Religion, science, and empire adds to the literature on European understandings of Indian society, culture and religion during the period of British rule, which has grown considerably since Bernard Cohn wrote his pioneering papers on the history of the study of India, and the role of the colonial census.^[1]  Most anthropologists of South Asia are probably familiar with the main lines of enquiry and argument in this literature: for example, the debate about the colonial construction of caste.  Anthropologists of the rest of the world, on the other hand, probably reckon that this literature has nothing to do with them.  That view, I suspect, stems partly from India’s near-absence from histories of anthropology, especially British social anthropology.  India is of course mentioned when Henry Maine and Aryan ‘ancient society’, or W. H. R. Rivers’s pioneering monograph on the south Indian Toda tribe, are discussed.  But otherwise the subcontinent mostly vanishes from anthropological history between the decline in Maine’s influence around the 1870s and the development of modern South Asian anthropology and sociology, which began after the Second World War and Independence in 1947.

According to George Stocking, the pre-eminent historian of British anthropology, the study of India, ‘like India in the British Empire, had always had a somewhat separate status’ in pre-modern social anthropology (1987: 296).  Stocking later mentioned the ‘well-established tradition of governmental ethnological work in India’, but his general conclusion about government anthropologists is that, especially as ‘producers of anthropological theory’, they were in ‘geographically distant, institutionally precarious, and intellectually marginal positions outside the metropolitan mainstream’ (1995: 381, 391).  To describe India, by far the most important and populous territory in the British empire, as ‘somewhat separate’ in status is curious, however.  Moreover, although official anthropologists in India were marginal vis-à-vis ‘the metropolitan mainstream’, they did produce a vast amount of ethnography and some significant theory, so that to marginalize them in historiography merely reproduces metropolitan bias and leads to the kind of Whiggish history ‘somehow leading inexorably up to modern … anthropology’ that J. W. Burrow rightly criticised (1966: 18-19).  India’s place in a more complete history of British anthropology, including its connection with imperialism and colonialism, is not a theme in Peter Gottschalk’s new book, but much of what he says is relevant.

Gottschalk, a professor of religion at Wesleyan University in Connecticut, published Beyond Hindu and Muslim: multiple identity in narratives from village India in 2001, an impressive monograph based on ethnographic fieldwork in 1994-5 in rural Bihar, north India.  In the main village and the ‘nexus’ of others linked to it, which were given the pseudonym Arampur, almost 90 per cent of the inhabitants are Hindus, the rest being Muslims.  As the book’s title indicates, people in Arampur identify themselves and others in a flexible variety of ways, not just as Hindu or Muslim, although the growing strength of Hindu nationalism was tending to sharpen those polar religious identities in the 1990s.  In Religion, science, and empire – in which the real name, Chainpur, is used – Gottschalk briefly describes the village in 2009 and the changes since the 1990s, especially improved education and the impact of the school curriculum in history and science.  He also refers to local nationalists with typically dogmatic views about Hindus and Muslims; nonetheless, he insists, ‘most Chainpur residents color their lives outside the two-dimensional borders drawn by religious communalists and simplistic scholarship’ (p. 303).

Until the final chapter 8, however, Gottschalk’s new book is about Chainpur in the past, or more exactly about changing understandings of India during the British raj as illustrated by different accounts of Chainpur, left to us by a succession of British and Indian travellers, officials and scholars.  So many people have visited Chainpur, and hence made Gottschalk’s book possible, because it is unusual: the village nexus is larger than most rural settlements in the area, it is close to the three major communication routes across northern India (the river Ganges, the Grand Trunk road and the main railway line), it is quite close to the sacred city of Benares, and it contains some notable religious sites, including a Hindu temple and Muslim tombs, which have always attracted pilgrims.  Even so, accounts of Chainpur have to be supplemented and put into a larger context by other writings about the surrounding district (Shahabad in the colonial period) and the wider region that is now western Bihar and eastern Uttar Pradesh.  Not surprisingly, some accounts are much more interesting and informative than others that were presumably included only because they actually are about Chainpur.  It’s a pity, for example, that a rare example of a personal journal by a British district officer, written by G. M. Ray in 1938-40, should amount to little more than rancorous, racially prejudiced and uninformative complaints about Bhabua, the miserable town nearby, to which he’d been posted as a junior member of the Indian Civil Service (pp. 154-61).^[2]

