Religion, Race, Language and the Anglo-Indians:
Eurasians in the Census of British India

Dr. Peter Friedlander, La Trobe University - Melbourne, Australia.

This paper looks at the creation of the categories of race and religion in the census of India in the 19th and early 20th centuries and how they impacted on the creation of a sense of self identity amongst communities in India including the Eurasians or Anglo-Indians. It draws upon work done at La Trobe University during 2001 on the digitization of the 1871-2 and 1891 Indian Census reports. I explore what the wealth of details in the census reports reveals about 19th and early 20th century perceptions of race, religion and the Anglo-Indians in India.


There is only one reference to the term "AngloIndian" in the Report on the 1891 Census of India. It occurs in the discussion of the range of "Leather Workers and the Lower Village Menials" castes:

The fourth item, Sakilia, is the caste known as Chucklers to AngloIndians in Madras.

This is a reference to British residents of India not to people who we would now call "Anglo-Indians". In the 19th census reports the term used to refer to "Anglo-Indians" was "Eurasians". For instance there are 21 references to Eurasians in the report on the 1891 census of India by J. A. Baines. By contrast the use of the term Anglo-Indian in place of Eurasian seems to have developed in the early 20th century. This is evident when in the 1919 Montague-Chelmsford Reforms the principle of communal representation was extended to various groups including Indian Christians within two categories, Europeans and Anglo-Indians.

Religion and Race

Cohn points that in the late 18th and early 19th century the notion of "statistical information" was understood as relating to much more than just numerical data and rather it implied:

collection of information thought necessary and useful to the state. Since the time of William Camden (1551-1623) information had been collected and published about current conditions, history, and antiquities of various localities in Great Britain. Central to this endeavor was the location and description of old buildings, ruins, sites of ancient settlements, collection of family histories and genealogies, as well as the description of local customs and laws, thought to be antique or unusual.

Due to this the census materials from the latter half of the 19th century contain a wide range of materials which nowadays might not form part of the kind of data an organization like the ABS might collect. In addition Cohn has pointed out that the census of India was part of an attempt by the British to enumerate their possessions.

As part of the imperial settlement project after the repression of the Indian uprising of 1857-1858, the Government of India carried out a series of censuses which they hoped would provide a cross-sectional picture of the "progress" of their rule. By 1881 they had worked out a set of practices that enabled them not just to list the names of what they hoped would be every person in India but also to collect basic information about age, occupation, caste, religion, literacy, place of birth, and current residence. Upwards of 500,000 people, most of whom were volunteers, were engaged in carrying out the census. The published census reports not only summarized the statistical information thus compiled but also included extensive narratives about the caste system, the religions of India, fertility and morbidity, domestic organization, and the economic structure of India. The census represents a model of the Victorian encyclopedic quest for total knowledge.( Cohn 1996: 8)

Cohn also proposed that the census needs to be seen in the light of the way in which it reified the categories it sought to enumerate.

It is my hypothesis that what was entailed in the construction of the census operations was the creation of social categories by which India was ordered for administrative purposes. The British assumed that the census reflected the basic sociological facts of India. This it did, but through the enumerative modality, the project also objectified social, cultural, and linguistic differences among the peoples of India. The panoptical view that the British were constructing led to the reification of India as polity in which conflict, from the point of view of the rulers, could only be controlled by the strong hand of the British. (Cohn 1996: 8)

Race and religion were important factors which influenced the way in which data came to categorized. One of these was a concern to classify the religions of India. In the report on the 1871 Census written by A. O. Hume Secretary to the Government of India in Simla on the 30th of September 1871 he noted:

The general division of the population adopted in the forms originally proposed was that of "Hindoo," "Mohomedan," and "others." It has been pointed out that in British Burmah Buddhists should be shown separately, and the Government of Madras supported the local board of Revenue to admit a class of "Europeans and Eurasians"

Eventually the simple suggestion of three categories was abandoned. In the end in the 1871-72 census the following categories for religious affiliations were lain down for how to draw up "Table V. A - Statement of Religions".

Hindu- Sections in as great detail as practicable including: 3 Shankar, or Smarta or Shaiva, 4 Madhava, or Vir Vaishava or Vaishnava, 5 Ramanuj, 6 Lingayat, 7 Swami Narayan, 8 Wallabachariya, 9 Kabir Panthis, 10 All those who simply worship some god or goddess without knowing anything of theology, 11 Ascetics, Religious Mendicants, etc., Mohomedans, 13 Sunni, 14 Shiah, Christians of all Types, 16 Armenian, 17 Baptist, 18 Episcopalian (COE), 19 Greek, 20 Presbyterian, 21 Roman Catholic, 22 Wesleyan, 23 Native Christians, 25 Jains, 27 Jews proper, 28 Beni Israel, 33 Sikhs proper, 34 Nanak Panthis, 36 Brahmos, 37 All others (Natrajan 1972: 29)

It can be seen that the notions of religion, race and caste were intertwined in the eventual system evolved for tabulating the results of the census data.

