir or settle a friend by giving him a share of his land. An heir of this kind, who was called the sonless man’s ‘earth-brother’ (bhunbhai), would become a member of his tribe and thapa by ‘the fiction of common descent’. [5] Ibbetson also explained that all tribes, such as the Jats (later described as the pre-eminent agricultural ‘caste’) or the Gujars (‘a notorious thieving tribe’) were endogamous, but each of them was divided into exogamous gots or gentes (clans), which supposedly included all descendants of one common ancestor. He also discussed alternative explanations for gentes’ names and in some cases thought there was evidence of ancient matrilineal descent or of totemism. [6]

Exogamy, endogamy and totemism were important, controversial topics for metropolitan evolutionist anthropologists who became preoccupied with the study of primitive society and its evolution in the late nineteenth century. But for the majority of them, kinship in general as opposed to marriage in particular, and the relationship between descent and locality (as manifested by thapas in Karnal), were not especially significant, although they certainly were for Henry Maine (1822–88). After publishing Ancient Law in 1861, Maine became the Law member of the viceroy’s council in India in 1862–9 and afterwards wrote Village-Communities in the East and West (1871), which he first delivered as lectures in Oxford. Ibbetson, in common with other official anthropologists in the Punjab, was greatly influenced in his thinking by Maine. [7]

In Chapter 7, Ibbetson discussed the proprietary body of landowners and the classification of land tenures, property rights and tenancy, which the British had originally misunderstood. He also carefully described the structure of the joint family, inheritance rules, and the restrictions on alienating family property, as well as the common expenditure and income of the village, the organisation of land revenue payments, and other aspects of the village’s economic affairs. Discussed in considerable detail, too, were the village moneylenders, who were much abused but actually provided an indispensable service, and the village accountants, who were generally trusted by local people. The village menials (kamin), who formed a vital part of the community, did all the work that was not paid for by the piece, but instead by a customary share of the harvested produce, or a fixed allowance on each plough or Persian wheel (used for irrigation), that was given to them by the cultivators. The kamins included the Chamars (tanners and cobblers), who were also agricultural labourers and were ‘by far the most important class of menials’, as well as the carpenters, blacksmiths, potters, sweepers, bearers or porters, barbers, various other tradesmen, and lastly the lowly musicians. [8] Karnal’s local economic system was, of course, a variant of the so-called jajmani system described by modern anthropologists of the Indian village.

Hardly any of the material in Chapter 7 and only some of that in Chapter 6 belonged to late-nineteenth-century ethnology or anthropology as they were normally conceived, whereas they commonly were included in updated form in the standard South Asian village ethnographies written in the mid- to late twentieth century. More conventional in their subject matter were Chapter 8, which was mostly about domestic living and life-cycle rituals, and Chapter 9, which discussed popular religion, primarily Hinduism because four-fifths of Karnal’s inhabitants were classified as Hindus. But Ibbetson’s account of the village community in Karnal was most impressive and original because, under the influence of Maine and his proto-functionalism, he showed how its component parts fitted together in a system in a manner resembling a functionalist ethnography written in the twentieth century. Hence the Karnal report often reads more like a modern text than the majority of colonial anthropological and ethnographic works on India, which generally used information collected in extensive surveys to address quintessentially Victorian questions like the evolutionary relationship between totemism and exogamy.

The 1881 Punjab Census Report

Ibbetson’s voluminous 1881 census report on the Punjab had a substantial chapter on religion, but one-third of it was a chapter on ‘Races, Castes and Tribes’, later republished separately, which became a foundational text for the anthropology of British India. [9] Ibbetson’s census report drew on his knowledge of Karnal, but the bulk of his information came from settlement reports, official publications and other works.