At the end of his first chapter, Gottschalk comments that religion always ‘remained a marker of difference for Britons in their perception of South Asians’ (pp. 51-2).  At first it was the distinction between Christians, and India’s Hindus and Muslims, but later between the more religious Indians and the more secular, rational Europeans.  The principal subject matter of Religion, science, and empire is the shifting character of this distinction in different fields of knowledge, as well as religion’s perennial centrality in most or all of them.  It is successively explored through cartography (chap. 2), ‘Christocentric travel writing’ by missionaries, clergymen and other fervent Christians (chap. 3), and then ‘humanist travel writing’ (chap. 4), which includes discussion of the famous landscape painters, Thomas and William Daniel, who visited Chainpur in 1790.  A new ‘epistemic regime’ based on scientific classification and enumeration is explored in chapter 5, which is about religion and caste in the modern census, inaugurated in 1871-2.  After looking at studies in folklore, ethnology and religion, partly by focusing on their treatment of a well-known myth from Chainpur (chap. 6), Gottschalk turns finally to the disciplines of history and archaeology (chap. 7).

According to Gottschalk, religion was not only the persistent ‘marker of difference’, but also ‘the primary category of comparison in British understandings of South Asia’ (p. 97).  Moreover, British classifications assumed both that communal (and hence oppositional) religious identities had existed since the rise of Muslim power in India in the twelfth century and that they ‘represented the foundation of the society and the basis of the individual’.  Indigenous, pre-British classifications, by contrast, were not mutually exclusive and ‘religious difference [was] an inconsequential issue’ in many of them (p. 105).  British rule’s ‘historical game-changer’, however, was not the assumption that Hinduism and Islam defined ‘mutually antagonistic communities but – less obviously and more perniciously – [that they defined] mutually exclusive categories of social belonging that bifurcated nearly every societal and cultural dimension of India’ (pp. 182-3).  In changing the game, the modern census was a key instrument, because it classified all Indians by religion, which was the primary category or ‘metacategory’, in turn subdivided into Hindu castes or Muslim sects (primarily Sunni and Shia).  Over time, the Indians’ own understanding of their religion and society was shaped by this classification system, described as ‘scientistic’ because it was believed to comply with superior, truer ‘science’ (p. 35).  The tragic, long-term consequences of this unfolding understanding have been communal violence, the partition of India and Pakistan, and the rise of aggressive, intolerant, politicised religious movements.  One small but telling example, which Gottschalk cites at the end, is a book published in 2005 entitled Religious demography of India, which is actually Hindu chauvinist propaganda made plausible by using census statistics that purportedly ‘validate an effort to read assertions of mutual exclusivity today onto a millennium of history in a manner echoing British assertions’ (p. 337).  By an unhappy coincidence, while I was writing this review, another telling example occurred when The Hindus: an alternative history by Wendy Doniger, Gottschalk’s teacher at Chicago, was withdrawn from sale in India because it was ‘disrespectful of Hinduism’, following a blustering campaign by a Hindu nationalist organisation.

Gottschalk contends that the ‘intersection of religion, science, and empire’ has been ‘relatively under-examined’ in the South Asian literature (p. 10).  More specifically, too, numerous writers on the census have overlooked the prioritisation of religion and the use of ‘Christian-informed, scientistic classification’ (p. 196), and several have instead portrayed ‘caste as the single most important trope for Indian society’, to quote Nicholas Dirks (2001: 49), who stresses the role of the census and other colonial ethnography in bringing this about (p. 201).  Whether Gottschalk’s claims are valid is debatable.  More than others, Gottschalk does emphasis how imperialism facilitated the development of modern disciplines of knowledge, including history of religions, as well as benefiting from them.  All the same, the relationship between religion, science and empire has been discussed fairly extensively, even if other writers have not reached exactly the same conclusions as Gottschalk.  Nonetheless, in one form or another, it has long been recognised that scientistic concepts and classifications of Hinduism and Islam as mutually exclusive religions, and Hindus and Muslims as mutually exclusive social groups, have played an important role in the development of religious communalism in colonial and post-colonial South Asia.  Yet, secondly, that fact in itself does not mean that religion, or a specifically scientistic concept of it, was or became the primary category for understanding India during British rule, as I shall try to show by focusing on the censuses and related literature, which are crucial for Gottschalk’s thesis.