One reason for this was there was an interaction going on here between the process of the census and the tabulation of the results. In particular because the census was carried out by getting individuals to return the name of their religion and these self-identifications were then fitted into religious types as worked out by the census officials. In the process many minor religious groups were probably excluded from tabulation. However, it is also very likely that individuals would have tailored their responses to fit what they saw as likely to reflect well on themselves, or to be to their advantage in one way or another.

The most striking thing about this process is not just the results obtained, but that it made people aware of the need to identify themselves as belonging to a particular group. Speaking about the emphasis on religion in the census Jones suggests.

Two conclusions are clear from an examination of the census reports. First, from the very beginning religion was a fundamental category for organizing data and for attempting to understand (Indians). Questions related to religions expanded steadily and cut across much of the other subjects discussed in the census reports, so that initial concern for religion and religious groups expanded with time. Secondly, the census reports provided a new conceptualization of religion as a community, an aggregate of individuals united by a formal definition and given characteristics based on qualified data. Religions became communities mapped, counted, and above all compared with other religious communities.(Jones 1981: 84)

What is more European notions of religion do not seem to have allowed for the idea that a person could be more than one thing at one time. For instance there were many aspects of religious practice in 19th century India that crossed the kinds of categories in the 1871 census schedule above. For instance, many Sikhs and Jains also worshipped what might be identified as Vaishnava gods. Indeed it can be argued that it was the process of being asked to identify your religion that began to make people regard themselves as members of a particular community in a new way which excluded the possibility of practicing aspects of multiple traditions at the same time.

In addition Jones points out that from 1891 onwards various groups tried to influence the census according to their own ends.

With the Punjab Census of 1891, a new development emerged as the subjects of the census sought to reshape it according to their own needs. Militant members of the Arya Samaj, a Hindu religions movement, had come to reject the name 'Hindu' feeling that it was a derogative label given them by the followers of Islam. Instead they wished to be known as 'Aryas', a name which to them symbolized a revived and purified Vedic religion. In the years 1890 and 1891 Arya Samaj newspapers instructed their followers to write 'Arya' in place of 'Hindu' in the religion column of the 1891 census questionnaires. At the same time they pressured the census officials to accept and record this change of religion. In this way they hoped to use the census to establish and legitimize their own vision of their own religion.(Jones 1981: 87)

Compare this with the passive sense of the people as the subject of inquiry reflected in the Memorandum on the 1871-72 census where the report writer regrets that race and religion are not enumerated with regard to Christianity and Native Christians, Europeans and Eurasians.

In Berar about 900 Christians are enumerated, but the Natives are not separated from Europeans or Eurasians. In British Burma the numerous Karen converts are not specified in the Census Report, and only 2,300 Native Christians have been entered; there are, however, 52,000 Christians in the province, and in the Administration Export the total number of Native Christians is stated to be 34,310. The Oude report does not distinguish between Native Christians and Europeans or Eurasians. (1871 Memoradum: 27)

It should be remembered in this regard that there were many reservations about the 1871-72 census as it was in fact a compilation of data from a number of provincial census reports made during the previous decade or more to a wide variety of standards.

By the 1881 census the manner of classification had been reorganized along these lines for the preparation of "Form III Distribution of Peoples according to Religion"

Hindu, Mohamedan, Aboriginal, Buddhist, Christian, Sikh, Jain, Satnami, Kabir Panthi, Nat Worship, Parsi, Jew, Brahmo, Kumbhpatia, Unspecified and Others (Natarajan 1972: 37)

Mean while the details for Christian denominations had grown to include in "Form IIIa Christians"

American, Armenian, COE, CO Scotland, Episcopalian, Greek, Lutheran, Others Protestant: - including: Baptist, Congregationalist, Methodist, Wesleyan Independent, Methodist Wesleyan and others, Roman Catholic, Syrians, Others and Unspecified. (Natarajan 1972: 37)

In the details of Mohomedans there were now included:

Sunnis, Shiahs, Farazis, Wahabis and Others without details. (Natarajan 1972: 37)

The 1891 census and the 1901 (Natarajan 1972: 81) and 1911 (Natarajan 1972: 106) censuses all adopted a similar schema to that taken up in 1891. In preparation for the 1891 census on the 28 August 1889 J. A. Baines was appointed Census Commissioner by the Home Department. In December of 1889 he held a conference in Agra of officers who had taken part in the 1881 census "for the discussion of the administrative details of the census of 1891".( Natarajan 1972: 53) This led to the drawing up of a schema for "Table VI Religions" which included the following categories.