Introducing Chapter 4 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. Classifying the people by their religion was extremely difficult and defining ‘Hindu’ was impossible. ‘Creed’, he wrote, is ‘rather a social than a religious institution’, and ordinary people’s religious practices, which were typically lax and eclectic, were very similar among Hindus, Sikhs and Muslims, ‘especially among the ignorant peasantry’. [10] Comparable comments were later made by other official anthropologists in every province of India. Ibbetson, who provided more ethnographic data on popular Hinduism than the other religions, particularly insisted that the ‘practical religious belief and life’ of almost all Hindus was far removed from the ‘ideal Hinduism’ described in books. Moreover, he asserted, every Punjabi village could supply far more valuable material on ‘primitive superstitions’ than Edward B. Tylor (1832–1917), John Lubbock (1834–1913) or John F. McLennan (1827–81) ever laboriously collected from ‘forgotten descriptions of little known tribes’. [11] But despite his warnings about misrepresenting religion, Ibbetson sweepingly and unconvincingly generalised about Hindus, Sikhs and Muslims as social groups, whose ‘external characteristics’ were primarily ‘thrift’, ‘bravery’ and ‘pride’, respectively. For reasons that are unclear, he also appeared to harbour prejudice against Muslims, especially converts from Hinduism, whereas most Britons in the Punjab tended to prefer Muslims and Sikhs to Hindus.

At the beginning of Chapter 6 on races, castes and tribes, Ibbetson insisted that ‘the whole basis of diversity of caste is diversity of occupation’, although he extended the meaning of ‘occupation’ to the broader functions of ‘ruling’ and ‘being ruled’. [12] He acknowledged that he used the terms ‘caste’ and ‘tribe’ loosely (and actually even more inconsistently than in his Karnal report), but he also insisted that ‘it is almost impossible to define a caste and difficult to define a tribe’ as discrete, bounded groups. [13] In general, he used ‘race’ as a synonym for ‘people’ and gave it no particular ethnological meaning.

Ibbetson identified different types of caste in the Punjab in several ways. [14] In his tables showing their geographical distribution, he primarily divided the province’s castes and tribes into ‘three great groups’ and discussed them further in his text. The first was the ‘landowning and agricultural group’ or ‘agricultural castes’, which made up half the province’s population and was ‘even more important socially, administratively, and politically than it is numerically’, a comment reflecting the great importance accorded to this group as the ‘loyalist’ bulwark of British power by Ibbetson and other officials. The first group was also subdivided into six sections : the ‘frontier races’ (notably the Biloches and Pathans) ; the Jat race ; the Rajput and allied tribes ; the locally restricted ‘minor dominant tribes’ ; the similar ‘minor agricultural tribes’ ; and the ‘foreign races’ (such as Shekhs and Mughals) claiming non-Indian ancestry. The second group comprised ‘the priestly, ascetic, professional, and mercantile castes’, which were generally described as ‘castes’, except for the Muslim ascetic orders. The third group of ‘menial castes’ – the ‘lower strata of society’ denoted by no consistent terminology – included ‘the vagrant, criminal, and gypsy tribes, the village menials [such as the Karnal kamins], and the industrial classes [artisans]’. [15] To organise his detailed descriptions, however, Ibbetson divided the communities rather differently into five sections : Biloch, Pathan and allied races ; Jat, Rajput and allied castes ; minor landowning and agricultural castes ; religious, professional, mercantile and miscellaneous castes ; and vagrant, menial and artisan castes. Ibbetson himself admitted that his census report was written too hastily and inadequately revised, but his discrepant terminology and classification also partly reflected the difficulty of condensing the Punjab’s extensive social and cultural diversity. This diversity was amply illustrated by the report’s copious ethnographic material, which Horace A. Rose (1867–1933) greatly relied on when later compiling the tribes and castes handbook for the Punjab as part of the ethnographic survey of India that Risley inaugurated in 1901. [16]

The Occupational Theory of Caste

Among the 1881 provincial census reports, Ibbetson’s was outstanding not only for its ethnography, but also because he presented a new occupational theory of caste. Baines in his census reports and Nesfield in his book set out slightly different theories that I also consider.