An initial point of some significance, however, is that every village in British India was the subject of a land settlement report, which described its land tenure systems and identified the proprietary landholders, who had to pay the revenue demand to the government.  Settlement was a vital part of colonial administration, for the simple reason that land revenue was such a vital source of government revenue.  In Bengal, which included Bihar and Orissa until 1911, a permanent settlement was made and the revenue-paying landholders were the great landlords or zamindars, from whom smaller landlords and finally the actual cultivators leased plots of land.  In most provinces, including Bengal, settlement reports were mainly confined to land settlement itself and contained little or no ethnographic information, for example, about the religion of villagers.  Moreover, as L. S. S. O’Malley’s Shahabad gazetteer (1906) typically illustrates, the chapter on land revenue included in district gazetteers examined the topic without any reference to religion, which would clearly have been seen as extraneous.  All this matters because settlement reports are important documents about the structure of Indian rural society in the colonial period.  They have previously been used to show that the tenurial hierarchy of landlords, tenants and labourers was formerly more critical than the caste hierarchy, but whether or not that is correct, the settlement reports plainly negate the claim that religion was always a primary category in British understandings of Indian society.

According to Gottschalk: ‘Religion was not the only concern of British Indian censuses, but it was the most central one’ (p. 199).  Thus, along with sex and age, religion was the sole question asked about every individual in every census from 1871-2 to 1941; in addition, arguments about the priority of caste advanced by Dirks and others overlook the fact that ‘the categorical supremacy of religion’ is already given in the taken-for-granted assumption that caste was ‘so quintessentially Hindu’ (p. 201).  Gottschalk misleads here.  First, the 1871-2 census was an unsatisfactory exercise, as officials quickly recognised, and in subsequent colonial censuses, from 1881 to 1941, the questions hardly varied and always included caste, occupation, language, and other topics, as well as sex, age and religion (pp. 342-3).  Secondly, Gottschalk develops his thesis by comparing the censuses of 1871-2, 1901 and 1941.  Most of the data collected in 1941, owing to financial and political problems at the time, were not published.  Among the three chosen for comparison, therefore, two were exceptional; furthermore, if the 1901 census is compared with its predecessors in 1881 and 1891, a different picture emerges.

The 1901 census commissioner was Herbert Risley, imperial India’s leading anthropologist.  Risley, in common with many others in his day, assumed that in ancient times tribesmen from central Asia belonging to a fair-skinned Aryan race invaded or migrated into India from the northwest, and progressively subjugated the dark-skinned, indigenous, Dravidian peoples.  Risley further believed that the hierarchical caste system grew out of this ancient subjugation, so that racial distinction lay at the origin of caste, although other racial groups migrated into India later and the system became more complex over time.  The eventual outcome, though, was a correspondence between the status of predominantly endogamous social groups and their physical features.  This could be proved scientifically by anthropometry, because bodily measurements made among different castes varied significantly; most notoriously, Risley claimed that one cranial measurement – the nasal index – showed that high-caste people tend to have narrow noses, characteristic of the Indo-Aryan race, whereas low-caste people tend to have the broad noses of the Dravidian race.  Risley’s racial theory features in Gottschalk’s book, both when discussing caste and the 1901 census in chapter 5 (pp. 212-5), and racial evolutionism in conjunction with the distinction between ‘primitive’, ‘animist’ tribes and ‘advanced’ Hindu Aryans in chapter 6 (pp. 253-65).  In my opinion, critiques of Risley’s theory and its manifest defects have become repetitively uncritical, but I shall not expatiate on this here and instead concentrate on a few points pertinent to Gottschalk’s general thesis.

Like some previous writers on whom he relies, Gottschalk misunderstands the relationship between Denzil Ibbetson’s occupational or functional theory of caste and Risley’s racial theory (pp. 214, 259-60).  Ibbetson was the 1881 census superintendent for Punjab and in his report, which included a very long chapter on the province’s ‘races, castes and tribes’ (reprinted separately as Panjab castes in 1916), he set out his evolutionist theory of caste, based on the axiom that the division of labour becomes more complex as society develops.  Probably depending mainly on Maine and Herbert Spencer, Ibbetson argued that ‘[t]he whole basis of diversity of caste is diversity of occupation’ (1916: 3).  Indian society differed from other societies at a comparable evolutionary stage, because the Brahman priesthood long ago exalted its religious office, developed as a hereditary group apart from the Kshatriya warrior-kings and asserted its primacy.  Over time, according to Ibbetson, the hereditary principle was imitated by all the lower grades of society, so that occupational groups, instead of evolving into open guilds (as happened in Europe), became closed hereditary castes; other groups with a different origin, such as tribal communities, also turned into castes of the same basic kind.  Athelstane Baines, who is undeservedly ignored by modern scholars, put forward a very similar occupational theory in his 1881 census report on Bombay, which he refined a decade later as the 1891 census commissioner.^[3]