Hindu, Musalman, Christian, Buddhist, Jain, Sikh, Parsi, Jew, Animistic.[add columns as required] (Natarajan 1972: 53)

Regarding how the census forms were to be filled out by the enumerators it is notable that race and religion are compounded in the way that Eurasians are to be enumerated.

Col. 4.* Enter the caste if Hindus or Jains, and the tribes of those who have not castes, and the races of Christians, Buddhists, &c. : As Brahmin, Rájput, Bania, Kúnbi, for Hindus. Pathán, Moghal, &c., for Musalmáns; Eurasian or Native Christian for Christians; do not cater vague terms, such as Hindustáni, Márwádi, Panjábi, &c. (Baines 1891: 187)

In the final report on the census the categories had grown very extensively and in particular the subdivisions of Christians had become very complicated indeed and were constantly compounded by references to race. Consider for instance these comments of Baines on " Minor forms of religion" from his report .

In every enumeration of forms of faith room has to be found for a residuum of special classes of belief which cannot be fitted into any of the corners of those more generally prevalent. In India this residuum is small but varied. It is not, however, complete, for we find samples of Unitarians, Theists, and Agnostics included with Christians, probably because the enumerator found them mostly in the schedules of Europeans. Similarly, there were a few instances, altered on scrutiny, of Buddhists amongst European and Eurasian Christians. Some of these related probably to the professors of the new Theosophy; others may be set down, perhaps, to the example of a high official, who pronounced for Buddhism, on the ground that there was less to be said against it, he thought, than against any other form of religion. (Baines 1891: 170)

What I find surprising here is that the instances of Europeans and Eurasians returning themselves as Buddhists were 'altered on scrutiny'. Apparently it was considered that Europeans and Eurasians were inherently Christians.

It is also apparent that there was a continuation of the attitude first expressed by the Madras Board of Revenue in 1871 to stress the difference between the Christianity practiced by Indians, Europeans and Eurasians. In this respect Baines writes in his report concerning Christians.

In the returns of the Christian population, the distribution by race is given, in addition to that by territory and denomination, and it will serve as a sort of introduction to the rest. From Supplementary Table A, at page 496 of the first volume of the returns, the community will be seen to be composed of 89.1 per cent. of natives of India, including a few negroes ; 7.4 per cent. of Europeans, including Americans and those from Australasia, and 3.5 per cent. of a mixture of these two, known by the general title of Eurasian. (Baines 1891: 177)

Regarding the European Christians Baines says:

The foreign element is composed of, first, the military, next the civil employés of the State, thirdly, those engaged on railways or mines, and then the professional and commercial classes chiefly congregated at the seaports and provincial capitals. There was also, at the time of the census a considerable contingent of seafaring people on board vessels in harbour. It is not practicable to separate these groups in the general return. Speaking roughly, the military section amounts to a little over one half. At the time of the census the strength of the European troops was about 67,800, with 2,550 officers, 3,120 women, 5,900 children, and approximately, 800 people engaged on the staff and in various military avocations in India. To these can be added about 1,200 wives and children of officers and those on the staff, &c., with 2,530 European officers attached to the native army, and about 1,100 persons representing their families, making in all about 85,000 souls. As to the other classes, we have still less of a foundation to build on. A recent estimate gives the civil employés and their families 10,500, and the railway servants 6.100, leaving about 66,400 to be distributed over the other groups, but this is admittedly no more than an approximation. Then, again, the distinctions between the three races is very shadowy, and there is a tendency for Eurasians to enter the European group, and for native Christians to be returned as Eurasians. As the return stands, however, the influence of the army is visible in the number of Europeans found in the Panjab, which heads the list. (Baines 1891: 177-178)

Whilst in regards to the Eurasians he comments:

Assuming the Eurasian return to be correct, Madras heads the list with 26,600 out of the total of 79,790. Bengal, which comes next, returns only 15,000. Bombay shows 8,500, amongst whom there are no doubt a good many Goanese or others with Portuguese patronymics. The NorthWest Provinces have always been remarkable for the number of their Eurasians, and Burma, which, in spite of, its small population has nearly the same number, owes its pre-eminence to its peculiar domestic institutions which foster the supply of children of mixed race. Haidrabad and Bangalore follow Madras in the frequency of the entries of this class. It would be interesting to compare the returns of the two last enumerations, so as to see if this community is numerically on the rise or not, but owing partly to defects in the schedules filled up, partly to the number in 1881 not filled up at all as regards race, the census operations give little help in this direction, and the matter can only be satisfactorily dealt with through special investigation. Since the community is mostly, congregated in the chief towns such an inquiry is not difficult, and can be made to include within its scope valuable details, such as the number of children per family, the age of parents at the time of marriage, and so on, which are impossible at a census. (Baines 1891: 178)

Regarding the subdivision of the Christians the 1891 census is very detailed, almost obsessive I might say, in the detail given.

TotalDistribution by Race. Denomination.Number.EuropeanEurasian.Indian. A. - Church of England, with Churches of India, Ireland, America, Anglican, and Episcopalian churches. 340,613103,14529,922 201,546B.- Church of Scotland, Presbyterian, United Presbyterian, Reformed Presbyterian, American or Irish Presbyterians, Irish Presbyterian Mission. 46,35110,5812,440 33,329C. -Baptist202,7462,907 2,352197,487D.- Wesleyan, Wesleyan Methodist, Methodist, Primitive Methodist, Episcopalian Methodist, Bible Christian. 32,1235,3622,349 24,412E.- Congregationalist, London Mission, Independent, Calvinist, Welsh Calvinist. 48,0361,285742 46,009F. - Nonconformist, Dissenter, Puritan557230180147G. - Plymouth Brethren, Open Brethren, Swedenborgian, New Jerusalem, Catholic Apostolic, Quaker, Friend, Salvationist, Anabaptist. 1,44517838 1,229H.-Lutheran, German Mission, Swedish Church, Reformed Dutch, Zwinglian, Moravian, German Church, Evangelical, Evangelist Church, Evangelical Union, Reformed Church. 69,4051,285195 67,925J.- Protestant14,2133,9384,0526,223 Total, Reformed Church755489 128,91242,270 584,307K. - Church of Rome1,315,263 35,64536,0891,243,529 L. - Syrian Church (Jacobite Section)200,467 153200,449 M. - Greek, Abyssinian, and Armenian1,258 726275257 Total, Older Churches1,5 16,998 36,38636,367 1,444,235Unsectarian and Unreturned11,903 2,1021,1538,048 Grand Total2,284,380 168,00079,790 2,036,690(Baines 1891: 179)

In regard then to the census during the period from 1871 to 1921 I would argue that it shows a consistent concern with categorising people on the basis of religion and race. Indeed as late as the 1921 census there is a "Table VI Religion" which included the following categories:

Indo-Aryan:- Hindu (Brahmanic), Kabir Panthi, Satnami, Hindu (Arya), Hindu (Brahmo), Sikh, Jain, Digambar, Swetambar, Swetambar Terapanthi, Sthanakbasi, Jain (Unspecified), Buddhist, Zoraastrian (Parsi),
Semetic- Musulman, Sunni, Shia, Christian, Jew
Primitive -Animist
Others- Indefinite Beliefs only (Natarajan 1972: 130-132.)

The 1931 Census marked a major change to collection of economic data and the collection of detailed data on religion was given up and Table XVI became simply a listing of broad categories.

Hindu - Brahmin, Arya, Brahmo, Hindu others, Sikhs, Jains, Buddhists, Zorastrians, Muslims, Christians, Jews, Tribal Religions, Minor Religions and Religion not returned.(Natarajan 1972:184)

Later census data also follows this pattern but it is notable that the release of data on religion tends to come out late in the cycle of results of census surveys and is not given prominence in contemporary Indian Census data.

However, a word of caution is necessary. The processes by which the categories were created and the errors in their compilation are, I suggest, as revealing as the eventual data. An example of this in the 1921 Madras Census was noted by J. Chartres Molony, the Provincial Superindentent of Census for the Madras Presidency in 1911.

Clever as the Madras clerks were, they could at times make quaint mistakes. I noted in the table of " religion " a large number of "Armenians" living together in the Kistna Delta. In Madras City there is an Armenian Cathedral, and here and there throughout the Presidency one may meet a rare Armenian; but a colony of Armenian farmers seemed to me incredible. I had the slips which gave the information traced out, and I found out that "Armenian" was a clerk's rendering of the faith of the Salvation Army. Here the clerk was at fault; but the perplexity of an Indian clerk on being called upon to assign " Truthists " to their proper place in the scheme of Christian sects is understandable and excusable. Roman Catholic Missions have always interfered less with caste than have the Protestant, and many Indian Roman Catholics mixed up difference of Christian sect with difference of Hindu caste.