Ibbetson and Baines worked independently at the 1881 census, as far as I know, but they both classified castes by occupation and explained the caste system in relation to the evolution of the division of labour. Like nearly all Victorian writers on India, including Nesfield and Risley, they also believed the subcontinent was invaded in ancient times by fair-skinned Aryans, who were more ‘civilised’ than the ‘primitive’ indigenous inhabitants, mostly dark-skinned Dravidians, whom they conquered, displaced and subordinated as their racial inferiors. Although there were differences between them, Ibbetson, Baines and Nesfield all accorded much less weight to the ‘Aryan invasion’ thesis than Risley, who explicitly developed his racial theory of caste as a critique of their occupational theories. [17]

According to the evolutionist social theories that were dominant in the late nineteenth century, the division of labour becomes more complex as the economy and society develop. Thus Ibbetson wrote that primitive societies are divided into tribal groups and no diversity of occupation exists, whereas 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 Herbert Spencer (1820–1903), but did not name his ‘modern authorities’ – the guild, like the tribe, was originally based on common descent, so that ‘the fundamental idea’ underlying it was ‘the hereditary nature of occupation’. As we have seen, for Ibbetson, caste diversity was based on occupational diversity, and ‘occupation’ or ‘function’ was defined broadly to include political dominance and subjection – ‘ruling’ and ‘being ruled’ by chiefly landlords, for example – 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’. The difference between social evolution in India and Europe, Ibbetson stressed, was ‘a difference of degree rather than of kind’, brought about in India by the singular exaltation of the Brahman priesthood, which turned into a hereditary group. The Brahmans then gave religious sanction to the hereditary principle, which was generalised throughout Hindu society. In short, occupational groups became closed, hereditary, ranked castes in India, instead of open guilds and eventually economic classes as in Europe. In the Punjab, however, the picture was complicated by the tribal and trades-guild types of caste, which Ibbetson explained as the outcome of Muslim influence on the pre-existing Hindu system. [18]

Ibbetson and Baines, who cited Auguste Comte (1798–1857) on ‘differentiation of employments’, presented their arguments about the division of labour somewhat differently, but agreed on the main points. [19] They also devised caste classification schemes in their census reports which were defined by traditional, hereditary occupation or function, so that they were consistent with their theories. But these schemes differed, partly because the Punjab was so different from Bombay. The main part of Bombay province was its western region (in modern Maharashtra and Gujarat), where 84 per cent of the population was Hindu ; the other non-contiguous part adjoining the Punjab was Sind (in modern south Pakistan), whose population was 78 per cent Muslim. Baines’s discussion of caste was almost entirely about the first region and its Hindus, not Sind and its Muslims, so that his classification scheme – unlike Ibbetson’s – was a predominantly Hindu one comprising thirteen ‘classes’ or groups of castes primarily defined by their occupations : Brahmans ; Rajputs ; writers ; traders ; artisans ; agriculturalists ; shepherds and other pastoralists ; fishers and seafarers ; personal servants ; minor professions ; devotees and religious mendicants ; depressed and unclean castes ; very low-status labouring and miscellaneous communities. After Brahmans and Rajputs (representing Kshatriyas), which are classically the highest Hindu caste categories or varnas, the other classes were listed in an order that was sufficiently ‘indicative of social position’, although it might be wrong in a few particular cases. Separate from these thirteen classes were three others : aboriginal and forest tribes, Jains, and Muslims. [20]

Baines, unlike Ibbetson, wrote concisely and included relatively little ethnographic detail in his census report. After the mid-1880s, Ibbetson did not do much anthropological work. Baines, on the other hand, succinctly refined the occupational theory when he wrote his report as the 1891 census commissioner and produced a classification scheme covering the whole of India for castes and tribes, and non-Hindu minorities, which was mainly based on ‘function’ – primarily traditional occupation – but also on ‘position … in social estimation’ or hierarchical rank, which varied considerably and had to be investigated locally or regionally. Baines claimed his scheme was the most practicable for an all-Indian classification, which in his report comprised twenty-one classes or groups, of which eleven were primarily occupational ; the rest included ‘forest tribes’, ‘indefinite’ castes, and religious and racial minorities. But he also acknowledged that his scheme had flaws, especially when function was given excessive weight relative to rank, so that, for instance, Brahmans were not placed in the first class, as their superior status would indicate, but in the ninth class of ‘professionals’ under ‘priests’, alongside the Muslim ulama and various ascetics of uncertain status. [21]