In 1891-2, in two anthropometric and two ethnographic volumes, Risley’s  The tribes and castes of Bengal was published.  His racial theory was set out in this work and an article in 1891, and repeated with little modification in his 1901 census report (co-authored with Edward Gait) and his book, The people of India, first published in 1908, which reproduces the census report’s chapter on ‘caste, tribe and race’.^[4]  Gottschalk implies that Ibbetson criticised Risley’s reductionist racial logic (pp. 259-60), although he obviously could not have considered Risley’s theory in 1881 and he did not write about it later.  In The tribes and castes of Bengal, Risley did not criticise the occupational theory directly, but he did so in the 1901 census report, although his criticisms of Ibbetson – a colleague whom he admired – were indirect and muted.  Nonetheless, Risley made it clear that he did not believe that the origins and development of caste could be explained by the evolution of the division of labour.  For the 1901 census he discarded the occupational classification used in 1891, which he criticised as a ‘patch-work classification’ (1969: 111).  Instead, he wanted castes systematically classified by ‘social precedence’ – i.e., hierarchical rank – and each census superintendent was asked to devise an appropriate set of tests to determine, for their own province, how each of its castes was ranked (ibid.: 111-16).

Risley’s plan ran into difficulties, so that in different provinces the ranking of castes was worked out differently, and in some of them not at all.  All the same, Risley’s focus on social precedence significantly altered the understanding of caste.  Social precedence as conceptualised by Risley was, first and foremost, a property of the Hindu caste system, so that Muslims, though not ignored, were discussed more cursorily (ibid.: 121-4).  Ibbetson and Baines assumed that caste was originally a Hindu institution, of course, but in their occupational theories caste categories often cross-cut religious divisions, because landlords or merchants, say, or priests or artisans, were placed in their respective categories irrespective of whether they were Hindu, Muslim, Sikh or Christian.  Thus religion was not the primary classificatory category, but a secondary one.  To cite one example: Ibbetson first divided the Punjabi castes into ‘landowning and agricultural’, ‘professional and mercantile’, and ‘vagrants, menials and artisans’, and then subdivided the three groups into sections, some of which are defined by religion, such as Biloch-cum-Pathan or Rajput in the first group, and Brahman or Saiyid in the second (1916: 28).  It is surely true that Risley’s emphasis on hierarchical rank does make caste ‘quintessentially Hindu’ in Gottschalk’s sense.  That probably influenced later scholarly writing on caste by officials such as E. A. H. Blunt (1931) and O’Malley (1932; 1934), but Ibbetson and Baines had understood caste otherwise in 1881 and 1891, and the change that took place undermines Gottschalk’s insistence on the perennial priority of religion in British understandings of Indian society.

On the question of religion itself as a category, Gottschalk’s thesis is more problematic than it appears.  The ‘historical game-changer’, as already stated, was the British assumption that Hinduism and Islam defined ‘mutually exclusive categories of social belonging’.  Although this looks true at first glance, the census reports repeatedly showed that the popular religion of the masses was highly eclectic.  Hence in regions such as Punjab – where there were large populations of Hindus and Muslims, as well as Sikhs – mingled religious customs were the norm, as Ibbetson explained.^[5]  Moreover, even if defining Islam was (too easily) deemed to be straightforward, defining Hinduism was always seen as very difficult or even impossible.  Particularly problematic was the distinction between Hinduism and animism, because the fuzzy relationship between them meant that Hinduism might be described as ‘animism more or less transformed by philosophy’ or even ‘magic tempered by metaphysics’, to quote Risley (1969: 233).  Pointed expressions of this sort were commonplace, so that as late as 1931, in his book on caste, Blunt quoted Risley and others to show that is there nowhere ‘a single satisfactory definition of Hinduism, which is not surprising, since it is impossible to define the indefinite’ (1931: 273).

Blunt, like others, was well aware of the problems caused by this vagueness for census operations, and knew that enumerators couldn’t just rely on people’s responses when asked their religion, because a lot of them said they didn’t understand the question or didn’t know the answer.  All this was exasperating for census officials, but the problem could not be wished away, so that if in British eyes Hinduism and Islam defined ‘mutually exclusive social categories’, they did so in great part because Hinduism – in contrast with Islam – could never actually be defined as a clear-cut category.  Whether such an oxymoronic conclusion supports Gottschalk’s thesis is unclear, but it’s certainly not obvious; perhaps he’d have considered the question if his book had included a chapter specifically on religion.  The absence of such a chapter, he admits, ‘may appear odd’, but because ‘religion features so prominently in British and Indian representations that are not the work of religion specialists … the salience of religion as an analytic category in the overall British episteme [is] all the more apparent’ (p. 332, italics original).  Maybe, but I’m not convinced and think that a fuller, critical discussion of how British scholarly officials understood Indian religions could have been very illuminating.