Such stories leave me with lingering doubts about the accuracy of some of the census data. Molony also tells an interesting story in this respect about the interpretation of the term "occupation".

We sought information as to religion, caste, infirmity, civil condition, occupation; and the information recorded was sometimes odd. Occupation was defined as the source from which a livelihood is drawn. With an excess of candour one gentleman returned " Honorary Magistrate " as his occupation. One would scarcely think that the occupations of " vagrant, prostitute, procurer, receiver of stolen goods, cattle-poisoner " would be openly avowed; yet the figures show that in 1911 there were in Madras nearly 162,000 " actual workers " in these arts. A lady entered in the "occupation" column the name of a gentleman to whom she certainly was not united in the bonds of holy matrimony; a village householder described himself as "married", and gave the name of his wife; a reflective enumerator added the marginal comment " not really " to the entry. (Molony 1926: 101)

What is more the 1891 report lists a number of other similar cases which suggest this was a reservation which was held by census officials at most census reports. Perhaps more to the point it also indicates the degree to which the data was "amended" in order to make the statements of those enumerated fit into the categories of the census.


Eurasians are referred as an English speaking community in the 1891 report.

We have now to consider the return of the languages of Europe and the west. Of these there is a great variety, though none but English is strongly represented. This last, which, for census purposes, is held to include Scotch though not Gaelic, is returned by 100,000 or so born in the United Kingdom, by scattered denizens of Australia, the United States, and Canada, &c., as well as by the increasing class of Europeans of British descent born and domiciled in India, and the Eurasians. (Baines 1891: 154)

It is striking in this passage that there is a distinction drawn between " Europeans of British descent born and domiciled in India" and Eurasians.


The issue of race has already been mentioned in regard to religious categories mentioned above. However, it also appears in its own right in census reports.

There is a reference to Eurasians in the 1871 report in relation to race.

There are 108,000 of mixed race, such as Eurasians and Indo-Portuguese. Of the 20,000 who are resident in Bengal, many are descended from the Portuguese, whose head-quarters were in Dacca and Chittagong. In the minor provinces very few have been returned, they having probably preferred to enrol themselves as Europeans. Of the 26,000 in the Madras Presidency, about half are found in the Madras and Malabar districts. Bombay contains about 48,000, three-fourths of whom are in the island of Bombay or the neighbouring district of Tanna; the number of Eurasians in the Presidency is not quite 3,700, while there are 30,000 Indo-Portuguese, and 14,000 who are entered as "others," without any description of the race to which they belong. (1871 Memorandum: 28-29)

This concern with the typology of race, classifying races and communities, was a major project of the British in India in the 19th century and not surprisingly this comes out in the census reports. By the 1891 census the subdivisions of race have become extremely complicated indeed. Division F. of the races deals with " Races and Indefinite Tribes " as follows.

L. Musalmán Foreign Races.
LI. Himalayan Mongoloids.
LII. Burmese and Chinese Mongoloids.
LIII. Western Asiatics.
LIV. Burmese, &c. Mixed Paces.
LV. Indefinite Indian Castes.
LVI. Europeans, &c.
LVII. Eurasians.
LVIII. Native Christians.
LIX. Goanese and Portuguese.
LX. Africans. (Baines 1891: 188)

The actual figures given for 1891 are: Eurasians 81,044 and Europeans 166,428 (Baines 1891: 208)


I have, I hope, shown that the Early Indian Census reports contain an interesting range of materials about the Anglo-Indian/Eurasian community in India. I would argue that the materials I have discussed here show the following points:

In conclusion then the broad picture of what the census of India says about who were the Anglo-Indians is not a surprise. Broadly speaking in the census of India to be an Anglo-Indian/Eurasian a person had to be an English speaking Christian of mixed racial background. What is surprising is the degree to which it is apparent that the census officials "corrected" data by changing details and glossing over grey areas in order to make the normative picture they wanted to see appear in the census reports.

Finally, I am sure that further study of the census reports for 1881, 1901, etc. and the provincial census reports for each census would reveal much more data on the position of the Anglo-Indian/Eurasian in India in the late 19th century and hopefully cast light on the question of who were the Anglo-Indians.