Twenty years later, while in retirement in England, Baines published Ethnography (Castes and Tribes) (1912) as a volume in a German encyclopaedia. Modern scholars have rarely mentioned this work, which was an impressive feat of ethnographic synthesis at the time. Baines’s text was primarily based on the 1901 census data and consisted of a kind of all-India glossary of castes and tribes arranged into five main groups by occupation or function. The first of these was ‘special groups’, comprising Brahmans, Rajputs, trading and writer castes, and religious devotees and mendicants, with Brahmans now ranked first, unlike in 1891. The second group was ‘village community’, which included castes of dominant landholders and peasants, as well as labourers, artisans and other menials. The third group was ‘subsidiary professional castes’, such as bards and temple servants ; the fourth ‘urban castes’, mainly tradesmen ; and the fifth ‘nomadic castes’, including carriers and shepherds. Baines’s sixth group was the ‘hill tribes’ and the seventh was ‘Muslim race titles’, such as Sheikh or Pathan. [22] In the ‘historical’ account of caste preceding the ethnographic description, however, he did not mention the evolution of the division of labour and paid very little attention to occupation. Instead, he emphasised the great importance of ritual purity and the way in which India’s uniquely powerful ‘sacerdotalism’ had enabled the Brahmans to prescribe elaborate rules for conduct that were accepted as ideal even by the castes ranked below them. Indeed, Baines’s exposition of the importance of race, hierarchy and Brahmanism moved him considerably closer to Risley’s position than he had been earlier. [23]

Nesfield’s version of the occupational theory was based on his analysis of the caste data collected in the 1881 census of the North-Western Provinces and Oudh. [24] Nesfield reclassified both the ‘casteless’ tribes and the Hindu castes listed in the census report by arranging them according to ‘function or occupation’ into nine groups mainly corresponding to the economic ‘stages’ of human history, which he also glossed as ‘progressive stages of culture’ and identified as hunting, fishing, pastoral, agricultural, commercial, industrial and, finally, religion and the arts of civilisation. The first group were ‘backward and semi-savage’ tribes. The second comprised castes connected with land, which were divided into five subgroups corresponding to economic stages : the first four were allied with hunting, fishing, pastoralism and agriculture, respectively, and the fifth was ‘landlords and warriors’. The third group included artisan castes, subdivided into ‘preceding’ or ‘coeval with’ metallurgy ; the fourth, fifth and sixth were the trading, serving and priestly castes, and the seventh included religious orders, such as ascetic renouncers. Nesfield contended, too, that a caste’s rank depended on whether it represented an economic activity belonging to ‘an advanced or backward stage of culture’ – so that cultivating castes were higher than pastoralist ones, for instance – and he concluded that function was therefore the true foundation of the hierarchical caste system. [25] Because rank ‘in the scale of human progress’ also determined caste rank and there was ‘unity of the Indian race’, no significant physical differences between the Aryans and the indigenous ‘Aboriginal’ inhabitants still existed. [26] Caste, moreover, was ‘a purely secular institution, and religion has had nothing to with it’, so that followers of different religions or sects may belong to the same castes, though it is only because arrogant Brahmans imposed endogamy on the ‘mesmerised’ lower classes that hereditary occupational classes were transformed into castes in India. [27] As the reader will see, Nesfield’s version of the occupational theory was more rigidly determinist than either Ibbetson’s or Baines’s.