The question of animism is briefly discussed in relation to Indian ethnology and its premise of racial evolution, primarily from ‘primitive’ animist tribe to ‘advanced’ Aryan Hindu caste in this context.   At this one place in his book, Gottschalk says that religion, in the discipline of ethnology, did become ‘a secondary analytic metacategory’ (p. 262).  His evidence derives from The tribes and castes of Bengal, William Crooke’s The tribes and castes of the North-Western Provinces and Oudh (1896), and several similar works – all vaguely characterised as ‘racialist’ (p. 258) – and Gottschalk appears to argue that racial evolutionism displaced religion in ethnology owing to British ‘national concerns’ (p. 262) for this discipline’s subject-matter.  This argument is unclear, but in any case the simplest explanation for why the primary category in the tribes and castes handbooks is not religion is that these works are self-evidently organised by the categories of tribe and caste, and each entry often (if not always) gives the group’s religious affiliation(s).  This is done with particular clarity in Crooke’s handbook.  Thus, to take just the first two main groups mentioned in it, the Agariya tribe ‘call themselves Hindus, but worship none of the regular Hindu deities’, and in the Agarwala trading caste, ‘most … are Vaishnavas; some are Jainas or Saraogis.  … A small minority are Saivas or Saktas’ (1896, 1: 8, 23).  In these texts, religion (like occupation, marriage system, etc.) is obviously a secondary category and no complicated explanation of why is needed.

The colonial census, despite its documentation of eclectic, popular religion, certainly played a key role in the development of mutually exclusive religious categories, primarily Hindu and Muslim.  Unlike its pre-modern antecedents, the modern decennial census depended especially on confidence in statistics and sound disciplinary bodies of knowledge, which made it ‘scientific’ or scientistic; in British India, indeed, the census best exemplified the drive to classify and count at the heart of Victorian positivism and science.  Data on religion, and caste and tribe, were most amply collected in 1881, 1891 and 1901; from 1911, when the commissioner was Gait (Risley’s colleague in 1901), the amount of data was reduced and, especially from 1921, other topics, notably industrial and economic ones, were given more attention (p. 215).  In his report, Gait was particularly sceptical about the censuses’ definition of religion, saying they made sense for Christians and Muslims, but none for others, including Hindus, so that the boundary between Hindus and Muslims was indefinite; Baines later congratulated Gait on showing so lucidly that the appellation ‘Hindu’ implied no uniformity.^[6]  No doubt, Gait and Baines partly wanted to make their scepticism clear precisely because separate Muslim electorates, which were introduced by the Morley-Minto reforms in 1909, had started to make population figures for religious communities politically sensitive and critical.  Gottschalk refers to similar scepticism about ‘the artificiality of the [religious] identities they required some groups to declare’ (p. 218), which were expressed by British officials before the 1941 census.  He rightly comments that the politicization of religious categories meant that officialdom had lost control of the categories by the 1940s, but one might equally argue that the scepticism is evidence against Gottschalk’s arguments that ‘mutually exclusive social categories’ became axiomatic for the British.  Moreover, as I’ve shown, in the three censuses of 1881, 1891 and 1901, as well as in related literature of the period, the relationship between religion and caste, and their relative priority as classificatory categories, was variable.  In the final analysis, I am unclear why Gottschalk doesn’t recognise this; if he had been less insistent on the priority of religion, his overall argument, in my judgement, would have been stronger, not weaker.

Religion, science, and empire is, however, a valuable book that contains a lot of interesting material and argument, and its focus on one small place – Chainpur and its neighbourhood – gives it a special originality.  It is a book that is well worth arguing with and it represents a major addition to the literature on the colonial understanding of India and, potentially, to India’s importance for the history of British social anthropology.

REFERENCES

Blunt, E. A. H. 1931. The caste system of northern India. London: Oxford University Press.

Burrow, J. W. 1966. Evolution and society: a study in Victorian social theory. Cambridge: Cambridge University Press.

Census of India 1881, Bombay, vol. 1, Text by J. A. Baines. Bombay, 1882.