Four features of the occupational theory should be highlighted as especially important. First, the theory explains the caste system by identifying it as a form of the division of labour within a society that was relatively advanced or ‘civilised’, except for a minority of ‘primitive’ hill and forest tribes, and by comparing it with the European form, which was based on ‘open’ economic classes, rather than ‘closed’ castes. Secondly, religion was only a secondary aspect of the caste system, because castes were primarily occupational or functional groups, to which people from different religions could all belong. Thirdly and relatedly, neither race nor religion had been determinant factors in its development. Fourthly, however, all the writers I have discussed also believed that the Brahmans in ancient times had successfully asserted their own superiority within society, as well as the rightfulness of their Brahmanical doctrine that the socio-religious order was properly constituted by hereditary occupational groups ranked in a hierarchy extending below their own caste. This special status of the Brahmans was what made the Hindu caste system unusual or even unique.

The Rival Theories of Caste and Their Legacy in Modern Anthropology

Risley, as mentioned above, criticised the occupational theory of caste and argued instead, first, that its origins lay in the racial inequality between Aryans and Dravidians and, secondly, that the system’s defining feature was the hierarchical ranking of castes according to ‘social precedence’ as recognised by Indian popular opinion. As the 1901 census commissioner, Risley therefore decided to classify castes primarily by social precedence and discarded Baines’s 1891 system, which he criticised as a ‘patch-work classification in which occupation predominates’, albeit modified by various other factors. Risley respectfully criticised Ibbetson and Nesfield by name and quoted their work at some length, but for some reason did not mention Baines’s name at all. [28]

As I explained in my BEROSE article on Risley, the most notable Western scholars to take a real interest in the official anthropology of India during the colonial era were the Indologist Émile Senart (1847–1928) and two pioneering sociologists, Célestin Bouglé (1870–1940) and Max Weber (1864–1920). These three scholars were particularly important because they were able to comprehend caste as a system qua system by looking beyond the classificatory empiricism of official anthropology, while also seizing on the mass of detailed, contemporary evidence that it contained. All three referred to Risley extensively, whereas they cited Ibbetson and Nesfield less frequently and Baines only rarely, so that these three men’s contemporary legacy is harder to detect than Risley’s. Senart, however, relied fairly heavily on Ibbetson’s 1881 Punjab census report and Nesfield’s book for much of his evidence on the contemporary system in his 1896 book on caste ; he also devoted one chapter to Ibbetson’s and Nesfield’s occupational theories of caste and another on Risley’s racial theory. Senart, who criticised each author for dogmatic overgeneralisation, especially Nesfield, also specifically objected to the occupational theory on several grounds, for example, that members of one caste often pursue multiple occupations, and one occupation is often taken up by people from different castes. [29]

In his essays on the caste system first published in 1908, Bouglé cited Senart and Nesfield repeatedly, though often critically, and Ibbetson and Baines occasionally, when he discussed occupational specialisation and the relationship between caste and family. But Bouglé also made considerable use of The Tribes and Castes of the North-Western Provinces and Oudh (1896) by William Crooke (1848–1923), who in turn referred to Ibbetson and Nesfield extensively and broadly endorsed their occupational theory of caste. [30] Weber discussed caste in his essay on Hinduism and Buddhism, originally published in 1917. In a bibliographical note, Weber listed Nesfield’s book as a useful older source and praised Baines’s Ethnography as the ‘best modern work’, but did not mention Ibbetson’s writings ; in his main text, however, he appears to have drawn on none of these men’s material directly. On the other hand, like Bouglé, Weber relied on sources, especially the census reports, which were themselves often dependent on the work of earlier official anthropologists, including Ibbetson, Nesfield and Baines, as well as Risley. Implicitly or explicitly, therefore, the influence of the rival occupational and racial theories of caste was present throughout the work of both early sociologists.

The official anthropologists of India, as I wrote in my article on Risley, have never been widely read by metropolitan social anthropologists. Nonetheless, through early European sociology and its pioneers, they have indirectly exerted a notable intellectual influence on post-colonial social scientists researching South Asia. It is difficult to construct a full history of ideas about caste since the late nineteenth century, but both Bouglé and Weber plainly found key evidence for their arguments about caste hierarchy in Risley’s work. Yet it is significant, too, that both sociologists developed their own arguments partly by rejecting the occupational theory. Thus Bouglé critically discussed Nesfield at length, as well as Senart more briefly, to reach his conclusion that among Hindus ‘religious beliefs above all, rather than economic tendencies, … fix the rank of each group’. [31] Weber, in his discussion of ‘caste and guild’, cited no specific authors, though he did quote some evidence from the 1901 census reports, but he clearly had the occupational theory in mind when he stated that caste and guild or any other ‘occupational association’ are fundamentally different types of group. And after examining ‘caste and status group’, he concluded that : ‘The caste order is oriented religiously and ritually to a degree not even partially attained elsewhere’. [32]

Bouglé, of course, inspired Louis Dumont (1911–98) who, in Homo Hierarchicus (1970), called his predecessor’s conclusion about the priority of religious beliefs ‘more or less our point of departure’ in discussing caste as the Indian form of the division of labour ; in this context, though rather allusively, Dumont cited Weber in support of his position as well. [33] Dumont’s ideas on caste were highly influential, especially in the late twentieth century, but that was always partly because they provoked so much critical attention, typically for exaggerating the importance of religion vis-à-vis economics and politics, and of caste and hierarchy vis-a-vis occupation, class and inequality more generally. All competent scholars recognise, of course, that castes are not solely occupational groups and that no one-to-one relationship between the caste system and the division of labour has ever existed, even in ‘traditional’ villages on the margins of the modern market economy. Nevertheless, Dumont’s critics, as well as a much wider range of modern anthropologists and sociologists, would also agree on the importance of political and economic factors for comprehending the caste system and, even if they have never read colonial anthropologists, many of the arguments they advance are the indirect and normally unacknowledged legacy of the occupational theory of caste formulated in late-nineteenth-century British India. Like Risley and other official anthropologists, Ibbetson, Nesfield and Baines have also left their mark on the modern anthropology and sociology of South Asia.


Books and Articles

Baines, J. Athelstane. 1893. ‘Occupation and Caste in India, as Shown in the Last Census.’ Journal of the Society of Arts 41 (March 31) : 457–75.

Baines, J. Athelstane. 1912. Ethnography (Castes and Tribes). Strasbourg : Karl J. Trübner.

Bouglé, Célestin. 1971 (1908). Essays on the Caste System (trans. with introduction by D. F. Pocock). Cambridge : Cambridge University Press.

Crooke, William. 1896. The Tribes and Castes of the North-Western Provinces and Oudh (4 vols.). Calcutta : Government Printing.

Dewey, Clive. 1991. ‘The Influence of Henry Maine on Agrarian Policy in India’, in Alan Diamond, ed., The Victorian Achievement of Henry Maine, 353–75. Cambridge : Cambridge University Press.

Fuller, Chris. 2016. ‘Anthropologists and Viceroys : Colonial Knowledge and Policy Making in India, 1871-1911’. Modern Asian Studies 50, 1 : 217–58.

Fuller, Chris. 2022. ‘Ethnography and Racial Theory in the British Raj : The Anthropological Work of H. H. Risley’, in BEROSE International Encyclopaedia of the Histories of Anthropology,

Goetzmann, Marc. 2020. ‘Le juriste-anthropologue du British Raj : Sir Henry Sumner Maine et son oeuvre’, in BEROSE International Encyclopaedia of the Histories of Anthropology,

Ibbetson, D. C. J. 1883. Report on the Revision of Settlement of the Panipat Tahsil and Karnal Parganah of the Karnal District, 1872-1880. Allahabad : Government of the Punjab.

Ibbetson, D. C. J. 1916. Panjab Castes [reprint of Report on the Census of the Panjab, 1881, vol. 1, chap. 6]. Lahore : Government Printing.

Maine, Henry. 1861. Ancient Law. London : John Murray.

Maine, Henry. 1871. Village-Communities in the East and West. London : John Murray.

Nesfield, J. C. 1885. Brief View of the Caste System of the North-Western Provinces and Oudh. Allahabad : Government Press.

Risley, H. H. 1891. The Tribes and Castes of Bengal : Anthropometric Data (2 vols) ; Ethnographic Glossary (2 vols.). Calcutta : Bengal Secretariat Press.