Census of India 1881, Punjab, vol. 1, Text by D. C. J. Ibbetson. Calcutta, 1883.

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

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

Census of India 1911, vol. 1, India, part 1, Report by E. A. Gait. Calcutta, 1913.

Cohn, Bernard S. 1987. An anthropologist among the historians and other essays. Delhi: Oxford University Press.

––––. 1996. Colonialism and its forms of knowledge: the British in India. Princeton: Princeton University Press.

Crooke, William. 1896. The tribes and castes of the North-Western Provinces and Oudh. Calcutta: Superintendent of Government Printing.

Dirks, Nicholas B. 2001. Castes of mind: colonialism and the making of modern India. Princeton: Princeton University Press.

Gait, E. A. 1914. The Indian census of 1911. Journal of the Royal Society of Arts 62 (June 5): 627-39.

Gottschalk, Peter. 2001. Beyond Hindu and Muslim: multiple identity in narratives from village India. New York: Oxford University Press.

Hunt, Roland, and John Harrison. 1980. The district officer in India, 1930-1947. London: Scolar Press.

Ibbetson, Denzil. 1916. Panjab castes. Lahore: Superintendent of Government Printing, Punjab.

Metcalf, Thomas R. 1995. Ideologies of the Raj. Cambridge: Cambridge University Press.

Nesfield, J. C. 1885. Brief view of the caste system of North-Western Provinces and Oudh. Allahabad: Government Press.

O’Malley, L. S. S. 1906. Shahabad (Bengal District Gazetteers). Calcutta: Bengal Secretariat Book Depot.

––––. 1932. Indian caste customs. Cambridge: Cambridge University Press.

––––. 1934. India’s social heritage. Oxford: Clarendon Press.

Risley, H. H. 1891. The study of ethnology in India. Journal of the Anthropological Institute, 20: 235-63.

––––. 1891. The tribes and castes of Bengal: anthropometric data. Calcutta: Bengal Secretariat Press.

––––. 1892. The tribes and castes of Bengal: ethnographic glossary. Calcutta: Bengal Secretariat Press.

––––. 1969 (1915). The people of India, 2nd ed. Delhi: Oriental Books.

Stocking, George W. 1987. Victorian anthropology. New York: Free Press.

––––. 1995. After Tylor: British social anthropology 1881-1951. Madison: University of Wisconsin Press.

1. ^[1]‘Notes on the history of the study of Indian society and culture’ (1968) and ‘The census, social structure and objectification in South Asia’ (1970) in Cohn (1987: chaps. 7, 10).  Cohn (1996), Dirks (2001) and Metcalf (1995) are just three of the more important contributions to the literature on colonial knowledge of India.↩
2. ^[2]Ray retired in 1947.  Reflecting on his early days in the 1970s, Ray was justifiably critical of the inadequate and antiquated training he had received, which meant that ‘the cultural and historical background of the people among whom we were to work and serve was almost entirely neglected’ (Hunt and Harrison 1980: 34). His ignorance no doubt partially explained the negative comments in his journal.↩
3. ^[3]Census 1881, Punjab, vol. 1, Text, chap. 6 [reprinted as Ibbetson 1916]; Census 1881, Bombay, vol. 1, Text, chap. 8; Census 1891, General Report, chap. 5.  Another version of the occupational theory of caste by Nesfield (1885) was also influential at the time.↩
4. ^[4]Risley was promoted to become Secretary in the Home department of the government of India before completing the general, all-India report and was replaced by Gait, who was also the 1901 census superintendent for Bengal.  Risley completed the long chapter 11 on ‘caste, tribe and race’, as well as parts of four other chapters (including the section on ‘animism’ in chapter 8 on religion), and Gait wrote the rest, except the chapter on language by George Grierson; see Census 1901, vol. 1, India, part 1, Report.  Chapters 1, 2, 4 (part), 5 (part) and 6 of The people of India are the census report’s chapter 11 and the passage on ‘animism’ rearranged, but not rewritten; nearly all the rest, including chapters 3 and 7, were written for the book.  It was first published in 1908 and reissued as a second, ‘memorial’ edition in 1915, edited with an introduction by William Crooke; my quotations from the census report are actually cited from the more easily accessible second edition of the book.↩
5. ^[5]Census 1881, Punjab, vol. 1, Text, pp. 100-101.↩
6. ^[6]Census 1911, vol. 1, India, part 1, Report, pp. 113-18; Baines’s comment as chairman of a talk on the census given by Gait (1914: 634).↩