Risley, H. H. 1915. The People of India (rev. edn by W. Crooke). Calcutta : Thacker, Spink ; London : Thacker.

Rose, H. A. 1911-19. A Glossary of the Tribes and Castes of the Punjab and the North-West Frontier Province, 3 vols. Lahore : Superintendent of Government Printing (vol.1, 1919) ; Civil and Military Gazette Press (vol. 2, 1911 ; vol. 3, 1914).

Senart, Émile. 1930 (1896). Caste in India : The Facts and the System (trans. E. Denison Ross). London : Methuen.

Weber, Max. 1967 (1917). The Religion of India : The Sociology of Hinduism and Buddhism (trans. Hans H. Gerth and Don Martindale). New York : Free Press.

Census Reports

Operations and Results in the Presidency of Bombay [Census, 1881], vol. 1, Text, by J. A. Baines. Bombay, 1882.

Report on the Census of the North-Western Provinces and Oudh, by Edmund White. Allahabad, 1882.

Report on the Census of the Panjab, 1881, vol. 1, Text, by D. C. J. Ibbetson. Lahore, 1883.

Census of India, 1891, General Report, by J. A. Baines. London, 1893.

Census of India, 1901, vol. 1, India, pt. 1, Report, by H. H. Risley and E. A. Gait. Calcutta, 1903.

[1On Ibbetson and Risley, see Fuller 2016 ; on Risley, see also Fuller’s BEROSE article (2022).

[2Ibbetson 1882 ; 1883 ; Census 1881 Panjab.

[3Nesfield 1885.

[4Comment by Lieutenant-Governor of the Punjab, in Ibbetson 1883, pp. 12–13 (prefatory pages).

[5Ibbetson 1883 : 74–6.

[6Ibbetson 1883 : 77–9, 81, 83.

[7On Maine’s influence, see Dewey 1991 : 363–73 ; see also Goetzmann 2020, esp. pt. 5. For official anthropologists outside the Punjab, such as Risley in Bengal, Maine the scholar-administrator was an inspirational figure, but his real influence on their anthropological thinking was quite minimal.

[8Ibbetson 1883 : 92–119.

[9Census 1881 Panjab ; Ibbetson 1916.

[10Census 1881 Panjab, pp. 100–101.

[11Census 1881 Panjab, p. 124.

[12Census 1881 Panjab, p. 173 ; Ibbetson 1916 : 3.

[13Census 1881 Panjab, p. 188 ; Ibbetson 1916 : 33.

[14Census 1881 Panjab, p. 178 ; Ibbetson 1916 : 13.

[15Census 1881 Panjab, pp. 185–7 ; Ibbetson 1916 : 26–30.

[16Rose 1911–19.

[17See discussion of Risley on ‘caste and race’ and ‘caste and hierarchy’ in Fuller 2022.

[18Census 1881 Panjab, pp. 172–4, 176–7 ; Ibbetson 1916 : 2–5, 9–10.

[19Census 1881 Bombay, pp. 120–23.

[20Census 1881 Bombay, pp. 126–7 ; cf. slightly different classification in table of castes on pp. 118–19.

[21Census 1891 India, pp. 189–90 ; Baines 1893 : 466.

[22Baines 1912 : 24–145.

[23Baines 1912 : 10–18.

[24Census 1881 North-Western Provinces and Oudh.

[25Nesfield 1885 : 3, 7–87.

[26Nesfield 1885 : 3–4.

[27Nesfield 1885 : 95, 115.

[28Census 1901 India, pp. 537–8, 549–51 ; Risley 1915 : 109–12, 263–7 ; ‘caste and hierarchy’ in Fuller 2022.

[29Senart 1930 : 8–88 (pt. 1), 156–74 (pt. 3, chaps. 2 & 3).

[30Crooke 1896, 1 : cxxxix–clix (chap. 3, ‘The Occupational Form of Caste’).

[31Bouglé 1971 : 39.

[32Weber 1967 : 39, 44.

[33Dumont 1970 : 26–7 & 289 n.41a, 105.